Building materials made out of fired clay can be divided into three classes depending on their mode of production. The first class is fired on site, often accidentally and includes daub, cob, clay floors and hearths and loom weights. The second class is fired alongside pottery vessels and includes ridge tiles, louvers and finials while the third class consists of bricks and tiles fired in kilns or clamps by specialists such as tilers, brickmakers or paviours.
The methods used to make these objects and their form and typology are described below while the historical development of the building material industry is reconstructed in chapter 9.
Very little fired daub has been found in the region, although it was probably the main walling material of the earlier middle ages. Two large collections of daub have been examined. That from 143-5 Bartholomew Street, Newbury, consisted mainly of debris from a burnt stake-built shed of medieval date while that from Victoria Street, Hereford, consisted of the collapsed superstructure of a corn-drying oven (Vince, forthcoming b; Shoesmith, 1982). In both cases the fragments were quite substantial and the Newbury daub showed sufficient impressions of the wattle framework for the size of the uprights to be calculated, 20-25mm diameter. This size was in agreement with that of the excavated stake holes. The average thickness of the walling was also calculated at c.86mm from the centre to the outside surface. The surfaces of the daub were lime-washed. The fabric of both collections of daub was examined and included no deliberate tempering material, such as quartz sand, horse-hair or chaff. The petrological characteristics of the daub, as of other daub examined in the study region, were compatable with a local origin.
Loom weights are not common in the study region. All the known examples are likely to be of late Saxon date, probably mainly earlier rather than later in the period. A classification of loom weights by Wheeler produced two main types, the annular and the bun-shaped weight, and a third intermediate category. The bun-shaped form, which has an assymetrical cross-section, is the most common in the study region, although an annular or intermediate weight has been found at Winchcombe.
Examples of loom weights from Hereford include some of probably late 9th to 10th century date, pre-dating the use of pottery on the site, while those from Gloucester are of later 10th to 11th century date.
The fabric of loom weights from Fladbury, Hereford and Gloucester has been examined in thin-section and in each case the petrological characteristics are in agreement with a local origin. The Fladbury and Gloucester weights are tempered, possibly naturally, with a quartz sand and a quartz and limestone sand respectively but the Hereford weights are made from an untempered silty micaceous clay.
The uses of unfired clay in building were widespread and various in the medieval period. Because of its impervious nature clay was often used in the construction of wall footings and dwarf walls and clay floors are a common feature of many medieval structures. At Bartholomew Street, Newbury, they were characteristic of the later medieval period. Clay was used from an earlier period for hearths.
Few samples of clay have been retained from the study region, although their analysis would have been very rewarding. Analysis of the clays from Bartholomew Street, Newbury revealed differences in composition between those of the late 11th to early 12th century and those of the later medieval period and neither clay was precisely similar to that found in patches above the the natural gravel on the site. However, it is not imagined that raw clay would have been tranported more than a couple of miles from its source. The main value of analysis of clays from site is that they reveal the local sources of clay utilised at that period and form a useful comparison with the pottery fabrics.
Ridge tiles are usually made on a sanded surface within a wooden mould. The dimensions of the mould govern the size and thickness of the tile. Tiles for which the length or breadth of the object can be measured are extremely rare but most have measurable thicknesses. These vary from c.8mm to c.20mm. Certain groups of ridge tiles are usually in excess of c.15mm thick, for example Worcester-type, Ashton Keynes, Stroat and North Devon while others are usually thinner, for example Malvern Chase and Hereford A7b.
It is notable that some ridge tiles have a more heavily tempered fabric than the pottery made in the same industry and most contain sparse large inclusions absent from the pottery, for example large rounded pebbles in Worcester-type tiles and fragments of Malvernian rock in Malvern Chase tiles. Tile fabrics also often contain lenses of different textured clay, showing that they were not so well mixed and wedged as clays destined for pottery manufacture.
