3 General description

Anglo-Saxon palaces have been known from documentary sources for a long time, but it has only been since the 1950s that archaeological evidence for these sites has come to light through aerial photography and excavation. Key excavations have been those at Yeavering, Northumberland (Hope-Taylor, 1977), Cheddar, Somerset (Rahtz, 1979), North Elmham, Norfolk (Wade-Martins, 1980) and at Northampton, Northants (Williams et al, 1985). As yet there have been no syntheses of the results.

The most prominent components of Anglo-Saxon palace sites are the large and elaborately built timber halls. These were typically between 18m and 30m in length and between 6m and 9m in width, although some were much larger as at Hatton Rock where one building appears to measure 50m x 9m (Rahtz, 1970). In common with timber buildings from other Anglo-Saxon sites, most halls were rectangular, often with simple length:width ratios (often 2:1 or 4:1); the use of the square, often in pairs, was very common (James et al, 1984).

The archaeological evidence for these buildings comprises the foundation trenches and postholes from which may be inferred the methods of construction and the basic superstructure of the building. Construction methods varied both between sites and within sites. The most common method was the 'post in trench' technique which comprised foundation trenches, in general up to 1m deep, although in the case of one hall at Yeavering between 2 and 3m deep, dug to hold the upright timbers. The impressions or "ghosts" of the posts left in the trenches show that the walls were either of solid vertical planks or of posts spaced apart and probably infilled with panels of wattle and daub. Another method of construction was the 'post in pit' method as used in the 10th-century halls at Cheddar; here square posts 0.30 - 0.60m across were set 2.3m apart into pits averaging 0.94m in depth, and the superstructure was probably either wattle and daub panelling or horizontal planks. The halls had between one and four entrances with two usually set symmetrically in the centre of the long walls.

In several cases the palace halls had annexes at one or both ends of the main building, or partitions within the building to create antechambers. Internal postholes suggest that some halls were aisled as at Yeavering and the possible episcopal palace at Bishop's Waltham, Hants (Lewis, 1985), and that some halls had an upper storey as at Cheddar. The timber walls and other internal timbers probably supported a thatched roof; there is no evidence of any other roofing material. In some cases, as at Yeavering, the structures were strengthened by supporting buttresses.

The second phase of the palace at Northampton (if it is indeed a palace site) had a large stone-built hall, the dimensions of which were 37.6m x 11.4m. The length was later extended to 43.4m with the addition of two rooms to the west end. The foundations were of ironstone and limestone slabs set in a sandy loam matrix in foundation trenches around 0.6m deep. The superstructure is likely to have been mortar-bonded and plastered and roofed in thatch. Fragments of tiles amongst rubble at Old Windsor suggest the possibility of a stone building with a tiled roof (Wilson, 1958).

At most sites there are ritual or religious buildings. These have been identified by their association with later churches and chapels or burials, and had several phases of rebuilding and development; at North Elmham there is evidence of a wooden church on the site of the ruined Late Saxon cathedral (Rigold, 1962) and at Cheddar the first chapel, thought to have been built of timber and ashlar construction with limestone boulder footings, was enclosed in the foundations of a later stone church. In some cases, as postulated at Yeavering, these were wooden buildings of similar construction to the great halls, only on a smaller scale; in other cases they were of stone as at Northampton and Cheddar. The church at Yeavering and the later chapel at Cheddar both measured in the region of 16.6m in length (including chancel) and 5.8m in width. At Yeavering there was not only evidence of a wooden church with an associated cemetery, but also a hall which has been interpreted as a temple, from its lack of domestic debris, mound of ox bones and cemetery centering on the site.

Associated with the religious buildings at Yeavering were two cemeteries. These contained East-West oriented inhumations, in predominantly unfurnished graves which were centered on the religious buildings. Those associated with the later church were confined within a fenced yard.

As well as the halls of the palace sites there are also smaller timber-framed buildings. These are either of the "timber in trench" type or of posthole/stakehole construction. Some of these buildings were similar to the great halls although on a smaller scale, others were of much lighter and less elaborate construction. The associated artefactual material from some buildings comprises general occupation debris suggesting that they were dwellings; others have produced evidence that they housed a specialised function, such as iron smelting and forging residue from a building at Cheddar.

