1In studies on the population history of Europe during the Upper Palaeolithic (UP, 45,000-10,000 BP), particular emphasis is placed on the initial UP phase (ca. 45,000-40,000 BP) in order to address the following questions: how and when the first anatomically modern human groups arrived in Europe (Benazzi et al., 2011; Higham et al., 2011), the timing of their expansion (Hublin et al., 2020) and the replacement of declining Neandertal populations and their partial absorption by modern humans (Slon et al., 2018; Villanea and Schraiber, 2019). Population dynamics during the UP, when only anatomically modern humans were present on the continent, are comparatively less studied, and little is known about the interactions of European modern human groups from 40,000 BP to the end of the Last Glacial Maximum (LGM, ca. 26,500-20,000 BP).
2Western Eurasian pre-LGM modern human remains are relatively abundant, mainly due to the large number of Mid Upper Palaeolithic (MUP, ca. 33,000-24,000 BP) primary burials, most being associated with the Gravettian technocomplex (Henry-Gambier, 2008a). The vast majority of pre-LGM specimens are from adults and adolescents (Henry-Gambier, 2008a; 2008b). Adult skeletons are relatively well documented (Trinkaus, 2007; Holt and Formicola, 2008; Villotte et al., 2017) and characterized by an overall "modern" (derived) human morphology, but "archaic" (plesiomorphic) traits and/or Neandertal features are frequently found in pre-LGM individuals (Trinkaus, 2007). No clear chronogeographical pattern has been detected so far for the distribution of these traits, and the morphology of Early Upper Palaeolithic (EUP, ca. 45,000-33,000 BP) and MUP specimens appears homogenous overall (Mounier et al., 2020), which is consistent with ancient DNA studies that suggested genetic continuity between at least some EUP and MUP individuals (Fu et al., 2016; Posth et al., 2016; Sikora et al., 2017).
3Unlike those of adults, well-preserved skeletal remains of pre-LGM children are rare (Zilhão and Trinkaus, 2002) and little is known about their morphology, growth and developmental patterns. One of the best preserved immature skeletons, the 4-5 year-old MUP individual Lagar Velho 1 from Portugal (Zilhão and Trinkaus, 2002), displays a pattern of maturation of its deciduous and permanent dentition and endostructural tooth organization that is absent in extant populations and documented only among Neandertals (Bayle et al., 2010). The immature mandible of El Castillo 2 from Spain (Garralda et al., 2019), belonging to another 4-5 year-old MUP child, shows a modern morphology overall (including its dental maturation pattern) combined with robust characteristics of the symphysis. Other well-preserved immature facial remains with mixed dentition are extremely rare (e.g. Sunghir 3, Kostënki 3 and 4) in the pre-LGM European fossil record (Trinkaus et al., 2014) and no data on their endostructural tooth structure have yet been published.
4Here we examine the mandible fragment of an immature MUP individual from the Gargas cave (Aventignan, France, figure S1), found in 2011 during excavations led by two of us (see SI-1 for details). The Gargas mandible (hereafter GPA-646) was discovered associated with an accumulation of faunal remains, lithic elements, used pebbles and colouring materials, and was sealed in indurated sediment, all of which corresponds to a palimpsest of repeated domestic occupations that took place during the Middle Gravettian and are dated to between 33,000 and 28,300 cal BP (Foucher et al., 2011; 2012; 2019). GPA-646 itself is directly dated to ca. 29,000 cal BP (Foucher et al., 2019; see table S1). Shortly after its discovery, a preliminary description of GPA-646 was undertaken whilst a large part of the mandible was covered by indurated sediment and concretions (Foucher et al., 2012). From macroscopic observations of directly observable aspects, this preliminary description indicates that it is a relatively complete mandible (the right ramus being missing) of a young child between two and five years of age, with a possible abnormal swelling on the lateral side (Foucher et al., 2012). In the present study, we used both macroscopic observations of the mandible (after thorough cleaning of the calcitic coating partially covering it) and imaging techniques applied to an X-ray micro-computed tomography (microCT) record of GPA-646, to characterize the external and internal morphology of the mandible and its erupted and unerupted teeth. Our aims were:
5i) to refine the age-at-death of this child and characterize its pattern of dental maturation;
6ii) to comparatively assess the mandibular and dental morphology and dimensions;
7iii) to quantify internal tissue proportions and enamel thickness of the deciduous first molar.
