The format for description of the floor tile industries is similar to that adopted for the pottery industries in Chapter 2. No illustrations of the floor tile designs are presented here since no complete corpus of the designs used in an industry has been prepared while the range of designs used is shown in many cases by Eames (1980).
Fabric: Hard, mainly reduced very dark grey (10YR 3/1) with inclusions of poorly sorted quartz grains, mainly subangular up to 0.6mm across. Rare fragments of iron ore and a large number of round vesicles caused either by the decomposition of limestone or organic inclusions (probably the former) up to 0.3mm across. The clay matrix is laminated and contains sparse very fine quartz.
Thin-sections: Keynsham; M38, M557.
Source: Tile waste in this fabric has been found at Keynsham Abbey although no kiln has been discovered (Lowe, 1978, 18-20). The clay is very similar to samples of Lias Clay from the area (for example M558).
Description: The tiles are mainly c.135mm square or are subdivisions of this quarry size. They average 27mm thick and have partially trimmed sanded bases and usually have no keying. The sides have a slight bevel.
Plain and white slipped tiles were made and used mainly as scored and snapped shapes, either quarter tile squares or eighth of a tile triangles. The tiles have a clear lead glaze.
Square Decorated Tiles
All these tiles are decorated with deep inlay. Eight designs are known on the 135mm square tiles of which one, 33, is commonly scored for snapping into four.
Four tiles are half the size of the square tiles - 65mm by 135mm. They are used as borders. These may have been fired as scored square tiles (one, 31, certainly was).
One tile, 32, was a quarter square, probably formed from tile 33.
Three border tiles are 135mm wide but are longer, either 175mm (two examples) or 190mm (one example) or 265mm (one example), this stamp is also found on a 265mm by 160mm tile.
Two other sizes of decorated tiles were found, 90mm square and 155mm by 225mm.
The Designs: These are without exception 13th century Wessex types, including several which are very close to the designs used at Clarendon Palace, Wiltshire. Design 2 might be made with the same stamp as one used at Clarendon but the Lion has lost its lower jaw. Other very similar designs are 4, 6, 10 and 11 (compare with Eames, 1957-8, plate XXXIV).
Dating: At least one of the dies used on the Keynsham tiles is grouped by Eames with those used in the earliest tile kiln at Clarendon Palace, dated between 1240 and 1244 (BM design 1271). Other dies were used on the Queen's Chamber pavement at Clarendon Palace, Wiltshire, dated by documentary evidence to 1250-2 (BM design 1967) whilst others still were used at Salisbury Cathedral Chapter House in c.1284 (BM design 2610). The majority of the Keynsham dies are similar but not identical to those used in the Salisbury area which might be an argument for a later date for the Keynsham tiles. The absence of keying, which is present on most 'Wessex' tiles is not a dating factor, since unkeyed tiles occur at Clarendon in c.1250-2 (Eames, 1980). A date in the mid- to late 13th century is likely for the Keynsham tiles but precise dating will not be possible without detailed comparison of the dies used at Keynsham and other sites to establish their relative date.
Distribution: (fig.3.1) Die-links exist between the Keynsham type I tiles and those at Clarendon Palace, Salisbury Cathedral Chapter House and Shaftesbury Abbey (BM designs 1967, 2610). There are no examples of tiles in Keynsham type I fabric from any sites except Keynsham Abbey and a possible example from Bitton Church. The collection of floor tiles from Keynsham Abbey is predominantly of type I tiles.
Fabric: Hard, oxidized reddish brown (2.5YR 5/4) with rare angular red iron ore, rounded micaceous sandstones and mudstones up to 2.0mm. The sandstones contain quartz silt up to 0.04mm across, black mica up to 0.2mm and a high quantity of brown amorphous material. The clay matrix is isotropic and contains angular quartz and some black mica up to 0.2mm and rounded grains of opaque iron ore up to 0.1mm.
Thin-sections: Gloucester Eastgate; M1046.
Source: The characteristics of this fabric are very similar to those of clays from the Devonian Marls and the sedimentary rocks could be of Old Red Sandstone age. The presence of what appears to be black mica both in the sandstones and in the matrix is distinctive. A Somerset source is more likely than one in Gloucester or its environs on the grounds that more sites with similar tiles have been found there than in Gloucestershire.
Description: The tiles have a sharply bevelled edge and knife-trimmed base with usually four shell keys. The designs are deeply inlaid and covered with a clear lead glaze.
The tiles are based on a 135mm square quarry (with examples 120mm and 145mm square). A larger series 200-210 mm square is also found. Rectangular tiles formed by scoring and snapping larger tiles occur in two sizes, 45mm by 135mm and 75-80mm by 140mm. The latter type is sometimes inlaid with a border pattern.
The Designs: The square tiles are inlaid with heraldic shields, bearing the Arms of England, De Clare and the Arms of Richard, King of the Romans, as Earl of Poitou and Cornwall (Ward Perkins, 1941 b, 41). Others bear the double-headed eagle (borne by Richard, King of the Romans and his son, Edmund), designs of tracery and one repeating pattern of interlocking circles containing four fleurs-de-lys. A few tiles are decorated with painted slip lines.
Dating: The heraldic shields must be later than 1252, when Richard Plantagenet became King of the Romans. However, they may be considerably later since the same arms are used at Worcester at the end of the 14th century. L J. Keen states (pers. comm.) that the Blackfriars, Gloucester, examples, if primary to the building date to the period 1260 to 1270 while Ward-Perkins and Eames believe that the most likely terminus post quem for the heraldic tiles is c.1272, the date of marriage of Edmund, son of Richard, King of the Romans, to Margaret de Clare (Eames, 1980, 195). A terminus post quem for the tiles found at the Carmelite Friary, Bristol, is 1267, the date of foundation of the Friary but the presence of vine leaf decoration in some of the designs suggests to Eames that they actually date to the 1280's or later (Eames, 1980, 194-5). An end date to the industry is probably provided by their presence at so many sites in Gloucester. Gloucester was supplied with tiles from the Bredon-type industry by the 1320's.
3.2 Gloucester
Distribution: (fig.3.2) Found at several sites in Gloucester, notably St. Bartholomew's Hospital and the Blackfriars. Also found in a visually identical fabric at Tintern Abbey and Bath Abbey. At Bath the two stamps are identical. Ward Perkins (1941 b) shows that the Cleeve Abbey tiles (which are die-linked to Gloucester) can also be die-linked to a number of sites in Somerset and Avon, as well as at Tintern. The two closest stamps to the Gloucester examples come from Glastonbury Abbey. Three other sites are die-linked; St. Ewen's Church, Bristol; Margam Abbey and Neath Abbey. The history of this group of tilers will not be clear until petrological analysis of the various tiles is undertaken.
Distribution: (fig.3.3) The following sites have tiles identified visually as being of this fabric: Gloucester (St. Bartholomew's Hospital, Cathedral, Blackfriars, Greyfriars, and Eastgate which was found in a post-medieval context); Tintern Abbey; Bath Abbey.
The remaining sites have tiles with die-links recorded by Ward-Perkins (1941b). The fabrics have not been checked personally by the author.
Fabric: Hard, usually incompletely oxidized with a grey core and oxidized sides and base. The fabric is tempered with rounded and angular quartz up to 0.5mm across and larger fragments of rounded and angular iron ore, a sandstone with an opaque iron-rich cement and a white sandstone with quartz and opaque inclusions (rare). The clay matrix contains a high quantity of fine angular quartz.
Thin-sections: (see chapter 2 for hollow wares) Newton St. Loe; M354-355.
Source: Nash Hill, near Lacock, Wiltshire. Two successive kilns were excavated in 1971 by M. R. McCarthy (McCarthy, 1974). Salvage work by the staff of Trowbridge Museum Service has shown that the area of pottery production was extensive and this may also be true for the tilery. The two were in fact closely related at Nash Hill since the two phases of tile kiln were separated by a pottery kiln.
Description: The tiles are of four sizes: c.110mm square; 125-145mm square; 160-165mm square and 200-210mm square. In addition some rectangular tiles were made (Eames, 1974, group III). Three of these tiles have the shorter side the same length as the 125-145mm tiles while a fourth has the long side the same length as the 200-210mm tiles. All have bevelled edges and between one and five shell keys.
Small border tiles, formed by scoring and snapping tiles of the four sizes, are found but in most cases it is difficult to be certain from which tile size they were formed. Oblong tiles made by dividing 133mm and 200-210mm tiles into three are found and square tiles made by dividing 130-145mm tiles into four and nine. Some of these were scored for division into triangles. Two tiles, also formed from larger tiles and scored for division into triangles were 50mm by 54mm and 46mm by 51mm. These may have been formed from 200mm tiles divided into sixteen squares.
110mm square tile designs
Three found at Nash Hill (Eames 1974, 50 to 52).
125-145mm square tile designs
Twenty five found at Nash Hill and many not found at the kiln site but found at Cirencester Abbey, Bath Abbey, Newton St. Loe and elsewhere.
160-165mm square tile designs
Two found at Nash Hill (Eames, 1974, 22 and 46).
200-210mm square tile designs
Eight designs found at Nash Hill (1 to 6 and 12 to 13). Others are known from Cirencester Abbey (Eames, 1974).
Dating: Mrs. Eames dates the latest kiln waste at Nash Hill to the first half (and probably first quarter) of the 14th century. By inference the first kiln must be late 13th century and this kiln contains decorated floor tile in its construction. The heraldic tiles, made in 160mm and 200-210mm square sizes, are datable to the late 13th century or later, and probably post-1280. One of these tiles was found in the phase 1 kiln wall. The tiles from Newton St. Loe include one design probably commissioned for the Castle, the Arms of the St. Loe family with a label of five points. Two possible owners of this crest exist: Sir John de St. Loe (mentioned 1313/4) or his namesake (mentioned 1375). Although the earlier date would fit well with the dating of the kiln site none of the four stamps used at Newton St. Loe are found amongst the kiln waste. There are few tiles of demonstrable later 14th century date in Wiltshire but there is the possibility that the Nash Hill tilery continued in production throughout the century. Eames has shown that the Nash Hill dies include some used first in the Clarendon Palace kiln of c.1240-4 and that in both method and style of decoration the tiles are closely related to other 'Wessex' tile groups (Eames, 1980, 196-9).
Distribution: (fig.3.4) The distribution of tiles with Nash Hill stamps in Wiltshire is discussed by Eames (1974). The largest collections are from Stanley Abbey and Lacock Abbey. Both sites are very close to Nash Hill and were certainly supplied from the kiln site. Amesbury is reported to have one tile with the same stamp as Nash Hill and three with the same designs but from different stamps (not examined in this thesis). In north Wiltshire, Gloucester (Blackfriars), Cirencester, the Bristol Avon valley and at Tintern Abbey tiles of Nash Hill fabric have been examined, including a large number of designs previously unreported and tiles with designs found at Nash Hill but not from the same stamps. There is an apparent lack of Nash Hill tiles in the south of Wiltshire, probably due to the presence of a tile factory in the vicinity of Salisbury.
Fabric: Soft to hard, either completely oxidized red (2.5YR 4/6), or oxidized with a reduced grey core and upper surface. Inclusions of rounded and subangular quartz, poorly sorted up to 1.0mm across, milky and red in hand specimen. Larger inclusions also common, mainly red sandstone, iron ore and quartzite (milky rounded fragments), sparse grey micaceous sandstone and coarse-grained white sandstone. These inclusions are mainly less than 2.0mm across but fragments up to 10mm across (mainly red iron ore) are also found The clay matrix contains fine white mica. The white slip contains few visible inclusions apart from white mica (and was not examined in thin-section).
