Navigation sidebar:



  1. Antique use of the term
  2. Archaeological definition
  3. Further characteristics
    1. Emergence of oppida
    2. Distribution and siting
    3. Size
    4. Fortification
    5. Duration of use
    6. Are oppida urban?
      1. Trade
      2. Industry
      3. Cult
      4. Coinage
      5. Central places for the area
    7. Population
      1. Population size
      2. Zoning
      3. Land utilization in oppida
    8. Why did oppida emerge?
  4. Footnotes

Antique use of the term

The word oppidum was used by Caesar in De Bello Gallico, in which he reports about his warfare against the Gauls. The sites referred to as oppida are described as fortified settlements, which represent the central place for a tribe. Some tribes had several of these places. Caesar describes Bibracte in more detail and mentions senate meetings and an election (Collis, 1997). According to this, it is assumed that the word oppidum in Caesars usage is intended to mean “town”, especially as some of these places are also mentioned as urbs. On the other hand he is using the word oppidum for merely defensive sites in Britain, which had no central place function. Collis thinks that this is due to translation from Gallic informants, who called these sites *dunon, which was translated oppidum, even though the meaning did not quite match.

Archaeological definition

In France the term oppidum is used for every fortified Iron Age settlement; while in the Czech Republic, Slovakia, and Germany the word oppidum is more a technical term for large (over 20-30 ha) fortified Late Iron Age settlements, thus excluding hillforts, which are smaller than 20-30 ha, and settlements like the Heuneburg, which is comparatively small and of Early Iron Age date. The use of the term oppidum is variable in Britain, as some authors rely on the Czech, Slovakian and German scheme, while others rely on Caesar's second hand description of British oppida, and thus include merely defensive sites with little or no habitation as well as small sites into their definition (Collis, 1984; Avery, 1976).

Further characteristics

The chief characteristics of oppida have just been mentioned, but there are more details to be added to the picture especially concerning regional variances. But a word of caution is necessary, because only few of the oppida have been excavated to any larger extent. Most of the assumptions are based on a few well excavated sites (like Bibracte, Manching, Staré Hradisko and Hrazany). Often enough only the ramparts have been investigated, because of the enormous size of the oppida, which makes settlement archaeology difficult, especially if a representative insight in the settlement history and the internal organization of the site are the aim (Wells 1984, 1987).

Emergence of oppida

It is assumed that the first oppida appeared in the Czech Republic, Slovakia, and southern Germany. Some of the oppida date into La Tène C21. In La Tène D1 more sites in central Europe emerge, for example in southern and middle Germany, Switzerland, Luxemburg, and France. Finally, around 50 BC and later, more sites in France and the first ones in Britain appear (Collis, 1984).

Distribution and siting

True oppida occur in France, middle and southern Germany, Switzerland, the Czech Republic, Slovakia, and Hungary (Wells, 1984). In adjoining areas we find similar developments, which do not exactly match the definition of oppida, but share some of their features. Also true oppida show regional variances in size, occupation, fortification etc. (Collis, 1995).

The oppida themselves are usually situated in easily defensible positions, like hilltops (e.g. Bibracte, Dünsberg) and peninsulae formed by river loops (e.g. Berne-Engehalbinsel, Altenburg-Rheinau)2. Often these sites were chosen for defence purposes, and not for their accessibility or their proximity to trade routes. In many cases this meant that the sites had no former “history”, i.e. they did not naturally evolve out of smaller settlements, but were consciously created (Collis, 1997). Their development was usually rapid. When an oppidum was founded the population of the surrounding area concentrated in it. Often this concentration lasted only for a few generations, then the sites were abandoned for more accessible places, or survived as little villages. Nevertheless some of the Gaulish oppida evolved into Roman towns (e.g. Bibracte or Paris) (Collis, 1995).


Woolf states that usually a minimum size of 20-25 ha is asserted for oppida. Collis sets his threshold at 30 ha, to separate oppida from hillforts. He maps 59 of these sites in Europe.

The differences in size can be enormous. Most sites are small, but the largest have 300 ha and more, for example Altenburg-Rheinau 315 ha, Manching 380 ha, Kelheim 600 ha, Heidengraben 1662 ha (Kuckenburg, 1993). Size is one of the criteria for oppida which is often neglected by authors—in particular when the site shows other criteria which are commonly associated with oppida. This is especially true for France (especially northern France has a large amount of smaller “oppida-like” settlements) and Britain3.


