Archaeological excavation adjacent to the Tanners' Hall, Gloucester, in 1997 and 1998, JO VALLENDER

Tags: Tanners, Gloucester, excavation, assemblages, Worcester Street, tanning industry, 17th-century, Heighway, 17th century, 14th-century, recorded, 14th century, assemblage, Tanners' Hall, 11th century, NISP, JO VALLENDER, Hare Lane, N Tanners Hall, Anglo-Saxon, Tanners Hall, excavation site, Roman road surface, earlier medieval period, animal husbandry, 17th centuries, medieval pottery, middle Saxon, archaeological deposits, archaeological excavation, post-medieval
Content: From the Transactions of the Bristol and Gloucestershire Archaeological Society Archaeological excavation adjacent to the Tanners' Hall, Gloucester, in 1997 and 1998 by Jo Vallender 2009, Vol. 127, 133-193 © The Society and the Author(s)
Trans. Bristol & Gloucestershire Archaeological Society 127 (2009), 131­193 Archaeological excavation adjacent to the Tanners' Hall, Gloucester, in 1997 and 1998 By JO VALLENDER With contributions by Richard Bryant, Julie Jones, Lynne Keys, Fiona Roe, Naomi Sykes, Jane Timby and Heather Tinsley INTRODUCTION Between November 1997 and March 1998 an archaeological excavation was undertaken during Stage 1B of the construction of Gloucester's inner relief road. The site (OS Nat. Grid SO 83361891) is located c.500 m north-west of Gloucester city centre, bounded to the north by a car repair business and to the south by a retail furniture shop. The eastern boundary was formed by Worcester Street and the western by Hare Lane (Fig. 1). The site is located on the Third Main Terrace alluvial deposits of the river Severn (BGS 1972) at an average height of 12.75 m above OD. Prior to the commencement of archaeological work the site was occupied by garage structures dating to the beginning of the 20th century and two terraced houses fronting Worcester Street dating to the 18th century. In the centre of those buildings, and almost completely obscured by them, were the remains of the 13th-century building known as Tanners' Hall (SM 28814; LB Grade II). The construction of the new section of the inner relief road was required to join existing sections in Gouda Way and Black Dog Way. During several seasons between 1978 and 1980 the interior of Tanners' Hall was excavated and the surrounding area identified as containing archaeological deposits dating to the Roman, medieval and post-medieval periods (Heighway 1983b). In 1994 a desk-based assessment of the archaeological implications of the road scheme was undertaken by the Gloucestershire County Council Archaeology Service (Hoyle 1994). It highlighted areas of archaeological potential and made recommendations for further archaeological evaluation. The field evaluation of the road corridor took place in 1995 (Vallender 1995) and consisted of three trenches, numbers 1, 2 and 4 (Fig. 2). Deposits of Roman, medieval and post-medieval date were found over the entire site, except where they had been removed by post-medieval truncation and 20th-century intrusions, such as building footings and inspection pits. The programme of archaeological investigation, as defined by the brief issued by the County Archaeologist, covered the drilling of boreholes for engineering ground investigations; the demolition of the existing structures on the site (Vallender and Jamfrey 1998); the excavation of the road corridor; and the archaeological monitoring of the realignment of the adjacent existing carriageways and drainage and of the insertion of street furniture. Prior to road building the engineers sank two boreholes adjacent to Hare Lane. These were archaeologically monitored and the cores examined. A layer of demolition debris was recorded beneath a concrete floor, and beneath this in both cores was a deposit of black humic soil 1.0­1.2
132
JO VALLENDER
N
N
ever
n River S
Cheltenham Gloucester GLOUCESTERSHIRE
10/89
5/89 34/79
37/93
see Fig 2
3/83
78/73
Cathedral
Gloucester
Approximate line of the medieval River Twyver
Based on Ordnance Survey GIS mapping with the permission of the Controller of Her Majesty's Stationery Office (Crown Copyright). Unauthorised reproduction infringes Crown Copyright and may lead to prosecution or civil proceedings. Glos CC LA 076627
Fig. 1. Site location plan (scale 1:5,000).
m thick. A large tanning pit recorded in one of the boreholes appeared to cut the natural sand; the other borehole contained a series of undated layers including a potential floor surface. A photographic survey of the external and internal elevations of the structures to be demolished was made (Vallender and Jamfrey 1998). The buildings recorded comprised the remains of garages and derelict houses. Following this, the demolition of the garage structures surrounding and attached to Tanners' Hall was monitored to ensure that the remains of the hall were not damaged during the process. The stability of the hall was uncertain and therefore buttresses from the adjoining walls were left in place. After the demolitions Tanners' Hall was enclosed within a fenced compound and supported with scaffolding to stabilise the structure and facilitate the detailed recording of the building. The elevations of the hall and its ground plan were drawn (Pyper 1999). The main excavation covered c.655 m2 (Fig. 2), an area smaller than the road corridor to allow for baulks against the adjacent structures. It was excavated by machine to the top of the first archaeological horizon, in spits of no more than 0.50 m, using a wide-bladed ditching bucket without teeth. The site was then hand-cleaned, a pre-excavation plan was prepared and a record of the visible contexts made. Subsequent investigation of archaeological levels was by hand, with cleaning, examination and recording both in plan and section and excavation to the depth required for the construction of the road. The road design included a fall from south to north and from east to west to compensate for the differing levels of Worcester Street and Hare Lane. This fall was reflected in the levels to which the archaeological excavation was undertaken, extending to between 0.96 m and 1.30 m below ground level. Approximately half (predominately the centre) of the site needed no archaeological excavation following the removal of the overburden and was
Park Street Hare Lane Worcester Street
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 133
Watching brief
N
Garage premises
A
B
C
Main excavation area
Trench 4
Tanners Hall
Trench 1
Trench 2
Based on Ordnance Survey GIS mapping with the permission of the Controller of Her Majesty's Stationery Office (Crown Copyright). Unauthorised reproduction infringes Crown Copyright and may lead to prosecution or civil proceedings. Glos CC LA 076627 Fig. 2. Trench location plan (scale 1:500).
134
JO VALLENDER
merely recorded at that level. The main reduction was at the west end and in the east only minimal excavation was required with the majority of the deposits being left in situ. The archaeologically excavated depths required to produce the correct levels varied between 0.02 and 0.30 m below the base of the machine excavation. Subsequent to the completion of the main excavation a watching brief was undertaken where areas in Gouda Way, Hare Lane, Park Street and Worcester Street were stripped of their existing road surfaces to allow their realignment and as new drains and traffic signals were installed. Three supports were constructed to retain the newly exposed wall on the northern boundary of the road.
THE EXCAVATION
Roman The main feature attributed to the Roman period was a road (779) located in the north-eastern corner of the site (Figs. 3 and 4). It was oriented NW­SE, extending westwards from the eastern section for 9 m with an exposed width of c.5 m and a cambered surface constructed of tightly packed small pebbles averaging 20 mm in diameter. As the road remained unexcavated its original dimensions were not revealed, but a partial section (Fig. 3) was recorded where it had been cut through by petrol storage tanks. Of the construction layers exposed, several bedding layers were recorded although the sequence was unclear. The earliest layer appeared to be of soil and rubble (766). Overlying it were layers of crushed limestone (777/706) and a distinctive red compacted gravel (782) interspersed with spreads of industrial waste (slag) (705), which was mixed with a yellow clayey material (770/787). Over the bedding material one or several pebble surfaces (778/779) had been laid. Overlying (779) were areas of a much rougher limestone surface (769/687) which were probably attempts to patch surface (779). The highest surviving point of the road was recorded at 11.58 m above OD. Although no dating evidence was recovered from the road surface itself, pottery from the bedding layers suggested a date of the 1st or 2nd century AD. A deposit (700) which contained industrial and household waste dating to the early to mid 3rd century had built up over the road. It was sealed by a surface (720) less compacted than the road and containing a considerable amount of reused material such as tile, burnt limestone and marl; it extended westwards and possibly southwards (725 and 726). The evidence for the relationship between the road and the overlying deposits and Roman features to the south and west, which are described below, had been removed by the later cutting of tanning pits. To the south of the road a truncated clay floor (666 and possibly 671) measuring approximately 10 Ч 6 m contained pottery of 2nd-century date. Cut into the floor was a series of shallow features [639, 640, 653, 655, 658, 659, 669 and 670], of which [653] appeared to be the base of a hearth and [669] an area of burning. It is not clear if the individual features were contemporary with one another or if they related to the floor surface through which they cut; the majority had been backfilled with slag and the few sherds of pottery recovered from the fills suggested a date of the 2nd century.
N
778
Fig. 3. Section of Roman road (scale 1:20).
779 770 766
S 11.55mOD
100/120
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 135 Worcester Street
725 830
782
770
779
782 779
Fig 3
779 770 782 779
205
209
200
203
164
726
Trench 4
725
830
609
720
700
259 671
697
694 691
688
692
689
685
671 655 651 660 671 658 659
683 640
670
653
669
680
671?
657
609
632 628 631
636
639 666
647 637
649 618
N Tanners Hall Fig. 4. Roman features.
Limit of excavation
Modern intrusions
Structures
Areas of burning
01
2
3
4
5m
525
520
519
100/100
136
JO VALLENDER
To the west of the floor was a possible structure measuring c.4 Ч 4 m (683). Although this was not excavated linear arrangements of masonry seen though the overlying deposits (685, 688, 689, 691, 692, 694 and 697) showed signs of having been burnt. The overlying deposits, which produced artefacts dating to the 3rd to 4th century, largely comprised ash and charcoal. A small entrance, stoke or drain hole, 0.2 m wide, located on the eastern side, appeared to link with a linear rubblefilled feature aligned north­south (657) which may have been a drain. It is thought this group of features and deposits may represent a kiln or bloomery. To the south of the clay floor (666) and possible kiln/bloomery was a partially revealed beam slot and posthole [649] adjacent to a tile-lined kiln (637). The posthole and beam slot were interpreted as the western corner of a structure aligned NW­SE and extending under Worcester Street; pottery of 2nd-century date was recovered from them. To the west of this building were the truncated remains of an unexcavated small kiln surviving to a height of 11.75 m above OD. The kiln appeared to have been built into an opus signinum floor. Its central part had been constructed from tile occupying an area of 0.3 Ч 0.4 m and encased in clay. Spreads of ash appeared to be associated with the kiln and the surrounding deposits also appeared to have been affected by burning. The relationship of the beam-slot structure and the kiln to the clay floor (666) was truncated to below the level of excavation to the north by a post-medieval culvert and to the south by a post-medieval wall and it therefore remained unresolved. To the south of this wall in the south-east corner of the site a Roman ditch [520], 1.5 m wide and aligned NW­SE, survived to a depth of 11.8 m above OD. Two postholes [525] and [526] also thought to be Roman in date had been cut through its fills. A surface or a series of surfaces (720, 725, 726) was recorded to the west of the main Roman features. These were not excavated for the most part, or fully exposed, but they appeared to continue beneath the post-Roman build-up (830). Undated in the main excavation these surfaces were sampled in evaluation trench 4 as (164 and 200) and produced pottery of 2nd-century date, possibly redeposited. At 11.58 m above OD they were thought to be the highest of a series of earlier surfaces. Spreads of demolition debris or collapse observed overlying the surface in several areas were also thought to date to the Roman period. Evaluation trench 4 lay centrally within the main excavation area and the stratigraphic sequence within it provides further information on the Roman deposits in this area (Fig. 10). At the base of the sequence was a large ditch [259], aligned NE­SW, and a possible bank (260) located on the eastern side. The ditch was c.3.4 m wide and c.1.2 m deep and the pottery from its primary fill suggested a date of the late 1st or 2nd century for its backfilling. The possible remains of the bank ran parallel to the ditch at a height of 10.79 m above OD; collapse, recuts and truncation had obscured the relationships. Cutting the backfilled ditch were three beam slots [203, 205, 209]. The dating from [203] spanned the 2nd to the early 4th century. The slots did not appear to be associated with any other excavated features or surfaces. They were probably heavily truncated and sealed by a surface, which produced residual dating of the 2nd century. Above this was a series of surfaces, the latest of which (200) continued into the main excavation as (725). Evaluation trenches 1 and 2 produced evidence of the continuation of Roman deposits to the south of the main excavation. The earliest Roman activity in trench 1 was pit [246] cut into the natural deposits and filled with cess, dated to the late 1st or 2nd century. Overlying it was a building of beam-slot construction [234] with a sand floor (232) and clay walls (231) thought to date to the 2nd century. In trench 2 was a cambered surface, possibly a road still in use during the 3rd century. It appeared to be aligned north­south, and gravel and sand (211) provided a base for a fine pebble surface (198). A ditch may have been present to the west of the possible road but it was poorly defined. Several new surfaces (189, 173 and 172) were laid on top of (198) and the camber maintained. Surface
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 137 (189) was primarily constructed of slag, perhaps derived locally. The area of the road was eventually levelled with waste and a surface (148) dating to the late 4th ­6th century extended across the whole trench at a height of 11.2 m above OD. Early Medieval (11th­13th century) A greenish sand deposit (830), varying somewhat in colour and texture, covered the majority of the site, overlying the Roman activity. It was cut by features dating from the medieval period onwards (Figs. 5­8). The ceramic dating evidence recovered from this deposit was ambiguous, comprising material from the 2nd to the 18th century as a result of intense medieval and postmedieval intrusion and truncation. Only a small amount of the deposit was excavated, but the stratigraphy suggests that it was formed during a period of disuse or levelling between the Roman and early medieval periods. A stratified series of features cutting (830) against the southern section, immediately to the north of Tanners' Hall, was partially excavated to a depth of c.0.2­0.3 m (Fig. 5). The earliest cut in the sequence [1002] was poorly defined but was thought to be a pit. Cutting it was a large, partially exposed, linear feature [1146], which was aligned NE­SW and was over 7.0 m long and 3.5 m wide. The depth of the feature is unknown as only the upper fill, comprising a huge dump of lias clay which contained pottery dating to the late 11th or early 12th century, was excavated. Deposits of ash and charcoal were visible beneath the base of the clay. It would appear that the clay had been deposited in an attempt to cap or consolidate the feature, which is thought to have been a ditch. Cutting through the clay cap at the feature's eastern end was a large pit [993] containing several fills. The earliest excavated fill, (978), contained pottery of 11th­12th-century date. Overlying it were the possible remains of a demolished structure (975) of the same date. The remains were heavily burnt and comprised more than 75 per cent daub. The presence of unburnt stone suggests that the material had not been burnt in situ. The daub was sealed by a mixed deposit of tanning waste (970) dating to the 12th­14th century. This top fill was cut by two possible tanning pits of medieval date, [900] and [940]. At the west end of the excavation area, cutting linear feature [1146], a charcoal-filled pit [1073] with four stakeholes in the base contained pottery of 11th­12th-century date. The pit also cut a deposit dating to the 12th­14th century suggesting the pottery recovered from the fill was residual. Fourteen possible tanning pits [865, 900, 907, 940, 957, 969, 981, 1003, 1054, 1067, 1142, 1145, 1148 and 1156], two linear features [1132 and 1138] and c.12 postholes (Fig. 5) contained fills with ceramic evidence suggesting that they might predate the 13th century when it is believed that Tanners' Hall was constructed. Clustered at the Hare Lane end of the site they possibly coincided with the commencement of tanning. The remains of what was thought to be a clay floor (1091) were recorded, running under the northern section. The floor had been heavily truncated and was cut by a large linear feature [1128] aligned north­south and filled with deposit (1076). This deposit was backfill or levelling of an area of robbed-out features, which remained undefined since the required depth of excavation had been reached. The floor was undated but finds from the backfill of the robber cut suggested a date of the 13th­14th century for the possible removal of a structure, perhaps suggesting a rebuilding phase during that period. An isolated feature interpreted as a hearth [1117] was recorded adjacent to the southern section at the west end of the site. These features would appear to represent buildings adjacent to the Hare Lane frontage but there was no evidence to suggest whether the structures were domestic or industrial. In evaluation trench 1 several layers of demolition debris dating to the 11th­12th century were present above the demolished or collapsed Roman building. The earliest layer, (227), was very clean
138
55/120
JO VALLENDER
1148 1003
981 957
1145 1156
1018
1132
1158
1138
1142
1054
Hare Lane
1128
55/110
878 1067
1007
1040
1117 1146
1063
1048 1035
1146
1002
993
969
940
900
1036
Limit of excavation
Modern intrusions
N
Structures
Areas of burning xxx* Clay deposits
0
1
2
3
4
5m
Fig. 5. Early medieval features: 11th­13th century.
907 865* Tanners Hall
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 139
50/120
70/120
953 963
9511010
947 972 968
922 957 995 921 920
958 1047
1050 1019 893
1031 1153
1142 1151 1142
1156 1077 890 1062
1123 1125
1000
1079 870 868
1060
Hare Lane
850
50/110
1109
1089
1137
1094 1073
1134
1071 1115 Possible structure alignment
1016
863 858
1006 1057
1085 913 1083
1042
1041
936 934
903 985 908 930
Limit of excavation N Modern intrusions Structures
Areas of burning xxx* Clay deposits
0 1 2 3 4 5m
898
980
861
885
929 876
889
855 831 Tanners Hall
Fig. 6. Medieval features, western end of excavation.
sandy clay, similar in nature to (830), and was probably part of the post-Roman dumping which appeared to have occurred in the area. Six possible layers overlying (227) were clearly redeposited building and domestic waste, perhaps imported to build up the area with a view to cultivation. In trench 2 the uppermost Roman surface (148) was cut by posthole [153], the dating of which is unclear. Filling the posthole and overlying the surface was a homogenous greenish sand deposit (107) dating to the 11th­13th century and similar in nature to (830). Cut through this deposit were two pits, [118 and 142], and a ditch, [129], of the same date. Above these features was a considerable amount of dumped soils, mostly sandy in nature. The date range for these deposits was the 10th­14th century. Truncating this material were a number of cuts and features relating to cultivation.
Late Medieval (14th­16th century) In view of the proximity of Tanners' Hall, it was anticipated that evidence of medieval tanning would be present within the area of the excavation. The evaluation had indicated that the medieval deposits had been extensively truncated and this was confirmed when machining took place at the start of the excavation. The medieval deposits consisted mainly of a range of cut features of which tanning pits were the most obvious. A number of postholes and several linear features were mainly
140
80/120 811
JO VALLENDER 759
100/120
822 819 949
812 181 Trench 4
698 674 642
Worcester Street
733 749*
Limit of excavation Modern intrusions Structures Areas of burning xxx* Clay deposits
608 600 N 0 1 2 3 4 5m
Fig. 7. Medieval features, eastern end of excavation.