Thin-sections: Moynes Court; M933 (white slip only, the body was too crumbly).
Source: Clays with similar inclusions were seen at Chepstow Park Wood and are probably boulder clays. Large pebbles of quartzite and red sandstone were visible. A south Welsh source is certain but precisely where is not yet known. No pottery or roof tiles are known with the same fabric.
Description: 115mm to 135mm square and 20mm to 29mm thick. All the tiles were trimmed with a knife after decoration and have four conical keys in the sanded bases. They have a clear lead glaze.
Five methods of treatment or decoration are found:
(1) Division into rectangles (45mm by 135mm) by scoring and snapping. Plain and white slipped examples known.
(2) Division into eight triangular tiles (sides of 68mm, 68mm, and 93mm) white slipped.
(3) Divided into sixteen triangular tiles (sides of 46mm, 46mm, and 68mm) white slipped.
(4) Decorated with inlaid designs.
(5) Decorated with inlaid designs and scored and snapped diagonally.
Designs: Five 'Wessex' designs found at Moynes Court while a tile with the Arms of England set diagonally with fleurs-de-lys in the corners was found at St. Mary's Chapel, Newport.
Dating: Probably late 13th or early 14th century.
Distribution: (fig.3.5) Found at Moynes Court, St. Mary's Chapel in Newport and at Llandaff.
Fabric: Not examined in detail.
Thin-sections: Not sectioned.
Source: Unknown. Some tiles were produced at Chertsey Abbey (Eames, 1980, Chapter 8) and other tiles were specially commissioned for Halesowen, but there is no evidence that they were made on the site.
Description: Thick square tiles with deep inlay and clear glaze.
Designs: Well-cut stamps of the Chertsey-Halesowen 'school'.
Dating: Late 13th to early 14th century.
Distribution: (fig.3.6) Die-links exist between Hailes Abbey, Bredon Church, Evesham Abbey, Bordesley Abbey, Tanworth-in-Arden and Kenilworth Abbey. No fabric analysis has been undertaken nor has any extensive search been made for this type.
Fabric: Hard, oxidized reddish-yellow (5YR 6/6) with a reduced light grey core, the fabric contains scattered angular and subangular quartz grains and red clay pellets up to 0.4mm across and a variable quantity of rounded sandstone and limestone fragments up to 2mm across. The sandstone has a high proportion of brown grains and a brown-stained matrix, the quartz grains in the sandstone range up to 0.1mm across. The limestone is rarer and consists either of fine-grained crystalline or micro-crystalline calcite with fine angular quartz inclusions The clay matrix contains a high quantity of angular quartz and white mica up to 0.04mm across.
Thin-sections: Hereford Berrington Street site 4; M620-621. Gloucester St. Oswald's Priory; M1050
Source: The sandstone and limestone inclusions are matched exactly with those found in Hereford Fabric A2 and in gravel samples from Bewell House, Hereford. The remaining inclusions are found in pottery and clay samples from a wider area of Herefordshire and beyond, although the inclusions in the clay matrix are finer than in many other fabrics of this group, where grains up to 0.1mm across are common. The distribution of the tiles is centred on Herefordshire and a source close to Hereford is reasonable on these grounds. However, no archaeological or historical evidence for a source in Hereford is known.
Description: Six shapes of tile are found, three of which are based on a quarry size of 160mm square. All have a slight bevel and knife-trimmed bases, sometimes trimmed at an angle to the top of the tile. The tiles vary from 21mm to 25mm thick. The bases are normally stabbed with a square-sectioned tool with sides not more than 5mm wide.
Border Tiles
Only one decorated border tile design has actually been found on a border tile, 80mm by 160mm. The design is based on a Chertsey pattern of alternating castles and fleurs-de-lys.
Two tiles from Ludlow show the probable method of manufacture of these tiles since they bear two impressions of the same stamps, one of two quatrefoils containing five-petalled flowers and the other a design of foliage. Therefore these tiles were either intended to be scored and snapped or the stamps were meant to be used on 80mm by 160 mm tiles and were reused on larger tiles.
160mm-square tiles
The majority of Bredon-type tiles are of this size Designs found include single patterns; four-tile patterns; one tile which may be from a sixteen-tile pattern and a large series of heraldic tiles with the coats of Arms set straight on the tiles.
Single tile patterns. There are even designs of this type, which could be used individually or laid together to form repeating patterns. Four designs are included within a lobed border which appear to have been direct copies from the Chertsey school of the West Midlands (see Halesowen-Chertsey-type tiles) examples of which are published from Kenilworth Abbey (Chatwin, 1936, plate II). Two tiles have patterns composed of smaller units, nine in one case (five-petalled flowers in circles) and sixteen in the other (three patterns; a white four-petalled flower against a plain ground, a plain four petalled flower against a white background and crossed vesicas against a white background, there being six examples of the first two patterns and four of the last on each tile). These tiles have very ungainly patterns and again it may be that the tiles should have been scored and snapped to form small border tiles of sides 53mm and 40mm respectively One example of the second stamp at Bredon Church is scored into sixteen but not snapped.
Four-tile patterns Five designs are from four-tile patterns. The centres of the designs are filled mainly with animal representations,such a lion's head, a stag or in one case a frieze of devolved grotesques (BM design 1396). Two designs make up a pattern with a quatrefoil at the centre within a diagonally set square with half circles at the corners bearing in one case a sexfoil and in the other a rabbit beneath a bush, possibly a vine.
Heraldic tiles. Three types of heraldic designs are found, which are all set straight to the tile In the first and most common the bottom corners are plain. In the second they have floral sprays and in the third they have circles filled with flowers. Large collections of these tiles occur at Bredon; Abbey Dore; St. Oswald's Priory, Gloucester, Tewkesbury Abbey and Ludlow and possibly significant differences exist between the distribution of the stamps at these sites. Abbey Dore appears to be the earliest pavement, while St. Oswald's Priory, Gloucester; Tewkesbury and Bredon contain a similar range of stamps. Recent work by A. R. Wilmott (Vince and Wilmott, forthcoming) on the identification of the heraldry casts doubt on some of the previous identifications (Porter, 1887, Holland-Martin, 1933).
Sixteen-tile designs. Two designs could form the centre tiles for sixteen-tile patterns although no examples of the outer designs have been found.
The Labours of the Months At Bredon (H. & W.) the pavement when complete must have contained twelve roundels in situ of diameter 180mm each set in a border of four 160mm tiles. The corners of the border tiles have been cut away in a quarter circle before firing. Only six roundels now exist at Bredon, none of which retains the inlaid design. However, a single circular tile at Colwall Church (H. & W.) of a man digging must be from a similar pavement. Several of the Malvern Chase tiles from Little Malvern are square tiles stamped with designs more appropriate to these roundels. One tile has the same design as the Colwall tile, but with a crack across the centre. This crack is visible at Colwall but is more pronounced at Little Malvern.
The border tiles at Bredon still retain a little of the original design, enough to show that the name of the month in crude Lombardic lettering was present with a motif in the corner. Two of the surviving roundels are surrounded by the same stamp border, suggesting that the paviours who laid the pavement could not read or were not supplied with the right tiles.
190mm-square tiles.
Hereford and Leominster have produced square tiles bearing more elaborate versions of the designs found on the 160mm square tiles. Four designs are known, one of which bears three letters in Lombardic script.
240mm-square tiles
One tile at Leominster is 240mm square, it bears a lion surrounded by a circular border with fleurs-de-lys in the corners.
Dating: The heraldic tiles at Bredon are dated by Wilmott to c.1300 to 1320, a dating confirmed by the architectural dating of the Chancel (Keyser, 1912, 4). Several of the 160 mm square tiles have patterns almost certainly copying the Chertsey school tiles found for example at Hailes Abbey and Evesham but not yet closer to Herefordshire (except for a single very worn example from Bredon itself). The prototype tiles do not survive at any of these sites but are well represented at Kenilworth Abbey (Chatwin, 1936). There is little evidence for any stylistic progression, although two or more stamps of some designs are known. This does not have to imply a long-lived industry however since at the Danbury tile factory several versions of the same design were in use at the same time (Drury and Pratt, 1975, ^^ ).
A large number of the Bredon-type stamps were later used at Malvern Chase (see below) but the larger tiles have not been found in that fabric. This suggests that they either went out of use before the movement of stamps to Malvern Chase or came into use afterwards.
There is no evidence that the Bredon-type tilery continued to exist after the movement of stamps to Malvern Chase and it is reasonable to see the larger and better-cut stamps used on the 190mm and 240mm square tiles as earlier tiles.
Distribution: (fig.3.7) All of the sites with Bredon-type tiles are easily accessible from Hereford. It is difficult to quantify the frequency of Bredon-type tiles since we do not know which other tile groups were contemporary. The absence of Bredon-type tiles from Evesham Abbey and Hailes Abbey is probably significant negative evidence, whilst the only site to produce them in Gloucester is St. Oswald's Priory, where all of the tiles come from the area of a single chapel.
Hereford has produced these tiles from four sites, one of which, Berrington Street IV, is a recent dump of tiles from an unknown source. To the west of Hereford the distribution is obscure because of lack of data.
Fabric: Hard, usually only partially oxidized, although completely oxidized tiles are found. The fabric is tempered with rounded quartz up to 0.3mm across with rarer clay pellets up to 0.4mm and chert up to 0.8mm. Malvernian rock fragments are rare but can be up to 10mm across. The clay matrix contains a high quantity of angular quartz up to 0.1mm across. This is coarser than that found in Great Malvern tiles (see below), although within the same size range. One tile from Bristol Greyfriars has a thick slip of fine micaceous clay over a typical body fabric. The tiles are inlaid with a white slip (not petrologically examined).
Thin-sections: Cirencester; M743. Ewenny Priory; M1006. Gloucester Cathedral; M1007. Gloucester St. Bartholomew's Hospital; M182-3. Hereford Blackfriars; M349.
Source: Samples of Malvern Chase tiles from Gloucester Southgate Street, Gloucester Cathedral (B.M. 11,459) Hereford, and Cirencester have been thin-sectioned and all have a similar fabric. This is quite different from that of most Malvern Chase hollow-wares (see Chapter 2). These have a higher quantity of rounded quartz sand and a lower quantity of angular quartz in the matrix. Until a kiln site is excavated it is not possible to say whether this fabric difference is due to the tiles being made at a separate site within the Chase or whether quartz-rich clays were being chosen deliberately at an establishment where both pottery and tiles were being produced. Waste tiles have been found at both the Gilbert's End and Hanley Swan sites (see chapter 2, fig. 00) but not in sufficient quantities to show that they were being produced there. A sample of floor tile from Ewenny Priory (M1006) has a very different fabric to that of the other floor tiles (which are all probably 14th century), but is similar to that of the 16th century pottery from Malvern Chase. No Malvernian rock fragments were present in the thin-section examined. No samples of tiles from Abbot Parker's monument were available for analysis and the origin of these 16th century tiles remains in doubt.
Description: Malvern Chase tiles are found in five sizes; 120-130mm, 130-140mm, 165mm, 190mm and 205mm. All tiles have a slightly bevelled edge and a knife-trimmed base. Square-sectioned stabbing occurs rarely on the bases. The glaze can be clear or green-flecked. Undecorated tiles can be either plain or have a brushed-on white slip under a plain or green glaze.
