Montjean and Maravielle and the Massalian connection

(1)

Montjean and Maravielle and the

Massalian connection

The effects of Greek colonisation on the fortification of indigenous settlements

in Mediterranean France during the Iron Age

Joost Pluijmen 27-07-2019

(2)

(3)

3 Montjean and Maravielle and the Massalian connection

The effects of Greek colonisation on the fortification of indigenous settlements in Mediterranean France during the Iron Age

By Joost Pluijmen

Master Thesis

MA Archaeology: Mediterranean Archaeology University of Amsterdam

VU University Amsterdam

Supervisor: Vladimir Stissi Second reader: Patricia Lulof

(4)

(5)

5

Preface

Every summer, my parents and I drove from the Netherlands to the south of France to visit my grandparents. During these trips we would visit a lot of archaeological sites, castles, Roman excava-tions etc. It was then that I fell in love with history and especially the history of an undiscovered past. After graduating from high school I went to study archaeology. Not only did I gain a good in-sight into the field of study, but I also became aware of the tremendous work that archaeologists need to do in order to find a small piece of the puzzle.

By visiting my grandparents in the historically important Provence region, I became interest-ed in the history and archaeology of the region surrounding their village, La Môle, in the Var de-partment. I discovered that two Iron Age sites called Montjean and Maravielle, named after the hilltops on which they are situated, are located within the municipal territory of La Môle. After some research it became clear that these two fortified sites or ‘oppida’ had not yet been subject to exten-sive academic research, except from the archaeologists who had excavated the sites. I discovered that the scholars who referred to these sites did not have a full overview of the results these sites had yielded. Therefore I decided to dedicate my thesis to these two sites.

I want to specially thank Esin and my family who encouraged me to continue with my research in times of struggle, and professor Vladimir Stissi who handed me the opportunity to write my thesis on this subject. Finally, I want to thank Jitte Waagen who helped me with the use of Arcmap and Gisèle deWaelsche, Digitisation Manager from the Direction régionale des affaires culturelles PACA, who gave access to the archaeological reports.

(8)

8

1. Introduction

Around 600 BC, present-day Marseille was founded by colonists from Phokaia who called it ‘Mas-salia’. Phokaia was an Ionian city nowadays located on the west coast of Turkey. Historical sources tell different stories about the founding of Massalia, but what is certain is that this foundation played an important role in the history of France. The indigenous population of Mediterranean France, the southern part of France, consisted of different tribes, of which two main tribes are known as the Celts and Ligures. Historical sources mention a difficult and sometimes violent rela-tionship between different tribes and Massalia (Dietler, 2010, pp. 4-8). Massalia remained an inde-pendent Phokaian colony until the Roman conquest around 125 BC, after which it became part of the Roman Empire. During the Phokaian period, Massalia founded some of its own colonies, with whom the Massalians would trade with the indigenous population.

The foundation of Massalia took place during a period which is now generally known as the ‘Iron Age’. Many archaeological sites have revealed layers of occupation in the Iron Age, but have only been partially studied due to the limits of archaeological research. The discoveries of Greek and Massalian pottery in Iron Age sites all over France triggered a debate (Dietler, 2005, pp. 15-24) about how the Phokaian colonists were connected to the surrounding indigenous tribes in terms of trade, intermarriage and war. Especially sites in Mediterranean France have shown traces of such a ‘Mas-salian connection’. A significant part of these sites were fortified during this period. The oppida of Montjean and Maravielle are two of these sites. The presence of Massalian, but also Greek and Etruscan, ceramics, dates these two sites to a period shortly after the establishment of Massalia. The close vicinity of the oppida of only 2,8 kilometres while facing each other on both sides of the same valley offers a quite unique situation. Moreover, the differences in style of fortification within the same period of occupation adds a notable aspect to this situation. The problem, however, is that the extensive excavations of the oppida were never fully published.

Aim

This thesis aims to place the oppida of Montjean and Maravielle in a historical and geographical con-text of Massalian influence in Mediterranean France during the Iron Age. The concepts ‘oppidum’, ‘Mediterranean France’ and ‘Iron Age’ respectively form the object, the region and the period of study and are subsequently provided with a clear context.

The period of study: the Iron Age

The period of study, the ‘Iron Age’, is the broadest level of analysis which serves the essential focus of the study. Different scholars have labelled different periods subdividing the Iron Age (Dietler, 2005, pp. 30-31). The duration of the Iron Age varies depending on the region of study. It is generally defined by the period in which the production of iron or steel was brought to the point where iron tools and weapons that were superior to bronze tools and weapons became widespread.

For the purpose of this thesis, the period that is meant with ‘Iron Age’ is limited to the peri-od of simultaneous occupation of the oppida of Montjean and Maravielle. Therefore, an artificial period of a few centuries from 650 to 350 BC forms the ‘Iron Age’. The period before the occupation of Montjean and Maravielle and after the appearance of the first oppida from approximately 650 to 600 BC is included to the period of study, because this period can be used to compare the occupa-tion of Montjean and Maravielle with the occupaoccupa-tion of the first oppida. As menoccupa-tioned before, an important event that took place within the boundaries of this period is the founding of Massalia by about 600 BC.

(9)

9 The concept ‘Mediterranean France’ denotes the region of study and mostly follows the definition of the ‘lower Rhône basin’ by Michael Dietler (Dietler, 2005, pp. 9-13). To understand the context of Mediterranean France, it is useful to take a look at a geographical height map of France, see Figure 2. The map shows the mountain ranges of the Massif Central and the Alps that are vertically sepa-rated by the Rhône river. These mountain ranges form, together with the Mediterranean coast, the shape of a triangle with lower lying terrain that roughly covers the river delta of the Rhône.

Dietler defines his concept of ‘lower Rhône basin’ on the basis of a more or less common material culture. Whereas the Alps and the Massif Central form clear natural boundaries, Dietler states that the sites along the Hérault river form a material culture boundary between sites in the western Languedoc and the eastern Languedoc. The material culture of the sites in the eastern Languedoc shows more similarities with the sites east of the Rhône than with the sites in the west-ern Languedoc. The narrowest passage between the Alps and the Massif Central forms the northwest-ern boundary. Dietler does not define a clear eastern boundary as he decides to include sites along the Côte-d’Azur as far as Nice. He places the Rhône central in his study because it is seen as the main communication corridor between the Mediterranean coast and West-Central Europe or the Hallstatt zone. Dietler suggests that the reason Massalia was founded near the eastern river mouth of the Rhône was to have access to the tin supplies in the northern Hallstatt zone.

For the purpose of this thesis, the concept of ‘Mediterranean France’ is based on the defini-tion of Dietler of the lower Rhône basin. The region that Dietler defined, can be marked with a circle with a radius of 150 kilometres around Massalia. In this way, the river mouths of the Hérault and the Var form the exact natural boundaries to the west and the east. The northern boundary of Dietler is also used in this thesis although it is more clearly defined by the oppidum of Malpas in Soyons. To make a distinction between impenetrable terrain and good accessible terrain the boundaries of the Alps and the Massif Central are here defined by an artificial boundary at a height of 750 metres above sea level, see Figure 3.

