top of page

Lopen's famous Roman Mosaic

In the autumn of 2001 in a small village in Somerset a digger driver was working on a new access road to an old mill

house. In his bucket he noticed some unusual stones he thought might be Roman.


When he contacted the Somerset County Council, the archaeologists agreed and so the story begins...


Please note that no images should be downloaded from these pages without the express permission of the owners.  In the first instance please use this email contact to obtain permission, stating how you wish to use the images.

The Discovery of the Mosaic

Lopen is usually a small dot on the map of Somerset in the South West of England, but in October 2001 it caught the imagination of the world when one of the largest and most important Romano-British mosaic pavements to be discovered was found there.

No one had guessed that in this small village of about 250 people, there had been a community of some importance long before Domesday book was completed in 1085-6 AD. One man with a mechanical digger and a sharp eye changed knowledge of the origins of our village pushing it back in time almost 1000 years, when this part of England was a province of the vast Roman Empire.

A new access road was needed to be put in to the back of Mill House and the previous farm from which the Osborne family ran their vegetable business. The field had been an orchard. Working with his digger late one evening, George Caton noticed small cubes of stone in the earth he was removing. He immediately recognised the significance of what he had seen and immediately stopped work. Next day in daylight, George and the family cleaned up a small area of the mosaic, confirming what they had thought; there it was, a large, elaborate and well preserved Roman mosaic pavement under the land that they had farmed for years.


​The Osbornes contacted Bob Croft, Archaeologist at Somerset County Council and a local archaeologist from a nearby village to look at what had been found. The importance of the discovery was immediately evident and negotiations began at once to allow time for the exposure of the mosaic floor and its subsequent recording, and to raise the funds to do this. Time was short. Frosts and winter weather break up soil surfaces and mosaics; it was already late autumn and the Osborne family needed their new access road, so all of the work would have to be done in three weeks (they were happy to have it last longer).

Thanks to the support of David Miles, Chief Archaeologist at English Heritage who helped with financing the project prepared by Somerset County Council, and archaeological contractors Terrain Archaeology, a three week excavation was possible.

The priority was to expose, clean and record the mosaic pavement, and by careful cleaning of the surrounding area, to understand something of the building in which it lay. In three weeks of intensive work these objectives were achieved. It had been agreed that none of the remains would be damaged and more detailed excavation was not needed.

There was no clear reason to assume that there had been an important Romano British Settlement on that site. It is just under 1km south of the Fosse Way, an important Roman Highway running about 400 miles across the Roman province . It linked Exeter (Isca), Ilchester (Lindinus), Bath (Aquae Sulis), Cirencester (Corinium), Leicester (Ratae) and Lincoln (Lindum). Then linking with Ermine Street to York.

Until recently no other buildings were known in the area, though South Somerset soils are some of the most fertile in England and must always have attracted settlement. A Roman villa had been found at Crimbleford Knap, some 3kms to the north west where excavations were carried out in (19 Century) and some more recently in the early 1990s

Map of Somerset showing Roman finds. Somerset Historic Environment Record

In those precious days of mercifully fine weather, experts and volunteers on hands and knees, with spades, trowels, buckets, wheel barrows, and brushes carefully ‘turned back’ the protective carpet of turf and soil and revealed foundations of walls and mosaic floors. There was no time for deeper or more extensive excavation.

The mill first recorded on the site in seventeenth century, and until boundary changes in 1982 was in the Parish of Hinton St. George, and on the Hinton Tithe map of 1841 shows he field as Hinton Mills. Since 1974 this is in the parish of Lopen.

A mill stream ran into the mill from the south. The remains of the Roman building lay between this and Lopen Brook, and probably

extended under the existing buildings. The older buildings of the mill have walls that contain worked Ham Hill stone and rubble that may have been reused from earlier buildings, probably even from the villa.

During the third week news of the find was released to the press. It was covered on television, in local newspapers, national newspapers (broadsheet and tabloid), in the American Time Magazine for kids and some national newspapers across the globe.  Crowds turned out to see the excavation works.


The Discovery: What was found?

We could all see the beautiful colours and patterns on the floor but needed help understanding what was there.


The walls were quite thick and the building could have had two storeys, its entrance probably approached from the Fosse Way to the north. The foundations of two rooms were clearly visible, significant parts of three more and evidence of three others.

