The Roman and Early Romano-British Landscape Between Pevensey and the Ouse

The stretch of downland and valley between Pevensey (Anderida) and the mouth of the Ouse has generally been treated as two separate stories: the naval headquarters at one end, and the Barcombe/Bridge Farm settlement at the other. Read together, however, the intervening landscape — road, field system, salt-working, villa estates, two coastal harbours, and two distinct defensive lines — describes a single working corridor, most plausibly organised to supply and defend the naval establishment at Pevensey through the late 3rd and 4th centuries, and still legible in the landscape as it passes into the Romano-British transition of the 5th century.

This page sets out that corridor as a connected hypothesis, built from excavated, surveyed, and place-name evidence at each of its component sites. As throughout this site, the argument is presented in a hedged register: what is established by excavation or documentary record is distinguished clearly from what is inferred from landscape logic, continuity, or pattern.

Two river systems drain the Weald through the South Downs into the Channel across this stretch: the Ouse, reaching the sea near modern Newhaven, and the Cuckmere, a shorter and shallower valley to its east. Between and around these valleys lies a belt of downland dip-slope — chalk grassland and cultivable interfluve — running from the Ouse in the west to the Pevensey Levels in the east. A single road, Margary 142/143, runs the length of this belt and physically connects a settlement hub at each end, with two watch points sited where it crosses the two river gaps.

At the tidal head of the Ouse, the Romano-British settlement at Bridge Farm, Barcombe, has been under excavation by the Culver Archaeological Project since the mid-2000s. The site was occupied continuously for around 350 years from the later 1st century AD, with enclosure ditches and industrial features consistent with sustained, organised activity rather than a short-lived farmstead.

Excavation seasons in 2024–25 exposed two roads converging on the site: one running north toward London, and one running east, directly toward Pevensey. This is the most significant single piece of physical evidence for the corridor argument: it is not an inferred route, but an excavated road surface, connecting the Ouse valley settlement to the Pevensey direction.

The Culver team's own working assessment is that the scale and longevity of the site may represent a small Roman town rather than a roadside settlement. This has not been formally published and should be treated as an informal inference by the excavators rather than an established classification, but it is worth noting that a site of sufficient scale for its own excavators to reach for "town" status is exactly the kind of hub that would be expected anchoring a state-directed supply corridor.

The road exposed at Bridge Farm running toward Pevensey aligns with Margary's numbered roads 142 and 143, the latter running via Stone Cross to Jevington, directly on the approach to Pevensey. This single road is the organising spine of the whole corridor: it carries the grain traffic from Ripe, passes the watch positions at Mount Caburn and Burlough where it crosses the two river gaps, and terminates at Pevensey itself. A separate branch, recorded in Margary's account of the wider Lewes road network, runs south over the Downs from the Ripe area to Seaford, connecting the same hub to the coastal watch chain discussed below.

Roughly midway along the eastern road, on the downland interfluve between the Cuckmere and Ouse valleys, Margary identified a field system at Ripe and Chalvington in 1940 (Roman Centuriation at Ripe, Sussex Archaeological Collections vol. 81). The fields form a rectangular grid on Roman actus/jugerum units — formally surveyed land division, not the organic, irregular field growth typical of unplanned farming. Later work (Peterson) has proposed situating this grid within a much larger hypothetical cadastral system radiating from Canterbury, some 80 km to the east; this wider claim is speculative and is noted here only for completeness, not relied upon.

A local record (SCATE East Sussex, drawing on Robert Masefield's Celts, Romans and Countrymen) already describes this same landscape in almost identical terms to the hypothesis developed here: the patchwork of small rectangular fields south of Ripe, centred on Hall Court Farm, is described as having been turned over to grain to feed the fleet at Pevensey, with produce carried down Langtye Lane to the boatyards. The same source notes a Roman villa to the southwest of Ripe and farmworkers' cottages, and records that Ripe itself had eight Domesday salthouses, separate from and in addition to Laughton's sixteen, with the Laughton Levels understood locally to have been a tidal inlet into the medieval period.

Planned, surveyed land division, an independently recorded delivery route to Pevensey, and Domesday-attested salt production at the same location together make Ripe/Chalvington the best-evidenced single component of the entire corridor.

Either side of the Ripe/Chalvington grid, the corridor is flanked by a run of known Romano-British sites. Beddingham and Plumpton villas sit in the Ouse valley near Lewes, close to the Barcombe hub. Bishopstone and Bullock Down are non-villa Romano-British settlement sites on the downland behind Seaford — working farms rather than high-status residences. Eastbourne villa, excavated by Sutton in 1952, sits at the eastern end of the corridor, a short distance from Pevensey itself.

