Vous l’avez sans doute déjà repéré : sur la plateforme OpenEdition Books, une nouvelle interface vient d’être mise en ligne.
En cas d’anomalies au cours de votre navigation, vous pouvez nous les signaler par mail à l’adresse feedback[at]openedition[point]org.

Précédent Suivant

Chapter 17. Sovjan and Lake Maliq (Albania)

A protohistoric site in its environment

p. 235-244

Texte intégral


1The southwestern Balkans are scattered with numerous lakes located within mountain basins typical of the area, and they have always attracted people from as early as Prehistory. The most common type of lake (Fouache, 2006) are large lakes like those of Ochrid and Prespa our Jannina, located in grabens that evolved in poljes over the Plio-Quaternary. With large limestone sides, these lakes are deep – they can be as deep as 280 m such as lake Ochrid, which is one of the deepest lakes in Europe. The also typically receive little alluvium and allocthonous organic matter. The relative stability of the level of these lakes together with the lack of development of the lake shore terraces may account for the fact that so these terraces have not revealed evidence of particularly dense prehistoric habitations.

2However, this is not the case with the numerous smaller, shallower marshy lakes like Lake Maliq, which, until about 1960, occupied the northern part of the basin of Korçë (Albania), or the lake that used to be at the foot of the prehistoric tell of Dikili-Tash in the plain of Drama in western Macedonia (Greece). In fact, such wet environments, which may appear at first sight to be less hospitable, seem to have been very attractive to prehistoric populations, as is shown by the discovery of several lake sites in the mid-twentieth century. Some of them have been investigated to a greater or lesses extent, such as Lake Maliq (Fig. 1) in the basin of Korçë, which was excavated by Albanian archaeologists in the 1960s. Another is the site of Dispilio on Lake Kastoria in northern Greece, where systematic excavations have been carried out for twenty years by a Greek team from the university of Thessaloniki. This list is far from complete, as shown by the fortuitous discovery of another lake settlement in the late 1980s less than 4 km to the northeast of Maliq near the modern village of Sovjan, as well by the more recent discovery of another similar kind of occupation site on the eastern shore of Lake Cheimaditis near the village of Anargyroi, 30 km to the south of Florina (Greece).


Fig. 1. Location of the studied area.


Fig. 2. Comparison of the reference chrono-stratigraphic sequence between Sovjan and other sites in the southern Balkans.

3The basin of Korçë is particularly rich in such discoveries since, out of the twenty sites or so that have so far revealed traces of more or less intermittent occupation between the Neolithic and the beginning of the Iron Age, more than half are located around the former Lake Maliq. Among those sites, at least three are “Palafittique” (pile-dwelling) – Maliq, which provides the reference chronostratigraphic sequence for the whole area (Fig. 2); Sovjan and Dunavec – while a fourth is known to lie on the south-western bank of the Great Prespa Lake near the villages of Kallamas and Goritsa (Albania).

The site of Sovjan: Research agenda and method

4Discovered fortuitously during the digging of a drainage canal, the site of Sovjan has been studied under a French-Albanian multidisciplinary research programme that combines archaeological exploration (jointly run by Gilles Touchais and Petrika Lera) and palaeoenvironmental studies (supervised by Éric Fouache). Sovjan is potentially the richest site for the study of the dynamics of human lakeshore occupation in the Balkans. Indeed, the variations in lake level and the extent of the lake together with the palaeoenvironmental information fossilised in the lake sediment provide one of the keys to the understanding of the history of relationships between man and the environment since the early Holocene.

5A local and regional geomorphological study was first carried out in parallel with the exploration of the site (Dufaure et al., 1999). The geomorphological map made it possible to understand the evolution of regional geomorphological dynamics since the end of the last glacial maximum and to identify the zones of sediment accumulation as well as the process of deposition. It also permitted us to identify possible locations for lake sediment archives. Then a drilling programme was carried out on lakes Maliq and Ochrid with three aims: (i) To reconstruct the vegetation dynamics along the transition between the Lateglacial and the Holocene (Denèfle et al., 2000); (ii) To reconstruct the climatic variability over the same time-scale (Bordon et al., 2009; Magny et al., 2009); and (iii) To reconstruct the variation in the level of Lake Maliq in relation to human occupation (Fouache et al., 2001, 2010). A preliminary drilling was carried out for palynological and palaeoclimatic studies It produced a sediment sequence covering the last 15000 years. A second, shorter, drilling, which covered the period of the site’s occupation was carried out near the tell of Sovjan. In order to reconstruct lake level variations, a transect of 10 drillings was carried out from the tell of Sovjan to the middle of the former lake.