634 633* 677 676 630* 503
100/100
located at the western end of the site, but the intensity of post-medieval activity adjacent to Hare Lane resulted in a confused sequence, difficult to interpret. At the eastern end of the site groups of intercutting pits were identified but the presence of overburden prevented individual pits being isolated. In the centre of the site a large area of post-medieval overburden obscured any underlying features. Approximately 50 individual tanning pits of varying dimensions were identified across the site (Figs. 5­8). Consideration of the site plans produced no evidence of groupings of specific types of pits for differing uses. The lack of conclusive dating evidence for the individual features gives the impression that the tannery yard had developed on an ad hoc basis rather than being formally planned. Six of the tanning pits, [630, 633, 749, 865, 953 and 1030], contained pink clay within their fills. These pits spanned the 11th to the 18th century and it is not at all certain whether the clay had been redeposited or represented in-situ linings. The only tanning pit [181] that was fully excavated was within evaluation trench 1 and was clay lined. Although partially obscured by the trench section, it measured 2.2 m in diameter and was 1.4 m deep. The pit had been lined with wood
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 141 and later clay which had been used as a sealant to prevent the escape of the tanning liquors, the pit having been cut into the permeable natural sand. Its primary fill dated to the late 14th­15th century and it was backfilled in the late 15th­early 16th century. The imprint of woven wicker, probably from a basket, was recorded in the surface of the clay. The tanning pits, about forty-three of them depicted (in some cases schematically) on Figures 5­8, are grouped on the basis of the date of their upper fills as the majority were unexcavated. An area containing intercutting pits [673, 698, 793, 794 and 811] was recorded in the north-eastern corner of the site and may indicate an increase in pit cutting activity towards the line of the river Twyver. Seventy-four postholes dating between the early medieval period and the post-medieval periods were excavated or recorded. Concentrated adjacent to Hare Lane and Tanners' Hall, their dimensions, spatial position and finds information were assessed to determine whether any structures could be identified. Postholes [893], [1000], [1071] and possibly [1109] in conjunction with [1142] may have formed a structure dating to the medieval period. However, their alignment does not relate to Hare Lane or Tanners' Hall, both of which were in existence by that time. The postholes were the largest recorded on site, having an average diameter of c.1.00 m and an average depth of 0.20 m. A suggested alignment is shown as a dotted line on Fig. 6. No floors or internal features had survived. Posthole [893] was found to contain, as packing, a fragment of a 9th-century Anglo-Saxon cross shaft (Fig. 13), although the surrounding fill dated to between the 12th and the 14th century. A number of linear features dating to the medieval period were recorded, again concentrated in the areas adjacent to Hare Lane and Tanners' Hall. Two alignments were apparent, one NE­SW [951, 1041, 1134, 1137], the other approximately north­south [850, 870, 889, 903, 922, 972, 1077]. It has not been possible to determine which alignment is the earlier as the dating from both ranged between the 11th and the 17th Century and there were no stratigraphic relationships between the two groups. The function of these features is unclear; they may have been tanning pits for a specific task, drainage features or footings for tanning apparatus. The NE­SW group may be structural in nature, representing beam slots and drainage for houses adjacent to Hare Lane. A substantial wall (503) was recorded at the eastern end of the site, oriented approximately east­west and continuing the alignment of the north wall of Tanners' Hall to the east. The wall was constructed of large neatly squared limestone blocks facing to the south. An area of the wall had been rebuilt (511), possibly in the post-medieval period, using the original stones. A doorway had been cut between the original wall and the rebuild. Traces of mortar remained between the stones; it appeared that most had been washed out. The limited excavation revealed the wall to the depth of 1.15 m and provided no dating evidence. The level from which the wall was constructed was unclear but a considerable depth of undated deposits had built up against the southern face. The wall was left in situ as its alignment suggested that it may have formed part of the tannery yard boundary on the southern side of the site. Future investigation may reveal a relationship to Tanners' Hall. To the south in evaluation trench 1 a hollow way was recorded. It spanned the medieval period, and the earliest fill from it contained pottery dating to the 10th­12th century. Within trench 2 the period between the 10th and the 14th centuries is represented by the dumping of deposits in the area to the south (front) of Tanners' Hall. Truncating this material were a number of cultivation features, which contained quantities of residual material, the latest of which dated to the 15th century.
142
JO VALLENDER
Post Medieval (17th­18th century) The tannery represented by the pits and other structures described above is thought to have continued in use at the start of the 18th century. The precise date of its closure is unclear. The work carried out by Heighway (1983b) within Tanners' Hall suggested that the circular tanning pits (17th century) were earlier than the square (18th century). Four tanning pits, [848, 927, 987 and 1037], were identified as dating to the post-medieval period (Fig. 7). In addition pit [939], measuring 1.8 by 1.2 m and 0.65 m deep, had been cut into the natural deposits at the west end of the site. It was irregular in shape and its eastern side was heavily undercut with visible tool marks. Its function is uncertain but it may have resulted from the removal of a large post. Pit [1037], aligned NE­SW and 4 m long and 1 m wide and of unknown depth, had been backfilled with horn-cores. It is not clear whether they were dumped during the operation of the tannery or during the truncation of the site. Several other recorded features were also thought to relate to the latest phases of tanning on the site. To the north-east of Tanners' Hall a group of linear features, [728, 732, 734, 735, 736, 737, and 738], dating to the 16th or 17th century appeared to have been set out in sequence (Fig. 8). Excavation provided no clue to their function. When tanning had ceased the site was truncated, probably to remove all the traces of the final phase of tanning, and then backfilled with between 1.0 and 1.3 m of imported soil (Fig. 10). It is possible that the depth of soil imported correlates with the depth of features and soil that were removed, but this could not be verified. The soil (627, 609, 606, 616, 502, 686 and 838) was black, humic, silty clay and contained a considerable quantity of Roman pottery. It is considered to be imported, as evidenced by the lack of medieval finds and demolition debris that might have been expected had the site merely been levelled. Over the majority of the excavation area no tip lines were observed within the imported soil and separate layers were only differentiated to the west of the site of a large house called Worcester Lawn. It has been suggested (Heighway 1983b) that the whole site may have been allotments in the post-tannery period until Worcester Lawn was built in the late 18th/early 19th century. The remains of a brick-built structure (602, 603, and 626) were recorded at the eastern end of the site, cutting through the imported soil (Fig. 9). The nature of this building was unclear. Maps showing Worcester Lawn and its outbuildings indicated no structures in this area. Therefore it is likely that the building relates to the use of the site as an allotment. Its dimensions were 8 Ч 10 m and it appeared to have reused the limestone southern boundary of the tannery yard. The surviving eastern wall was visible in the section against Worcester Street. No traces of flooring or internal features within the building remained and the majority of the foundations had been robbed out, possibly when Worcester Lawn was constructed as it was cut by a culvert from that house. At the eastern end of the excavation to the south of the tannery yard boundary the remains of a post-medieval house (505) were recorded. The original date of this structure is unclear, although, from map evidence it would appear to have predated the construction of the houses fronting Worcester Street and to have been demolished by the time garages were built on the site c.1920. Historically, it occupied a parcel of land to the east of Tanners' Hall, bounded to the north and south by walls extending from the hall. The structure had reused the remains of the tannery's stone boundary wall. It was not excavated. Within evaluation trench 1 the medieval hollow way was superseded by a cobbled lane known by the 18th century as Upper Tanners' Yard (later Tanners' Yard) and still in existence at the start of the 20th century. The lane provided access to the front of Tanners' Hall and possibly to the yard to the rear. The yard was flanked by buildings on both sides, the earliest of which dated to the 17th century. Finds collected during the evaluation suggested that pin making or the
50/120
1106
50/110
Hare Lane
927 974
1037
960 1030*
987
939
866
847 1105
884 880
1075
1104 848
1044 1043 1114
1014
887
1058
992
852
1064
948 846
896 859
Tanners Hall
Undated features Limit of excavation Modern intrusions N Structures Areas of burning xxx* Clay deposits
0
2
4
6
8
10m
Fig. 8. Post-medieval and undated features.
817 818 814 813
793 774
763 789 780 759 757 761 775
783 791
803
797
727 728
Trench 4 734 735
732 736
738
737
693 690 678 681
100/120 664 699 622
511 505
100/100
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 143 Worcester Street
50/120 745 50/110
835 outbuilding
Limit of excavation
N
Modern intrusions
Structures
Areas of burning xxx* Clay deposits
Tanners Hall
0
2
4
6
8 10m
Fig. 9. Worcester Lawn and associated features.
Trench 4 626
culvert 613
603
602
100/100
JO VALLENDER
144
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 145 large-scale use of pins took place on the southern side of the yard. By the 18th century the lane appears no longer to have been used as a thoroughfare as a large well was cut through its cobbled surface. This may have coincided with the closure of the tanning industry on the site. By 1809 a large house called Worcester Lawn (745) (Fig. 9) and its outbuildings had been constructed on top of the imported soil (Heighway 1983, 83). The house fronted Hare Lane and was set back c.8 m from the street front, with dimensions of c.9 m east­west and c.14 m north­south. Its gardens to the rear extended to the modern line of Worcester Street and were bounded to the south by the line of the tannery boundary. Outbuildings stood to the east and south of the house. The surviving remains of the house comprised the basement, which was apparently designed to be only partially below ground. It cut through the post-medieval imported soil, which appeared to have been largely removed to the front of the building before being reinstated as evidenced by tip lines. The basement, sub-divided into five rooms, was constructed from unfrogged bricks, ranging in dimensions from 0.70 Ч 0.16 Ч 0.12 m to 0.70 Ч 0.24 Ч 0.12 m and bonded with a light grey mortar. Repairs had been made with a yellow sandy mortar. There was no apparent pattern in the coursing and differing brick sizes were mixed together randomly; some areas of English bond could be recognised. In place of conventional foundations the walls were supported by a series of wooden stakes driven into the underlying deposits. The voids created by the decomposed stakes suggested that they were c.0.10 m in diameter. The depth to which these piles sank was not clear. A number of floors constructed from handmade bricks were in situ, while in the rest of the rooms the imprint of the floor bricks could be clearly seen in the surviving mortar base. Associated with the house was an external toilet (formerly an earth closet), a porch to the front of the house and a culvert which ran into Worcester Street. To the rear (east) of the house the foundations of a large outbuilding (832), measuring 7.5 m east­west, were recorded adjacent to the northern boundary wall. The building is shown on the 1902 Ordnance Survey map for the area (OS 1902) and it would appear to have been a glasshouse. The plan of the building was not fully exposed as it remained in the baulk supporting the modern boundary wall. A brick culvert (613), constructed with flat sides and base and an arched roof, ran east­west across the excavation area. Falling to the east towards Worcester Street, it appeared to have taken waste from Worcester Lawn and other buildings on the site to the main sewer in Worcester Street. The culvert was three quarters full of black silt and probably went out of use when Worcester Lawn was demolished in the early 20th century. The provision of sewerage then appears to have switched towards Hare Lane, as major drain cuts were present within the area fronting Hare Lane. Modern After the demolition of Worcester Lawn most of the activity within the excavation area related to the construction and demolition of garage buildings on the site. These comprised underground petrol storage tanks, car inspection pits and foundations for the various buildings. The earliest garage appears to have fronted Worcester Street, on the former gardens of Worcester Lawn. The last replaced Worcester Lawn on Hare Lane. The rubble from the house itself was spread across the site and used as hard core for the garage floor. On excavation it was found that the majority of the pits and tanks associated with the garages had been backfilled with oil and scrap car parts. On the north-east corner of Tanners' Hall a basement presumably associated with one of the garages also had been backfilled with waste. A feature unrelated to the garages and post-dating their closure was a huge posthole (620) cut to support an advertising hoarding fronting Worcester Street. A rebuild (511) of the southern tannery boundary also appeared to have been carried out during the 20th century.
146
concrete rubble
W
841
11.96mOD
839
[1167]
1168 842 10761091
949 1107
[1106]
838
1076
1091
942
1112 [947]
1169
941
964 944 [951] 916 1011 96150911173[11728]43[922]
1008 1174
959 1076 [1128] 1120
1009
[1178]
[1010]
1170
1176
1175
1133 1001
[1177]
830 [1132]
839 840 838 [1003]
745 830
concrete rubble mortar 838 1084 [1085]
835 830
E 11.96mOD
JO VALLENDER
W
745
11.96mOD
concrete 745 rubble 835
rubble
mortar
745
834
835 timber pile
rubble
832
concrete
835
627
rubble 627 [832]
E 11.96mOD
W 11.96mOD
rubble 832 rubble 627 garage foundations
W
11.96mOD
804
[805]
rubble 627
inspection chamber rubble
627
95/126 105
48
[74]
79
[104]
132
180
163
230
169
179
[181]
627
762 [763]
764 767
788
764 [789]
781 [780]
764 787
772 787
764 [771]
inspection chamber
reddish sandy clay
627 [807]
reddish gravel
E 11.96mOD
reddish sand
806 774
66
182
190
199
202 204 208 [209]
214 206
[203] [205]
220
213 [186]
221 236 243
260
237
238
240 170
187 164 216 206
251 [217] 239 241
251 252 253
200 206 247 248 250 249 225 256 natural sand
242
[259]
Limit of excavation
712 718
627
E
11.96mOD
665
663
665
[664]
Modern intrusions Structures Areas of burning
0
1m
Fig. 10. South-facing section from the excavation and evaluation trench 4 (scale 1:75).
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 147 THE WORKS IN HARE LANE AND PARK STREET After the completion of the main excavation the focus of the archaeological works moved to the line of the new road where it crossed Hare Lane and Park Street to the west to link with Gouda Way. Existing road surfaces were removed to a depth of c.0.9 m and relaid to achieve the correct camber and drainage. This section of the new road was to be carried on a reinforced concrete raft above the archaeological deposits. On removal of the tarmac it was apparent that those deposits were well preserved beneath the formation layer of the road and an archaeological record was made of them. No excavation took place and only surface finds were collected. Part of the exposed area was hand-cleaned for clarification and the area was then recorded by plan, photograph and written record. The stripped area was dominated by the modern sewerage system, which was left in situ to form baulks across the area. No firm evidence for Roman or Anglo-Saxon activity was recorded in this area due to the presence of the medieval and later deposits. Deposit (1508) contained 10th­11th-century pottery and an adjacent gravelly deposit (1505), possibly a base layer for a medieval road surface, contained 2nd-century material. It is probable that both layers had been redeposited or that the finds were residual. Features possibly indicating medieval activity comprised a probable structure in the north of the area and an extensive surface, at an average level of c.11.3 m above OD. The structure was suggested by a possible beam slot (1502) associated with postholes and a pit (1511). The area to the south of it appeared to have been surfaced with pebbles or fine cobbles, repaired with various materials such as limestone and tile. The surface (1500), mostly obscured by post-medieval overburden, appeared in patches over the majority of the stripped area. Evidence of wheel rutting was recorded to the east of the modern sewer baulk, on the line of Hare Lane. The finds recovered from the surface suggested a date between the 12th and the 16th century. Evidence for an earlier surface (1546) was recorded beneath (1500) in the south of the area. Considerable post-medieval activity had taken place. The area between the medieval Hare Lane and Park Street, which ran northwards and parallel to one another, was heavily built up, mostly with terraced housing, in the late 19th and early 20th century (OS 1884­86, 1902, 1923). The majority of the road surfacing of both streets had been destroyed and jumbled remains of extensive cobbling were mixed with the modern debris that formed the modern road sub-base. Some intact areas of cobbling had survived, especially in the areas adjacent to the modern pavements. The cellar of a public house (possibly the Worcester Arms) was clearly visible between the two streets. It had been partially flooded and as a result contained a layer of black silt. The cellar was constructed with faced stone, which appeared to have been reused. The foundation cut for these walls was much larger than necessary and backfilled with lias clay, possibly in an attempt to keep out damp. The cellar steps were intact, as was the ramp to roll the barrels into the cellar, and both were constructed from brick. A barrel shelf was built into the west wall. Above cellar level the building had been completely removed. Between the two streets, and to the north of the cellar, were the remains of another post-medieval structure. The pattern of foundation and drains recorded did not indicate clearly the building's plan but map evidence suggests that an end-terrace property once existed in this location and shows that a lane ran between the two structures (OS 1884­86, 1902, 1923). Several inter-cutting features were recorded on the eastern side of the stripped area located in Hare Lane. These appeared to be post-medieval and modern in date and were thought to be service trenches. Large areas of the site remained covered by post-medieval overburden, possibly resulting from the considerable demolition which took place in the area before 1955 (OS 1955).
148
JO VALLENDER
The stratigraphy in the north-western area (west of Park Street) remained unclear. The remains of the terrace of houses which fronted Park Street until the 1930s should have been evident. However, only post-medieval overburden and a layer of black silt were recorded.
THE WATCHING BRIEF After the road corridor across Hare Lane and Park Street had been recorded, a watching brief was carried out during the installation of drainage and sewerage connections and new street furniture. In addition, three trenches (A, B and C: Fig. 2) were excavated to install bracing for the southern wall of the garage premises adjacent to the northern edge of the new road. Sketch plans and sections were recorded where appropriate, and finds, all unstratified, were collected. Trenching to the west of Park Street revealed only a 2.1 m deep layer of post-medieval black silt overlying natural sand. The evidence suggested that the post-medieval houses in this area had been completely removed after their demolition possibly when Gouda Way was constructed. Excavations in Hare Lane adjacent to the west end of the main excavation area recorded patches of post-medieval cobbling and, 0.8 m below this, another cobbled surface thought to correspond to (1500) to the west. Below this surface was a layer of greenish sandy clay similar in nature to (830). From this deposit three fragments of medieval pottery were retrieved. Further excavations in this area recorded an undated linear feature running along Hare Lane, possibly a roadside ditch. Trenches inserted in Worcester Street, at the eastern end of the main excavation area, recorded 0.5 m of black silty soil, the continuation of the imported post-medieval soil. Beneath it was a greenish sandy deposit, possibly (700), overlying a layer of orange sand and gravel at a height of 11.47 m above OD and a fine pebble surface. The orange gravel was considered to be a continuation of the make-up layers of the Roman road and the pebbles a continuation of its surface (779), which extended at least 1.2 m past the eastern edge of the excavated area. The continuation of the post-medieval imported soil to the east may be evidence of the clearance of the former tannery yard, which perhaps had extended east of the Worcester Street boundary towards the river Twyver. A baulk left adjacent to Tanners' Hall during the excavation was machine excavated to the same level as the main excavation. A southwards continuation of the series of late medieval linear features aligned NW­SE (736 et al.) was observed. The post-medieval culvert running east­west across the excavation site was found to continue to the immediate north of Tanners' Hall, ending in the area of the toilet in the outbuildings associated with Worcester Lawn. The line of the culvert provided the new section edge for the road and was left in situ. Among the three trenches excavated adjacent to the north side of the road, the post-medieval soil in trench A was 0.9 m deep below 0.2 m of concrete. Below this was a large greenish deposit (830?) through which a tanning pit had been cut. The pit was 1.86 m deep and c.1.8 m of its diameter was exposed in the section. The latest fill was dark green organic material, probably tanning waste, overlying a dump of stony material, which sealed a layer of black silt in the base of the cut. Deposit (830?) was more than 1.6 m deep and directly overlay the natural sand, the top 0.2­0.3 m of which appeared to have been either redeposited or heavily affected by leaching from the tanning pit. In trench B 1.2 m of post-medieval soil was recorded beneath the concrete. It overlay a deposit of green sandy silt, also thought to be (830), which sealed what appeared to be a linear feature filled with black green sandy silt, running roughly north­south. A layer of stone, possibly the remains of a surface, was visible at the top of the cut. The linear feature cut the natural sand at 2.6 m below the concrete garage floor.
