● Documented
Engleton is a villa
Excavation revealed a winged courtyard building with principal rooms, a bath-house wing and portico, with occupation from at least the 2nd to 4th centuries.3
Featured Investigation · How We Know Things
The Missing Villas of Western Britain
Engleton, Ffrith and what counts as a Roman villa. On a modern map, the villa dots thin dramatically as we move west. That pattern matters — but before we turn it into a story about wealth, resistance or Romanisation, we have to ask a harder question: what exactly created the dots?
Our question: if Engleton is securely recognised as a villa, while Ffrith has masonry, heated rooms, plaster, opus signinum and a substantial high-status building yet remains functionally uncertain, what are we really mapping when we map “Roman villas”?
Evidence first
What survives the first test?
Three facts must be kept separate. Engleton is a securely identified villa. Ffrith contains substantial Roman buildings but its function remains debated. And the Rural Settlement of Roman Britain map is an exceptional research synthesis that explicitly warns users that it is not an exhaustive record of all Roman evidence.
● High-status building
Ffrith is harder to name
Excavation records masonry, heated space, painted plaster and opus signinum. The HER has described the building as “villa ?” with “bath house ?” as an alternative; later work still found no secure function.46
● Map with limits
A dot is a database decision
The ADS resource is built from excavated evidence selected by project criteria. Its own guidance says it is not intended as an exhaustive or authoritative record of every Roman site.12
The investigation is therefore not “Can we put more villas on the map?” It is: what has to happen to an ancient place before it becomes a modern villa dot?
How we know things
Historical confidence is built in layers
A distribution map is the end of a long chain of decisions. To understand an apparent absence, we have to work backwards from the symbol on the screen to the archaeology beneath the ground.
Step 1Read the pattern
What does the map actually show — all Roman sites, excavated settlements, or only sites assigned to one category?
Step 2Check the rules
What evidence qualified a site for the dataset, and what kinds of investigation were excluded or insufficient?
Step 3Return to the trench
Look at walls, floors, heating, finds, chronology and context before accepting a label.
Step 4Test the label
High-status architecture can belong to a villa, bath-house, mansio, military facility or another complex.
Step 5Keep uncertainty
“Villa?”, “bath-house?” and “high-status building” are evidence statements, not failures of archaeology.
The apparent pattern
“There were hardly any villas in the west”
A villa-filtered map makes a powerful visual argument. Dense clusters across parts of southern and eastern Britain thin towards Wales and the western uplands. There is a genuine regional story here, but the first question is not why the west “rejected villas”. It is how the map was assembled.
● Visual impression
Few dots = few villas?
The eye naturally turns absence of symbols into absence in the ancient landscape.
● Dataset
Excavated evidence selected by criteria
The project prioritised evidence capable of supporting settlement analysis and notes that small-scale work often lacked enough morphological information for classification.1
● What must be tested
HER records and individual excavations
The map itself directs researchers back to Historic Environment Records for a fuller and more current archaeological picture.2
Filter 1What survived
Stone robbing, ploughing, erosion and later building alter what remains.
Filter 2What was recognised
A heated room or wall fragment must first be noticed and recorded as Roman.
Filter 3What was excavated
Modern development and research history determine which sites receive trenches.
Filter 4How it was classified
The final label decides whether a site appears in a particular map query.
That means our modern map is created from: what survived + what was recognised + what was excavated + how archaeologists classified it.
The control case
Engleton: what a secure villa looks like
Engleton, near Brewood in Staffordshire, gives us a useful control. The label is not being inferred from one luxurious fragment. Excavation in the 1930s recovered enough of the plan and associated features for the site to be securely understood as a villa complex.3
● Plan
Winged courtyard
At least four principal rooms were arranged within an east-facing winged courtyard layout, with a portico.3
● Comfort
Bath-house wing
A bath-house formed part of the complex — exactly the sort of architectural investment often associated with elite rural residence.3
● Duration
At least two centuries
Three building phases were identified, with artefacts indicating occupation from at least the 2nd through 4th centuries.3
The challenge case
Ffrith: the building that refuses a simple label
Ffrith in Flintshire sits on the Chester–Ffrith–Caer Gai Roman route and has produced Roman structures and artefacts over generations of discovery. The most important building sequence at Arosfa and Emdyklin began with timber occupation around AD 80 and was replaced by masonry around AD 120–125. The surviving evidence is unmistakably Roman; its exact function is not.45
● Excavated
Seven or more rooms
The 1967–69 work revealed a masonry building with at least seven rooms. Recorded features included wall plaster, an opus signinum floor and evidence for a heated room.4
● Status
Comfort and long-distance contacts
Later evaluation confirmed at least one substantial, well-constructed building and described the settlement as relatively high status; imported and regional pottery point to wide-ranging contacts.6
● Function uncertain
Villa? Bath-house? Something else?