8In 2012, the two mandibular fragments were imaged at high resolution using the X8050-16 Viscom AG X-ray microcomputed tomography (microCT) equipment at the University of Poitiers. The scans were made using the following parameters: for the larger fragment GPA-646a, 150 kV voltage, 0.4 mA, 32 integrations per projection and a projection every 0.2°; for the smaller fragment GPA-646b, 90 kV voltage, 0.35 mA, 31.5 integrations per projection and a projection every 0.3°. Because of the presence of indurated sediment and a flat triangular pebble (ca. 8 cm for each side, and ca. 4.5 cm of thickness) firmly attached by a thin layer of concretion around the larger fragment GPA-646a (figures 1, S2), the resolution of its microCT record was lower than expected. The final volumes were reconstructed using DigiCT v.2.4.2 (DIGISENS) with an isotropic voxel size of 82.7 μm for GPA-646a and of 24.9 μm for GPA-646b. Subsequent 2D and 3D virtual imaging and morphometrics were performed at the IC2MP of the University of Poitiers and at the PACEA imaging laboratory of the University of Bordeaux. Around 50% of the lateral surface of the main fragment GPA-646a was originally covered by concretion. In 2015, this calcitic coating was nearly completely removed using gentle mechanical and chemical techniques (figure 1).
Figure 1
The Gargas mandible GPA-646. a) Left lateral view of GPA-646a before cleaning of the concretion and b) After cleaning. c) Supero-anterior view of GPA-646a and GPA-646b joined together. d) Anterior view of GPA-646a after cleaning |
La mandibule GPA-646 de Gargas. a) Vues latérales gauches de GPA-646a avant le nettoyage de la concrétion et b) Après le nettoyage. c) Vue supéro-antérieure de GPA-646a et GPA-646b réunis. d) Vue antérieure de GPA-646a après le nettoyage
9Macroscopic observations were made after removal of the thin layer of concretion on the main fragment. The linear measurements and angles followed the Martin (M-#) system (Bräuer, 1988), completed by measurements defined by Trinkaus (Trinkaus, 2002). They were taken either directly on the fragments or on the 3D model using Avizo, Fiji and TIVMI (Treatment and Increased Vision for Medical Imaging) software (see SI-4). The measurements on 3D models were repeated three times at least one week apart. As the differences between repeated measurements were negligible, the values obtained for each measurement were averaged and used in this study (table S5).
10In order to virtually extract bone, teeth and dental tissues, semi-automatic threshold-based segmentation with manual corrections was conducted on the microCT record following the half-maximum height method (HMH, Spoor et al., 1993) and by taking repeated measurements on different slices of the virtual stack using Avizo 7.1 (Visualization Sciences Group Inc.). Because of the presence of indurated sediment and concretions around the larger fragment GPA-646a, the resolution of its microCT record was lower than expected, and the segmentation of the bone and teeth was technically difficult and time-consuming. While we were able to virtually segment and extract each tooth, an unambiguous distinction between enamel and dentine was not possible for this fragment (figures S2-S3), precluding any investigation of the dental tissue proportions for the GPA-646a fragment. On the other hand, no particular technical difficulties were encountered in segmenting the smaller fossil fragment GPA-646b, which is characterized by a rather distinct endostructural signal. In this case, the resolution of the microCT record is high enough to distinguish between tooth tissues. The study of dental tissue proportions was therefore conducted only on the smaller GPA-646b fragment. As only the LRdm1 has a complete crown on this fragment, this tooth is the only one for which we extracted dental tissue proportions and enamel thickness variables (see SI-4). The crown was digitally isolated from the roots (Olejniczak et al., 2008), and 3D surface models of the outer enamel surface (OES) and the enamel-dentine junction (EDJ) were generated using a constrained smoothing algorithm. The microCT record was used to examine the mineralization stage of each tooth, to assess molar crown morphological variation on both the OES and EDJ and to measure dental tissue proportions and enamel thickness. All the raw data obtained for the Gargas specimen are available in this paper and its supplementary information.
