When the Romans arrived in the area of the future Gloucester in the 40s AD the Severn followed a course considerably different from today's east and west channels. There was at the time only a single main channel, which flowed farther east than today's east channel. Although sources vary, a strong theory is that a cross channel, the forerunner of today's west channel, flowed from Over directly across today's Alney Island and joined the main channel on the north side of today's gyratory around The Island, the area now occupied by the Westgate Retail Park just before Westgate Bridges.
The first Roman fort was established at Kingsholm, north of the present day city centre, close to a ford across the Severn. This was abandoned in the 70s AD in favour of a new fort, located on the site of present day Gloucester city centre. The rectangular fort, which became the Roman colony of Glevum at the end of the 1st century, was enclosed by wooden, later stone walls, and the west gate was located close to the entrance of today's Berkeley Street (modern Westgate Street runs slightly north of its Roman ancestor).
Roman River and Quay
The east bank of the Severn in the Roman period lay some 240 metres (260 yards) beyond the west gate and followed the line of today's Lower Quay Street, some 100 metres (110 yards) east of the modern river. Archaeological discoveries of wall structures in 1846 and 1928 interpreted as possible parts of Roman quays on the east and west banks, and identification of the Roman riverbed on the west bank in 1976, indicate the Severn to have been some 30 metres (100ft) wide (though there is some doubt as to whether the stonework found on the west bank is actually part of a quay or even Roman – one source suggests it is more likely part of the western approach to the Norman castle). Later discoveries, such as those found in the 1970s construction of the present Dukeries housing complex and a 2019 borehole, tend to confirm the ancient line of the east bank, with Roman quay walls or evidence of them unearthed in various locations on a line from the north-west corner of the Gloucester prison site northwards along Lower Quay Street to Clare Street.
Roman Harbour/Dry Dock Theory
Excavations during building work on the now long since demolished gasworks between Gloucester prison and Quay Street in the period 1820–1834 unearthed evidence of a man-made basin. Later works revealed a wall lining the southern side of a presumed entrance to the basin. Archaeological excavations conducted during the construction in 1938 of the Co-op Creamery on Westgate Street (now no. 85, sandwiched between the Old Crown and the former White Lamp pubs, and the modern development behind, between Westgate Street and Quay Street) revealed a curved foreshore lined with wooden stakes. A 1942 report interpreted these finds as a Roman harbour dug out of an inlet on the east side of the Severn.
The harbour theory was supported by the discoveries in 1805 of the remains of an ancient boat nearby (though its exact position, recorded only as "at the bottom of Bearland", i.e. modern Quay Street, has been lost and the boat may have been either Roman or early medieval) and a buried barge found in 1910 on the site of the Petty Sessions Court, in the area of today's Shire Hall complex. An alternative explanation for these discoveries, however, is that the vessels were abandoned in the moat of the Norman castle which bordered Quay Street to the south, on the site of which Gloucester prison was later built.
The theory that a Roman harbour existed next to the river in the area of today's lower Westgate and Quay Streets subsequently fell out of favour. On the basis that such a facility would have silted up as fast as it could have been dug, it was theorised that the area was actually a Roman dry dock. A 1978 article detailing the evolution of the Severn around Gloucester suggests that the basin cut out of the clay next to the river was dammed at the river bank and drained. Boats would be constructed or repaired in the dry dock thus formed, and when ready the dam breached to fill the dock and allow the vessel to be floated out onto the river. The dam would then be remade and the basin drained ready for the next boat.
Reclamation Theory
With the demolition of the Co-op Creamery, another archaeological excavation on the site in 1989/1990 resurfaced again a feature that could have been a 1st-century foreshore with timber revetments. The discovery led to a new theory: that the Roman river initially ran even farther east, on the line of today's Three Cocks Lane and Upper Quay Street, but was pushed westwards by reclamation during the 2nd century, after which the quay in the area of today's Lower Quay Street found by archaeologists was built.
Fixed Shoreline Theory
The problem with the reclamation theory is that there is no evidence anywhere other than the Co-op Creamery site of alluvial deposits that would confirm a river once flowed over the ground so far east. A more recent theory is that the foreshore found there belonged to a creek (perhaps, as a personal observation, similar to the pills that can be found farther downriver, such as Berkeley Pill and Bullo Pill).
According to the fixed shoreline theory, the timbers – dendro-chronologically dated to c.74 AD and c.95 AD – were part of a wharf or walkway that lined the creek. This landing stage was used for a short period in the pre- and early-colonia period, and over the 2nd century only the creek was subject to significant land reclamation. There is, however, plenty of evidence that a thin strip of land immediately behind the Roman quay, which remained in the same position along the line of today's Lower Quay Street throughout the Roman period, was raised by reclamation of riverside wetland to shore up the waterfront.
