The Roman quay at St Magnus House, London

Louise Miller, John Schofield and Michael Rhodes     Download (75 MB)


The work conducted by the Museum of London at New Fresh Wharf – now the site of St Magnus House, Lower Thames Street – took place in three stages: two area excavations in 1974 and 1975, and a watching brief in 1978. Although the site stretches from Thames Street to the present river wall, only the northern third contained surviving archaeological strata. This report deals with finds of the Roman period, from the 1st to the 4th centuries AD; finds of the 10th to 12th centuries are described in Aspects of Saxo-Norman London 3: the Bridgehead and Billingsgate to 1200.


The excavations uncovered structures of three successive periods:

a 2nd-century river embankment wall, showing the limit of land reclamation at the end of that century;
an ambitious two-part quay installation, dated principally by dendrochronology to the decades AD 225-245;
the riverside city wall, running along the back of the 3rd-century quay. This was probably built in the years AD 255-70, and may have had grave consequences for the useful life of the quay.
In addition, many portable finds, especially those recovered from in and around the 3rd-century quay, have considerably advanced our knowledge of public, private and commercial life in Roman London.

The Roman quays and embankments

In the northernmost part of the site, actually under the southern pavement of Thames Street, traces of a timber embankment wall, made of piles and planks, were recorded. It was erected some time during the 2nd century, and marks a phase of land reclamation beyond the line of the 1st-century quays which are to be found on the north side of Thames Street.

Four metres to the south lay a second revetment. Though both revetments were of similar construction, river silts lay against the first but not the second. This suggests that the first formed a river embankment but that the second was merely the preparation for a new quay and further phase of reclamation. The quay itself was built out beyond the revetment, a further 5m into the contemporary river. The wall consisted of five tiers of large oak beams held in position by a framework of braces and piles. Sufficient of the structure remained in position to suggest at least fourteen stages of construction, starting with pile-driving and ending with the laying of the topmost tier of beams. The wooden framework was further stabilised by tons of earth and rubble, tipped around the timbers. The upper layers would have been rammed down to create a working surface. This dumped material – which included numerous imported goods as well as general household rubbish – was evidently culled from sources in the immediate area, some of which may have been warehouses.

Dating and parallels for the quays

Precise dating of the 3rd-century quay involves the evaluation of several different indicators, and the consideration of parallels both in this country and abroad. Tree-ring dating suggests that the latest of the quay timbers themselves were cut down after AD 209 but probably before AD 244. It is likely – to judge from what is known of Roman building practice – that such large timbers as these would have been used soon after felling. The felling date should therefore be close to the date of construction. On the other hand, a considered date for the latest pottery from the quay infill – samian ware from Rheinzarben in Germany – is AD 235-45. This discrepancy is not easily resolved. Delays in building, or the possibility that infilling was completed some time after the quay had been constructed in timber, can be considered but mostly rejected. On balance, the best that can be proposed is that the quay was built at some point in the bracket AD 225-45.

The form of construction, relying on piles and tieback braces, is found upstream of London Bridge at the same time (late 2nd/early 3rd century) and on the eastern part of the Custom House site slightly earlier (late 2nd century). The closest parallel, however, is in the 2nd-century quay at Xanten on the Rhine; the arrangement and size of timbers, details of jointing, and the use of consolidation material at least around the lower beams are very similar. The form generally appears elsewhere in late 1st and 2nd-century contexts, and so the St Magnus House structure is the latest known instance of it.

A second and contrasting type of construction, based on beams arranged in boxes, characterises the 1st-century quays to the north of Upper and Lower Thames Street, and certain 2nd-century works at the west end of the Custom House site and in Dover. The two types existed side by side and do not represent stages in an evolving technology. In fact, differences in the stability of the underlying strata may have governed the choice of design. Stable London Clay largely underlay the box-constructed quay at Custom House, whereas potentially unstable river silts and gravels were to be found at St Magnus House; here too the quay was built further out into the Roman river, at a point near the bridgehead where reclamation was more intensive.

Samian and other pottery

The monograph includes an extensive series of reports on the portable finds, chiefly the pottery. Approximately 88% of the pots dumped in and around the quays were imported - an astonishingly high proportion which, taken with the fact that many vessels are complete and apparently unused, suggests that the dumps may represent clearance from nearby shops or warehouses. Many of the vessels are in samian or black colour-coated ware and can, in turn, be divided into two groups: a Central Gaulish group, datable to c. AD 170-80, and an East Gaulish group of c. AD 235-45. Less familiar - and in much smaller, though still significant, quantities - are coarsewares and finewares from other production sites along the Rhine and Moselle: for example, mortaria from Speicher, near Trier, or Soller, near Bonn, and hunt beakers from Cologne. So too are there wares from the Pas-de-Calais/Picardy region of north-west Gaul and amphorae which, as a group, show a bias towards sources in north Africa and the east Mediterranean. Such Romano-British pottery as there is comes almost entirely from four sources: south Devon, Dorset, the Nene Valley and Colchester.

Other portable finds

Besides pottery, many other finds demonstrate that London’s prime purpose – at least up to the early 3rd century – was as a port. Some objects would have come by sea from other parts of Britain: roofing slates from north Wales and the Bristol region, jet pins from Whitby in Yorkshire, and coal from either south Wales or the Durham coalfield. A pilaster capital may have been imported from northern France. The business of the port was partly recorded, as letters, bills of account or transactions, on wooden writing tablets made from woods such as cedar, silver fir and Norway spruce.

Other finds reflect domestic life and the appearance of the Roman city more generally. Building stone was mostly imported from Kent, but fragments of Carrara marble and veneers of onyx marble, together with fragments of several different wall-paintings, show that some buildings were richly decorated. Large numbers of leather shoes were recovered, of which some are gilded and others have exaggerated broad toes - a style evidently fashionable in the early 3rd century. Clothing is represented by parts of leather jackets and, possibly, breeches, while inscribed offcuts are remnants of the leatherworking industry itself.

Dismantling of the quays, construction of the Riverside Wall

In the 240s London’s prospects must have looked encouraging, but within fifty years this had changed dramatically. Political and economic changes in the Roman empire brought about the collapse of the main north European markets which supplied London and Britain. From about AD 250 Saxon pirates began to disrupt the trade routes. The level of the river may have been dropping in relation to the land, causing silting and difficulties of access in the ports of south-east England.

The lifespan of the quay at St Magnus House was short. There is no evidence for use after about AD 260, barely twenty years after its construction. At about this time, moreover, a large wall was built to run along the full length of the quayside, about 8m behind the river frontage. The Riverside Wall here resembles the lengths found at Blackfriars and the Tower of London, and is dated by dendrochronology – as on the other sites – to AD 255-70. Although this was the time when some of the Saxon Shore forts were being built, the wall here is closer in its style of construction to the earlier traditions of architecture embodied in London’s early 3rd-century landward wall. The quay was no longer required on the same scale as before, and parts were dismantled.

The first post-Roman structures on the site were rubble banks and stakes of the 10th century, laid against the decayed and eroded base of the Riverside Wall.

[Special Paper 8 (1986); abstract by Francis Grew, 26-Oct-1997]