11Bivariate plots of mandible dimensions, crown dimensions and dental tissue proportions were used to compare the morphometric characteristics of the Gargas mandible with those of the comparative specimens. A principal component analysis based on the mandible dimensions was performed for the specimens without missing data.
12Measurements obtained for the mandibular corpus, dental crown diameters, tissue proportions and enamel thickness were compared to those of Neandertals (Neand), Upper Palaeolithic modern humans (UPMH) and Holocene modern humans (HMH). Details of all comparative samples are provided in SI-5. For the mandible, the comparative sample comprised 90 immature specimens (table S6). For the permanent and deciduous teeth, the crown dimensions of GPA-646 were compared with those of 317 individuals (table S7). For the dental tissue proportions and enamel thickness of the LRdm1 of GPA-646, the comparative sample comprised 27 specimens (table S8).
13E. Trinkaus (Washington University) kindly provided most of the comparative data for the external measurements. Comparative data on dental tissue proportions and enamel thickness were acquired directly in this study.
14The human immature mandible GPA-646 (figure 1) is now free of the thin layer of concreted sediment that covered around 50% of its lateral surface (figure 1a-b). It is composed of two mandibular fragments, GPA-646a and GPA-646b, that are joined along a longitudinal post-mortem fracture at the interdental septum of the right lateral incisor (LRdi2) and right deciduous canine (LRdc) (figure 1c). The larger fragment GPA-646a corresponds to the left half of the mandible and a small portion (ca. 1.5 cm) of the right half. The fragment is well preserved and only displays minor damage: the left coronoid process is missing, the left gonial angle is broken, and the condylar process and the inferior part of the symphyseal region are eroded (figure 1b,d). The lingual surface of GPA-646a is firmly attached to a flat triangular pebble by a thin layer of concretion (figures 1, S2). GPA-646b is a smaller fragment (22 × 14 mm) of the right half of the body (figures 1c, 2). Matrix infill of the cancellous bone is visible on the microCT record for both fragments (figures 2, S2).
Figure 2
a) 3D rendering of GPA-646a and GPA-646b virtually joined together showing the bone (in semi-transparency) and the teeth in occlusal view, b) In situ deciduous and permanent teeth of GPA-646a in posterior view, c) Close-up of GPA-646b and its in situ teeth in right lateral view, d) Virtually extracted and 3D rendered deciduous and permanent tooth elements of GPA-646b in occlusal (O), mesial (M), distal (D), lingual (L) and buccal (B) views. Enamel appears in yellow, dentine in blue, and the pulp cavity in red |
a) Rendu 3D de GPA-646a et GPA-646b réunis virtuellement montrant l’os (en semi-transparence) et les dents en vue occlusale, b) Dents déciduales et permanentes incluses de GPA-646a en vue postérieure, c) Gros plan sur GPA-646b et ses dents incluses en vue latérale droite, d) Éléments dentaires déciduaux et permanents extraits virtuellement et rendus 3D de GPA-646b en vues occlusale (O), mésiale (M), distale (D), linguale (L) et buccale (B). L’émail apparait en jaune, la dentine en bleu et la cavité pulpaire en rouge
15The left and right first deciduous molars (respectively LLdm1 and LRdm1) are the only erupted teeth still present on the mandible (figures 1-2). The alveoli of the six anterior deciduous teeth are empty (figures 1c-d, 2a). A portion of the enamel on the lingual and buccal surfaces of the LRdm1 has been lost. The germs of the right lateral permanent incisor (LRI2) and the right permanent canine (LRC) are visible in their crypts at the break between GPA-646a and GPA-646b. The microCT record allows identification of other unerupted tooth germs (figure 2): the central and lateral left permanent incisors (respectively LLI1 and LLI2), the right central permanent incisor (LRI1), a supernumerary incisor causing crowding in the anterior alveolar region (figures 2a-b, S3a), the left permanent canine (LLC), the right and left permanent third premolars (respectively LRP3 and LLP3), the left deciduous second molar (LLdm2, its occlusal surface being just at the alveolar margin) and the left permanent first molar (LLM1). Thus, a total of twelve teeth or germs are present on the GPA-646 mandible (see SI-2 for the complete description): the LRdm1, the LRP3 and the LRC are preserved in the fragment GPA-646b (figure 2c-d), whereas all the other teeth (including the supernumerary incisor) are preserved in the larger fragment GPA-646a (figure 2a-b).
