Navigation – Plan du site

AccueilNumérosvol. 32/1Differentiation of Bovinae teeth ...

Differentiation of Bovinae teeth from the Portel‑Ouest Mousterian cave (Loubens, Ariège, France) by near‑infrared spectroscopy and chemometrics

Différenciation des dents de bovinae de la grotte moustérienne du Portel-Ouest (Loubens, Ariège, France) par spectroscopie proche infrarouge et chimiométrie
Cecile Levasseur‑Garcia et Régis Vézian
p. 11‑27

Résumés

Le site du Portel, situé sur le piémont pyrénéen en France (Ariège), comprend deux grottes : la grotte du Portel‑Est, célèbre pour ses peintures rupestres principalement magdaléniennes, et la grotte du Portel‑Ouest correspondant essentiellement à un habitat moustérien (Couches L à B1a : 135 000 à 36 300 ans avant le ; Ajaja, 1994, Tissoux, 2004). Les fouilles stratigraphiques de plus de cinq mètres de profondeur, réalisées dans le Portel‑Ouest par Joseph et Jean Vézian entre 1949 et 1987, ont révélé 33 vestiges néandertaliens associés à plus de 200 000 autres vestiges. Dans ces vestiges, 1483 restes de Bovinae ont été trouvés. Dans une étude paléontologique précédente, la différenciation entre Bos primigenius et Bison priscus s'est avérée difficile. En tenant compte de critères morphologiques, biométriques ou morphométriques, seuls 21 % des restes ont pu être attribués à l'un des deux genres, avec une nette prédominance pour Bison priscus. Ainsi, 79 % des restes de Bovinae n'ont pu être attribués à l'un ou l'autre des deux genres, soit parce qu'ils avaient des formes composites, soit parce qu'ils étaient trop fragmentés ou trop usés pour tenter une attribution spécifique. Le but du présent projet est donc de développer et d'exploiter l'outil rapide de spectroscopie proche infrarouge pour discriminer ces genres, en se concentrant sur les dents. Chaque dent de la base de données a été scannée à l'aide d'un spectromètre proche infrarouge. Pour 53 dents, la base de données contenant les spectres infrarouges a été complétée par l'expertise paléontologique désignant la dent comme appartenant à Bos ou à Bison. Des outils chimiométriques ont ensuite été utilisés pour construire un modèle permettant d'attribuer la dent à un genre, en fonction de son spectre infrarouge. Dans une dernière étape, le modèle a été utilisé pour prédire le genre de 76 dents qui ne pouvaient pas être attribuées au genre Bos ou Bison par les paléontologues. Cette étape a permis d'obtenir une projection de la population animale vivant à cette époque sur le site de Portel. Enfin, les résultats ont été comparés aux analyses paléogénétiques réalisés sur ces vestiges.

Haut de page

Texte intégral

1 ‑ Introduction

1The paleontological study of Bovinae with the Bos‑Bison differentiation is of significant interest on the paleoclimatic and paleo‑environmental levels. The Bos originates generally from humid locations with temperate‑to‑cold temperatures, favorable to the recolonization of forests, especially deciduous. Bos prefer a landscape of meadows or clearings in semi‑wooded environments sufficiently sheltered from the winds in a rather cool climate (Fontana, 2003; Auguste, 2012). In historical times, Bos was found in Europe, even in regions with a rather cold climate such as the Baltics or Poland (Guintard & Rewerski, 1999; Van Vuure, 2002, 2005).

2In today’s breeding farms, domesticated aurochs can, given shelter, withstand temperatures only as low as −20°C. However, if they adapt well to the cold, domestic cattle cannot tolerate heat above 24°C, which is a major handicap (Flaba et al., 2014). In fact, the metabolism and rumination of Bovinae produce significant heat, which obliges them to regulate their internal temperature to avoid a rapid onset of heat stress. The ideal ambient temperature for domestic aurochs is between ‑5°C and 24 °C, with a relative humidity between 50% and 70% (Flaba et al., 2014). These factors strongly suggest that the original aurochs were not adapted to extreme temperatures and dry conditions, unlike reindeer, nor to a cold steppe environment wind swept, unlike Bison priscus (Auguste, 2012).

3In previous work, we studied 1483 remains of Bovinae from the Portel‑Ouest cave (mainly from isotopic stage 3 layers; layers G to B1; from 44.9 ± 6.7 ka to 36.3 ± 5.4 ka : tab. 1). Differentiation between Bos and Bison proved difficult because the bones had been fragmented either by anthropic action or by carnivores, leaving us to make a distinction only by the teeth that had not been excessively worn. For these reasons, only 21% of the remains could be attributed to Bos primigenius and Bison priscus mediador (Vézian, 2014). Following this paleontological study, ancient DNA was extracted from Crocuta spelaea coprolites in the caves of this collection (Utge et al., 2020). This identification required a high level of expertise and led to the identification of Crocuta spelaea prey at levels corresponding to isotopic stage 3 (tab. 1). DNA from Crocuta spelaea and forest bison (in the lineage of Bison schoetensacki) were also identified, making this study complementary to the paleontological studies that had initially identified Bison priscus and Bos primigenius.

4According to a paleogenetic study based on mitochondrial DNA analysis, the three species come from a common lineage that separated at around 928 ka into two lines, that of Bison priscus and that common to the Bison schoetensackiBos primigenius lines, the latter separated into two lines at around 768 ka. Bos primigenius is thus genetically closer to the lineage of Bison schoetensacki and its descendants than to Bison priscus (Massilani et al., 2016).

