Version classiqueVersion mobile

La géoarchéologie française au xxie siècle

Nathalie Carcaud
Gilles Arnaud-Fassetta

Partie III. Alluvionnement, peuplement, stratégies et formes d'adaptation/Section 3. Alluviation, settlement strategies and adaptation

Chapter 13. Aggradation, climate and settlement in the plain of Roussillon (France) in the last millennium

A geoarchaeological approach

Jean-Michel Carozza, Carole Puig et Thierry Odiot

Texte intégral


1The geoarchaeology of the last thousand years is a discipline still in its infancy. Yet in this period the landscape changed under the combined effects of significant demographic shifts, with population growth in the central Middle Ages, demographic crises of the 14th c., changes of land management techniques and land-holdings, as well as climate change during the Mediaeval Climate Anomaly and Little Ice Age. These events profoundly altered the relationship between society and its environment. This is particularly the case in the Mediterranean lowlands, where the pioneering work of Vita-Finzi (1969) recognised a major development during the post-Roman period, the “Younger infill”. While significant progress has been made in understanding early human impacts between the Neolithic and Bronze Age and during the Classical period, there has been less work on the Mediaeval and early Modern period. A survey of geoarchaeological work on the Middle Ages cited in the major online databases Scopus, ISI Web of Knowledge, Springer and Ingenta, shows little activity for this period, although a slight increase has been noticeable since the early 2000s, with a lag of nearly 10 years compared to geoarchaeology as a whole (Fig. 1). Even if this balance does not demonstrate the true value of the work being carried out, it indicates that only a small proportion deals with the Mediaeval period. Thus, as noted by Durand (2010) with respect to Mediaeval bioarchaeology, the geoarchaeology of the Middle Ages and the Modern period remains a discipline to be developed.

2The proliferation of written and iconographic sources from the 12th c. opens new perspectives on the history of the environment (Beck and Delort, 1993), but does not exhaust this topic, especially as this documentation is often biased. Legitimate other approaches remains, just as for earlier periods, even in the face of specific challenges such as dating. This is the only period that can be documented by both written and sometimes iconographic archives and pedo-sedimentary archives, which are generally well preserved. The wealth of possible interactions between these data sources leads to an increase in spatial and temporal resolution (Carozza and Puig, 2012). However, it is also an opportunity to test, validate and better constrain hypotheses concerning the relationship between climate variability, population pressure, changes in patterns of land use and their impact on the environmental operating systems on one hand and geomorphological conditions and environmental changes that past societies adapted to. This methodology that integrates geoarchaeological approaches to recent alluvial formations sensu largo, field archaeology and history, was used to further our understanding the evolution of a Mediterranean coastal lowland, the Roussillon plain, during the last millennium. In particular, we focused on the period of oscillation between the Mediaeval Climate Anomaly (MCA; 10th-13th c. AD; Xoplaqui et al., 2011) and the early Little Ice Age (LIA; 14th-17th c.; Le Roy-Ladurie, 1967).

3This work addresses several questions. Firstly, it examines the possible impact of the MCA and LIA oscillation as dominant controls on the geomorphological evolution of the lowlands. This is not simply in order to go beyond a regional phenomenon, which is generally apprehended continental or global scales, but also to determine the extent, timing and effects locally. Second, we consider the response process of downstream hydro-systems, especially the architecture of the deposits associated with this event and the potential impact on the organisation of landscapes, territories and societies. The emergence of new practices or methods of land management, especially in riparian areas, can then be understood from an environmental perspective and not just from sociopolitical or technical perspectives. Finally, it allows a re-interpretation of the phenomenon of settlement re-organisation between the 13th and 16th c. in the broader context of environmental history.

Fig. 1. Co-occurrence of “geoarchaeology” and “Middle Age” or “Mediaeval” in the titles, keywords and abstracts of the following databases: ISI Web of Knowledge, Science Direct, Springer, and Ingenta. Duplications have been removed.


4This study is based on work with a wider chronological framework on the evolution of the Roussillon plain during the Holocene (Carozza et al., 2005; Carozza, 2011) for which all data from the sub-soil (BDGSS) have been exploited and a series of core samples taken. They revealed the existence of widespread post-Roman alluvial cover, virtually omnipresent throughout the lowlands (Carozza et al., 2009). Its thickness seems highly variable, as indicated by previous work of Serrat (1999) in the plain of Agly, and Marichal et al. (1997) and Calvet et al. (2002) in the plains of Salanque and Tech (Carozza et al., 2011). Moreover, its sedimentary architecture does not seem to be part of a single progradation of alluvial lobes. The Carte Archéologique Nationale (i.e., national map of archaeology) shows large areas where archaeological remains prior to AD 1000 are rare or non-existent or perhaps undiscovered because they are deeply buried (Fig. 2).

