The rise of a villa landscape

Lees deze text in

Population before the arrival of the Romans


kaart1ROBAmersfoortIn the last half-century BC, new population groups arrived in the southern Netherlands, and mixed with the local population. Around the beginning of the first century AD there were at any rate two such new population groups in the southern Netherlands whose name we know for certain: the Batavians in the river area, and the Cananefates along the coast of Zuid-Holland province. It is however difficult to pinpoint exactly which group lived where. For more information, see ‘The province’s indigenous population’.


001paleogeografieNederlandROBAmersfoortArchaeological remains repeatedly highlight the differences (and similarities) between regions in the Netherlands. These regions correlate well with the different types of landscape, as the natural environment had a strong bearing on whether an area was suitable for occupation. In the Roman period large parts of our country were wet and barely accessible. However, this did not mean that they were deserted; they were simply occupied differently to drier areas. The Netherlands can in fact be divided into a high, dry Pleistocene part in the south and east of the country (the yellow and orange areas on the map) and a wet Holocene part in the north and west (shown in blue).
The land in the Pleistocene area formed during the last ice age, and consists mainly of sandy soils. During the ice age, glaciers pushed up large quantities of earth, forming ice-pushed ridges which give the land some relief. Only the very south of Limburg province, with its loamy loess deposits (orange areas on the map) is really hilly. People have always been attracted to loess, as it produces such fertile soil.
The Holocene part was formed after the last ice age ended, some 10,000 years ago, in the delta of several rivers, and by the sea. The rivers meandered through the low-lying landscape on their way to the sea. This part of the country consists largely of clay and peat. Small rivers with higher banks wound their way through the peatlands. However, most of the land was wet and virtually inaccessible.

kaart 50 na

Farmers on wet and dry soils

It is interesting to note that virtually all pre-Roman settlements are similar in size and in the archaeological finds they yield. The fact that there were few, if any, central places in the Netherlands was probably due among other things to the traditions and way of life of the indigenous population. The wet areas and the sandy soils of Brabant and the east were populated mainly by stockbreeders who reared cattle and horses. Arable farming probably produced too small a yield in these areas. Only in the very south of the country, on the loamy loess soils of southern Limburg, was arable farming the mainstay of the population.

Cattle farmers

Indigenous stockbreeders kept cattle in their farmhouses, which included both living quarters and animal stalls under one roof. Grazing cattle required a lot of land, so settlements lay far apart, and it would be impossible for a stockbreeder to live in a town or larger settlement. Indeed, there were probably no such settlements at that time. The region around Kessel and Lith nevertheless seems to have played a central role in the river area, though it should be noted that the eastern river area is one of the most intensively investigated parts of the country. Kessel and Lith are located at the spot where two major rivers – the Waal and the Maas – meet. This was an important location. An important shrine here attracted people from far and wide even before the arrival of the Romans. Similar shrines were also built in Elst and in the village of Empel in Brabant. However, there was probably no such thing as a capital at that time.

Unique Batavians

We know from historical sources that the Batavians had a special status in the Early Roman period. They had probably befriended and concluded agreements with the Romans before they moved to the Dutch river area. This friendship continued after the Batavians moved. The agreements, under which the Batavians were not required to pay taxes in the form of money or goods, remained in force. They were however obliged to provide soldiers, and Batavian warriors were highly prized. Large numbers of young men left their families and joined the Batavian units in the Romany army. The other population groups in the Netherlands, including the Cananefates and probably also the Sturii and Marsaci, also provided troops for the Romans. All soldiers came into contact with Roman culture during their military service. They were obliged to adopt a Roman soldier’s way of life, and many of them therefore learnt to read and write.

Settlements prior to the arrival of the Romans

Changing settlements

Many of the settlements that have been excavated date from around the beginning of the present era. However, some of them appear to have been occupied as long ago as the mid-first century BC, remaining in use into the Roman period.

Shifting farmsteads

In the Iron Age, land division took the form of ‘shifting farmsteads’. Iron Age farmhouses, which combined animal and human accommodation under one roof, were in the middle of the fields. Small barns for storing the harvest stood beside the main house. The farmstead was regularly moved, so that other fields could be brought under cultivation, while the farmer remained close to his land.

A permanent pitch

From the Late Iron Age onwards, people began to settle at fixed spots in the landscape. When a house needed replacing, it would be rebuilt on more or less the same site. A shallow ditch (probably with a hedge or fence) would be dug around the farmhouse and outbuildings. This clearly divided the farmstead from its surroundings, and prevented animals – wild or domesticated – from entering or leaving the farm. The ditch also seems to have served to distinguish each farmstead from the rest, as people apparently became more conscious of their possessions. Some larger settlements were in fact entirely fenced in.


The Iron Age occupation differed in a number of respects from the occupation in the Roman period. The changes to houses and settlements at the end of the Iron Age, in the last fifty or hundred years BC, are striking. Because people were living for longer on the same spot, their houses became more robust. The roof-bearing central pillars became thicker, and stood in deeper pits. The orientation of the houses also seems to have changed. For the first time, larger settlements comprising more than a few farms began to emerge, as people developed a tendency to live in greater proximity to each other as they became more settled.

Regional traditions

056nederzetting50vChrThe majority of settlements consisted of no more than three or four contemporaneous farms with outbuildings and wells. The people still lived in farmhouses combining human and animal living quarters. If we map the occurrence of certain house types and different types of handmade pottery, we find that certain customs dominated in certain regions. The boundaries between the different regions appear to coincide with the boundaries between different landscapes and soil types. The regions are also reasonably consistent with the territories of certain population groups as described in Roman texts, although such apparent links should be viewed with caution. In this way, three regions can be distinguished in what was later to be the Roman part of the Netherlands: the southeastern Netherlands (now the provinces of Noord-Brabant and Limburg), the coastal area (Zuid-Holland and Zeeland) and the river area (part of the present-day provinces of Gelderland and Utrecht). Though slight differences can be discerned within this rough division, there seems to have been little difference between regions in terms of the size and layout of settlements.


The southeastern Netherlands: arable farmers and stockbreeders

059tweeschepigWe know that the Texuandri and the Cugerni lived in the southeastern Netherlands around the beginning of the present era. The region divides into two main areas: the fertile loess soils in the hills of southern Limburg, and the poorer sandy soils of Brabant and northern Limburg. A continuous development in house building can be seen in this region. The farmhouses were two-aisled, a row of roof-bearing pillars in the middle of the house dividing it in two. The pottery also shows continuous development from the Iron Age into the Roman period. Smaller houses are also found sporadically in this region, mainly in the loess area. They were probably separate houses or animal stalls, instead of the combined farmhouses found elsewhere.

058tweeschepigdehorden   060beegden

Along the coast: living with water

061plattegrondEllewoutsdijkThe coastal region was mainly characterised by wet peaty areas and higher coastal barriers. The latter would have been the only parts of this landscape that really leant themselves to occupation. Prior to the Roman period, large tracts of land behind the dunes were inundated and in some areas the soil was for a long time too wet to live on. This can often be clearly seen in excavations around the Maas estuary. A layer of clay deposited during flooding is often found between Iron Age and Roman features. In the peaty area of Midden-Delfland (just north of the Maas estuary), however, signs of continuous occupation have been found. Remarkably, settlements on the peat appear to have been continuously occupied, while those on clay were not.

We know from historical sources that the Cananefates lived in the coastal area between the Maas and the Rhine at the beginning of the Roman period. The Marsaci, Sturii and Frisiavones lived to the south of the Maas estuary, in present-day Zeeland and the west of Brabant province.

The pottery and houses of all coastal inhabitants, including those north of the Rhine, are very similar. The differences between the housing and pottery of the Frisians to the north of the Rhine (who were never really part of the Roman empire) and those of the inhabitants of the coastal zone to the south became more marked during the Roman period.

Midden Delfland 2-3-4However, around the beginning of the present era, the houses in both regions were largely two- or three-aisled, although single-aisled houses have also been found in several places south of the Rhine. The three-aisled houses to the south of the Rhine had an extra feature: the outside posts stood diagonally, forming a kind of ‘A’ shape. Furthermore, in Midden-Delfland (an area that lies between Rotterdam, Delft, The Hague and the coast), ‘wall-ditch houses’ from the first half of the first century AD have been found. Traces of such houses are found in the soil in the form of a ditch surrounding a turf embankment. This type of house had already been found in Assendelft in Noord-Holland province, well north of the Rhine.