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 149 In trench C several archaeological features were recorded below the post-medieval overburden. This trench was within the area previously occupied by Worcester Lawn and the north wall of the building was observed in the section. Two truncated tanning pits c.1.3 m deep were recorded. One had sides sloping to a rounded base and had mainly been backfilled with a green sandy silt containing lenses of lime. The other, square in section with a flat base, was backfilled with similar material. Both pits had c.0.2 m of black silt in their bases. Deposit (830) was not evident in the section but at the equivalent level were layers of redeposited natural sand, which overlay a possible surface at c.2.4 m below ground level. This sealed layers of redeposited natural sand with a layer of silt beneath, possibly in the base of a cut. THE FINDS Pottery by Jane Timby The excavation resulted in the recovery of more than of 8,376 sherds of pottery weighing 125.8 kg and dating to the Roman, late Saxon, medieval and post-medieval/modern periods. Approximately 5,000 sherds (62%) came from identified archaeological contexts; the remainder was unstratified. As a result of the disturbed nature of the site and the constraints of the archaeological work, which have limited detailed interpretation of the individual deposits, the pottery has not been analysed to the level of detail that would perhaps normally be expected with such a large group of material from the centre of an important Roman and medieval town. Although there have been numerous archaeological investigations in and around Gloucester over the past two decades most of the documentation of the Roman, medieval and later pottery sequences was undertaken some time ago and there has been little apparent recent maintenance or updating of the ceramic record. Much of this ceramic work remains unpublished and essentially inaccessible. It is fortunate that our knowledge of the broad ceramic history of Gloucester from Roman times has been well established through published excavation reports (e.g. Ireland 1983; Vince 1983; Goudge 1982). Following a description of the methodology employed this report summarises the pottery from each of the main periods by each excavated area. Methodology The complete pottery assemblage was subjected to a standard assessment. The assemblage was sorted into broad fabric types corresponding to the Gloucester City Excavation Unit Type Fabric (TF) series, but no attempt was made to identify any of the less common, unfamiliar wares ­ largely glazed medieval sherds and some of the post-medieval types ­ and they were given general codes. A sherd count and weight quantification were made for each recorded context and a date range provided. The unstratified pottery without context numbers and the watching brief material were scanned for any unusual pieces but not sorted or quantified in detail. It was clear from the assessment that there was considerable mixing of deposits with sherds of wide ranging date occurring alongside each other. This was particularly apparent for the Roman assemblage, which continued to manifest itself throughout the medieval and post-medieval sequence possibly as a result of soil being imported into the site from elsewhere in Gloucester. Analysis using the stratigraphic matrix demonstrated how severe this problem was for the post-Roman deposits with numerous instances of later material being stratified below what initially appeared to be earlier groups. Despite the post-depositional disturbances the pottery is quite well preserved with an overall average sherd weight of 14.5 g. There are no complete or reconstructable vessels.
150
Table 1. Roman pottery.
JO VALLENDER
Imports
Glos TF NRFCC Description
Number stratif
3B 8A-C 12H 12J 12K 12N 12Q 12S 9C 9I 10A 10B 10C 10 10
LYO CC CNG BS MOS BS ARG CC KOL CC ARG RS NOG WH RHL WH BAT AM GAL AM NAF AM
imported mica-slipped fineware 5
South, Central, East Gaulish samian 155
Lyons ware
0
Central Gaulish colour-coat
0
Moselle colour-coated ware
0
Argonne ware
1
Cologne colour-coat
0
North Gaulish colour-coat
1
North Gaulish mortaria
0
Rhenish whiteware mortaria
0
South Spanish Dressel 20 amphora 14
Gaulish wine amphora
11
North African amphora
4
hollow foot amphora, Kapitan II 0
unclassified amphora
2
Weight (g) stratif 32 1,703 0 0 0 15 0 3 0 0 3,211 163 401 0 22
Number u/s 2 328 1 3 2 0 1 0 2 2 43 25 0 1 5
Weight (g) u/s 19 3,313 10 18 6 0 1 0 659 316 3,477 792 0 40 302
Number % overall * 11 * * * * * * * * 1 * * * *
Weight % overall * 7 * * * * * * 1 * 5 1 * * *
Regional 4 DOR BB1 Dorset black burnished ware
279
3,184
757
8,924
24
175
1A OXF PA Oxfordshire parchment ware
0
0
6
48
*
*
12A OXF RS Oxfordshire colour-coat
17
135
124
1,114
3
2
9X OXF RS Oxfordshire colour-coated mortaria 1
27
17
258
*
*
13 OXF WH Oxfordshire white ware
25
259
26
215
1
*
9A OXF WH Oxfordshire whiteware mortaria 10
489
19
959
*
2
9W OXF WS Oxfordshire white-slipped mortaria 0
0
2
75
*
*
12B LNV CC Lower Nene Valley colour-coat 2
29
51
380
1
*
9E LNV CC Lower Nene Valley mortaria
0
0
2
89
*
*
12C NFO CC New Forest colour-coated ware 0
0
2
6
*
*
15A SOW WS South-west white slipped ware
2
28
16
190
*
*
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 151
15B
South-west red burnished ware
1
2
3
35
*
*
12D
South-west colour-coated ware
0
0
1
7
*
*
9 SHM RS Shepton Mallet red slip mortaria 0
0
1
3
*
*
6 SAV GT Savernake ware
7
194
2
122
*
*
201
Wilts wheelmade black burnished 28
149
0
0
*
*
231
Wiltshire grey sandy wares
3
21
0
0
*
*
232
Wiltshire oxidised ware
1
2
0
0
*
*
22 ROB SH late Roman shelly ware
2
15
4
106
*
*
241 PNK GT pink grogged storage jar
0
0
5
177
*
*
Gloucester 3A 7 11A 9B 25 39
mica-slipped Gloucester ware
0
0
3
9
*
*
white slipped Gloucester kiln ware 52
588
56
673
2.5
2
oxidised/grey Gloucester kiln ware 221
2,694
267
3,136
11
8.5
Gloucester mortaria
3
62
3
153
*
*
sanded Gloucester kiln ware
4
43
7
94
*
*
Kingsholm grey sandy ware
1
10
2
40
*
*
Local
5
micaceous greyware
35
805
110
1,500
3
3
11B SVW OX Severn Valley ware
315
5,141 1,138 19,678
33
36
17 SVW OX Charcoal- tempered SVW
2
44
1
8
*
*
23 SVW OX handmade SVW storage jar
1
10
0
0
*
*
19
Malvernian ware
3
32
16
943
*
1.5
Native
2A
33
grog-tempered handmade
4
62
1
8
*
*
Malvernian limestone-tempered 2
54
0
0
*
*
Unknown 00 9 Total
unclassified unclassified mortaria
19
245
73
944
2
1.5
1
53
2
51
*
*
1,234 19,927 3,132 48,898 100
100
Key: * = less than 1%
152
JO VALLENDER
For the publication report the basic data compiled during the assessment was used. Further work was carried out on some of the unidentified wares from the stratified contexts. The evaluation and watching brief collections were not looked at again. A small group of material was selected for illustration. No detailed fabric descriptions are given as these have generally been described in detail elsewhere. For Roman see Goudge 1982; Ireland 1983; Rawes 1972; Timby 1991; for Saxon Vince 1979; 1982; and for medieval and post-medieval Vince 1979; 1983.
Roman Roman pottery, amounting to some 4,366 sherds (68.8 kg), accounts essentially for just over half the total stratified assemblage, 54 per cent by sherd count. The fabrics present are listed in Table 1. The original Gloucester type fabric codes have been retained augmented by the national Roman reference collection codes which relate mainly to well-known regional types or traded wares (Tomber and Dore 1998). In terms of the types represented the pottery suggests intense activity in the locality from the late 1st/early 2nd­3rd century. However, as is apparent from Table 1, 72 per cent of the 4,366 sherds quantified are redeposited in post-Roman horizons. No detailed specialist work has been carried out on the samian of which 483 sherds were recovered, 70 per cent from post-Roman layers. The majority of this is plain with only 30 decorated sherds noted and at least 10 stamps. A light scatter of potentially pre-Flavian material is present. It may derive from earlier activity or, more likely, may have been imported into the site. Included are typical wares one might associate with the military occupation at Kingsholm, for example Lyons ware, South Gaulish samian and Kingsholm grey sandy ware (TF 39). The earliest Roman activity appears to date to the latter part of the 1st century AD into the early 2nd century. The pottery evidence is mainly represented by wares from the local Gloucester kilns (Rawes 1972; Timby 1991). Those kilns, found on at least three sites dating from the late Flavian­Trajanic periods, produced a range of oxidised (orange), reduced (grey) fine (TF 11A), sanded (TF 25), white-slipped (TF 7) and mica-slipped (TF 3B) vessels. There is at least one stamped Gloucester mortaria in the Tanners' Hall group (Fig. 11, no. 16) from a later context. Gloucester products account for 14 per cent by sherd count, 11 per cent by weight. From the 2nd century increasing amounts of local Severn Valley ware gradually replace the Gloucester wares. Regional imports include Oxfordshire white wares, both flagons and mortaria, and Dorset black-burnished ware. Amongst the continental imports are Dressel 20 olive oil amphora from southern Spain, wine amphora from Gaul, abundant samian tablewares and Gaulish colour-coated beakers. The incidence of samian begins to decline from the early 3rd century to be replaced in the second part of that century by colour-coated wares from the Oxfordshire industry. Other imports of note include a Central Gaulish (Lezoux) barbotine decorated cup from (627) generally dated to between 180 and 210/220 AD (Fig. 11, no. 17). In the 4th century products from the large regional industries based in the Severn Valley, Oxfordshire and Poole Harbour, Dorset, dominate the market. Other typical late wares include a small number of sherds of New Forest colour-coated beaker and late Roman shelly fabric typical of the later 4th century. Of particular note are fragments from later Roman amphorae including a hollow foot Aegean type (Kapitan II) from (966) probably imported in the 3rd­4th century. This is the first recorded instance in Gloucester of a type which shows a mainly coastal distribution in Britain especially around the Thames estuary (Tyers 1996, fig. 83). At least four sherds, including a handle, are an African amphora probably dating from the later 4th to 6th century. African amphora sherds have been documented in Gloucester at St Oswald's priory (Goudge 1982, 49)
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 153 and the New Market Hall (Hassall and Rhodes 1974, fig. 33.37). A single fragment of a cйramique а l'йponge flanged bowl (Raimbault 1973, type VI) from ditch (506) is likely to date to the 4th century or later. Despite the fact that it is well represented elsewhere in Gloucester, there is perhaps surprisingly little of the late Roman shelly ware indicative of occupation in the last quarter of the 4th century. This could suggest that the later Roman levels have been truncated or that there was little activity in this area in the later Roman period. The main excavation The main excavation area shows a pottery sequence starting with early Roman (late 1st­2ndcentury) material. The road surface (779) produced no pottery but a total of 30 pieces was recovered from associated deposits (702), (705) and (770) sealed below this. These include types spanning the later 1st to early­mid 2nd century; mainly products of the Gloucester kilns, but also including a Central Gaulish dish Curle 35/6 from (705). Deposit (671) south of the road produced a small assemblage of 21 sherds with a terminus post quem (tpq) in the late 2nd century. The assemblage was mixed and included sherds of Oxfordshire whiteware, black-burnished ware (acute lattice), Dressel 20 amphora, Central Gaulish samian (Drag. types 31, 33), Severn Valley ware, Gloucester kiln ware, micaceous grey ware (TF 5) and south-west white-slipped flagon. Only three of the features cutting floor (666) produced pottery of 2nd-century date; (669) and (670) included 2nd-century pottery and (634) a mixture of Roman and medieval sherds. The series of deposits west of (671) associated with possible structure (683) collectively produced an assemblage of some 33 sherds. They are mainly of 2nd-century currency but with some slightly later material from (685) and (689), including an Oxfordshire colour-coated beaker from the latter likely to date to the later 3rd or 4th century. Posthole (649) south of floor (666) also produced a late 2nd-century collection of 21 sherds. The opus signinum floor (628) produced a single Central Gaulish dish (Drag. 31) and three Severn Valley ware sherds. No pottery seems to have been associated with kiln structure (637) or ditch (520). The post-Roman build-up (830) produced a large assemblage of some 398 sherds (5.75 kg) of mixed chronology despite the fact only c.5 per cent was excavated. Medieval sherds spanning the mid 11th­15th/16th century account for 60 per cent of the group. The remaining sherds are Roman. Evaluation trenches The earliest activity in trench 1 dates back to the late 1st­early 2nd century with a group of 25 sherds from cess pit (246). It includes local Gloucester wares, a Central Gaulish samian cup (Drag. 33), Severn Valley ware and Savernake ware. Possible beam slot (234) produced pottery of 2ndcentury currency from its fill and is immediately succeeded by horizons containing medieval wares with no surviving late Roman levels. The sequence in trench 2 commences with Roman pottery dating mainly to the 3rd century. The make-up/levelling (171) for the uppermost Roman surface (148) produced a handle from a late Roman north African amphora imported from the late 4th­6th century. The small group of wares from surface (148) includes a sherd of late 4th-century Roman shelly ware suggesting that some form of activity may have been under way in this area in the very late Roman/subRoman period. Trench 4 also produced a good pottery sequence. It started with an early Roman ditch (259) containing wares of late 1st- to 2nd-century date from the primary fill. Little material came from
154
JO VALLENDER
the backfilling and consolidation other than late 1st­early to 2nd-century material from (221/220). Small groups of 2nd- to late 3rd-/early 4th-century pottery were recovered from cut (203). Exclusively Roman pottery recovered from features (217), (215) and (151) may be redeposited material.
Catalogue of illustrated sherds (Fig. 11) 1. Necked bowl, TF 11B (SVW OX). (506). 2. Hemispherical bowl. TF 11B (SVW OX). (506). 3. Thickened rim globular jar. Grey TF 11A. (670). 4. Everted rim jar. Grey TF 11A. (705). 5. Hemispherical bowl with a trace of rouletted decoration. Grey TF 11A. (705). 6. Straight-sided bowl. Grey TF 11A. (665). 7. Everted rim jar, TF 11B (SVW OX). (665). 8. Globular jar/beaker with a slightly moulded rim. Grey TF 11A. (665). 9. Small hemispherical bowl. Oxidized TF 11A. (665). 10. Short everted rim beaker. Oxidized with a grey core. TF 11A. (665). 11. Shallow flat rim bowl with burnished lattice decoration. TF 4 (DOR BB1). (665). 12. Wide-mouthed beaded rim bowl decorated with a fine burnished lattice. TF 4 (DOR BB1). (665). 13. Sharply everted globular-bodied jar. Oxidised TF 25. Partly burnt. (645). 14. Necked jar, well-fired. Dark grey with brown streaks, burnished on the exterior. TF 11B. (700). 15. Rim from a hollow-foot amphora, Kapitan II probably from the Aegean. dark brownish red fabric with sparse white angular quartz grains. (966). Typical of 3rd to 4th-century assemblages. 16. Gloucester mortaria, TF 9B, with a stamped flange. (616). 17. Carinated, handle cup with barbotine decoration. Central Gaulish black colour-coated ware (TF 12J; CNG BS). Date range 180­210/220. A similar vessel is recorded from New Fresh Wharf, London (Richardson 1986, fig. 1.97). 18. Whiteware bowl with red-painted decoration, probably an Oxfordshire product. (616).
Anglo-Saxon A single organic-tempered handmade sherd was redeposited in (830). A scatter of organic or grasstempered sherds has been noted in Gloucester, notably two sherds at St Oswald's priory (Vince 1982) and others from unpublished excavations conducted elsewhere, between 1968 and 1973 (Vince 1984, 240). The tradition is quite a long-lived one and recent finds from Bishop's Cleeve suggest that it was still in use in late Saxon times. A small group of sherds typical of the late Saxon period (10th­11th century) is present. These vessels, mainly oolitic limestone-tempered wares (Gloucester TF 41A) thought to have been introduced in the 10th century, have been relatively well documented from elsewhere in Gloucester, in particular from Westgate Street (Vince 1979). A handmade coarse shell-tempered jar from (140), probably redeposited, may also date to this period. The forms identified in TF 41A include a lid-seated jar (Fig. 12, no. 19), a handmade jar with an everted rim and thickened neck (Fig. 12, no. 21) and wheel-thrown everted rim jars/cooking pots. Evidence that the late Saxon material indicates late Saxon occupation on the site is scant. A number of layers succeeding the Roman layers apparently contain small numbers of Roman sherds and single pieces of TF 41. Further analysis suggests that with the exception of linear fill (983) and deposit (1508) all these pieces are TF 41B dating from the later 11th century.
Medieval Stratified medieval wares account for approximately 23 per cent by sherd count of the overall assemblage, some 1,893 sherds weighing 26 kg (Table 2). An additional 780 sherds (12.7 kg) came
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 155 Fig. 11. Roman pottery. from post-medieval deposits. The stratified material contains pottery typical of the 11th­12th century through to the 14th­15th century. The group is dominated by the Malvernian industry, TF40/52, which accounts for 45 per cent by sherd count. Gloucester oolitic-limestone-tempered ware (TF 41B) accounts for 25 per cent and Minety ware for 7.5 per cent. All the other wares account for 4 per cent or less of the total. The range of material is fairly typical of Gloucester. A high proportion of the deposits contained single medieval sherds alongside Roman material, in particular sherds of TF 41B. The dating of these could, therefore, lie anywhere from the 11th
156
JO VALLENDER
Table 2. Medieval pottery.