(10)

10 The object of study: the oppidum

At the beginning of the Iron Age there was a sudden explosive growth in the fortification of sites. For the purpose of this thesis, these sites are referred to as ‘oppida’. These oppida existed with different types of fortification systems and in different sizes. Not all excavated oppida show signs of perma-nent inhabitation in the form of stone or mudbrick walls, which does not mean that these oppida were never permanently inhabited. Whereas the size of the oppida does not depend specifically on its natural location, the fortification system that protects it does depend on it. A part of these sites were already occupied before they were fortified, suggesting a sudden need for fortification. By 350 BC, oppida formed a very significant part of the Mediterranean French landscape, whilst some of the earliest oppida like Montjean and Maravielle seem to have been abandoned before this period.

Figure 3. The region of study: Mediterranean France.

Studies

In order to achieve te aim of this thesis to place the oppida of Montjean and Maravielle in a histori-cal and geographihistori-cal context of Massalian influence in Mediterranean France during the Iron Age, three independent studies are combined:

Study 1

The first study comprises a review of the archaeological research at the oppida of Montjean and Maravielle. The main question in this study is:

‘Can geographical and archaeological features explain the simultaneous occupation of the oppida of Montjean and Maravielle?’

The answer this question, different aspects of the oppida, which can each attribute from their own perspective, are examined. The history of the archaeological research shows the development of

(11)

11 how researchers considered the dating and function of the two sites. The geographical properties of the sites give an insight in how the sites can be regarded in connection with their environment. The habitational and fortification structures as well as ceramic assemblages of the two sites tell some-thing about their function and dating and show similarities and differences between them. This study will mostly be based on the reports and publications on the oppida. The study contributes to the aim of this thesis by presenting an archaeological context of the oppida and an image of coexist-ence of the two sites.

Study 2

The second study comprises a historical and archaeological study of the Massalian colonisation of Mediterranean France. The main question in this study is:

‘What information do the classical authors provide on the location of Greek and Massalian colonies, and can these locations and their dating be confirmed by archaeological evidence?’ To answer this question, first, all passages in classical texts that refer to Greek and Massalian colo-nies are examined in order to localise the colocolo-nies that were located in Mediterranean France. The Massalian colonies that can be localised within Mediterranean France will be matched to archaeo-logical sites that are located in the vicinity of the assumed locations of the colonies. These matches are based on the classical texts and linguistic and archaeological evidence. This study will contribute to the aim of this thesis by providing a context of colonisation for the oppida that were occupied during the same period.

In an additional study, the steps above are undertaken to localise settlements in the Massif des Maures, the mountain chain in which Montjean and Maravielle are located, that were not necessari-ly described as Massalian colonies. This extra study is undertaken to see if these other settlements can tell more about the function and occupation of the oppida of Montjean and Maravielle.

Study 3

The third study comprises a quantitative study of the archaeological and geographical characteristics of oppida in Mediterranean France. The main question in this study is:

‘What can ceramic assemblages, fortification structures and geographical properties of op-pida tell us about Greek influence in Mediterranean France during the Iron Age?’

To answer this question, the named characteristics will be compared in varying manners in order to find differences and similarities between the oppida. For the purpose of this study, the geospatial processing program ArcMap has been used to visualise the characteristics of the different oppida in distrubution maps. In these distribution maps, the features of the different oppida are first com-pared individually. Different features are then combined and comcom-pared to find out if multiple as-pects can reveal patterns in the distribution of the features of the oppida. Furthermore, the oppida are placed in a context of the colonies that were localised in study 2, to see if there is a connection between the locations of the oppida and the locations of the colonies. This study creates a context of the occupation of oppida throughout Mediterranean France. It contributes to the aim of this thesis by placing the oppida of Montjean and Maravielle in this context and comparing them with the other oppida.

The results of the three studies will provide a subsidiary context for the existence of Montjean and Maravielle. This means that the first study focusses on a very local scale on the archaeological and geographical features of Montjean and Maravielle. The second study on the other hand, examines the period of colonisation from an ancient Greek and Roman historical ‘global’ view, in order to find

(12)

12 meaning in the colonisation of the whole region of Mediterranean France. The last study will then regard Mediterranean France from an indigenous perspective and examines if the fortification of settlements during the period of study can be connected to an influx of Greek and Massalian colo-nists. These three different perspectives will then be combined to find a meaning of the fortification of the settlements during a time that Massalia had an impact on Mediterranean France.

Structure

The remaining part of this thesis will focus on the three studies as specified above. Chapter 2 offers a detailed study on the oppida of Montjean and Maravielle. The study includes a geographical descrip-tion, a history of research, the details of the excavations of the sites and an analysis of the archaeo-logical data that have been assembled during those excavations. Chapter 3 reviews all surviving clas-sical texts that refer to Massalia and its colonies in relation to Mediterranean France. These colonies will, if possible, be localised geographically. Archaeological evidence at the geographic locations may subsequently provide a dating to these colonies. In addition, sites that can be localised near the Massif des Maures will be archaeologically reviewed. Chapter 4 focusses on the comparison of the different oppida in Mediterranean France. First, the outline of the methodology that is used in the comparative research of the different oppida will be explained. The second part focusses on the results of the analysis of all the archaeological data with the help of distribution maps, displaying for example the spread of imported ceramics throughout the region. Chapter 5 focusses on the discus-sion and concludiscus-sion of this thesis in which the results of the three studies are combined in order to work out the aim of this thesis.

(13)

13

1. Montjean and Maravielle

This chapter covers the first study of this thesis; a review of the archaeological research on the oppi-da of Montjean and Maravielle. The study examines the more or less simultaneous occupation of the oppida. Most of the reviewed archaeological research was conducted by Denis Wallon, who super-vised the excavations of both sites, see Appendix A: Excavations of Montjean and Maravielle. The first section describes the geographical context of the two oppida. The second section examines the history of research of the oppida. The fourth section provides an overview of the archaeological sites and presents the archaeological contexts. The fifth section contains a description of the assembled ceramic material and presents a comparative analysis of this material. The sixth section focusses on the functioning of the oppida and their simultaneous occupation according to Wallon. The chapter ends with a conclusion that attempts to answer the main question of this study; can geographical and archaeological features explain the simultaneous occupation of the oppida of Montjean and Maravielle?’

1.1. Geographical context

1.1.1. Massif des Maures

Figure 4. Geographic overview of the Massif des Maures.

The region in which Montjean and Maravielle are located is known as the Massif des Maures. The Massif des Maures is a small crystalline mountain range that stretches along the Mediterranean coast between Hyères and Fréjus, see Figure 4 and Figure 5. The Massif is shaped like a horseshoe, whose open side faces east, with an approximate maximum length and width of 60 km and 30 km respectively, see Figure 6. It is limited by the Maures plain in the north, the Argens river in the east,

(14)

14 the Mediterrenean coast in the south and the plain of the Gapeau river in the west. The ‘open’ part of the horseshoe consists of the Grimaud basin and the gulf of Saint-Tropez.

Figure 5. Topographic overview of Montjean and Maravielle.

The Massif forms a quite inaccessible area because of its geology and its almost impenetrable na-ture. The highest points of the northern branch of the Massif are the Signal de la Sauvette (780 m) and the Notre Dame des Anges (767 m), of which the latter can also be seen in Figure 6. While the hills do not exceed 780 metres, the passes lie rather high near steep ravines. For this reason, the main roads have always been led around the Massif following the Maures plain and the valley of the Argens from Toulon to Fréjus. During the Roman era, the Via Julia Augusta followed this trace be-tween Fréjus and Le Luc. Nowadays, there are only two national roads that pass the Massif from the Grimaud basin to the Argens valley, the D558 via La Garde-Freinet and the D25 from Sainte-Maxime towards Le Muy via the Col de Gratteloup. Additionally, there is only one national road that passes the Massif from east to west, the D98 via another Col de Gratteloup.