Fortification is one of the main characteristics for an oppidum, but as mentioned by Woolf, probably not a very good one. The fact that fortification seemed to be a central feature of oppida goes back to Caesar and his description of these sites in De Bello Gallico.

The ramparts show regional variation, and usually four different construction techniques are mentioned: Preist-Altkönig, Hollingbury, murus gallicus and Kelheim (Collis, 1975). In addition to these there is the Ehrang construction (like the murus gallicus but without nails (Audouze, 1992)) which is usually incorporated into the group of muri gallici. These are only principle construction schemata, which in reality appear with different kinds of variation. Almost every oppidum has a special type of rampart construction—if not several (Leicht, 2000). Nevertheless there are differences in the distribution of the main fortification types (see Audouze, 1992: fig. 50), (Woolf, 1993). In France muri gallici prevail, while in the more eastern regions (Germany, Slovakia and the Czech Republic) various layouts of the timber-framed wall are preferred (Endert, 1987).

Gates are another important feature of the fortification, they differ roughly in their use through time. The main types according to Collis are: the simple gap, the overlapping entrance, the inturned entrance and the Zangentor4, and hornworks. Of these gate types all but the hornworks %I am not too sure %whether one of the gates might have a tutulus in front of it\ldots appear at the Dünsberg.

The last thing left to mention about the fortification is the complexity of enclosures. Collis mentions that outside Britain most of the complex enclosures with more than two ramparts are confined to the Mittelgebirge5, to which belongs the Dünsberg.

Coming back to the problem of defenses as a necessary attribute for oppida—some authors like Hill think it would be better to judge sites by the action that took place in them and not on the fact whether or not they had fortifications. He is taking up arguments offered by Woolf, who stated that the fortified nucleated settlements were not really different in use, size etc. from their unfortified neighbours. On the other hand he acknowledges the communal efforts to establish these huge fortifications, and assumes that these were just reactions to a thread or a crisis. As a confirmation for this thesis the development of Manching could be cited, or other oppida where the buildings were not confined to the area inside the walls, but where the settlement continued outside as well6. On the other hand we know of deliberate foundations (see section siting).

Collis also mentions sites, which lie outside the boundaries of the “oppida civilization”, for example the Zemplín type settlements in Slovakia and Hungary, which are concentrated habitation sites around a small fortification. They have otherwise all the characteristics of oppida: trade, industry, nucleation etc.. In France on the other hand we have small fortified and unfortified settlements, which show the traits of oppida only at a smaller scale. He even adds fortified sites in Spain and Portugal to the picture, which are the size of Gaulish oppida, but were already in existence before the Roman conquest in the 2nd cent. BC. He continues to describe the British “oppida” and mentions their unusual siting in the valleys and the lowland, as well as their small settlements in the often huge embanked area, where the fortification seems to have been a prestige object. Their role in trade and crafts is unknown.

All these peripheral sites share some aspects of the oppida but not others. It is certainly difficult to decide in which way to broaden (or not to broaden) the term oppida, or to come up with more meaningful definitions, although this is not the aim of this work.

Duration of use

In the east oppida usually stayed in use for about a century or longer (Wells, 1984), but Collis mentions that some places in France had a use of less than a generation. They were then abandoned for some other settlement. He uses an example where four sites (Corent, Gondole, Gergovie, and Augustonemetum) followed each other in a rapid succession and other regions witnessed similar developments. Nevertheless some French sites had more continuity as they lived on as Roman towns (see section siting).

Are oppida urban?

One of the most controversial points related to oppida is the question whether they really were urban settlements, as Caesar described them to be. This question is hard to answer, because only a few sites have undergone modern large-scale excavation, and the surrounding area of these places is even less investigated, so that their role as central places stays obscure. The best excavated and most cited oppidum is Manching, which is in many respects not a typical site. The same accounts for Bibracte, which developed into a Roman town, and in turn effaced the underlying Pre-Roman remains. So the focus for the following questions has to rely on less well excavated sites, which obviously leaves a lot of questions unanswered.


It is not entirely clear how important trade was for the development of oppida. Remains of imported goods such as amphorae have been found in settlements, but long distance trade with the Mediterranean existed long before the oppida came into being.