The river area: between north and south

The river area was home to the Batavians in the last half-century BC. Their houses clearly reflect the area’s position between the north and the south. Some are two-aisled, resembling those in the southern sandy areas. Other houses are three-aisled, as commonly found to the north of the Rhine. Many houses were in fact a combination of two traditions, with a three-aisled and a two-aisled part. The former probably housed the animals. Smaller houses with no animal stalls also occurred in the region in the Early Roman period. The settlements were sometimes separated from the outside world by ditches. These ditches also provided essential drainage, of course. In terms of finds, too, the Batavians differ from the other population groups, for example in terms of coins that are found only in their territory.

n nl drieschepig 064TweeDrieschepigHuisDeHorden

Continuation of indigenous settlements

Dating problems

054hawRijwijkRMOLeidenIndigenous settlements continued to exist throughout the entire Roman period. Some of them were continuations of Iron Age settlements. It is however difficult to establish when the first people came to live in a settlement, as most of the pottery found there tends to be handmade. This kind of pottery is difficult to date accurately, so such settlements are generally held to have been established some time between 50 BC and AD 50. Only when a settlement is comprehensively excavated does it become easier to determine an accurate start date, on the basis of the buildings found, among other things.

Same people, new settlement

kaart5It is rarely possible to demonstrate continuous occupation from the Late Iron Age into the Roman period. This does not however mean that the people disappeared, or that the transition to the Roman period brought new people to the settlement. It is quite possible that settlements moved, so a settlement from the beginning of the Roman period that looks new might in fact already have existed for some time.

Small settlements grow larger

Most small settlements of no more than three or four contemporaneous farms underwent few changes. However, some grew in the first century into larger, sometimes enclosed settlements. Examples of this include the settlements at Wijk bij Duurstede-De Horden , Houten-Overdam, Oss-Westerveld, Hoogeloon and Voerendaal-Ten Hove.

 Wijk bij Duurstede-De Horden, ca 50 BC-25 AD   Wijk bij Duurstede-De Horden ca 50 AD  

  Houten Overdam overzicht  huizen houten overdam overzicht

New farming techniques

A number of settlements changed under the influence of the Romans. They developed in the second century into villae, or villa-like settlements. This development is examined in detail in the section entitled  ‘Advent and development of Roman villae’. However, the form of the settlement was not the only thing to change. Though these further changes may not have had much to do with the Romanisation of the population, they were an indirect consequence of the Roman presence in the country. The large demand for agricultural produce prompted farmers to seek ways of achieving higher yields. One way of doing so was to fertilise less productive land. Stalls were now built with a lowered floor that would be covered with sods of peat. This bedding would collect the animals’ dung and urine. Once a layer of bedding was well impregnated with slurry, a new layer was added. As a result, the floor of the stall grew thicker and thicker. The slurry-soaked bedding was later applied to the soil in the fields, making them much more fertile. This type of animal stall was called a ‘deep litter house’ or ‘potstal’. Farms with this type of stall occurred only from the Roman period onwards.


New houses

In some settlements, the people also attempted to design their homes more efficiently. One or more of the heavy roof-bearing pillars in the middle of the house would be replaced by load-bearing pillars in the walls, thus removing an obstacle in the middle of the structure, and creating more usable space.

New pots

073importawThe Romanisation process can be seen in the finds made during excavations. Despite the presence of the Romans, in the first century AD the inhabitants of the entire southern Netherlands continued to use handmade pottery they produced themselves. Once the Romans had been in the Netherlands for a while, however, the indigenous inhabitants used more and more thrown pottery made on a potter’s wheel at a central workshop. This technique had been introduced by the Romans. The potters sold their wares in large parts of the province, and handmade pottery was used less and less. In the course of the second century handmade pottery in fact largely disappeared from most settlements. Only the Cananefates in the coastal area to the north of the Maas estuary stuck to the old pottery for longer. Even when thrown pottery was readily available, they still preferred to use handmade pots.

New settlements based on indigenous traditions


Rust na de opstand

Tijdens de Romeinse tijd werden er veel nieuwe inheemse nederzettingen op het platteland gebouwd. Vooral nadat de Bataafse opstand in 70 na Chr. was neergeslagen. Na deze opstand werd het rustig langs de Rijn. De handel kwam goed op gang en met de economie ging het steeds beter. De invloed van de Romeinse aanwezigheid op het platteland werd steeds zichtbaarder bij het verschijnen van enkele grote herenboerderijen in 'Romeinse' stijl (de villae) en kleine plattelandscentra. De bevolking groeide onder de gunstige omstandigheden. Oude inheemse nederzettingen groeiden en er werden nieuwe nederzettingen gesticht.

Peace follows uprising

Many new indigenous settlements were built in rural areas during the Roman period, particularly after the Batavian revolt was suppressed in AD 70. After this revolt, the area along the Rhine enjoyed a long period of peace. Trade developed and the economy flourished. The impact of the Roman presence on the countryside became more and more visible, as several large ‘Roman-style’ homesteads (villae) and small rural centres appeared. The circumstances were conductive to population growth, and old indigenous settlements expanded at the same time as new settlements were being established.

New settlements based on old traditions

075venrayThese new settlements barely differed from the old ones, if at all, and old traditions were upheld. The presence of the Romans did not stop most indigenous country dwellers from building their houses according to their own traditions. One example of a new indigenous settlement has been found in Venray-de Hulst. There was already a settlement there in the Late Iron Age, but it was abandoned by the Early Roman period. In the second century, a new settlement was established a short distance away. This new settlement probably existed for 150 years. The excavation plan shows many houses, though they would not all have existed at the same time. The settlement actually consisted of a few contemporaneous farms with outbuildings and wells.

Southern Netherlands


067liesfig0701Lieshout has some good examples of the type of settlement found in the southern Netherlands. Three settlements have been found here within a distance of 400 metres. The largest was established around the beginning of the present era, and was abandoned sometime during the second half of the second century. It probably consisted of only two contemporaneous farms during the early and final phases. During its heyday, in the second half of the first century and early second century, however, there were probably four farms there.There was at least one farm some 400 metres to the west of the large settlement during the same period, around the end of the first century. Around AD 200, when the larger settlement had already been abandoned, two smaller settlements stood no more than 250 metres apart there. They probably did not last long, however, and had at any rate been abandoned by the middle of the third century. This is not unusual, in fact, as the majority of the indigenous-Roman occupation appeared to cease in the southern Netherlands around AD 225.

Population growth

Excavations at Lieshout-Beekseweg showed that two settlements could exist simultaneously in a relatively small area. Finds from the area surrounding these excavations suggest that the entire area was densely populated in the Roman period. Even fairly inconvenient locations were occupied and cultivated.

The river area

Settlements in the river area were similar to those in the southern Netherlands. Only the building methods differed. Farms also included small barns alongside the farmhouse. As in the southern Netherlands, wells provided fresh water if there was no surface water nearby.

Along the border of the empire

66 locatie goudse rijpadNotably, there were also indigenous settlements near to the Rhine, the border of the Roman empire. Some archaeologists believe that there was a military zone along both sides of the border which was entirely controlled by the army. If there was indeed such a zone in the Netherlands, this apparently did not mean that no indigenous people were allowed to live there. Several indigenous settlements have been excavated along the Roman border in the last few years, so the idea of a military zone along the border which was exclusively reserved for the military would not appear to apply to the Netherlands. Some of the settlements date to between AD 40/50 and AD 200. This means that they were built in the period when the border was finalised, and several new fortresses were built. The settlements ceased to exist at more or less the same time as a period of unrest along the border. The settlements near the border were small, like those in the rest of the country, with no more than three or four contemporaneous farms. The farms were exactly like those in other indigenous settlements. Farmers and their animals lived under one roof, in a farmhouse combining both human and animal quarters. They made most of their pots themselves, though later they made grateful use of the attractive, robust thrown pots that the Romans had imported. There were probably therefore no soldiers living in these settlements, as archaeologists have been known to conjecture, just ordinary indigenous farmers. The farmers in the border area could sell their surplus grain and cattle to the Roman military camps, which meant they would have encountered Roman goods and customs.