Glos TF Description
Roman
redeposited sherds
Number 1,214
Late Saxon 0 misc. handmade
3
41A oolitic limestone-tempered
18
Medieval 40 Malvernian unglazed ware
686
41B oolitic limestone-tempered
475
42 misc. sand-tempered unglazed 61
43 sand and oolite-tempered
82
44 Minety ware
145
50 misc. sand-tempered glazed
5
52 Malvern Chase ware
177
53 Ham Green jug
9
83 Brill-Boarstall type
25
90 Worcester-type glazed ware
48
91 Worcester-type cooking pot
46
98 Oxford fabric OX Y
2
99 coal-measures glazed ware
1
110 sandy-limestone-tempered
3
M00 unsorted wares
78
Late Med 54 59 65 Sub-total Total
Herefordshire border ware later Surrey wares Tudor Green ware
17 27 6 1,893 3,128
Key: * = less than 1%
% Weight (g) % 15,305
20 216
36
9,837
38
25
4,922
19
3
857
3.5
4
872
3.5
7.5
2,992
11.5
*
37
*
9
2,529
9.5
*
140
*
*
315
1
2.5
636
2.5
2.5
842
3
*
58
*
*
43
*
*
110
*
4
1,105
4
*
183
*
*
500
2
*
8
*
100
25,986
100
41,527
century onwards. Indeed the high level of redeposition suggests that any dating involving single or small groups of sherds must be treated with caution. The earlier (late 11th-/12th-century) levels are particularly dominated by jars/cooking pots in TFs 40, 41B and 43. To them can be added glazed spouted and tripod pitchers in Minety ware (TF 44), oolitic ware (TF 41 B) and Malvern Chase ware (TF 52). A rare example of a stamped TF 41B jar was noted from gully (872). A single example of a jar (TF 110) from the Ross-on-Wye area was noted in context (855). These wares, containing calcareous sandstone, are rare in Gloucester (A. Vince pers. comm.). Regional wares become more apparent from the 13th century and include glazed jugs from Worcester, Bristol (Ham Green), the Surrey­Hampshire border industry and Brill-Boarstall in Buckinghamshire. To these can be added a sherd of developed Stamford ware from (808) and sherds of glazed white-ware jug (TF 99) whose petrology suggests a source on the coal measures. In the later medieval period are a few sherds of Tudor Green dating to the 15th­16th century and Herefordshire border wares. The former includes a lobed cup and jugs, the latter bowls, pipkins, jars and jugs. These are augmented by Spanish olive jar, Cistercian ware and Raeren
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 157 stoneware from the post-medieval layers. A sherd from a possible late medieval Italian glazed jar was recovered from slot (735). The fabric is oxidized, fine textured, red firing clay with moderate to abundant round voids with calcareous reaction rims. Similar vessels are known from Southampton and London (A. Vince pers. comm.). The main excavation The fill of the large linear feature (984/1146) produced 86 sherds. The assemblages from fill (983/1136) contained mainly redeposited Roman sherds accompanied by two sherds of TF 41A, five of TF 41B including a spouted pitcher, and one of Malvernian ware TF 40 suggesting a tpq in the late 11th­early 12th century. The large pit (993) contained two sherds of TF 41B from its primary fill. Another sherd of the same from (975) is again indicative of a late 11th- to early 12thcentury date. Deposit (970) contained more TF 41B accompanied by two sherds of Worcester plain jar (TF 91) and Malvernian ware (TF 40) suggesting a 12th-century date. Pit (969) cutting the deposit produced 37 sherds, mainly Roman accompanied by seven medieval pieces (TFs 40, 41B, 43 and 44), which all again could be of 12th- or early 13th-century currency. The charcoal-filled pit (1072/1073) adjacent to linear feature (984/1146) produced eight sherds amongst which were five pieces of TF41B (late 11th/12th century). The four stakeholes in the base of this feature were cut into gully fill (1118) which produced twelve sherds of TF 41B, including a spouted pitcher and a jar, and pieces of TF 40 and 43 indicative of a late 12th- to 13thcentury date and suggesting that the pottery stratigraphically above this is redeposited. Many of the tanning pits contained more Roman than medieval material. The range of medieval pottery present appears to span the 12th to 15th/16th century. Possible earlier groups dating to the early 12th/13th century with plain jars/cooking pots or spouted pitchers in TF 41B, 40 or 43 are typically from pits 1123 (1124), 1054 (1053), 1003 (1001), 1010 (1009), 981 (982), 957 (924), 812 (fill and cut), 865 (864) and 671 (673). Pit 1148 (1147) has a typical 12th-century assemblage but is accompanied by a 14th- or 15th-century glazed Malvernian tile, perhaps intrusive, illustrating the dangers of dating small groups. Possibly slightly later groups dating to the 13th­14th century are typified by a wider range of wares and include plain ware sherds of Malvern Chase TF 40/52, Minety ware TF 44, oolitic limestone ware TF 41B, sand- and limestone-tempered ware TF 43 and Worcester jars TF 91. These are accompanied by smaller numbers of tablewares in the form of decorated glazed jugs or pitchers from Ham Green TF 53, Minety TF 44, Brill-Boarstall TF 83 and Worcester TF 90. Pits with this type of assemblage include 1130 (1129), 1123 (1101), 1062 (1061), 959 (947), 929 (928), 749 (fill and cut) and 74 (48). Slightly later pits are suggested by the presence of Tudor Green TF 65 and/or Herefordshire border wares TF 54 alongside the same range of material. Amongst later pits is 1092 (1081) with a tripod pipkin (Fig. 12, no. 36). Evaluation trenches A small amount of medieval material was recovered from the hollow way in trench 1. No pottery was retrieved from the cobbled surface (146/149) sealing this. In trench 2 the clay deposit (107) sealing the latest Roman surface contained Roman wares and two handmade sherds of TF 41B suggesting a date in the 11th­12th century. Similarly deposit (111/112) and the fill of pit (118) sealed by (107) contained single sherds of TF 41B. Although the sherds are themselves non-diagnostic and could date from any time between the 11th and the 13th centuries the absence of other medieval ware strongly suggests a pre 12th-century date. Further sherds of TF 41A and 41B from (103) and (76) included a spouted pitcher from the latter. Build-up (39) included sherds from a TF 41B tripod pitcher and a typologically early rim from a
158
JO VALLENDER
Fig. 12. Medieval pottery.
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 159 Malvern Chase (TF 40) cooking pot indicating a probable early 12th-century date. Ditch fill (75) produced a single glazed Ham Green jug sherd (mid 13th century). The garden features contain a mixture of redeposited Roman and 11th­12th-century sherds, nearly all TF 41B or Malvern Chase (TF 40). The only context to contain different material is pit 26 (25) which, with two small sherds of Tudor Green (TF 65), suggests a date from the 15th­16th century. This is succeeded by 18th-century and later activity. In trench 4 Roman pottery recovered exclusively from features (217), (215) and (151) may be redeposited. The large tanning pit (181) produced a good assemblage of 13th- to 15th-century pottery. Sherds broadly of similar date came from the post-tanning deposits. No post-medieval material appears to be present in this trench. Catalogue of illustrated sherds (Fig. 12) 19. Small lid-seated, handmade jar. TF 41A. Deposit (1508). 10th­11th century. 20. Handmade wide-mouthed jar. TF 43. 11th century rubbish pit (1072). 21. Handmade cooking pot with slight sooting on the rim exterior. Fettling marks can be seen at the join of the rim. TF 41A. Pit (1147). 10th­11th century. 22. Handmade jar. TF 41B. Pit (1147). 11th­12th century. 23. Wheel-turned jar. TF 40. Pit (1147). 12th­13th century. 24. Simple rim, wheel-turned jar. TF 41B. Deposit (1112), 12th­13th century. 25. Jar. TF 41B. Fill (1034). 12th century. 26. Simple everted rim jar. TF 43. Fill (1034). 12th century. 27. Handmade cooking pot with sooting on the exterior. TF 41B. Fill (1034). 12th century. 28. Handmade jar. TF 40. Posthole (1108). 12th century. 29. Handmade spouted pitcher with a thin external green glaze. Traces of alternating incised and applied vertical strip decoration on body. TF 44. Gully (1139). 12th century. 30. Wide-mouthed cooking pot with a slightly sooted, black exterior. Wheel-turned. TF 40. Tanning pit fill (1124). 31. West-Country style handmade jar with a single wall perforation, TF 40. ?Tanning pit fill (1001). Late 12th­13th century. 32. Burnt dark green glazed jug with an applied face medallion. TF 90. Deposit (634). Mid 13th century. 33. Wide-mouthed, wheelmade jar. TF 41B. Fill (954). 12th­13th century. 34. Small cooking vessel with a sooted exterior. TF 91. Fill (954). 12th­13th century. 35. Green glazed jug. TF 90. Fill (954). 13th century. 36. Pipkin with a sooted exterior and burnt interior. Slight pouring lip. Speckled green glaze at the interior of the rim interior and base. Scars where feet were once attached. TF 52. ?Tanning pit (1081). 15th­16th century. Post Medieval Pottery dating from the 16th/17th­20th century (Table 3) accounts for 12.5 per cent of the overall assemblage by sherd count. Of the pottery recovered from the post-medieval horizons, 72 per cent is redeposited Roman or medieval material and only 28 per cent is of actual post-medieval date. This group is itself very much dominated by wares dating to the 18th and 19th centuries. Pottery dating to the 16th and 17th centuries is not particularly common but potentially includes Cistercian ware, Malvern Chase ware, Surrey­Hampshire Border ware and a few sherds of North Devon ware. A 17th-century Surrey porringer occurred in (808). The later ware includes numerous industrial glazed white earthenwares, plain and hand- and transfer-decorated, salt-glazed wares, slipdecorated dishes, various iron-glazed kitchen wares, imported Westerwald stoneware, a small amount of porcelain, English stoneware and various local glazed red earthenwares.
160
JO VALLENDER
Conclusion Generally speaking the assemblage presents a sequence fairly typical of those recovered elsewhere in Gloucester. The group dates from the Roman period through to the modern with an apparent absence of recognisable activity or cultural material dating to the early­middle Saxon period (5th­9th century). The Roman evidence points to a relatively early phase of activity beyond the 2nd-century defensive circuit but essentially along the route linking the settlement in Gloucester with the slightly earlier Roman fortress at Kingsholm. It is possible that the reason for Roman occupation here is connected with industrial activity, more desirable away from the centre of the Colonia. Among the few comparable sites outside the north-west defences is the Roman colonial tilery at St Oswalds dating from the later 1st century through to the 3rd century (Heighway and Parker 1982). After the tilery's abandonment its site was used as a Roman cemetery so there is little evidence there of 4th-century pottery. Previous excavations at Tanners' Hall (Heighway 1983b) recovered a small assemblage of just 121 sherds of residual Roman pottery ranging in date from the Flavian to the late Roman periods with nothing diagnostically later than early 4th century. Similarly the late Saxon/early medieval deposits lie outside the town's north gate, probably rebuilt in the 11th century, but inside its Alvin gate first mentioned in the late 12th century. It has been postulated that Hare Lane once formed part of a late Saxon extra-mural cattle or stall market (Heighway 1983b, 7). Such a use might suggest little accumulation of domestic debris. The pottery from Tanners' Hall appears to suggest a possible change of activity/use in this area around the 11th­12th century. At most Gloucester sites 12th­13th-century pottery is the most common medieval pottery found. This site is no exception. The assemblage appears to contain a high proportion of cooking wares and spouted tripod pitchers along with a marked number of glazed table wares, including some highly decorated jugs/pitchers. The assemblage differs considerably to that recovered from the previous excavations at Tanners' Hall which produced some 1,809 sherds (Ireland and Goudge 1983). Then, however, the Roman levels were not excavated, a fact which is reflected in the overall composition of the group by sherd count which can be broken down in percentages to 6.5 Roman, 11.5 medieval (11th­13th century) and 82 post medieval compared to 54 Roman, 33.5 (late Saxon/medieval) and 12.5 post medieval from the more recent work. The much larger recent assemblage shows a greater chronological span.
Acknowledgements I am particularly indebted to the late Alan Vince for his assistance in identifying some of the medieval wares and commenting in particular on the Italian import and to Paul Tyers for identifying the Aegean amphora.
Ceramic Building Material by Jane Timby Recovered ceramic building material (cbm) amounted to some 1,650 fragments weighing over 108 kg. Although it was associated with some 232 contexts much of the material appears to be in secondary contexts. It includes Roman, medieval, Tudor and post-medieval pieces. As part of the assessment the material was rapidly sorted into types based on diagnostic features where possible and a crude count made. Many pieces were in a fairly degraded condition and could only be categorised as indeterminate. No further detailed analysis of the material has been undertaken.
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 161
Table 3. Post-medieval pottery.
Glos TF Description
Roman
redeposited
Number 1,918
% Weight (g) %
51
33,593
56
Late Saxon 46 grass-tempered ware
1
*
12
*
Medieval 40 Malvernian unglazed ware
308
41 oolitic limestone-tempered
236
42 misc. sand-tempered unglazed
8
43 sand and oolite-tempered
18
44 Minety ware
74
50 misc. sand-tempered glazed
3
51 Stamford ware
1
52 Malvern Chase ware
57
53 Ham Green jug
2
83 Brill-Boarstall type
18
90 Worcester-type glazed ware
26
91 Worcester-type cooking pot
2
99 coalmeasure glazed ware
1
M00 unsorted wares
25
8
4,878
8
6
3,066
5
*
85
*
*
268
*
2
1,404
2
*
20
*
*
4
*
1.5
2,170
3.5
*
5
*
*
397
*
*
175
*
*
14
*
*
10
*
*
220
*
Late Med 54 59 60 65 68 117
Herefordshire border ware later Surrey wares Cistercian wares Tudor Green ware Raeren stoneware Spanish olive jar Italian imported jar
51
1.5
1,036
1.5
13
*
143
*
3
*
12
*
9
*
54
*
1
*
5
*
1
*
12
*
1
*
12
*
Post Med 61/74 Staffordshire iron glazed
90
2.5
742
1
62 tin glazed ware
22
*
162
*
63 unglazed red earthenware
9
*
286
*
70 Devon gravel-tempered ware
7
*
169
*
69/71 late white-firing ware/transfer 628
16.5
5,386
9
decoration
75 Staffs black glazed kitchen ware 20
*
172
*
77 Whieldon ware
3
*
9
*
80/103 glazed red earthenwares
155
4
4,272
7
94 Westerwald stoneware
11
*
144
*
95/96 English stoneware
18
*
294
*
PM00 unsorted/unidentified wares
25
*
738
1
Total
3,765 100 59,969 100
Key: * = less than 1%
162
JO VALLENDER
Roman Many of the pieces are Roman in origin, mainly roofing material, particular tegulae and to a lesser extent imbrices. Several fragments of pilae were noted. Only three possible fragments of box-flue were identified on the basis of combing. None of the Roman pieces showed evidence of stamps or other structural features worthy of note. A pila from (897) had an impressed animal pawprint. Part of a fired clay bar, possibly kiln furniture, of rectangular cross-section came from (688). Much of the Roman cbm is redeposited in later levels, possibly indicative of the possible reuse of material or simply reflective of the disturbed nature of the deposits or the dumping of soil brought in from elsewhere in Gloucester. The proximity of the Roman tilery at St Oswalds (Heighway and Parker 1982) may also have bearing on the quantity of ceramic tile recovered.
Medieval and Post Medieval Medieval and post-medieval cbm was limited to bricks, roof tiles and a few glazed ridge tiles. A stone roof tile with a nail hole came from (25). Evidence from elsewhere in Gloucester suggests that ceramic roofing tile started to be used in the late 12th century (Vince 1983, 214). The presence of several medieval imbrices, some associated with 12th-century pottery in, for example, (165, 989 and 1034), could suggest the presence of a relatively high-status early building close to the site. In the 12th century high-status buildings were roofed in a style imitating the Roman use of tegulae and imbrices (A. Vince pers. comm.). From the 13th to the 16th century flat ceramic or stone roof tiles along with ridge tiles were used. Several of the ceramic pieces were partially glazed. Eight handmade bricks, two complete, were recovered from post-medieval context (745).
Slag by Lynne Keys A total of almost 46 kg of slag was recovered during evaluation and excavation. The slag had not been washed, making identification sometimes difficult, particularly since much of the assemblage consisted of broken fragments. For assessment the slag was visually examined and categorised on the basis of morphology and colour. Each category of slag within each context was weighed. Additionally the smithing hearth bottoms were individually weighed and measured to obtain their length, breadth and depth (Table 4).
Table 4. Smithing hearth bottoms.
Weight (g) Length (mm) Width (mm) Depth (mm)
Range 58­1,234 55­135 45­110 20­85
Mean 368 87 67 43
Std. deviation 290 20 15 15
Roman When the material from Roman levels was assessed it became apparent that, although some smelting slag was present, most of the slag was probably generated by smithing activity. The largest smithing hearth bottoms from the site were from contexts dated to the Roman period and some of the other large examples found in medieval levels could well be redeposited from Roman levels. The smelting slag was very fragmentary, indicating it had been broken up before deposition.
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 163 Within evaluation trench 3 context (136) contained a lot of slag as well as some fragments of hearth superstructure. The amount of hammerscale was small. Context (147), a build-up or deposit dated to the 3rd century, contained three smithing hearth bottoms, smithing slag lumps, some tap slag, and undiagnostic iron slag. Roman road surfaces (135) and (172) comprised in part slag. The latter contained three smithing hearth bottoms, smithing slag, some tap slag and undiagnostic iron slag. It is probable that much of the rest of the assemblage from the Roman period may have been deposited as road or surface dumps to give a hard and compact surface. Within the excavation trench context (685) was a tip of industrial waste related to a furnace. Since the only slag present was a smithing hearth bottom, it may be a smithing hearth or just a piece of slag thrown into a cut with other burnt material. Context (689) contained an iron lump, which may be part of a bloom or another iron piece being worked. Saxon, Medieval and Later More smelting slag was present in these periods. The later the dating becomes, however, the more difficult it is to distinguish what may genuinely be contemporary and what may be redeposited from earlier levels on site or brought in from elsewhere. A piece of slag in context (1066) was very similar to blast-furnace slag but the context was dated to the 12th­13th century. The tanning pits produced, in relative terms, a significant amount of slag. It could represent redeposition from deposits cut through when the pits were dug or be an indication that iron working took place near by and that metal workers used disused tanning pits to dispose of their slag. Conclusion Much of the slag was undiagnostic of either smelting or smithing, but slags diagnostic of both processes were present in contexts dated to the Roman period and onwards. The smelting slag, where present, was in relatively small pieces. Either it was broken up for dumping (Roman road surface) or it had been moved about and damaged before being dumped at the Tanners' Hall site. All the slag was generally redeposited in cuts and dumps. None came from any features which could be interpreted as smelting furnaces or from any building which might be interpreted as a smithy. Worked Stone by Fiona Roe The stone assemblage consists of 6 objects and 36 pieces of building stone. The objects, from both Roman and later contexts, comprise 3 quern fragments, 1 whetstone fragment and 2 hollowed pieces of Lias. The building stone consists mainly of pieces of roofing tile, but includes a segment from a Roman column, part of a carved Saxon cross, part of a possible medieval tombstone and 4 fragments of moulding probably from a medieval window. Wide use was made of local Jurassic limestone, six different varieties being used for structural stone, and three varieties of imported sandstone were used for objects and roofing tile. Some slate was also imported. Objects Roman Only one object came from a Roman context, a 3rd-century ditch fill (519). An unshaped piece of Lias with a basin-shaped hollowed depression, it seems best explained as a casually made mortar. A similarly worked piece of Lias from medieval context (1076) may perhaps also be of Roman origin. Another possibly deliberately hollowed Lias fragment was found in a Roman context at
164
JO VALLENDER
Upper Quay Street, Gloucester (Roe unpublished a). A large stone with a basin-shaped hollow was reported from Vineyards Farm, Charlton Kings (Rawes 1991, 55), and others that seem to be of the same general type have been reported from Frocester (Price 2000, 131, fig. 7.1, 4 and 5). Such casually made items may be relatively common.