16The reported prevalence of supernumerary teeth in the permanent dentition of the living population ranges from 0.1% to 3.8% (Rajab and Hamdan, 2002). Cases in the lower incisor region are the least common, representing 0% to 5.1% of all supernumerary teeth identified (Rajab and Hamdan, 2002; Leco Berrocal et al., 2007; Celikoglu et al., 2010; Demiriz et al., 2015). The supernumerary tooth of GPA-646 therefore appears to be exceptional in the light of modern clinical data. Supernumerary teeth in the fossil record are also relatively rare and, interestingly, the other Eurasian Upper Pleistocene cases described so far are for one EUP and four MUP specimens from Russia, Moravia and south-western France (table S2). As heredity is believed to be an important etiological factor in the occurrence of supernumerary teeth, their presence in six European pre-LGM modern human individuals (including Gargas) – whereas no cases have been reported so far for older or later prehistoric groups – supports the hypothesis of a genetic singularity common to at least some EUP and MUP individuals. Interestingly, in several mammalian taxa, supernumerary teeth are found very frequently in hybrids (i.e. the product of interbreeding between individuals from genetically differentiated lineages: for a review, see Ackermann, 2010; Ackermann et al., 2019). Ackermann and colleagues wrote (2019: 97) that "The consistency of these findings across taxa strongly suggests that the presence of such nonmetric traits in relatively high frequencies is a general indicator of hybridization", and they indeed suggested this hypothesis for some of the MUP supernumerary tooth cases.
17The symphysis of GPA-646 appears fully fused, both macroscopically and through microCT examination. Complete fusion of this region usually occurs during the first post-natal year in extant humans (Scheuer and Black, 2000). Only the deciduous teeth were erupted at the time of death of the Gargas child. Complete occlusion of the first deciduous molar and alveolar eruption of the second deciduous molar occur on average between 1 and 2 years in extant humans (AlQahtani et al., 2010). The dental mineralization stages (AlQahtani et al., 2010) for each tooth element of GPA-646 point to an age between 1 and 3 years (table S3). According to the results of the Bayesian analysis (table S4), the mineralization sequence (C/C/C/A/0/C/0) displayed by the permanent teeth of GPA-646 has a frequent occurrence in extant humans (see SI-3), and this specific mineralization sequence is shared with a girl aged 1.48 years. All the maturational aspects are compatible with an estimated age-at-death of 1 to 3 years for the Gargas child. Our assessment thus gives a slightly younger and narrower range for the age-at-death of this specimen than the previously published estimation.
18The mandibular dimensions for GPA-646 and the comparative samples are given in table 1 (see also table S9). The external cortical surface of all of the bone is highly porous, as is typically seen in very young individuals. A clear chin is observable on its anterior surface (figure 1), formed by a mental protuberance (tuber symphyseos), ca. 11 mm in width, that continues laterally on each side as a thickening of the inferior margin of the corpus, forming an inverted "T"-shaped structure. To each side of the mental protuberance lies a well-marked depression (incisura mandibulae anterior). Even though the inferior part of the area is eroded, this morphology clearly corresponds to the derived modern human anterior symphysis (mentum osseum stage 5 – Dobson and Trinkaus, 2002), which does not occur in Neandertals or in other late archaic humans (Schwartz and Tattersall, 2000; Dobson and Trinkaus, 2002; Liu et al., 2010).