The coastal area

New land

Many new settlements were built in the coastal area around the beginning of the present era, partly because more areas fell dry and became accessible for occupation and farming. The number of settlements rose sharply during the Roman period.

Different building styles

063eenschepigkustThe predominant building style along the entire coast was the two- or three-aisled house. Single-aisled houses were also found to the south of the Rhine. Despite the major similarities within the coastal area to the south of the Rhine, the Maas estuary appears to marked a boundary. The differences between the areas to the north and south of the estuary became more pronounced during the course of the Roman period. This can be seen above all in the pottery. Inhabitants of the area to the north of the estuary continued to use their own indigenous pottery for a long time, even when others had already switched to using pottery introduced by the Romans.

Farmers on the peatlands

069KethelTo date, most of the settlements found in Midden-Delfland, a large peaty area north of the Maas estuary, have been new. There is some evidence of a settlement on the peat, which was occupied from the Late Iron Age into the Roman period. However, most Late Iron Age farmers left their old home and moved their settlements to sandy creeks that had recently dried up and were higher and drier than the surrounding wetlands. Occupation on peat causes the ground to slowly sink under the weight of the house and as a result of drainage through ditches. In the course of time, a dried-up sandy creek will therefore ‘rise’ above the surrounding peatlands. Only small settlements have been excavated to date, usually consisting of only one house at a time. The settlement site would be defined by natural boundaries. In this way, a strip of settlements formed along the higher ridges in the landscape during the second century. People often built a small mound (terp) on which to build their house. The terps consisted mainly of turf and dung. The buildings in Midden-Delfland were often built of turf. Sometimes a house would be repeatedly rebuilt at the same spot by each successive generation, causing the site to be raised further and further. Hearths and ovens have been found both inside and outside the houses.

Midden delfland prof

Everything in its place

Farmsteads were laid out using a ditch system, which also provided drainage. The ditches created a number of blocks, each of which had its own function. The house, the vegetable garden, the outdoor animal shelters and the well were all surrounded by their own ditch. An example of this system can be seen at the Woudse Polder settlement.

Living on a terp

070schipluidenVandenBroeke1995The remains found in the Dorppolder near Schipluiden were so well preserved that many details of the structure of the houses could be seen. The oldest house dated from the beginning of the first century AD. It stood on a small terp made of clay sods and dung. Each successive house was built on the remains of the previous one, and the most recent one dates from around AD 120. The farmstead was surrounded by a fence. The wet soil had also preserved remains of the paths leading to the house, which were made of branches. The floor of the house consisted of layers of dung and reeds. A path of reed or straw matting had been laid in the stalls, and the cattle stood between wattle partitions.

Midden delfland 2-2-3Connecting ditches

A large number of drainage ditches were dug to keep the settlements in Midden-Delfland dry. They flowed into a main ditch that was well maintained and regularly dredged. This main ditch formed part of a much wider parcelling system consisting of ditches, which divided the land outside the settlement into blocks of arable land and pasture. The settlement was therefore situated within the parcelling system, whose ditches connected several settlements. The division of land became stricter during the Roman period, and it has been demonstrated that standard dimensions were used, indicating that land use was well organised, perhaps even by the Roman authorities.

Advent and development of Roman villae


What is a villa?

074VoerendaalMany new settlements were founded during the Roman period. The most striking of them were the villae, which first appeared around the beginning of the second century. A Roman villa was a large homestead in the countryside. It took the form of a complex, with a main building and auxiliary buildings. One or more of the buildings would often feature architectural elements introduced by the Romans, such as stone walls, pillars and roof tiles. A villa produced a surplus of agricultural produce that could be sold in the towns or to the army. Sometimes the owner would live at the villa, though some were also run by a bailiff or tenant. Villae were usually large farms, though some were also involved in non-agricultural activities.

From wood to stone

The new villae were sometimes built as simple wooden structures. Only later, in the course of the second century, were they rebuilt with stone foundations – primarily the main building, but also often some of the outbuildings. This development can be seen in the villa at Kerkrade-Holzkuil, in the fertile loess area of southern Limburg.

Holzkuil hoofdgebouw

Settlements become villae

Besides the new villae, a number of larger enclosed settlements were rebuilt in the late first and early second centuries. A large Roman-style main building with several auxiliary buildings would be built on the site of the old farms. The structure of the settlement was sometimes radically altered to create an open central courtyard, making it a true villa. Examples include the villae at Voerendaal-Ten Hove and Neerharen-Rekem (Belgium). Both are in the fertile loess area of the south.

Neerharen1 Neerharen2 Voerendaal fasering

Villa-like settlements

Not quite a real villa

In other settlements, though an actual villa was not built, there was clearly some Romanisation of the buildings in the late first and early second centuries. One example is the enclosed settlement at Oss-Westerveld, where there was already a single homestead using specially imported Roman goods in the Early Roman period. 

079porticushuiswesterveldA special house was built at this homestead at the end of the first century. Though it was still a traditional two-aisled structure, it had extra posts around the edge, which have been interpreted as a colonnade (porticus). Such features were not found in indigenous building traditions; they had been adopted from Roman architecture. This special house was surrounded by its own ditch. Beyond this ditch the rest of the settlement comprised traditional farmhouses combining living quarters and animal stalls. Though the entire settlement is reminiscent of a Roman villa, the traditional buildings and layout are clearly not. Such settlements, similar to but not quite real villae, are known as proto-villae or villa-like settlements.

westerveld fase 4

From settlement to villa: a big step

The biggest difference between indigenous settlements and Roman villae lay in the fact that villae had to make a profit from the sale of their agricultural produce. They therefore no longer produced merely for their own consumption, but for trading. This was a new phenomenon in the Netherlands. The villa-like settlements were probably also commercial operations, but the inhabitants were less concerned with the appearance of the settlement. ‘Real’ villa buildings differed considerably from traditional buildings in terms of their appearance, as they included Roman architectural features. Villa-like settlements or proto-villae did not therefore have the appearance of a ‘full-blown’ villa. It is not certain whether such a settlement would ever have been able to develop into a ‘real’ villa. It took a lot of money to build a villa in the true Roman style, and most settlements would probably not be able to afford it. However, they might be able to pay for a colonnade around a traditional wooden house, to give it a more Roman appearance. Another possibility is that, though these people could afford to build a villa, they chose not to, as they had no desire to adopt Roman culture.

The indigenous wealthy

It would appear that the indigenous people of the Netherlands themselves often undertook the construction of a villa, as suggested by the Late Iron Age or Early Roman period settlements found immediately beneath a number of villa. Villa-like settlements might also be interpreted as a sign that this occurred. But how did the indigenous population know how to build a villa? They would probably have turned to the towns, where people from other parts of the Roman empire lived, for help. The architecture of the Romanised buildings was almost entirely unknown to the indigenous population, though the idea of enclosing and structuring a settlement was not. The enclosed settlements from the Late Iron Age were also often structured to some extent around an open area.

Examples of villa-like settlements

Hoogeloon: villa inside an indigenous settlement

082platHoogeloonThe settlement at Hoogeloon represents an intermediate stage between a real villa and a villa-like settlement. The settlement was first established around the beginning of the present era, with four contemporaneous traditional farmhouses. At the end of the second century AD a villa building was erected in the settlement. This building was highly Romanised, and included a heated bathing facility, a porticus and its own palisade right from the outset. Recent research has revealed a number of new aspects that are not included in the plan shown here. A second courtyard containing only a well and a cattle pen was created outside the palisade around the main building. Outside this area there were three traditional farmhouses. The plan shows many more houses, though they were probably not contemporaneous. The villa building was abandoned in the late second century, having existed for no more than a hundred years. The unique thing about the settlement in Hoogeloon is the combination of a truly Romanised villa building and a completely traditional indigenous settlement. The villa building was clearly given its own special place in the settlement. It was built over part of the old settlement enclosure, allowing the farmhouses in the rest of the settlement to remain, and was also oriented in the same way as the existing buildings.

Rijswijk: a villa in the coastal region

083rijswijkAnother example of this type of settlement can be found in Rijswijk in the province of Zuid-Holland. Here, the main building even had wall paintings and walls built partly of stone – highly recognisable Romanised features. The auxiliary buildings were however built according to the old methods and the settlement has the appearance of a not entirely Romanised villa. The ground plan of the Rijswijk villa (on Tubasingel) has been reconstructed in a local park, allowing visitors to trace its position and layout.