Medieval A rotary quern fragment (1076) is made from May Hill sandstone. The piece is very similar to two fitting fragments used as packing in undated posthole (1074). May Hill sandstone was used in Gloucestershire for saddle querns during the Iron Age, as for example at Gilder's Paddock, Bishop's Cleeve (Roe 1999, 109). During the Roman period Upper Old Red Sandstone was preferred locally as a quern material, but the use of May Hill sandstone seems to have revived during the medieval period. One such medieval rotary quern was found at Moreton-in-Marsh (Langton et al. 2000, 21), while another came from a medieval context at Stoke Road, Bishop's Cleeve (Enright and Watts 2002). Further finds from Gloucester, in Longsmith Street (Gloucester Museum) and Upper Quay Street (Roe unpublished a), are of unknown date but seem likely to be Saxon or medieval rather than Roman.
Post Medieval A whetstone fragment (844) is from a long, tapered form with an oval cross-section, a distinctive post-medieval variety. It appears to be made from a variety of Upper Greensand known as Devonshire Batts found near Honiton on the Devon­Somerset boundary (Ussher 1906, 11). This material is known to have been quarried even more recently than the 18th­19th-century date range implied by the other finds from the same context (Moore 1978, 62). Another possible example of a whetstone made from Devonshire Batts came from an 18th-century context at Lower Quay Street, Gloucester (Roe unpublished b).
Building Stone Roofing Materials The building stone from Tanners' Hall is typical of finds from sites in Gloucester. Much of it consists of relatively small fragments of roofing tile. Lower Old Red Sandstone from the Forest of Dean occurs in deposits of all periods, although only one of 14 pieces came from a Roman context, (628). The use of Old Red Sandstone as a Roman roofing material has been recorded from elsewhere in Gloucester, as for example the Bon Marchй site (Hunter 1963, 61), the east gate (Heighway 1983a, 213 and fig. 121) and, more recently, Lower Quay Street (Roe unpublished b). It is not possible to know to what extent the finds of Old Red Sandstone tilestone from later contexts at Tanners' Hall may be redeposited Roman material. It seems quite likely that roofing tiles were also brought by boat from the Forest of Dean during the medieval period. Old Red Sandstone tiles of medieval type have been found at Woolaston Grange Quay (Fulford et al. 1992, 114), destined presumably for places that could well have included Gloucester. Roofing tiles made of Cotswold stone appear consistently in Gloucester from medieval times onwards. There are 11 fragments from Tanners' Hall, the earliest being from an 11th­12thcentury context, (1141). The majority of these tiles were made from a shell-fragmental limestone, such as that found at High Brotheridge, near Cranham, only 7.4 km (4.6 miles) south-east of Gloucester (A. Price pers. comm.). A few tiles made from a sandier tilestone (e.g. 1020) may have been obtained around Througham, near Bisley, some 13.5 km (8.6 miles) south-east of Gloucester. There tilestone from the same stratigraphical position as the Stonesfield Slate series
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 165 was still quarried early in the 20th century (Richardson 1972, 160). Similar Cotswold tilestone has been recorded from medieval contexts elsewhere in Gloucester, dating from the 13th century at the east gate (Vince 1983, 214) and from the 12th to 13th century at Lower Quay Street (Roe unpublished b). Structural Stone The only piece of Roman building stone, a segment from a small column, came from a medieval context, (1081). It is made from oolitic freestone of Painswick type, a well-known building stone which forms part of the Lower Freestone (Richardson 1972, 82). The column fragment has a diameter of 170 mm and thus falls within the size range for porch and portico columns suggested by Blagg (2002, 189). In later Roman Gloucester many buildings seem to have had colonnades, perhaps partly because of the availability of good freestone. They are known, for example, from the Bon Marchй site (Hunter 1963, 36), the New Market Hall (Hassall and Rhodes 1974, 26), the east gate (Heighway 1983a, 221 and fig. 124), and perhaps also under the Norman castle (Darvill 1988, 7) and Upper Quay Street (Roe unpublished a). At Tanners' Hall the excavations of 1978­80 produced part of a limestone column capital (Heighway 1983b, 102 and fig. 10, 49). A fairly highstatus Roman building in the vicinity would seem to be indicated. Medieval contexts (857 and 894) also produced fragments of a Saxon cross with carved ornament (below). Four ornamented Saxon cross shafts were found near by at St Oswald's priory (Heighway and Bryant 1999, 154). The limestone used for this fifth cross is light coloured, oolitic freestone, similar to that used for the Roman column described above. It is possible that this cross was made from Roman masonry, which would have been readily available, perhaps in the form of ashlar blocks used for walling (e.g. Fullbrook-Leggatt 1968, 16 and illus). However, such stone would have been `dead', making it difficult to carve to such a high quality (R. Bryant pers. comm.). A harder variety of local Jurassic limestone, shelly with some oolitic grains, was used for part of a possible tombstone from an 11th­13th-century context (1108). Pieces of window tracery from a post-medieval context (946) probably came from a medieval window, quite possibly from Tanners' Hall itself. The limestone used for these was again of Painswick type (A. Price pers. comm.), oolitic but with higher shell content than the pieces described above. The walls of Tanners' Hall are mainly of Lower Lias. The one Lias fragment in this assemblage, from a postmedieval context (627), represents the sixth example of local Jurassic limestone in use here over the centuries. Acknowledgements I am very grateful to Arthur Price for his most helpful advice on the building stone found at Tanners' Hall. Thanks are also given to Roger Howell for his assistance with the fieldwork. Anglo-Saxon Cross Shaft by Richard Bryant A fragment of an Anglo-Saxon cross shaft (Fig. 13) was recovered from a 12th­14th century posthole, (893), during the excavation. It was originally part of a freestanding cross with two broad faces (A and C) and two narrow faces (B and D). The bottom of the fragment is flat and is almost certainly the surface joined to the lower part of the cross. The piece, of which the top section is broken obliquely, has a maximum height of 255 mm. The surviving part of face A tapers from 195 mm to c.170 mm, while Face C tapers from 180 mm to c.157mm. Face D is 211 mm wide and straight-sided. Face B is missing.
166
JO VALLENDER
The fragment is made of a fine-grained oolitic limestone and has broken along the line of a vertical drilled fixing-hole or dowel-hole 30 to 35 mm in diameter. A downward sloping hole 25 mm in diameter has been drilled through from Face C to meet the vertical hole and was presumably intended for pouring lead in and around a metal fixing dowel located in the vertical hole. A similar arrangement was observed in two of the standing crosses (33 and 34) and one of the decorated arch fragments (stone 46) from St Oswald's priory (Heighway and Bryant 1999, 161, 162, 171; figs. 4.10, 4.11 and 4.20). Each face of the shaft is edged with a flat inner moulding and an outer cable moulding. Face A is dominated by a large, finely carved, human figure in a frontal position. The top part of the head is missing in a line from just above the right ear to below the left ear. The break line actually passes through the right eye and the remains of a drilled pupil are visible. Most of the nose has been lost, but the outline of the right nostril and the right side of the nose survive. The upper lip is well defined, as is the slightly open mouth that turns down at the corners. There is no beard or moustache, but the chin is fairly sharp and the cheeks rounded. Just below the break on the right side of the face, behind part of the figure's ear or hair, there is a slightly raised area with a curving outer edge. The rest of the background is cut back into pronounced vertical ridges. The figure wears a round-necked tunic and an over-garment. The tunic is decorated with pelleting or beading at the neck and vertical panels at the front. The over-garment is draped across the right shoulder in well-modelled folds. A median-incised fold or border runs diagonally across the right side of the figure from shoulder to waist. About 65 mm below the shoulder, an area of the folds of the over-garment has been flattened by chisel or adze blows. The shoulder itself has been cut by two deep, `V'-shaped and slightly curving grooves. Face C is decorated with a median-incised tree scroll, broken along the central vertical stem. Each volute terminates in an upward-pointing veined leaf. Lobed leaves or fruit spring from the junctions between the volutes and the central stem and overlap each volute just below the leaf terminal. Face D carries two volutes of a simple, median-incised scroll. Each volute terminates in a veined drop leaf. A similar leaf drops from the outer edge of the upper volute, and part of another drops
Fig. 13. Sculpture on cross shaft (scale 1:4).
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 167 from the outer edge of the lower volute. From the node between the two volutes grows a bud on a stiff stalk. Discussion The Tanners' Hall cross shaft is one of seven Anglo-Saxon stone cross shafts from Gloucester. Four were from the site of St Oswald's priory and one from Wotton Pitch on Ermin Street about a kilometre north-east of the city centre. The other fragment was recovered from 10th-century destruction levels during excavations at St Mary de Lode (Bryant and Heighway 2003). The Wotton cross dates from the 9th century, and it has been suggested that it could be associated with the hundredal meeting place for Dudstone hundred (Heighway 1980, 61). The four crosses from St Oswald's are of late 8th- and 9th-century date. There is some evidence to suggest that they stood together in a high-status area, possibly a cemetery, on the banks of the river Severn on or near the site of the later priory (Bryant in Heighway and Bryant 1999, 155). The Tanners' Hall cross shaft was found in a posthole dated to the 12th­14th century. It was almost certainly brought to the site as building stone or levelling material. Its original provenance is unknown, but it may have come from St Oswald's priory or from St Peter's abbey, sites that are equidistant from the Tanners' Hall site. The cable moulding that forms the outer border for each face of the cross is wider and flatter than that on cross 34 from St Oswald's and on the Wotton shaft. It is more like the border on the early 9th-century Lechmere stone possibly from Hanley Castle in Worcestershire (Webster and Backhouse 1991, 245). It is also similar to the border on the late 8th- or early 9th-century cross 3 from Hexham (Cramp 1984, pls. 174 and 175). Part of an early 10th-century arch (stone 44), reused in a late 10th-century foundation at St Oswald's, is decorated with oval pelleting and a flat cable but the sides of each twist of the cable are concave rather than straight as on the Tanners' Hall stone (Heighway and Bryant 1999, 170­1; figs. 4.20 and 4.21). The scroll on Face D and the tree scroll on Face C are typical of 8th- and early 9th-century carving. Although it is slightly unusual to find leaf forms used as the central terminal of the volutes, the earliest cross (stone 30/31) from the St Oswald group also uses terminal leaves, albeit in a rather more adventurous way (Heighway and Bryant 1999, 156­7). Among other examples are those recorded at Ramsbury and Britford, both in Wiltshire (Kendrick 1938, pls. lxxvi and xcix). The lobed leaves and veined leaves are found widely throughout Northumbria, the centre from which many of the new ideas spread during the 8th and the 9th centuries. Examples can be found at Rothbury 1, Hexham 3, Lowther, Bewcastle, Ruthwell, Jedburgh and elsewhere (Cramp 1984, 177, 217­19, pls. 174­5, 211­15, 264, 265; Bailey and Cramp 1988, 61­72, 127­9, illus. 90­117, 119, 427­43, 465­8). More locally these elements occur on the Lechmere stone and on the early 9thcentury cross head from Cropthorne, also in Worcestershire. They do not occur on any of the other Gloucester crosses. The bud on a stiff stalk, that grows from the dividing point of the two volutes on Face D of the Tanners' Hall cross shaft, can be paralleled in Hexham 3 and also the late 8th- or early 9th-century cross at Simonburn (Northumberland) (Cramp 1984, pl. 219). The most unusual aspect of the Tanners' Hall cross is the human figure of Face A. In Gloucester none of the St Oswald's cross shafts display human figures. There is a figure on the Wotton cross shaft, but it is of a more stiff and stylised type. At Gloucester cathedral a panel bears a roundel in the centre of which is a badly damaged figure of Christ (Talbot Rice 1952, 98­100). Apart from this, the nearest Anglo-Saxon carvings of similar quality are the full-length figures on the Lypiatt cross shaft, which stands near the Cotswold village of Bisley, about 17 km (10 miles) south-east of Gloucester (Bryant 1990). Not enough survives of the Tanners' Hall figure to assess whether it was full length or not, but it is to the same scale as the Lypiatt figures which are about 800 mm
168
JO VALLENDER
high. The figures on the crosses from Ruthwell and Bewcastle are similar in height. Half-length figures or `portrait' busts are fairly widespread in Mercia. Examples include the figure of the Virgin from Bredon (Cramp 1977, 210, fig. 58), the Angel from Deerhurst (Rahtz and Watts 1997, 43­4, figs. 29 and 30, table viii), the damaged Christ from Gloucester cathedral, and a figure from St Andrew's, Pershore (King 1992, 129­34). However, the treatment of the face on the Tanners' Hall cross shaft is most similar to that of Christ from the top of the shaft on Rothbury 1 (Cramp and Miket 1982, cover and pl. 11; Cramp 1984, pls. 213 and 214). The Tanners' Hall figure's features are rounded and youthful, with a slightly pointed chin and no beard or moustache. The mouth is firm and slightly open. There are traces of the modelling of the lower part of the right eye, and the pupil is drilled. The curving edge to the carving behind the right ear may indicate that the figure had a small halo, rather tight to the head. However, in most contemporary representations the halo rises from the shoulders and is much larger. The curving edge may, therefore, simply be part of the figure's hair. In this case the figure should probably be interpreted not as Christ but perhaps as a saint. The Virgin on the Bredon stone does not have a halo. A very worn figure on a cross shaft from Hackness in east Yorkshire also has no halo, and could represent Abbess Oedilburga (who might be equated with St Ethelburga) (Lang 1991, 138­40, illus. 454). This cross has been given a late 7th­early 9th-century date. On the Lypiatt cross one figure (D) has no halo but is interpreted as one of the Evangelists (Bryant 1990, 42, figs. 1 and 2). A late 8th-century panel from Bredon contains a pair of male figures without halos, one of whom carries a book in his left hand (Cramp 1977, fig. 55). A late 8th-century date has been suggested for much of the Bredon material (Dominic Tweddle in Webster and Backhouse 1991, 239). On the Bewcastle cross the figure normally identified as John the Baptist has no halo, nor does the figure with a bird who might be St John the Evangelist (Bailey and Cramp 1988, 68­70, illus. 92, 94 and 95). At Ruthwell the figures of Mary and Martha do not have halos and neither do St Paul and St Anthony (Cassidy 1992, illus. 24, 39). Several of the figures on the base of the late 8th- or early 9th-century cross shaft at Aukland St Andrew in Durham wear garments similar to those on the Tanners' Hall figure (Cramp 1984, 37­40, pls. 1, 2, 4 and 5). However, the closest parallel for the clothes occurs on a portrait of St John Chrysostom that is inserted into an early 9th-century Gospel Book in Salzburg (Hubert, Porcher and Volbach 1970, 180). St John, archbishop of Constantinople from 398 to 404, wears an overgarment and a tunic that has vertical panels across the front. The over-garment, presumably a chasuble, falls from the shoulders in heavy fold and has a deep `V'-shaped opening at the front, bordered by two, narrow, parallel bands. This may be intended to indicate the `Y'-shaped pallium of an archbishop or pope. The pallium is shown more clearly on a mosaic portrait of Pope Paschal I (817­24) in the apse of Santa Prassede in Rome (Backes and Dolling 1969, 60, 61). On the mosaic Pope Paschal has the square nimbus that denotes a living person. It would be possible to interpret the carved vertical ridges behind the Tanners' Hall figure in a similar way and to suggest that the figure depicted was a living saint. It is, however, more likely that the carving is intended to indicate curtains or drapes behind the figure. It is equally possible that the deep, rather coarse, `V'-shaped cuts were not part of the original design. They could have cut away the background and even, perhaps, removed part of a larger halo. Certainly the curving `V'-shaped cuts across the figure's shoulder are later, and they make little sense in relation to the folds of the over-garment. The chisel or axe damage to the front of the folds could have occurred after the cross shaft had been broken up, but the cuts into the shoulder look deliberate. They may have been made to remove something, but the cuts create an oval depression with a raised centre and may equally have been carved to form the seating for something like a shoulder brooch. Could this be intended for the display of an object associated with the saint depicted on the cross?
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 169 Dating and Provenance The style and quality of the Tanners' Hall cross shaft suggest that the fragment is part of a late 8th- to early 9th-century freestanding cross. The fact that all of the elements of its design are different from those of the St Oswald's priory crosses might indicate that a different patron commissioned the Tanners' Hall cross. In Gloucester another major patron must have been St Peter's abbey (now Gloucester cathedral). In the 9th century, the abbey commissioned the roundel depicting Christ mentioned above. This is the only other figure carving of similar quality actually from Gloucester. Perhaps, therefore, the Tanners' Hall cross also originally stood in the abbey. Other Finds A total of 389 ferrous objects were recovered from the site. The majority were highly corroded nails but thirty-eight items were considered of sufficient interest to be recorded as registered finds. They included a collection of knife blades, none of which could be specifically associated with the tanning process. Two hundred and seventeen iron nails were collected from the excavation and 134 from the evaluation. These were common to all periods and were mostly heavily corroded and fragile. Fifteen bone pins were recovered. None showed signs of any fine carving and two (49 and 144) were more probably tools. The 21 coins recovered from the site all appear to be Roman in date. They are scattered through deposits of all periods highlighting the highly disturbed nature of the stratigraphy across the area of both the evaluation and the excavation. A considerable number of copper-alloy pins and the wire used to make them were recovered. The majority of the pins were found during the evaluation and were possibly associated with a structure on the southern side of Upper Tanners' Yard within the area of evaluation trench 1. It was thought that a post-medieval pin factory existed in the immediate area. A total of 239 fragments of glass was recovered. Eight whole vessels from the semi-basement rooms within Worcester Lawn were dated to the 18th century. The rest of the assemblage comprises window and vessel fragments. Five fragments of painted wall plaster and three tesserae were collected from the site. Colours recognisable on the plaster were dark red; bright orange with dark red painted over it (possibly part of a pattern); dark reddish brown with a dark yellow line c. 5 mm wide painted over the red; and white and pale pink (which may be faded red). However, none came from dated Roman contexts and it is unclear whether this material originated on the site or was imported during levelling. THE ENVIRONMENTAL EVIDENCE The Animal Bone by Naomi Sykes The animal bone was assessed in 1999 by the Centre for Human Ecology and Environment, University of Southampton. It was determined that small sample sizes and poor dating limited analysis of the Romano-British and early medieval assemblages. Thus in this report the material from these periods is discussed only briefly. By contrast it was felt that the substantial numbers of horn-cores recovered from the 13th­14th- and, in particular, 16th­17th-century deposits warranted additional investigation and these are accordingly examined in more detail. The horn-cores were
170
JO VALLENDER
analysed and recorded using Serjeantson's (1996) `zones' system. Specimens were identified to species, with sheep and goat being differentiated using Boessneck's (1969) criteria. Cattle horncores were classified tentatively to `type', age and gender following the systems of Armitage and Clutton-Brock (1976), and Armitage (1982). Measurements were taken in accordance with the standards set by von den Driesch (1976). Specimens that showed signs of butchery were recorded using Lauwerier's (1988) system.