Wallon stated that the Massif contains around twenty sites like the oppida of Montjean and Maravielle (Wallon, 1984c) of which only a few have shown evidence of occupation during the Iron Age. Wallon proposed that the presence of abandoned terraces on the slopes of the hills carrying the oppida were cultivated by the populations that occupied the oppida. These terraces would then have been irrigated by the springs in the hills. Wallon emphasised that the Grimaud basin, that is now the most arable part of the Massif because of the confluence of the Môle and the Giscle rivers, was still marshy when Port-Grimaud was established in the 1960s. Port-Grimaud was built near the most inland point of the gulf of Saint-Tropez, where the Giscle and the Môle reach the sea. The presence of abandoned metal mines and terraces indicates that the population of the villages in the Massif was once significantly larger than the small population in 1910.

(15)

15 Grimaud basin  Col de Gratteloup (D25)  Le Muy  Nartuby river  Jabron river  Verdon river  Col d’Allos  Ubaye river  Drac river  Isère river

, of which there was no archaeological evidence for a comparable route north of Fréjus. He therefore wondered if this route was not a north-south transalpine route, such as the Rhône and the Durance rivers, that were already in use in the Iron Age and functioned in times of disputes between indige-nous peoples and the Massalians.

Figure 6. Overview of Montjean and Maravielle (Wallon, 1984c, p. 6; Wallon, 1988). 1.1.2. Montjean

The oppidum of Montjean lies at an altitude of 460 metres on the summit of one of the highest peaks of the coastal chain of the Massif des Maures, Le Montjean. To the southeast, the oppidum dominates the bay of Cavalaire-sur-Mer at a distance of 3 kilometres to the beach which extends over a distance of 2 kilometres. The advantage of the bay of Cavalaire-sur-Mer is the proximity to several springs, including the Eau Blanche creek at the foot of Montjean. Additionally, the beach is located between two inhospitable coastlines to the east and the west. On the other side of the hill, the oppidum overlooks the valley of the Môle, the oppidum of Maravielle and the Grimaud basin, and has a direct view of the summits of the northern branch of the Massif of which some carry other oppida like San Peire and Peigros. To the east, the oppidum has a good view on the Italian border and the Alps in the Massif du Mercantour. To the southwest, the view is partly blocked by a higher peak, Les Pradels. The disadvantage of this peak, however, is the interrupted view of the bay of Cavalaire-sur-Mer and the gulf of Saint-Tropez (Wallon, 1984c).

(16)

16 The oppidum of Maravielle lies at an altitude of 279 metres on top of the Maravielle plateau which forms a basaltic, volcanic island in the centre of the further mainly crystalline ‘horseshoe’. The oppi-dum overlooks the valley of the Môle and has a direct view on the oppioppi-dum of Montjean in the southeast. It looks over the gulf of Saint-Tropez and towards the Alps in the northeast. While it is not situated on top of one of the highest hills, its location on the end of the plateau is defended by a steep cliff on the eastside. Its location also gives the oppidum a good position in the centre of the cultivable Grimaud basin. There are numerous water reserves in close proximity to the oppidum, including two wells on the east slope, a well in the centre of the plateau, and a source in the north (Wallon, 1988).

1.2. History of research

This section describes the history of research at Montjean and Maravielle. The first subsection ex-plores the statements of various authors on Montjean and Maravielle in chronological order before the 1960s. The endnotes refer to the original statements which are cited at the end of this thesis. The second subsection gives an overview of the historical findings the excavations that were con-ducted between the 1960s and the 1980s.1

1.2.1. Before the 1960s

The two volumes of the Dictionnaire historique et topographique de la Provence ancienne et modern by Étienne Garcin in 1835 contain the oldest references to Montjean and Maravielle. Garcin de-scribed Maravielle as a refuge of the ‘Saracens of Fraxinetum’ that inhabited the ‘Mons-Maurus’ (Garcin, 1835a, p. 476; Garcin, 1835b, p. 212)iii_{. These ‘Saracens of Fraxinetum’ are historically}

known as Andalusian Moors who arrived on the coast of the Massif des Maures around 887 AD (Ballan, 2010). The Moors established an Islamic ‘frontier-state’ in Carolingian France from which they were able to invade and raid the Frankish territory for a few centuries. Garcin stated that Mont-jean was a ‘Celto-Ligurian’ enclosure around the top of one of the highest mountains of the Massif des Maures (Garcin, 1835a, pp. 486-487)iii_{. He added that it could have been the most important}

defensive work of the tribe of the Catamulici. Albert Germondy agreed with Garcin by stating that Maravielle was a Moorish encampment enclosed by a stone defensive wall and additionally referred to Montjean as such (Germondy, 1865, p. 179; p. 202)ivv_.

In 1850, Adolphe d’Archiac wrote a geological description of the hilltop of La Bauduffle in La Môle, in which he described that ‘the summit is a circular, crateriform depression, surmounted over three-quarters of its circumference by a ‘roll’ that looks like a bulwark, which is divided on the north and so represents an actual crater whose lava currents still dominate the plateau’ (d'Archiac, 1850, p. 344)vi_{. Baron Gustave de Bonstetten identified Maravielle in 1873 as a ‘square enclosure of dry}

stone walls’ (Bonstetten, 1873, p. 29)vii_{. In 1906, Adrien Guébhard referred to the descriptions of}

d’Archiac and Bonstetten by stating that they were both referring to Maravielle (Guébhard, 1906, p. 381)viii_{, on which he, a year later, elaborated by stating that he found it astonishing that d’Archiac}

did not identify this ‘roll’ with clearly visible built walls (Guébhard, 1907, p. 175) ix_.

In the same work, Guébhard presented an overview of descriptions by other hiostorians of prehistoric enclosures in the Var department, including extensive descriptions of Maravielle and Montjean. According to its description, the enclosure of Maravielle consists of inaccessible parts at the east and the south because of steeply descending rocks. The northern part is defended by a large rampart, whereas the western part consists of a deep ditch which itself is also defended by a less important rampart (Guébhard, 1907, p. 175). The site of Montjean is described as an enclosure made up of two semi ellipses that are attached to the peak, defended by another wall 120 metres

1_{See Appendix A: Excavations of Montjean and Maravielle for an overview of the excavation reports,} publica-tions etc.

(17)

17 ahead. The exterior wall reaches heights of up to 2 to 3 metres and a thickness of 3 metres (Guébhard, 1907, p. 181)x_.

In 1923, Commandant Laflotte published a renewed list of the prehistoric enclosures of the Var department and drawings of the sites of Montjean and Maravielle made by Martial Imbert in 1905, see Figure 7. In his description of Montjean, Laflotte referred to the statements of Germondy and Garcin about the origin of the enclosure by stating that there is no doubt about its prehistoric foundation and thus refuting the idea of a Moorish establishment (Laflotte, 1923, p. 241)xi_.

Figure 7. Montjean and Maravielle drawn by Martial Imbert (Laflotte, 1923). 1.2.2. 1960s - 1980s

In 1962, Denis Wallon and his brother in law, Bernard de Trévillers, conducted their first survey of the oppidum of Montjean. After two other years of surveying, Wallon and de Trévillers started exca-vating in 1965. The excavations continued until 1979 under the supervision of Denis Wallon with only two breaks, respectively in 1968 and 1974. In 1965 and 1967, Jacques Gautier undertook two surveys at the oppidum of Maravielle. Near the end of his excavations at Montjean, Denis Wallon, resumed Gautier’s surveys in 1978 and 1979 before he decided to start excavating in 1980, which he continued until 1985.