5Given the difficulty of differentiating between Bovinae genera due to their relatively close kinship, another method that would allow real‑time screening of teeth would be of great interest to palaeontologists. This tool could be near‑infrared spectroscopy (NIRS). NIRS is fast (around 1 mn per sample), non‑destructive, easy to use, does not require chemicals, offers the possibility of numerous repetitions, and has many applications, both in situ and in the laboratory (Margaris, 2014). This method is used in other fields to differentiate wood species, bacteria... (Bertrand & Dufour, 2006; Pasquini, 2018).

6The objective of the present study is thus to develop a method for real‑time screening by infrared spectroscopy of Bos and Bison teeth.

7Teeth consist of a tough inner core, called ‘dentin’, and a hard, glossy, outer substance that covers the tooth crown, called ‘enamel’ (Kohn et al., 1999). Enamel consists of approximatively 96% calcium phosphate, 3% water and 1% organic matter, whereas dentine consists of approximatively 70% to 75% calcium phosphate, 20% organic matter and 5 to 10% water (Kohn et al., 1999). In contrast, the mineral phase of fossil bones and teeth consist of hydroxy‑carbonate apatite (Lee‑Thorp & van der Merwe, 1991; Michel et al., 1996), which is the mineral fraction of vertebrate skeleton (Huang et al., 2007).

8Relatively few infrared spectroscopy studies of fossils exist (Furedi & Walton, 1968; Suzuki, 1975), and those that do mainly involve mid‑infrared spectroscopy (Termine & Posner, 1966; Weiner et al., 1993; David et al., 1996; Dauphin & Lange‑Badré, 2000). For example, authors (Michel et al., 1996) studied the chemical and structural changes that occur during fossilization in bones and dentine from Cervus elaphus jaws. By using infrared spectroscopy, among other techniques, they concluded that fossil dentine is less stable than enamel—a hypothesis confirmed by another research (Lee‑Thorp et al., 1997; Lee‑Thorp & Sponheimer, 2003; Botha et al., 2004; Dauphin et al., 2007).

9The characteristic bands in the spectral profiles of fossil teeth are found in the mid‑infrared regions 500‑650 cm‑1, 950‑1200 cm‑1, 1400‑1500 cm‑1 and 1640‑1660 cm‑1. They correspond to the organic (amides I, II and III) and mineral (phosphate from the apatite, and carbonate substituting in the apatite lattice ‑ PO4 and CO3) parts of teeth (Dauphin & LangeBadré, 1997; Miller et al., 2001; Botha et al., 2004; Cazalbou et al., 2004; Huang et al., 2007; Lebon et al., 2010; Paschalis et al., 2011). The purpose of this study is to collect spectra directly from previously unground teeth, and to predict the genus of the remains. For this purpose, the chosen vibrational spectroscopy is near infrared.

Tab. 1: Chronological, archaeological, and paleo‑environmental context of the remains of the Portel‑Ouest site.

Tab. 1: Chronological, archaeological, and paleo‑environmental context of the remains of the Portel‑Ouest site.

We grouped ungulates according to their ecological preferences (Griggo, 1996) by distinguishing deciduous forest from grassland ungulates in a more‑or‑less open environment (red deer, roe deer, aurochs, megaceros, wild boar), grassland and cold steppe ungulates in a frankly very open environment steppe bison, horse, woolly rhinoceros, woolly mammoth), cold steppe and tundra ungulates, taiga, with more or less migratory behaviour between the taiga and open spaces (reindeer) rupicolous ungulates (chamois, ibex).

2 – Material and methods

2.1 ‑ Archaeological, chronological, and paleo‑environmental context of the Bovinae teeth studied

10The Bovinae teeth studied herein come from the Portel‑Ouest cave, which corresponds to a Mousterian habitat. This cave is in the department of Ariège in the northern piedmont of the Pyrénées, 10 km west of the town of Foix (fig. 1). The cave corresponds to a blocked porch of the Portel‑Est cave, which is famous for its cave paintings (Breuil & Jeannel, 1955). From 1949 to 1987, Joseph and Jean Vézian excavated a gallery measuring 12 m by 6 m and dug a 5 m deep hole near the north‑western wall, which proved to be excessively rich with thirty‑three Neanderthal human remains at a depth of about 3.4 m (fig. 2) (Vézian, 1953, 1964, 1972, 1989a,b; Becam & Chevalier, 2019; Becam et al., 2019).

11In 1987, after extracting more than 200 000 remains with his father, Jean Vézian stopped the excavations to begin an in‑depth study in collaboration with Professor Henry de Lumley and his team. Since 1987, the publication of about a hundred articles and academic works have led to a better understanding of the subsistence strategies of Neanderthal man living at the foot of the Pyrénées.

12During the excavation, 20 layers were distinguished over a depth of 5 m. These layers were grouped into five archaeostratigraphic sets (tabs. 1 & 2). Layer A is not shown in the table 1 and corresponds to the Holocene serie.

Fig. 1: Location of the Portel site, commune of Loubens (Ariège).

Fig. 1: Location of the Portel site, commune of Loubens (Ariège).

Fig. 2: Stratigraphic survey of section 16/17 of the sondage and plan of the PortelOuest.

Fig. 2: Stratigraphic survey of section 16/17 of the sondage and plan of the Portel‑Ouest.

Tab. 2 : Summary of sedimentary characteristics of the Portel‑Ouest stratigraphy.

Tab. 2 : Summary of sedimentary characteristics of the Portel‑Ouest stratigraphy.

Becam & Chevalier, 2019.

2.2 ‑ Prehistoric teeth

13The deposit yielded 12 963 specifically determinable wildlife remains, including 1486 Bovinae remains corresponding to a minimum of 77 adults and 11 juveniles (tab. 3, Vézian, 2014). Among these remains we counted 611 permanent and deciduous teeth. We chose to study the first, second, and third upper and lower definitive molars because they were present in sufficient number in each layer. We had no teeth in the basic set. The study considers a total of 129 teeth from layers F3 to A.

