to late 17th century population explosion.
Apart from a scatter of Roman material, the earliest evidence for occupation on the site of Norwich comes from one, or possibly two, Pagan Saxon cemeteries. These lie on the edges of the walled area, adjacent to two of the possibly three Roman roads (one leading south to the town and major pagan cemetery at Caistor) that cross the site. It was around the Roman crossroads on a gravel spur running down to the river, that the Middle Saxon settlement that was to become Northwic developed. A considerable amount of 7th—9th century Ipswich ware has now been recovered from this area, around the Cathedral Close, both from excavation (in 1956 and
1971-72) and from building sites. Unfortunately this does not apply to the other three Middle Saxon sites which are suspected from place name evidence to lie within the later walled area. The most important of these is undoubtedly Conesford, the original settlement to the north of which the 'wick' (suburb or trading settlement) that was to give its name to the town developed. It is tempting to locate this settlement in the vicinity of St. Etheidreda's church, which in addition to its apparently early dedication has a totally anomalous tithing system, but there is no archaeological evidence to support this. North of the river the topographical element of Coslany (Cost's Long Island) helps narrow down the possible area of its site,
but this is covered by 19th century factories which will not be demolished for another ten years or more.
Westwick, the fourth of the suspected Middle Saxon sites, has proved almost as elusive as Conesford and Coslany. Excavations in 1971 and 1972 to the east of St. Benedict's Gates added only a few sherds of Ipswich ware to the handful found by John Hurst in the 1950s, and it looked as if the focus of Westwick might lie further to the east (around St. Gregory's church). This would have supported an interpretation of the 'wick' element in the name as meaning 'suburb' rather than the 'trading settlement' that John Hurst favoured. In 1973 however, sufficient work was done