Houten: villae in the river area

Villae or villa-like buildings also existed in the river area. Small sections of a number of villae have been identified. One example is the villa in Houten-Molenzoom. Here, two rows of foundation pits – pits filled with rubble that provided foundations for wooden posts – have been found, suggesting that a building with a colonnade once stood here. Remains of two Late Iron Age buildings have also been found in the immediate vicinity, so the spot had been inhabited for a long time. Even before the building with the colonnade was erected, another building nearby was decorated with wall paintings, remains of which have been found among the rubble in the pits. We do not however know what this earlier building would have looked like.

Houten burg wallerweg villa (ROB)The ground plan of another villa in Houten can be seen in the paving on Burgemeester Wallerweg in the town centre. Several Late Iron Age or Early Roman period features have been found here, but the most interesting are the remains of a Roman villa building. Three successive buildings stood on this site in the Roman period. The first farm was built somewhere between AD 50 and 75. It was a wooden building, only the northern half of which has been excavated. Between approximately AD 110 and 120 a second farm was built, again from wood. In this case, too, only a small part of the farm has been excavated. The third building was built somewhere between AD 150 and 175. It had stone foundations and a porticus (colonnade) on the north side. Glazed windows, wall paintings and hypocaust heating tell us that this was a luxurious home. Unfortunately, the rest of the villa complex has not been excavated, so we do not know whether this was a ‘real’ villa or a villa-like settlement, similar to that found in Hoogeloon, for example.

Different types of Roman villa

Villa urbana and villa rustica

Country estate or farm?

There were in fact two types of villa in Italy and the old provinces of the Roman empire: the villa urbana and the villa rustica. A villa urbana was inhabited by the owner himself, though perhaps not always continuously, as he would also have a house in town. Although a villa urbana was a farm, it functioned mainly as a peaceful retreat in the countryside. At a villa rustica, by contrast, the emphasis was on farming. It was often inhabited and run by a bailiff or tenant, and the owner would have visited only on rare occasions. Excavation results in the Netherlands do not allow us to identify the difference between a villa urbana and a villa rustica. It is therefore assumed that a villa urbana, where a wealthy owner would spend a lot of his time, would be more luxurious, while a villa rustica would have been more practical. The term villa rustica is therefore use for villae that were clearly largely oriented towards farming. Villa urbana is reserved for very large, luxurious villae.

A villa as a business

Most of the villae in the Netherlands were probably villae rusticae. They are unlikely to have been very spacious and luxurious, and the entire homestead would have been designed with practicality and production in mind. However, it is assumed that the owner would have lived there himself, unlike at the original villae rusticae.

Luxury villa with fine views

085PLasmolenvillaOnly one villa in the Netherlands is regarded as a villa urbana. It is the ‘Plasmolen’ villa on St. Jansberg near Mook, to the south of Nijmegen. This large villa, built around the beginning of the second century, stood on a specially created terrace on a hillside.  From its high vantage point, the villa commanded a magnificent view over the Maas valley and the surrounding area. Inside it was decorated with wall paintings. A special heating system that had been installed in some rooms supplied heat under the floor and through pipes behind the walls. There were also heated baths in the house itself. Because the building stood on a hillside there was no room for any outbuildings. They may well have been at the bottom of the hill, though none have been found to date. Plasmolen is regarded as a villa urbana because of its size and location, which is still a noticeable feature in the landscape. The villa has been partially reconstructed on its original site, on the flat terrace halfway up the side of St. Jansberg hill, which is covered in trees nowadays.

Villae in the Netherlands

Early investigations

086BocholtzVlengendaalMost of our information about Roman villae in the Netherlands comes from the south of Limburg. Many Roman villae have been excavated there, largely in the 19th and early 20th centuries. Since archaeology was not practised according to the same standards then as it is today, our knowledge of these villae is limited, however. During excavations, the focus was largely on the architectural and art historical aspects of the buildings. Often, only one building would be excavated, whereas villae consisted of several buildings. One exception was the villa at Bocholtz (also known as Vlengendaal), which was excavated around 1912. There, two outbuildings were also excavated.

Southern Netherlands

087MaasbrachtplatIt was not until relatively recently that two villae were virtually completely excavated: Voerendaal-Ten Hove and Kerkrade-Holzkuil in the southern Limburg loess area. An almost complete excavation of the villa in Neerharen-Rekem, just over the Belgian border, has taught us more about the homestead itself and the auxiliary buildings. Analysis of seeds and pollen has added to our knowledge of the agricultural villa economy. Unfortunately, we know little about the Dutch villae further to the north. At Maasbracht, in the Maas valley, however, the main building of a villa has been excavated, giving us more information about villae in this area. The Maas valley was probably full of villae, but very few have been excavated.

Holkzuil alle sporen Holzkuil fases

The river area

Things are much the same in the river area. Surface finds suggest that there were a considerable number of villae there. However, only a few have been extensively excavated: of the 24 (probable) villae in the area, only nine have been partially excavated. The single almost completely excavated villa or villa-like settlement in the river area is Druten-Klepperhei, where a number of wooden buildings and buildings with stone foundations have been found. A wooden house surrounded by a wooden porticus stood at the site of what is suspected to have been the main building. A stone cellar and bath house were also found there, along with many fragments of wall paintings. The main building was probably erected in the late first century, and abandoned in the second century, having existed for barely a hundred years. The rest of the settlement remained for longer, however. The excavation data from the villa at Druten-Klepperhei are being re-examined. It has been found that our old image and the ground plan of this villa probably require some adjustment, though exactly what kind of adjustment is not yet clear.

Ground plan of a villa

What did a villa look like?

Studies – both old and new – of villae in the Netherlands and just over the border have revealed a number of general features that are common to Dutch villae. A villa was an enclosed homestead, which included a main building with auxiliary buildings, a well or water storage facility and perhaps a pond. The site was enclosed by a ditch, hedge, fence, palisade or walls. In some cases, a shrine has been found at or near the homestead. The main building and several auxiliary buildings would usually have had stone foundations. They stood in a fairly ordered pattern around an open area in the centre of the homestead. Villae where the buildings were dispersed over the site are referred to by the German name ‘Streuhof’. On an ‘Axialhof’ the buildings were much more neatly arranged, with the main building on the main axis, and the majority of the auxiliary buildings on the two perpendicular axes, creating a central courtyard. Most of the villae excavated so far in the Netherlands have been of the ‘Axialhof’ type.

Auxiliary buildings

Little is generally known about the function of the auxiliary buildings. They are likely to have been used for storage, housing animals, workshops and possibly staff accommodation. Sometimes a separate bath house was built, though bathrooms were often simply added to the main buildings. It was not uncommon for several of the outbuildings to be built partially of stone.

The main building

089villatypenThe main building or house generally consisted of a rectangular core, to which a porticus (colonnade) and corner pavilions were often added. The house would be fairly symmetrical. Inside, some walls would be decorated with wall paintings. A number of rooms would be heated by means of hot air flowing under the floor and through the wall cavity (hypocaust heating). Many villae lost their original symmetrical form as they were repeatedly expanded.

Visiting a Roman villa?

090KrichelbergA reconstruction of a Dutch villa in the municipality of Kerkrade is open to visitors. At Kaalheide, on the Krichelberg plateau (currently on the edge of a residential area), the foundations of the main building of a villa have been recreated in a local park, clearly revealing the ground plan of the building. The walls had stone foundations that protruded above the ground, probably supporting wattle and daub walls.

Work and private life

A villa traditionally consisted of two parts: the pars urbana (also known as the pars domestica) and the pars rustica. The pars urbana was the residential part of the villa. The pars rustica was the commercial part, where the work – usually farming – was done. This distinction is very difficult to make in Dutch villae. Often no more than a few buildings have been excavated at each site, so it is difficult to determine what purpose they would have served. Furthermore, the two sides of villa life were often not entirely separate. The main, residential, building might also have contained rooms that would be used for commercial purposes, for example.