The largest room, and the one in which the major mosaic pavement was found was in room 1. It is thought that this was an extension built with rooms 2, 3 and 4 (also with a mosaic) in about A.D. 360. On the north side corridor 5 runs from the east, three rooms 8, 7 and 6, are south of the corridor. Room 1 extended beyond the northern wall and to the south. To the west from room 1 lay room 4 with 2 and 3 to the south. The largest room 1, 12.20m north to south, was entered from the slightly lower north corridor and separated from it by a step; giving an added sense of grandeur. It lead firstly into a smaller ante chamber 4.5m x 4.5m, and ultimately into the larger southerly opulent space 7m x 7m of this bi-partite room. The east side of room 1 had a recessed area floored with larger tesserae, where the owner may have sat or reclined to greet guests, or perhaps where furniture, such as a display cabinet stood. This very imposing room was 12 metres in length (approx. 39ft.), with the elaborate mosaic, and probably , from the fragments found, painted plaster walls. It is thought that the room was a dining room or triclinium 8. This may have looked towards Hinton St George.

The room’s undecorated edge of larger plain tesserae would have allowed the placing of furniture and shelves against the walls whilst leaving the decorated areas clearly on view.

There were some signs that there was a heating chamber to the west of this room, so it may have been centrally heated. The room appears to be a passage with a staircase of which the first step survived.

The excavated remains of the buildings are thought to be part of a much larger building. A geophysical survey showed that it did not extend to the North or the West. Further remains probably lie beneath the standing buildings to the east and south.. The Halstock Villa about 30 km away in Dorset covered an area well in excess of 1000 square metres. It had ranges of buildings and a bath house grouped around a courtyard had similarities not least its mosaics. The more recent discovery of the Dinnington Villa evcavated by Time Team ( in 2002 Broadcast 12 January 2003 on Channel 4 showed another vast building both from the evidence of an aerial photograph taken in 1976 and geophysical survey conducted during Time Teams weekend. These would confirm the likelihood of the Lopen Villa being much larger than is known so far.

Local oolite was the predominant building stone but Ham Hill stone was also found indicating that it would have been used as cut stone around doors and windows. The roof was probably of blue lias stone tiles. Photographs of the stone use and its probable origin is included in the materials page.

Careful records need to be made during every archaeological excavation so that information recorded can be made widely accessible. Dr David S. Neal and Stephen Cosh ( 11 ), internationally recognised experts on Romano-British Mosaics were involved and made careful drawings of the mosaic from which Dr Neal completed a scale painting of the whole pavement.

The Discovery: Mosaic materials

The mosaic is made from small cubes of natural stone and terracotta possibly cut from tile or brick. Each of these small cubes is called a tessera (several tesserae) from the latin word for dice.The mosaic is made from small cubes of natural stone and terracotta possibly cut from tile or brick. Each of these small cubes is called a tessera (several tesserae) from the latin word for dice.









Section diagram showing how the layers were built up to support pavements and possibly this one.

The tesserae were laid on a surface layer of fine mortar that was applied freshly for each days work. Below that was a layer containing terracotta, nucleus. This rested on a coarser layer of well compacted mortar, rudus, which in turn lay on a layer

of compacted rubble, statumen.

Lime mortar becomes harder more quickly with the addition of terracotta and was called opus signinum, from the Italian city of

Sigma were it is thought to have been developed.

Colours used in the mosaic and their origins




























The Discovery: Mosaic Patterns and Designs

​The mosaic was believed to be the work of the Saltire School based in Corinium, now known as Cirencester. The main characteristics of these designs are large squares and saltires (St Andrew's crosses), with central roundals. The mosaic is multi-coloured or polychrome (Poly = many, from the Greek polus - much and chrome from Greek khroma - colour). The use of the various tones of blue, white, buff and brown, and the red of the terracotta, gave great richness.

The southerly end almost 7m x7m and the north end 4.5m x 4.5m. The design at first glance seems symmetrical but gradually it becomes apparent that that is not the case.

The largest and best preserved area is in the south. The outer border is as elsewhere, set with the larger tesserae, within which the main decorative panel was placed. I will describe it as approached in the way of a visitor walking through the ante chamber and moving towards the south. On your left there is a fish heading south (sadly much of the body missing) within the leaf and guilloche border.