6The reconstruction of past variations in lake level is based on changes in the lithology: silty lake-marl is deposited in lake water, whereas organic deposits (clay, coarse gyttja, peat, “anmoor”) reflect nearshore areas (eulittoral zone, littoral mire). The estimation of the past positions of the water table (expressed as elevation in metres a.s.l.) refers to the absolute elevation at which the main lithological markers have been observed in the sediment profiles.

The environment of the site

Geomorphological dynamics

7Lake Maliq is located where the graben of Korçë has experienced its greatest subsidence since the Pliocene. From the Holocene onwards, that is for 11650 years, global warming has contributed to the lake functioning as a peat bog so that the peat deposit reaches 10 metres in depth. The possible extension of the lake’s waters to the north of the basin is limited by the confluence of the Devoll (coming from the east) and the Dunavec (coming from the south), while the Devoll flows on towards the west to the Adriatic.

8Lake Maliq was thus a rather shallow lake, with a maximum of a few metres depth, all the more so since the presence of an outlet ensured that the overflow of the floods could be evacuated. The lake, which was drained in the early 1960s within the framework of a vast drainage plan, tended to be silted up by alluvium from the Devoll, which in times of flood carried high sediment loads from its upstream catchments, which are rich in terrigenous clastic rocks such as molasse. The flooded areas, however, could double in size from winter to the summer, as is depicted on the topographic map of 1948.

Climatic and vegetal variations

9While climatic variations were clearly experienced in the region, the average altitude of the basin, at 820 m a.s.l., the massive winter rainfall (600-800 mm yr-1), and sometimes snow, guaranteed the presence of water in Lake Maliq, even in dry years. These conditions account for the richness of the ecosystem, which is characterised by a high natural productivity, whether it be biomass or fauna. The pollen and spore diagram analysed from a 9.3 m coring into the peat of the former lake shows that those conditions were equally favourable during the last glacial maximum – when the basin was a refuge for tree species as well as during the first half of the Holocene, when a dense oak forest existed, which only began to be cleared in the Bronze Age. As far as the climate is concerned, the palynological studies made from a core into the sediments of Lake Maliq allowed the teams from the Laboratories of Climate and Environment Studies (Gif-sur-Yvette) and of Chrono-Ecology (Besançon and Dijon) to reconstruct temperature and rainfall curves over the past 15000 years (Fig. 4). Those studies reveal that while the average seasonal temperature has been rather constant from 10000 years BP, the average rainfall has been more variable, especially in its seasonal distribution, and this is probably the primary cause for lake level variation (Bordon et al., 2009).

The site of Sovjan and its occupation

The chrono-stratigraphic sequence of the site

10The protohistoric habitation site is made up of a very low, elliptic-shaped tell of a very modest surface area (about 2.5 ha; Fig. 5). The excavations have so far centred on the southeastern quarter of the tell (Sector A) and have allowed us to recognise a sequence of 13 anthropogenic levels corresponding to several main periods of occupation (Fig. 2). The earliest settlement (Level 13) apparently dates back from the very Early Neolithic (around 7000 BC according to the radiocarbon dating) while the most recent level (Level 5), which precedes the final abandonment of the site, dates back to approximately 700 BC. At the time, a rise in the lake level caused the inundation of the site, which remained submerged until the early 1960s when the lake was largely dried out. During this long period of submersion, which lasted almost 2000 years, a thick layer of peat was formed at the bottom of the lake (Level 4), sealing the archaeological levels. This layer of peat is itself covered with sterile sediments (Level 3: settled-out (decantation) clay; Level 2: fluvial deposits while Level 1 is made up of rubble from the canal.


Fig. 3. Geomorphological map of the basin of Korçë.


Fig. 4. Reconstruction of the curves of palaeotemperatures and palaeorainfall on the basin of Korçë. Done by Bordon et al. (2009) from a pollen diagram of Lake Maliq.