Certain industries produced glazed but otherwise undecorated ridge tiles, for example Gloucester TF89 in the late 12th to 13th century and the 17th century Herefordshire and Ashton Keynes centres. These tiles often have a thick glossy covering of clear lead glaze.
Most ridge tiles are decorated with applied strips along the crest. Sometimes, when the crests have fallen off, it can be seen that the surface of the tile was roughened up with the fingers to help adhesion.
Two basic groups of crests exist. In the first decoration is applied as a single strip along the crest of the tile and in the second individual knobs (usually two per tile) are added.
The most common design in the first group is when the strip is worked up into a ridge which is then cut with a knife into a coxcomb pattern (fig 8.1). This type is found, for example, at Bristol and in Hereford A7b. Sometimes one or both sides of the crest are stabbed with a knife to give wedge-shaped holes. These may help to attach the crest to the tile as well as allowing excess water to escape on firing. A variant of this type consists of more elaborate patterns cut out of the crest (fig 8.2). These 'fretwork tiles' occur in Hampshire (Platt and Coleman-Smith, 1975) and occasionally in Berkshire (in Newbury tile fabric).
In the second group the knobs are either tall as in Worcester tiles (fig. 8.3) or short as in Malvern Chase tiles (fig. 8.4). Some tiles seen in Worcestershire and Warwickshire have knobs which fold back to touch the tile.
Another method of decorating the single strip tiles is to mould a coxcomb by hand. Such tiles often make use of thumbing to decorate the crests (fig. 8.5). The only medieval examples known are Minety tiles, which have single thumb impressions on either side of each crest. North Devon tiles in the late 17th century have shallow grooves in the same position, possibly imitating the knife-stabbing found on knife-cut coxcomb tiles. Stroat tiles have heavily thumbed crests with thumb impressions at both sides and on top of each crest.
An extremely rare type of decoration takes the form of handmoulded animals or occasionally humans. One example has been found in the region, a bear from Miserden Castle (fig. 8.6). It may be a Minety tile (Dunning, 1979).
The use of glaze on ridge tiles varies from type to type. Clear lead glaze is found on Bristol and Minety tiles while copper-flecked glaze is found on Worcester, Malvern Chase and Hereford A7b. Glaze cover is normally total except on Malvern Chase and some Hereford A7b tiles where it is restricted to streaky patches along the top of the tiles.
Ridge tiles also occur in tile fabrics not made alongside pottery, for example Newbury tile fabric (fig.8.20). Most of these ridge tiles are undecorated but glazed although crested tiles, both with knife-cut coxcomb and fretwork crests exist.
Decoration on the sides of the ridge tiles is rare (fig. 8.7). Bristol tiles often have applied thumbed stips criss- crossing the sides while Minety tiles are sometimes combed or grooved. Some South Welsh tiles, for example from Caerleon, are decorated with roller-stamping (of 'complex rouletting' patterns).
Two sorts of finials exist. Wheelthrown examples fitting into a hole in the ridge tile (fig. 8.8) and wheelthrown or handformed examples luted onto the ridge tile (fig. 8.11). Some finials are found with mortar around the base and were presumably permanently attached to the roof while others, including one still in use in Hanley Castle, merely sit on the roof. It has been suggested that these finials form a stopper for a simple form of ventilation, allowing smoke to disperse quickly in summer but keeping it, and the heat, in during the winter. Wheelthrown examples are known in Malvern Chase and Hereford A7b fabrics and a luted example is known in Minety ware. Ridge tiles with a shallow flange are known in Malvern Chase and Hereford A7b fabrics (fig. 8.9). One tile from Newbury is either part of a similar flanged tile or perhaps the base for an attached finial. Minety tiles often have a circular or oval hole between two of the crests and this may perform the same function as the flanged holes, although no separately made finials are known in this ware (fig. 8.10).
Large handmade louvers are known only in Bristol and Hereford A7b wares (fig. 8.12). They acted as ventilators and would have sat in the roof above the open fire. Louvers are much larger and more impressive than finials and it is probable that their use was restricted to the houses of the upper classes.