The occupation dates of these sites span the 1st to 4th centuries broadly; whether each shows continuous activity into the Carausian period specifically (late 3rd century) has not been separately verified here and would need checking against individual site reports before being asserted as a synchronised system.

Domesday Book records sixteen salthouses at Laughton, a marshy manor on the edge of the Honour of Pevensey a few miles from Ripe, held by Earl Godwin in 1066 and valued at £15. Combined with Ripe's own eight salthouses, this gives the corridor a substantial, geographically concentrated salt-production capacity at Domesday. Salt-working sites are tied to fixed physical requirements — access to saltmarsh or tidal edge, and a fuel source for evaporation — that persist across periods more readily than settlement patterns do. Following the same continuity logic already applied elsewhere on this site (the Norfolk and Hastings salt pages), an operation of this scale is more plausibly a continuation or reoccupation of a long-worked landscape than a wholly new Anglo-Saxon or Norman development, though no Roman-period archaeological evidence has yet been identified at either Laughton or Ripe to confirm this directly.

If the inference holds, salt production here represents a third economic input into the corridor alongside grain at Ripe and timber/iron at Ashburnham on the eastern network — specifically, preservation capacity for meat and fish supplying a garrison and shipyard workforce.

The position of the Ouse's outlet to the sea has moved substantially over the historic period. Coastal geomorphological research indicates that in the Roman period the river reached the sea along the western margin of the valley, beneath the chalk cliffs of Castle Hill — essentially the site of modern Newhaven. It was only in the medieval period that a drifting shingle spit deflected the outlet eastward toward Seaford, and only in 1539 that a new cut restored the mouth to close to its Roman-period position, giving the modern town its name.

Newhaven Fort, on Castle Hill, is recorded as occupying the site of an earlier Roman fortification. This gives the corridor a physical western terminus: a fortified position on the western headland of the Roman-period Ouse mouth, at the point where the river, and by extension the Barcombe/Bridge Farm hub upstream, met the sea. Castle Hill itself is a further example of the place-name pattern already established elsewhere on this site — the Cold Harbour/castle-/borough-/bury- markers along the Classis Britannica boundary — a "castle" name attached to a site with a genuinely recorded Roman fortification beneath it, at a defensible harbour-mouth headland.

The corridor appears to be watched by two distinct systems, doing two different jobs. The first follows the road itself. Margary 142 crosses the Downs from Pevensey to Lewes, and in doing so crosses two natural chokepoints where the Cuckmere and Ouse cut through the chalk escarpment. Both crossings are overlooked by defensible high ground.

Mount Caburn, an Iron Age multivallate hillfort above Glynde, overlooks one of the few natural gaps along the whole length of the Downs escarpment, where the Ouse cuts through, directly above the approach to Barcombe/Bridge Farm. Nineteenth-century excavation by Pitt Rivers recovered Roman pottery from the upper ditch fill at the associated Ranscombe Camp, confirming some Roman-period activity on the hill, though this should not be read as confirming a maintained Roman-period garrison or refortification without further evidence — an older theory of Saxon-period remodelling (Curwen) has been revised by more recent excavation and is not relied upon here.

Burlough sits opposite Alfriston at the point where the Cuckmere breaks through the Downs, centred around 50.817, 0.171. The visible earthwork here is conventionally dated to the Norman period, built to guard the Cuckmere crossing, and its form as a genuine motte-and-bailey has been disputed by at least one assessment (King), which considers it a natural feature. However, Iron Age/Romano-British pottery sherds and worked flints, Mesolithic and Neolithic, have been recovered in the immediate vicinity, evidencing long-term activity at this specific chokepoint well before the Norman-period structure. The name itself is consistent with the burh + hlēow, "refuge fort," reading already applied extensively elsewhere on this site's Burghal Hidage page — ten of roughly thirty Wessex burhs are read there with the same hlēow element, including Hastings, Lewes, Southampton, Pilton, Axbridge, Lyng, Eashing, Worcester, Warwick and Bridport, with Lyng's Burrow Mump (Burh hlēow munt) the closest direct parallel to Burlough's construction. On this reading, the name may preserve an older naming layer for a fortified refuge point at this crossing, independent of the specific, and later, earthwork visible today. Alfriston itself, on the opposite bank, is associated with the early Anglo-Saxon cemetery at High and Over, one of the more significant pagan-period burial sites in this part of Sussex, so whatever its origin, the crossing was clearly a place of continued significance across periods.