The large villa at Voerendaal

091voerendaalOne of the most comprehensively excavated villae in the Netherlands is in Voerendaal, southern Limburg. Here, it is possible to distinguish between the pars urbana and the pars rustica, although living and working spaces do seem to have been combined here and there. The pars urbana consists of several separate buildings connected by a colonnade: the house in the north of the site, the horreum (granary) in the northwest, the bath house to the south of the horreum and, finally, the walled garden, which also formed a clear boundary between the pars domestica and pars rustica. On the northern side of the house, in the back yard, there were probably two shrines that can also be regarded as part of the pars domestica. Another slightly larger building whose function is unknown also stood here. The location of the horreum – strongly associated with the commercial side of the villa – in the pars domestica can be explained by the fact that the fruits of an entire year’s labour would be stored there. The prosperity of the villa depended almost entirely on these stocks, so it was important that they were not accessible to all and sundry. The other buildings and structures at the front can be regarded as part of the pars rustica. The animal stalls or barns were probably on the eastern side. The staff of the villa might have lived in their own separate quarters close to the large house. In the southeast corner of the complex there was a building where grain was processed. The smithy probably stood in the southwest corner.

Interior of a villa

Central heating

092hypocaustMost villae in the Netherlands were not particularly luxurious. However, one luxury people did permit themselves was central heating. Traditional farmhouses were heated by a central hearth, which was also used for cooking. The animals in the adjacent stalls also provided some heat. In a villa, the rooms were heated by the hypocaust method. Warm air from a furnace room flowed through a hollow space under the floor (the hypocaust), rising from there through the wall cavity. The cavity was created using square or rectangular tubes (tubuli). Usually only a few rooms would be heated in this way. It is generally assumed that these were the living quarters or reception rooms. Bathrooms were usually also heated.

Bath house

The bathing facility in a villa was usually a smaller version of a public bath house. It generally consisted of cold, lukewarm and hot baths, a changing room and a furnace room. In Kerkrade-Holzkuil the bathing facility was found to be in a good state of preservation. Beside the furnace room was the hot water bath, and probably also a dining room or reception room, which would also be warmed by the heat from the hot bath. There was also an unheated room (probably the changing room) and a cold water bath. The bath house was surrounded by a corridor with a door to the outside, so that people from outside (probably staff) could use the facilities without having to pass through the main building.

093badhuisholzkuil 212balsamarium

Wall paintings

The décor in a Dutch villa was probably nothing special, though interior wall paintings would have clearly distinguished it from an indigenous farmhouse. Such paintings were otherwise found only in military or public buildings, or in town. Those found in the countryside were probably fairly simple and abstract. The walls were often divided into several framed panels of different colours. Ceilings and walls were sometimes decorated with a flower motif or rosettes. Many fragments of this kind of decoration have been found at the proto-villa in Rijswijk. Few remains of paintings depicting animals or people have ever been found, except in the villae at Kerkrade-Holzkuil and Maasbracht, where a number of fragments of wall paintings depicting people have been found. One of the people depicted at Maasbracht is shown with a writing tablet (tabella). The owner of the villa might have commissioned a portrait of himself to show that he was a civilised man with the ability to read and write.

muurschildering vlengendaal 096muurschilderingrijswijk Maasbracht schrijfplank b


Mosaics, generally so common in the Roman world, were rare in the Roman Netherlands. We know from reports of 19th-century excavations that some mosaics have been found, but none has been preserved. Individual mosaic pieces among the finds at several villa sites suggest that there were probably mosaics there, but we do not know what they would have looked like.

Distribution of villae

Distribution of villae

The importance of a good location

kaart6ROBAmersfoortThe variety of villae and villa-like settlements probably derives from the different types of soil and cultural traditions, and their geographical position within the Roman empire. By far the majority of Dutch villae were in the south of the present-day province of Limburg (represented by orange dots on the map). This was due primarily to the fertile soil found there, which would ensure the best possible yield. The loamy loess soil in the south of Limburg was fertile and easy to work. However, the locations of villae suggest that stockbreeding also played a role in the villa economy. Less fertile ground was fine for grazing cattle, and the wetter soils of a brook valley or river meadow provided perfect pasture. There were also slight differences in fertility in the loess soils themselves. Villae often stood more or less on the dividing line between fertile and less fertile land, on a plateau or slope down towards a stream or river. This allowed crops to be grown on the dry, fertile soil, while cattle would graze the lower-lying wetter land. Besides the favourable soil conditions at such a location, the aesthetics of the landscape might also have been a consideration. If a villa was built on a higher spot, it would be visible for miles around, certainly with its orange roof tiles. And of course the inhabitants of the villa would have a fantastic view.

Infrastructure and markets

It was not only the fertile loess soils that ensured the villa economy in southern Limburg flourished. Thanks to the road and water infrastructure, the loess area and Maas valley, in particular, were able to develop rapidly. There were markets close at hand, so produce could be sold quickly and at little cost. Towns such as Heerlen, Maastricht, Tongeren (Belgium) and Aachen (Germany) needed a lot of food. Goods could also be transported quickly and cheaply along the Maas to the north, to supply the large numbers of soldiers stationed on the border. The major road between the Roman town of Boulogne-sur-Mer (France) and the provincial capital in Cologne (Germany) ran right through this area, bringing plenty of trade and traffic. The villae in the Maas valley were ideally located for the Maas, guaranteeing them good connections to both the north and the south. A major road also ran along the left bank of the Maas, from Nijmegen to Tongeren. There were probably many villae along this road, too.

Indigenous settlements among the villae

098kerkradewinckelenIndigenous settlements continued to exist, too. In Kerkrade, an indigenous settlement and several villae have been found within a three kilometre radius. The settlement and villae were probably occupied at the same time. We do not know what kind of relationship existed between indigenous settlements and villae. The inhabitants of the settlement may have worked for the villa owner, but it is equally likely that settlements and villae led separate existences and had very little contact.

Germania Inferior or Gallia Belgica?

Although the villae on the loess soils of southern Limburg have played an important role in Dutch villa research, we are not entirely certain that they were actually in the province of Germania Inferior, since we do not know the precise position of its southern border. The Dutch river area and a large part of the Dutch coastal region did at any rate belong to Germania Inferior. It is possible that the southern Limburg loess area was part of the province of Gallia Belgica, which lay largely in present-day Belgium and northern France. This province was placed under civilian administration at quite an early stage, and the army’s presence was not quite as pronounced there as it was in Germania Inferior. Given that it was a military province until the end of the first century, the villae in Germania Inferior may have emerged and developed differently from those in the neighbouring province of Gallia Belgica. However, it is not clear whether the administration or the presence of the army had such an impact on civilian life.

Distribution of villa-like settlements

Cattle farmers on sandy soils

kaart 6 detailVilla-like settlements (represented by yellow dots on the map) appear to occur mainly on the less fertile sandy soils and along the coast. The sandy soils between the rivers Maas, Demer and Scheldt were in a much less favourable geographical position in the province. Farming was less profitable there, as the soil was less fertile than in southern Limburg. It is therefore assumed that stockbreeding was more important in this region. Breeding cattle probably brought with it an entirely different way of life, and a large-scale farm like a villa might not have suited the way people lived there.

Little contact with the Romans

The sandy area was also somewhat distanced from events during the Roman period. The military-controlled border area was quite far away, and there were no large towns in the region. There were probably no major thoroughfares either. As a result, the inhabitants of the region were less likely to come into contact with Roman culture. Developments in the sandy area therefore differed from those in the loess area of southern Limburg. The villa-like settlement in Hoogeloon, for example, focused mainly on stockbreeding for its subsistence. However, observations of Roman stone buildings and various surface finds suggest that there are more villae (or villa-like settlements) in the sandy soils of Brabant than have been discovered to date.

River area: close to the border

The river area was a formidable competitor of the sandy area when it came to stockbreeding. The Batavians, who lived among the interwoven rivers flowing through the Netherlands, were closer to the military sites and therefore also to the main markets for the sale of cattle and horses. This region was heavily influenced by the military sites along the empire’s border. The Batavians quickly became Romanised after the Batavian revolt had been defeated in AD 69/70. There were probably many villae or villa-like settlements in this region, though only a few have been excavated.