The first of the fine tesserae form a wider band around the whole of that part of the room. Both panels are surrounded by a band of ‘cornice’ or egg and dart motif. This series of circles, here separated by ‘apple-core’ shapes, each contains a heart-shape leaf. The leaves are alternately inward and outward pointing. Within this band a further border and the first of the guilloche (a plait or braid).  This one appears double stranded.



With the exception of this border, the design lacks symmetry, although the geometry remains important for our understanding of how the design was laid out.















The double heart - some with a simple intertwined pair of links- a guilloche knot with a leaf tip at each apex. Some have diamonds and triangles or hour glass forms, others split squares and one a meander.




In B3 the middle panel is a saltire, pointing towards the corners of the room. It is centred by a roundel with an indefinable middle.

The L shape formed by line D1-D4 and to A4 has some of the other motifs included and some new. D1 shows a half of an interlinking square star of double cord guilloche and in the space contained within, a half octagon with three heart-shaped leaves pointing towards a half circle.

D2 appears have contained what is also in B4, a rectangle containing a 3 cord guilloche plait, flanked by two sets of lozenges, as two cubes or perspective boxes set together.

In D3 the image that was flashed across the world by the world press, the dolphin swimming west.

In the south west corner a cube had on its 'flat' surface a guilloche  knot with hour glasses on the lozenges. Turning the corner to C2 the half octagon contains a cup from which spring two leaves, linked by their stems.

Ante chamber Northern section of room 1

The ante chamber was badly damaged but has a scheme of irregular octagons developing small squares. There is a recessed area to the east,the wall following the line of the east wall in Room 1a. This has a floor of the large tesserae. The mosaic in the south rectangle links into a panel running north south. The ‘leaf in circle’ design continues as the boarder but the band of guilloche appears to change from two strand to a four strand plait or guilloche.

To the south west of the panel is a small block of dark and light grey checker pattern, there is a similar area in cream and blue to the east and red and cream to the north. A segment of the umbrella roundel was visible on the south and in-discernible to the south. To the north west ahalf hexagon contains a half circle and a severely damaged interior. To the east another roundel resembles the one in B2.

The narrow room 4 running to the west has a central panel that differs from the other rooms in a number of respects. The side panels of large creamy grey are wider throughout the preserved length. The first border in finer tesserae is of blue triangles set in light grey. A simple blue square at the eastern end contains an eight point petalled flower but atnits centre two interwoven links of a guilloche knot. Beyond this to thewest a larger area of meander in dark blue and then finally to a square ncontaining three (and probably a fourth) heart-shaped leaves with stems to the corner.

It is believed that the pattern continued beyond the limits of the excavation.

The Discovery: The other finds

The site produced the usual range of finds from a Romano-British dwelling. Besides the building debris of stone and tiles and some wall plaster (some painted); pottery was found, some fine vessel glass fragments and a total of three bronze coins. The sparceness of the artifacts reflects the lack of invasive excavation, and that the rubbish deposits from the villa, in surrounding pits and ditches were not found.

The reconstructed panel: The Idea

During archaeological excavations all sorts of bits and pieces are found in the soil, fragments of pottery, coins and bits of stone. It is possible to say what these pieces might be and the period they were made but not necessarily why they were there. Pots get broken, and brooches lost, but all the bits of stone....

​From Lopen there were about 20 bags of stones (possible 6-7000 bits) that were roughly cut into cubes, from tiny fingernail sized to 6cm blocks. These were all thought to have become dislodged by tree roots and the digging of a well during the 1600 years below the surface from the mosaic floors of the villa.


​The mosaic had to be recovered by soil again for its own protection from weathering, frost and other damage and we had nothing tangible left to remind us of its beauty.

What could we do?

​Could we try to reconstruct a part of it, working in the way of the Roman craftsmen, using the same materials? We would need funds so we could get the skilled people to work on the project which included producing a film and developing a web site. Lopen is a small village so The Lopen History Group supported an application for money from The Local Heritage Initiative, who awarded the substantial grant.

The Reconstructed Panel: Cutting to Size

We were pleased to find that we had just about enough to complete the cantharus. Some pieces had to be cut down to fit. We only used larger tesserae, and some pieces of broken Roman roofing tiles from the site.


​The mortar used in the Villa floor had been analysed. Brick or tile dust was used in the lime based mortar to give strength by the Romans and us. The test block gave a chance to work out how to smooth the upper surface and if we were likely to need to polish it.