During the years of surveys and excavations, the progress of the research was reported by means of short descriptions in the French archaeological magazine Gallia. These descriptions were part of a periodic overview of the archaeological research in a specific region or ‘circonscription’ with a gradually changing name. In 1964, Fernand Benoit was the first to report on the surveys of Mont-jean by Denis Wallon and Bernard de Trévillers in the Circonscription d’Aix-en-Provence (région sud) (Benoit, 1964, p. 593). Three years later, in 1967, Maurice Euzennat reported on the first surveys of

(18)

18 Maravielle and described the progress of the first excavations at Montjean in the Circonscription de

Provence-Côte-d'Azur-Corse (région sud) (Euzennat, 1967, p. 423). Christian Goudineau presented

the results of excavations at Montjean in the following years, first in the Circonscription de

Côte-d'Azur - Corse Gallia (Goudineau, 1971) and later in the Circonscription de Côte-Côte-d'Azur (Goudineau,

1973, pp. 560-561; 1975, p. 563; 1977, p. 501; 1979, p. 559). In 1981, Goudineau described the first excavations at Maravielle and wrote a concluding report on Montjean in the Circonscription de

Côte-d'Azur (Goudineau, 1981, p. 538). Eventually in 1986, Marc Gauthier wrote a concluding report on

the oppidum of Maravielle in the Circonscription de Provence-Alpes-Côte-d'Azur (Gauthier, 1986, pp. 476-477).

In addition to the periodic reports in Gallia, there have been occasional publications on the progress of the archaeological research at Montjean and Maravielle. The first publication on the first surveys of Montjean appeared in 1965 (de Trévillers & Wallon, 1965). In the following years of the excavations, Wallon published three articles in which he elaborated on new discoveries and the adapted hypotheses on the chronology and function of the site (Wallon, 1967b; 1969b; 1973b). In 1979, Wallon published an article about the necks of Massalian amphorae that were discovered at Montjean (Wallon, 1979c). After he started excavating at Maravielle, Wallon wrote an article for the archaeological magazine Histoire et archéologie in which he described the discoveries at Montjean and Maravielle and commented in particular on the simultaneous existence of both sites (Wallon, 1981b). On the remaining excavations at Maravielle, Wallon only wrote two short texts, describing the latest updates on the excavations (Wallon, 1982b; 1984b). Under the supervision of Jean-Claude Echallier, Wallon performed petrographic analyses on some of the excavated ceramics, which was published in two articles (Echallier & Wallon, 1982; 1985).

1.3. Overview of the oppida and the excavations

This section gives an overview of the oppida and of the excavations that were conducted between the 1960s and the 1980s.2

1.3.1. Montjean

The oppidum of Montjean consists of two concentric semi-circular ramparts on the west side that faces the valley of the Môle of which the inner rampart forms the main enclosure. These ramparts both end up at the ridge of the hill on the north and the south. In addition to the two semi-circular ramparts there are three more walls that seem to defend the entrance from the north. The walls were built with blocks of mica schist that seem to be cut from the rock of the hill, see Figure 8. On the east side, facing the bay of Cavalaire-sur-Mer, the oppidum is protected by a steep slope on which a few smaller walls were built as watch towers. Therefore, the oppidum is not visible from the bay.

The inner rampart was built as a wall with double sidings with a trapezoidal cut. It reaches a maximum height of 3 metres on the external side and a height of 1,8 to 2 metres on the internal side, which demonstrates the steepness of the slope at approximately 25%. The width at the top of the wall is 2,5 to 4,2 metres and the width at the base is larger than 5 metres. The wall was built on top of a natural rock bottom which does not seem to have been modified. Although no towers or redans were added, on two places, a parapet is visible on top of the rampart. The total length of the parapet is about 12 metres. The parapet is 1 metre thick and approximately 0,6 metres high. The exterior siding of the parapet follows the exterior siding of the rampart, whereas the interior siding rests on top of it. On a couple of places along the rampart there seem to be ramps to ascend the rampart (see Figure 13 left). At the lowest point of the sloped terrace, in the middle of the main en-closure, the northern and southern part of the wall overlap each other with a width of 2,5 metres

2_{The overview is based on the reports and publications by Wallon, see Appendix A: Excavations of Montjean} and Maravielle

(19)

19 and form the main entrance of the oppidum. Here, the southern part of the wall is 4,2 metres wide while the northern part is just 2,5 metres wide. Because the entrance was built on the lowest point of the terrace, it also functioned as a drainage channel. Near the entrance is a small corbelled niche built in the interior siding of the rampart, which could have been a small funerary structure.

The second rampart is no more than 2 metres wide and does not have a preserved height of more than 2 metres. The wall does not have a clear entrance nor parapets attached on it. It only has a reinforcement on the south side, the ‘south bastion’ where it has a rounded angle which seems to be built to protect the south side of the oppidum with the pass, Col cote 341, that separates the Montjean hill with the Pradels hill. The wall probably functioned as a means of glacis to unmask pos-sible invaders before they reached the main enclosure. The third wall was also built with double sidings and is about 40 metres long. It is located 50 metres north of the second rampart. It seems to have functioned as a defence for a spring that lies just underneath. The fourth wall is actually a re-taining wall that was built to enforce a plateau that starts after the third wall. It probably functioned, like the second rampart, as some kind of glacis. Another small wall with just a single siding is located a small distance to the north. The fifth wall is short and also has double sidings. This wall probably served as surveillance post for the pass, Col cote 328, on the north side of the oppidum.

Figure 8. Overview of Montjean (Wallon, 1984c).

The choices for the location of the trenches were mainly made on the basis of the vegetation of the terrace on which the oppidum of Montjean is located. Most of the trenches inside the oppidum of Montjean were dug against the interior of the rampart, because the vegetation revealed deeper soil layers than in the centre of the terrace. The centre of the terrace has a lot of spots where the mica schist rock is visible on the surface with only low maquis between it. This maquis grows on small holes that are filled up with soil of only 30 to 40 centimetres deep. Against the interior of the ram-part that lies lower on the slope of the terrace, the soil was raised to heights of up to 1,4 metres (Wallon, 1984c).

During the early stage of the research at Montjean between 1962 and 1966, Wallon and de Trévillers opened trench 1 to 14 (see Figure 10 above left and right). Trench 9 consisted of a main trench against the internal siding of the main rampart and a long, narrow trench to the east inside

(20)

20 the oppidum. Trenches 12 and 13 consisted of multiple small trenches in the south part of the oppi-dum. During the years 1967-1973, Wallon opened trenches 15 to 22 of which 18, 18bis, 19, 20, 21, 21bis and 22 were all located on the southeast slope of the oppidum (see Figure 10 below left). Ad-ditionally, Wallon continued excavating trench 5 in which he was mainly interested since the earlier period of research. Next to trench 5, most of the research was conducted in trenches 17 and 18bis. The trenches that were opened during the last period of excavating, between 1975 and 1979, were mainly located on parts of the site that were not yet investigated (see Figure 10 below right). Anoth-er trench 22 was opened in the northeast part of the site, accompanied by trenches 22bis and 25 (see Figure 9 left). Additionally, trenches 23, 24A, 24B, 26 and 27 were opened near the third and fourth walls of the oppidum (see Figure 9 right). However, during this last period, Wallon was mainly interested trenches 11/ 17 and 18bis.