5The monitoring of development-driven archaeological operations and the analysis of the reports of earlier excavations were carried out on a wide area of 400 km2. Key areas were identified for coring and geoarchaeological trenches in order to identify and date the sedimentation of alluvial lobes. The methodology used differs from that which is generally applied and implemented by systematic transects. Here, in view of the surface under study and the difficulty of dating, it was based on known archaeological sites on the surface or subsurface, which provided markers making it possible to calibrate levels of soil associated with these occupations, then work outwards stratigraphicaly. Usually these are occupation sites or religious buildings for which there is historical data. They provide a tentative time-frame for occupation – date of foundation or first mention of religious buildings, successive legal status, hearth tax, signs of decline, the local elements –, which can be confirmed by archaeology or 14C dating. The stratigraphic context of these sites allows a first approach to the chronology of the alluvial deposits. Furthermore, this method lies at the heart of the approach, which places archaeological sites in their environments. Written sources can also document some features of the territories around these sites. Thus, Mediaeval documentation sometimes mentions status of names of rivers and land parcel adjacent to the river and these can often be easily located on the current land register. The systematic exploitation of this information led to the reconstruction of old plots and the dating of the operation or abandonment of these channels. Thus, the mention of a river as ‘confront’ of sediments provides a terminus post quem for the abandonment of the channel. Similarly, references to old channels, as Tech Vell or Viel, provide a date ante quem for the abandonment. The cartographic interpretation made from the Napoleonic cadastre or current cadastre can then provide the locations of these channels that can then be found on the ground and studied using a geomorphological, sedimentological and/or palaeohydrological perspectives.

Fig. 2. Archaeological sites dating from the last millennium in the lowlands from the National Archaeological Map (from Service Régional de l’Archéologie Languedoc-Roussillon).

Study area

6The Roussillon plain is an area of extensive coastal lowlands related to the development of a Neogene trapezoidal sedimentary basin (Calvet, 1994; Carozza, 1998). In the north, the fault of the Tet bounds the basin and the Corbieres limestone whereas in the south, the fault of Tech separates the Palaeozoic land from the axial zone. Between the two developed a hillside cut into Neogene to Pleistocene formations. From a functional viewpoint, the plain has been divided into two entities (Mussot and Benech, 1995). In the upstream part, the plain is eroding (i.e., alluvial forms are stepped) whereas downstream, it is accumulating (i.e., aggrading or prograding). The transition between the two domains is indicated by the intersection between the profiles along the most recent Pleistocene terrace and Holocene formations (Carozza, 2011). This segmentation point generally corresponds to the apex of the Holocene fan-delta system. Furthermore, it is not coincident with the maximum Holocene transgression, which is located furthest downstream and whose position depends mainly on the importance of the pleni-glacial incision. Upstream, floodplain development has been dominated by fluvial processes whereas downstream prograding and aggrading systems have developed.

7The Roussillon plain is the outlet of four main catchments of unequal size: the Agly (1045 km2), the Têt (1550 km2), the Réart (160 km2), and the Tech (750 km2) that together form two major sedimentary units. North of Perpignan, the rivers Agly and the Tet have a vast common plain called the Salanque. The interfluve between two rivers is complicated and difficult to identify from a topographical standpoint. In contrast, the petrographic signature of the “black sands” of the Agly River, reworked from the Albian marls, provide a reliable criterion for differentiating between the fluvial influence of the two systems. To the south, despite the presence of a line of high hills, the rivers Tech and Réart developed a common plain between the pond of Canet and Argelès-sur-Mer. Both sectors have been over the past eight years subjected to systematic archaeological monitoring, and other systematic operations.


Mapping alluvial dynamics in the last millennium

8Sediment dynamics over the last millennium can be seen across the plain as a whole. At nearly 23 locations throughout the lowland it is possible to assess the importance of aggradation post-13th c. (Fig. 3). If the record of recoveries is highly heterogeneous, its spatial pattern can be explained by the study of the chronology of their deposition across each of the lobes.