19The symphysis dimensions are relatively small (table 1), especially the height of the symphysis compared to those of El Castillo 2 and Lagar Velho 1 (figure 3a). The lateral corpus decreases slightly in height from the symphysis to the ramus (figure 1; table 1). A single ovoid (ca. 2.0 mm in breadth and 1.0 mm in height) mental foramen is visible on each side (figure 1), directly below the anterior root of the LLdm1. This position is common for immature Neandertal and UP individuals with only deciduous teeth erupted (Coqueugniot, 2000). The corpus does not appear extremely robust at the mental foramen compared to Neandertal individuals at a similar stage of maturation (Arnaud, 2015). More posteriorly on the lateral surface of the mandible, a distinctive bulge is present around the crypt of the LLM1 (figures 1b, S2b). Previously described as abnormal (Foucher et al., 2012), this bulging shape appears to be directly linked to the developing tooth germ contained within it. The ramus is low compared to the corpus. The gonial angle is ca. 150° (figure 1; table S9), which is consistent with the general stage of maturation (in extant humans the gonial angle is usually between 150 and 130° before complete eruption of the deciduous dentition (Jensen and Palling, 1954). Unfortunately, the medial surface of the ramus is firmly attached to the block, precluding any direct observations, but based on the 3D model no prominent relief was detected, such as a medial pterygoid tubercle frequently found in Neandertals and in some Early and Middle Pleistocene specimens (Rak et al., 1994; 1996; Bermúdez de Castro et al., 2015). The condyle is medially placed relative to the crest of the mandibular notch (score II, Jabbour et al., 2002), which is common in recent Homo sapiens of a similar age (Jabbour et al., 2002). The dental arcade of GPA-646 appears wide (table 1). The bi-external deciduous canine and first molar breadths (figure 3b-d) are well above the mean values for both UPMH and HMH (table 1; figure S4). In particular, both values for dimensions are higher than the maximum values for UPMH individuals with only deciduous teeth erupted and recent Europeans aged 1 to 3 years, and are roughly similar to Neandertal values (figures 3d, S5). While the two corpus dimensions show an overlap between Neandertals and modern humans (figure S6), it has been shown that plotting mandibular breadth against lateral corpus height clearly distinguishes Neandertals from early modern humans (Verna et al., 2012). Interestingly, the morphology of GPA-646 appears to be closer to that of Neandertals than other UPMH children (figures 3b, S5), i.e. a broad mandible compared to its height. Despite the Gargas child being very young, its mandible displays large breadth dimensions (figure 3b-c) that are comparable to Neanderthal specimens. The results of the principal component analysis based on these six mandibular variables (figure 4) confirm that the Gargas mandible differs from contemporaneous specimens: the Gargas mandible appears to be closer to the morphology of Neandertals, while other MUP specimens display a morphology that is similar to the Holocene sample (figure 4).