120-130mm square tiles. Only three designs are known on this size of tile. One, found on an unprovenanced tile in the Gloucester City Museum is based on a Great Malvern design bearing the arms of the De Clare family. This must date to the second half of the 15th century or later. Another, with a pattern of interlocking circles is a copy of a design found on Droitwich-type tiles.
130-140mm square tiles. Six designs and plain tiles with a white slip are found on this tile size. Four of the designs are made with Great Malvern stamps (GM11, 78, 81 and 82). Three of these designs are used in the Gloucester Cathedral Lady Chapel pavement. The remaining two are not Great Malvern stamps and include the corner of a 16-tile pattern and an incomplete design of foliage.
165mm square tiles. At least 20 designs are known of this size of tile and many more fragments probably are from similar sized tiles. Seven single and 4-tile pattern designs are made with Bredon-type stamps. In one case it can be shown that the Malvern Chase tiles are later because a crack has developed across the stamp. Several heraldic designs are also shared with the Bredon-type tiles, also made with the same stamps. Finally five of the Bredon 'labours of the months' roundel stamps are re-used on Malvern Chase tiles. Here too one of the stamps has developed a crack after its use on the roundels. Of the remaining tiles, several of the designs are very similar to surviving Bredon-type tiles and may also be Bredon-type tile stamps reused on Malvern Chase tiles. This implies that the Bredon-type tile prototypes still have to be discovered.
The Gloucester Greyfriars pavement and loose tiles from the same site bear a number of designs which are close to the Bredon-type designs but not made with the same stamps. These may be a later series of stamps made solely for Malvern Chase after the original Bredon-type stamps had worn out. In three cases these degenerate copies can be linked with their prototypes. In the process of copying the Arms of Bishop Trilleck of Hereford, they have been reversed.
190mm square tiles. Plain and white slipped tiles are known in this size of tile. Some are scored for snapping diagonally.
205mm square tiles. Ten designs can be recognised on tiles of this size but most a represented only by small fragments. Unlike all of the preceding groups no tiles are known in other fabrics and it therefore appears that the stamps were cut at Malvern Chase. One of the designs is loosely based on a Great Malvern design. Two others are diagonally set shields bearing in one case the Arms of St. Peter and St. Paul and in the other the Arms of Malmesbury Abbey, possibly bearing the initials of Abbot Thomas Bristol, 1436-1456 (see Malmesbury-type tiles). Three tiles are from four-tile patterns with a design of twisted stalks. In one case the centre of the pattern is composed of a chained bear climbing a tree trunk. All these designs have similarities and are probably the work of one craftsman. The use of Great Malvern motifs must date the series to the 1450's or later.
Dating: Only three medieval pavements containing Malvern Chase tiles have been found. The Gloucester Cathedral Locutorium (Treasury) pavement contains 190mm square plain and white slipped tiles set parallel to the walls of the passage. It can be dated to the late 14th or early 15th century or later since it lies in a passage extended at the time of the construction of the Singing School and Library above. The Gloucester Greyfriars pavement was excavated in 1974 and only small patches survived. All the tiles were 165mm square inlaid tiles and were set diagonally with bands of tiles set parallel to the walls to separate the panels. This pavement can only be dated stylistically, except that it was earlier than the surviving early 16th century structure and was also cut by a series of graves. Comparison with the Bredon-type tiles shows that the pavement must be of early to mid-14th century date. The third pavement is that surrounding the Monument to Abbot Parker in Gloucester Cathedral. This must date to the early 16th century. Abbot Parker died c.1539 but his Monument may well have been under construction during his lifetime.
Another two fixed points in the chronology of the Malvern Chase tiles are the re-use of the Bredon-type tile stamps, some time after c.1327, and the re-use of stamps from the Great Malvern tilery probably in the 1480's.
The following chronology is most likely-
i) c.1330 or later. Start of industry (by Herefordshire tilers?).
ii)c.1350. Gloucester Greyfriars pavement and other new stamps replacing Bredon-type originals.
iii)c.1350-c.1450. Gap in industry or a switch to plain tile production (eg. Gloucester Locutorium pavement). Possibly some 120mm square tiles with designs copying Droitwich-type tiles.
iv) c.1450-c.1540. Copying of Great Malvern pattern on 120mm square tile.
v)c.1480+. Re-use of Great Malvern stamps on 130-140mm tiles. 205mm square tiles made. Abbot Parker's monument pavement (and Ewenny Priory pavement?).
There is no evidence that the production of decorated tiles survived the dissolution of the Monasteries.
Distribution: (fig.3.8) The distribution of Malvern Chase tiles is heavily biased in favour of archaeological collections. This may be due to the poor quality of the tiles, both artistically and technically. The Gloucester Cathedral pavement was in a passage which for many years had been blocked at its western end but which was unblocked in the mid-19th century. In the century in which the passage was reopened considerable wear took place. The absence of Malvern Chase tiles in South Worcestershire is almost certainly due to the absence of archaeological collections in the area. The same is probably true of Herefordshire. Where the tiles are found, for example Little Malvern, Worcester, Hereford, Breinton and Much Marcle, they are in archaeological contexts rather than relaid in church floors. Too few findspots are present to plot any changes in the distribution with time.
Fabric: Hard, oxidized reddish yellow to red (5YR 6/6 to 2.5YR 5/6) or sometimes overfired. The fabric is tempered with a variable quantity of sand, consisting of rounded grains of quartz, brown chert, silicified sandstone and micaceous sandstones and siltstones. These inclusions are mainly up to 0.6mm across with some much larger rounded fragments sometimes present.
The clay matrix contains angular and subangular quartz up to 0.2mm across and some white and black mica up to 0.1mm across.
Thin-sections: Keynsham Abbey; M40, M561. Gloucester Telephone Exchange site; M42. Tewkesbury Holm Castle; M381-3. Gloucester St. Oswald's Priory; M1047. Hereford Berrington Street site 4; M1131.
Source: The composition of the rounded sand in Droitwich-type tiles is very similar to that found in Worcester-type jugs and Canynges-type tiles. Both of these fabrics are thought to have a south Worcestershire source.
Subdivisions of the Droitwich-type tile group on fabric grounds have been attempted at Gloucester, where the tiles were divided into two type fabrics, Gloucester TF86 and Gloucester TF114. The latter fabric had a much higher quantity of sand and in consequence a poorer quality impression of the design. At Keynsham Abbey the Droitwich-type tiles were divided into seven sub-groups (Lowe, 1978, 30-35, Types IVa to VIII). Despite these differences in appearance and quantity of inclusions the similarity in inclusions and in the designs used probably indicates that a single 'factory' was responsible for the manufacture of all the Droitwich-type tiles.
One kiln has been found in which tiles of this type have been found, that excavated in the 1830's at St. Mary Witton, Droitwich. However, as will be shown below, there is some doubt about the attribution of the surviving tiles to the kiln.
The tile kiln was discovered in 1837 in the graveyard of St. Mary Witton. This land had then only recently been consecrated and may have had no ecclesiastical connections in the medieval period. Soon after its discovery the site was re-excavated by Jabez Allies..
The kiln consisted of two parallel rows of circular arches. 'The arches were rather flattened, and there were several of them in each row, and each was two feet six inches high, two feet four inches broad, and six and a half inches deep; that is the depth of the bricks of which they were built; these bricks are five and a half inches broad, and one inch and three quarters thick, and are squared at the corners on the one side. The intervening space between each arch was five inches; that is the diameter of the encaustic tiles which filled up the sides between the arches to the bend of them, and which tiles were cemented horizontally upon one another with red cement. The whole resembled the skeleton of the back of a horse or an ox' (Allies, 1856, 103-5).
Allies goes on to state that he and his companions had originally thought that the structure was used for boiling brine and that its true nature was pointed out to them by Albert Way, Director of the Society of Antiquaries. He also lists a number of the designs found, examples of which were donated by the rector to the Worcestershire Museum.
'... one of them contains the representation of an archer with a long bow, dog, trees, and something like an owl; another has two birds with their backs towards, but looking at each other; another has a lion; another has the first half of the alphabet in Longobardic characters; another has fleur-de-lis; and another has the ancient symbol of the Christian faith, viz., a fish enveloped in its own bladder, like an oval ring, styled by antiquarians "vesica piscis"..'
From this description there can be little doubt that a small collection present in the Worcester City Museum is that donated by the Rev. Mr. Topham (Worcs. City Museum Acc. No.1977/504a to f). However, these tiles are complete and not unduly warped or blistered. They may therefore have been tiles incorporated into the structure of the kiln. A tile in the Victoria and Albert Museum is recorded as being from the Droitwich kiln (Acc. No.5660-1901). Two tiles in the British Museum collection were also recorded as being from the Droitwich kiln. One of these tiles is now lost while the other is a Canynges-type tile (BM cat 11431, design 2953). Both the Victoria and Albert and British Museum tiles are in good condition and not obviously wasters. While it is fairly certain that the Worcester City Museum tiles were found on the site even these cannot be used as conclusive evidence for the products of the kiln.
No documentary evidence for medieval tilers operating in Droitwich has been found, although tilers were operating at the right date, the late 14th to 15th centuries, in Worcester (see Canynges-type tiles, below).
Description: The Droitwich-type tiles vary from c.110mm square to c.155mm square but all are made by the same methods. The tiles have sanded bases and slightly bevelled sides. The sides are often covered in excess glaze and white slip runnels, often with traces of other tiles stuck on to the sides. The tiles are either plain lead glazed (appearing black or brown), plain glazed over a white slip (appearing yellow) or decorated with various patterns, probably applied by the stamp-on-slip method.
In many cases the method used is not discernible. All that can be said is that a very thin inlay, often no lower than the surrounding body, was used. A possible chronology for the use of the various sizes and shapes of tiles has been worked out but the tiles are described below by type of design and shape.
Plain tiles. Most of the plain tiles found are between 115mm and 125mm square or were produced by scoring a tile of this size before firing and then snapping the tile into the required shape after firing.
A: Snapped into two oblong tiles, c.120mm by 60mm.
B: Snapped into four square tiles, c.60mm.
C: Snapped into nine square tiles, c.40mm.
D: Snapped into three oblong tiles, c.120mm by c.40mm.
E: Snapped into two oblong tiles, c.120mm by c.40mm and c.120mm by c.80mm.
F: Snapped into six rectangular tiles, c.60mm by c.40mm.
G: Snapped into eighteen triangular tiles with sides of c.40mm, c.40mm and c.55mm.
H: Snapped into two triangular tiles, along one diagonal. Sides c.120mm, c.120mm and c.170mm.
I: Snapped into four triangular tiles, along both diagonals. Sides c.85mm, c.85mm and c.120mm.
J: Snapped into eight triangular tiles. Sides c.670mm, c.60mm and c.85mm.
White slipped tiles. Tiles covered with a white slip and clear glaze can be found in all the shapes described above but are less common than plain brown or black glazed tiles.
Green glazed tiles. Tiles with an overall white slip but a green glaze, due to the presence of copper, have been found but are rare.
Mosaic tiles.
Table showing the incidence of decoration on the Holm Castle, Tewkesbury, tile mosaic.