The southern sector

9The work has helped to clarify the dynamics of alluvial lowlands. In the case of the lower Tech, the site of Sainte-Eugenie de Tresmal is emblematic of Mediaeval villages located in the current floodplain, which were abandoned between the late 13th c. and the start of the Modern period (Carozza et al., 2008, 2011; Carozza and Puig, 2012). It allows us to reconstruct the evolution of the southern sector over the last two millennia and particularly after the founding of the Mediaeval village in AD 951, which followed an earlier occupation between the 1st and 4th c. AD. At this site, we can consider the possible role of changes in the hydrological functioning of the plain in the process of village abandonment (Fig. 3). During the 1st millennium, the hydrodynamic conditions were calm and vertical silt accretion rates were low (ca. 1 mm yr-1). The alluvial soils associated with these two phases of occupation show regular but moderate accumulation. The development of the Mediaeval village around the church is an extension of this phase. The first signs of change occur before the digging graves dated between AD 1163 and 1284, which are dug into a first level sandy loam. But the major break in the sedimentary sequence happens after that date, when massive sand mounds are formed (units 4a and 4b, then 5 and 6). The interpretation of such an evolution in terms of allocyclicity (i.e., under climate control) or autocyclicity (displacement of the main channel to the site) is difficult to demonstrate. However, analogous formations of the same date north of the town of Elne on an alluvial lobe of the Tech were formed by progradation of a marsh area, clearly indicating a broadening of the influence of deposition and a widespread tendency to aggradation (Carozza et al., 2011). Thus, disruption of sediments caused by floods is superimposed on a historical breakdown with streams changing course at the time of these brutal societal events.

Fig. 3. Stratigraphy of the Sainte-Eugénie-de-Tresmall early Roman chapel.

The northern sector

10In the Salanque, the main results are from the lower plain of Agly River, the dynamics of the Têt River for the time being remaining more difficult to reconstitute. Monitoring road works along the RD83 has tracked the evolution of the meander waveforms of the Agly between the 4th and 14th c. (Fig. 4). During this time interval, the Agly River flowed north from its current position, between Saint-Laurent-de-la-Salanque and Saint-Hippolyte. It flowed into the lagoon of Leucate, which then extended much further south. A first lobe, associated with a Visigothic occupation site Figuera Mola (Passarrius et al., 2010) developed between the 4th and 9th c. Then the Agly River changed its course and took a more southerly course. This route is clearly marked in the landscape by a meandering path whose toponym Agly vell is explicit. The cutting-offof the meander has not been dated directly, but post-abandonment infill suggests a date between the 14th and 15th c. This is consistent with the historical position of the Agly River. The northern lobe was thus functional between AD 1195 and 1279. The first mention of a toponym indicating the abandonment of the channel of the Agly vell is dated AD 1280. Over this period, at least four secondary channels have been identified.

Chronology of the alluvial crisis

11The dating of major changes of course, as well as the timing of the various alluvial lobes, indicates a change in alluvial dynamics from about AD 1260-1280, reaching a first peak between AD 1330 and 1380. This phase is synchronous with the mega-floods of the north-western Mediterranean that characterise the first episode of the Little Ice Age (Treton, 2007). It is also corresponds with rivers changing their course in central Languedoc, notably of the lower reaches of the Aude River (Verdeil, 1970) associated with the flood of AD 1316-1320. In Roussillon, initiation of change in alluvial dynamics appear before AD 1300, which is atypical, but does correspond with a cooling trend observed in lake deposits in the Continental regions (Magny, 2004) as well as the Mediterranean (Martin-Puertas et al., 2009). A recent summary of data from Spain (Morellon et al., 2011) confirms the existence of this wet period. As for the Eastern Pyrenees, the dendrological data of Buntgen et al. (2008) clearly indicate a cold phase at this time, without specifying the impact of rainfall. The late 14th c. and the first half of the 15th c. seem to have been less cold, but a new phase of degradation occurred about AD 1450, and then again shortly before AD 1600.

Fig. 4. The Agly course changes according to geoarchaeological, archaeological and historical data.
Fig. 5. Alluvium plots limits in Alenya (end of the 14th c.).

12So it seems there is a strong link between the alluvial dynamics and climate variability during this transitional phase between MCA and LIA and population dynamics including in the mountains settle into the rhythm of the climatic changes (Ejarque et al., 2009). However, it is clear that changes that affect water systems have a direct impact on the reorganisation of settlement in the plain.