Table 1
Mandibular dimensions for GPA-646 and the comparative samples of Neandertals (Neand), Upper Palaeolithic (UPMH) and Holocene (HMH) modern humans |
Dimensions mandibulaires de GPA-646 et de l’échantillon comparatif d’individus Néandertaliens (Neand), du Paléolithique supérieur (UPMH) et de l’Holocène (HMH)
Figure 3
Bivariate plots of the mandibular dimensions of Gargas compared to the values for Neandertals (triangles), Upper Palaeolithic (diamonds) and Holocene modern humans (crosses) for a) Symphysis breadth and height, b) Corpus height and bicanine breadth, c) Bicanine breadth and bi-dm1 breadth (with only comparative S1 specimens for a, b and c, see SI-6) and d) bi-dm1 breadth against age (with all immature specimens) |
Graphiques bivariés des dimensions mandibulaires de Gargas comparées aux valeurs des individus Néandertaliens (triangles), du Paléolithique supérieur (losanges) et de l’Holocène (croix) pour a) La largeur et la hauteur de la symphyse, b) La hauteur du corps et la largeur bicanine, c) La largeur bicanine et la largeur bi-dm1 (avec seulement les individus S1 pour a, b et c, voir SI-6) et d) La largeur bi-dm1 selon l’âge (avec tous les individus immatures)
Figure 4
Results of the principal component analysis for the mandibular dimensions (see table 1) of Gargas compared to Neandertals (triangles), Upper Palaeolithic (diamonds) and Holocene modern humans (crosses) |
Résultats de l’analyse en composantes principales pour les dimensions mandibulaires (voir tableau 1) de Gargas comparé aux individus Néandertaliens (triangles), du Paléolithique supérieur (losanges) et de l’Holocène (croix)
20It has been suggested that the large breadth of the dental arcade in Middle Palaeolithic fossils may be related to large anterior dental dimensions (Mallegni and Trinkaus, 1997). This hypothesis can be rejected for GPA-646, as its anterior teeth are not especially large (see infra). It seems possible that the morphology of the GPA-646 arcade is the by-product of the presence of a supernumerary tooth in the incisor region. In any case, the proportions of the GPA-646 mandible appear to be "archaic", i.e. closer to the earlier groups (Neandertals and MPMH) than to UPMH (including the contemporaneous Lagar Velho 1 and El Castillo 2 specimens) or the comparative Holocene sample. This result is not surprising considering that previous analyses have shown the presence of such "archaic" traits in early modern human specimens from Europe, Asia and Africa (Trinkaus et al., 2003; Crevecoeur and Trinkaus, 2004; Shang and Trinkaus, 2010). These traits have been used to underline that the biological processes through which humans became "modern" were more complex than usually thought and/or reflect admixture between modern human and regional late archaic human groups (Trinkaus et al., 2003; Shang and Trinkaus, 2010).
21Dental non-metric scores for the deciduous and permanent molars (see SI-7) reveal that no supplementary cusps are present in GPA-646 and the hypoconulid on the LLM1 is small (table S10). The morphology of the deciduous and permanent molars as observed on both the outer enamel surface (OES) and the enamel-dentine junction (EDJ) appears simple, especially compared to more recent Upper Palaeolithic individuals (from Lafaye and La Marche, see Le Luyer, 2016). Overall, the morphology of the Gargas teeth displays modern characteristics.
22Crown mesiodistal and buccolingual diameters for GPA-646 and the comparative samples are provided in table 2. The dimensions for the anterior permanent teeth of GPA-646 fall within the variability of the comparative samples and are closer to the mean values for UPMH specimens. Conversely, the crown dimensions of the GPA-646 deciduous and permanent molars are very large (figure 5), larger than the averages obtained for the Neandertal, UPMH, and HMH samples (table 2). GPA-646 is especially impressive for the mesiodistal diameter of its deciduous molars (figure 5): the values for the left and right Ldm1s exceed all the individual values from the comparative samples, and the value for the Ldm2 is only equalled and exceeded by two EUP specimens. Both deciduous and permanent molars of GPA-646 display larger mesiodistal dimensions than the two other contemporary immature Lagar Velho 1 and El Castillo 2 specimens (figure 5). For the buccolingual dimensions, only the Ldm1s of GPA-646 clearly exceed those of Lagar Velho 1 and El Castillo 2 (figure 5a) while the Ldm2 and LM1 of the three MUP immature specimens show relatively similar buccolingual dimensions (figure 5b-c).