Not entirely Romanised

099DrutenThe only virtually completely excavated villa or villa-like settlement is in Druten. Though this settlement had a Romanised main building, the other buildings appear to have been traditional farmhouses. The complex existed for only a hundred years or so before it was abandoned. The few remains of other villae suggest that they too were not entirely Romanised, at least not in the way that the villae in southern Limburg were. Despite the proximity of Roman culture at nearby military sites, there were still a lot of indigenous settlements in the region, several of which have been excavated.

Where are the villae in the coastal region?

Little is known about the final region, along the coast. It would seem that there were no fully Romanised villae there, though a number of villa-like settlements have been excavated. They were probably concentrated near the Roman town in Voorburg and around the Maas and Scheldt estuaries. The Maas estuary was an economically important area, and a number of stone buildings were erected there in the second half of the second century. No more than a few fragments of them have ever been found, so we do not know whether they were villae. Around the same time, a number of wooden buildings clearly inspired by Roman architecture appeared. We also have little information about these buildings, though it is quite possible that both the stone and wooden buildings were the remains of villae or villa-like settlements along the Maas estuary. The Scheldt estuary was probably also important. Unfortunately, however, the sea has eroded away much of the land that was there in the Roman period, though we do know that the estuary was an important point of departure for ships sailing to Britannia. Many merchants dedicated altars to the goddess Nehalennia here, in the hope that she would ensure them safe passage across the stormy North Sea. The presence of this important trade route will have encouraged occupation in the region. However, since the soil (mainly wet peat bogs) was not very suitable for farming, it is not clear whether there would have been any villae there.

Rijswijk: villa-like settlement

Rijswijk VillafaseThe villa-like settlement in Rijswijk – where a small indigenous settlement developed into a villa-like complex – is a good example of what a villa or villa-like settlement in the coastal region might have looked like. It all began around the beginning of the present era with a single small farm. The settlement soon grew to three contemporaneous farmhouses with outbuildings. The northeastern farmstead – the site of the original farm – developed differently from the others. There, a house with stone foundations was built in the early third century. None of the other houses in the settlement had stone foundations. The house had other Roman features, too, including a bathroom, wall paintings and hypocaust heating. The other buildings and the layout of the settlement changed little. All the farmsteads had their own commercial operation associated with stockbreeding or arable farming. This suggests that, though the settlement did not have the appearance of a villa, it nevertheless operated like one. The settlement site in Rijswijk was divided up by ditches in the mid-second century, creating three separate farmsteads. The entire settlement was also surrounded by ditches, which were connected to a much larger system of ditches. The parcels of land (arable land and pasture) created by this ditch system covered some 2 to 5 ha. Together, the ditches created a parcelling system based on standard dimensions. We now know that an indigenous settlement further to the south, in Midden-Delfland, was part of the same parcelling system.

Katwijk: villa-like settlement on the border

There was probably also a villa-like settlement in Katwijk, very close to the border of the Roman empire. Here, finds of roof tiles, wall paintings and heating pipes suggest that a special building stood on this site, built in the first phase of the settlement and surviving to the end. Though the villa-like settlement in Katwijk was probably indigenous, it appears to have had strong ties to the nearby military site in Valkenburg. This is substantiated not only by finds, but also by the fact that the settlement was founded and abandoned at almost exactly the same time as the military camp at Valkenburg (AD 40-200). The settlement in Katwijk thus clearly differs from that in Rijswijk.

The economic basis of the villae

Farming constituted the main economic basis of villae in the Netherlands, although stone quarrying, iron ore smelting and brick and tile production would also have been possible sources of income.

Arable farming

Transformation of arable farming

Farmers mainly grew cereal crops, particularly on the fertile loess soils. They had already done so in the Iron Age, albeit on a subsistence basis. In the Roman period demand for agricultural produce grew steadily, and farming had to become more intensive to meet the demand. This was achieved using new techniques, such as fertilisation, harrowing and improved ploughing methods. Crop rotation was also introduced, allowing poorer soils, which in prehistory had to be left fallow for long periods, to be used for longer consecutive periods. The harvester developed in Gallia, which speeded up the harvest, might also have been used in some regions.


101speltThe villa economy on the loess soils probably specialised in growing spelt, which has been found at the villa at Voerendaal in an analysis of burnt grains and pollen. Although other types of grain and crops have been found, they were probably grown only for use at the villa itself. Spelt, and also common wheat, have been found in excavations of large granaries at several military camps and in towns. They were probably the most important staple crops.

Fertile land is priceless

It has been calculated that seven villae in the ‘Heerlen basin’, where the Voerendaal-Ten Hove villa is situated, would produce a surplus that could feed an extra 2800 people (besides the villa inhabitants and staff). The less fertile farmland in the loess area would have produced a smaller surplus. There, seven villae could have fed up to 861 extra people. Loess soil degraded as a result of farming and slope erosion, gradually making it less suitable for growing cereals and other crops. It is therefore likely that the production of the villae gradually shifted to stockbreeding in places where the soil had declined in fertility.


Stockbreeding had always played a greater role on the sandy soils of Brabant and in the river area. Although small buildings for storing grain have been found at virtually all villa-like settlements – which appear to be most common in stockbreeding areas – stockbreeding would have been their main source of income. Cattle were kept in the villa-like settlements in traditional farmhouses combining living quarters and animal stalls under one roof. It is not entirely clear whether cattle would also have been kept in the main building itself. At the villa in Hoogeloon there is strong evidence that the settlement revolved around stockbreeding. The position of the cattle pen, within the perimeter fence right beside the main building, indicates the importance of cattle to the settlement. The presence of various combined farmhouses at the villa-like settlements at Rijswijk-De Bult and Oss-Westerveld suggests that they too lived from stockbreeding.


Stone quarrying

Stone suitable for building is found in southern Limburg, though it is not as good as the stone from the German Rhineland or Belgian Ardennes. The Limburg stone has been found at several villae in the region. Interestingly, it was rarely used outside southern Limburg, where people seemed to prefer better quality stone from further afield. The quarries in southern Limburg, which mainly produced limestone, were probably run by villae, however. One possible example of such a villa is Kerkrade-Holzkuil, which was close to a quarry known to have been worked in the Roman period. Half-finished sections of pillar made from stone quarried there have been found in a well at the villa. The villa residents at any rate used the stone from the quarry for their own purposes. There is no evidence that the quarry was exploited commercially, so we do not know whether this would have been the villa’s main source of income. We do however know that arable farming was an important economic activity there.

Production of brick and tile

102dakpanstempelIt is quite possible that there were also villae in the Maas valley or river area that produced their own bricks or roof tiles. The river clay and the clay found in the many brook valleys there was ideal for the purpose. Brick ovens existed along the Maas well into the twentieth century. Stamps on roof tiles tell us who made them, so we know that the majority of roof tiles found in the Netherlands were produced by the Roman army. However, roof tiles with other stamps have also been found in villae. Some of them seem to have originated in Belgium, where they were made by civilian producers. However, no tile works that produced the tiles found has ever been excavated there, and it is quite possible that they were actually made in the Netherlands.

Iron production

Metal slag has been found at most settlements and villae, evidence of a smithy at the site. Most of them will not have been very large, and would probably have made and repaired things for the villa itself and perhaps for people in the immediate vicinity. Nevertheless, the possibility of iron production cannot be entirely discounted. In several areas of the Netherlands there would have been enough bog iron or iron ore for some level of iron production to exist. However, no smelting furnace has ever been found at a villa complex.

Shrines at villae

Possible shrines have been excavated at several villae. Figurines of gods and goddesses are also regularly found during excavations, indicating that cults played a role in the daily lives of the inhabitants. It is difficult to ascertain which gods were worshipped at the shrines found, though it would appear that they included both Roman and indigenous gods.

Secundio of Voerendaal

104voerendaalgraffittiTwo small buildings that are believed to have been shrines have been found in the back yard of the Voerendaal villa. Stronger evidence has indeed been found in association with one of the buildings, in the form of unusual earthenware pots with names scratched on them. They include the forename of a number of people: someone known as Secundio, another whose name begins with Sever(…), a name beginning with Cu(…). A fourth inscription is difficult to decipher. It could be that it included the name of a goddess, which has since become illegible. The name ends in ‘…na’. The following sentence can be reconstructed from the remaining letters: ‘Iu(cundus, -lianus, -lius) hath dedicated this to (the goddess) …na’.