11The chronology of the levels of occupation (Levels 5 to 14) rests both on archaeological data and on a series of 35 radiocarbon dates, some of which still need to be confirmed. Dendrochronology hasn’t yet provided absolute dates but the analysis of 1300 wood samples taken from the site already provides a clearly and continuous relative sequence covering more than four centuries.

12The most interesting levels from an archaeological point of view belong to the Bronze Age. They correspond to Levels 7 to 9 attributed to the Middle Bronze (Maliq IIIc phase) but they do not seem to be earlier than the last centuries of the 3rd millennium (the end of the Early Bronze Age) and also Levels 5 and 6, dating back to the Late Bronze (Maliq IIId phase, second half of the 2nd millennium BC). The early phases of the Early Bronze so far are only represented by a few pieces of ceramics at Sovjan while the Chalcolithic and the Late Neolithic are completely absent from the site.

Structures of the settlement

13The site, which was buried by a high lake level when it was abandoned, allowed numerous remains of wooden structures to be preserved, which is one of the main interests of the site lies.

14In the Neolithic levels, which have only been excavated by sample cores over a very small surface area, those remains are limited to a few fragments of oak or ash-tree planks, fragments of poles and perforated soles of wood piles. However, only some of timber remains were in place (in situ).

15The best-preserved structures, both horizontally and vertically, often in situ, are to be found in the Middle Bronze levels. Apart from several floors, which differ mainly in the choice of wood varieties used (Fig. 6), and innumerable stakes (almost all of which of oak), the most interesting structure is a large apsidal house of more than 15 m in length and 4 m in width. This house, which was nicknamed the ‘Canal House’, had wattle and daub walls preserved to a height of about 50 cm (Fig. 6). Inside the house to the north a load-bearing wall separates the small apse room from the main room to the south, which people used to enter through a door, which opened through the long eastern wall along the lake. A large bark bed several metres square, which also probably served as insulation, lay beneath the ground of the main room. Less than 2 metres to the north of the house, a floor of logs laid on long cross-ties was interpreted as a path. Across this path on the other side lay another house named the Fisherman’s House, of which only the door and a small segment of wattle and daub wall have been uncovered. This not only provides us with preliminary data on the construction techniques and the ground plan of the houses, but also on the arrangement of the domestic space.

The site and its relationship with its environment

16Encouraged by the good preservation of organic matter, a first for this region, the palaeoenvironmental studies allowed us to assess the impact of anthropogenic activities on the local ecosystem. Although the results are preliminary, they already provide precious information both on the faunal environment and the vegetational environment. Generally speaking, it appears that, throughout the 2nd millennium, the impact of human activities was characterised by relatively important use of the natural environment but no significant alteration of this environment, unlike what can be observed in the Aegean region, for instance. This means that beside a production economy a rather intense economy of hunting was maintained, but their combined efect on the environment remained limited.


Fig. 5. Map of the excavations on the site of Sovjan.


Fig. 6. Photograph of the “Canal House” dated Middle Bronze.

The faunal environment

17The initial results of the study of faunal remains collected at Sovjan showed that, while the meat supply comes essentially from the exploitation of domestic animals – mainly cows, sheep/goats and pigs, but also dogs were eaten –, the wild fauna also provided a substantial input of meat until the end of the Bronze Age. Among these mammals are Cervidae and Suidae, which make up 8 to 9% of the total remains, to which should be added birds – almost all of them water birds such as grey herons, moorhens or Eurasian coots – and ish – chub, amur bitterlings (small carp) and common roaches — which were found in huge numbers in several areas of the site and particularly in the Fisherman’s House, hence its name. The study of these remains showed that most of them were the remains from cooking (Gardeisen et al., 2002).

The vegetational environment

The exploitation of forest resources

18In the present state of the dendrochronological study, only preliminary observations are possible. First, as conirmed by the palynological data (Denèle et al., 2000; Fouache et al., 2001), we note a massive predominance of oak trees, as early as the Neolithic. In the Middle Bronze, many oak trees used by the inhabitants of Sovjan to build their loors were more than 350 years old when they were felled. Thus they most likely came from the primary forests highlighted by the pollen diagram (cf. supra).

19Together with oak trees, the protohistoric population also used a number of non-dominant species as timber, mostly at the beginning of the Middle Bronze. Most of them are species from wet environment such as alder, ash or willow, which must have been abundant around the site.

