primary 2

Roman Villas

This article is by courtesy of Dr J Dicks who retains the copyright.

Rowland’s Castle had both an industrial and agricultural tradition during the Roman occupation of Britain from the Claudian invasion in AD 43 under the command of Titus Flavius Vespasianus, who later became the Emperor Vespasian, until the departure of the Army in AD 410. The fine red Reading bed clays, the timber from the local woods and water from the stream provided the natural resources necessary for the manufacture of ubiquitous coarse kitchen grey wares so synonymous with the Romano-British era. The site of the pottery manufacture and its associated kilns were situated in the gently sloping valley which runs through the modern day village (SU735 103). The geology of the surrounding area is a head of variable deposits of silty clay, locally gravelly with chalky and flinty deposits of clay with flints (British Geological Survey, 1996). On the top of a small promontory overlooking the valley is a thick layer of Reading Clay which was deposited during fluviatile conditions consisting of red mottled clays over Upper Cretaceous Chalk (Chatwin, 1948). It was the high quality of the clay ideal for the manufacture of ceramic vessels that attracted the Roman potters to Rowlands Castle in the First century AD (Dicks, 2009). The river Lavant, which now only flows during wet winters, would have been a constant source of water during Roman times when the local water table was higher. Large quantities of ceramic vessels manufactured at Rowlands Castle have been from both Chichester (Down and Rule, 1971) and Fishbourne Roman Palace (Cunliffe, 1971). This would indicate that the kiln site was a major and significant pottery production centre. Rowland’s Castle Romano-British pottery has been found on many sites in Southern England.

Above the kilns to the west in Mayse Copse was a Romano-British villa which by its close proximity must have been associated with the pottery industry. This may have been the residence of the owner of the potteries or his bailiff (Figure 1).

The major Roman Road from Chichester (Noviomagus Reginorum) to Winchester (Venta Belgarum) was only 4.5 kilometres to the south and it has been postulated that there was a road from the kiln site joining it at modern day Havant. The major Roman roads that connected the principle local administration centres, Civiates, such as Chichester and Winchester were created by the Army. The development and construction of an infrastructure of ports, maritime and river travel, bridges and roads by the Romans was part of an integrated communications structure which facilitated the transportation of goods and people across the empire (Adams, 2012, 231). This transport network was put in place to facilitate the movement of the army and provide imperial access to the provincial centres in order to administer and control the collection of taxes and duties. This infrastructure of roads in Britannia not only facilitated communications within the province but also provided a route to the urban markets for the local farmers. Any connecting tracks would have been built by the local inhabitants and it would seem highly probable that there was a road running south from the kiln site to the major road and as such providing access to the markets at Chichester and possibly Winchester.  It has been suggested that the road ran south from the potteries and the villa crossing the modern day Whichers Gate Road and continued down Prospect Lane. There was another Romano-British villa situated just metres from this track now under the Havant Academy playing fields at Wakeford’s Copse (Figure 1). Whether this villa had any connection to the potteries is not known and it was probably merely an agricultural establishment.

The majority of rural Romano-British villas were agricultural establishment and it is highly probable that a third villa in the Rowland’s Castle area was in a field west of Finchdean Road (SU737 114). Pottery and building material have been found near this site (Figure 1).

The definition of what constitutes a villa has been debated by archaeologists for many years but the definition that has been adopted here is any domestic structure that exhibits Romanised architectural features such as stone mortared walls, internal plastered and painted walls, tessellated pavements and floors laid with mosaics, hypocausts and bath houses. Other features include roofs tiled with either ceramic or stone tiles or both and evidence of window glass. Obviously not all ‘villas’ will contain all these features and the opulence of the structures will have depended on the wealth and social status of the owner of the property. Over four centuries of the Roman occupation of Southern England from the Claudian conquest in 43 AD to the end of the Roman administration at the start of the fifth century AD many of the villas would have by necessity been modified either to enhance and extend their structures or in some cases have been abandoned and demolished. Whatever the history and events that befell these villas they can be divided into two basic categories; either rural or urban villas. These two categories of villas were built to perform different functions. The rural villa would have served as the farmstead and base of an agricultural enterprise whilst the urban villa would have operated as the domestic home of town inhabitants. As such these differences in functional use would have been reflected the architecture of the two villa types.