Shape |
a |
b |
c |
d |
e |
Total |
M1 |
3 |
3 |
|
|
|
6 |
M2 |
|
5 |
1 |
3 |
|
9 |
M3 |
|
2 |
1 |
|
|
3 |
M4 |
|
1 |
2 |
|
3 |
6 |
M5 |
|
1 |
|
|
1 |
2 |
M6 |
|
1 |
2 |
2 |
1 |
6 |
M7 |
|
|
1 |
|
1 |
2 |
M8 |
|
1 |
|
|
|
1 |
M9 |
2 |
|
3 |
|
|
5 |
Total |
5 |
14 |
10 |
5 |
6 |
40 |
a: plain lead glaze
b: green glaze over a white slip
c: plain glaze over a white slip
d: plain glaze over a stamp-on-slip fleur-de-lys
e: uncertain
One group of mocsaic tiles has been found in this fabric, at Holm Castle, Tewkesbury. Nine different shapes were found, of which two were decorated with stamp-on-slip fleurs-de-lys.
M1. A equilateral triangle.
M2. A diamond.
M3. A triangular tile with the shortest side concave.
M4. A segment from a circular band.
M5. A shape with two parallel or slightly tapering sides and one short concave side.
M6. A shape with two concave and one convex side.
M7. An identical shape to M6 but with a central hole.
M8. A shape with one straight side, possibly snapped along a pre-firing scored line, two concave sides and probably two straight sides.
M9. A shape very similar to M8 but smaller. This shape has not been snapped along the scored line, which is present on all examples but one.
Two dies were used on these tiles, both fleur-de-lys. Neither is found on a square tile. No other examples of tile mosaic have been found in this fabric.
Decorated tiles. Three large collections of Droitwich-type tyiles have been examined and all the designs recorded. They are the collections at St. Oswald's Priory, Gloucester, Little Malvern Court (from the site of the Priory) and Holwm Castle, Tewkesbury. Smaller collections from Evesham Abbey and various sites in Gloucester have been examined. The information on the Worcester Cathedral pavements, Hailes Abbey and Keynsham Abbey is derived mainly from published or forthcoming series of tracings, although the fabric of some Hailes and Keynsham tiles has been examined (Sassoon, forthcoming; Lowe, 1978). Many more collections of Droitwich-type tiles exist but I have not been able to catalogue them in detail. This includes a number of collections relaid in parish churches, where there is the possibility that the tiles were not made with the same clay. Where it has been possible to examnine and compare tracings and tiles it is clear that several designs are represented by more than one die However, there is a large amount of subjectivity involved in deciding whether or not the designs on a pair of tiles decorated by the stamp-on-slip method were made using the same die.
100mm tiles. These tiles mainly have single or repeating patterns, although some 4-tile patterns are known. Many of these designs occur on 120mm tiles but made with different, larger dies. 100mm square tiles vary from 22mm to 33mm thick. 17 designs have been recorded, including BM designs 2479, 2397, 2316, 2303, 2243, 2284, 2289, 2209, 2186, 2036 and 2126.
Many of these designs, quite possibly made with the same dies, are found in the London area on 'Westminster' tiles and in Warwickshire on Stoke (Coventry) or Chilvers Coton tiles (Eames, 1980, catalogue). The 'Westminster' tiles are unlikely to later than the early 14th century since the London market in the late 14th century was flooded by Penn products.
120mm tiles.
120mm square tiles are the most common type found and include the widest range of designs.
i) Lombardic characters. Made on square tiles with four letters to a tile, snapped after firing. Four separate designs have been recorded, all from Evesham Abbey.
ii) Animals and figures within a circle having fleurs-de-lys in the corners. Designs include a lion facing left, a lion facing right, two lions back to back, a knight on horseback and a griffin. This type of design is typical of the 'Wessex' school and was in use at Clarendon Palace in the mid-13th century but dies of this type were still being used in the late 14th century at Worcester Cathedral.
iii) Designs set within a diagonally placed square, the corners of the tile being formed of a contrasting colour. The central designs of this type are either flowers or fleurs-de-lys. Six dies have been recorded.
iv) Other single and repeating patterns. The remaining single and repeating designs include a fish in a diagonal vesica (two dies known, one from the Droitwich tile kiln), two birds facing a tree with their heads turned away (three dies known, BM design 1961), interlocking arcs from circles (three dies known, BM design 2291), interlocking vesicas (two dies known), two birds set diagonally on the tile with their heads facing (two dies known, one from the Droitwich tile kiln), single fleurs-de-lys (seven dies known) and four fleurs-de-lys separated by a white cross (similar to BM design 2205, found snapped into four squares at Fladbury church, Hereford and Worcester).
v) Heraldic designs with a shield set square to the tile and parallel lines in the bottom corners. Arms represented include De Clare (three dies known), the Arms of England (2 dies known) and six others.
vi) Heraldic designs, as (v) but with other patterns, mainly floral, in the bottom corners. Arms known include those of Beauchamp (three dies known), St. Peter and St. Paul (two dies known, probably used for Gloucester Abbey), Berkeley (two dies known), Despencer, Beauchamp of Holt, the impaled arms of Beauchamp and Ferrers, the dimidiated arms of Beauchamp and Mortimer, Mortimer, Cantilupe and the See of Worcester.
vii) 4-tile patterns needing four different dies to make up the design. Four designs are known; the Arms and badge of Richard, King of the Romans, the Arms of Robert Fitzhamon, the founder of Tewkesbury Abbey (still laid in the Founder's Chapel, Tewkesbury Abbey) and the Arms of Sir John Talbot, 1st Earl of Shrewsbury, quartering his first wife (Furnival) and impaling his second wife (Beauchamp). A complete set of the last design is in the British Museum collection (BM designs 1711-4).
viii) 4-tile patterns made by repeating one die four times. Eighteen dies are recorded including BM designs 2647, 2676, 2710, 2752, 2762, 2799 and 3000. Some of the plainer designs are very similar to those used in the late 13th to early 14th century on tiles of the 'stabbed Wessex' type.
ix) 9-tile patterns. One 9-tile pattern is found but at least three sets of dies were used to produce the outer tiles, each consisting of two dies. The centre of the design would be made of a single tile. Only one tile has actually been found used with this pattern.
x) 16-tile patterns. A number of 16-tile patterns are found. The most common incorporates the arms of Beauchamp. Other designs include a quatrefoil within a circle (two sets of dies known). Dies represented on tiles in the British Museum collection include designs 2162, 2916, 2982, 3003-4, 3006, 3010, 3035 and 3041-2.
135mm tiles.
Although there is some degree of variation in the size of the quarries of the 120mm tiles there is a series of tiles with a slightly larger quarry, between 130mm and 140mm square. The tiles include single, 4-tile and 16-tile patterns, many of which are larger versions of dies used on 120mm tiles. Tiles in the British Museum collection of this size are decorated with designs 1844, 1886, 2260, 2668, 2683, 2770, 2811, 2908-10, 2931-4 and 2986 but the largest published collection is that from Keynsham Abbey (Lowe, 1978). None of the simple patterns found on the 100mm tiles occurs on a 135mm tile, neither is there much overlap with the late 14th century 120mm designs used at Worcester Cathedral Singing School.
155mm tiles.
Three designs are known on 155mm square tiles: a fleur-de-lys, a 4-tile pattern and the corner tile from a 16-tile pattern. All three are larger versions of designs found on 135mm tiles.
Dating: The earliest sound dating evidence for Droitwich-type tiles is provided by the Worcester Cathedral Singing School pavements. Documentary evidence for the construction of the buildings in which the tiles were laid dates the pavements to c.1377 or later. The tiles from these pavements include both 100mm and 120mm tiles decorated with dies which are stylistically identical to those of the late 13th to 14th century 'Westminster' and 'Wessex' schools. Any earlier tiles made in the same tilery are therefore unlikely to have stylistically earlier designs and therefore the only evidence which can be used to date the inception of the industry is provided by archaeological evidence alone.
The Holm Castle tiles have two claims to be earlier than the Worcester Cathedral pavements. Firstly, they include a collection of mosaic tiles, which are unlikely to be later than c.1350, since tile mosaic is exceptionally rare after the first half of the 14th century (Eames, 1980). Secondly, the site of Holm Castle was apparently abandoned in the early 14th century and certainly declined in status. Holm Castle, Tewkesbury, and St. Oswald's Priory, Gloucester have both produced quantities of small snapped tiles, of rectangular and triangular shape. The use of these small border tiles is also thought to be predominantly an early 14th century and earlier feature.
Conversely, it is unlikely that the Droitwich-type tiles would have been made at the same time as decorated tiles were being produced in Bredon-type and Malvern Chase fabrics, since these tiles had a wide distribution over exactly the same area covered by Droitwich-type tiles. It is known that the shift from Bredon-type to Malvern Chase fabric must be later than c.1330 and the Malvern Chase tiles were probably not made for less than 10 years, otherwise they would probably have a limited distribution. Therefore a date late in the 1330s is the earliest at which one may suggest that Droitwich-type tiles were made.
Although some of the designs, and possibly the actual dies, used on Droitwich-type tiles were probably made in the late 13th century this cannot be used as evidence for such an early date for the industry since dies could have a long life.
The two lines of evidence therefore combine to indicate a date c.1340 for the start of the Droitwich-type tilery. Since the Holm Castle, Tewkesbury, and Worcester Cathedral Singing School designs are very similar only close study of dies is likely to distinguish tiles of the mid-14th century from those of the late 14th century. One important group of tiles of this general date is that found at Gloucester Cathedral, both relaid in the eastern chapels (St. Edmund's Chapel, St. Edward's Chapel and St. Stephen's Chapel) and in situ surrounding the monument to Edward II. Although all of these tiles must be later than c.1327 their exact date is not known. Reconstruction of the east end of the Abbey church was in full swing in the mid- to late 14th century.
Late 14th to mid-15th century pavements are illustrated by those in the Founder's and Beauchamp chapels at Tewkesbury Abbey. These pavements are dated c.1395 and c.1437 respectively and both consist of 120mm tiles laid diagonally to the walls of the chapel. Only five dies each occur on tiles in these pavements, a 4-tile pattern of the founder's arms and a fleur-de-lys in the former pavement and a 16-tile pattern and a rose in a garter in the latter.
The latest collections appear to be those found at Hailes Abbey, Keynsham Abbey and Fladbury church. None of these tiles are stratigraphically associated but the absence of such early features as small border tiles or dies used in the Worcester Cathedral Singing School pavements is sufficient to indicate that most of the tiles in these groups are later than those at Worcester. Only the 9-tile and one 16-tile pattern at Keynsham have any parallels at Worcester and even there different dies were probably used. These tiles were in fact slightly smaller than the average Droitwich-type tiles at Keynsham which are 135mm square. Hailes, Fladbury and Keynsham are linked together by die-links and the use of the same designs but made with different dies. Although there are no die-links between this group and the tiles at Evesham Abbey and Little Malvern Priory both sites have produced 135mm tiles, which seem to be a late feature. The latter site also produced 155mm tiles, which on the evidence of their dies must also be 'late'.
It is not known quite how late the Droitwich-type industry may be. The Beauchamp chapel pavement at Tewkesbury is the latest dated collection but some of the 16-tile patterns may have been the prototypes for the Canynges-type 16-tile patterns and it is possible that these two industries were sucessive or even slightly overlapped in date. However, the date of the inception of the Canynges-type industry itself is in doubt. If the dubiousd evidence for a commission for the Hungerfords in the mid-15th century is omited then the earliest dated tiles may be later than c.1480. It is perhaps more economical to imagine a succession of industries in south Worcestershire with the Droitwich-type tilery in decline or having ceased production by c.1450 when the Great Malvern tilery started and the Canynges-type tilery taking over from Great Malvern in the 1480's.