Taphonomy Whilst much of the bone material was well preserved (37% being identifiable), its surface condition was poor, perhaps explaining the lack of apparent gnawing and butchery marks. These findings are borne out by the horn-core study, which recorded a low incidence of taphonomic indicators (Table 5). The absence of gnawing marks on the horn-cores is to be expected since these elements are of low palatability to dogs and their deposition in pits would have made them less accessible to scavengers. The lack of cut and chop marks is, however, surprising since commonly other horncore assemblages, such as that from Greyfriars in Chichester (Armitage 1993), show high frequencies of marks associated with skinning and horn detachment. Their dearth may be related to the specimens' surface condition but the few cut marks present were well defined, suggesting that their low incidence was actual. Horns can be removed without the use of a knife if they are first left to rot. This breaks the bond between the sheath and core allowing the sheath simply to be pulled away (Prummel 1978, 43). It seems possible that this practice was utilised at Tanners' Hall, thereby accounting for the lack of butchery marks.
Table 5. Incidence of butchery marks: codes (in parentheses) after Lauwerier (1988).
Date 13th­14th century 16th­17th century
Taxa Cattle Goat Sheep Cattle
% Butchered Poleaxe (1) Chop (3) 22
Cut (6) 25
4
0.6
Total (NISP) 9 4 1 162
Taxa Representation The relative frequencies of the different taxa suggested by the 1999 assessment are presented in Table 6. It must be remembered that the material was not subject to full investigation and, therefore, the data should be viewed tentatively. Despite this, and the small sample sizes, the results provide an indication of the way in which Gloucester's animal economy developed between the Romano-British and post-medieval period. In all the phases cattle, sheep and pig are clearly the most represented taxa. Horse and dog are occasionally present in low frequencies. Other animals, such as birds, fish and small mammals, are not represented by more than a few fragments each. Inter-period variation in the main domesticates' frequencies is consistent with both regional and national trends. For instance, the Romano-British material is dominated by cattle remains, a situation typical of assemblages dating between the 2nd and 4th century (King 1978). Indeed other sites from the Gloucester area, such as Denmark Road (Ayres and Clark 1995; Powell and Clark 1996) and Westgate (Maltby 1979), all demonstrated a preponderance of cattle. Originally this widespread reliance on cattle was interpreted as evidence for Roman dietary preferences (King
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 171 1991) but now it is thought to represent the wider agricultural regime, with cattle being managed for their secondary uses, in particular for traction (Hamshaw-Thomas 2000). Cessation of occupation between the 5th and the 10th century means that the Tanners' Hall deposits provide no evidence of animal management practices for early medieval Gloucester. The late Saxon bone recovered indicated that cattle continued to be the dominant animal. Little confidence should, however, be placed in such a small sample, especially when the material is viewed in a wider context: whilst cattle are a feature of late Saxon assemblages from towns in the north and east of England, such as Lincoln (Dobney et al. 1995; O'Connor 1982) and York (O'Connor 1989), sheep are generally more numerous in assemblages from the south and west of the country. Certainly sheep fragments outnumbered those of cattle in the late Saxon deposits from both Westgate (Maltby 1979) and Winchcombe (Levitan 1985). According to the Tanners' Hall material, sheep/goats gradually attained higher relative frequencies, becoming the dominant taxon in the 13th­14th-century deposits (Fig. 14). Again this rise is in accordance with national trends, which suggest that through the medieval period increasingly large flocks of sheep were maintained to manure arable fields and supply wool for the growing textile industry. The evident decline in pig representation is a widely recognised pattern commonly attributed to the growth of arable farming. As more land was put to the plough, availability of woodland pannage, and thus the ability to manage large populations of pigs, was reduced (Grant 1988; Noddle 1975). Small sample sizes for the 15th­16th-century assemblages render any calculations of taxa abundance meaningless. Evidence pertaining to taxa ratios is equally scarce for the 16th­17thcentury material because, although the assemblage contains substantial quantities of identifiable specimens, it is composed almost entirely of cattle horn-cores, skull fragments and limb extremities, a selectivity of taxa and element type that is indicative of a craft deposit. Serjeantson (1989) has shown that it is often difficult to discern whether such assemblages represent accumulations of
60
50 40 % 30
cat t le sheep pig
20
10
0 10th-11th
12th-13th
13th-14th
Fig. 14. Interperiod variation in the relative frequencies of the main domesticated species (medieval assemblages with NISP of > 100 only).
172
JO VALLENDER
butchery, tanning or horning waste but, in this instance, the contextual evidence leaves no doubt that the horn-core deposits are associated with the tanning industry. Until recently skins were sent to tanners with the heads and feet of the animals attached, a practice that can be dated back to the mid 11th century (Shaw 1996; Sykes n.d). The reasons for this are uncertain but Schmid (1973) has suggested that heads and feet may have been retained either because their weight stopped the hide from shrinking or because they enabled the tanner to ascertain the age of the animal. Whilst tannery assemblages do not reflect the wider animal economy, potentially they can provide substantial information concerning the age, gender and conformation of the animals utilised by the industry. The Tanners' Hall horn-core assemblage therefore allows an examination of the types of cattle present in the town's hinterland, an opportunity made more significant by the fact that no other deposits of this kind have, to date, been recovered from medieval or postmedieval Gloucester.
Horn-Core Study In total 176 horn-cores were examined (Table 7). The majority came from the 16th­17th-century levels, a small number from contexts dating to the 13th­14th century. Whilst the post-medieval assemblage was composed exclusively of cattle horn-cores, both sheep and goat were represented in the earlier deposits. This variation seems to form part of a wider trend. Most horn-core assemblages from early medieval sites, such as Thetford (Jones 1984), Norwich (Jones 1994) and Exe Bridge, Devon (Levitan 1987), contain a mixture of cattle, sheep and, in particular, goat specimens, whereas those from later periods tend to be more species specific: assemblages from Greyfriars, Chichester (Armitage 1993), St Albans, Hertfordshire (Saunders 1977), and The Green, Northamptonshire (Harman 1996), were composed entirely of cattle horn-cores. Reasons for this variation are uncertain but it seems possible that, in the earlier medieval period, craft activities were less specialised, with tanners dealing with all types of animals. By the later period there were two types of tanners (Serjeantson 1989,129): those who dealt with large `hides' (from cattle and horses) and those who processed `skins' from smaller animals (sheep, goats and calves). If by the post-medieval period Tanners' Hall had become a `hide' tannery this may account for the absence of caprine remains. It might be expected that the development of specialised `hide' tanning would also have impacted upon the cattle age structure, with juvenile individuals becoming less well represented since their `skins' would be sent to a different location. In neither the 13th­14th- nor the 16th­17th-century assemblages were infant animals (those less than 1 year of age: Stage 0) present (Table 8). To some extent this is to be expected for the medieval material, since in this period the need for plough animals meant that few cattle were slaughtered young. It is widely accepted, however, that in the 16th and the 17th centuries there was a nation-wide change in cattle husbandry, with a drop in the average slaughter age and, in particular, greater numbers of calves being sent for slaughter (Albarella and Davis 1996; O'Connor 1993). Table 8 demonstrates a post-medieval decline in slaughter age but shows no evidence of very young individuals; indeed, even animals between approximately 1 and 2 years of age (Stage 1) are poorly represented. This supports the idea that calf skins were not sent to Tanners' Hall during the post-medieval period. Instead, most (92%) of the animals represented in the post-medieval deposits were aged between 2 and 7 years (Stages 2­4). A similar pattern was apparent for the 16th-century horn-cores from Greyfriars, Chichester, where all of the specimens were classified to that age range (Armitage 1993). The material from the 13th­14th-century deposits at Tanners' Hall, although a smaller sample, indicates a wider distribution of ages. Again this may be evidence that the tanning industry became more specialised in the later period.
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 173
Table 6. A: taxa representation (NISP); B: relative frequencies of the main domesticates for sample with a NISP >100.
A
Romano-British
Medieval
Post Medieval
Total
1 2 3 Late Saxon 10th­11th 11th­12th 12th­13th 13th­14th 14th­15th 15th­16th 16th­17th 17th­18th
Cattle
1 9 39
10
47
26
62
67
11
14
392
22
700
Sheep/goat
2
13
3
46
14
63
81
18
8
11
259
Pig
124
13
5
5
5
35
Horse
2
1
3
1
1
1
9
Dog
2
2
4
Galliform
4
2
1
4
1
2
1
15
Other bird
3
6
4
13
Fish
1
1
Other
2
4
1
1
1
1
10
Unidentified
232
239
66
40
531
110
1,784
Total
13 37 308
41
282
138
371
403
98
No. Identifiable 4 11 62
13
114
49
139
164
32
% Identifiable 30 30 20
32
40
36
37
41
33
65
941
133
2,830
25
410
23
1,064
38
44
17
38
B Total (NISP) % Cattle % Sheep/goat % Pig
10th­11th 106 44 43 12
12th­13th 130 48 49 3
13th­14th 153 44 53 3
16th­17th 403 97 3
174
JO VALLENDER
Table 7. Taxa represented in the 13th­14th- and 16th­17th-century horn-core assemblages.
Period 13th­14th century 16th­17th century Total
Cattle 9 162 171
Goat 4 4
Sheep 1 1
Total (NISP) 14 162 176
Table 8. Age distribution of the 13th­14th- and 16th­17th-century cattle horn-cores from Tanners' Hall: age groups follow Armitage (1982).
Period 13th­14th century 16th­17th century %
Age Group
Total
1
2
3
4
5 (NISP)
1
1
3
3
1
9
11
55
60
34
2
162
7 33.9 37 20.9 1.2
Having ascertained age estimations for all of the horn-core specimens, it is possible to consider their size and shape as evidence pertaining to the type and gender of the individuals represented. Because horn-core morphometry varies greatly with age, metrical analysis was undertaken only for specimens from adult animals: for cattle those that had attained Stage 3 (Armitage's 1982 `Young Adult' group with a suggested age range of 3­7 years) and beyond and for sheep/goat those that exhibited compact bone. The single sheep horn-core was from a juvenile animal and thus could not be classified to gender. One of the goat specimens was also sub-adult but the other three were adult. Goat horn-cores are highly sexually dimorphic and, even in a group of mixed genotype, their basal measurements often form separate male and female clusters. Because of this the Tanners' Hall goats can be compared to those from other similarly dated sites, with a high possibility of ascertaining their gender. The 13th-century assemblage from Exe Bridge (Levitan 1987) produced a substantial number (442) of goat horn-cores which formed two discrete groups when their basal measurements were plotted: those with a greatest basal diameter (BA) measuring between 20 and 35 mm (females) and those measuring between 40 and 60 mm (males). All three adult goat horncores fall firmly within the female range. This may indicate that nanny goat skins were preferentially imported or were simply more available than those from billy goats, but from such a small sample this cannot be claimed absolutely. Certainly the horn-cores recovered from 12thcentury Skeldergate, York, demonstrated no such selection, with males and females represented in roughly equal numbers (O'Connor 1984). Determining gender from cattle horn-cores presents a greater problem. Grigson (1974, 1975 and 1976) has demonstrated that whilst their size and shape vary with sex, those features are also influenced by breed. As a result classification of archaeological horn-cores to males, females or castrates cannot be undertaken without first attempting to sort the specimens into `type' groups. Armitage and Clutton-Brock (1976) developed a system whereby the length of the outer curve (OC) was used to place horn-cores into one of four categories: labelling those measuring less than 96 mm `small horns', those between 96 and 150 mm `short horns', those between 150 and 220 mm `medium horns' and those over 200 mm `long horns'. Figure 15 shows the 13th­14th-century horn-cores from Tanners' Hall plotted against those from the post-medieval deposits, placed in Armitage and Clutton-Brock's cattle types. All of the
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 175 earlier specimens fall within the `short horn' range, whereas the post-medieval sample demonstrated the presence of all four types. This inter-period variation may again reflect the development of the tanning industry. In the earlier period animals may have been drawn from the local area, thereby accounting for the lack of type-based variation. By the 16th­17th century, if the industry at Tanners' Hall had indeed become more intensive, the catchment area may have increased, with skins from various cattle types being imported from a wider geographical region. Alternatively, the wide variety in the post-medieval horn-core types may reflect the 16th­17thcentury advances in animal husbandry, which gave rise to many new breeds of cattle. Because of these different breeds Armitage (1993, 216) has admitted that his original classification system (Armitage and Clutton-Brock 1976; Armitage 1982) is inadequate to deal with post-medieval assemblages and has stressed that a revised typology is required. Indeed, in his 1993 study he relabelled many of his original groups. However, for the purpose of this study, the 1976 categories will be used. With all of the Tanners' Hall cattle specimens having been classified to type, appraisal of their gender can be undertaken. Visual discrimination was carried out at the first stage of recording, using Armitage and Clutton-Brock's (1976) criteria. The results, presented in Table 9, appear unsatisfactory. For example, they suggest that bulls were more numerous than oxen. This is highly unlikely in terms of herd management, since it is widely accepted that, if not slaughtered young, most males were castrated. Although not entirely analogous, Hatting (1975) demonstrated that in sheep the effects of castration depend on the age at which it occurs: if rams are castrated within a few months of birth their horn-cores become similar to those of ewes, but if castration takes place later horn-core growth is less affected. Other studies of sheep horn-cores have shown how variable these elements are, being highly susceptible to environmental factors, and that horn-core shape
230
210
190
BC (mm)
170
small
150
13-14th
short 130 medium
110
long ?
90
70
50 50
100
150
200
250
OC (mm)
Fig. 15. Scatterplot (OC by BC) for 13th­14th- and 16th­17th-century cattle horn-cores.
176
JO VALLENDER
Table 9. Gender composition of the 13th­14th- and 16th­17th-century horn-core assemblages as suggested by Armitage and Clutton-Brock's (1976) criteria.
13th­14th century 16th­17th century
Bull
Cow
2
3
16
41
Ox Uncertain Total
1
3
9
5
100
162
is not always a useful indicator of gender (Sykes n.d). It seems possible that the same is true for cattle and it may, therefore, be difficult to determine gender using purely shape-based criteria. According to Grigson (1975 and 1978), skull and horn-core proportions are more profitable indicators of gender. Armitage and Clutton-Brock (1976) also suggest the size of the basal circumference in proportion to that of the the outer curve as a gauge of gender: the horn-cores of bulls having a large basal circumference (BC) proportional to the outer curve (OC), those of cows having a longer OC to BC, and those of castrates falling mid way between the two. Scatterplots of these two factors seldom produce obvious trimodal or even bimodal distributions suggestive of sex (Fig. 15) but clustering of the data is slightly more apparent when the basal circumference and outer curve measurements are plotted as an index figure (calculated using the equation BC/OC Ч 100). This provides an indication of the specimens' proportions: a high figure represents a specimen whose outer curve is small in proportion to its basal circumference (indicative of males) and a low figure suggests a horn-core that is longer in proportion to its basal circumference (females). Again this type of index is affected by cattle type: long horn cattle will generally have lower index values than small horn cattle. Nevertheless, Figure 16 shows that within each type group there are specimens that fall outside the main cluster, having significantly higher index values. It may be conjectured that these specimens are bulls. Differentiation between cows and oxen is less obvious. The gender composition indicated by Figure 16 fits existing models of medieval and postmedieval husbandry better than the ratios proposed in Table 9. The 13th­14th-century distribution suggests a cow:bull:ox ratio of 3:2:1, as might be expected from a period when most cattle, being used primarily for traction, were raised to adulthood with the males being castrated. The postmedieval move towards dairy production would have seen a greater number of males being slaughtered as calves. Under such a regime the cow:bull:ox ratio of 6:1:1, as shown by the 16th­17th-century Tanners' Hall material, would be fitting for the adult population. Conclusion From the animal bone assemblage it would seem that the Tanners' Hall area developed considerably between the Romano-British and post-medieval periods. The zooarchaeological evidence also supports the notion that before the 13th century occupation of the area was largely domestic. Certainly the taxa representations are similar to those of other urban sites. By the early 13th century the area is reported to have become a tannery, as is attested by the presence of cattle, sheep and goat horn-cores. Originally the scale of production appears to have been localised, with the skins of cattle, sheep and goats of all ages being imported from near by. This idea is supported by the fact that all the cattle horn-cores from the 13th­14th-century deposits are of the short horn type. The post-medieval assemblage indicates a wider variety of cattle types. This, combined with the ageing and taxa representation data suggests that Gloucester's tanning industry was more specialised by the 16th century, with the cattle skins from animals aged between 2 and 7 being imported selectively from a wider hinterland.
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 177
BC (mm)
240 220 200 180 160 140 120 100 80 60 60
Oxen ? Cows ?
Bulls ?
small 13-14th short medium long ?
80
100
120
140
BC/OC Index
Fig. 16. Scatterplot (BC/OC index by OC) for 13th­14th- and 16th­17th-century cattle horn-cores.
Although tanning deposits provide only limited information concerning the greater animal economy, data derived from the Tanners' Hall assemblage fits into accepted models of medieval and post-medieval animal exploitation. For example, the ageing evidence reflects the 16th-century shift towards the slaughter of younger cattle. The wider range of cattle types present in the post-medieval assemblage may also reflect advances in animal husbandry. Determining gender composition of the assemblage was more problematic and results varied depending upon the method used. This demonstrates the inaccuracies of existing classification systems and highlights the need for them to be tested on modern material of known gender and breed. Nevertheless, gender groupings based on metrical analysis seemed more satisfactory than those achieved by visual inspection. They suggest that by the 16th century the frequency of cows increased relative to bulls and oxen, again conforming to known patterns of animal husbandry, in particular the post-medieval increase in dairy production. Plant Remains by Julie Jones Only seven samples were taken from the site for environmental evidence since during the excavation very few primary fills were reached. They produced sparse finds of charred grains of wheat, barley and oats, plus a few weed seeds and charcoal, and are likely to be of a secondary nature and not related to the original function of the pits. Several of the contexts examined produced reasonable quantities of both large and small mammal bones, as well as fish bones, and these were subsequently examined with the rest of the bone assemblage. During the 1995 evaluation nine samples were taken and they were more reliably dated due to the greater depth of stratigraphy. The sample sizes varied from 5.5 kg/4 litres to 31.7 kg/17 litres. These were sieved in a flotation tank, the residues to 500 microns and the float to 250 microns. The residues were dried and scanned
178
Table 10. Charred and waterlogged plant and other remains.