14Of these 129 teeth, a sample of 53 molar teeth was studied by infrared spectroscopy to establish a model. Next, 76 teeth, mostly undetermined, were tested to attempt a specific determination. For each tooth, the layer in which it was found was noted (A, B, B1, B1A, C1, D, F, F1, F1A, F2, and F3). The layer in which 7 undetermined teeth were found is not indicated. The characteristics of these teeth are gathered in tables 4‑6. Most teeth come from the F2 layer.

Tab. 3: General inventory of the fauna of Portel‑Ouest in number of remains.

Tab. 3: General inventory of the fauna of Portel‑Ouest in number of remains.

Vézian, 2014. R= reworked.

Tab. 4: Layered distribution of all teeth studied.

Tab. 4: Layered distribution of all teeth studied.

Vézian, 2014. A = Holocene.

Tab. 5: Designation of the teeth of the studied Bovinae.

Tab. 5: Designation of the teeth of the studied Bovinae.

Tab. 6: Degrees of wear of the teeth of the studied Bovinae.

Tab. 6: Degrees of wear of the teeth of the studied Bovinae.

1=slightly worn, 7=very worn.

2.3 ‑ Infrared spectral acquisition

15Near infrared (NIR) spectral data were collected from 129 teeth by using a Fourier transform NIR spectrometer (MPA, Bruker Optics, Ettlingen, Germany) equipped with an integrating sphere. The tooth is placed on the quartz window. The scanned surface is about 3 cm². Data were collected from 12 490 to 3 600 cm−1 at 8 cm−1 increments and at room temperature. Each spectrum was obtained by averaging 64 scans. Infrared spectra were collected from each side of each tooth (collar, internal and external faces, top of tooth and root; fig. 3).

Fig. 3: Prehistorical Bison tooth.

Fig. 3: Prehistorical Bison tooth.

2.4 ‑ Data analysis

16The study is composed of two parts: (1) the development of a predictive model of the genus Bos or Bison from the infrared spectra collected from 53 teeth, appraised by Régis Vézian; and (2) the use of this model to predict the genus of the 76 unknown teeth for which the morphological analysis was inconclusive. In step 1, a multivariate data analysis was applied to attribute the teeth to the Bos or the Bison genus. Prior to analysis, the spectral data were mathematically pre‑processed to remove the effects of light scattering and to compensate for baseline offsets and bias. Different types of pre‑processing were tested. These included no treatment (i.e. the raw data were used), first (D1), second (D2) derivatives transformations (Savitzky‑Golay) and multiplicative scatter correction (MSC) ± D1. The derivatives of spectra were used to emphasize small bands and resolve overlapping peaks. MSC was used to compensate for additive and/or multiplicative effects in spectral data. This method also corrects the variation in baseline shift (Levasseur‑Garcia et al., 2015).

17After preprocessing, a principle components analysis (PCA) was done to compress the spectra (1153 variables) into a set of linearly uncorrelated variables called principal components (PCs) (Martens & Naes, 1989; Esbensen et al., 2002). The first PC accounts for the maximum variability in the infrared spectra, with subsequent PCs accounting for ever‑decreasing variability (Tufféry, 2007).

18The first seven principal components (PCs) were used as features to input into the classifiers, which consisted of linear discriminant analysis, quadratic discriminant analysis and Mahalanobis discriminant analysis (Medina‑Gutierrez et al., 2005; Cozzolino et al., 2006; Naes et al., 2017).

19Data processing and modelling were accomplished by using The Unscrambler® (v. X; CAMO A/S, Oslo, Norway). The best model was selected based on the criterion of accuracy calculated based on a confusion matrix (Levasseur‑Garcia & Kleiber, 2015).

20In Step 2, the best model from Step 1 was used to predict the genus of the 76 unidentified teeth based on their infrared spectrum. The identification takes the form of an assigned value for the Bos genus and a second value for the Bison genus. Higher values serve to assign the corresponding genus to the tooth. When the infrared spectra differ significantly, the assignment to Bos or Bison genus is simple; however, it becomes more complicated when the spectra are similar. A third step was used in this case to determine the final assignment. The assigned values for Bos and Bison were categorized in an Ascending Hierarchical Classification (AHC). The AHC was analyzed by using XLSTAT version 2020.1.1, configured to use the Euclidean distance for the dissimilarity parameter, Ward's aggregation method, and automatic truncation threshold by entropy.

3 ‑ Results and discussion

3.1 ‑ Description of the dataset

21129 teeth were analyzed in this study: 53 teeth assigned to the Bos or Bison genus and 76 teeth for which the morphological identification is uncertain. Of the 53 teeth identified, a total of 237 infrared spectra were collected on the teeth that were identified as Bos or Bison. For the analysis, 51 spectra were set for testing and remained 186 spectra for training (78% of the data). All the spectra collected on the same tooth are in the same set (training or testing). The accuracy of the classification models was assessed based on the percent of correct classification. The best model was the one with the highest accuracy. The test set was used to estimate how well the models would perform on new data (Tufféry, 2007; tab. 7).

22Figure 4 shows raw mean spectra from a) collar, face, root and top of teeth, b) Bos and Bison teeth; slight differences appear between the spectral profiles.

23The percentage of Bos and Bison teeth is approximatively the same in both sets and is representative of the database. There are 4 times more Bison teeth than Bos.

Tab. 7: Repartition of infrared spectra collected on aurochs and bison teeth, according to the two sets (training and testing).

Tab. 7: Repartition of infrared spectra collected on aurochs and bison teeth, according to the two sets (training and testing).