Distribution: Roughly speaking, the 100mm square tiles are early, and the 135mm and 155mm square tiles are late while the 120mm square tiles occur throughout the operation of the tilery. Therefore a study of the distribution of tiles by their size may reveal differences in distribution between the mid-late 14th century and the early to mid-15th century.
Two features are revealed by this study (figs.3.9 to 3.11). Firstly, the distribution in south Shropshire, Hereford and Worcester and Gloucestershire is not significantly different between the two periods. However, the tiles found in the lower Avon, in Gwent, Avon and beyond and consistently 'late'. This suggests firstly that the industry increased in scale during the late 14th to early 15th century and secondly that Bristol and the lower Avon must have been supplied by another tilery at this time. One possible candidate for this late industry is that at Nash Hill, Lacock.
Fabric: Not described in detail, fine sandy.
Thin-sections: None.
Source: Unknown. The method of decoration and the style of decoration suggests a connection with the Droitwich-type tiles or the 'Newbury' tiles of west Berkshire.
Description: Stamp-on-slip decorated tiles, including examples with the arms of Berkeley, Beauchamp and Despencer (BM designs 1609, 1594, 1696) and the griffin, the arms of Malmesbury Abbey flanked by initials identified by Brakespear as those of the Abbots and Abbey or Malmesbury (BM designs 1474-6), also a border tile with a monkey examining a glass urinal (BM design 1329). A 9-tile pattern was found (BM design 2856) and a 16-tile foliage pattern (BM designs 2914-5).
Dating: The heraldry apparently dates this group to the late 14th and early 15th centuries (Brakespear, 1913 & 1914).
Distribution: (fig.3.12) Malmesbury-type tiles have a limited distribution. They occur at Malmesbury Abbey and a small number are present in the Corinium Museum, Cirencester, where they are recorded as being found at Cirencester Abbey.
GREAT MALVERN TILES
Fabric: Hard, Oxidized red (2.5YR 5/6). Rounded clay pellets with few inclusions up to 0.6mm across. Sparse angular Malvernian rock fragments (orthoclase felspar, often altered, quartz, hornblende) up to 2.0mm across and rare rounded quartz grains up to 0.4mm across.
The clay matrix contains abundant quartz fragments, between 0.02mm and 0.1mm across.
Source: Great Malvern tiles were made in a tilery situated to the east of Great Malvern Priory church. A kiln was discovered there in 1833, about 200 yards east of the church and a description was immediately published (Eginton, 1833). The structure was set into the hill-side and consisted of a long chamber with a flue underneath it. The chamber had a removable floor. A discussion of the technical aspects of the kiln was published in 1887 (Anon, 1887). Amongst the tiles found in 1833 in or around the kiln was one tile whose die can be identified (GM36). The remaining tiles were simply said to correspond with those in Great and Little Malvern churches. The identification of this structure as a kiln has been doubted, notably by Haberley (1937).
An overfired tile now in the British Museum has been thin-sectioned and has a similar petrology to that of tiles from Great Malvern church, Abbot Sebroke's pavement in Gloucester Cathedral and elsewhere.
Thin-sections: Gloucester; Gloucester Cathedral; Great Malvern Priory; Bristol Horsefair; Hailes Abbey;
Description: The British Museum collection includes large numbers of Great Malvern tiles and tiles of other fabrics stamped with Great Malvern dies. Each die recognised on the British Museum collection tile has been given a BM design number while the tiles themselves have BM catalogue numbers. However, the British Museum collection includes only a half of the dies encountered by the author, who did not have the facilities to draw all of them. Instead, black and white photographs were used to compare dies. It is to be hoped that the Census of Medieval Tiles for Hereford and Worcester and for Gloucester will provide the vehicle for the publication of a single corpus of Great Malvern tiles. In the interim, this thesis uses an unpublished series of numbers, prefixed GM, except where a BM design number has been assigned to the die.
All the Great Malvern tiles have a slight bevelled edge and untrimmed bases, coated in quartz sand (not Malvernian gravel). Clear glaze covers the top and often the sides of the tiles and is often found as dribbles on the base of the tiles. Several distinct series of tiles can be recognised, both on the basis of size or shape and on the basis of design.
The Large Rectangular Series. (GM1-4, 40-1). Six dies are known which were used on large wall tiles. The only examples known now surround the altar at Great Malvern Priory but none are apparently in their original position. The top tile of the series is inscribed 'Anno d m cccc liii' (1453AD). The bottom tile of the design was formed by two alternative dies, GM3 and GM40. The remaining tiles could fit anywhere in the design, although as reproduced by Shaw the tip of the crown on GM41 is present on GM1. This detail is not clear on the surviving tiles. Four of the designs bear shields, whose heraldry is described by Porter (1890, 155-6). The shields bear the arms of England; the coat of Edward the Confessor (later the Arms of Westminster); the Arms of Bohun, Earl of Hereford; the Arms of Mortimer of Kyre Wyard; Le Despencer; Beauchamp of Powyke; De Clare, Earl of Gloucester; Beachamp, Earl of Warwick; Skull of Wichenford and Stafford of Grafton.
No plain or white-slipped tiles are known in this series.
The Sebroke Pavement Series.
The pavement commissioned by Abbot Sebroke of Gloucester Abbey was laid in the presbytery and choir of the Abbey church. An area of this pavement, in front of the high altar survives intact but much was removed during alterations to the choir. Examples of Sebroke Pavement tiles in the British Museum and from Northgate Street, Gloucester (site 74/68 Tr.IV layer 2), probably derive from this lost portion of the pavement.
The surviving portion contains tiles laid diagonally with single, 4-tile, 9-tile and 16-tile patterns separated by bands of plain tiles. Four sets of tiles incorporate either inscriptions or heraldry which show that their dies were made specially for this pavement.
The first of these is a nine-tile pattern bearing an inscription which includes the date 1455. A transcription of this inscription reads "Ecce qua bonu et quam iocundum habitare fratres in unum fiat amen. R. Brugg, J. Appl'bi, W.Farlei, R.Hullei, nefact'hui' loci A D M CCCC L V. Dompnus Thomas Sebrok Abbas G." This set dates the commissioning of the pavement to c.1455. (BM design 1473 is the centre of this pattern, other tiles are GM50-3).
The second set is a nine-tile pattern bearing the Arms of Gloucester Abbey and Westminster. (British Museum designs 2565, 2859-60). The third set is a 16-tile pattern bearing the Arms of Westminster (GM74-77; British Museum designs 2897-9) and the fourth set is a repeating 4-tile pattern bearing the arms and motto of Abbot Sebroke "Fiat voluntas dei". (GM68). In addition, there are two dies which may have been used on other parts of the Sebroke pavement but which do not appear on the surviving part of it. One is a repeating 4-tile pattern similar to GM68 but bearing the Royal Arms (GM21; BM design 1480) and the other is a 4-tile pattern similar in design to the 9-tile pattern of GM65-7.
A number of tiles are found with designs incorporating a quarter of a quatrefoil in each corner, or on the edge tiles of the design, in two of the four corners. There are six designs in the series, but one has a variant used on a rectangular tile (GM15, BM design 2578; GM15a, BM design 2579). One die is used in a cracked and uncracked condition (GM16 and GM17; BM designs 1241, 1243). Three dies include flowers in the corners and could be an early stage in the use of the dies in this series, rather than separate stamps (if this is so then GM47 would have become GM16, GM48 would have become GM18 and GM49 would have become GM27). Only one of the stamps with flowers in the corners remains in the Sebroke Pavement, GM49, which is used as a 4-tile pattern. Two of the remaining dies were not used on the surviving pavement, GM20 (British Museum design 1519) and GM55 (BM design 1693), and the remainder were used to form a 16-tile pattern in the centre of the surviving pavement (GM18, BM design 1418; GM27, BM design 27; GM19, British Museum design 1551). The rectangular die, GM15a, is not found on tiles in the pavement but is used on a tile from Gloucester Cathedral and now in the Victoria and Albert Museum (Lane, 1939, Pl.21a). A variant die of BM1418 was used on the Monmouth-type tiles at Monmouth and Llangattock-nigh-Usk but has not been found on a Great Malvern fabric tile (Rushforth, 1924; Griffinhoofe, 1894, No.8). A four-tile pattern which might also be used as a corner tile in this series occurs on Monmouth-type tile fabric, Great Malvern fabric and at the Lenton Priory tile kiln (British Museum design 1425).
There are several other designs which have no special connection with Gloucester and may have been in use before 1455. GM24 is a 4-tile tracery pattern (BM design 2689), GM26 is a 4-tile pattern bearing the Arms of De Clare (BM design 1656) and GM59 is a 4-tile pattern bearing the Arms of Beauchamp (BM design 1601).
The Square wall-tile series.
The 5 dies of this series were used on tiles now surrounding the high altar at Great Malvern Priory Church. There is a top tile to the design, GM10, but no bottom tile (GM10-14; BM designs 2577, 1420, 1718, 2576, 2575). With the exception of the top tile all the tiles could as easily be used to form borders.
The Rectangular Wall-tile series.
The tiles of this series are the most famous of the Great Malvern tiles (GM 5-9, BM designs 1321, 1322, 1324, 1323 and 1325; GM83). All except GM83 are represented in the surround of the altar at Great Malvern. GM83, which shows the resurrection, is only found on a single tile, now in the Ashmolean Museum but thought to come from Great Malvern Priory Church. A set in the British Museum is illustrated in colour by Eames (1968, Pl.D). The top tile of the series has an inscription, transcribed "Anno r r h vi xxxvi" (36 Henry VI, or 1457-8).
The Gloucester Lady Chapel Series.
The Gloucester Lady Chapel was constructed between 1460 and 1480 and would probably not have been ready for paving until late in this date braket.
The pavement of Gloucester Cathedral Lady Chapel survives in places at the west end of the chapel, in one of the two side chapels, where it is cut by a 17th century monument and at the east end, where a repair patch can be seen that corresponds to the position of the original altar. There is no doubt that this is the original late 15th century pavement.
The pavement is laid diagonally with 4-tile and 16-tile patterns separated by plain tiles. GM69 is a 4-tile pattern with the black-letter inscription "Ave Maria gr[atias] pl[ena]" and GM78 has a similar pattern with the inscription "Dme jsu miserere" (BM design 1441). Another 4-tile pattern, present in one area near the altar, carried the inscription "Orate pro anima Johannis Hertlond" (GM82). Two 16-tile patterns are found, GM70-73 and GM79-81. The former has an untranscribed inscription and the latter has an inscription "Domine (jesu) miserere".
The sombre nature of these patterns in comparison to the earlier tiles is quite notable, although the dies are as well-cut as the earlier ones. It is likely that they were commissioned for the Lady Chapel, although John Hertlond, who may have been the benefactor, has not been identified.
Black-letter tiles.