Most of the traded goods stayed near the border to the Mediterranean cultures, so that in central and southern Gaul large amounts of wine amphorae were found, while their quantity decreases in the inner regions of France (Collis, 1995). Antique sources mention the trade of iron objects between the Romans and the the residents of Noricum. More information about this trade comes from graffitis in Magdalensberg. Further we are told about the exchange of slaves for wine by antique authors, while trade goods like Campanian fine wares, or silver bowls appear at the oppida themselves (Collis, 1997). The question is whether this trade really played such an important role as is often suggested—even the whole existence of oppida is ascribed to it (Wells, 1987). I think that the trade with the Mediterranean was not so important.

Besides the long distance trade, which was according to classical authors organized by Italians7, there is also local trade and trade between oppida (Collis, 1997). Raw materials for production in the oppida could come from considerable distances. For example, the graphite clay for the potteries in Manching was transported over 200 km from Passau. On the other hand the trade of raw iron seems to have ceased, and now the finished goods were traded. Goods manufactured at the oppida were distributed to the surrounding areas, like the painted wares from Stradonice which can be found in the whole of Bohemia. Finished graphite ware cooking pots were also widely distributed (Collis, 1997). How this trade was organized is unclear.


There is plenty of evidence for manufacturing in the oppida. Iron working took place in all of them, and also ‘bronze casting, glass manufacture, pottery making, coin minting, textile production, bone and antler carving, and jewelry manufacture’ (Wells, 1984). Collis even goes so far as to mention mass production of for example brooches and belthooks, which now had stereotypical patterns. On the other hand a lot of specialized tools were produced: for agriculture, crafts, personal adornment, household equipment, wagons, horse gear, and warfare (Collis, 1984). Závist contained 85 different blacksmithy products, while in Hrazany about 65 different objects were produced (Drda, 1995). Pottery production also became more “industrialized”. For example almost 75 percent of the pottery in Manching was manufactured on the potters wheel (Collis, 1997).


Cult related finds in oppida are very rare or not easily identified. In Manching some round buildings with a rectangular ditch enclosing them are taken for temples, one of these buildings has hoards associated with it. A cult-tree (a gilded branch with leafs, together with a gilded sheath) was found, as well as remains of a horse statue made from iron. The horse statue was discovered close to a hoard of weapons, and the fact that the horse was destroyed and its remains scattered is taken as further evidence for a battle at the end of the 2nd century BC in which some of the sanctuaries were looted and destroyed (Sievers, 1999, 1993). During the excavations at Manching huge amounts of weapons and human bones were found. Some of the weapons were too bent to be the remains of a fight, and the bones were mainly skulls and longbones. The skulls might represent trophies, which were taken from defeated opponents, while the longbones were cut from decomposing bodies, freed from the flesh and finally the joints were removed. Both kinds of bone were kept for some time, before they were discarded. That sculls were kept as trophies is reported by classical authors. The longbones could belong to some kind of ancestry cult (Sievers, 1999). There are a few cases of cult related sites found inside oppida, for example there are a few rectangular enclosures (Viereckschanzen)8 in the oppidum of Závist, but most of the sites are outside the settlements (Wells, 1990).

On the other hand it is remarkable that a few of the oppida lost their importance as occupational sites, but continued in use as gallo-roman temple (e.g. the Martberg)9. This indicates that some of the sites must have had some religious function already in Celtic times which continued to be of importance.


Gold coins were known from earlier Celtic contexts, but they were rather a means for storing wealth than for payment. The situation changed in the late second century BC when low value silver and potin coins were appearing (Collis, 1995). Almost every oppidum seems to have minted coins (even though there are coins, which were not produced in oppida, but came from open settlements (Wells, 1990)), but they usually did not leave the area immediately surrounding the oppida. The coins minted in Stradonice, for example, are normally not found outside of a radius of 30 km around the oppidum, and thus we can assume that trade was not managed on a monetary basis (Collis, 1997). The origins of the coins can often be traced, because many of them bear legends naming the local rulers (Wells, 1990).

Central places for the area

One of the criteria Woolf mentions to be essential for his concept of “urbanism” is that there has to be a functional differentiation between sites. This would mean that the oppida, if they have any function as a central place at all, would need smaller sites, which they dominate. For one thing not enough research has gone into this, but what has been observed so far is that there are no secondary settlements near oppida. What is rather the case, is that minor settlements were abandoned in favour of the oppidum, as soon as it was established. If there are larger open settlements they seem to have had the same role as the defended oppida and were not subordinate to them, that is if trade can be used as an indicator for economic dependency. Products from the oppida were either evenly distributed in their hinterland10, or they did not leave the immediate area surrounding the oppida at all11.

