Subscribe now for full access and no adverts
In recent decades, development-led excavations on the outskirts of Cambridge have offered vivid insights into the area’s Roman past, revealing a once-intensively occupied landscape striped with roads and scattered with settlements and cemeteries, farmsteads and villa estates. One of the latest additions to this increasingly detailed map emerged during archaeological investigations immediately to the west of the village of Milton, about 4km (2.5 miles) north-east of Cambridge’s historic core. There, in 2023, Archaeological Research Services Ltd carried out a large, open-area excavation exploring 3.56ha (8.8 acres) ahead of the construction of the new Milton Police Station. As the recently published monograph A Landscape of Plenty: excavations on a Roman estate, Cambridgeshire (see ‘Further reading’) reports, this project uncovered a large and complex network of intercutting enclosures that are thought to represent part of the agricultural working area associated with a late Roman villa estate, as well as giving diverse insights into its inhabitants’ activities.
It is easy to understand why this would have been an appealing location for agricultural activity during the late Roman period. The site nestles on a low, sheltered terrace of the Cam valley, and was well-connected, with the River Cam flowing around 1.75km (1.1 miles) to the east and Akeman Street (a Roman road, also known as the Mere Way, which ran north-east from Cambridge into the Fens) passing about 500m (0.3 miles) to the west. Perhaps unsurprisingly, then, the wider area has already proven to be rich in archaeological remains. Extensive fieldwork on adjacent sites such as Milton Landfill (in 1994-1998 and 2007-2013) and Milton Park and Ride (2017) revealed evidence of Bronze Age and Iron Age occupation, as well as Roman features including ditches and planting trenches, a barrow containing both inhumation and cremation burials, and building material from a possible villa. As we will explore here, excavation of the Police Station site has been equally illuminating.
An evolving landscape
The site’s story begins long before the Roman period: the earliest hints of a human presence came in the form of seven worked flints – certainly prehistoric, and probably produced during or after the late 4th millennium BC – though these were all residual finds recovered from Roman or later features.
There was also evidence of rather less ephemeral activity (referred to as Period 1 in the monograph) dating to the middle-to-late Iron Age, with a handful of features possibly representing a peripheral part of a larger settlement. Most prominent among these was a meandering ditch that ran for at least 64m (210ft) in the south-eastern part of the site and, just 1.5m (4.9ft) south of its line, a possibly contemporary ring ditch that may have once surrounded a roundhouse with an internal diameter of 9.5m (31ft). This circular gully contained barley and wheat remains, as well as weed seeds, providing evidence for small-scale, localised crop-processing. Over in the north-eastern part of the site, the team found an isolated, shallow pit containing fragments of middle-to-late Iron Age pottery and charcoal, while, further to the south (about 41m/135ft from the western limit of the meandering ditch), there was a 1.2m-deep (3.9ft) well with vertical sides. Its lower fills produced a number of pieces of waterlogged worked wood, while its upper level was formed from a dump of rubbish including middle-to-late Iron Age pottery and a cattle bone that was radiocarbon dated to 398-209 BC.
There were no certain early-to-middle Roman features, perhaps suggesting that the area was covered by open fields at that time, but the team did recover individual finds of this period (again, residual discoveries in late Roman features), pointing to the presence of a contemporary settlement lying somewhere nearby but outside the excavated area. These comprised a scatter of pottery, including imported types such as 49 fragments of Samian ware and a few fragments of Spanish olive oil and Gaulish wine amphorae; a couple of late 1st-century to early 2nd-century copper-alloy coins; a copper-alloy bow brooch of Colchester Derivative type; and several pieces of glass vessels.
By contrast, this patchy picture becomes clearer – and much busier – during the later Roman period, when the landscape’s archaeology explodes into vibrant life. At this time, the apparently open environment was transformed into a complex web of enclosures, ditches, trackways, small timber structures, pits, and a diverse array of other features including wells or waterholes, a pond, and an oven. We will unpick these in more detail below, but analysis of pottery, coins, and other finds, combined with radiocarbon dating, indicates that this activity probably began in the mid-3rd century (Period 2), peaked in the mid-to-late 4th century (Period 3), and continued through the late 4th century, possibly extending into the 5th century (Period 4).