The HER preserves “Villa ?” with “Bath house ?” as an alternative. Wessex Archaeology later concluded that no excavation had produced a clear indication of the settlement's function.46
A useful correction came in 2005. A curving wall previously interpreted as an apsidal end was found to be irregular and crudely built; Wessex Archaeology judged it much more likely to be an enclosure wall. Yet the same evaluation strengthened the case for a substantial high-status building nearby. Good archaeology can weaken one attractive interpretation while strengthening the wider site.6
The archaeology of disappearance
How does a villa vanish from the map?
● Stop and test the absence
A blank area on a distribution map can represent ancient reality — but it can also contain sites that were destroyed, buried, poorly excavated, incompletely published, or never confidently classified.
The important discipline is not to use these biases as an excuse to invent missing villas. It is to recognise that the number of symbols on a map is not identical to the number of elite households that once existed.
Survival: Were walls robbed for later buildings, damaged by quarrying, ploughed flat or built over by a modern village?
Discovery: Has the landscape been exposed by development and rescue archaeology as intensively as regions with heavier modern construction?
Excavation scale: Did a trench expose enough of the plan to distinguish a villa from a bath-house, roadside building or military facility?
Publication: Were early discoveries recorded to modern standards, with finds tied securely to stratigraphy?
Classification: Does a database require a confidently assigned settlement type before a site appears under a “villa” query?
Regional difference: Did western communities genuinely express wealth and status through settlement forms that were less villa-like than those of the south and east?
Ffrith demonstrates several of these filters at once. Its Roman remains lie within a later village; early work was fragmentary; the 1967–69 records were basic and many finds became effectively unstratified; and modern evaluation could confirm status more confidently than function.56
Competing interpretations
Three models fit the western villa pattern differently
Model A · too simple
The dots are the landscape
Western Britain possessed almost no villa-style elite rural settlement, and the modern map can be read as a direct census of that ancient reality. This asks the dataset to be more complete than its creators claim.
Model B · plausible
Many villas are simply missing
The apparent regional difference is largely produced by survival and research bias. Those biases undoubtedly matter, but this model risks explaining every absence away and flattening genuine regional diversity.
Model C · best working model
Real difference, filtered evidence
Western settlement traditions genuinely differed from the villa-rich lowlands, while preservation, excavation history and classification also affect the number and type of dots we see. Engleton and Ffrith show both processes at work.
The best interpretation does not choose between “the map is right” and “the map is misleading”. It asks what the map is right about, what it cannot measure, and which archaeological decisions created its apparent certainty.
What can be seen today?
Two sites, mostly beneath the present
Neither Engleton nor Ffrith offers a standing villa ruin in the way a visitor might imagine from a guidebook. Their importance lies in buried archaeology, old excavation records and the relationship between modern settlement and ancient remains.
● Engleton
A protected buried complex
The villa survives as a Scheduled Monument west-north-west of Engleton Hall. Historic England notes that archaeological remains survive despite quarry damage at the southern end.3 Treat it as protected archaeology rather than a public ruin.
● Ffrith
Roman archaeology beneath a village
Recorded Roman buildings, findspots, historic excavations and the Chester–Ffrith–Caer Gai road are distributed through and around the modern village. Two areas have Scheduled Monument status.67
A field research checklist
Start with the map, never finish with it. Record what layer or filter you are looking at and what qualifies a dot for inclusion.
Open the HER record. Look for question marks, alternative site types and the date of the underlying excavation.
Ask what was actually found. Plan, wall construction, heating, floor surfaces, pottery, coins and building materials may carry different levels of confidence.
Separate status from function. A high-status Roman building is an archaeological conclusion; “villa” may be a further interpretation.