20Lastly, pine, which belongs to the vegetation of the foothills bordering the plain, was only used for construction at the end of the Middle Bronze and mostly in the Late Bronze.


21As an indication of the use of the surrounding environment for food we have identiied and listed several berries and wild fruit such as hazelnuts, beechnuts blackberries, raspberries, Cornelian cherries, elder berries, apples and pears but also, in the Middle Bronze levels, rather high concentrations of acorns (Allen, 2002). As the acorns were mostly roasted we can assume that they were eaten by the inhabitants. However, they become very rare in the Late Bronze levels, which may be explained either by a change in the ecosystem or by a change in the food preferences, but at present we have no way of deciding between these two possibilities.


22In this field too, the palaeoethnobotanical study has already allows us to understand some change during the Bronze Age, both concerning cultivated species and cultivated spaces.

23As far as cereals are concerned, einkorn wheat (Triticum monococcum) dominated throughout the Bronze Age. But while in the Middle Bronze we also found emmer wheat (Triticum dicoccum), barley (Hordeum vulgare distichon) and millet (Panicum miliaceum), two new species appeared in the Late Bronze: spelt wheat (Triticum spelta) and common wheat (Triticum aestivum). The spectrum of legumes also broadened in the Late Bronze, with chickpeas (Cicer arietinum) being added to lentils (Lens culinaris) and bitter vetch (Vicia ervilia) both of which had been cultivated earlier (Allen, 2003).

24Concerning the agricultural landscape and its variations during the Bronze Age, information was provided by examining the herbaceous plants associated with cereals. It is indeed noticeable that in the Middle Bronze Age these plants are from a wet environment whilst in the Late Bronze Age they are more varied, including Gramineae that can resist more arid conditions (such as ryegrass). This seems to suggest that in the Middle Bronze Age – as probably in the Neolithic, a period for which we unfortunately have no palaeobotanical data, the cultivated area was in fact limited to the immediate surroundings of the site, while in the Late Bronze Age it extended to the foot of the hillslopes (Allen, 2004) but probably not beyond, judging from the absence of important terrigenous inputs in the Late Bronze levels. It is only after the Iron Age that the archaeological levels were covered by a thick, clayed layer, which is indicative of the erosion of the slopes following the cultivation of the molasse hills. It is difficult to date this phenomenon precisely but studies carried out in other Balkan regions showed that it did not occur prior to the Roman era, or even the Mediaeval or post-Mediaeval times (Fouache, 2002). Thus it seems that while forest clearing starts in the Neolithic in the basin of Korçë, when sedentary settlements are established around the lake, it remains quite limited until the end of the Middle Bronze as the inhabitants only cultivated the wet zone near the site until the middle of the 2nd millennium BC. This zone may vary in its extent due to changes in climate but it never disappeared completely, at least for part of the year, even in periods when the lake level was particularly low. This is a mode of exploitation of the land that was common with the Neolithic populations in the plain of Drama, as was recently revealed by Lespez (2003). For contrary to what is commonly thought, it appears that the Neolithic people from the north Aegean regions did not cultivate “light soils” but, on the contrary, very heavy, naturally fertile soils such as those of the marshy areas that were constantly wet (Dufaure, 2003).

25From this perspective, one aspect of the Sovjan excavations will be to demonstrate the survival of this north Aegean “model” until the end of the 2nd millennium whilst in the south Aegean the exploitation of the land was marked by massive forest clearings as early as the 3rd millennium if not a little earlier, and which have had a long-lastingly mark on the landscape (Zangger, 1992; Willis, 1994). This contrasting record of agricultural practices in the northern regions is in keeping with what can be observed in Sovjan as well as in Maliq in every domain related to material culture, whether it be ceramic, tools or settlement form (Prendi, 1966; Touchais, 2008). In other words, the Ancient Bronze Age and the Middle Bronze Age in southern Albania presents characteristics that could be defined as sub-Neolithic, and it would be interesting to know the geographic extent of this phenomenon.

The impact of the environment on the evolution of the site

26The results of the corings carried out in the basin, particularly in the area of the former lake Maliq and near the site of Sovjan, together with the data collected from the archaeological excavations allowed us to propose a reconstruction of variations in the lake’s level and to observe their impact on the habitation site of Sovjan throughout its occupation from the beginning of the Neolithic to the beginning of the Iron Age. The results have also and more generally provided data on the dynamics of population change over the whole basin of Korçë from the end of the Lateglacial to modern times.