Two objects in Newbury Group B fabric have been interpreted by Dunning as Chimney pots (fig. 8.13, Dunning, 1961 b). Both are cylindrical, handmade and unglazed and are decorated with stabbing and circular holes c.10mm diameter. One however has a very narrow diameter, c.40mm.
Many roof tiles, floor tiles and bricks occur in fabrics not used for hollow ware vessels. Even allowing for differences in clay preparation, it is most likely that these objects were not made alongside pots.
The split in the ceramic industry, into hollow ware potters and brick- or tile-makers, is certainly not complete. Flat roof tiles may have been made in the Laverstock pottery kilns while floor tiles and pots were produced in the same industries, if not the same kilns, at Malvern Chase and Nash Hill. However, the division is sufficiently general to be a useful method of classification.
As noted above, ridge tiles were made in tile fabrics in the Kennet valley and also in the London area (Pritchard, 1982). These tiles, while glazed, are usually undecorated. However, in the Severn Valley from the late 16th century onwards ceramic flat roof tiles were made but ridge tiles continued to be made in pottery fabrics.
Tiles used as a roof covering rather than as decorative crests are rare in the study region, occuring sporadically in the 12th and 13th centuries in the Severn Valley. Within the study region, only in Berkshire and possibly south-east Wiltshire was the use of flat roof tiles in the medieval period general. Four main types of flat tile roof were used in the study region between the late 12th and the mid-17th centuries. These are flanged and curved tiles, shouldered pegtiles, standard pegtiles and nibbed and pegged tiles.
Excavations at Reading Abbey revealed a collection of glazed roof tiles in the foundations of a probably 12th century outbuilding of the Abbey. Analysis of these tiles showed that they are of two types; a flat, flanged tile with glaze along the centre and a curved tile. Similar tiles have been found at Southampton (Platt and Coleman-Smith, 1975, nos. 1386-1390), Scarborough (Drury, 1981) and London (Armitage et al., 1982). The Reading examples were weighed and although there were more flat tiles than curved the proportion of one to the other makes it likely that they were used in the Roman manner with the curved tiles locking over the flanges of adjoining flat tiles and being held in place by a slight tapering of both the flat tiles and the curved tiles, as well as by a single peg or nail at one end. No fragments from Reading were large enough to show this tapering, although it is present on a complete tile from London and on a fragment from Scarborough (Figs. 8.14 and 8.15). The Reading tiles were in a different fabric from the Newbury tile fabric pegtiles in the same contexts. A possible corner of a flanged tile was found in an early 13th century context in Hereford in Hereford A2 fabric.
Shouldered peg tiles were first recognised in London in 12th century material from Swan Lane (fig. 8.16). They are thick rectangular pegtiles in which the two top corners have been removed. No evidence for the use of a knife for removing the corners was found, neither are clear traces of the use of a mould present on these tiles (normally the sanded base of a moulded tile curves up into the side, which often has vertical striations where the mould has been lifted off). Instead it appears that the tiles are formed by hand on a sanded surface. Glaze is present on the lower third of the tile and mortar is used to join the middle third of an underlying tile to the bottom third of the one above. Therefore they have an overlap of two-thirds. Both one or two peg holes are found.
Only one fragment of shouldered pegtile has been recognised in the study region, in Gloucester TF89 from a late 12th century context at the Eastgate site. This was only recognised for what it was after the London examples were identified and long after the main survey was completed. It is therefore possible that this type is more common than it now appears, although good negative evidence comes from Hereford and Chepstow where all types of 12th century roof tiles are absent.