Table 2
Dental crown dimensions for GPA-646 and the comparative samples of Neanderthals (Neand), Upper Palaeolithic (UPMH) and Holocene modern humans (HMH) |
Dimensions des couronnes dentaires de GPA-646 et de l’échantillon comparatif d’individus Néandertaliens (Neand), du Paléolithique supérieur (UPMH) et de l’Holocène (HMH)
Figure 5
Bivariate plots of the mesiodistal (MD) and buccolingual (BL) dimensions for a) The Ldm1, b) The Ldm2 and c) The LM1 of Gargas compared to the values for Neandertals (triangles), Upper Palaeolithic (diamonds) and Holocene modern humans (crosses) |
Graphiques bivariés des dimensions mésiodistales (MD) et buccolinguales (BL) pour a) La Ldm1, b) La Ldm2 et c) La LM1 de Gargas comparées aux valeurs pour les individus Néandertaliens (triangles), du Paléolithique supérieur (losanges) et de l’Holocène (croix)
23UP individuals (including Lagar Velho 1 and El Castillo 2) tend to display relatively large deciduous and permanent molars, but their dimensions remain smaller on average than those of the Gargas child. Indeed, the deciduous molars of GPA-646 exhibit crown dimensions that are rarely seen, if at all, in the entire UP sample. Interestingly, greater tooth size has been considered as a possible indicator of hybrids in non-human primates (Jolly et al., 1997; Ackermann, 2010). The unusual dental dimensions of the Gargas molars may thus be related to undescribed admixtures of different early European populations of anatomically modern humans or, less likely considering the direct date obtained for GPA-646, to interactions between late Neandertals and early modern humans. In any case, our study of GPA-646 confirms that the phenotypical diversity of the European pre-LGM modern human fossil sample tends to be very high (Zilhão and Trinkaus, 2002; Teschler-Nicola, 2006; Trinkaus et al., 2012). Previous studies have reported a significant decrease in crown size for permanent molars (expressed for both diameters) between pre and post-LGM groups (Brace, 1967; Frayer, 1977). When plotted against chronological variables (i.e. C14 dates; figure S7), the LM1, Ldm1, and Ldm2 crown dimensions of the UPMH specimens included in this study do not appear to follow this trend and no decrease in crown sizes between 15,000 and 10,000 BP has been identified. Indeed, greater LM1 dimensions are present in the MUP and in the terminal phase of the LUP (figure S7), and while a slight decrease seems noticeable for the deciduous molars, no clear trends over time are identified as large teeth are also shown to occur at the end of the UP (figure S7). However, geographical patterns have emerged from our analysis: the Ldm2 dimensions of (south)-western European UP fossils appear to be larger than those of (north)-eastern UP specimens (table S11). Similar results have been obtained with a subsample of specimens dated between 45,000 and 20,000 BP, and this trend is also seen for the MUP only, although the sample size is relatively small (table S11). Our analysis is the first to report such a geographical pattern for UP dental dimensions. GPA-646, as an MUP specimen from western Europe, thus presents extreme dental dimensions that nevertheless fit broader, and previously undescribed, chronological and geographical patterns. As tooth crown size is governed by genetic, epigenetic and environmental factors (see Hughes and Townsend, 2013 for a review), our results suggest complex interactions of these factors through the European UP and may indicate distinct populations during the MUP, not identified on the basis of cranial morphology (Mounier et al., 2020). Further analyses are required in order to clarify these interactions.
24The internal dental characteristics of GPA-646 have very high values (tables 3-4): the enamel area and volume of its LRdm1 exceed all of the individual values from the comparative samples (figure S8). Moreover, its crown area and volume and its EDJ length and area exceed those of the other modern humans and are more similar to Neandertal values. However, the tissue proportions within the crown sets GPA-646 apart from Neandertals, with particularly high values for enamel area and volume but lower dentine and pulp values (figure S8). Therefore, the LRdm1 of GPA-646 exhibits a pattern closer to UPMH individuals with a low percentage of the crown volume and area that is dentine and pulp, and high AET and RET values both in 3D and 2D. The visual representation of the RET (figure 6), clearly distinguishing Neandertals from modern human specimens, shows the Gargas LRdm1 to be closer to the latter group but at an extreme position, especially for 2D variables (figure 6b).