Evidence of occupation
Features dating to Period 2 were concentrated along the southern edge of the site, perhaps suggesting that the main focus of activity initially lay further south still, beyond the limits of our investigation. There were some interesting discoveries from this phase, however, including the near-complete burial of a large dog (comparable in size to a modern golden retriever) which had been carefully laid in a ditch. Another notable find assigned to Period 2 was an isolated, keyhole-shaped oven that was located in the south-western part of the site, apparently outside any known enclosure. Measuring 2.64m (8.7ft) in length, it had a chamber at the north-west end and a stokehole to the south-east, and while it is uncertain what the oven was used for, its fills contained quantities of charred cereal grain and fragments of oak- and birch-charcoal, while radiocarbon dating suggests it was operating in the first half of the 3rd century.
Helpfully, in Period 3 the main area of occupation appears to have extended further to the north, bringing it within range of the southern and central parts of our excavation. New enclosures were created at this time, typically cutting across the lines of the Period 2 features, though some of the ditches followed, or were influenced by, the layout of their predecessors. This trend continued into Period 4, with new enclosures and boundary ditches often recutting or following/extending the same or similar lines as those from Period 3. The focus of activity also appears to have continued its drift northwards, with most Period 4 enclosures concentrated in the central and particularly the north-central parts of the site.
As well as ditched enclosures, the team identified rather smaller outlines: the footprints of a few possible timber structures, picked out by post-holes and possible beam-slots. These were all small and very simple, but they are significant finds, as the remains of definite Roman timber buildings are rarely encountered during archaeological work in Cambridgeshire – with the notable exception of Cambridge Archaeological Unit’s excavations at Camp Grosund, near Earith, which discovered a Fenland port settlement on the River Great Ouse, boasting a remarkable 62 timber buildings of platform, beam-slot, and post-built type, located c.18km (11 miles) north-west of the Police Station site (see CA 295). The general lack of such structures in this county and many other parts of England is presumably because they were typically of sill-beam construction, with shallow footings that are easily damaged by ploughing and perhaps also by machine-stripping prior to excavation. To explore one of the Milton examples in more detail, Structure 3 was discovered in the western-central part of the site and is thought to date to Period 2. Measuring c.14m by c.3.5m (46ft by 11.5ft), the structure had no internal partitions and was defined by a fairly unsubstantial subrectangular ditch; it is tentatively interpreted as a timber beam-slot building, which may have housed both humans and livestock.
Interpreting the site
The layout of the late Roman enclosures was strikingly compartmentalised, suggesting that corralling livestock had been a key role for the site – and the importance of animal husbandry was abundantly clear, too, from the substantial and well-preserved assemblage of more than 5,000 animal bones recovered from Roman contexts. Of those that could be identified, the majority (65.7%) came from cattle, with a smaller but still significant proportion of sheep (22.4%), some horses and possible donkeys or mules (9.4%), and a very small amount of pig (2.5%) – relative proportions that remained fairly similar throughout Periods 2-4. Examination of the bones indicated that the cattle were primarily raised for traction (that is, pulling ploughs and carts) and presumably manure to support arable farming, while they and the sheep would have also provided meat and milk/cheese, probably for fairly local consumption. Adding to ideas of dairy production, we found four fragments of ceramic vessels with holes drilled through their base, which were probably cheese-presses or strainers.
The site’s inhabitants were also growing crops and, in contrast to the largely local consumption indicated by the animal remains, environmental samples produced cereal grains and chaff in such quantities as to suggest that spelt wheat and, to a lesser extent, emmer wheat were being produced for export to a wider area as well as being eaten closer to home. These crops were being processed on site, too, as attested by the recovery of many worn fragments of probable rotary querns.