Figure 9. Montjean. Left: northeast part of the oppidum, right: third wall (Wallon, 1976).

The trenches outside the main enclosure, 23, 24A, 24B, 26 and 27 turned out to be completely free of archaeological remains, which was the same case with trenches 7, 10, and 15. The trenches inside the oppidum, 1, 2, 3, 4, 6, 8, 9, 9 T, 12, 13, 14, 16 and 28 yielded just a few sherds but were not valu-able for dating. The excavations on the southeast slope, trenches 18, 19, 20, 21 and 25 did reveal a layer of sherds mixed with soil and demonstrated that this area was probably well visited. The areas that yielded the most intact stratigraphic layers included the area of the main entrance with trench-es 11 and 17, the foundations of three huts in trench 5, and two trenchtrench-es on the southeast slope with a hut in trench 18bis and a fire place in trench 29. These three areas will be discussed in more detail in the following three subsections.

(21)

21

Figure 10. Oppidum of Montjean. Above; left: in 1965 (Wallon & de Trévillers, 1965), right: in 1966 (Wallon & de Trévillers, 1966). Below; left: in 1973 (Wallon, 1973a), right: at the end of the excavations (Wallon, 1984c).

(22)

22

1.3.1.1. The main entrance

The vicinity of the entrance of Montjean was excavated in two parts, trenches 11 and 17, which were enlarged during the years of research and finally combined in the middle (see Figure 12).

When Wallon and de Trévillers started their excavations, the opening of the entrance was blocked by a three metre wide pile of stones which had caused the accumulation of eroded soil to be piled up in a layer of almost one metre against the rampart. Wallon first thought that the spot of the rampart formed a redan or a staggering, but he found out that the stones were piled up on top of a fourth century BC layer, while the north and south parts of the rampart were built on top of the bare rock. Underneath the fourth century BC layer, there was only a layer of yellowish clay that could have been the run-off of the internal siding of the north part of the rampart (see Figure 11).

Figure 11. Entrance of Montjean in 1979 (Wallon, 1979a).

The pile of stones was too high to have resulted from the collapse of the north and south parts to-wards the middle, which suggested that the entrance was blocked intentionally. An unusually large stone within the pile of stones was assumed to have been placed as a base for the construction of the barricade. Since it was not entirely possible to find the exact layout of the two banks, Wallon was unsure how the ends of the north and south parts were shaped.

A part in the north sector (see Figure 12 right) and a trace of 2 metres along the wall with some kind of irregular pavement which rested on top of a layer with a mix of dark soil and sherds, formed the stratigraphically most interesting parts of trench 11/ 17. Wallon did not find traces of foundations or clay on wattle but he did find several fire hearths of which some were encircled by stones to the east of the pavement. He interpreted this arrangement as if the occupants constructed ‘huts’ against the rampart and placed their fire places more to the interior of the oppidum.

1.3.1.2. The three ‘huts’

The excavations at trench 5 against the internal siding of the rampart yielded an interesting set of constructions (see Figure 13), which Wallon described as ‘huts’. This 3,5 metres wide set of

(23)

construc-23 tions includes, from north to south respectively, huts I and II, a stone pavement and hut III. The as-semblage was constructed on top of a sort of small thalweg formed in the mica schist which runs down the slope to the west. It seems that huts I and II were built first on the north side, after which hut III was built on the south side of this thalweg. Eventually, the depression in between huts II and III was paved with stones.

Figure 12. Overview of trench 11/ 17. Left: in 1975 (Wallon, 1975), right: in 1979 (Wallon, 1979a).

As can be seen in Figure 13, the constructions are made up of walls A, B, C, D and E. Walls B and C form a band of 3,5 metres placed on a rim that was created when a part of the rock substratum was cut away. The walls are approximately 40 centimetres high and have a siding on the west side. Walls A and D were built perpendicular to the rampart.

(24)

24 Hut I is formed by walls A and B and therefore only closed on the south and the west sides. This form indicates that it functioned as a shed or an entrance for hut II. Hut I contained several fragments of baked clay on wattle and carbonised grains. Hut II is formed by walls A, C and D. The excavations revealed three layers of occupation. Wallon dated the oldest layer to the first half of the fifth century BC, based on the ceramic material. The second occupation layer was based on levelling the older layer and by adding a layer of baked clay on top of it. On this layer, Wallon found several fragments of clay on wattle, including two pieces that still contained an angle of 90 degrees. The third layer was built on top of a layer of sherds which made it impossible to date the second layer. The latter layer consisted of material from the fourth century BC and was also covered with pieces of clay on wattle. Hut III is only limited by wall E on the east side which stretches parallel to the rampart with a length of about 5 to 7 metres. The hut is not closed off by walls on the north and the south side. The discovery of big shards of amphorae suggests that hut III could have served as a granary or a cellar. In the northern area of the hut, Wallon discovered a thick layer of ash which could indicate that this hut functioned as a workshop at some point.

Figure 13. The three huts of Montjean (left (Wallon, 1974), right (Wallon, 1984c)).

1.3.1.3. The southeast slope

On the southeast slope, Wallon discovered the remains of a hut in trench 18bis and a fire hearth in trench 29 (see Figure 14). The hut in trench 18bis is placed between the interior of a crescent shaped wall and a descending slope. Facing the east, the hut has a complete view bay of Cavalaire-sur-Mer and the path ascending towards the oppidum. The wall against which the hut was built is over 2 me-tres wide and 5 to 6 meme-tres long. The hut is 1,5 meme-tres wide at the entrance and 1 metre at the deepest point. This narrowness indicates that the hut was probably a watch post instead of a real living space. Remarkable were two almost complete broken urns of respectively 0,54 and 0,35 cen-timetres high. The amphorae retrieved from this hut dated to the last quarter of the sixth to the fourth century BC.

(25)

25 Trench 29 was situated against the interior of a low retaining wall and contained a fire hearth of 2,5 by 1,5 metres. The soil against the wall dated trench 29 to the last quarter of the sixth and the beginning of the fifth centuries BC. The wall has double sidings and probably served as a wedge for a roof of a hut on lower the west side.

Figure 14. South-eastern slope of Montjean in 1979 (Wallon, 1979a). 1.3.2. Maravielle

Unlike the oppidum of Montjean, the oppidum of Maravielle consists of only one large rampart that encloses the end of the plateau of Maravielle from north to south. The enclosure has a crescent shape with the open side facing the east and overlooking the valley of the Môle (see Figure 15). The outline of the oppidum has a length of about 420 metres and encloses a terrain of about 0,7 to 0,8 hectares.

The central west part of the rampart is the widest, with a width of up to 16 metres at the point of the ‘donjon’ of the oppidum. The wall is 1,6 metres wide at its narrowest points. The north and south sides of the oppidum are steep and quite well defended naturally. The terrace is figura-tively cut in half by a diagonal ridge of basalt. It is possible that the south side of this ridge has func-tioned as a quarry for the rampart because the rocks on that part of the terrace are bare and show many irregularities. In the northwest part of the rampart just after the so-called donjon, two over-lapping parts of the rampart form the main entrance, which Wallon referred to as a ‘chicane en-trance’ and a ‘key ring enen-trance’. The overlapping walls form a channel of about 15 metres and are further defended by the strongly descending slope on the exterior of the rampart. About 40 metres south of the donjon, Wallon found a small hole of a square metre which showed a siding that was perpendicular to the rampart. This opening turned out to belong to a complex assemblage consisting of two walls that formed a narrow entrance through the rampart from a steep slope on the west side of the rampart to the terrace on the east side. The assemblage seems to have functioned as some kind of postern.