Another component of palaces sites, but not confined to them, are the sunken featured timber buildings (SFTB). These have been recognised at a number of palace sites but vary a great deal in size, shape and function. One such building at Northampton measures 3m x 2.6m at Yeavering 12m x 5.7m and a possible SFTB at Hatton Rock measured 24m x 9m.

Some palace sites have sunken featured timber building varients. A posthole building at Cheddar with a central annular area, sunken to 30cm, has been interpreted as a fowl house, or a mill (Holt, 1988). Smaller posthole structures with sunken floors containing large pits at North Elmham have been interpreted as latrines.

A courtyard may be associated with the buildings as at North Elmham where the L-shaped hall formed two sides, each c 15m long.

At Yeavering foundation trenches of nine concentric arcs of a circle have been interpreted as an assembly place. These trenches which, in general, increased in depth outwards from around 1m to around 1.5m, held closely set vertical timbers with larger terminal posts. Four postholes showed that the ninth row was supported by buttresses. In front of this arrangement were eight small postholes suggesting a platform or dais behind which was a 4m deep pit which would have held a vertical post. Both were set within a semi-circular setting of six postholes possibly holding a screen.

A large enclosure was also a separate feature of the palace complex at Yeavering. This underwent several phases of development and in its final form comprised two parallel foundation trenches for palisades, the inner of which was approximately 1.8m deep and the outer approximately 2.1m deep. The main entrance through these palisades was flanked on either side by a circular trench, measuring around 18m internally, one of which contained a rectangular timber building.

Fences and palisades comprising lines of postholes or stakeholes have been found on several sites, sometimes forming enclosed areas and often related to specific buildings. At Yeavering one palisade was trench-built, with deeper post holes dug to hold the main uprights which were split timbers of half -round section. This has usually been interpreted as an animal stockade, but a defensive fort has also been suggested (Alcock, 1988).

Ditches appear to be fairly common features on Anglo- Saxon palace sites, acting either as drainage or boundary ditches. At Cheddar a ditch comprising linked depressions carried storm water from the site to the river; this was later replaced by a steep-sided, flat-based ditch 1.2m deep, which probably acted as a soak-away pit. Another ditch marked the eastern boundary of the site and was broken in two places by entrances. At Milfield crop marks suggest that the palace site was situated within a twin-ditched enclosure probably originally holding palisades.

The remains of water mills have been recovered from the sites of Tamworth (Rahtz and Sheridan, 1971-72) and the probable palace site of Old Windsor. Waterlogged conditions at Tamworth, which preserved the timber, showed that the mill was constructed from horizontal oak planks slotted into sill beams, and that the timbers of the base were joined by mortice and tenon joints. Plank walls survived to up to 0.50m above the floor. Two rectangular boxes were set at different levels; the upper one, packed around with clay, acted as a mill pond. There were also fragments of twenty millstones with a diameter of between 0.60m and 0.70m.

Two Wells have been found at the palace site at North Elmham; One was dug 12m, the other 6.3m into the underlying boulder clay, and both were timber-lined. Due to waterlogging the wells have provided very good preservation conditions and have yielded information on Middle-Saxon carpentery as well as environmental evidence. The preserved timber has been used for radiocarbon dating and dendrochronological analysis and has been dated to the late 8th and early 9th centuries (this dates to the phase before the site was in use as an episcopal residence).

At Northampton, in association with the stone hall, there were the remains of two mortar mixers; these were 2-3m in diameter and were cut into the ironstone and lined with wattle work, the impressions of which were left pressed in the sand. A central post would have supported a beam which rotated a portable paddle mechanism resting on a bearing.

Due to the small number of Anglo-Saxon palace sites, there can be no detailed typology. They can however be provisionally classified as follows:

This simplified classification, however, does not take into account the diversity of individual sites and thus has a great effect on the diversity (form) score.

Simplified ground plans of palace sites are shown in Figure 1.

Some of the palace sites continued in use as palace sites into the post-Conquest period, as at Cheddar; other sites have such as North Elmham and Northampton continued as settlement sites after the high status of the palace was lost. In some cases settlement has continued into the post- medieval period.

From the archaeological and documentary sources, Anglo- Saxon palaces are thought to have been the residences of royalty or bishops. These palaces also served as places of assembly as is suggested by the structure of concentric arcs at Yeavering; documentary evidence shows that the Witan met at Cheddar in 941, 956 and 968. It is likely that these palace sites also served as administrative centres for the collection of tribute or tax as well as other functions.