Taken together, the features and finds recorded on the Police Station site speak of an intensive agricultural working area, which may well have formed part of a late Roman villa estate. While an actual villa structure remains elusive, many fragments of Roman ceramic building material, including tegulae, imbrices, brick, and flue tiles, were found scattered in ditch fills across the site, suggesting that this kind of complex could lie nearby, most likely in the unexcavated area immediately to the south. Several fragments of probable building stone – including two stone roof tiles (one with a nail hole), a possible flagstone, and a large (c.27kg/59.5lb) piece of roughly shaped limestone that would have been brought to the site from further afield – add to this picture. Tantalisingly, a previous excavation carried out just 200m (0.1 miles) to the south at Milton Landfill in 1994 revealed roof tiles, box tiles, and worked stone, also thought to represent destruction material from a villa. These were found in the fills of Roman ditches in the northern part of that site; might the missing building lie between the two excavations?
Three infants of the same age, very likely triplets, had been placed in a pit cut into the side of a Period 4 enclosure ditch.
If so, our site could represent a new addition to the many late Roman villa estates that are already known to have existed in the area around Cambridge and to the south of the Fens. Excavated evidence testifies to the Police Station site being involved in the production of crops on a considerable scale, and the site was well-connected to potential distribution and consumption centres by road and water – not just via Akeman Street and the River Cam as mentioned above, but also by a Roman canal (today known as the Old Tillage/ Car Dyke) which passed c.3km (1.9 miles) east-north-east of the site and connects the Cam near Waterbeach to a former tributary of the Great Ouse. As for where the estate’s produce was destined, the nearby nucleated settlement or small town of Roman Cambridge (Duroliponte) could have served as a distribution centre for agricultural products from the surrounding area via state-controlled supply and marketing networks, which probably ultimately linked to military and urban consumption centres elsewhere in Britain, and perhaps also on the near Continent.
Diverse discoveries
Complementing the features described above, what can artefacts add to our understanding of the site’s function and its possible status? The excavation recovered large quantities of pottery – nearly 7,000 sherds, weighing about 140kg/309lb, and while much of this suggested basic, utilitarian domestic activity, there were a few examples that hinted at higher-status habitation somewhere nearby, including a rare costrel (flask)-like vessel and fragments of four flagons decorated with human faces. There were a few glass vessels as well, including fragments of two 4th- and 5th-century beakers.
The fairly large number of 68 copper-alloy coins found on the Police Station site seems more in keeping with a villa site than a farmstead, and while coin use seems to have only become widespread there at a relatively late date (sometime between AD 330 and 350 instead of the late 3rd century as is typical on other nearby rural sites), this use also extended rather later than at most of its neighbours, continuing on a considerable scale into the late 4th century and down to the end of the Roman period – a pattern that is seen at some other known villa sites in the wider region. Of the 61 Roman coins from the site that were closely datable, 19 were issued AD 364-378 and 11 were issued AD 388-402, the latter being among the latest Roman small-change supplied to Britain.
Coins were not the only copper-alloy Roman finds produced by the excavation: among the more notable objects of this material, we found a well-preserved octofoil brooch with a pronounced central boss and red and blue enamel decoration (probably datable to the 2nd to 3rd centuries); several late Roman bracelet fragments; and a 2nd- to 4th-century dining spoon that may originally have been silver-plated. While these items are by no means exceptional, they could point to higher-status activity close to the site, and it is possible that they might have been worn or used by the inhabitants of a hypothetical Roman villa.
As for ironwork, this was mostly fairly consistent with what we might expect from a working agricultural rural site, including dozens of nails and hobnails, as well as a few structural pieces, tools, and knives. There were also several bone and antler objects, mostly fairly utilitarian items such as handles and bone cases for iron sewing needles (the latter being a very rare find-type), but there was a simple, double-sided composite comb of a well-established local type, too, made from antler and datable to c.350-425. We also found an antler pestle and possible wall hook, and evidence that bone and antler were being worked on the site, with working waste indicating small-scale production focused on the manufacture of pins and handles respectively.