Villa Map

Figure 1; Location of the three villas

The Villa at Mayse Copse

There are very few records of this villa and no evidence of it exist today. It was situated at SU 7342 0988 to the south of Woodberry Lane on the northern slopes of the small valley that runs through the modern day village at a height of 40m above Ordnance Data. The earliest recorded evidence is to be found in a ‘Topographic account of the Hundred of Bosmere’ (Bingley, 1817) in which it states that the remains of a Roman buildings were found on the edge of Mayse Coppice. The villa complex would seem to have consisted of at least three separate buildings. One building contained ‘an apartment’ 18 feet (5.48m) by 14 feet (4.27m) which had a red tessellated floor. The walls were plastered with lime tempered with powdered brick (opus signinum) and ornamented with coloured ‘frescoes’. This was assumed to be the principle room as the adjoining one was of smaller dimensions. It was, however, heated by a hypocaust and whilst Bingley assumed it to be sudatorium (a vaulted sweating-room) it was probably a triclinium, (a heated dining room). Two other buildings were recorded by Bingley but with no details apart from stating that one building was twenty feet square (6.10m square) and the other was to the east.

It is probable that the villa buildings created a courtyard which was a typical form of arrangement for rural villas (Smith, 1997). The close proximity to the pottery kilns must indicate that the resident of the villa were either the owner or his bailiff. The occurrence of tessellated floors, painted wall plaster and a hypocaust would seem to indicate that this villa owner had sufficient disposable income to invest in his residence and was, therefore, of some status.

Five coins were recorded as having been recovered from the site: Gallienus; 253-268 AD, Diocletian; 284-305AD, Constantine; 306-337AD, Valens; 364-375AD, Valentinian; 364-375AD. All these coins are of the late third to late fourth century AD suggesting that the villa and the potteries were operating in this time frame. Examples of pottery recovered from the 1963 excavation of the kiln site indicate that the site was also in production during the first and second centuries AD (Dicks, 2009).

A further intriguing observation by Bingley was that from the villa complex there was a causeway with a fosse either side extending in a southerly direction towards the sea coast. This would seem to endorse the hypothesis that there was indeed a road running south from the villa and the potteries to the Chichester-Winchester Roman Road.

Unfortunately Bingley records that the foundations are now torn up and the site of the building sown with corn and, therefore no further discoveries can be expected to be made. Such is archaeology!

The Villa at Wakeford’s Copse

The villa at Wakeford’s Copse was situated at SU727 091 on the north-east side of a gentle sloping hill at a height of 45m above Ordnance Datum not far from the supposed Roman track from Rowland’s Castle to Havant and Haying Island. The local geology is a tertiary strata of London Clay overlaid by river gravels (British Geological Survey, 1996). The possible location of a villa had been established in the early 1960s (Stanley, 1961) (Norton, 1963) when several small trial trenches discovered the flint footing of the villa walls. The site was fully excavated by Portsmouth City Museum in 1968 and again in 1970 prior to its being levelled to create the playing fields for Wakeford School (present day Havant Academy). The 1968 excavation was never fully published but documentation from the excavation are archived at Portsmouth City Museum.

The main building was a rectangular strip or row type house (Smith, 1997) approximately 14m by 8m the foundations of which consisted of mortared flints 0.5m wide. The internal partitioning of the building consisted of wooden walls resting on wooden sleeper beams the position of which were clearly evident through the floor cobbling. The internal arrangement of the villa was five rooms grouped around a possible small corridor 6m wide. There were several patches of red tessellated floor particularly in the north-west room and the small corridor. To the south of the villa there was an enclosed courtyard which contained an oven which would have been used for domestic purposes. To the north-west of the courtyard was an enclosure that was delineated by a series of at least five large stone packed post holes approximately one metre in diameter and about one metre deep. This could had been a stock yard. No coins were recorded but the pottery recovered from the site dates from the mid second century Samian through to the late fourth century AD New Forest and Oxford fine wares. Roof tiles (tegulae) and and a cameo head of Medusa, cut from an onyx and lost from a pendant or brooch setting were also found.

The heavy clays of the site were unsuitable for an arable farming regime so it likely that this was a pastoral enterprise. The villa was, however, a simple rural structure probably owned by a relatively low status farmer.