Impact of settlement and adaptation strategies

13The period between the 9th c. and the 15th c. is a pivotal time for settlement and territorial organisation in the lowlands, which manifests itself in two stages. Between the late 9th c. and the 10th c., written sources indicate the first step in the establishment of a network of settlement on the Roussillon plain. Some of these, particularly small rural settlements, seem to have disappeared. The continuity of these sites with ancient villae is not the normal situation. This new settlement pattern that is being established in the high Middle Ages inaugurates a series of transformations all the more important because they are accompanied by hydrological interventions (drainage, irrigation and drainage of wetlands; Caucanas, 1994) and the establishment of new agricultural structures (Puig, 2003). From the late 13th c. or early 14th c., the network undergoes a new phase of restructuring with the abandonment of some settlements. Abandoned or deserted villages are a ubiquitous European cultural phenomenon that is not specific lowlands. The entire Roussillon s.l., count of churches in Catalunya Romanica indicates 230 buildings in AD 1300, and 90, or 40%, disappeared within the next two centuries (Catafau and Passarrius, 2007). These abandonments were attributed to political and demographic phenomena that mark the beginning of the 14th c. and are the well-known “crises of the 14th c.”. However, in both in the plain and mountainous area (Carozza and Puig, 2008), the role of environmental-hydroclimatic factors needs to be considered. Thus, in the case of Salanque, the cross-checking of archaeological and historical data indicates a dropout rate over the same period of almost 65%. All abandoned sites are located in the lowlands during the central Middle Ages, moreover, the timing of abandonment seems to reveal a specific context, villages declining between early AD 1260-1280 and 1330. Synchronism with the first hydrological events and their effects, suggests that there may be a direct link. Here also, the crossreferencing of geoarchaeological data, archaeological and historical allows us to propose the mechanisms or causes of abandonment.

14This is particularly the case around Alenya, where several archaeological excavations near the present village centre, have revealed fragmentary buried horizons under the alluvial deposits of the Late Middle Ages (Fig. 5). These horizons are associated with a parcel of ground that contains fragments of pottery and ancient material. This system was abandoned and covered by almost 1.1 m of sandy alluvium between the late 14th c. and early 15th c. These depsoits pose two questions: that of a permanent plot one hand and the possible agronomic impacts. The permanence of these plots is shown by the discovery of specific boundary markers (Carozza and Puig, in press), and the decrease in land values due to the effects of sedimentation appears to have been a constant concern from the 14th c. This problem seems general and is seen in written sources from the early AD 1330s. A document of that dates indicates an agreement between Pierre de Puig, Lord of Mossellon (one of the settlements now deserted in the lower valley of the Tech River) and the population of Saint-Cyprien giving the land in freehold because floods made it impossible to use (Alart, C M, VII-31, quoted by Cazes, 1973). Several years later, in AD 1338, the monk Arnaud Tafanel gives to Sainte-Marie-de-la-Salanque, in the low plain, a land he describes as “deteriorata et pauci valorus propter inundacionem aquarum” (Puig, 2009). Crop destruction and sterilisation of the land seems also to have been the motive for the first canalisations of the rivers Agly, Têt, Réart and Tech by Jacques II (cited in Aragon, 1919). The immediate danger represented by flooding does not appear to drive the abandonment directly, as exemplified by the case of Saint-Martin-de-Tura where, from AD 1332, the villagers asked in vain for their village to be moved because of the frequent floods, but in reality the abandonment process took place over a long period of nearly a century and a half.


15This first synthesis of geoarchaeological and historical data from the Roussillon plain during the last millennium shows the magnitude of environmental changes in these lowlands. The importance of these changes both in terms of palaeogeography and alluvial dynamics seems incommensurate with the earlier developments. The transition from a highly aggrading-prograding dynamics during the late 13th c. and early 14th c. has strongly contributed to the current appearance fo the plain and its geomorphology as earlier traits were largely obliterated by Mediaeval activities. In this sense, the current landscape is mainly inherited from this unpublished, morphogenic event.