27Three specific periods in the sequence are particularly remarkable when it comes to relating the dynamics of occupation to variations in the level of the lake.


Fig. 7. Variations in the level of Lake Maliq in relation with human occupation of the site of Sovjan.

28The first period occurred right at the beginning, indeed it seems that the silt layer reached at the bottom of the sequence (Level 14) corresponds to the level of the ancient water table. This layer occurs everywhere beneath the oldest level of occupation (Level 13) and suggests that people settled in Sovjan at a time when the surface area of the lake was reduced due to a low lake level. This settlement phase dating back to the very beginning of the Neolithic (from radiocarbon and a few ceramics with an impresso decoration sheds some light on the issue of neolithisation in the southwest of the Balkans. The presence of this settlement, of which scattered pieces of construction have been found, at least proves that lakeside habitats did exist near villages established on firm ground, such as Podgorie or Vashtëmi, in the basin of Korçë as early as the initial phase of the Neolithic.

29The second period is the Late Neolithic, Chalcolithic and beginning of the Ancient Bronze, which provides no occupation evidence at the site. It is possible, however, such levels exist in areas of the site that have not been explored yet where the habitation could have moved temporarily. But as long as they have not been found, we must consider the possibility of a gap, which would then be between 5000 and 3000-2500 years BC. At the time, the site of Maliq, less than 4 km away, was flourishing, as shown by the numerous stakes of the Chalcolithic site (phases IIa and IIb), the image of which has long been emblematic of Albanian prehistory. Considering the closeness of the two sites, it is clear that the natural conditions were the same. If then the continuation of research at Sovjan confirms this gap, the cause would have to be found in reasons other than the natural environment.

30The third period of particular interest corresponds with the abandonment of the site, which was submerged at the beginning of the Iron Age when the lake was most extended and it even reached the foot of Mali Thate. The peat layer (Level 4) reflects to this phase of submersion, which sealed the underlying archaeological levels.


31Given the present state of research, we can say that the site of Sovjan was occupied continuously between the end of the 3rd millennium and the 7th c. BC. The sequence of Maliq is very similar and partly complementary. Thus today at least two lake sites with long stratigraphies are known from southern Albania, which shows that this type of habitat persisted for several millennia until the historic period in the southwest of the Balkans. This environmental situation is actually evoked in a famous passage from Herodotus concerning the tribes of the region of Pangaea and lake Prasias, some of which, according to the historian, “build their dwellings on the lake itself… on a platform that rests on raised stakes connected to the firm land only by a narrow walkway” (V, 16).

32Since 2007, the new PALM research programme (Prospection Archéologique du Lac Maliq – Archaeological Prospection of Lake Maliq) has enabled us to locate almost sixty new sites around the lake and in the piedmont area, whose chronology now extends from the Neolithic to the Mediaeval period and whose establishment is linked to variations in the extent of the lake’s surface area, as at Sovjan. So we now know that the basin was occupied continuously from the beginning of the Holocene to the present day.


Des DOI sont automatiquement ajoutés aux références bibliographiques par Bilbo, l’outil d’annotation bibliographique d’OpenEdition. Ces références bibliographiques peuvent être téléchargées dans les formats APA, Chicago et MLA.


Allen S.E., «The palaeoethnobotany of Sovjan, Albania: Preliminary results», in Touchais G., Renard J., L’Albanie dans l’Europe préhistorique. Actes du colloque international organisé par l’École française d’Athènes et l’université de Bretagne-Sud, Lorient, 8-10 juin 2000, BCH Suppl. 42, 2002.

Allen S.E., «Prehistoric wetland agriculture at Sovjan, Albania», Paper presented at the 68th Annual Meeting of the Society for American Archaeology, Milwaukee, Wisconsin, April 10-13, 2003.

Allen S.E., «Food, fodder and farming in Southern Albania: Palaeoethnobotanical research at Sovjan», Paper presented at the 105th Annual Meeting of the Archaeological Institute of America, San Francisco, California, January 2-5, 2004.

10.1016/j.quaint.2008.05.014 :

Bordon A., Peyron O., Lézine A.-M., Brewer S., Fouache E., «Pollen-inferred Late-Glacial and Holocene climate in Southern Balkans (lake Maliq)», Quaternary International, 200, 1-2, 2009, p. 19-30.

