Although undoubtedly made in several counties in the study region chaff-tempered and other early to mid-Saxon wares are best described in one section.
Soft to hard.
Black, sometimes with patches of oxidation on the surfaces.
All chaff-tempered pottery contains abundant chaff tempering (not examined botanically) but examples differ in the quantity of chaff, its probable identity (a neutral term, vegetable-tempered ware has been proposed), the presence of other inclusion types and in the character of the clay matrix (Russel, 1980). There are a few vessels known in which the chaff- tempering may not be abundant. In some instances, as at Hatton Rock, Warwickshire, there is a continuum of textures between fabrics with almost pure chaff-tempering and those with sand-tempering (see Other early to mid- Saxon wares, below). The characteristics of individual groups of chaff-tempered pottery are described below.
Gloucester, Gloucestershire. Fragments of chaff-tempered pottery from St. Oswald's Priory occur in two fabrics: i)Abundant rounded and subangular quartz up to 0.5mm and a fine clay matrix containing a little fine white mica. ii)Sparse quartz (as (i)) and abundant quartz and mica silt.
Frocester Court Roman Villa, Gloucestershire. At Frocester the sherds contained few inclusions other than chaff, except for sparse rounded quartz and limestone fragments. There was no evidence for any real division in fabric amongst this collection. The Frocester vessels contain iron ore, angular quartz and white mica up to 0. 1mm.
Wycomb, Andoversford, Gloucestershire. Sherds from Wycomb, near Andoversford, include examples with a quartz-free micaceous fabric (Rawes, 1980).
Trowbridge Castle, Wiltshire. Chaff-tempered pottery from Trowbridge Castle contains abundant very fine quartz sand (Smith, R. forthcoming a).
Others. 'Large' collections of chaff-tempered pottery, in excess of two or three sherds, are also known from Ogbourne, Ramsbury, Swindon, and Westbury in Wiltshire. Smaller collections are known from sites in Somerset and Gloucestershire. The fabrics of the Ramsbury and Swindon vessels have been compared but the remaining types have not been petrologically examined, except for one thin-section of Cheddar fabric A (Timby, forthcoming; Russel, 1980; Peacock, 1979).
Thin-sections: Gloucester; None (too little material available), Frocester; M700, Trowbridge; M1218.
There is no evidence for any trade in these vessels but conversely, because of the absence of diagnostic inclusions there is little evidence for local production. The petrology of the Gloucester and Frocester sherds shows that they are from different sources and the appearance of the Wycombe sherds suggests that they too are from a separate source. However, these sites are over 10 miles apart. Russel concludes that one of the Ramsbury fabrics is indistinguishable from the Swindon fabric, but given a probable difference in date and the undiagnostic nature of the fabrics this cannot be used as positive proof that the two sites obtained pottery from a single source. Peacock concludes that the Cheddar ware was probably not locally produced.
Cooking pots. Handmade bag-shaped vessels with short, slightly everted rims and rounded or flattened bases. The Wycombe sherds include a small lug, and a pedestal base. Bowls. The Wycomb and Bourton-on-the-Water collections include poorly finished bowls with slightly everted rims.
It is difficult to put a precise date on any Chaff-tempered pottery from the study region, although sherds from Barnsley Park and Wycomb could belong to the late 4th to 5th centuries since they occur in association with Romano-British pottery. Several hypotheses can, however, be put foreward for testing:
Although it is important for the archaeology of the region to distinguish between hypothese (a), (b) and (c), the latter two hypothese are most important for this study (see Ch.11). At Old Town, Swindon, the sequence of structures and finds suggests a date-range of 5th-6th to 8th century for chaff-tempered ware, but both ends of this range are the result of absence of dated contexts rather than definite evidence that the ware was not in use. Evidence from Cricklade suggests that chaff-tempered ware was in use before the construction of the defenses in the late 9th century but this pottery may have been associated with a much earlier occupation, in which case the constructors of the defenses left no pottery at all. Chaff-tempered pottery in the construction levels of a 10th century Church at Trowbridge was probably residual from an earlier occupation on the site, which is undated. At Ramsbury, chaff-tempered pottery was associated with iron-working of late 8th and 9th century date, and in this instance cannot be residual since the sherds occur in a occupation deposit overlying an accumulation of hill-wash which itself is of mid-Saxon date. These then are the latest stratigraphic occurences of the ware in the region. Outside the region, in Warwickshire, a chaff and sand-tempered ware was in use at Hatton Rock in the Mid-Saxon period, associated with a radio-carbon date of ad 906+_88. This determination is unlikely to result from a real date earlier than the 9th century. Evidence for an early date for the ware is easier to come by, and it is easy to show that chaff-tempered pottery was at one time the only ware used in Wiltshire. This does not, in itself, prove that the ware was not also used in the mid-late Saxon period and indeed the evidence increasingly points to chaff-tempering having been used as a potting technique into the 10th century. If, however, we compare the archaeological and documentary evidence for occupation at Hereford and Gloucester in the 9th century it seems clear that no chaff-tempered pottery was in use in either settlement.
Dunning (1932), Ireland (forthcoming a), Fowler (1966, 1970), Rawes (1980), Russel (1980), Timby (forthcoming).
Vessels of early to mid-Saxon date in fabrics which are not chaff-tempered are extremely rare. Sherds of stamped pagan Saxon jars, mainly cremation urns, from Berkshire have been examined visually and samples of cremation urns from Beckford and Burn Ground, Hampnett, and domestic vessels from Signet Fields, Kidderminster, Maiden's Bower, Solihull, and Fladbury have been thin-sectioned. In no case is is likely that a pair of samples come from the same source but two samples from Burn Ground, Hampnett, contain a sandstone and quartz sand which is petrologically identical to those found in the Maidens Bower, Hatton Rock and Signet Fields samples. The same sandstone and quartz sand was identified visually amongst a small collection of Saxon pottery from the Cirencester bypass in the possession of R. Reece. The Beckford and Fladbury samples have moderate inclusions of ill-sorted subangular and rounded quartz.
Hard.
Black.
Hatton Rock, Warwickshire.
Signet Fields, Kidderminster, Hereford & Worcester.
Burn Ground, Hampnett, Gloucestershire.
Fladbury, Hereford & Worcester.
Beckford, Hereford & Worcester.
The petrology of the Cirencester and Burn Ground sandstone and quartz tempered sherds indicates a non-local source, probably in Warwickshire or Worcestershire, although the inclusions are not sufficiently distinctive to show that the vessels were made at the same site.
Cremation urns. Jars.
The Beckford and Burn Ground urns must be of 6th or 7th century date since they are in a pagan context. Some of the Cirencester bypass sherds are stamped, which should date the collection also to the 6th or 7th century. The Fladbury vessel is probably of 8th or 9th century date and the Signet Fields and Maiden's Bower vessels are undated.
Grimes (1960, 113-125), Hirst & Rahtz (1973), Meaney (1964), Peacock (1967 b).