Colchester is the largest town in Essex by population, although it is neither the county town nor does it have a cathedral, both distinctions belonging to Chelmsford. It is in the NE of the county, 50 miles NE of London to which it is linked by the A12 and the Great Eastern main line into Liverpool Street station.

The castle was built using the podium of the Roman Temple of Claudius as foundation for the keep, and is built of rubble including septaria and stone and tile quarried from Roman buildings around the town. The keep is the largest ever built in England with a footprint of 152 by 110 feet (46m × 34m). In form it is a hall keep (rather than a tower keep, like Castle Hedingham). Its long axis runs from N to S, with the main entrance at the W end of the S wall. At the S end of the E wall is an apsidal projection that housed the chapel. The castle is now 2 storeys high, but might have been higher when first built. The only Romanesque sculpture surviving is on the S doorway.

Colchester is one of five towns in Britain with claims to be the oldest in the country. Colchester’s claim is that it is the oldest recorded town, since it was mentioned by Pliny under the name Camulodunum in 77 AD. It certainly existed as a town before that, as coins minted in 20-10 BC by King Tasciovanus of the Catuvellauni mention the Celtic name of the town Camulodunon. The castle is traditionally dated to 1076 and was built on the orders of the king, probably by Eudo Dapifer, his steward. It remained in royal hands until 1101 when Henry I granted it, with the town, to Eudo Dapifer. On his death in 1120 it reverted to the crown, where it remained until Henry IV granted it to his youngest son Humphrey of Lancaster in tail male (i.e. to Humphrey and his male heirs) in 1404. It remained in royal hands until 1629, and thereafter the process of dilapidation began.

Colchester castle and the White Tower of the Tower of London are unique among English Norman keeps, although Fernie (2000) finds continental parallels with the great towers of Ivry-la-Bataille and Loarre in Arargon. The issue of primacy between the two English buildings is difficult to decide. Drury (1982) argues that the Norman castle was built in two phases; in 1074-76 by the crown, and from 1101 by Hugo Dapifer, and he dates the wall in which the doorway stands at this height to the second phase. Clarke (1974) has argued that the doorway was a later insertion in an 11thc wall, while Fernie (2000) has no difficulty in accepting the doorway as work of the 1080s, arguing that the portcullis slot, which courses with the masonry of the doorway, must have been an original feature. In terms of the sculpture a date in the late 11thc is preferred here, given that this is a royal building and would not be expected to be old-fashioned in its details. The Corinthianesque capitals are comparable with those of Scolland’s Hall in Richmond Castle (1071-89) foe example, and the cushion capitals are ubiquitous from the 1070s onwards.