Roman Riverside Settlement and Bridge
There is a suggestion that in the late-Roman period the quay wall served a defensive function for a 'lower colonia' outside the main Roman colony. Archaeological evidence suggests the existence of a prosperous Roman riverside suburb. A mosaic-floored building has been found on Quay Street, masonry foundations on the Shire Hall site, a high-status 1st/2nd-century house which was replaced in the mid/late Roman period by a temple or baths complex beneath St. Mary de Lode, more buildings and a tilery at St. Oswald's Priory, and evidence of substantial buildings found on modern day Clare Street, the Westgate Flats and on Westgate Street opposite the creamery site.
Although sources differ as to whether the Romans crossed the Severn by bridge or ferry, recent studies suggest that what was initially interpreted in 1973 as part of the quay wall on the site of the Archdeacon flats is actually part of a Roman bridge.
Post-Roman River
Little is known of the waterfront in the immediate post-Roman period, but by the time of the Saxon re-occupation in the 10th century the east bank had shifted westwards by some 15 metres (16 yards). Saxon Gloucester extended to the west of the former Roman colony as far as the river; its western defences rested on the Roman quay walls, which were not demolished until the 12th century. The long-lost, Saxon-named Powke Lane, first recorded in the early-13th century, may have marked the Saxon riverfront. It ran north from Westgate Street through the middle of today's Dukeries housing complex roughly along the line of Archdeacon Court to St. Oswald's Priory. Property boundary lines south of Westgate Street that crept ever closer to today's river bank betray a Severn that continued a westerly migration in the late-Saxon and early-Norman periods.
The first definitively known bridge across the Severn, later known as Foreign Bridge, was constructed in 1119. Its six stone arches were progressively blocked and built against as the river channel it crossed dried up over the centuries. The remains of Foreign Bridge still lie under the roads to the south and south-west of Westgate Street car park.
The channel spanned by the bridge had long been shrinking due to increased silting. Its flow had been significantly reduced after the Severn spawned a second channel, today's east channel, farther upstream (believed to have occurred in the 6th or 7th centuries, one or two centuries after the Romans had abandoned Britain). The new channel initially joined the cross channel above today's Westgate Bridges, which still joined the original channel in the north-east corner of The Island.
Shortly after the construction of Foreign Bridge, and possibly as a result of it, the new channel broke its banks in the area of today's Westgate Bridges. It began flowing south, not north of The Island, thus completing the course of the east channel we know today. The new course made necessary for the first time a bridge across the Severn at the far side of The Island. In it's modern form, Westgate Bridge comprises a footbridge flanked on each side by road bridges.
The original channel continued to silt up, becoming known progressively as Little Severn and Old Severn, while the cross channel disappeared completely after the west channel we know today was formed in the late 15th century. By the modern era the original channel had become a pale shadow of the wide river that had flowed in Roman times. It ended its days as Dockham Ditch, which in the 19th century was culverted below the site of the by then buried Foreign Bridge and filled in and paved over above it.
According to one source, present-day ground water level is some six metres (twenty feet) above the minimum river level of the Roman period. The river level at Gloucester was raised by 1.4 metres (four and a half feet) alone with the construction in 1871 of Llanthony and Maisemore weirs. This meant that modern Gloucester no longer had to deal with the high tides and Severn bores that made life precarious for Roman vessels, though by then the dangerous vagaries of the river had finally been bypassed as a major trade route by the opening of docks and canal earlier in the 19th century.
Sources
Gloucester's Roman Waterfront: A Reconsideration, Transactions of the Bristol and Gloucestershire Archaeological Society, Vol. 138, 2021, Andrew Armstrong
Glevum and the Second Legion, Journal of Roman Studies, Vol. 32, 1942, Charles Green
Waterfront archaeology in Britain and northern Europe: Gloucester, Council for British Archaeology, 1981, pp. 123–124, Carolyn Heighway and A.P. Garrod
The River Severn at Gloucester, Gloucester and District Archaeological Research Group, 1978, F.W. Rowbotham
Archaeology in Gloucester 1989, Transactions of the Bristol and Gloucestershire Archaeological Society, Vol. 108, 1990, pp. 187, 190, M. Atkin & A.P. Garrod
Archaeological Review No 15, Transactions of the Bristol and Gloucestershire Archaeological Society, Vol. 109, 1990, pp. 228–229, Malcolm Atkin
Gloucester Archaeology, 1900-1990: an Historical Review, Transactions of the Bristol and Gloucestershire Archaeological Society, Vol. 110, 1992, p. 27, M. Atkin
Anglo-Saxon Gloucester: c.680 - 1066, Victoria County History, 1988, ed. N.M. Herbert
The Archaeology of Gloucester Castle, Transactions of the Bristol and Gloucestershire Archaeological Society, Vol. 102, 1984, pp. 83 & 95, Henry Hurst
The Golden Minster, Council for British Archaeology, 1999, pp. 6, 9, Carolyn Heighway and Richard Bryant
Gloucester: Street names, Victoria County History, 1988, ed. N.M. Herbert
Base map credit: https://www.openstreetmap.org/