3.2 ‑ Principal Component Analysis (PCA) of the infrared spectra

24The entire spectra were submitted to PCA analysis (not shown). The first main component represents 97% of the variability of the spectra. The spectra of the different places of the tooth are superimposed on the representation, thus, the spectra collected on the different faces of the teeth will be used in the same database for the rest of the study. No distinct clusters appeared in the score plots of all pairwise combinations of PC1 to PC7.

3.3 ‑ Discriminant analysis models and performances

25We developed 30 PCA‑DA (Principal Component Analysis‑Discriminant Analysis) models to predict teeth genus (Bos or Bison): five types of pre‑processing (raw, D1, D2, MSC, MSC+D1), three PCA‑DA models (linear, quadratic and Mahalanobis), each of the models was developed for the averaged spectra (average of the spectra obtained on all tooth surfaces) and for the non‑averaged spectra. Table 8 shows the results of the models.

26When considering the 237 individual spectra, collected on all tooth surfaces, the best model is obtained without pre‑processing the spectra ("raw"), for a discriminating PCA‑DA Quadratic analysis. The percentage of well‑identified tooth spectra is 82% during training, and 84% when this model is applied to the test set spectra. When averaging the spectra per tooth (root, neck, face and top of tooth spectra), the best model is obtained without pre‑processing the spectra ("raw"), for a discriminating PCA‑DA linear analysis. The percentage of good ranking reaches 93% in training and 75% in testing. Table 9 presents the matrix of confusion for the model developed on averaged spectra.

27The test set was used to estimate how well the models would perform on new data. In training, for the averaged database, the percentage of Bos teeth recognized as such is 91%. The percentage of Bison teeth well recognized is 93%. These performances deteriorate in testing. Only 50% of Bos teeth are recognized, compared to 88% of Bison teeth. When validating on the 12 teeth independent data set, 3 teeth are incorrectly classified. They correspond to two Bos teeth in layer F and F1, with a low degree of wear, and a Bison tooth in layer B1, with a medium degree of wear.

28The database contains four times fewer Bos than Bison, which means that the model is less efficient in predicting Bos. Given the relatively small number of teeth, the performance in testing should be viewed with hindsight.

Tab. 8: Performances of the different near‑infrared spectroscopic models developed to predict the genus corresponding to the prehistorical tooth Bos or Bison (percentage of correctly classified spectra).

Tab. 8: Performances of the different near‑infrared spectroscopic models developed to predict the genus corresponding to the prehistorical tooth Bos or Bison (percentage of correctly classified spectra).

Tab. 9: Confusion matrix of the best PCA‑DA model, for training and independent tests, when considering averaged infrared spectra.

Tab. 9: Confusion matrix of the best PCA‑DA model, for training and independent tests, when considering averaged infrared spectra.

3.4 ‑ Identification of unidentified teeth based on the model developed

29In addition to the 53 teeth used to develop the previous model, 76 teeth from the same cave were analyzed by infrared spectroscopy. Given that their morphological identification was uncertain, the infrared model previously developed from the 53 teeth was therefore applied to tentatively identify the genus to which each of the 76 teeth belongs.

30For the mean spectrum of each of the 76 teeth, the model predicts the genus (Bos or Bison) and assigns a discriminant score. The genus is determined based on the largest discriminant score. For example, if tooth 1 has a discriminant score of −10.8 for the Bison genus and of −17.9 for the Bos genus, the largest discriminant score is −10.8, so the tooth is assigned the Bison genus (fig. 5). When the Bos score is greater than the Bison score, the tooth (whose numbers are represented all around the diagram, from 1 to 76) is assigned to the genus Bos (green example). If it is the opposite, the genus of the tooth to be determined will be Bison (red example).

31Although the genus assignment is immediate when the discriminants scores differ significantly, it is more subtle for very close discriminants’ scores (purple example). Thus, for each of the 76 teeth, we calculated the difference between the two discriminants’ scores; this difference varies between −10 and 17, but is close to zero for some teeth, in which case assigning a genus is difficult (purple example). In order to determine which discriminant score value serve as boundaries between the Bos and Bison genders, we applied a hierarchical clustering (Agglomerative Hierarchical Clustering ‑ AHC, Ward’s method, proximity type: Euclidian distance; automatic truncation with entropy) to the Bos and Bison discriminant scores for the 76 teeth (2 scores per teeth). This method is part of the unsupervised tools (no indication of the desired number of output classes). The AHC grouped the 76 teeth into three groups comprising 6, 18, and 52 teeth (8%, 24%, and 68%, respectively). These groups are labelled I–II‑III, respectively. Based on the models developed from the infrared spectra collected directly from the teeth and based on this data mining approach (AHC), a third species is present in the Portel collection.

32This would confirm the results of the paleogenetic analysis (Utge et al., 2020).

33The majority genus in Portel being the steppe Bison, this could correspond to 68% of the teeth (group III), while the Bos could correspond to 24% of the teeth (group II) (Utge et al., 2020). The six remaining teeth could correspond to forest Bison remains (group I).

Fig. 5: Summary chart of the discriminant scores for assigning gender to the 76 teeth (Bison, Bos or unknown).

Fig. 5: Summary chart of the discriminant scores for assigning gender to the 76 teeth (Bison, Bos or unknown).

4 – Discussion

34Most of the teeth come from layers deposited during the isotopic stage 3, which was characterized by about fifteen warm and cool phases during the Dansgaard Oeschger and Heinrich events. These variations corresponded to sudden drops in temperature and then rapid rises in temperature, which severely tested both humans and animals for nearly 25 000 years. This was certainly the case for Bos, which had to migrate to warmer regions during the cool phases (perhaps to Spain via the Mediterranean or the Atlantic Ocean) and return to the foot of the Pyrénées during the warm phases.