During his first surveys of the oppidum of Maravielle in 1965 and 1967, Jacques Gautier opened four test trenches. Denis Wallon opened four other test trenches in 1978 and 1979. Since

(26)

26 trench 1 of Gautier was centred in the main entrance, Wallon decided to enlarge this trench and continued the research of trench 4 of Gautier. During the following six years (1980-1985), Wallon opened trenches 4 N to 18. Trench 4 N was an extension of his trench 4 but against the interior sid-ing of the rampart. Trench 8 belonged to the complex structure of which Wallon believed it was a postern (see Figure 16).

Figure 15. Overview of Maravielle (Gautier, 1965).

Because the remaining structures in Montjean were sheltered against the internal siding of the ram-part, Wallon decided to search this area of Maravielle first but he did not find any such structures. Therefore, he dug some trenches in the centre of the oppidum of which only trenches 9 and 10 showed traces of occupation. Trench 9 showed a terrace with a few fire hearths that were circled by arranged stones that functioned as a foundation of the terrace. This trench confirmed an occupation of the site before the establishment of the rampart. Trench 10 was located on a small terrace which was also supported by blocks of basalt. Here, most of the clay on wattle material was found, stem-ming from both the chalcolithic and protohistoric periods of occupation. The only place against the internal siding of the wall that revealed a structure was trench 18, where Wallon found a small stone wall that was perpendicularly attached to the interior siding of the rampart. Trench 7 only provided chalcolithic material and trenches 4, 14 and 15 presented protohistoric material but no occupation layers. Therefore, during the last period, Wallon prioritised the excavations of the main entrance and the postern. Wallon preferred to excavate these two entrances because the other trenches did not reveal any remaining structures. The initial trenches were combined in these areas, such that at the postern, trenches 8, 12, 13 and 17 met, whereas at the main entrance, trenches 4 N, 5, 6, 11 and 16 met. The excavations of these two areas will be discussed in more detail in the next two subsections.

1.3.2.1. The main entrance

Trenches 4 N, 5, 6, 11 and 16 were all dug in the vicinity of the entrance as can be seen in Figure 16. Whereas trenches 6 and 11 were inside the entrance, trenches 4 N and 5 examined the internal sid-ing of the southwest bank of the entrance, and trench 16 was a small trench outside the oppidum at

(27)

27 the bottom of the donjon. Wallon divided the main entrance of the oppidum of Maravielle in three segments (see Figure 17) with from the outside to the inside of the oppidum:

- Segment I: A long straight, converging corridor - Segment II: A diverging curve

- Segment III: The siding of the donjon with an ascending slope

Figure 16. Oppidum of Maravielle. Above; left: in 1978 (Wallon, 1978b), right: trench 7 with the location of Gautier’s trench 4 (Wallon, 1980). Below; left: in 1980 (Wallon, 1980), right: in 1985 (Wallon, 1988).

(28)

28

Figure 17. Main entrance of Maravielle (Wallon, 1988).

The two banks of the ramparts in segment I are about 1,8 to 2 metres high. However, since the en-trance was filled with a significant amount of stone material, this height could easily have reached a height of up to 3 metres originally (see Figure 19). The northwest bank is 3,5 metres wide at the en-trance and reaches a width of 15 to 16 metres at the donjon, whereas the southeast bank has a width of 3 to 4 metres and shrinks to 1,6 metres in the south end. The internal and external sidings were interrupted by several old collapses which were interspersed with well-preserved pieces of wall. This sharp contrast between well-preserved and collapsed pieces of the sidings made Wallon believe that the entrance was purposely dismantled. The ceramic material in segment I was quite poor except for some sherds of indigenous pots and Massalian amphorae as well as a Greek ampho-ra button. In 1980, Wallon found a bronze jug filled with 1742 Antoniani from the third century AD. The jug was buried in the stone mass of the collapsed walls, which shows that by this time, the en-trance was already dismantled.

The junction of segments I and II seems to have been reworked on multiple occasions be-cause of the reuse of stones in the sidings. The manner of reconstruction was interpreted to have been hasty since some of the reconstructions are not founded on a solid underground. In the same junction, Wallon found some blocks that slipped down and rested upon a protohistoric lenticular hearth which he dated by a shard of an amphora neck.

Segment II starts at the point where the entrance is the narrowest and where the southeast bank bends at an obtuse angle which was almost entirely crumbled. Here, the northwest bank curves in a concavity and widens the entrance to a width of 3 metres. In the south part, the structure is doubled by a small wall D (see Figure 17) of 0,5 metres high which probably strengthened the last angle of the southeast bank. This last angle marked the end of the southeast bank and continues into the internal siding of the rampart which was only traceable for 5,5 metres after which it seems to be totally collapsed.

The concavity in the northwest bank seems to have underwent the same kind of reconstruc-tion as the northwest bank in segment I. The most northern part of the concavity rests on top of a horizontal flat monolith of 1,6 metres wide and has a corbelled structure of 40 centimetres deep on

(29)

29 a height of 1,6 metres which could not be rightly interpreted because of the collapsed siding. South of this northern part, there are two parts that are posterior to the northern part while the most southern part of the concavity is contemporary to the northern part. Therefore, the middle section of the concavity seems to be a later reconstruction of this part of the wall. In the same northwest part of segment II is another deep siding (siding A in Figure 17) which continues the northwest facing of segment I.

Underneath parts of compacted earth that were found under different parts of the wall, Wallon found three fire hearths of which one was located beneath wall C which was the oldest ele-ment in segele-ment II. Underneath wall C, there was ceramic material that dated the main entrance to the fifth century BC.

Segment III is defined by the end of the doubled siding of the northwest bank. It opens to-wards the terrace of the oppidum on an ascending slope. To the west, wall C follows a length of 6,2 metres along the donjon with a height of approximately 50 centimetres. After these 6,2 metres a row of large blocks, at former trench 6, follows a corner of 90 degrees to the east and there is no visible siding of the donjon which bends to the southwest from here. The area between the row of stones and the end of the southeast bank remains unexplored. The donjon stands on top of an oval platform of 19 metres north to south and 5 metres east to west. Since the donjon consists of a mass of collapsed stones, it is difficult to reconstruct its original dimensions (see Figure 18). However, it was probably constructed with ascending steps.

Figure 18. Cut E-F (see Figure 16 below right) with trench 16 (Wallon, 1988).

Figure 19. Cut A (see Figure 16 below right) with trenches 4 and 4 N (Wallon, 1988).

Wallon examined the internal side of the wall with trenches 4 N and 5 in order to find a siding, to examine the thickness of the rampart and to find habitational structures. Despite digging through one metre of crumbled stones Wallon could not find an internal siding (see Figure 19). The slope against the interior of the rampart was flattened with dry stone to a depth of up to 1,5 metres on top of which a compacted clay soil was established. This layer of clay presented many protohistoric sherds that could be dating from the early sixth to the fifth century BC. Additionally, there were pieces of clay on wattle which showed the nearby presence of a habitat. Since Wallon dated the rampart to the fifth century BC, he suggested that the earlier material stemmed from a period

(30)

be-30 fore the construction of the rampart and was used as building material for the earthworks that were contemporary to the rampart. The latest material was dating from the end of the fifth century BC and could be related to the last occupation of the oppidum.