16Society in the central Middle Ages, the late Middle Ages and early Modern period have had to adapt to new conditions: Sediment aggradation caused widespread difficulties for the maintenance of property, for land recovery, and for the hydraulic structures that had constantly to adapt to new flow conditions. These difficulties seem to have been overcome in the first instance, between AD 1260 and 1330, which could be said to be a resilient phase. From AD 1332 to 1350, the resilience of the communities seems to have been exceeded and the Mediaeval society of Roussillon entered a phase of adaptation. This resulted in a series of innovations, both technical (embankments), legislative (exclosure for riparian forest protection – coppice-wood), political and territorial through the reorganisation of the network of settlements and territories on a large scale. The phenomenon of Mediaeval villages abandonment must therefore be considered in the context of environmental history.



Aragon H., « L’alignement du lit de la Basse et de la Têt à Perpignan au xive siècle », Revue Catalane, no 13, 1919, p. 119-174.

Beck C. et Delort R., Pour une histoire de l’environnement, Paris, CNRS Editions, 1993.

Büntgen U., Frank D., Grudd H. et Esper J., « Longterm summer temperature variations in the Pyrenees », Clim. Dynam., no 31, 2008, p. 615-631.

Calvet M., Morphogenèse d’une montagne méditerranéenne, Les Pyrénées-Orientales, Orléans, BRGM Editions, 1994, 3 tomes.

Calvet M., Serrat P., Lemartinel B. et Marichal R., « Les cours d’eau des Pyrénees Orientales depuis 15 000 ans : état des connaissances et perspectives de recherches » in Bravard J.-P. et Magny M., Les fleuves ont une histoire, Paris, Errance, 2001, p. 279-294.

Carozza J.-M. et Puig C., « Les changements de tracés des cours d’eau d’après les sources historiques et géomorphologiques dans la plaine du Roussillon depuis le xiie siècle: approche théorique et premiers résultats », in Ropiot V., Puig C. et Abbé J.-L., Les plaines littorales en Méditerranée nord-occidentale, Montagnac, Mergoil Editeur, 2012, p. 297-311.

Carozza J.-M., Odiot T., Passarius O., Puig C., Péquignot C. et Valette P., « Impact du Petit Âge Glaciaire sur la plaine alluviale du Roussillon », in De Dapper M., Vermeulen F., Deprez S., Taelman D., River, O., Geoarcheological aspect of rivers and alluvial plains, Gant, Belgéo, 2009, p. 185-197.

Carozza J.-M., « L’évolution géomorphologique des basses plaines littorales méditerranéennes: un outil de compréhension des biais taphonomiques en archéologie? », Méditerranée, no 116, 2011, p. 43-49.

Carozza J.-M., Odiot T. et Puig C., « Le Petit Âge Glaciaire dans la plaine du Roussillon: apport des données géomorphologiques et historiques », Archives du Midi Médiéval, no 27, 2009, p. 207-215.

Carozza J.-M., Pous M., Odiot T., Jorda C. et Carozza L., «Évolution paléogéographique de la plaine du Roussillon. Conséquence sur la répartition des sites néolithiques», in Berger J.-F., Bertoncello F., Braemer F., Davtian G. et Gazenbeek M., Temps et espaces de l’homme en société, analyses et modèles spatiaux en archéologie, XXVe rencontres internationales d’archéologie et d’histoire d’Antibes, Antibes, APDCA, 2005, p. 419-423.

Carozza J.M., Évolution des systèmes géomorphologiques en contexte orogénique: l’exemple de la plaine du Roussillon, Thèse, Université Toulouse, 2005.

Catafau A. et Passarrius O., « La restructuration du peuplement au 10e-11e s. en Roussillon. L’apport de l’étude des habitats ruraux abandonnés à la chronologie de la formation du village », Domitia, no 8-9, 2007, p. 90-121.

Caucanas S., Moulins et irrigation en Roussillon du ixe au xve siècle, Paris, CNRS éditions, 1995.

Cazes A., « En parcourant les cartulaires », Massana, no 17, 1973, p. 7-11.

Durand A., « L’émergence d’outils empruntés aux sciences biologiques végétales en archéologie médiévale en France », in Chapelot J., Trente ans d’archéologie médiévale en France, Un bilan pour un avenir, Publications du CRAHM, 2010, p. 25-38.

Ejarque A., Julià R., Riera S., Palet J.-M., Orengo H.A., Miras Y. et Gascón C., «Tracing history of highland human management in the Eastern Pre-Pyrenees (Spain): an interdisciplinary palaeoenvironmental approach», The Holocene, tome 19, no 8, 2009, p. 1241-1255.