Six tiles are linked only by the substantial use of black-letter inscriptions. One tile carries a text which was apparently used as a talisman against fire "Mentem sanctem spontanem honorem deo at patrie liberacionem" (Nichols, 1845, viii, and No.75, BM design 1429). A 4-tile pattern bears an inscription "Pax Christi inter nos sit semper. Amen" (Nichols, 1845, No.74 and pviii; GM56; British Museum design 1469). Another 4-tile pattern bears the inscription "MARC : MATHE : LUCAS : JOH. A : D: MCCCCLVI MISEREMINI : MEI : MISEREMINI : MEI : SALTEM : VOS AMICI. MEI : QUIA : MANUS. DNI. TETEGIT : ME". This tile incorporates the date 1456 (GM63, BM design 1468). A 4-tile pattern, GM64, is inscribed "BENEDICTUS DEUS IN DONIS SUIS" (Blessed is God in his gifts). This tile also depicts two shields showing various implements and what might be antlers (Nichols, 1845, xiv). A fragment from St. Oswald's Priory, Gloucester, has not be transcribed (GM58) and a single tile with an inscription in English is known, the 'Executors tile', GM36, which has eight lines of verse (Nichols, 1845, No.72).
Other tiles.
The remaining patterns cannot be easily divided into sets made at one time and are described below by type of pattern.
Heraldic designs. One tile bears two unidentified coats of arms (GM22, BM design 22) and one tile bears the arms of Newburg and Despencer (GM42, BM design 1723). Both are 4-tile patterns.
Two tiles, bearing the same type of border, have heraldic badges. One depicts the nave of a wheel with the Stafford knot (GM37, BM design 1739) and the other a collared and chained swan (GM38). Both badges also occur on the Thornbury Castle tiles made for the Duke of Stafford in Canynges-type fabric in the early 16th century.
4-tile patterns. One tile has a tracery and foliage design (GM23, BM design 2687) one includes two birds and flowers (GM25, BM design 2688) and a third incorporates a crowned 'M' (GM43, BM design 2686).
Several single-tile patterns are known, mainly using foliage and tracery (GM28-34, 45-6, BM design 2439, -, 2462, -, 2435, 2533, 2543, - and -).
Two tiles are linked only by their comparative crudity of die-cutting. One bears the name "WHILLAR" (GM39, Nichols, 1845, xiv) and the other has a coat of arms, said to be that of the Russell family (GM57).
Dating: Four Great Malvern tiles incorporate a date in their design. It is assumed here that this dates the initial die-cutting, both of that die and those in the same set, although the dies may have continued in use for a long time. The large rectangular tiles therefore date to 1453, the Sebroke tiles date to 1455, an isolated black-letter tile dates to 1456 and the small rectangular wall tiles date to 1457-8. Documentary evidence for the construction of the Lady Chapel at Gloucester Cathedral may be used to give a date of c.1460 to 1480 for the construction of that pavement and possibly for the cutting of the dies used.
Cracked dies can be used to discover the relative dating of various pavements but very few cracked dies were used, nor have duplicate dies of the same pattern been noted. The best example of a cracked die is that used to make GM16 (the uncracked version) and GM17 (the cracked version). The crack was obviously developing whilst the Sebroke pavement tiles were being made, although much worse cracking is visible on some of the tiles at Great Malvern Priory. Since tiles made with the uncracked die are found at both sites it is clear that the two sites were being supplied simultaneously. The uncracked stamp was used on the White Lion, Bristol, collection which includes other Sebroke tiles. The cracked stamp is used at Bath Abbey in a collection which also includes Gloucester Cathedral Lady Chapel tiles. GM27 is uncracked at the White Lion, Bristol; Great Malvern and Gloucester but cracked at Little Malvern Court. GM59 is uncracked at Little Malvern and Gloucester but cracked at Bath Abbey. GM21, a Sebroke pavement style design not found in the surviving pavement, is found uncracked at the White Lion, Bristol, Great Malvern and Stratford-on-Avon (not checked personally). Cracked examples are found at Bath Abbey and Llangattock-nigh-Usk. At the latter site it is possible that the stamps were used on Monmouth-type tiles. The tile fabric has not been examined. Cracks have also been noted on the dies used for GM51 (a Sebroke pavement tile, from a 9-tile set) on a Gloucester example and GM63, the black-letter apostle tile bearing the date 1456, for a tile found at Bayham Abbey. The evidence of cracked dies shows the Bath Abbey collection to be the latest collection known, a conclusion confirmed by the presence of the Gloucester Cathedral Lady Chapel tiles. Some of the Little Malvern tiles have been shown to be earlier than those at Bath Abbey but later than the White Lion, Bristol, Great Malvern and Sebroke tiles.
Both the Bath Abbey and Little Malvern collections include tiles from the Gloucester Cathedral Lady Chapel series and at Bath at least there is no reason to believe that tiles from more than one pavement were present. Therefore dies in use c.1455 were still in use, but showing signs of wear, closer to c.1480.
The evidence of dies and cracks can also used to show that the Monmouth-type tiles were probably made in a short-lived tilery, later than c.1456 but earlier than the Gloucester Cathedral Lady Chapel pavement and the Bath Abbey tiles.
The only other documented movevents of dies are from Great Malvern to Lenton Priory, Nottinghamshire and from Great Malvern to an unknown centre supplying Cirencester Abbey (dies GM17 and GM18). GM18 is also found on a Monmouth-type tile showing that this die, and possible others also, had a long and complicated use.
The Lenton Priory Kiln site produced tiles made with dies used at Gloucester in Abbot Sebroke's pavement, but none of the dies used need have been specifically made for that pavement, so that they could date earlier than 1455. Some of the Lenton tile dies were also used on Monmouth-type tiles and one was used on both Monmouth-type tiles and Malvern chase tiles (GM11, BM1420). Two of the Lenton Priory dies are found on Great Malvern tiles with cracks in 'late' collections (Bath Abbey and Little Malvern Priory). The cracks have not been noted on the Lenton tiles and it is likely therefore that the Lenton Priory Kiln and the Monmouth-type tiles were both the work of Great Malvern tilers in the 1460's to 70's in the lull between the first series of Great Malvern pavements in the 1450's and the construction of the Lady Chapel pavement for Gloucester Cathedral in the 1470's or 1480's.
Four Great Malvern dies to date have been recognised on Malvern Chase tiles; GM11, GM78, GM81 and GM82. These dies were used in the construction of the Gloucester Cathedral Lady Chapel pavement and therefore the movement of dies to Malvern Chase must post-date this pavement. It is likely therefore that the use of Great Malvern tile dies at Malvern Chase occurred at the end of the Great Malvern industry. It may be significant that these Malvern Chase tiles were made with knife-trimmed bases, probably by Malvern Chase tilers rather than Great Malvern ones.
At present the following scheme for the chronology of the products of the Great Malvern 'school' of tilers seems likely:
c.1453. Start of production at Great Malvern, large rectangular wall-tiles.
c.1455. Production of Abbot Sebroke's Pavement for Gloucester Cathedral at Great Malvern.
c.1457-8. Production of small rectangular wall-tiles for Great Malvern.
c.1460-70. Production of Lenton Priory and Monmouth-type tiles.
c.1470-80. Production of Gloucester Cathedral Lady Chapel pavement and pavements at Little Malvern Priory and Bath Abbey.
c.1480+. Movement of dies from Great Malvern to Malvern Chase. Possible movement of tilers from Great Malvern to Canynges-type industry.
Distribution: The distribution of Great Malvern tiles proved by thin-section analysis includes nine sites, three of which are in Gloucester (fig.3.13). With the exception of Bayham Abbey and Raglan Castle all are in the Severn Valley to the south of Great Malvern.
The distribution of tiles stamped with Great Malvern dies is much wider and would repay petrological investigation (fig.3.14).
Fabric: Hard, oxidized with inclusions of sparse rounded and angular clay pellets and rare micaceous sandstone up to 1.0mm across. Sparse subangular quartz grains and sandstone fragments up to 0.3mm also occur. The sandstone contains quartz grains up to 0.1mm across with interstitial brown amorphous material. The clay matrix is isotropic and contains numerous fragments of angular quartz and white mica up to 0.1mm across.
Thin-sections: Monmouth; M995-998. Llanthony Prima; M1124.
Source: The Welsh borderland, probably Gwent or Herefordshire. Petrologically the fabric is part of a group that includes wares made in the Forest of Deerfold in North Herefordshire, the Upper Wye valley, the Usk valley and the Lower Wye valley (including Monmouth itself).
Description: The tiles vary from 24mm to 33mm thick and were probably all c.135mm square. They are decorated with shallow inlay.
Some of the stamps used are identical to those used on Great Malvern tiles (including GM18 and GM63) whilst others might be the same (the tile fragments being either worn or too incomplete to tell). One die however is not known amongst surviving Great Malvern tiles. That die was used to produce an oblong tile with a memorial inscription to Thomas Coke and his wife Alice (BM design 1326, cat 11470). Despite the fact that the approximate date of death is known, c.1450 to c.1480, and that the tile has been published since 1908 no attempts to identify Thomas Coke have been sucessful.
Dating: The Monmouth-type tiles must post-date 1456, since this date is incorporated into GM63, found in Monmouth-type Fabric at Llanthony Prima. Since some of the dies used on Monmouth-type tiles occur in 'late' Great Malvern tile assemblages such as that at Bath Abbey, and since only one die may have been cut in the Welsh Borderland, it is likely that the industry was short-lived, probably during the 1460's to 70's.
Distribution: (fig.3.15) Only two sites have produced Monmouth-type tiles (which can only be distinguished from Great Malvern tiles by fabric analysis); Monmouth (salvage excavation WCC) and Llanthony Prima. It is however quite likely that the tiles from Monmouth Priory relaid on the tower wall at Monmouth Church are of the same fabric (Griffinhoofe, 1894) as well as the rectangular memorial tile now in the British Museum (Eames, 1980, 250).
Many of the other instances of Great Malvern stamped tiles in the area may be from this tilery rather than that at Great Malvern (for example, Llangattock-nigh-Usk (Rushford, 1924), Stretton Sugwas, Hereford All Saints, Croft, and Newland).
Fabric: Hard, grey core with brown surfaces. Abundant angular and rounded quartz up to 0.5mm across, abundant rounded iron ore up to 0.5mm across and rare rounded possible white clay pellets up to 2.0mm across.
Thin-sections: Only one example, not thin-sectioned.
Description: A square tile 126mm by 126mm by 28mm with no bevel and a sanded base. The tile has a shallow inlay design under a clear glaze, which is opaque in places.
Design: is very close to those on Great Malvern tiles (compare with GM28, 31 and 77) but cannot be precisely paralleled.
Dating: Probably mid-15th century or later.
Distribution: (fig.3.16) Only one tile in this fabric is known, from Tintern Abbey and now in the British Museum (cat. no.7655, design 2792)
CANYNGES-TYPE TILES
Fabric: Hard, oxidized red (2.5YR 5/6). Inclusions are sparse in the hand specimen and the overall texture of the clay is smooth. The clay is poorly mixed and lenses of different texture are visible in thin-section. The most common inclusions are subangular and rounded quartz grains, mainly less than 0.5mm across with some up to 1.0mm across. Grains of felspar, light brown chert and quartzite of similar size are sometimes seen. Large rounded and angular clay pellets of different types are common. The most frequently found type is dense, with few inclusions. A less common variant of this type is stained black, possibly by manganese. Rarer types contain a high proportion of carbonate in the form of small crystals, less than 0.02mm across and mostly fired or leached out or a mixture of carbonate, clay minerals and angular quartz grains up to 0.1mm across. Other inclusions are rare and include red micaceous sandstone fragments, chert or quartzite fragments, usually rounded and up to 5.0mm across and opaque black iron ore up to 0.4mm across. The clay matrix is optically isotropic and contains sparse angular quartz grains from 0.02mm upwards, and variable quantities of carbonate and white mica. The quantity of carbonate varies, but without recognisable pattern whereas the quantity of white mica varies regularly. None of the samples of 'stock' late 15th century tiles had a high white mica content but samples of the St. Augustine's Abbey series, the Carew series, the Llanthony series and the Melton series all had a high white mica content.