Open-air shrines

105heiligdomHoogeloonSmall shrines have also been found at the villa-like settlements at Hoogeloon and Oss-Westerveld. These open-air shrines, with no building around them, were already in use before the Roman period. In Hoogeloon the shrine was surrounded by a ditch with a row of posts and possibly a (sacred?) tree.

Villa inhabitants

Veterans and indigenous elite

It was originally thought that most villae were established by Roman colonists or army veterans. However, recent excavations have revealed that many villae were in fact owned by indigenous people. This would at any rate appear to have been the case with the indigenous settlements that were later converted to villae, such as Voerendaal-Ten Hove, Neerharen-Rekem and the villa-like settlements at Hoogeloon and Rijswijk. We do not know for certain whether the newly established villae also had indigenous owners, though it is believed that most of these villae will have been founded by indigenous people who had achieved a high status within the new society. They would have had enough money to afford to build a villa.

Local officials from Xanten

106bronzenplaatjeBronze plaques with inscriptions including names have been found in one of the villae in southern Li.mburg (Houthen-Sint Gerlach). The first indicates that a certain Julius gave one Marcus Vitalinius, a high-ranking official and one of the ‘mayors’ of Colonia Ulpia Traiana (the large town in Xanten, Germany), an assurance of friendship. On the back of the plaque is a more recent inscription in which the tribe of Catualium pays honour to Titus Tertinius, an ex-official (one of the law enforcers) from Xanten. The same name also appears on a second plaque, which also honours Titus Tertinius, stating that he was also a ‘council member’ and ‘mayor’. The inscription on the third plaque is very incomplete, although one might discern the name Titus Tertinius here, too. These bronze plaques suggest that villae were owned by wealthy local officials. The villa is more than 100 kilometres from Xanten, so the officials would probably not have lived there all year round. A bailiff or tenant will have been responsible for the actual running of the villa.


Indigenous settlements

Before the Romans came, people lived in rectangular farmhouses that combined both living quarters and animal stalls. Cattle would be housed in one half of the house, while the family occupied the other half. A settlement would generally consist of no more than three or four contemporaneous farmhouses with a number of outbuildings, granaries and wells. But farms often stood alone, too. Sometimes a ditch would be dug around a settlement or around individual farmsteads. Such small settlements continued to dominate in the countryside throughout the Roman period. Finds made there provide incontrovertible evidence that they were situated within the Roman empire, however.

New implements

During the course of the second century almost the entire population switched from their old familiar handmade pots to the thrown pots introduced by the Romans. Other find material also suggests that indigenous people no longer made everything themselves, and that they adopted more and more objects brought by the Romans. The Batavian revolt in AD 69/70 was a turning point. After the rebellion had been put down, the population seems to have submitted to Roman rule, and the ‘Romanisation’ of settlements and objects really took hold.

Romanisation: everyone in their own time

Romanisation was clearly a matter of choice for the local population. Some people adopted more Roman ways than others. Indigenous settlements have therefore been found alongside almost completely Romanised villae. The degree of Romanisation depended, among other things, on the economic potential of the various regions. Southern Limburg, for example, was exceptionally fertile, and farming flourished there. This tied in seamlessly with Roman culture, which had always been strongly geared towards arable farming. The sandy soils, wet peatlands and river meadows in the rest of the country were more suitable for stockbreeding. This was economically less interesting, so occupation developed less rapidly in those areas. Finally, cultural differences between the Dutch regions will also have created differences in the development of and adaptation to Roman culture. Thus, while ‘real’ Roman villae were able to develop in southern Limburg, the most important family in a settlement on the sandy soils of Brabant or in the coastal region will have sought to distinguish itself with nothing more than a Romanesque colonnade. Such settlements are therefore referred to as villa-like settlements.

Roman villae

A Roman villa (villae in the plural) was a form of rural enterprise whose layout and style included elements of Roman architecture and culture. It was a complex with a main building and auxiliary buildings. A villa produced a surplus that could be sold in the towns or to the army. Sometimes the owner would live at the villa, while others would be run by a bailiff or tenant. Most villae were large farms, though some ran quarries or tile works. A villa consisted of an enclosed homestead which included a main building and several auxiliary buildings. The main building, and often also several outbuildings, had foundations made fully or partially of stone, supporting wattle and daub walls. This clearly distinguished them from the indigenous building tradition, which used only wattle and daub. The buildings stood in a more or less ordered fashion around a central yard, where there would normally be a well, and sometimes also a pond or water storage facility. The main building would usually be symmetrical. Most Dutch villae feature a colonnade on the front of the building, with two square rooms at the corners. The interior walls would be decorated with paintings. One or more rooms, including bathrooms in some cases, would be heated by means of a hypocaust, which drew warm air through the floor and wall cavities. Since villae were continually being expanded, their original symmetrical design was often lost.

From indigenous settlement to Roman villa

Both the ‘real’ villae and the villa-like settlements in the Netherlands probably arose out of indigenous settlements. They would have been the slightly larger settlements consisting of more than the usual three or four farmsteads. Several such settlements include one house that was clearly more important than the rest. Archaeologists have found special items that suggest that the occupant of such a house would have been largely responsible for contacts with the Romans. 084porticushuisdeHordenAt the end of the first century many of these main buildings were adapted to the Roman style, as the occupants decided to adopt more Roman customs. Some owners could afford a colonnade and a tiled roof. Meanwhile, the rest of the settlement would barely change, if at all. These villa-like settlements were situated on the less fertile sandy soils and in the wet peat and clay areas along the coast. Examples of such settlements have been found in Rijswijk-De Bult, Oss-Westerveld and Hoogeloon. A few indigenous settlements in southern Limburg were given a radical makeover, however. All the old houses were demolished to be replaced by a main building with separate outbuildings: a real Roman villa. These villae were probably concentrated in the south of Limburg and the Maas valley. Examples of settlements radically remodelled to become villae include those at Neerharen-Rekem (Belgium) and Voerendaal-Ten Hove. We do not really know whether there were ‘real’ villae in the river area, the territory occupied by the Batavians.

New villae

Many new villae were also built, particularly on the fertile soils of the southern Netherlands. It is assumed that these, too, were built by indigenous people, around the beginning of the second century – the same period as the remodelling of Voerendaal and Neerharen, for example. Most villae, both new and remodelled, would have had a provincial Roman interior. In other words, the owners could usually afford the luxury of their own bath house and central heating (hypocaust) to remove the need for a smoky open fire. The walls were often painted. However, very few traces of mosaics – found almost everywhere in the Roman empire – have been found in the Netherlands.


We can generally conclude that the inhabitants of the Roman Netherlands responded to the presence and dominance of the Romans in their own way. It was not until after the Batavian revolt that part of the population really began to adopt Roman customs. People in the fertile areas of southern Limburg were the most widely Romanised, while the majority of rural settlements in the Netherlands were not particularly Romanised. Though the people there were relatively prosperous, they might have deliberately decided not to build a villa, in order not to fully adapt to Roman culture.

Margje Vermeulen-Bekkering


Es, W.A. van, 1981, (3e herziene druk), De Romeinen in Nederland. Haarlem

MUST & B. Colenbrander (red), 2005, Limes Atlas. Rotterdam.

Percival, J., 1976, The Roman villa: an historical introduction. London

Stuart, P. & M.E.Th. De Grooth (ed.), 1987, Langs de weg. Heerlen/Maastricht.

Indigenous population:
Bridger, C., 2000, Zur römischen Besiedlung im Umland der Colonia Ulpia Traiana/Tricensimae. In: Germania inferior. (Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde Band 28). Berlin/New-York 185-211

Heinrichs, J., 2000, Römische Perfidie und germanischer Edelmut? Zur Umsiedlung protocugernischer Gruppen in den Raum Xanten 8 v.Chr. In: Germania inferior. (Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde Band 28). Berlijn/New-York. 54-92

Hiddink, H.A., 1999, Germaanse samenlevingen tussen Rijn en Weser. 1ste eeuw voor-4e eeuw na Chr. Amsterdam.

Roymans, N., 1990, Tribal societies in Northern Gaul. An anthropological perpective. (Cingula 12) Amsterdam

Roymans, N., 1996, The sword or the plough. Regional dynamics in the romanisation of Belgic Gaul and the Rhineland area. In: N. Roymans (ed.), From the Sword to the Plough. Three studies on the earliest romanisation of northern Gaul. Amsterdam. 9-126.

