The 'castle' at Burgh is thought to have originated as one of the Roman forts of the Saxon Shore.  It may have been mentioned as Garianonum in the late Roman period in Britain, the name possibly derived from the River Gariennos of Ptolemy (d.c.170AD).  This is now known as the River Yare and joins the Waveney just north of the fort.  The history of the Roman site must be mainly worked out from the 183 unstratified coins found at the site.  These range in age from Domitian (81-96) to Honorius (395-423).  Although 30 of these were illegible, the next 30 examples take the coin record from Domitian up to the start of the reign of Constantine (306-337).  The other 123 coins date from 306 to 423 AD with there being a definite spike in numbers in the early 320s.  This might suggest that the fort was built some time after the crushing of the rebel Allectus in 296 by Constantius and certainly by the end of the reign of the Emperor Constantine (306-37AD).  The evidence suggests that the fort was abandoned in the early fifth century, but like Pevensey, it might have lived on in a cashless society that has left little trace.  To support this supposition, the pottery finds also seem to run from the late third century onwards.  The fortress itself and its garrison of late Roman cavalry (equites stablesiani) of the Notitia Dignitatum (dated to 400-425 AD) suggests it was still functioning in the early fifth century.  Certainly the coin record continues within the fort with a silver sceatta of c.700AD and another of King Ceowulf of the Mercians, Kent and East Anglia (821, deposed 823).

Burgh castle, like Caister on Sea, has been associated with the Cnobersburgh mentioned by Bede (d.735) as a possession of King Sigebert of East Anglia (c.631-635AD).  This has him build the monastery for the Irishman Fursa on a site:

pleasantly situated close to the woods and the sea in a castle (castro) which in English is called Cnobheresburg, ie the city of Cnobheri.  Later King Anna of that realm with many noblemen adorned it with more stately buildings and more splendid donations.

King Anna died in 653/4.  Wattle and daub huts excavated in the north part of the fort overlying the Romano-British huts found with substantial quantities of Ipswich ware are thought to have been remnants of this.  Early churches were also founded in the old Roman forts of Bradwell, Reculver, Richborough and just possibly Walton which may have been Dommoc, although this is usually thought to be Dunwich. 

Burgh fort belonged to Bishop Stigand of Norwich under Edward the Confessor (1042-66) and was a parcel of the 4 carucates that made up the manor which also contained a church, the whole manor being worth 100s.  By 1086 the manor of Burch was held by Ralph the crossbowman (Balistarii) and was worth 106s.  Ralph also held 4 other vills lying south and south-east of Burgh forming an arc towards Lowerstoft.  Presumably Ralph was the ancestor of the Burgh family of whom another Ralph held Burgh for the sergeanty of finding a crossbowman for 40 days by 1212.  By 1246 his grandson, another Ralph, had granted the manor to Gilbert Wesenham who in turn passed the land on to Bromholm priory in Norfolk.  From this it would appear that the castle was no longer a going concern by the thirteenth century.

Burgh Roman fort stands hard against the muddy banks of the River Waveney in Suffolk which runs past some 30' below the interior of the fort.  In fact, perhaps as much as half of it has been washed away to the west, leaving a site 640' long by 410' wide.  Of the remainder the east side still carries the remnants of 4 boldly projecting, 15' diameter, round towers, 2 equidistantly placed from the destroyed east gate and 2 at the chamfered off corners.  Presuming the fort had similar north and south walls only 1 of the equidistant towers remains to the north, while that to the south, like the main gates, have fallen, although the remains of the single tower still lies on its side, giving good access to the hollow in its central core.  The towers at first appear to have been added to the defences as the lower 7' of their construction is not bonded to the wall.  However, although only the upper halves are bonded, the tile lines appear at similar levels throughout and the towers have not settled differently to the walls, as would have been expected if they were of different phases.  Therefore it seems that the first 7' were not bonded as the foundations were known to have been insufficient and the builders did not want to strain the walls until the first building phase of 7' had completed.  Consequently they left the lower portions of the towers unbonded.  The 2' diameter holes, 2' deep in the centre of the towers, are thought to have been an aid in mounting war engines, maybe for swivelling powerful crossbows, although the narrow diameter of the towers, only 15', would make them cramped to operate.  The only surviving gate, the east one, is now just a featureless gap.  However, excavation in the 1850's found some 2' thick walls running back from the 11½' wide gate.  These are now interpreted as the sides of 2 internal guard rooms which may have made an internal gatetower.

The fort walls still stand over 15' high and some 10' thick.  They are a mixture of coursed flint rubble set between regular 3 courses of tile in typical Roman fashion.  This style is repeated in the Roman defences at Colchester, Dover lighthouse (one or 2 tile courses) Lincoln, London, Lympne (2 tile courses), Pevensey (twin tile courses), Portchester (twin tile courses), Richborough (twin tile courses), Silchester (twin flat stone courses), Venta Icenorum and probably Walton.  Such a tile reinforced style seems confined to areas of poor quality building stone, like flints and so is not present in sandstone or limestone built forts like those along Hadrian's Wall, or surprisingly Chichester which is flint built, but without tiles.

The remnant of the south wall at Burgh is much collapsed, but at the south-west corner of the present site are the shattered remnants of a castle motte, demolished around 1770 and levelled in 1839.  This was protected by a V shaped ditch which has been infilled.  The ditch enclosed a motte with a diameter of some 235' north to south by 170' east to west.  Excavation found it had been 13' deep and wet, presumably fed by the nearby river.  Clay pads disturbed Saxon graves and slots in a separate wall fragment were thought to be where a substructure once stood to help stabilise the motte so that it could carry a timber Norman tower.  Radio-carbon dating of the bones in the Saxon graves returned results dating from 630 to 1170 which might suggest the fort was used as a Saxon graveyard from the seventh century until the Norman Conquest.  Certainly the late dating bone sample which returned 810-1170 was from a grave that had disturbed an early burial.

To separate the motte from the rest of the fort a 60' wide section of fort wall was demolished where the motte ditch passed through the old defences.  As the west side of the fort had been washed away an earth bank was constructed on this side to make the rest of the fort walls the castle bailey, while the broken western end of the north wall was strengthened by placing earth against it up to 20' high.  That this was necessary shows that the western portion of the Roman fort was washed away by the eleventh century, as indeed most of its replacement castle bank has also been.  Traces of the earth bank were confirmed by the excavation which also found pottery dated to the eleventh and twelfth century.  Presumably, unlike Walton, the Norman castle was soon abandoned.


Copyright©2020 Paul Martin Remfry