35These frequent climatic variations help explain the minority presence of the Bos compared with the more permanent Bison, which is better adapted to the cool phases. During the isotopic stage 3, the recolonization of the forest during the warm phases favored the forest bison as well as the bos and red deer.

36The present feasibility study focuses on a small number of teeth. Among these teeth, Bison teeth are much more represented than Bos teeth. The same result is found for all remains from the Portel site. It would be interesting to add Bos teeth and re‑compute the model.

37It would also be interesting to use separately the spectra collected from the faces of the teeth (i.e. from the enamel) and the spectra collected from the top of the tooth, where abrasion could reveal dentine. In fact, according to several authors, diagenesis decreases enamel more than it does dentine.

38Furthermore, the teeth were taken from different geological layers corresponding to the Mousterian and Upper Paleolithic. However, several authors use the composition of fossil teeth and bones to study prehistoric environments (Kohn et al., 1999), diets, diagenesis or for dating samples (Weiner & Bar‑Yosef, 1990; Michel et al., 1996; Lee‑Thorp & Sponheimer, 2003; Reynard & Balter, 2014). A model based only on teeth from a given layer is also of interest, especially the F2 layer, which has the most vestiges.

39Note that, although this method has proven effective on fossilized molars from the Portel‑Ouest, the conclusions remain to be verified based on other teeth and on other sites with different conditions of conservation.

5 ‑ Conclusion

40The paleontological study of Bovinae animals from fossilized remains is always difficult and often presents uncertainties regarding the specific determination because so many intermediate forms are possible. Although paleogenetics provide valuable information when possible, testing all the remains of the various lineages is impossible because it would require a minimum of material sampling. Compared with this method, infrared spectroscopy has the advantage of being non‑ destructive and rapid and of being able to quantify the fractions of the different genera. This is the case for Portel‑Ouest and in this feasibility study: the results obtained from training and independent test sets show that the proposed model suitably differentiates between the teeth of Bos and of Bison. A more in‑depth study using a more complete database should be undertaken shortly. The present study focuses on the upper and lower molars; by analyzing a greater number of remains, we could extend the sampling not only to all molars but also to all premolars, which would allow a specific differentiation and would define more precisely the relative fraction of each genus of Bovinae per layer.

41The results could provide further information on paleo‑climatic and paleo‑environmental variations. In any case, this pilot study opens the door to new research perspectives on the paleontological material of the PortelOuest Bovinae. The resumption of excavations, as of 2019, of this very rich deposit, notably at the level of the first set of Mousterian (layers F to G), should provide us with more teeth to study and for testing the models and refining the method.

Haut de page

Bibliographie

AJAJA O., 1994 ‑ Datation de quelques sites moustériens de Catalogne et du Languedoc par la méthode U‑Th : comparaisons avec la méthode ESR. Thèse de Doctorat. Université de Paris, Muséum national d'histoire naturelle, Paris. 149 p.

AUGUSTE P., 2012 ‑ La grande faune de Conty : taxinomie, écologie et palethnographie. Quaternaire, Hors‑série 5, 2012, 95‑124.

BECAM G. & CHEVALIER T., 2019 ‑ Neandertal features of the deciduous and permanent teeth from Portel‑Ouest Cave (Ariège, France). American Journal of Physical Anthropology, 168 (1), 45‑ 69, doi: 10.1002/ajpa.23719.

BECAM G., VÉZIAN R., TESTU A., MOIGNE A.M., CHACÓN NAVARRO M.G. & PERRENOUD C., 2019 ‑ Le site moustérien de la grotte du Portel‑Ouest (Loubens, Ariège). Rapport annuel de fouille programmée 2019. DRAC Occitanie, Toulouse, 94 p.

BERTRAND D. & E. DUFOUR., 2006 ‑ La spectroscopie infrarouge et ses applications analytiques. Éd. Tec & Doc, Paris, 692 p.

BOTHA J., LEE‑THORP J. & SPONHEIMER M., 2004 ‑ An examination of Triassic cynodont tooth enamel chemistry using Fourier Transform Infrared spectroscopy. Calcified Tissue International, 74 (2), 162‑169, doi: 10.1007/s00223‑003‑0124‑3.

BREUIL H. & JEANNEL R., 1955 ‑ La grotte ornée du Portel à Loubens (Ariège). L’Anthropologie, 59, 197‑204.

CAZALBOU S., EICHERT D., DROUET C., COMBES C. & REY C., 2004 ‑ Minéralisations biologiques à base de phosphate de calcium. Comptes Rendus Palevol, 3 (6), 563‑572, doi: http://dx.doi.org/10.1016/j.crpv.2004.07.003.

COZZOLINO D., FASSIO A., FERNÁNDEZ E., RESTAINO E. & LA MANNA A., 2006 ‑ Measurement of chemical composition in wet whole maize silage by visible and near infrared reflectance spectroscopy. Animal Feed Science and Technology, 129 (3‑4), 329‑336, doi: 10.1016/j.anifeedsci.2006.01.025.

DAUPHIN Y. & LANGEBADRÉ B., 1997 ‑ Composition chimique de quelques Mammifères fossiles d’un milieu continental phosphaté (Aubrelong I, Lot, France, Oligocène inférieur). Paläontologische Zeitschrift, 71 (3), 257‑266, doi: 10.1007/BF02988495.

DAUPHIN Y. & LANGEBADRÉ B., 2000 ‑ Estimating fossil bone preservation; integrating different levels of observation. Paläontologische Zeitschrift, 74 (3), 441‑457.