Pegtiles are oblong tiles, between c.10 and c.20mm thick with two circular, or more rarely square, holes at one end and normally a clear glaze over the lower half of the tile (fig. 8.17). Such tiles were found in the study region only in Berkshire, at Reading and Newbury. On both sites tiles were present from the late 12th century onwards and were used in walling and hearth construction as well as for roofing. Quite wide variations were found in length, breadth and thickness within the Newbury collection but were not distinct enough to enable sub-groups to be defined. There was no apparent correlation of size with period of use. Pegtile waste was found in the foundations of a 15th to 16th century house at 143-5 Bartholomew Street, Newbury, and there is documentary evidence for the production of tiles in the area (Eames, 1980).
Rectangular tiles with a small rectangular nib worked up from the body at one end and sometimes a single peg hole are found in Hereford A10, Gloucester TF88 and Malvern Chase fabrics. Only the Malvern Chase tiles are sometimes glazed on the lower half (fig. 8.19). One Gloucester TF88 tile had a circular stamp, possibly a monogram 'W'.
Hip tiles were found only in Newbury-type tile fabric at Newbury. One had a peg hole at the narrower end (fig. 8.21).
Hearth tiles are large rectangular tiles without glaze and with moulded rather than knife-cut edges (fig. 8.22). Examples from mid-14th century contexts in Newbury are in a flint-tempered fabric and measure 310-30mm by 267-270mm by 30-1mm. An example from Upton Warren is in a sandy fabric and was found in a 13th century context and hearth tiles are both recorded amongst the products of the Laverstock kilns and from domestic buildings at Gomeldon DMV (Musty et al., 1969). With these exceptions, hearth tiles are rare in the study region. Hearth tiles were used flat to form hearths, at Newbury one hearth was formed from four tiles set in mortar.
Oven tiles are square tiles with rows of conical or thimble- shaped scoops in the underside (fig. 8.23). Small holes are pierced through the scoops to the upper surface of the tile. No complete tiles are known and only one fabric, Ashton Keynes ware, has been identified.
Since the only examples found were in a late 18th century context, this type may be solely 18th century and thus outside the scope of this study.
19th Century moulded oven tiles of similar design are still found forming the floors of Malt Kilns in the Severn Valley but they could as easily have been used in domestic ovens.
Ornate moulded tiles for use in large stoves have not been found in archaeological contexts in the region. These tiles were almost certainly all imported from Germany or the Low Countries and are mainly of late 16th to early 17th century date (the latest types bear the Arms of James I). An object with similar decoration is illustrated by Rackham (1972, Pl.96). This was a cistern, decorated on one side with the Arms of Henry VII and Queen Elizabeth of York, possible from Windsor Castle, Berkshire, and of early 16th century date.
Handmade, moulded bricks are not found in the region until the 16th century. With the exception of bricks from Reading Abbey, Wigmore Abbey and Thornbury Castle all of the bricks could belong to the late 16th century, post-dating the dissolution of the monasteries. The Thornbury bricks are still in situ, including some elaborate decorated chimneys. They can be dated to the second decade of the 16th century but could not unfortunately be examined. Floor tiles made for the Castle are thought from their petrology to have been produced in South Worcestershire and it would be interesting to see whether the bricks have a similar petrology or were made on site.
Most of the 16th to 17th century bricks examined were made in a coarse sandy fabric. This texture often makes thin- section preparation difficult and the results bland. At both Reading and Newbury it could be shown that different fabrics were used for the bricks and roof tiles although no evidence for the source of the bricks was discovered.
The Wigmore Abbey bricks, however, are in a mudstone- tempered fabric identical to that of the late 15th to early 16th century floor tiles (see Ch. 3) and some are splashed with accidental lead glaze.
Brick is found in Malvern Chase, Gloucester TF88 and Hereford A10 fabrics and in each instance the same fabric was used for the production of roof tiles. Brick was being made in Worcester in the 16th century and Worcester brick was used in the construction of the Bishops Palace at Hartlebury (Dyer, 1973). Examples of brick from Sidbury, Worcester, and Evesham have been examined in thin-section and contain a coarse rounded sand, consisting mainly of quartz grains.