Pevensey, Burlough and Mount Caburn happen to fall on a striking near-straight west–east line: Pevensey and Burlough sit at almost identical latitude, roughly seven miles apart, with Caburn a little to the north-west. Rather than treat this as a deliberately surveyed alignment, the more economical explanation is that all three points are simply sited on or immediately astride the same road, Margary 142, wherever it meets defensible high ground at a river-gap chokepoint. The apparent line is a function of the road's route, not evidence of independent intervisibility or coordinated design.

A separate system watches the sea approach rather than the inland river gaps, running along the coastal edge of the Downs from Pevensey west to the Ouse mouth. Pevensey itself, Anderida, is the naval base and the eastern anchor of both lines. Combe Hill, above Eastbourne, is a scheduled Neolithic causewayed enclosure, securely dated by excavation to c. 3500–3300 BC; no Roman-period evidence has been identified here, and its inclusion in this line rests on its commanding position on the scarp above the Eastbourne approach rather than on direct period evidence, a distinction worth keeping clear wherever the line is presented. Belle Tout, at Birling Gap, is an Iron Age promontory enclosure with an earlier Late Neolithic/Bronze Age settlement inside it, now substantially lost to cliff erosion; a 1971 cliff fall exposed a hand-cut shaft or ritual well of Iron Age or Romano-British date near the site, since itself lost to further erosion, and this is the strongest direct period-evidence for Roman-era activity anywhere on the coastal line.

Seaford Head sits on the eastern side of the Roman-period Ouse mouth, directly across the estuary from Castle Hill. Its clifftop platform is reconstructed as having stood between 0.6 and just over a mile further out into the Channel in the late 3rd century than today, based on measured chalk erosion rates of 0.5–1 metre per year projected back roughly 1,735 years; cliff retreat is episodic rather than linear, so this is a first-order range rather than a precise reconstruction. Nineteenth-century antiquarian sources record "the remains of a large Roman camp" on Seaford Heights, once speculatively, and incorrectly, linked to Anderida itself — that identification is superseded and should not be repeated, but the underlying record of Roman-period remains at the site is worth noting as a separate, independent thread. Castle Hill at Newhaven forms the western terminus, with its recorded Roman fortification discussed above.

Combe Hill's evidence is the weakest link in this line and should be flagged as such wherever the line is presented; Belle Tout and Castle Hill carry the strongest direct period evidence, with Seaford Head's case resting on its geographic relationship to Castle Hill and the reconstructed cliff position rather than on-site Roman finds.

Grain, salt, and a single spine road link two fortified maritime positions — Pevensey in the east, a port at Castle Hill in the west — while two distinct watch systems cover the approaches: an inland line following the road across the river gaps at Caburn and Burlough, and a coastal line following the cliff edge at Combe Hill, Belle Tout and Seaford Head. Evidence quality varies significantly across the corridor. Bridge Farm's excavated roads, Ripe's centuriation, and Castle Hill's recorded fort are the strongest anchors; Combe Hill's role in the coastal line and the dating of Burlough's earthwork are the weakest and should be presented with appropriate caution throughout. The corridor is best understood as a supply and defence system organised around Pevensey's needs as a naval base, most plausibly reaching its fullest organised form in the Carausian period in the late 3rd century, and persisting, in some form, into the 4th century.

What changes across this corridor through the 4th century in terms of villa abandonment, the end of centuriated management at Ripe, and the fate of Bridge Farm remains to be developed with excavation-dated evidence and is not addressed here. What can be added now is a layer of Saxon-period place-name evidence, tested individually against landscape and documentary fact rather than taken on trust, which shows the same corridor continuing to function as an organised, water-linked agricultural and landing-stage landscape into the Saxon period, whatever continuity or discontinuity there was in who was managing it.

Each name discussed below was checked against early recorded spellings from Domesday and later, the physical landscape including elevation and watercourse position, and, where available, independent documentary or archaeological corroboration, rather than accepted on the strength of a single gloss. Readings are described as confirmed where independently corroborated, corrected where an initially proposed reading was revised against the evidence, or plausible but unconfirmed where a reading is consistent with known infrastructure but lacks direct period evidence.

A single connecting watercourse ties much of this cluster together. Glynde Reach, the tidal-influenced tributary of the Ouse already noted at Beddingham, rises from streams recorded near Laughton, near Ripe, near Selmeston, and on the outskirts of Ringmer, before flowing past Glynde itself down to the Ouse. This means Ringmer, Laughton, Ripe, Selmeston, Beddingham and Glynde all sit on one connected drainage and tidal-influence system, rather than standing as separate, coincidentally similar sites. This is independently attested through Glynde Reach's recorded catchment and is the strongest single piece of landscape evidence tying the whole hām, mere and landing-stage cluster together.