DAUPHIN Y., MONTUELLE S., QUANTIN C. & MASSARD P., 2007 ‑ Estimating the preservation of tooth structures: towards a new scale of observation. Journal of Taphonomy, 5 (1), 43‑56.

DAVID H., DAUPHIN Y., PICKFORD M. & SENUT B., 1996 ‑ Conservation de sucres dans les phases organiques d'os de bovidés fossiles. Bulletin du Muséum national d'histoire naturelle, Section C, Sciences de la terre, paléontologie, géologie, minéralogie, 18 (2‑3), 403‑415.

DE LUMLEY H., DEGUILLAUME S., LAURENT M., MENZHI M., 1991 ‑ Campagne de relevés de coupes réalisées sous la direction de Henry de Lumley. Centre Européen de Recherches Préhistoriques de Tautavel, 12 p.

ESBENSEN K.H., GUYOT D., WESTAD F., HOUMØLLER & L.P., 2002 ‑ Multivariate data analysis: in practice: an introduction to multivariate data analysis and experimental design. Camo, Oslo, 5th edition, 598 p.

FLABA J., GEORG H., GRAVES. R.E., LENSINK J., LOYNES J., OFNER‑SCHRÖCK E., RYAN T., VAN CAENEGEM L., VENTORP M. & ZAPPAVIGNA P., 2014 ‑ The Design of Dairy Cow and Replacement Heifer Housing Cattle Housing. Report of the CIGR Section II Working group No 14 Cattle Housing, 63 p.

FONTANA L., 2003 ‑ Le Renne, l’Aurochs et les volcans. Hommes, gibiers et reconquête forestière en Grande Limagne entre 12 300 BP et 7 500 BP. PALEO Revue d’Archéologie Préhistorique, 15, 87‑104.

FUREDI H. & WALTON A.G., 1968 ‑ Transmission and Attenuated Total Reflection (ATR) Infrared Spectra of Bone and Collagen. Applied Spectroscopy, 22 (1), 23‑26, doi: 10.1366/000370268774383679.

GRIGGO C., 1996 ‑ Etablissement de courbes climatiques quantifiées à partir des communautés animales pléistocènes suivi d'une application aux gisements de l'abri Suard (Charente) et la grotte de Bois‑Ragot (Vienne). PALEO, Revue d'Archéologie Préhistorique, 8 (1), 81‑97.

GUINTARD Cl. & REWERSKI J., 1999 ‑ Disparition de l’aurochs en Pologne au xviie siècle, et projet de “réintroduction” de l’aurochs‑reconstitué en Mazurie. In L. Bodson (éd.), Animaux perdus, animaux retrouvés : réapparition ou réintroduction en Europe occidentale d’espèces disparues de leur milieu d’origine, journée d’étude, université de Liège, 21 mars 1998, Colloques d’histoire des connaissances zoologiques, Liège, 57‑104.

HUANG C.M., ZHANG Q., BAI S. & WANG C.S., 2007 ‑ FTIR and XRD analysis of hydroxyapatite from fossil human and animal teeth in Jinsha Relict, Chengdu. Spectroscopy and Spectral Analysis, 27 (12), 2448‑2452.

KOHN M.J., SCHOENINGER M.J. & BARKER W.W., 1999 ‑ Altered states: effects of diagenesis on fossil tooth chemistry. Geochimica et Cosmochimica Acta, 63 (18), 2737‑2747, doi: http://dx.doi.org/10.1016/S0016‑7037(99)00208‑2.

LEBON M., REICHE I., BAHAIN J.J., CHADEFAUX C., MOIGNE A.M., FRÖHLICH F., SÉMAH F., SCHWARCZ H.P. & FALGUÈRES C., 2010 ‑ New parameters for the characterization of diagenetic alterations and heat‑induced changes of fossil bone mineral using Fourier transform infrared spectrometry. Journal of Archaeological Science, 37 (9), 2265‑2276, doi: http://dx.doi.org/10.1016/j.jas.2010.03.024.

LEE‑THORP J. & SPONHEIMER M., 2003 ‑ Three case studies used to reassess the reliability of fossil bone and enamel isotope signals for paleodietary studies. Journal of Anthropological Archaeology, 22 (3), 208‑216, doi: 10.1016/s0278‑4165(03)00035‑7.

LEE‑THORP J.A. & VAN DER MERWE N.J., 1991 ‑ Aspects of the chemistry of modern and fossil biological apatites. Journal of Archaeological Science, 18 (3), 343‑354, doi: https://doi.org/10.1016/0305‑4403(91)90070‑6.

LEE‑THORP J., MANNING L. & SPONHEIMER M., 1997 ‑ Problems and prospects for carbon isotope analysis of very small samples of fossil tooth enamel. Bulletin de la Sociéte Géologique de France, 168 (6), 767‑773.

LEVASSEUR‑GARCIA C. & KLEIBER D., 2015 ‑ A method for the allotment of maize contaminated by toxins. Journal of Near Infrared Spectroscopy, 23 (4), 255‑265, doi:10.1255/jnirs.1168.

LEVASSEUR‑GARCIA C., BAILLY S., KLEIBER D. & BAILLY J.D., 2015 ‑ Assessing Risk of Fumonisin Contamination in Maize Using Near‑Infrared Spectroscopy. Journal of Chemistry, Article ID 485864, 10 p, doi: Artn 48586410.1155/2015/485864.

MARGARIS A.V., 2014 ‑ Fourier Eransform Infrared Spectroscopy (FTIR): Applications in Archaeology. In C. Smith (ed.), Encyclopedia of Global Archaeology. Springer New‑York, New York, 2890‑2893, doi: 10.1007/978‑1‑4419‑0465‑2_343.