Roymans, H.G.A.M, 2004, Ethnic identity and imperial power the Batavians in the early Roman empire. Amsterdam

Swinkels, L. (ed), 2004, De Bataven verhalen van een verdwenen volk. Amsterdam/Nijmegen

Wolters, R., Germanische Mobilität und römische Ansiedlungspolitik: Voraussetzungen und Strukturen germanischer Siedlungsbewegungen im römischen Grenzland. In: Germania inferior. (Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde Band 28). Berlin/New-York. 146-168

Indigenous settlements:
Bloemers, J.H.F., 1994, Die social-Ökonomischen Aspekte der ländliche Besiedlung an Niederrhein und Niedermaas in Germania Inferior und das Limesforveld von Christi Geburt bis zum 5. Jahrhundert nach Christi. In: Ländliche Besiedlung und Landwirtschaft in den Rhein-Donau-Provinzen des Römischen Reiches. (Passauer Universitätsschriften zur Archäologie 2)

Broeke, P.W. van den & H. Van Londen 1995: 5000 jaar wonen op veen en klei. Archeologisch onderzoek in het reconstructiegebied Midden-Delfland. Utrecht

Bruin, J. de, 2005, Een eerste aanzet tot de ontrafeling van de bewoningsgeschiedenis van zuidelijk Zuid-Holland in de Romeinse tijd. In: E. van Rossenberg et al (ed.), SOJAbundel 2002/2003. Amsterdam/Leiden p. 27-28

Enckevort, H. van, 2000, Bemerkungen zum Besiedlungssystem in den südostlichen Niederlanden während der späten vorrömischen Eisenzeit und der römischen Kaiserzeit. In: Germania inferior. (Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde Band 28). Berlin/New-York 336-396.

Hiddink,H.A., 2005, Archeologisch onderzoek aan de Beekseweg te Lieshout (gemeente Laarbeek, Noord-Brabant). (Zuidnederlandse Archeologische Rapporten 18) Amsterdam

Londen, H. van, 2000, Landscape and water management: Midden-Delfland, a region south of the Limes. In: Germania inferior. (Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde Band 28). Berlin/New-York.169-184

Roymans, N. & S. Heeren, 2004, L'archéologie des habitats ruraux romains aux Pays-Bas bilan de cent années de recherches. Archéologie de la Picardie et du Nord de la France, Revue du Nord. Vol. 86, 19-35.

Schinkel, C., 1998, Unsettled settlement, occupation remains from the Bronze Age and the Iron Age at Oss-Ussen. The 1976-1986 excavations. In: H. Fokkens (ed.), The Ussen project. The first decade of excavations at Oss. (Analacta Praehistoria Leidensia 30). Leiden

Slofstra, J., 1991, Changing settlement systems in the Meuse-Demer-Scheldt area during the Early Roman Period. In: N. Roymans & F. Theuws (ed.), Images of the past: studies on ancient societies in northwestern Europe. (Studies in pre- en protohistorie 7) Amsterdam. 131-199

Stoepker, H. et al, 2000., Venray-Hoogriebroek en Venray-Loobeek: Nederzettingen uit de prehistorie, Romeinse tijd en late Middeleeuwen. (ROB Rapportage Archeologische Monumentenzorg 46). Amersfoort. 33-87

Velde, H.M. van der & M.C. Kenemans, 2003, Een nederzetting uit de IJzertijd en Romeinse tijd op De Hulst (Oostrum, gemeente Venray) (ADC rapport 202). Bunschoten

Verwers, W.J.H., 1998, North Brabant in Roman and Early Medieval Times: Habitation History. Amersfoort

Willems, W.J.H., 1986 Romans and Batavians a regional study in the Dutch eastern river area (originally Published in: Berichten van de Rijksdienst voor het Oudheidkundig Bodemonderzoek 31(1981) and 34 (1984))

Villa-like settlements:
Bloemers, J.H.F., 1978, Rijswijk (Z.H.), 'De Bult'. Eine Siedlung der Cananefaten. Amersfoort.

Hulst, R.S., 1978, Druten-Klepperhei. Vorbericht der Ausgrabungen einer römischen Villa. Berichten van de Rijksdienst voor het Oudheidkundig Bodemonderzoek 28, 133-151

Slofstra, J., 1982, Een inheems Romeinse villa op de kerkakkers bij Hoogeloon. In: J. Slofstra, H.H. van Regteren Altena, N. Roymans & F. Theuws, Het kempenproject. Een regionaal-archeologisch onderzoeksprogramma. (Bijdragen tot de studie van het Brabants Heem 22; I.P.P. publikatie 306). Waalre.102-112

Slofstra, J., 1991, Changing settlement systems in the Meuse-Demer-Scheldt area during the Early Roman Period. In: N. Roymans & F. Theuws (ed.), Images of the past: studies on ancient societies in northwestern Europe. (Studies in pre- en protohistorie 7) Amsterdam. 131-199

Wesselingh, D.A., 2000, Native neighbours: local settlement system and social structure in the Roman period at Oss (The Netherlands). (Analecta Praehistorica Leidensia 32) Leiden

Villae in the Netherlands:
Boe, G. De, 1988, Die Siedlungsgeschichte der Villa rustica zu Neerharen-Rekem (Belgien). In: M. De Grooth et al, Villa rustica; römische Gutshöfe im Rhein-Maas-Gebiet. Freiburg i. Br.

Braat, W.C., 1953, De grote Romeinse villa van Voerendaal. Oudheidkundige Mededelingen uit het Rijksmuseum van Oudheden te Leiden 34, 48-79

Dierendonck, R.M. van et al, 1987, Rijke hereboeren uit Maasbracht. In: M. De Grooth et al, Villa rustica; römische Gutshöfe im Rhein-Maas-Gebiet. Freiburg i. Br.

Es, W.A. van, 1981, (3e herziene druk), De Romeinen in Nederland. Haarlem

Hazenberg, T. & W. K.Vos, 1999, Aanvullend Archeologisch Onderzoek (AAO) in Mook en

Middelaar, 'villa Plasmolen'. (ADC Rapport 6). Bunschoten.

Tichelman, G., 2005, Het villacomplex Kerkrade-Holzkuil. (ADC rapport 155). Amersfoort

Willems, W.J.H., 1987, De grote villa van Voerendaal, In: P.Stuart & M.E.Th. De Grooth (ed.), Langs de weg. Heerlen/Maastricht. 46-50

Willems, W.J.H., 1988a, Voerendaal. Publications de la Société historique et archéologique dans le (duché du) Limbourg 124. 404-424

Willems, W.J.H., 1988b, Die grosse Villa rustica von Voerendaal. In: M. De Grooth et al, Villa rustica; römische Gutshöfe im Rhein-Maas-Gebiet. Freiburg i. Br.

Different types of villa:
Gaitzsch, W., 1986, Grundformen römischer Landsiedlungen im Westen der CCAA. Bonner Jahrbücher 186, 397-427

Gaitzsch, W., 1988, Grundformen römischer Landsiedlungen im Rheinland. In: M. De Grooth et al, Villa rustica; römische Gutshöfe im Rhein-Maas-Gebiet. Freiburg i. Br.

Lenz, K.H.,1998, Villae Rusticae: zur Entstehung dieser Siedlungsform in den Nordwestprovinzen des Römischen Reiches. Kölner Jahrbuch 31, 49-70.

Villa interiors:
Moormann, E.M. & L.J.F. Swinkels, 1980, Wall-Painting Fragments from a Roman Villa at Vlengendaal (Bocholtz). Berichten van de Rijksdienst voor het Oudheidkundig Bodemonderzoek 30, 347-365

Economic basis of villae:
Kooistra, L.I., 1996, Borderland farming possibilities and limitations of farming in the Roman period and Early Middle Ages between the Rhine and the Meuse. Assen

Villa residents:
Mann, J.C. 1983, Legionary Recruitment and Veteran Settlement during the Principate. (Institute of Archaeology Occasional Publication 7). London.


Imperium Romanum (Duits) (Engels)

Excavations around Houten (Utrecht province):
Archeologisch Pakhuis Houten

Examples of settlements and villae in the river area:
Archeologisch Pakhuis Houten