Thin-sections: Bristol Canynges Pavement; M1005. Cardiff; M1012-3. Carew; M1014-5. Gloucester St. Oswald's Priory; M999, M1048-9. Gloucester Eastgate; M187, M1125-30. Hailes Abbey; M1028-33. Neen Savage; M1002-4. Slebech; M1011. St. David's; M1001, M1008-10.
Source: The characteristics of this fabric can all be matched in the Keuper Marl. A sample of post-medieval brick from a brick kiln at Droitwich (M934) was similar in the presence of carbonate in the clay matrix and in the presence of marl pellets. It contains a lower quantity of rounded quartz and a higher quantity of angular quartz in the matrix than any Canynges-type sample. The Keuper Marl has a wide outcrop. The Great Malvern tilery, for example, used clays derived from the Keuper Marl. Distribution evidence however confines the search for a source to South Worcestershire, whilst the Great Malvern area can be excluded since none of the clays used there are calcareous.
Droitwich is an obvious possibility since a tile kiln is known there and it is thought that a tile-making industry was based on the town from at least the late 14th century until at least the mid-15th century. It is possible that the Canynges-type tiles represent a change in technique and fabric at the Droitwich tilery. This is however unlikely since certain tiles found at Halesowen Abbey and Hailes Abbey and decorated with stamp-on-slip designs are made in the Droitwich-type fabric and are dated by Mrs. Eames to the early 16th century (1980, Late Worcestershire Group).
The other possibility is Worcester itself. Tilers are recorded in the town from the late 13th century. In 1299 tilers were recorded at both Northwich and Whitestones, where one is named as Johannes Tegulator, (Hollings, 1950, 7). There were still tilers active in the town in the 16th century. Dyer, records wills and inventories of part-time tilers in 1560 and 1611 (Dyer, 1973, 132). This was no doubt mainly a roof tile and later also a brick-making industry. The 1467 Ordinances of Worcester deal with the tilers in some detail. No Guild could be formed, every tiler was free to come and go as he liked and no meeting in a union or parliament was allowed under pain of loss of franchise and a fine of 20s. Every tiler was to mark his tiles (Smith, 1870). These rules were reinforced in 1497 (quoted in V.C.H. WORCS II, 275). The meaning of the last rule is unclear, since neither floor nor roof tiles in the area bear recognised makers marks. In 1641 Habington recorded that the clay pits of Losemore, owned by the Bishop, had been used to make brick and tile from the reign of Henry IV. Habington also recorded that a house outside St. Martin's Gate was anciently called the Tilehouse (Habington, ed. Amphlett, 1899, 45). The known products of the Worcester pottery industry; cooking pots, jugs, and ridge tiles, are made in a fabric with no carbonate inclusions and few clay pellets. They do however contain a similar quartz sand, although in much larger quantities. It seems therefore that petrological analysis will not be able to pin-point the Canynges-type industry more precisely and that final proof of the source of the tiles will have to wait for further documentary or archaeological information.
Description: The tiles are made to four sizes; 120mm square, 120mm by 165mm, 150mm square and 160mm square. They are between 23mm and 34mm thick and have sanded bases. They have a slight bevelled edge and a clear glaze, often found on the sides and bases of the tiles. The glaze of the Canynges pavement has a relatively high tin content, rendering it opaque in patches (Eames, 1972 a). Decoration is produced by a shallow inlay (possibly the slip-over-impression method).
120mm square tiles. Plain and white slipped tiles are found, but their incidence, except in the Canynges pavement, is difficult to determine since in most of the relaid pavements only the decorated tiles have been kept.
Single tile patterns can be divided into abstract patterns, those with heraldic significance and shields, set diagonally. Abstract designs include Canynges XII (BM design 2008), XX, XXI (BM design 2490), and XXVII (BM design 2366). Those with heraldic significance include Canynges XXVIII (BM design 1523), which is apparently one half of the Arms of Henry, Prince of Wales 1405-1413 (Porter, 1887). Canynges XXIII (BM design 1737) which is a badge of the Hungerfords (Brakespear, 1894, no.4), a crest in the form of an elephant's head (BM design 1727) and tiles 85 (based on the Arms of Bishop Carpenter of Worcester, Porter, 1887, 158), 87 (based on the Beauchamp Arms), 88 (BM design 1993), 92, 93, and 102 (the last three of which have connections with the Hungerford Family, Brakespear, 1894, nos.3, 7, and 9). Two tiles (nos.00 and 00) are based on the Arms of Evesham Abbey. Shields include the Beauchamp Arms (Canynges XVI, BM design 1606, and another stamp), Wyatt of Tewkesbury (no.81), Craddock (no.89) and Hungerford (Canynges XIV). An unusual set is made up of four stamps (nos. 94 to 97, BM designs 2580, 2581, 1417 and one other) and makes a cross with the sacred monogram at it's centre.
Four-tile patterns can be divided into purely ornamental patterns, heraldic patterns and a series with black letter inscriptions. The ornamental designs include Canynges XIII (BM design 2249), XV (BM design 2697), XVII (BM design 2796), XXII (BM design 2809), XXV (BM design 2011), XXVI (BM design 45), and XXX (BM design 2698, together with a variant of XXX, BM design 2700). The heraldic designs include Canynges X (BM design 1440), which has the inscription "fidelium defunctorum" with two mullets and the letter 'A' in the corner, XVIII (BM design 1716) and XXIV (BM design 1464), which has the inscription "Sir John Talbot" with a talbot. The black-letter tiles have the inscriptions "Domine Jesu misere" (Canynges XI, BM design 1442) and "Ave Maria g. p." (No.86). These two designs are based on the 4-tile patterns used in the Great Malvern tile pavement at Gloucester Cathedral Lady Chapel (GM69 and 78). Canynges XI is an accurate copy whilst no.86 takes the inscription from one tile and the design from the other.
The Canynges pavement contains nine 16-tile patterns (Canygnes I to IX), many of which have variant stamps. One design, Canynges VII, is present in two versions in the Canynges pavement itself. In several cases the centre tiles can be altered or used on their own as 4-tile patterns. One pattern, Canynges IX, is based on the 16-tile pattern used in Abbot Sebroke's pavement which was made at the Great Malvern tilery. Five 16-tile patterns are not represented in the Canynges Pavement, although some are variants of patterns which do occur (BM designs 2980-3, 2991-4 and 2886). (Canynges I = BM designs 2295-8; Canynges II = BM design 2878; Canynges III = BM designs 2919-22; Canynges IV = BM designs 2969-71; Canynges V = 2953-6; Canynges VI = 2973-6; Canynges VII = BM designs 2987-90 and a second series of dies, BM design Nos.2991-4; Canynges VIII = BM designs 2900-3; Canynges IX = BM designs 2893-6).
The St. Augustine's Abbey series. At least fourteen stamps incorporate the initials 'R.E.' or 'I.N.' together with heraldry which links the tiles to St. Augustine's Abbey, Bristol (Eames, 1980, 247-8). 'I.N.' stands for John Nailheart, Abbot from 1481 to 1515 and 'R.E.' stands for Robert Elyot, Hosteller at the Abbey during the Abbacy of Nailheart and later Abbot himself. Eames has suggested that this series dates from the Abbacy of Nailheart rather than Elyot since Elyot acquired a crozier and mitre in his Arms when he became Abbot. There were probably at least four sets of 4-tile patterns, each one composed of four separate stamps. A complete set from Gloucester Cathedral is in the British Museum and has either the initals 'R' or 'E' in the corners. These tiles bear coats of Arms of the Berkeleys, the rebus of John Nailheart, the Arms of Robert Elyot (actually the R.E. monogram with the mullets of St. John) and the mullets of St. John (BM designs 1499-1502). Two similar sets must have existed and are represented by one design each. They have the initials 'I' or 'N' in the corners but similar shields to the Gloucester Cathedral set. It is evident that the designs come from different sets because the black-letter inscriptions are from different sentences (Nos. 74 and 102). The fourth set has the monogram R.E. in the centre and the initials 'R' or 'E' in each corner with a black-letter inscription "In te d[omi]ne s/peravi non / confundar / in eternum". One tile of this series is found in the Canynges pavement (Canynges XXIX, BM design 1467). Two single tile patterns have the monograms of Robert Elyot or John Nailheart in the centre and a black-letter inscription around the edge (nos. 78, BM design 1427, and 103, Nichols, 1845, xiv). A final tile is possibly from this series since it has the initial 'R' in the corner. It is probably part of a commemorative tile set, similar to the Lygon tiles described below (no.82, BM design 1481).
The Lygon tiles. These tiles are closer to 130mm square than 120mm square. Two tiles in Great Malvern Church go to make up a 4-tile pattern with the inscriptions "Orate pro aia Th/ome Lygon mil" (Pray for the soul of Thomas Lygon, Knight, BM design 1463).
A Thomas Lygon of Madresfield Court died in 1507 and is very probably the person commemorated by these tiles. Madresfield is very close to Great Malvern and is quite likely to be the burial place of the Lygon family.
The Carew tiles. Three rectangular tiles, 120mm by 165mm, have the coats of Arms of Pembrokeshire dignitaries (Eames, 1980, 249). These include Sir Thomas Lloyd, Precentor of St. David's between 1534 and 1547, and Sir Rhys Ap Thomas, who was prominent at Carew Castle between 1485 and 1528 (BM design 1547).
The Llanthony Priory tiles. At least forty-one stamps are used on a series of 150mm square tiles found at various sites in Gloucester but probably made for the Priory of Llanthony, just outside the town to the south. Three 16-tile patterns are made with 'stock' 120mm tile stamps to which 30mm wide strips have been added. These utilise Canynges I (BM designs 2010 and 2999) and variant stamps of Canynges III (BM designs 2924-6) and the less common version of Canynges VII (ie. BM designs 2991-4). The remaining stamps were probably made at one time for this series. Two 16-tile patterns are known, with outer tiles only which suggests that they might have formed a border around any of the 4-tile patterns (BM designs 2882-3 and 2884-5). Five 4-tile patterns are known, all but one of which were made using four separate dies (BM designs 1706, 1704-5 and 1496-7). Four single-tile patterns are known, all shields (BM designs 1561, 1559, 1553-4). Two of these shields show the same heraldry but one is a red design against a white slip background and the other is a white slip design against a red background (BM designs 1553-4).
The series is dated by its heraldry to the early 16th century and includes the Arms of Prior Forrest (1501 onwards) and Archbishop Henry Dean of Canterbury, 1501-3, who was Prior at Llanthony until 1494 (Eames, 1980, 255-7).
The Melton tiles. The Melton tiles, now mostly in the British Museum, were found at Hailes Abbey and, relaid, at Southam De La Bere (Eames, 1980, 257-263). They were made for Hailes Abbey during the Abbacy of Abbot Melton, between 1509 and 1527. The tiles are 160mm square and include single, 4-tile and 16-tile patterns.