MARTENS H. & NAES T., 1989 ‑ Multivariate Calibration. Wiley, New York, 419 p.

MASSILANI D., GUIMARAES S., BRUGAL J.P., BENNETT E.A., TOKARSKA M. ARBOGAST R.M., BARYSHNIKOV G., BOESKOROV G., CASTEL J.C., DAVYDOV S., MADELAINE S., PUTELAT O., SPASSKAYA N.N., UERPMANN H.P., GRANGE T. & GEIGL E.M., 2016 ‑ Past climate changes, population dynamics and the origin of Bison in Europe. BMC Biology, 14 (1), 93, doi: 10.1186/s12915‑016‑0317‑7.

MEDINA‑GUTIERREZ C., QUINTANAR J.L., FRAUSTO‑REYES C. & SATO‑BERRU R., 2005 ‑ The application of NIR Raman spectroscopy in the assessment of serum thyroid‑stimulating hormone in rats. Spectrochimica Acta. Part A: Molecular and Biomolecular Spectroscopy, 61 (1‑2), 87‑91, doi: 10.1016/j.saa.2004.03.016.

MICHEL V., ILDEFONSE P. & MORIN G., 1996 ‑ Assessment of archaeological bone and dentine preservation from Lazaret Cave (Middle Pleistocene) in France. Palaeogeography, Palaeoclimatology, Palaeoecology, 126 (1), 109‑119, doi: http://dx.doi.org/10.1016/S0031‑0182(96)00074‑0.

MILLER L.M., VAIRAVAMURTHY V., CHANCE M.R., MENDELSOHN R., PASCHALIS E.P., BETTS F. & BOSKEY A.L., 2001 ‑ In situ analysis of mineral content and crystallinity in bone using infrared micro‑spectroscopy of the ν4 PO43−vibration. Biochimica et Biophysica Acta (BBA) ‑ General Subjects, 1527 (1), 11‑19, doi: http://dx.doi.org/10.1016/S0304‑4165(01)00093‑ 9.

NAES T., ISAKSSON T., FEARN T. & DAVIES T., 2017 ‑ A user‑friendly guide to multivariate calibration and classification. IM Publication Open, 2e edition, 338 p.

PASCHALIS E.P., MENDELSOHN R. & BOSKEY A.L., 2011 ‑ Infrared Assessment of Bone Quality: A Review. Clinical Orthopaedics and Related Research, 469 (8), 2170‑2178, doi: 10.1007/s11999‑010‑1751‑4.

PASQUINI C., 2018 ‑ Near infrared spectroscopy: A mature analytical technique with new perspectives – A review. Analytica Chimica Acta, 1026, 8‑36.

PRINCE G., 2000 ‑ Contribution à l'étude des industries de la grotte du Portel‑Ouest (Commune de Loubens, Ariège, France) : étude techno‑typologique des industries moustériennes. Thèse de Doctorat, Université de Perpignan, Perpignan, 214 p.

REYNARD B. & BALTER V., 2014 ‑ Trace elements and their isotopes in bones and teeth: Diet, environments, diagenesis, and dating of archeological and paleontological samples. Palaeogeography, Palaeoclimatology, Palaeoecology, 416, 4‑16, doi: http://dx.doi.org/10.1016/j.palaeo.2014.07.038.

SUZUKI M., 1975 ‑ Studies on the physicochemical nature of hard tissue. Infrared, NMR, X‑ray diffraction investigation of hydroxyl‑radical, crystalline water and carbonate substitution in biological apatites. Colloques Internationaux du CNRS, Physico‑chimie et cristallographie des apatites d’intérêt biologique, 230, 77‑83.

TERMINE J.D. & POSNER A.S., 1966 ‑ Infra‑Red Determination of the Percentage of Crystallinity in Apatitic Calcium Phosphates. Nature, 211 (5046), 268‑270.

TISSOUX H., 2004 ‑ Datation par uranium‑thorium et par résonance paramagnétique électronique de quelques gisements paléolithiques du Pléistocène supérieur de Catalogne (Espagne) et du sud de la France. Thèse de Doctorat, Université de Paris, Muséum national d'histoire naturelle, Paris, 262 p.

TUFFÉRY S., 2011 ‑ Data Mining and Statistics for Decision Making. John Wiley & Sons, Hoboken, 533 p.

UTGE J., SEVEQUE N., LARTIGOT‑CAMPIN A.S., TESTU A., MOIGNE A.M., VEZIAN R., MAKSUD F., BEGOUEN R., VERNA C., SORIANO S. & ELALOUF J.M., 2020 ‑ A mobile laboratory for ancient DNA analysis. Plos One, 15 (3), e0230496, doi: 10.1371/journal.pone.0230496.

VAN VUURE C., 2002 ‑ History, morphology and ecology of the aurochs (Bos primigenius). Lutra, 45 (1), 16 p.

VAN VUURE, C., 2005 ‑ Retracing the aurochs, history, morphology and ecology of an extinct wild ox. Pensoft éd., Sofia‑Moscow, 431 p.

VÉZIAN J., 1953 ‑ Gisement moustérien de la grotte du Portel à Loubens (Ariège). Bulletin de la Société d'Histoire Naturelle de Toulouse, 88, 211‑217.

VÉZIAN J., 1964 ‑ Fouilles à l'entrée ouest de la grotte du Portel. Bulletin de la Société Méridionale de Spéléologie et Préhistoire, 11, 4‑10.

VÉZIAN J., 1972 ‑ La grotte du Portel. Commune de Loubens, (Ariège). Bulletin de la Société d'Études et de Recherches Préhistoriques des Eyzies, 21, 1‑16.