​We chose to lay the panel on a slab of blue lias, a very stable stone and the type from which the blue tesserae had been made, and possibly from a neighbouring quarry.

The design was pricked through from the traced plan onto the fresh bed, the pricked holes joined up using red ochre. The surfaced was scratched through so the next layer of mortar could be keyed onto it. Image from larger photo showing the underlying drawing. With Gina Wrights help, small areas of the red ochre pattern were covered with fresh mortar and the tesserae from the laid out pieces were transferred to the final panel.

The surface needed to be levelled before the mortar set hard. In the way that the lines and forms in the design were copied Gina and Nick felt that at least two mosaisists had been involved in it's original construction. There were very few spare pieces available and this lead to tension during the last day of the construction.

The grouting gave it's thick final covering, when sponged off were at last able to get the sense of the glory of the floor when first completed in the 4th century A.D.

For the revealing to the general public it was covered with silver sand, so in brushing this away those involved with the early stages of the initial excavation could relive those amazing moments less than a year before.

Our Reconstruction



























Our reconstructed panel is available to see in Lopen Church. The completed cantharus. Photo Copyright © Pauline Rook

The Future for Lopen's Roman Villa Site

The protection of the mosaic is already in place; it was carefully covered after the recording was completed. However, the finding of the remains of this Romano-British villa and its beautiful mosaic floor gave us an idea of the importance of this house, but very little of the size of the building or the community it supported.  It is hoped that if money can be found further archaeological excavations may be carried out.

Protection against future damage of historical sites

The County Council Archaeologists are required to check if any previously known sites of archaeological importance are subject to planning applications, to prevent unauthorised damage.

Countryside Stewardship Schemes can be appropriate to enable payment to be made to compensate landowners for a 10 year period to use the land in a way that will not damage an historic site.


Tarmac Finders Award

The landowners have had their actions in allowing this important site to be excavated and recorded professionally has been recognised nationally. George Caton on behalf of himself and Nigel Osborne, went to Liverpool to receive the British Archaeological Award sponsored by the Tarmac Group on 7th November 2002. The Tarmac Finders Award is given for the best non-archaeologist who by chance discovers archaeological remains and reports them to the appropriate authorities.

Further Resources


The following reports are available as PDF files. To view these documents, Adobe Acrobat Reader is required and can be downloaded free from

Click on the PDF icon or file name to open the document or to download, open the document and click on file>save or right-click on the link and choose "Save target As" for Internet Explorer browsers or for Netscape users, choose "Save Link As".


Photogrammetry and the Lopen Mosaic

Mosaic plans: drawings to scale of the site and room plans and the mosaic layout


Websites of Interest

​The site of ASPROM, The Association for the Study and Preservation of Roman Mosaics. For membership and other information including news on the publication for the Society of Antiquarians of the four volumes illustrating all of the mosaics found in Britain. All are the work of David Neal and Steve Cosh, Volume II covering South West Britain expected to be published in April 2003.

​Covers a range of topics related to the Romans including and acess to a make your own mosaic.

​Gives an opportunity to view some of the images found in mosaics at Bignor and Fishborne.

Introduces the publication Current Archaeology, page for number 157, May 1998.

Time Team Video of the Dinnington Mosaic

Further Reading

Roman Food in Britain. Joan Alcock Tempus Press ISBN 0-75241924-2

Roman Somerset, Peter Leach. The Dovecote Press ISBN 1-874336-93-8

Roman Gardens and their Plants, Claire Ryley ISBN 0-904973-16-6

Roman Clothing and Fashion Alexandra T Croom. Tempus Press ISBM 0-7524-2512-9

Doctors and Diseases in the Roman Empire, Ralph Jackson British Museum Press. ISBM 0-7141-1398-0

Ancient Mosaics, Roger Ling. British Museum Press ISBN 0-7141-2218-1

The Romano-British Villa at Halstock, Dorset. Excavations 1967-1985. Natural History and Archaeological Society, R.N.Lucas. 1993 Monograph Series: Number Thirteen


Lopen Mosaic Gallery

Please note that no images should be downloaded from these pages without the express permission of the owners.  In the first instance please use this email contact to obtain permission, stating how you wish to use the images.


Credits & Thanks

Without the help of Stephen Cosh, Alan Graham and Hugh Prudden and the photographers, this web site would not have been possible. To them I owe a debt of gratitude.  Angela Naunton-Davies

Reconstruction Project Team

bottom of page