Table 3
3D dental tissue proportions and enamel thickness of the LRdm1 for GPA-646 and the comparative samples of Neandertals (Neand), Upper Palaeolithic (UPMH) and Holocene modern humans (HMH) |
Proportions des tissus dentaires en 3D et épaisseur de l’émail de la LRdm1 de GPA-646 et de l’échantillon comparatif d’individus néandertaliens (Neand), du Paléolithique supérieur (UPMH) et de l’Holocène (HMH)
Table 4
2D dental tissue proportions and enamel thickness of the LRdm1 for GPA-646 and the comparative samples of Neandertals (Neand), Upper Palaeolithic (UPMH) and Holocene modern humans (HMH) |
Proportions des tissus dentaires en 2D et épaisseur de l’émail de la LRdm1 de GPA-646 et de l’échantillon comparatif d’individus néandertaliens (Neand), du Paléolithique supérieur (UPMH) et de l’Holocène (HMH)
Figure 6
Bivariate plots of the internal proportions of the LRdm1 of Gargas compared to values for Neandertals (triangles), Upper Palaeolithic (diamonds) and Holocene modern humans (crosses). a) 3D average enamel thickness (log) against crown dentine and pulp volume (log) and b) 2D average enamel thickness (log) against crown dentine and pulp area (log) |
Graphiques bivariés pour les proportions internes de la LRdm1 de Gargas comparées aux valeurs pour les individus Néandertaliens (triangles), du Paléolithique supérieur (losanges) et de l’Holocène (croix). a) Épaisseur moyenne de l’émail en 3D (log) contre le volume de dentine et de pulpe dans la couronne (log) and b) Épaisseur moyenne de l’émail en 2D (log) contre l’aire de la dentine et de la pulpe dans la couronne (log).
25Enamel thickness can change rapidly over short evolutionary periods (Alvesalo and Tigerstedt, 1974; Potter et al., 1976; Hlusko et al., 2004; Horvath et al., 2014) and a thick enamel is associated with strong masticatory forces (Schwartz, 2000; Lucas et al., 2008). It may therefore be suggested that the high enamel thickness in GPA-646 indicates high biomechanical loads for this MUP population. Endostructural analyses of other MUP teeth from south-western France are required in order test this hypothesis.
26The GPA-646 mandible belongs to an MUP child who died between 1 and 3 years of age. Its characteristics differ from the two other contemporaneous individuals from western Europe, namely Lagar Velho 1 from Portugal and El Castillo 2 from Spain. The Gargas child displays modern morphological traits, such as a true chin and a modern pattern of dental maturation. However, our examination of both external and internal characteristics reveals that this Late Pleistocene child possessed a robust mandible with a wide dental arcade. The deciduous molars are large and have an enamel quantity exceeding all reported values so far for modern humans. It also displays a very rare condition, namely a supernumerary lower permanent incisor, which can be added to the list of Late Pleistocene congenital disorders and rare anomalies. Interestingly, these last two conditions have been identified as possible indicators of hybridization in the human fossil record (Harvati et al., 2007; Ackermann, 2010; Ackermann et al., 2019), and their presence in GPA-646 may indicate large-scale movements of individuals from previously isolated populations during the EUP and/or the MUP. Moreover, the differences identified between the three contemporaneous MUP immature individuals from Spain, Portugal and France raise the question of MUP biological variability and the subsequent role of genetic, epigenetic, and environmental factors.
Acknowledgements: Thanks to the municipality of Aventignan, the owner of the cave, and the not-for-profit organizations "Archéologies" and "Association pour le Rayonnement de l’Art Pariétal Européen" for their help during our excavations. Thanks also to Monique Drieux (Materia Viva) for her help during the cleaning of the fossil. SV thanks Dominique Henry-Gambier (University of Bordeaux, CNRS, UMR 5199 PACEA) for giving him the opportunity to study this fossil. Thanks to Arnaud Mazurier (CNRS, UMR 7285 IC2MP) for the acquisition of the 3D data and to Erik Trinkaus (Washington University) for sharing his data and for his helpful comments on the manuscript. This research was funded by the French National Research Agency (ANR GRAVETT’OS, ANR-15-CE33-0004). The excavations in Gargas were funded by the French Ministry of Culture (DRAC Occitanie) with support from the Hautes-Pyrénées département council. MLL’s work was supported by the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement no. 796499.