1.3.2.2. The postern

Between 1981 and 1985, Wallon excavated the area of the structure which he referred to as the postern. He first opened trench 8 (see Figure 16) and later started excavating trenches 12 (see Figure 21), 13 and 17 (see Figure 20) in the vicinity of the postern. After his last research, Wallon described the area of the postern as a structure consisting of four different defensive elements:

- The postern - A two-level rampart - An artificial terrace

- The slope at the foot of the postern

Figure 20. Overview of the postern. Left: in 1982 with trench 13 (Wallon, 1982a), right: in 1985 with trench 17 (Wallon, 1988).

Figure 21 shows the layers on which the area of the postern was built. The oldest layers, 9a and b, showed a chalcolithic occupation with a first retaining structure. The layers on top of these layers, clearly stem from the Iron Age.

The postern consists of a corridor of 6 metres long and 0,8 metres wide which is curved at its opening on the east side, probably to mask its existence. The corridor ascends on the slope towards the east and stretches up to a retaining wall A of 1,25 metres high. The walls of the corridor were built on top of embankments that also continue underneath the two banks of the rampart. Before

(31)

31 wall A, the corridor makes an angle to the north where it was blocked by two large stones after which it bended to the east. Here, wall B joins a large flat stone that forms a step onto the terrace of the oppidum (see Figure 18). On the other side of the south-north part of the corridor is siding C of the rampart with an undefined structure.

The filling of the corridor shows that it was already established soon after the construction of the postern and close to the period of the abandonment of the oppidum. This layer could be dat-ed to the fifth century BC. Layer containdat-ed a mixture of chalcolithic as well as Iron Age elements of which one could be dated to the fifth century BC and a few others were considered to have been probably older (see Figure 21).

Figure 21. West-east cut. Above: in 1984 with trench 12 (Wallon, 1984a), below: in 1985 (Wallon, 1988).

The two-level rampart (see Figure 20 right) is different to the north and south sides of the corridor. The upper level of the northern part is about 2 metres wide with a crumbled siding. The upper level of the southern side, however, follows the level of the terrace of the oppidum like a retaining wall. The lower and upper levels seem to meet each other at 15 metres to the north and south of the postern.

The slope from the foot of the postern to the donjon is about 8 metres high and 15 metres wide. It was covered with a scree of stones from the rampart that completely covered the underlying relief. The function of the scree was not clear; it could have been there on purpose to support the rampart or it could have been the result of the collapse of the rampart. With the help of trench 17 at the foot of the slope, Wallon discovered that the ditch mentioned by Guébhard did not exist. Trench 17 also showed that the scree ended up perfectly on top of a levelled primitive soil (see Figure 23).

(32)

32 The scree formed a layer of at least 80 centimetres but could not be removed to a deeper extent because of the instability of the stones and the natural slope could thus not be reconstituted.

Figure 22. Postern in 1985 (Wallon, 1988).

Figure 23. Cut L of the postern with trench 17 in 1985 (Wallon, 1988).

1.4. Ceramics and relative dating

1.4.1. Comparison of the ceramic assemblage

Wallon published solely the numbers of significant sherds, i.e. rims, handles etc. (see Appendix D: Ceramic assemblages).3_{The numbers of significant fragments of the CNT ware and the Massalian}

amphorae from Montjean are based on the quantity of individual vessels that Wallon identified.

(33)

33 Therefore, it is uncertain whether the individual vessels corresponded exactly to the quantity of sig-nificant fragments. The biggest difference between the ceramic material of Montjean and Maravielle can be found in the quantity of collected material. Whereas Wallon collected approximately 1080 significant fragments at Montjean, he collected only 196 significant fragments at Maravielle. Alt-hough the total amounts of sherds are in neither case significantly large, it can still be useful to ex-amine the ratio between different ceramic types found at Montjean, compared to Maravielle.

Figure 24. Left: the ratio of amphorae to cooking and tableware (C&T-WARE), right: the ratio of wheel-thrown ware (W-T WARE) to hand-modelled ware (CNT).

The ratio of amphorae to cooking and tableware is significantly higher at Montjean than at Maraviel-le (see Figure 24, on the Maraviel-left). The ratio of hand-modelMaraviel-led ware to wheel-thrown ware is considera-bly higher at both places, a phenomenon that corresponds to most of the contemporary indigenous settlements (see Figure 24, on the right).

Figure 25. Percentages of different provenance of wheel-thrown ware.

The percentages of different types of wheel-thrown wares shows a clear difference between Mont-jean and Maravielle (see Figure 25). The excavations at Maravielle clearly did not yield any other type of wheel-thrown ware than Massalian cream-ware (CL-MAS). Although this type is also domi-nant at Montjean, its ratios to Attic ware and to a lesser extent to grey-monochrome ware are con-siderable. Therefore, it can be assumed that Montjean was susceptible for a larger variation of wares than Maravielle.

A close look at the percentages of different amphorae shows another interesting picture (see Figure 26). The imports of Greek and Punic amphorae are negligible in comparison to the im-ports of Etruscan and Massalian amphorae. In addition, the ratio of Etruscan to Massalian amphorae is much larger at Maravielle than at Montjean. An important aspect of the Etruscan amphorae,

how-0,00% 20,00% 40,00% 60,00% 80,00% 100,00% AMPHORAE C&T-WARE Montjean Maravielle 0,00% 20,00% 40,00% 60,00% 80,00% 100,00% W-T WARE CNT Montjean Maravielle 0,00% 20,00% 40,00% 60,00% 80,00% 100,00%

ITAL-VN ATTIC GREC-OR GR-MONO CL-MAS

Montjean Maravielle

(34)

34 ever, is that Wallon dated seven out of eight edge fragments to the second and third quarter of the sixth century BC, a period during which Montjean was not yet occupied, whereas the Etruscan am-phorae fragments at Montjean dated mainly from the end of Etruscan trade (Wallon, 1988, p. 35).

Figure 26. Percentages of different provenance of amphorae.

Figure 27 shows the percentages of imported, hybrid and CNT ware. Whereas the percentages of imported ceramics differ slightly, the ratio of hybrid to local ceramics is significantly higher at Mont-jean.

Figure 27. Percentages of imported, hybrid and local ceramics.

Figure 28. Percentages of different provenance of ceramics.

0,00% 20,00% 40,00% 60,00% 80,00% 100,00%

A-PUN A-ETR A-GRE A-MAS

Montjean Maravielle 0,00% 20,00% 40,00% 60,00% 80,00% 100,00% IMPORT HYBRID CNT Montjean Maravielle 0,00% 20,00% 40,00% 60,00% 80,00% 100,00%

PUN ETR ITA-GR GREEK

Montjean Maravielle

(35)

35 Figure 28 shows a clear difference between the ratios of different imported ceramics. Whereas the ratio of Punic and Italo-Greek imports to Etruscan and Greek imports is very low at Montjean or even null at Maravielle, the ratio of Etruscan to Greek imports stands out when comparing Montjean and Maravielle. At Maravielle, Wallon found only one Greek fragment in the form of an amphora button, whereas the other imported ceramics consisted only of Etruscan imports (Wallon, 1988, p. 35). At Montjean, on the other hand, the quantities of imported Greek and Etruscan ceramics are almost equal.

1.4.2. Dating of the sites

Wallon used the relative dating of the ceramic material to date the occupation of the sites (Wallon, 1984c, pp. 59-61; Wallon, 1988).