The Thornbury series. The Thornbury tiles were made of the Duke of Stafford between 1510 and 1520 for his castle at Thornbury. The tiles are larger than any other Canynges-type tiles and the designs used are not found on any other group of tiles. However, two of the 4-tile patterns are stock designs with no connection with the Duke of Stafford, one of these is a fleur-de-lys within a quarter-circle (BM design 2251) and the other is the same design as BM 2249, one of the Canynges pavement designs.
The remaining designs include a 4-tile shield within a circular band bearing the motto "HONYSOIT * / QUI * MAL * Y / PENSE *" (BM designs 1482-5), three single-tile shields and three single-tile badges, showing the antelope (BM design 1726), the chained swan and the nave of a wheel with the Stafford knot (Eames, 1980, 257).
Dating: The tiles at Heytesbury (Wiltshire, Brakespear, 1894) may be part of a commissioned pavement for the Hungerfords (whose seat was at Heytesbury). This series should therefore date from 1421 or later, since it incorporates a badge only adopted by Lord Walter Hungerford in that year (Eames, 1980, 242). Since Lord Walter Hungerford died in 1449 it is suggested by Eames that the Heytesbury tiles should date from between 1421 and 1449.
Canynges IX must be later than c.1455 since it is based on a design from Abbot Sebroke's pavement (see above, Great Malvern tiles). Two designs, Canynges XI and no.86, copy Gloucester Cathedral Lady Chapel tiles and must date from the 1470's at the earliest and probably from the 1480's (see above, Great Malvern tiles). Eames suggests that the Canynges pavement dates to c.1481 or later, since it incorporates one of the St. Augustine's Abbey series tiles, which must date from the Abbacy of John Nailheart, 1481 to 1515 (Eames, 1980, 247).
Only two pavements have survived in their original position long enough to be recorded; the Canynges Pavement from a house in Bristol known as Canynges House (This pavement is now in the British Museum, Eames, 1951, 1980) and the St. David's Abbey pavement at St. David's Cathedral, Dyfed (Fryer, 1903). Both are useful in giving 'assemblages' of stamps known to have been used at one time. It is also fairly certain that the commissioned series tiles would all be produced at one time. Certain other collections of Canynges-type tiles have the appearance of being cohesive sets, for example those at Acton Court, Bath Abbey, Ledbury and Rudford Church (Glos.). Other collections may contain tiles from earlier pavements or later replacements and the dating of one tile in such a collection cannot be used to date the others.
The Canynges pavement contains stamps used on the Heytesbury tiles as well as tiles from the St. Augustine's Abbey series (Canynges XXIX) which must date after 1481. Other tiles in the pavement must date to c.1455 or later (Canynges IX) and to the 1470's or 80's (Canynges XI). On two counts, therefore, it is likely that the pavement was laid down in the 1480's at the earliest.
The St. David's pavement, although perhaps repaired and relaid in places is essentially in its original position and is laid up to the Throne of Bishop John Morgan, 1496-1504 (Fryer, 1903, 177-8). Of the 43 stamps found in the St. David's pavement 27 are also found at Canynges pavement (including Canynges XI but not Canynges XXIX). Some of the same patterns are present with variant stamps and two of the variant stamps are the same as those used in the Llanthony series post-1501 to make the enlarged patterns. This evidence shows that the St. David's Pavement is later than 1496 and that it is later than the Canynges pavement. Therefore where other variant stamps are found it is probably the St. David's rather than the Canynges versions which are later. Thus we find that Canynges V and VIII are replaced by variants. Wherever the Canynges versions are found tiles from the St. Augustine's series are also found, confirming that the St. Augustine's series tiles are contemporary and earlier than the Canynges pavement, rather than Canynges XXIX being a replacement tile.
Only two stamps with cracks have been seen, and in neither case is a variant stamp known. These are Canynges I (uncracked at Canynges and in the Llanthony series but cracked at Strencham) and Canynges XX (BM design 2166, cracked at Strencham and Bath Abbey). The evidence of the Canynges I stamp also shows that the stamps were re-used on 120mm square tiles after being used on the larger Llanthony series tiles, so that the St. David's pavement is not necessarily earlier than the Llanthony series.
The evidence accumulated to date suggests that none of the collections of Canynges-type tiles found can date before the 1470's and that the Canynges pavement dates to c.1481 or later, whilst the St. David's pavement is later and dates to c.1496 or later. The commissioned tiles are dated as follows:
Hungerford tiles- c.1421 to 1449
St. Augustine's series- 1481 to 1515 (but probably earlier in this bracket than later)
The Lygon tiles- c.1507
The Carew Tiles- c.1485 to 1528
The Llanthony series- c.1501 to 1539
The Melton tiles- c.1509 to 1527
The Thornbury tiles- c.1510 to 1521.
On this basis, it would be possible, if not for the Hungerford tiles, to erect a short chronology for the industry, from c.1480 to c.1510, encompassing all of the available evidence. It is worth considering therefore whether the Hungerford tiles might not have been made for a descendant of Lord Walter Hungerford, for example Sir Walter Hungerford who was active in Wiltshire from the 1470's until his death in 1516 (D.N.B., 1908, X, 257).
Similarly a long chronology might be erected. The industry may have preceded that at Great Malvern and have been contemporary for a time with that postulated at Droitwich. There is no reason why the industry should have disappeared at the dissolution and pavements such as that at Strencham Church might therefore date to the mid-16th century if, for example, the Llanthony series was actually constructed immediately before the dissolution of the Monasteries. A long chronology would also enable tilers from this industry to use their dies for the Lacock Abbey tiles in the 1550's.
Distribution: The distribution of Canynges-type tiles may be considered series by series. The 'stock' patterns are found over the widest area, with a cluster of sites in South Worcestershire and a scatter of sites in the Severn Valley with two strings of sites leading off from there along the South Welsh coast and along the Bristol Avon into Wiltshire (fig.3.17). The south Welsh finds actually form two clusters, one around Cardiff and the other in Dyfed. This sort of pattern certainly suggests that the tiles were transported by water, while the clustering of the outlying finds is probably due to the tilers gaining orders whilst laying an initial floor.
The distribution of St. Augustine's Abbey series tiles is very similar to that of the 'stock' patterns, including Carew in Pembrokeshire (fig.3.18). The tiles probably entered the 'stock' of designs available to any buyer because they are of the same 120mm square size.
The remaining commissioned series were of a different size. The Carew tiles are found at Carew and Bath Abbey, the Lygon tiles only at Great Malvern (an an unprovenanced find in the British Museum) and the Melton tiles only at Hailes Abbey. A similar sized tile with decoration in the same style but not paralleled at Hailes was found at Gloucester Blackfriars.
The Llanthony series, including the enlarged 'stock' patterns, is found at several sites in Gloucester (fig.3.19) with an outlier at Caerleon (an unprovenanced Museum find), while the Thornbury series tiles are found at a number of sites in the Lower Severn valley with an outlier in Gloucester (fig.3.20). It might be argued that both series represent the dispersal of tiles after the original buyer has for some reason been unable to complete the purchase, in one case because the Priory of Llanthony was dissolved before the general dissolution (I am indepted to L. J. Keen for this suggestion) and in the other because the Duke of Buckingham was beheaded in 1521 and his Castle forfeited to the Crown (Jeffcoat, 1932). It is also possible however that both series represent the same pattern of supply as the Pembrokeshire sites - the tilers arrived in the area with more than enough tiles for the job and sold off the excess to neighbouring establishments. The interpretation is important not only for the light it sheds on the organisation of the industry but also because if the first hypothesis is correct then the tiles must be dated to the end of their date brackets, c.1521 and c.1539, but if the second hypothesis is correct they could be 10 to 30 years older.
Fabric: Not described, fine sandy usually with a light grey core.
Thin-sections: None
Source: Unknown. Fragments with the surface in perfect condition and a few possible wasters were found at St. George's Church, Fordington, in 1907 (Eames, 1980, 264-6). It is, however. quite common to find tile waste at a site in small quantities since it could be present in the batch sent to the site.
Description: Shallow inlaid designs, with inlay c.1mm thick, and a stabbed base. The designs include a Great Malvern style crowned 'M' (BM design 1422), a 4-tile pattern with parallels in the Canynges-type tiles (BM design 1465) and a fragment of a tile which appears to copy the large tiles found at Ewenny Priory and possibly made at Malvern Chase (BM design 1479). The remaining tiles are derived from late 13th to 14th century types.
Dating: The use of a 'Malvern School' design must date this group to the mid-15th century or later. The presence of designs which appear to copy tiles of the Canynges-type and the large Malvern Chase groups suggests an even later date, probably in the early 16th century (Eames, 1980, 264).
Fabric: Soft to very hard, oxidized light brown to red (7.5YR 6/4 to 2.5YR 4/6). Inclusions of fine-grained sandstone up to 6.0mm across in a micaceous clay matrix. Identical sandstone fragments form the sand on the bases of the tiles and the tempering of the white mortar clinging onto some of the specimens.
Thin-sections: None. Examined in detail under the binocular microscope.
Source: The same fabric was used to make Tudor bricks at Wigmore and this, together with the petrological similarity of the sand used in the mortar indicates a very local origin.
Description: Tiles. The tiles average 130mm square and 24mm thick. Most are plain glazed or covered with br5shed white slip. Five counter-relief tiles were found, of which three were plain lead glazed and two had a white slip brushed-on before stamping (including BM design 391). A few small scraps of inlaid tile were found but these were not necessarily made in the same fabric. The counter-relief tiles were of two designs. One was a 4-tile pattern of intertwining foliage and the other possibly the centre of a 16-tile pattern.
Bricks. The bricks were made in a sanded mould with a trimmed upper surface. Two examples were accidentally glazed. Thicknesses of 50mm and a width of 120mm were recorded.
Dating: None of the brick fragments were stratified but the similarity in fabric and the presence of accidental glaze suggests a similar date to the tiles. Since the tiles were stratified below the 16th century destruction rubble they must be of this date or earlier, definitely pre-1539. Similar counter-relief tiles in Staffordshire have been dated by Eames to the 15th century (Eames, 1968, 31 and Pl.1 nos. 6 & 7).
Distribution: (fig.3.21) Found at Wigmore Abbey and Wigmore Grange (BM 11,158). No similar tiles have been found in the study region but the counties of Shropshire and Staffordshire, to which Wigmore is very close, have not been searched for similar fabrics or stamps.
Fabric: Not described, fine sandy
Source: Unknown
Description: Shallow inlaid tiles with sharply bevelled sides. Shell-like keys on the knife-trimmed base. Designs found include several renaissance motifs and the monogram of the post-dissolution owner of Lacock Abbey, William Sharington and his wife, Grace. Other tiles found at Lacock Abbey are apparently of the same series but use designs derived from the Great Malvern and Canynges-type series (Eames, 1980, 266). The Lacock tiles have not been examined by the author but it seems as if these late tiles use actual Great Malvern and Canynges-type stamps. If so then this could account for the occurrence of tiles, apparently of these types being found in Dorset, since a Lacock-type tile has also been found at one of the find spots, Glanville's Wooton.
Dating: Mid-16th century. William Sharrington married Grace in 1550 but died in 1553. This probably limits the period of manufacture of these tiles.
Distribution: (fig.3.22). Lacock-type tiles have been found at Lacock Abbey itself, Bath Abbey and sites in Dorset.