Several readings in this cluster are now confirmed by independent source. Beddingham's church sits at Glynde Reach and is described in an independent parish record as marking the former centre of a settlement that was "a modest port," consistent with a hām originally sited at a landing stage. Hamsey is a genuine island in a loop of the Ouse, recorded as a Saxon-period port where King Æthelstan held council, with its population later moving to Offham as the ground there proved better; the second name-element is possibly the Norman de Say family rather than Old English ieg, "island," and is worth checking further. Glynde derives from glind, "an enclosure or fishing frame" and sits on the tidal inlet in Glynde Reach. Ringmer derives from hring mere, "the fortified settlement near the sea" . Alfriston derives from Ael Fyrse ton the eel village near the furse on the hill. Barcombe derives from bere and combe, "barley valley" a directly agricultural name at the tidal-head hub itself. Berwick derives from bere-wīc, "barley vicus," implying a Roman fortification nearby possibly Burlough Castle. Rodmell derives from the rod derived from hroða – hides -and mell derived from mela – meal or flour, so the cattleand meal settlement

Other readings remain plausible but unconfirmed. A local record independent of this analysis already describes the centuriated fields south of Ripe as grain land supplying Pevensey via a named route, Langtye Lane, and Ripe itself carried eight Domesday salthouses alongside Laughton's sixteen. Offham and Southerham both combine documented chalk quarrying with river and wharf access on the Ouse, and since lime-burning is a Roman-introduced technology in Britain, a Roman-period kiln at either site is plausible. No Roman-period kiln evidence has been found at either, though continuous later quarrying, documented from the 18th century onward, would plausibly have destroyed earlier phases, on the same pattern already established for Wealden iron-working sites.

One further pattern is worth flagging as an open, untested hypothesis rather than an established finding. A cluster of hybrid -ing-tūn names — Wilmington, Lullington, Litlington and Jevington — sits around the rim of the Cuckmere valley rather than at its centre, close to Burlough and the Alfriston/High and Over cemetery. One possible reading is that such names mark a phase of Saxon territorial expansion, sited at the edges of an area as it was being organised rather than at its core. This is consistent with Dodgson's argument that -ingas and -ing-type names represent a secondary phase of estate organisation rather than the earliest settlement layer, but it has not been tested here against independent dating evidence, principally whether early Anglo-Saxon cemetery evidence clusters at proposed boundary zones more than in the interior, and should be treated as a flagged hypothesis for future testing rather than an established pattern.

One further feature is noted but unresolved. Mays Farm, centred around 50.851, 0.16 to 0.17 (point observations vary across the platform's footprint), sits on raised ground at roughly 100ft or 30m — well above the c. 4.5m contour of the surrounding Ripe/Chalvington field system. First-edition Ordnance Survey 6-inch County Series mapping from the 1860s–80s survey marks a "(site of)" feature here with no further label. Personal site inspection confirms a raised platform approximately 6ft above the surrounding land, covering roughly 14,630 square metres, around 3.6 acres — a substantial, deliberately-scaled earthwork rather than an incidental rise. The platform is now built over by modern farm buildings, meaning no ground-level archaeological assessment is currently possible without geophysical or LiDAR survey. Its position, at the edge of the field system with a sightline over the Roman road corridor, would suit a watch or defensive function, but this is presently unconfirmed: the site does not appear in any source checked for this page, its date and function are unknown, and no early spelling of "Mays Farm" has been traced, so the proposed reading of the name from Old English mece, "sword," is noted here only as a working hypothesis, not an established etymology. It is worth flagging to local contacts, such as HAARG, or checking against the East Sussex Historic Environment Record before further use.

Taken together, this gives the corridor a genuine, if partial, Saxon-period layer: the same Ouse and Glynde Reach system organising grain at Ripe, Berwick and Barcombe, salt at Laughton and Ripe, and landing stages at Beddingham and Hamsey that the Roman-period evidence already describes, with the boundary and frontier question at Wilmington, Lullington, Litlington and Jevington left open for further work.

This page presents a hypothesis built from independently sourced evidence at each site. Established facts — excavation results, Domesday entries, measured erosion rates, Margary's survey — are distinguished from inference, such as continuity arguments, functional interpretation, or the dating of the corridor as a unified system, throughout. Evidence quality varies significantly between sites: Bridge Farm's excavated roads, Ripe's centuriation, and Castle Hill's recorded fort are the strongest anchors, while Combe Hill's role in the coastal line and the dating of Burlough's earthwork are the weakest and should be presented with appropriate caution. Further work, particularly confirming Carausian-period occupation phases at the individual villa sites, and any Roman-period evidence at Laughton, would strengthen or qualify specific elements of the argument.