1.4.2.1. Montjean

The ceramic material found in the oppidum of Montjean indicates that the site was only sporadically visited in the period before the first quarter of the sixth century BC. The ceramic material from be-fore the sixth century BC represents only a dozen elements mixed with later material along the ram-part. This could mean that there is no occupation layer from this period. It could also mean that the declivities in the terrace were filled up with other remains, since the interior of the oppidum was not extensively excavated. The distribution of the older elements along the wall can be explained by the removal of dirt from the inside of the habitable areas. This explanation is supported by the fact that more recent material was found near the primitive soil.

The last quarter of the sixth century is characterised by a sharp rise in the quantity of a large range of ceramic material found along the rampart. However, the only object that was precisely dated, an Attic black-figure sherd, was not covered by an occupation layer. The only sealed occupa-tion layer of the site is ‘layer 3b’ of hut II in trench 5 against the rampart, which was dated to the first half of the fifth century BC based on the embedded Etruscan and Massalian amphora fragments. This dating indicates that the rampart was built during the last quarter of the sixth century BC.

The abandonment of the site is easier to date because of the disappearance of imported Greek or Italic ceramics. The latest occupation layer yielded an almost intact amphora of the late fourth or early third century BC. Some amphora necks showed traces of fire which indicates an im-mediate abandonment of the site. The fourth century BC sherds of trench 11/17 confirmed this da-ting.

The ceramics yielded no sign of a significant period of discontinuity between the first and the last moments of occupation, although the excavations revealed only a few habitation zones. The north bank of the main entrance seems to have been built in a later phase in the fifth or even the fourth century BC, which indicates that the area of the entrance was probably reshaped in a later stage.

1.4.2.2. Maravielle

The oppidum of Maravielle was clearly built on top of the remnants of an older Chalcolithic occupa-tion based on the sherds from this period. The chalcolithic layer was levelled and burnt in order to form a foundation layer which was clearly visible in the excavation of the postern. The layers under-neath the postern also showed an extra levelling of basalt stones on top of which wall A was built as some kind of retaining wall. Behind this wall, the terrace was levelled with more layers of stones. The levelled layers underneath wall A included some Etruscan sherds from the late seventh and sixth centuries BC, which dates this first occupation to the beginning of the sixth century BC.

The fact that wall B was later covered by the rampart showes that there were various stages in which the postern was created. The material found in the layer belonging to the last stage of level-ling before the construction of the postern and in the layer belonging to the fillevel-ling of the postern

(36)

36 showed an hiatus of less than a century. The deepest elements in this layer and the layers under-neath the main entrance suggest a construction of the rampart during the second or third quarter of the fifth century BC. Thereafter, the site was abandoned in just a few decades around the end of the fifth century BC.

1.5. Functioning and coexistence of the oppida according to Wallon

The fact that two contemporary sites with such impressive fortifications were occupied on two sides of the same valley convinced Denis Wallon that these sites were somehow involved with each other. 1.5.1. Montjean

Since Wallon excavated Montjean before Maravielle, he first formed an opinion on the function of Montjean (Wallon, 1984c, pp. 108-114). Due to its location, Montjean could have functioned as a perfect maritime observatory. This function permitted the oppidum to overlook the maritime route from the Etruscan coast, via the island of Elba and the Cap Corse on Corsica. Therefore, observers could for example have protected themselves from a possible attack or ambush the sailors them-selves. The view towards Marseille is blocked by the hill of Les Pradels, which suggests that Mont-jean could have mainly been used as an observatory to the east. Therefore, Wallon found it logical if Montjean was connected with another oppidum which did have a good view on the west.

Wallon, recognised three different interpretations of the site:

1. The oppidum oversees a large bay with a sufficient supply of drinking water to function as a resupply stop for sailors. Additionally, the oppidum is invisible from the sea and holds many storage jars. This could mean that the oppidum functioned as a camp of refuge in times of in-vasions from sea. Its close distance to the beach could also suggest that it served as a safe habitat for traders.

2. The oppidum could have formed a so-called ‘citadel/ indigenous harbour’ couple with a Greek harbour in the bay of Cavalaire-sur-Mer, which Fernand Benoit believed would then have been the harbour of ‘Heraclia Caccabaria’ (Benoit, 1965, pp. 96-97), a theory that will be examined later (find below, Heraclia Caccabaria).

3. The abundance of Massalian amphorae indicates that Massalian trade influenced the func-tion of Montjean.

Wallon considered the third interpretation as leading in his observation of the oppidum. According to him, the relationship of the amphorae with the function of Montjean depended on where the amphorae were produced:

A. In Massalian workshops

If the amphorae were produced in Massalian workshops, this would imply that they were transport-ed to Montjean. Wallon recognistransport-ed two functions for Montjean in this capacity:

1. Montjean functioned as a native market

Increasing trade with Massalia led the indigenous people towards the coasts where they built a forti-fied market place. This market place served as a distribution point for the hinterland of Montjean, maybe as far as the alps, where the Massalians seemed to have had less influence than near the navigable waterways. The abandonment of the site in the fourth century BC could then be associat-ed to a disturbassociat-ed contact with Massalia.

(37)

37 The Massalians built a distribution point themselves to extent their influence over the coastal region. Montjean was used as a fortified defence for the harbour against the interior region or even against Etruscan traders. The abandonment of the site would however be difficult to explain because of the growing Massalian power in the fourth century BC.

B. Locally somewhere in the Massif des Maures or in the Var region

Wallon argued that if the amphorae were produced in local workshops, which Jean-Claude Echallier suggested on the basis of a petrographic analysis of the amphorae (Echallier, 1982), it would imply the presence of a Greek colony. Although it seemed clear that indigenous potters started to copy the Greek methods, it was the Greeks that started the production of the amphorae. Wallon suggested that Montjean could then have functioned in two ways:

1. The vessels were produced in Massalians workshops

The Massalians not only built a distribution point as was suggested in the former hypothesis, but they also colonised a part of the region where they established workshops and viniculture. If the Massalians preferred the Var region over the region near Massalia for their viniculture, it does not explain why they abandoned Montjean in the fourth century BC.

2. The vessels were produced in other Greek workshops

It could be that other Greek settlers colonised a part of the Var region and established a stable rela-tionship with the Massalians, which would have been necessary due to the Massalian control over the sea. This idea would support the abandonment of Montjean when the Massalians were to take over control of the region.

2.5.2. Maravielle

In his analysis of the function of Maravielle, Wallon mainly examined its relation and its differences to Montjean. For the oppidum itself, the defensive system was his main focus of attention. The two entrances and the donjon of the rampart had different functions. The location of the main entrance on the northwest benefitted from the natural slope of the hill, whereas the location of the donjon facilitated it to function as a watchtower for the western plateau and at the same time as a defence for the main entrance. The location of the postern was difficult to determine, but could have func-tioned as an opening to the nearest growing area. The layout of the rampart seems to have been significantly more advanced than the rampart of Montjean. It could be that this refinement originat-ed in the difference in building material, which consistoriginat-ed of dolerite slabs at Maravielle and the mica schist rock at Montjean.

Wallon never determined the function of Maravielle as extensively as he did with Montjean, other than the major differences with Montjean.

1.6. Conclusion

To answer the main question of the first study: ‘How can geographical and archaeological features explain the simultaneous occupation of the oppida of Montjean and Maravielle?’, it is required that all the features of the oppida as examined in the subsections above are considered.

Both oppida were located on hill tops of the Massif des Maures, which offered them both a decent natural protection as well as strategic observation points. The locations of the oppida in close proximity to each other gave the occupants the possibility of direct communication in terms of fire or sound signals. Looking at the locations separately, it seems that Montjean could have been in

Montjean and Maravielle and the Massalian connection