Look for later disturbance. Modern houses, quarrying, roadworks and stone robbing can determine which part of a Roman complex survives to be recognised.
Respect access and designation. Scheduled or privately owned archaeology should be researched from public rights of way, records and published excavation rather than entered without permission.
Evidence audit
What can we responsibly say?
| Claim | Evidence | Verdict |
|---|---|---|
| Villa distributions are much denser in parts of southern/eastern Britain than in the west. | Large settlement syntheses identify strong regional differences in rural settlement traditions. | Documented pattern |
| The villa dots form a complete census of ancient elite rural residences. | The ADS resource explicitly states that it is not exhaustive and uses defined inclusion criteria. | Not demonstrated |
| Engleton is securely classified as a Roman villa. | Winged courtyard plan, principal rooms, bath-house wing, portico, multiple phases and long occupation. | Documented |
| Ffrith contained substantial, relatively high-status Roman architecture. | Masonry, hypocaust evidence, opus signinum, wall plaster, ceramic building material and imported/regional pottery. | Strong |
| Ffrith was definitely a villa. | HER records “villa?” / “bath house?”; later evaluation found no clear indication of function. | Possible, unproven |
| All western “missing villas” are simply undiscovered. | Would ignore genuine regional differences in settlement traditions and use absence bias as a universal explanation. | Too strong |
| The best reading combines ancient regional difference with archaeological and classificatory filters. | Regional syntheses plus the contrast between secure Engleton and ambiguous Ffrith. | Best working model |
Investigation verdict
The missing dots are not missing proof.
Western Britain does appear to have had a different rural settlement history from the villa-rich landscapes of parts of the south and east. But a villa map cannot, by itself, tell us how many wealthy rural households existed, how they expressed status, or how many substantial buildings sit at the edge of our modern categories.
Engleton shows what happens when plan, chronology and function align strongly enough to make “villa” a secure label. Ffrith shows the opposite intellectual virtue: substantial Roman architecture can be confidently recognised while its function remains open. The uncertainty is not a gap to hide. It is evidence about how archaeological knowledge is made.
How we know things: an archaeological map is not the ancient landscape. It is the ancient landscape after survival, recognition, excavation and classification have each had their turn.
Sources and evidence
What this investigation rests on
1. Allen, Blick, Brindle, Evans, Fulford, Holbrook, Lodwick, Richards & Smith, The Rural Settlement of Roman Britain: an online resource — Introduction/Overview. Explains the project dataset, the inclusion criteria and the limits of small-scale investigations for settlement classification. ADS overview.
2. Archaeology Data Service, RSRB Map Interface guidance. Explicitly states that the resource is a synthesis of excavated evidence and is not intended to be an exhaustive or authoritative record of all Roman evidence; recommends consulting HERs. Map guidance.
3. Historic England, Roman villa 380m WNW of Engleton Hall, NHLE 1006082. Records the winged courtyard plan, four principal rooms, bath-house wing, portico, three building phases and occupation from at least the 2nd to 4th centuries. Historic England record.
4. Heneb / former CPAT HER, PRN 100024, Ffrith Roman site, Building VI, Arosfa & Emdyklin. Records the masonry building, at least seven rooms, plaster, opus signinum, heated-room evidence and the alternative interpretations “villa ?” and “bath house ?”. HER record.
5. Heneb HER, PRN 38257, Ffrith excavation 1967–69. Important for the timber-to-masonry sequence and also for the limitations of the excavation record: basic sections and finds not securely related to stratification. Current HER record.
6. Wessex Archaeology, Ffrith, Flintshire, Wales: Archaeological Evaluation, Time Team fieldwork (2005). Reassessed the curving wall as more probably an enclosure than an apse; confirmed at least one substantial relatively high-status building, possible ancillary structures, and concluded that the site's precise function remained unclear. Wessex report PDF.
7. Heneb / former CPAT, Ffrith community archaeological index. Shows the concentration of recorded Roman buildings, excavations, finds and Roman-road records within the modern village and its surroundings. Ffrith site index.
8. Fulford & Holbrook et al., The Rural Settlement of Roman Britain. The wider synthesis identifies major regional contrasts, including western and northern traditions with stronger continuity of roundhouse and enclosure forms. Project volume, Chapter 1.