The Villa at Wellsworth Lane

Although no buildings have been found a concentration of ceramic building material including flanged roof tiles (tegulae) and Roman pottery have been discovered in a field west of Finchdean Road at SU 737 114 at a height of 60m above Ordnance Datum. The pottery sherds included examples from Roman Gaul (Samian) as well as from the New Forest. This would seem to indicate that the artifacts found were from a villa with a ceramic roof dating from the second to fourth century AD. Tesserae were also recovered but as none of the material can now be located it is impossible to say if they came from a simple red tessellated floor or a mosaic.

The remains of any buildings that may have existed have now probably been destroyed by deep ploughing as the field has for many year been under an arable farming regime. However a recent geophysical survey of the area by Winchester University may have identified the possible remains of an aisled building. The villa would have been situated on the Tarrant Chalk Member (white chalk with flints) of Upper Chalk (British Geological Survey, 1996). The surface soils contained relatively large quantities of organic matter incorporated into the mineral horizon and were humic rendzinas (Smith, 1980) ideal for the cultivation of cereals in the Romano-British period and which may have resembled the modern landscape.


At the time of the Claudian invasion and conquest in AD 43 Britain was under the control of the local Iron Age tribes. In the South of England the leaders of the Atrebates and other tribes would seem to have submitted to the Romans without any significant struggle. The principle Roman towns (Civites) in Southern Hampshire and West Sussex were Noviomagus (Chichester), Venta Belgarum (Winchester) and Calleva Atrebatum (Silchester). These civitas capitals were created in the first century AD to administer each tribal area and the rural villas constructed during the first and second century AD were largely owned by the native aristocracy who had adopted the Roman fashions of living (White, 2007, 125). The civitas capitals would have been a focal point for laws, markets, social and recreational activities as well as religion and required an administrative structure to ensure that life was regulated and the town remained habitable.

The ownership of land in Roman Britain was complex and a villa and its immediate lands might have formed only a part of a much larger estate. Each separate villa unit could possibly have been devoted to various agricultural activities and as such contributed to the productivity and wealth of the estate owner. This could explain the variation in the size and Romanisation of the different villas in this area. The deliberate production of an agricultural surplus that could be actively marketed to a wider population needed a market where the products could be exchanged. The civitas capitals would have provided that market. The relationship between the markets, the location of the villas and the communication links to these markets was vital to the success and economic development of the villas.

The existence of such a high concentration of Romano-British activity in Rowlands Castle is a clear indication that this area was a very active during the Roman occupation of Britain. Many other examples of Roman pottery, coins and quern stones have been found over the years in Rowlands Castle most of which were probably dispersed as a result of manure spreding over the fields.

The villas at Wakeford’s Copse and at Finchdean Road were most probably agricultural establishments whilst the villa at Mayse Copse was probably associated with the industrial production of pottery. All three villas would, however, have taken advantage of the road and track network to market their products at the local civitas capital of Chichester.


ADAMS, C. 2012. Transport. In: SCHEILDER, W. (ed.) Roman Economy. Cambridge: Cambridge University Press.

BINGLEY, W. 1817. A Topographical Account of the Hundred of Bosmere, Havant, Havant Press.

BRITISH GEOLOGICAL SURVEY. 1996. Fareham Sheet 316. Keyworth, Nottingham

British Geological Survey.

CHATWIN, C. P. 1948. British Regional Geology, The Hampshire Basin, HMSO.

CUNLIFFE, B. W. 1971. Excavations at Fishbourne 1961 – 1969, Volume 2: The Finds

London, The Society of Antiquaries.

DICKS, J. 2009. The Rowland’s Castle Romano-British Pottery Industry. Journal for Roman Pottery Studies, 14, 51 – 66.

DOWN, A. & RULE, M. 1971. Chichester Excavations 1, Chichester, Phillimore.

NORTON, A. B. 1963. Wakefords Copse. Portsmouth City Museum.

SMITH, C. J. 1980. Ecology of the English Chalk, London, Academic Press.

SMITH, J. T. 1997. Roman Villas: A Study in Social Structures, London, Routledge.

STANLEY, C. 1961. Wakefords Copse. Portsmouth City Museum.

WHITE, R. 2007. Britannia Prima, Stroud, Tempus.