Collis is also aware of the fact that the oppida did not form any kind of economic, trading and defensive network. He bases his assumption on the fact that some sites in the the German Mittelgebirge (among them the Dünsberg) were still inhabited in Augustan times even though in France and Southern Germany the oppida had already been abandoned for a generation or more.


Population size

There have been different attempts to judge the size of the population. A major drawback is that usually no cemeteries belonging to the oppida can be detected. Thus the estimates are highly variable, according to the methods used to make them12. Numbers mentioned range from 3000-5000 inhabitants per oppidum (Wells, 1990). For Bohemian oppida several thousand persons are estimated (Závist: 3400, Staré Hradisko: 5000), while the population for Manching, based on the quantity of meat represented by animal bones, was estimated at 1700 (Wells, 1984). Other authors propose population sizes of 1000-2000 persons per oppidum. %who? e.g. The more moderate estimates seem in this case the more likely guesses (Wells, 1984: 171).


Some of the oppida show signs of a deliberate layout with a rectangular grid pattern, where houses and streets follow the major axes (Wells, 1984). These patterns are attested for Staré Hradisko and Manching. The latter site even provides evidence for a major change in orientation to this grid pattern. This in turn gives us proof of the internal organization at work inside the oppida (Sievers, 1999). Both afore mentioned sites have traces of fences, which divide the area into units. These units are interpreted as farmsteads which are joined together to bigger settlement blocks. Wells assumes that 95 percent of the population in the oppida were farmers, and that they may have worked as craftsmen and merchants during the winter. Still it is obvious that there must have been a small number of professional craftsmen as well, especially regarding the high specialization in tools found inside the oppida, and regarding the diversity of tasks carried out. Drda mention smithies, mints, jewellery, and other metallurgical workshops, which are specialized, tannery, coopery, casketry, joinery, wheelwright's work and lathe turning. Further crafts can be assumed of which no evidence survived (e.g. basket-making, dye-making).

Drda state that there are only a few places which were obviously used as sites of workshops, but it can be observed that certain productions took place in the same areas in different oppida. These workshop places can be found either: ‘in the vicinity of gates, in the homesteads and, in special cases, extra muros.’ And further: ‘The presumption that these craftsmen worked in the enclosed areas of farmsteads, in smaller homesteads or in dwellings bordering the main trackways (roads) in the oppidum is valid.’

Wells proposes that there are two types of sites—sites with a distinction between residential areas, workshops and areas of political activity (Bibracte and Manching are named as examples), and sites (like Staré Hradisko) where farmsteads are grouped together as units, but no differentiation of agriculture, craft, and the public sector is evident—every farmstead incorporates all of these criteria.

Woolf refuses to see any evidence of zoning in oppida at all. In this model a town needs distinct quarters for residence, crafts and agriculture, as well as zoning by social class. This differentiation is not evident in oppida, especially as a division by only a few hundred meters between differently used areas is not enough for him.

Evidence for zoning has been found in Manching where an area predominantly used for crafts was discovered . Nevertheless the excavators seem inclined to believe that the area was used exclusively for crafts, because the moist soil was not suitable for residential buildings (Sievers, 2000: 391).

My opinion on the question of zoning is that Woolf is thinking in too modern terms, spatial division between living and working areas is a rather recent development, which I would attribute to the dawn of the industrial revolution and the emergence of manufactures. In Roman towns there might have been residential areas for the elite, but this is due to the fact that the Roman elite did not work in agriculture or crafts. So I would rather agree with him that we are here looking at ‘a local variation on urbanism peculiar to Iron Age Europe’.

Land utilization in oppida

Oppida do not only consist of occupied regions, they usually include empty areas as well. The use of this land could have been diverse, at Manching it is assumed, that wet areas were used as pastures for cattle (Kuckenburg, 1993), while drier areas might have provided arable land for some of the farmers inside the oppidum (Sievers, 1999). This area could also provide space for the population of the surrounding settlements at times of war (i.e. act as an refuge (Dehn, 1962)). For both of these uses lower ground had to be included into the fortification. It is also attested that sources of water were deliberately incorporated into the oppida (Avery, 1976). One example would be the Grinchesweiher and the Schulborn at the Dünsberg.