10.1006/qres.2000.2179 :

Denèfle M., Lézine A.-M., Fouache E., Dufaure J.-J., «A 12,000 years pollen record from Lake Maliq, Albania», Quaternary Research, 54, 2000, p. 423-432.

10.3406/edyte.2003.860 :

Dufaure J.-J., «Regard sur les bassins intramontagnards sud-balkaniques – morphogénèse, milieux, sociétés», in Dynamique et vulnérabilités des milieux montagnards méditerranéens et alpins, Mélanges offerts au professeur René Lhenaff, Collection EDYTEM – Cahiers de Géographie, 1, 2003.

Dufaure J.-J., Fouache E., Denèfle M., « Tectonique et évolution géomorphologique: l’exemple du bassin de Korçë (Albanie) », Géomorphologie : relief, processus, environnement, 2, 1999, p. 111-128.

Fouache E., « Dynamiques paléoenvironnementales en Albanie à l’Holocène », in Touchais G., Renard J., L’Albanie dans l’Europe préhistorique, Actes du colloque international organisé par l’École française d’Athènes et l’université de Bretagne-Sud, Lorient, 8-10 juin 2000, BCH Suppl. 42, 2002.

Fouache E., « 10 000 d’évolution des paysages en Adriatique et en Méditerranée Orientale », Travaux de la Maison de l’Orient Méditerranéen (TOM), vol. 45, 2006, Lyon, diff. De Boccard, Paris.

10.1007/PL00006922 :

Fouache E., Dufaure J.-J., Denèfle M., Lézine A.-M., Léra P., Touchais G., «Man and environment around Lake Maliq (southern Albania) during the late Holocene», Vegetation History and Archaebotany, 10-2, 2001, p. 79-86.

Fouache E., Desruelles S., Magny M., Bordon A., Oberweiler C., Coussot C., Touchais G., Lera P., Lézine A.-M., Fadin L., Roger R., «Palaeogeographical reconstructions of Lake Maliq (Korçë Basin, Albania) between 14000 BP and 2000 BP», Journal of Archaeological Science, 37-3, 2010, p. 525-535.

Gardeisen A., Garcia Petit L., Piquès G., « La recherche archéozoologique en Albanie: un état de la question à Sovjan (bassin de Korçë) », in Touchais G., Renard J., L’Albanie dans l’Europe préhistorique, Actes du colloque international organisé par l’École française d’Athènes et l’Université de Bretagne-Sud, Lorient, 8-10 juin 2000, BCH Suppl. 42, 2002.

10.1016/S0341-8162(02)00164-9 :

Lespez L., «Geomorphic responses to long-term land use changes in Eastern Macedonia (Greece)», Catena, 51, 2003, p. 181-208.

10.1177/0959683609337360 :

Magny M., Vanniere B., Zanchetta G., Fouache E., Touchais G., Petrika L., Coussot C., Walter-Simonnet A.-V., Arnaud F., «Possible complexity of the climatic event around 4300-3800 cal. BP in the central and western Mediterranean», The Holocene, 19, 2009, p. 823-833.

Prendi F., «La civilisation préhistorique de Maliq», Studia Albanica, 3-1, 1966, p. 255-280.

Touchais G., « Sovjan et l’Âge du Bronze en Albanie », in Guilaine J., Villes, villages, campagnes de l’Âge du Bronze, Séminaire du Collège de France, Errance, Paris, 2008.

Touchais G., Fouache E., « La dynamique des occupations de bord de lac dans le sud-ouest des Balkans : l’exemple de Sovjan, bassin de Korçë (Albanie) », in Richard H., Magny M., Mordant C., Environnements et cultures à l’Âge du Bronze en Europe occidentale, Actes du 129e Congrès national des sociétés historiques et scientifiques, Besançon, 19-24 avril 2004, PUB, Besançon, 2007.

10.1016/0277-3791(94)90104-X :

Willis K.J., «The vegetational history of the Balkans», Quaternary Science Reviews, 13, 1994, p. 769-788.

Zangger E., «Neolithic to Present soil erosion in Greece», in Boardman J., Bell M., Past and Present Soil Érosion, Oxbow, Oxford 1992.

Précédent Suivant

Le texte seul est utilisable sous licence Licence OpenEdition Books. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.