Objects like these paint a vivid picture of the eclectic activities that were being carried out on the site – and, more poignantly, we found physical traces of some of the agricultural community’s youngest members: the burial of three infants of the same age, very likely triplets, who had been placed in a pit cut into the side of a Period 4 enclosure ditch, probably in the late 4th century. They were accompanied by a number of iron nails, perhaps suggesting that the babies had been laid in a coffin or box before they were committed to the ground. If these were triplets, they would be only the second set recorded from Roman Britain; the other known example, dating to the late 1st century, was found at the Roman small town of Baldock, Hertfordshire, and they are also thought to represent stillborn babies or infants who had died soon after delivery, highlighting the dangers of multiple births before modern medicine.
Infant burials were discovered in other Period 3 and 4 ditch fills, too; this informal approach to the burial of babies is typical of Roman funerary customs, which often used such liminal locations for interment of the very young, and no other human remains were identified on the site. It is interesting, however, that a Roman barrow containing both inhumations and cremations of adults and children was discovered during the Milton Landfill excavation in 1995, just to the south of our site. Perhaps this was the formal burial location for the estate’s older occupants.
The end of the estate
Although the agricultural working area appears to have continued in use as late as the end of the 4th to early 5th century (for example, a sheep frontal bone from the fill of a Period 3 enclosure ditch produced a radiocarbon date of AD 348-535 at 92.9% probability and AD 379-433 at 65.5% probability, which strongly suggests that this ditch was filled in the mid-4th century or later, probably in the late 4th to early 5th century), it does not seem to have survived the end of official Roman administration in Britain. Excavated evidence suggests that the late Roman agricultural working area went out of use around AD 400 or shortly afterwards; all of the enclosure and boundary ditches were infilled at about this date or in the following decades, either through gradual silting, which might suggest abandonment, or in a deliberate act that could mark a shift to larger, open fields.
This development probably related to the wider political, military, and economic upheaval that marked this period, as state payments to Britain and military salaries ceased in a withdrawal of imperial support that could have prompted economic collapse and a breakdown of the villa-estate system. This opens up important questions about how the landscape of the Police Station site was used during the immediately post-Roman period and into the Anglo-Saxon period. Did local woodland regenerate, or were fields still tilled or given over to grazing? The infilling of the ditches suggests that land divisions – and potentially the ownership or tenure that they signalled – were deliberately changed as new systems of control, governance, coercion, and military and political dominance took hold.
For now, we can only speculate as to what this transition may have looked like. Although the team identified a large number of parallel ditches and/or furrows indicating that the site had served as an agricultural field during the medieval/post-medieval period (Period 5), no features or finds of Anglo-Saxon date were recorded anywhere in the excavation. Perhaps tellingly, though, the alignment of the Period 5 features was similar to that of their Roman predecessors. This might reflect the enduring influence of the line of nearby Akeman Street on adjacent field- and property-boundaries – but it could also indicate that the earlier fields had continued in use in some form, shaping the open field systems that emerged long after the Roman period ended.
Acknowledgements:
We are grateful to Cambridgeshire Constabulary for their funding and support. The constabulary also funded a series of value-added work including well-attended public open days towards the end of the excavation. These featured information panels about the site, stalls with objects and finds experts on hand to discuss them, a Roman history expert dressed as a Roman soldier, site tours, and craft activities such as as making and decorating pottery. We gave presentations to a local society and school, too, outlining the results of the excavation, while a small exhibition of artefacts from the site was shown at the nearby Denny Abbey and Farmland Museum. Display boards detailing the project’s findings will remain on the excavation site. Further scientific analysis on the probable triplets is planned, which will lead to a journal article.
Further reading:
F M Morris and J E R Davey (2025), A Landscape of Plenty: excavations on a Roman estate, Cambridgeshire. Available in hardback (£60) and as an open-access .pdf eBook from the publisher at http://www.archaeopress.com. It is also free to download at https://archaeologicalresearchservices.com/projects/cambridgeshire-southern-police-station-milton.
All Images: Archaeological Research Services Ltd
You must be logged in to post a comment.