JO VALLENDER
Cess Pit
Context
246
Sample number 13
Sample size
6/
(kg/litres)
3.4
Float size (ml)
50
Charred plant remains
Cereals
Triticum sp (grain)
Wheat
Hordeum sp (grain)
Barley
Avena sp (grain)
Oat
Avena sp (grain-sprouted)
cf. Secale cereale (grain)
Rye
Cereal indet (grain)
Triticum spelta (glume base) Spelt Wheat
Weeds
Bromus sp
Brome
Carex spp
Sedge
Corylus avellana L. (nut frag) Hazel
Eleocharis palustris/uniglumis Spike Rush
Galium aparine L.
Cleavers
Lathyrus Vicia spp
Pea/Vetchling
Raphanus raphanistrum spp
raphanistrum (pod)
Wild Radish
Vicia faba L.
Celtic Bean
Waterlogged plant remains
Brassica Sinapis spp
Mustard/Rape/
Cole etc.
Carex spp
Sedge
Conium maculatum L.
Hemlock
Corylus avellana L.
Hazel
Eleocharis palustris/uniglumis Spike Rush
Ficus carica L.
Fig
Roman
Ditch
Slag deposit
Fill
235
242 189 59
12
14 9 11
5.85/
12.7/ 23/ 31.7/
4.3
6.2 12 17
93
7
30
9
2
1
1
1
1
1
1
1 1
1
1
1
1
1
1
1
4
2
1
1
2
2
2
Medieval
Tanning Pits
Ha
179 132 169 213
8
6
7 10
5.5/ 27.1/ 27.6/ 28.4/
4 23 16 25
40 44 14 22
2
4
4
1
3
2
3
2
1
1
2
3
1 11 1 2 1
C
G
H
M
C
2
D
C #
C G B H M #
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 179
Fumaria sp
Fumitory
Hyoscyamus niger L.
Henbane
Montia spp
Blinks
Ranunculus acris/repens/bulbosu Buttercup
1
Ranunculus sardous Crantz
Hairy Buttercup
1
Raphanus raphanistrum spp
Wild Radish
1
raphanistrum(pod frag) frag)
Sambucus nigra L.
Elder
1
Urtica dioica L
Common nettle
2
Valerianella dentata (L.)Pollich Narrow-fruited Cornsalad
Other remains
Charcoal (approx. no. frags >2 mm overall
150
500
dimensions)
Animal bone (approx no. frags) 5
6
Fish bone/scales
F
Shellfish
f
f
Pottery
7f
9f
Tile
Glass
Slag
Metal objects (?nails)
2 1 1 1 f 3 1 20 100 40 10 <10 2f Ab Ab 1
C B W DG C C
4
1+f 1
D
25
D
C
<20 >200 C20 100
150 100 100
Frq F
FF
2f 40f 20f 30f
6f 5f 10f
2f 2f 4f
p
p
1
9
4
Key F = Few: 1; Frq = Frequent: 2­10; Ab = Abundant: 10+; f = fragments; p = present Habitats (Ha): B: bankside; C: cultivated/arable; D: disturbed; G: grassland; H: hedgerow; M: marsh; W: woodland # = cultivated plant/of economic importance
180
JO VALLENDER
and the finds within them were quantified. The floats produced both charred and waterlogged plant macrofossils but numbers recovered were low. The results are shown in Table 10 with the plant macrofossils quantified on a scale of abundance. Nomenclature and habitat information follow Stace (1991).
Roman Two of the samples were recovered from the lower (246) and upper fill (235) of a cess pit. The floats from both contexts were small and consisted primarily of charcoal fragments. The only waterlogged seed recovered from (246) was a single elder (Sambucus nigra), although, as well as the charcoal fragments, a limited number of animal and fish bones and shellfish fragments were noted. Context (235) produced a limited number of waterlogged weeds of different habitats, among them damp ground species such as sedge (Carex spp) and spike rush (Eleocharis palustris/uniglumis) and disturbed ground types such as hairy buttercup (Ranunculus sardous) and common nettle (Urtica dioica). The only indications of food remains are a few fig (Ficus carica) seeds and several hazel nut (Corylus avellana) fragments. Charcoal fragments were again abundant with some animal bone and shellfish fragments. The primary fill of a ditch (context 242) described as greyish silty clay sand was sampled. Although little dating evidence was recovered later deposits higher in the section were of Roman date. The very small float recovered (7 ml) contained charcoal fragments, single charred oat and barley grains and a charred sedge nutlet. The waterlogged plant remains included weeds of disturbed ground such as common nettle and elder. A few of the species recovered will grow in damp conditions as on riverbanks and included henbane (Hyoscyamus niger) and hemlock (Conium maculatum) as well as sedges and blinks (Montia spp) which thrives in many types of wet places. This may suggest that the ditch was wet for at least parts of the year.
Medieval Two tanning pits were located in trench 4. One had been lined with clay and contained a primary fill (179) dated to the late 15th century. This and two overlying fills (169 and 132) were sampled. The primary fill of the other, smaller pit (213) was also sampled. The floats from all four contexts ranged from 14 to 44 millilitres and consisted primarily of charcoal fragments with low concentrations of charred wheat, barley and oat grains. Context (213) also yielded a few charred weed seeds. The only waterlogged seeds surviving were elder, many of which were fragmentary. Elder seeds are robust and often survive where other material will have decayed. The residues produced many fragments of animal bone. Many were unidentifiable but they did include some fragments of horn-core and metatarsals. These may represent by-products of butchery used by both horn workers and tanners in their crafts, the two industries often seeming to have been found together. Fish bones, including some scales, were also noted. Primary fill (179) also contained lumps of concreted mineralised material with stem fragments, small slivers of bone and impressions of fly pupae embedded in it. This may represent dung or faeces and the pollen assessment (below) showed the presence of whipworm eggs (Trichuris sp), an intestinal parasite of a wide variety of vertebrates. The processes in tanning include a preliminary stage in which the animal hides are immersed in one of a number of mixtures, including dog dung and warm water to make them soft and porous before the actual tanning process begins (Williams 1979).
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 181 Conclusion There was no indication from the plant macrofossils recovered from the cesspit to suggest they related to the primary function of the feature. The fills therefore may derive from backfilling of the pit. The pollen results also give no indication of the presence of cess although the possible introduction of hay or straw in the fills may have been deliberate in an attempt to mask unpleasant odours. The fills of the tanning pits also produced a limited assemblage of plant macrofossil remains. The charcoal and few charred remains are likely to be of a secondary nature and not related to the original function of the pits. The presence of horn-cores may be a by-product of the tanning industry. It seems unlikely that the tanning process itself would leave many obvious traces of the original function of the pits, although the presence of mineralised coprolites may relate to one of the stages in the tanning process. Pollen Assessment by Heather Tinsley Four spot samples were analysed from the Tanners' Hall excavations. They came from the lower fill of Roman cesspit (246); Roman cesspit (235); medieval tanning pit (179); and Roman ditch (242). Methodology The samples were prepared for analysis using standard techniques (Moore, Webb and Collinson 1991). Two tablets of an exotic marker (Lycopodium) were added to each at the start of the laboratory procedures to help assess pollen density. Initial digestion in 10 per cent potassium hydroxide was followed in turn by treatment with cold 40 per cent hydrofluoric acid for one week and by treatment with hot 10 per cent hydrochloric acid. All samples were acetolysed and finally stained with safranin and mounted in glycerol. The samples were counted at a magnification of Ч400 and for critical determination at Ч1000 magnification. The aim was to count +/- 100 pollen grains per sample to give an assessment of the potential of this material for full analysis. Spores were not included in the count, nor were any degraded or crumpled grains. The pollen density in (179) (the tanning pit) was very low and in this case the count was stopped at 80 grains when 26 traverses of the slide had been made. The presence of charcoal fragments was noted, and eggs of the nematode genus Trichuris were counted. Plant nomenclature follows Stace (1991) and pollen types follow Bennett (1994). Results (Table 11) Context (246). Pollen preservation was fair. Generally grains were rather thin and crumpled, although the proportion which was not identified owing to poor preservation was low. The density of pollen in the sample was high. Of the 119 pollen grains identified none was tree pollen. The pollen assemblage was dominated by Poaceae (grasses) (51% of total pollen) of which 11 per cent were Hordeum type which includes barley as well as some wild grasses such as Glyceria fluitans (sweet grass) and Elytrigia repens (couch grass). Cyperaceae (sedges) contributed 10 per cent of the pollen assemblage, with the remainder being made up of flowering herbs, principally Cichorium intybus type (e.g. dandelion and hawkbit, 22.0%) and Centaurea nigra (common knapweed, 8.4%). Two grains of Plantago lanceolata (ribwort plantain) were noted and also one grain of Typha angustifolia
182
JO VALLENDER
Table 11. Pollen assessment.
Sample
Context Context
246
235
Context Context
179
242
Tree pollens
Pollen counted
Corylus (hazel)
Quercus (oak)
Sorbus type (rowan)
Non-tree pollens
Poaceae undifferentiated
Poaceae Hordeum type
Cyperaceae
Typha angustifolia type
Plantago Lanceolata
Chenopodiaceae
Solidago virgaurea type
Centaurea nigra
Centaurea cyanus
Cichorium intybus type
Apiaceae undifferentiated
Polygonum
Fabaceae undifferentiated
Brassicaceae
Rosa type
Ranunculaceae
Silene type
Counted outside pollen sum
Filicales undifferentiated
Polypodiaceae
Pteridium aquilinum
Sphagnum
Tilletia sphagni
Crumpled grains
Broken grains
Degraded grains
Traverses
Trichuris eggs
Charcoal fragments
119 48 13 12 1 2 10 27 3 3 8 1 12 3 11 6 abundant
118 80 (26 traverses) 120 1 1 1
60
13
20
8
30
1
2
1
2
2
1
2
2
2
1
2
1
2
7
2
2
24
23
85
1
1
5
1
4
1
1
1
2
3
1
7
1
8
4
8
2
6
2
3
1
2
10
6
4
30
5
26
8
37
abundant occasional
type (includes lesser bulrush and bur-reed). A few spores of Pteridium aquilinum (bracken) were found and one of Sphagnum (bog moss). Abundant fragments of charcoal were noted.
Context (235). Pollen preservation was good and the concentration of pollen high. As in the previous sample, the pollen spectrum was dominated by Poaceae (58% of total pollen) of which 7 per cent were Hordeum type. A wide range of flowering herbs occurred in this sample. The most frequent was Cichorium intybus type (20% of total pollen). Seven grains (6%) of Centaurea nigra and five grains (4%) of
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 183 Polygonum (knotweed) were noted. The remaining types ­ Plantago lanceolata, Chenopodiaceae (goosefoot family), Solidago virgaurea type (includes daisy and related Compositae), Apiaceae (the umbels), Fabaceae (pea family), Ranunculaceae (buttercup family) and Rosa type (rose) ­ were represented by one or two pollen grains only. A single grain of tree pollen of Sorbus type (rowan) was found. A few spores of Pteridium were noted. Charcoal was again abundant. Context (242). The pollen concentration in this sample was good but the preservation generally was poor. Of the 120 grains identified a large number (30) showed signs of degradation. This suggests that there may have been differential losses of less robust pollen types. The pollen assemblage was dominated by Cichorium intybus type (71%), a distinctive grain resistant to decay. Poaceae made up 17.5 per cent of the assemblage and occasional grains of Solidago virgaurea, Centaurea nigra, Plantago lanceolata, Chenopodiaceae and Brassicaceae were found. There were a few spores of ferns (undifferentiated) and of bracken. No charcoal was noted. Context (179). Pollen preservation was reasonable, though the large Poaceae (Hordeum type) tended to be very crumpled. The pollen assemblage was again dominated by Poaceae (54%) and of these more than 30 per cent were Hordeum type. As in the other samples the principal herb pollen found was Cichorium intybus type (29%). Other herbs represented by just a few grains were Solidago virgaurea, Apiaceae, Chenopodiaceae, Ranunculaceae, Plantago lanceolata, Brassicaceae (cabbage family) and Centaurea cyanus (cornflower). Again a few bracken spores were noted and two spores of the fungus Tilletia sphagni. There were occasional small charcoal fragments. Thirty-seven eggs of the parasitic nematode Trichuris were counted and these were identified with reference to published illustrations (Bayer Pharmaceutical 1968). Different species of Trichuris infest humans, dogs and pigs; they are similar in appearance but differ in size. On the basis of their size the examples counted in this sample appear likely to be the human whipworm. Interpretation The assemblages from all four samples are virtually devoid of tree pollen. All have high percentages of grasses and of Cichorium intybus type (e.g. dandelion and hawkbit), together with a range of pollen of other herbs associated with disturbed or open ground, e.g. plantain, knotweed, daisy and knapweed, and spores of bracken. Such a pollen assemblage is local in origin, the lack of tree pollen indicating the absence of any regional component. It is an assemblage which could be associated with weed, with colonised disturbed ground within an urban area or with hay or straw. The presence of two grains of Centaurea cyanus (cornflower) in the sample from (179) fits the latter interpretation. If the former interpretation is correct the pollen assemblage would have reached these fills by the usual process of dispersal and deposition. If the latter interpretation is correct the pollen could have entered the fills when hay or straw was deposited in them. The exceptionally high percentage of Cichorium pollen in the sample from (242) is probably a function of differential preservation which has resulted in losses of less robust pollen types. It is not therefore to be regarded as particularly ecologically significant. The presence of Trichuris eggs was only noted in the sample from (179), the tanning pit. This could be significant in terms of the use of faeces in the tanning process or it could simply be due to cess entering the pit at some stage.
184
JO VALLENDER
DISCUSSION
The evaluation and excavation at Tanners' Hall produced evidence for human activity from the 2nd century AD. Several factors affected the survival of the stratigraphic sequence across the site. Two episodes of truncation were recorded. The first may have occurred in the late Roman period and could account for the paucity of late 4th-century pottery and activity on the site. The second, and perhaps the more significant, was the removal of all the medieval and post-medieval tannery surfaces and above ground structures as part of a cleansing of the site, probably towards the end of the 18th century in preparation for the reuse of the site as an allotment. The base level of this activity coincided with the upper surviving Roman levels at the east end of the site and for the most part with the levels required for the construction of the new road. In an archaeological investigation that was severely limited, only the upper fills of the majority of features were investigated and many other features remained unexcavated and were dated from surface finds only. Although evaluation trench 4, located centrally within the area of the excavation, provided a good stratigraphic sequence from the 1st to the 18th centuries, the phasing of the stratigraphic sequence presented below must be regarded as provisional.
Prehistoric There was no evidence for prehistoric activity within the excavated area. However, sandstone quern fragments recovered from a medieval context and a single damaged piece of worked flint may indicate activity of this period within the locality.
Roman The excavation site is located between the Roman settlement at Kingsholm (c.600 m to the north) and the later Colonia founded at Gloucester in the late 1st century AD on the site of the modern city centre (c.250 m to the south). The combination of the results of evaluation trench 4 and the excavation produced a stratigraphic sequence, 2.5 m deep, spanning the 1st to the 4th centuries. Information from evaluation trenches 1 and 2 provided the evidence for activity to the south of the site. The presence of pre-Flavian pottery within the assemblage may indicate early activity associated with the fortress at Kingsholm. The earliest Roman activity recorded was represented by a ditch [259] aligned NE­SW and by a possible cesspit (246) within trench 1. The material which formed the base of road (779) also produced pottery of this date. The upper edge of the ditch survived to a height of 10.39 m above OD whilst the lowest recorded levels of the road base were at 11.55 m above OD, suggesting that the two were not contemporary and that the finds within the road base were redeposited. Very little domestic or environmental evidence was present within the ditch fills. The cesspit was more productive and may indicate domestic activity to the south of the site towards the Colonia. Examination of the pollen and charred plant remains from these contexts suggests an environment of disturbed ground colonised with weeds, perhaps with boggy areas and standing water for at least part of the year. The inhabitants of the area were eating meat, predominately cattle, and also fish, shellfish, barley, figs and hazelnuts. A phase of later Roman activity is represented by the possible remains of smithing activity at the eastern end of the excavation area. The evidence, associated with a clay floor (671), suggests a broad 2nd­4th-century date. Other structures, including a tile-lined kiln, are possibly contemporary. From these features quantities of iron slag and some copper slag, hammerscale and
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 185 smithing hearth bottoms were recorded. Large quantities of iron nails recovered from fills of all periods may have originated from the Roman smithing activity. Also associated with this period were structures of posthole and beam-slot construction both within the main excavation area and in trenches 1 and 4. A fragment of a Roman column recovered from a medieval context and part of a limestone column capital found during the earlier excavations within Tanners' Hall (Heighway 1983b) may indicate a building of some status in the locality. No further evidence for this building was found within the areas of the evaluation or excavation with the exception of a number of fragments of Roman glassware, painted wall plaster and tesserae. Further evidence for industrialisation of the area to the north of the Colonia also exists at St Oswald's priory (325 m to the north-west) where a tilery with a date range of the 1st to the 3rd century has been excavated (Heighway and Bryant 1999). The proximity of the tilery may have influenced the amount of ceramic building material recovered from both the evaluation and excavation, which included many roofing tiles, some pilae and possible fragments of box-flue. A final phase of activity was possibly represented by a fine cobbled surface observed across the western end of the excavation area. The extensive surface was partially excavated within evaluation trench 4 and was found to comprise a sequence of surfaces. Although these produced dating evidence ranging from the 2nd to the 4th centuries, they sealed beam slots which contained 4thcentury material. Surfacing was also recorded as a final Roman phase in evaluation trench 2. In this area the surface, which had a cambered, road-like form, appeared to be aligned north­south, and its base comprised quantities of slag, including smithing hearth bases, perhaps the residue of earlier Roman smithing found elsewhere on the site (see above). Further attempts at phasing the recorded features and placing them within the wider context of Roman Gloucester were made by studying the alignments of the early fort at Kingsholm and of the later Colonia. The fort was aligned north­south and the Colonia NE­SW. Ditch [259], which has been interpreted as a land division or boundary, was aligned NE­SW, and this may confirm the evidence from the pottery assemblage that this area was first occupied in the early stages of the Colonia rather than associated with the Kingsholm fort. Comparison with other earlier work in the area indicates that at the east end of the site the level of the Roman deposits at 11.58 m above OD (c.0.9 m below the former garage floor level) was considerably higher than had previously been recorded. Previous observations (GCEU 1973­93) suggested that the top of the Roman deposits lies at between 1.6 and 2.7 m below ground level in Worcester Street. These discrepancies may reflect differential truncation of the Roman levels in this area of the town or simply may have resulted from the nature of the earlier archaeological work, which consisted of watching briefs that have generated limited stratigraphic and dating evidence. Ditch [259] survived at a level of c.10.39 m above OD, which is a level broadly comparable with the earlier observations. Due to the limited depth of the excavation it is unclear why the Roman activity was confined to the east end of the site as evidence exists in Hare Lane for Roman deposits at a depth below ground of c.1.6 ­ 2.0 m (GCEU 1973­93). It is possible that ground levels across this area of Gloucester varied considerably to those that exists today and that the Roman deposits are buried more deeply in the vicinity of Hare Lane. However, the intensive medieval activity thought to have taken place in Hare Lane may have significantly truncated any Roman deposits. Anglo-Saxon/Early Medieval Little is known of this area of Gloucester in the immediate post-Roman period. The evidence for occupation during the 5th to the 8th centuries is largely restricted to the city centre. Towards the end of the 9th century Жthelflaed, queen of Mercia, established a burh or defended settlement at
186
JO VALLENDER
Gloucester. A royal palace at Kingsholm and the priory of St Oswald were founded at the same time to the north of the burh; it is not clear to what extent other domestic settlement was present in the vicinity. The extent of the burh in the 10th century and therefore the line of its defences is uncertain, but it probably included the area within the Roman defences, suggesting that the excavation site was at that time located outside of the burh boundary. It is suggested (Hurst 1986) that the river Twyver may have marked a north Anglo-Saxon defensive line (Fig. 1). This stream was diverted and channelled on a number of occasions throughout the medieval period making it difficult to assess its original course (Heighway and Bryant 1999). It has also been suggested that the layout of Hare Lane and Park Street, aligned south­north together with a third parallel street (Bride Lane) to the west, may represent a late Saxon extra-mural cattle or stall market within a suburb north of the burh (Heighway 1983b). During the Anglo-Saxon period the excavation site appears to have been cleared for cultivation as evidenced by the lack of later Roman stratigraphy and the importation of soil (830). Although well stratified between the Roman and medieval features the dating of deposit (830) has proved difficult, since its upper surface coincided for the most part with the post-medieval truncation horizon and the level at which excavation ceased. The small amount of the deposit that was excavated produced a pottery assemblage ranging from the 1st to the 19th century, indicating a considerable degree of disturbance from medieval and post-medieval activity on the site. The evidence from the evaluation trenches is somewhat clearer. Deposits similar to (830), compared by soil type and their position in the stratigraphic sequence, were recorded in trench 1 (227) and trench 2 (107) and were dated to the 10th­11th century. A very similar deposit, (108), was also recorded during the earlier excavation within Tanners' Hall. Analysis of its particle size, coarse fraction and molluscs in addition to its soil chemistry and micro morphology (Bell and Macphail 1983) concluded that the sediment had been dumped by human activity rather than deposited by water. It suggested that there was evidence for the mixing of two soil types and the inclusion of organic matter. The possible micro morphological evidence of cultivation suggested that the deposit may have been purposely dumped for urban cultivation. The deposit within Tanners' Hall was considered to be a single homogenous deposit, but the similar deposit (107) in trench 2 of the evaluation was composed of several layers, probably the result of tipping. To the rear of 15 Park Street, 30 m east of the excavation, a greenish sandy loam recorded above Roman deposits may be of the same origin as (830), perhaps indicating that the area of post-Roman cultivation extended across Hare Lane and Park Street (GCEU 1973­93). This period of cultivation was followed by a phase of activity, possibly in the 11th century, represented by demolished structures, backfilled features and structural and household debris some of which was heavily burnt and used as levelling deposits. It is possible that this is evidence for a fire in 1088 which apparently destroyed most of Gloucester (Hurst 1986, 129­32).