VÉZIAN J., 1989a ‑ Le Moustérien et le Châtelperronien dans le haut bassin de la Garonne et de l’Ariège. L'homme de Néandertal, 8, 251‑256.

VÉZIAN J., 1989b ‑ Les fouilles à l'entrée du Portel‑ouest. Bulletin de la Société Préhistorique de l'Ariège, 44, 225‑261.

VÉZIAN R., 2014 ‑ Étude paléontologique des Bovinae et des Equidae de la grotte moustérienne du Portel‑ouest (Ariège, France) : Cadres biostratigraphique, biochronologique et paléoenvironnemental. Thèse de Doctorat, Université de Perpignan, Perpignan, 446 p.

WEINER S. & BAR‑YOSEF O., 1990 ‑ States of preservation of bones from prehistoric sites in the Near East: A survey. Journal of Archaeological Science, 17 (2), 187‑196, doi: https://doi.org/10.1016/0305‑4403(90)90058‑D.

WEINER S., GOLDBERG P. & BAR‑YOSEF O., 1993 ‑ Bone Preservation in Kebara Cave, Israel using On‑Site Fourier Transform Infrared Spectrometry. Journal of Archaeological Science, 20 (6), 613‑627, doi: http://dx.doi.org/10.1006/jasc.1993.1037.

Haut de page

Table des illustrations

Titre Tab. 1: Chronological, archaeological, and paleo‑environmental context of the remains of the Portel‑Ouest site.
Légende We grouped ungulates according to their ecological preferences (Griggo, 1996) by distinguishing deciduous forest from grassland ungulates in a more‑or‑less open environment (red deer, roe deer, aurochs, megaceros, wild boar), grassland and cold steppe ungulates in a frankly very open environment steppe bison, horse, woolly rhinoceros, woolly mammoth), cold steppe and tundra ungulates, taiga, with more or less migratory behaviour between the taiga and open spaces (reindeer) rupicolous ungulates (chamois, ibex).
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-1.jpg
Fichier image/jpeg, 94k
Titre Fig. 1: Location of the Portel site, commune of Loubens (Ariège).
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-2.jpg
Fichier image/jpeg, 61k
Titre Fig. 2: Stratigraphic survey of section 16/17 of the sondage and plan of the Portel‑Ouest.
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-3.jpg
Fichier image/jpeg, 48k
Titre Tab. 2 : Summary of sedimentary characteristics of the Portel‑Ouest stratigraphy.
Crédits Becam & Chevalier, 2019.
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-4.jpg
Fichier image/jpeg, 81k
Titre Tab. 3: General inventory of the fauna of Portel‑Ouest in number of remains.
Légende Vézian, 2014. R= reworked.
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-5.jpg
Fichier image/jpeg, 86k
Titre Tab. 4: Layered distribution of all teeth studied.
Légende Vézian, 2014. A = Holocene.
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-6.jpg
Fichier image/jpeg, 36k
Titre Tab. 5: Designation of the teeth of the studied Bovinae.
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-7.jpg
Fichier image/jpeg, 29k
Titre Tab. 6: Degrees of wear of the teeth of the studied Bovinae.
Légende 1=slightly worn, 7=very worn.
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-8.jpg
Fichier image/jpeg, 42k
Titre Fig. 3: Prehistorical Bison tooth.
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-9.jpg
Fichier image/jpeg, 18k
Titre Tab. 7: Repartition of infrared spectra collected on aurochs and bison teeth, according to the two sets (training and testing).
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-10.jpg
Fichier image/jpeg, 28k
Titre Tab. 8: Performances of the different near‑infrared spectroscopic models developed to predict the genus corresponding to the prehistorical tooth Bos or Bison (percentage of correctly classified spectra).
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-11.jpg
Fichier image/jpeg, 83k
Titre Tab. 9: Confusion matrix of the best PCA‑DA model, for training and independent tests, when considering averaged infrared spectra.
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-12.jpg
Fichier image/jpeg, 50k
Titre Fig. 5: Summary chart of the discriminant scores for assigning gender to the 76 teeth (Bison, Bos or unknown).
URL http://journals.openedition.org/quaternaire/docannexe/image/15000/img-13.jpg
Fichier image/jpeg, 28k
Haut de page

Pour citer cet article

Référence papier

Cecile Levasseur‑Garcia et Régis Vézian, « Differentiation of Bovinae teeth from the Portel‑Ouest Mousterian cave (Loubens, Ariège, France) by near‑infrared spectroscopy and chemometrics »Quaternaire, vol. 32/1 | 2021, 11‑27.

Référence électronique

Cecile Levasseur‑Garcia et Régis Vézian, « Differentiation of Bovinae teeth from the Portel‑Ouest Mousterian cave (Loubens, Ariège, France) by near‑infrared spectroscopy and chemometrics »Quaternaire [En ligne], vol. 32/1 | 2021, mis en ligne le 01 janvier 2021, consulté le 08 février 2025. URL : http://journals.openedition.org/quaternaire/15000 ; DOI : https://doi.org/10.4000/quaternaire.15000

Haut de page

Auteurs

Cecile Levasseur‑Garcia

University of Toulouse, INPT, INP‑PURPAN, 75 voie du T.O.E.C., FR‑31076 TOULOUSE. Email: cecile.levasseur@purpan.fr

Régis Vézian

University of Toulouse, INPT, INP‑PURPAN, 75 voie du T.O.E.C., FR‑31076 TOULOUSE. Email: regis.vezian@purpan.fr ; UMR 7194 HNHP. CERPT Avenue Léon Jean Grégory, FR‑66720 TAUTAVEL

Articles du même auteur

Haut de page

Droits d’auteur

Le texte et les autres éléments (illustrations, fichiers annexes importés), sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search