Leroy-Ladurie E., Histoire du climat depuis l’an mil, Paris, Flammarion, 2 vol., 1967.

Magny M., «Holocene climate variability as reflected by mid-European lake-level fluctuations and its probable impact on prehistoric human settlements», Quaternary International, no 113, 2004, p. 65-79.

Martin-Puertas C., Blas L. Valero-Garcés, M. Pilar Mata, González-Sampériz P., Bao R., Moreno A. et Stefanova V., «Arid and humid phases in southern Spain during the last 4000 years: the Zoñar Lake record, Córdoba», The Holocene, no 18, 2009, p. 907-921.

Marichal R., Rebe I. et Treton R., « La Transformation du milieu géomorphologique de la plaine du Roussillon et ses conséquences sur son occupation: premiers résultats. La Dynamique des paysages protohistoriques, antiques, médiévaux et modernes », in Burnouf J., Bravard J.-P., Chouquer G., XVIIIe Rencontres internationales d’archéologie et d’histoire d’Antibes, actes des rencontres, Sophia-Antipolis, 1997, p. 272-283.

Morellon M., Perez-Sanz A., Corella J.-P., Buntgen U., Catalan J., Gonzalez-Samperiz P., Gonzalez-Trueba J.J., Lopez-Saes J.A., Moreno A., Pla-Rabes S., Saz-Sanchez M.A., Scussolini P., Serrano E., Streinhilber F., Stefanova V., Vegas-Vilarrubia T. et Valero-Garces B., «A multi-proxy perspective on millennium-long climate variability in the Southern Pyrenees», Clim. Past, no 8, 2012, p. 683-700.

Mussot R. et Benech C., « L’influence des interventions humaines sur l’écoulement et sur le transport solide. L’exemple des Pyrénées-Orientales (France) », Annales de Géographie, no 581-582, 1995, p. 105-118.

Passarrius O., Carozza J.-M., Kotarba J., Illes P. et Porr a-Ketuni V., Elargissement de la RD83. Communes de Saint-Laurent-de-la-Salanque et du Barcarès, DFS, DRAC languedoc-Roussillon, 2010.

Puig C., « Les prémices du Petit Âge Glaciaire en Roussillon à travers le prisme des sources écrites », Archéologie du Midi Médiéval, no 27, 2009, p. 191-205.

Puig C., Les campagnes roussillonnaises au Moyen Âge : dynamiques agricoles et paysagères entre le xiie et la première moitié du xive s., Thèse, Université Toulouse 2, 2003, 4 vol.

Serrat P., Genèse et dynamique d’un système fluvial méditerranéen: le bassin de l’Agly (France), Thèse, Université de Perpignan, 1999.

Treton R., « Crues et inondations dans les Pyrénées Méditerranéennes aux xive et xve siècles: état des sources et perspectives de recherches », Domitia, no 8-9, 2007, p. 213-226.

Verdeil P., « Données nouvelles sur le Quaternaire de la basse vallée de l’Aude », Bull. Soc. Geol. Fr., no 7, 1970, p. 413-425.

Vita-Finzi C., The Mediterranean Valleys: geological changes in Historical Times, Cambridge University Press, 1969.

Xoplaqui E., Fleitmann D., Diaz H., von Gunten L. et Kiefer T., «Medieval Climate Anomaly», Page, volume 19, no 1, 2011, p. 1-40.

Table des illustrations

Légende Fig. 1. Co-occurrence of “geoarchaeology” and “Middle Age” or “Mediaeval” in the titles, keywords and abstracts of the following databases: ISI Web of Knowledge, Science Direct, Springer, and Ingenta. Duplications have been removed.
Fichier image/jpeg, 62k
Légende Fig. 2. Archaeological sites dating from the last millennium in the lowlands from the National Archaeological Map (from Service Régional de l’Archéologie Languedoc-Roussillon).
Fichier image/jpeg, 366k
Légende Fig. 3. Stratigraphy of the Sainte-Eugénie-de-Tresmall early Roman chapel.
Fichier image/jpeg, 155k
Fichier image/jpeg, 851k
Légende Fig. 4. The Agly course changes according to geoarchaeological, archaeological and historical data.Fig. 5. Alluvium plots limits in Alenya (end of the 14th c.).
Fichier image/jpeg, 218k


© CNRS Éditions, 2015

Conditions d’utilisation :

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search