Why did oppida emerge?

Most of the researchers think that the emergence of oppida was in some way related to the trade with the Mediterranean (Wells, 1984, 1987; Collis, 1993, 1995). A competition started for the luxury goods imported from the south so that the iron industry was intensified, slaves could be obtained through warfare (which would stimulate the need for defence). Social and political changes came to pass, which in turn lead to settlement nucleation in form of the oppida.

If we would only look at France, we could at least agree that trade was a factor in the process of oppida development, because there the oppida emerged at about the same time when the trade with the south began to flourish, but having a look at the wider picture we see that the trade theory has not much footing indeed. Collis included maps of imported Mediterranean goods from the Iron Age. They show that most of the imports (mainly wine) reached southern and central France, but at the same time almost no imports reached Slovakia, the Czech Republic or Germany, and these were the first places where oppida emerged. In addition the oppida were built before the trade with the Mediterranean became important. Even though Collis realized this discrepancy he offered no other explanation.

Other theories are concerned with threat from outside, like the Dacians or the Germans who were challenging the security of the Celtic settlements, driven by overpopulation after a period of migration.

This might have been of concern for the people living near these tribes, but hardly for the Celtic population living close to the Alps. But something obviously must have happened if suddenly large settlement concentrations appeared. Certainly there was an ‘increased social power’ (Woolf, 1993), which made it possible that fortifications could be build and large settlements evolved in them. Audouze (1992: 242) have a different explanation:

However, we believe that we can distinguish in the characteristics of the oppida the signs and motivations that go beyond the need for defence. By going back to earlier hillforts or installing themselves in similar upland locations the Gauls resumed an older tradition.
They see in the creation of the oppida the wish to hold on to old traditions, but also the ‘wish to delineate an urban space, separated from the countryside’. I am not sure whether I could agree that the oppida were planned as towns, but certainly they were planned, and the vast area they can include seems to indicate that a lot of settlement activity is expected. At the Dünsberg we also have evidence that a formerly fortified site was chosen again to be used as settlement area. The inner fortification is most probably of Late Bronze Age / Early La Tène date. Settlement remains dating to this period have been found inside the fortified area. Then there was a break in the settlement activity, which started again, maybe in the Middle, but certainly in the Late La Tène period.

Surely this does not explain everything, but new and more likely ideas are needed to find the answer to why the oppida emerged.

1 I will follow the terminology of Reinecke, when mentioning relative chronology. A lookup table for his terminology will be presented in appendix chronology.

2 Manching is here as so often an exception, it is situated in lowland area, and existed at first as an undefended settlement nucleation, which was only enclosed by a murus gallicus after a major military disaster.

3 This may also be grounded in the different nomenclature. See section definition (Woolf, 1993).

4 Zangentore are according to Collis definition over 20 m long, but Dehn and Endert define them as follows: ‘rampart ends were turned back at right-angles into the interior to edge the entrance ways’, the length of the entrance way is not mentioned.

5 Outside these regions only the Magdalensberg and Závist provide complicated fortifications.

6 For example on the Dünsberg, where Reeh maps platforms to the west and south of the fortified area (even though it is not attested that these platforms are contemporary with the oppidum).

7 This is also supported by findings from the Magdalensberg (Collis, 1975).

8 The question whether Viereckschanzen are really cult related sited is vividly discussed. K. Schwarz had based on his excavation in Holzhausen (1957-1963) argued that Viereckschanzen are cult related, this had been the standard interpretation for some time, until recently. Webster already doubted the cultic significance of these sites, and in a new publication (Günther Wieland (ed.), Keltische Viereckschanzen—einem Rätsel auf der Spur, Stuttgart: Theiss Verlag 1999.) they are rather interpreted as places for living, production, storage and shelter. Unfortunately I could not get hold of this publication, my information is based on

9 Information provided by C. Nickel in private correspondence.

10 For example painted pottery from Stradonice in Bohemia, see section trade.

11 We saw this in relation to the coins.

12 Estimates were made on the amount of animal bones found in Manching, otherwise the population size may be estimated by regarding the amount of buildings on the site. These methods can only give a rough indication of the real population size as most of the sites are not sufficiently excavated.