Medieval During the medieval period the site was within the northern suburbs of the town which were defined to the north and east by the river Twyver (Fig. 1). The pottery from the late 11th/12th century is domestic in nature, being dominated by jars and cooking pots, although after this time there may have been a change in use on the site. This change is also reflected in the animal bone assemblage which also suggests that, before the 13th century, activity in the area of the site was largely domestic, The excavated evidence from the interior of Tanners' Hall suggests that its construction was between the 11th and the 13th centuries (Heighway 1983b). The evidence from the two phases of building recording which have been carried out (Heighway 1983b; Pyper 1999) give a firmer date of the 13th century, and none of the
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 187 possible 11th-century activity recorded in the excavation could be positively linked to any activity on the site of Tanners' Hall itself. Documentary evidence records that the area was occupied by tanners by the early 13th century, Hare Lane then being known as the tanners' street (Herbert 1988, 366), and excavation of the interior of Tanners' Hall (Heighway 1983b) has confirmed the presence of tanning activity. The extent of the industrial area is uncertain. Tanners' Hall became the meeting place of the Company of Tanners but there is no evidence that the Company existed in the 13th century. None of Gloucester's trade companies is recorded before the mid 15th century (Heighway 1983b). The majority, if not all, of the medieval features recorded (pits, postholes and linear features) appears to have been associated with the tanning process. The date of the commencement of this activity on the site is unclear. Several pits and postholes contained pottery assemblages dated to between the 11th and the 13th centuries and may coincide with the commencement of tanning on the site. The industry thus may predate the construction of Tanners' Hall. The combination of medieval imbrices and 12th-century pottery suggests a medieval structure of relatively high status in the vicinity which might relate to the site of the hall. Over 50 pits were recorded. Only one, [181] in evaluation trench 4, was fully sectioned and no other complete profiles were recorded. Therefore neither could their functions, dimensions or dates of backfilling be ascertained nor their sequence established. It is probable that the pits represent a variety of activities within the tanning process, which would have included soaking hides in alkaline liquor comprising wood ash or lime to encourage putrefaction and loosen the hair. Other sites such as Exe Bridge and High Street, Exeter (Cherry 1991), have provided evidence of pairs of tanning pits dating to the late 13th century, one of each pair containing lime. No deposits of lime were recorded within the excavation area. However within evaluation trench 1 successive layers of lime were recorded at the base of the hollow way, perhaps evidencing the transport of the material to the front of Tanners' Hall. Pit [181] provided the only detailed evidence for the tanning process. It had been lined with pink clay (also observed elsewhere on the site) presumably to prevent the tanning liquors from leaking. Impressed into this was the imprint of wicker work and could have represented a basket, which could have aided the lifting of the soaked skins in bulk from the pit allowing the liquors to drain through the weave of the wicker. Clay pit linings were also recorded at a medieval tannery site in Northamptonshire (Cherry 1991). The primary fill (179) of pit [181] dated to the 13th to 14th century and contained coprolites, thereby indicating that the pit was perhaps used in the processes of bating, puering or mastering. They involved the use of dog faeces to open up the grain of the hide prior to tanning. However, the environmental evidence from the deposit produced fragments of horn-cores, cattle metatarsals, fish scales, wheat, barley and oats, perhaps indicating the pit was also used for waste disposal. The noxious substances involved in the tanning process may explain the location of the industry on the north side of the medieval town, as the prevailing winds come from the south (Heighway 1983b). No evidence was recorded across the site for other specific tanning activities. As a result it has not been possible to make comparisons with excavated tannery sites such as those in Northampton, Exeter, Winchester and York (Cherry 1991). The pollens from the base of the tanning pit [181] suggest that the area around the tannery during this period was still relatively undeveloped, 54 per cent of the pollens comprising grasses with some bracken and fungus spores. However, a similar pollen assemblage could have derived from the inclusion of hay or straw within the fill of the pit. Adjacent to Hare Lane a possible structure was identified from the remains of a clay floor and robbed cut features dating to the 12th­14th century. The survival of a floor surface at this level may suggest that it predated tanning activities on the site as all floor surfaces associated with tanning
188
JO VALLENDER
had been destroyed by later activity. The building also appeared to be respected by broadly contemporary tanning pits in the immediate vicinity. Although numerous postholes and linear features were excavated on the site little could be learned regarding the activity to which they may have related. A possible structure was suggested by a linear arrangement of postholes aligned NE­SW (Fig. 6). This, however, did not relate to the alignment of either Hare Lane or Tanners' Hall, both of which were in existence during this period. Few finds which can be attributed to the tanning process were recovered from the site. The range of medieval pottery was typical for Gloucester and demonstrated high levels of redeposition. Many of the tanning pits contained more Romano-British than medieval pottery. The majority of the features considered to be associated with the tanning industry could not be dated any more closely than to the 12th­15th/16th century due to the small assemblages produced by the limited excavation. Where earlier medieval material was recovered, it appeared to indicate domestic activity rather than tanning (see above). Regional wares become more apparent in the assemblage by the 13th century and may indicate the start of regional trade connected with the tanning industry. Despite this apparent increase in activity there was little pottery dating from the 16th/17th century. This may be due to the industrial nature of the site at that time with domestic activity being located elsewhere. The animal bone assemblage confirms that tanning had begun on the site by the 13th century, albeit on a localised and small scale with a diverse cross-section of animals of all ages being processed. It is clear from the bone assemblage that the hides were supplied with the horns attached, as hundreds of cores were present on the site, mainly in backfilled pits. Analysis of the cores revealed very little signs of butchery and it has been suggested that their cores were left to rot, thereby facilitating the easy removal of the horn. A variety of knives would have been used in the various tanning processes: the hides would have been scraped over beams with two handled knives; blunt single-edged unhairing knives, two-edged fleshing knives and blunt scudding knives were also used. Where identifiable the blades recovered were similar to large kitchen knives that were possibly used to cut the hide into relevant sections. A significant number of bone pins was collected from medieval contexts but it is not clear whether they were related to tanning or to another industry whose waste was used to backfill the disused tanning pits The excavation could not confirm the boundaries or the extent of the tanning yard adjacent to Tanners' Hall, apart from possibly a length of the south boundary wall. Cartographic evidence suggests that the tannery lands extended to the river Twyver to the east and that Hare Lane probably delineated the western boundary (see Fig. 1). The site at Tanners' Hall would have undoubtedly utilised the Twyver as the nearest source of water. In Colchester the river Coln was used by the tanners to wash the hides (Cherry 1991). Elsewhere, as in York, tanneries have been excavated some 500 ft from the nearest river and it is suggested that well water was also utilised (Cherry 1991). A well was also recorded on the Tanners' Hall site. Although it remained undated, it may indicate that the demand for water was so great that two sources were required. Possible tanning pits recorded 75 m to the west into the area of the former Bride Lane (GCEU 1973­93) were probably not related to the Tanners' Hall site. Evidence from trench 2 suggests cultivation of the ground in front (south) of Tanners' Hall at this time but it contradicts the evidence from trench 1 which recorded a hollow way and a later cobbled yard giving access to the front of Tanners' Hall from Hare Lane. The medieval activity outside the tannery yard in the areas of Park Street and Hare Lane was unclear due to the presence of in-situ post-medieval deposits.
Post Medieval The post-medieval development of the area is well-documented. Leases from this period exist for the tannery (Heighway 1983b) and detail some of the goods and implements associated with the
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 189 works. Although cartographic evidence commences in 1610 with Speed's map of Gloucester, the mapping of the site remained largely schematic until Causton's map of 1843, by which time Tanners' Hall had been converted into two cottages. No maps exist to show the tannery precinct or any structures which may have occupied its interior. Tanning is thought to have continued on the site until the 18th century but the date at which it ceased is not clear. This phase of activity is perhaps the most ill-defined, as practically all traces of the layout of the tannery yard were deliberately removed in order to reclaim the land after its closure. A small number of the features recorded were considered to relate to the post-medieval tanning industry. A large rectangular pit or vat (1037) of unknown depth, backfilled with horncores, and several other pits were recorded adjacent to Hare Lane (Fig. 8). A small group of postholes dating to the post-medieval period was recorded to the north of Tanners' Hall; they were not suggestive of a structure. Study of the horn-cores recovered from the site indicates that only cattle were being processed during the post-medieval period. The horn-cores also indicate a diversity in animal type perhaps indicating a wider catchment area or a diversity of animal breeds. Also demonstrated was a shift towards the use of younger animals after the 17th century and the increased frequency in cows coincided with the post-medieval growth in dairy production. In the 18th century most of the tannery deposits appear to have been removed from the site and no surfaces, structures and very few pits or vats survived. This clearance, to remove the noxious deposits from tanning and to introduce clean soil as a replacement, not only destroyed the medieval and post-medieval stratigraphy on the site but also caused a great deal of disturbance to the underlying deposits. Once the reclamation had been completed there is a record (Heighway 1983b) that the site was used as allotments. This phase of activity could relate to the structure recorded adjacent to Worcester Street, before the construction of Worcester Lawn in the late 18th/early 19th century. Modern Worcester Lawn was constructed during a period in which this area of Gloucester was significantly re-shaped. Worcester Street was constructed in 1822 and most of the housing to either side of it built within the next a few years (Herbert 1988, 167). Map evidence from the 19th century suggests that by the late 1880s the whole area had been completely built over, leaving no medieval buildings in the area between Hare Lane and Park Street (OS 1884­86). By this time Tanners' Hall had been converted into two cottages and the area of Tanners' Yard had at least in part been rebuilt. In 1919 A.C. Stretton, the owner of Worcester Lawn and surrounding properties, Tanners' Hall among them, conveyed the whole to his garage business, Strettons Garage Ltd. It is thought that the post-medieval house south of the tannery boundary was demolished at this time. By the early 1920s Worcester Lawn had been demolished and Tanners' Hall and the adjacent cottages gutted and converted for use as garage buildings (Heighway 1983b). By 1936 Tanners' Hall was almost totally surrounded by garages (OS 1936), although these gradually went out of use during the 1970s and 1980s, some being used for different purposes until the land was purchased in the 1980s in conjunction with the scheme proposed for the inner relief road. Acknowledgements The author would like to thank Richard Waters of Gloucestershire County Council for commissioning the works and Sir William Halcrow and Partners, in particular Gordon Jones and his staff, for their support and co-operation. County Archaeologist Jan Wills edited this report; Toby Catchpole managed the excavation; Caroline Jamfrey illustrated the sculptural fragment; and Karen Derham and Richard Barrett provided the other illustrations that support the text. Finally
190
JO VALLENDER
thanks are owed to the numerous site assistants who worked on all aspects of the site, particularly Paul Nichols.
Bibliography Albarella, U., and Davis, S., 1996. `Mammals and Birds from Launceston Castle, Cornwall: decline in status and rise in agriculture', Circaea 12, 1­156. Armitage, P.L., 1982. `A system for ageing and sexing the horn-cores from British post-medieval sites (17th to early 18th century) with special reference to unimproved British longhorn cattle', in B. Wilson, C. Grigson and S. Payne (eds.), Ageing and sexing animal bones from archaeological sites (BAR British Series 109), 37­54. Armitage, P.L., 1993. `Report on the cattle horn-cores from the Greyfriars site, Chichester, 1984', in A. Down and J. Magilton, Chichester Excavations VIII (Chichester Dictrict Council), 211­18. Armitage, P.L., and Clutton-Brock, J.,1976. `A system for Classification and Description of the Horn Cores of Cattle from Archaeological Sites', Jnl. Archaeol. Science 3, 329­48. Ayres, K., and Clark, K., 1995. `The animal bones from Denmark Road High School, Gloucester' (Centre for Human Ecology and Environment unpublished rep.). Backes, M., and Dolling, R., 1969. Art of the Dark Ages (Harry N. Abrams, New York). Bailey, R., and Cramp, R., 1988. Corpus of Anglo-Saxon Stone Sculpture II: Cumberland, Westmorland and Lancashire North-of-the-Sands (Oxford University Press). Bayer Pharmaceutical 1968. The Microscopic Diagnosis of Tropical Diseases (Bayer, Germany). Bell, M., and Macphail, R., 1983. `The "dark earth" deposit', in C.M. Heighway, `Tanners' Hall, Gloucester', Trans. BGAS 101, 107­8. Bennett, K.D., 1994. Annotated catalogue of pollen and pteridophyte spores of the British Isles (Department of Plant Sciences, University of Cambridge). BGS 1972. British Geological Survey map 1:50,000, Sheet 234, Gloucester. Blagg, T.F.C., 2002. Roman Architectural Ornament in Britain (BAR British Series 329, Oxford). Boessneck, J., 1969. `Osteological Differences between Sheep (Ovis aries Linnй) and Goat (Capra hircus Linnй)', in D. Brothwell, E. Higgs and G. Clark (eds.), Science in Archaeology (Thames and Hudson, London), 331­58. Bryant, R.M., 1990. `The Lypiatt Cross', Trans. BGAS 108, 33­52. Bryant, R.M., and Heighway, C.M., 2003. `Excavations at St. Mary de Lode Church, Gloucester, 1978­79', Trans. BGAS 121, 97­178. Cassidy, B., 1992. The Ruthwell Cross (Princeton University Press, New Jersey). Cherry, J., 1991, `Leather', in J. Blair and N. Ramsay (eds.), English Medieval Industries (London), 295­318. Cramp, R., 1977. `Schools of Mercian sculpture', Mercian Studies (Leicester University Press). Cramp, R., 1984. Corpus of Anglo-Saxon Stone Sculpture I: County Durham and Northumberland (Oxford University Press). Cramp, R., and Miket, R., 1982. Catalogue of Anglo-Saxon and Viking Antiquities in the Museum of Antiquities, Newcastle upon Tyne (University of Newcastle Upon Tyne). Darvill, T., 1988. `Excavations on the Site of the Early Norman Castle at Gloucester, 1983­84', Medieval Archaeol. 32, 1­49 Dobney, K.M., Jaques, S.D., and Irving, B.G., 1995. Of Butchers and Breeds: Report on vertebrate remains from the city of Lincoln (Lincoln Archaeol. Studies 5). Driesch, A. von den, 1976. A guide to the measurement of animal bones from archaeological sites (Peabody Museum Bulletin 1, Harvard). Enright, D., and Watts, M., 2002. A Romano-British and medieval settlement site at Stoke Road , Bishop's Cleeve, Gloucestershire (Bristol and Gloucestershire Archaeol. Rep. 1, Cirencester, Cotswold Archaeol.). Fulford, M.G., Rippon, S., Allen, J.R.L., and Hillam, J., 1992. `The Medieval Quay at Woolaston Grange, Gloucestershire', Trans. BGAS 110, 101­27. Fullbrook-Leggatt, L.E.W.O., 1968. Roman Gloucester (Glevum) (Stroud/BGAS).
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 191 GCEU. 1973­93. `Gloucester City Excavation Unit unpublished site records'. Goudge, C.E., 1982. `The Roman pottery', in C.M. Heighway and A.J. Parker, 38­57. Grant, A., 1988a. `The Animal Resources', in G. Astill and A. Grant (eds.), The Countryside of Medieval England (Blackwell, Oxford). Grigson, C., 1974. `The Craniology and Relationship of Four Species of Bos. I Basic Craniology: Bos taurus L. and its Absolute Size', Jnl. Archaeol. Science 1, 353­79. Grigson, C., 1975. `The Craniology and Relationship of Four Species of Bos. II Basic Craniology: Bos taurus L. Proportions and Angles', Jnl. Archaeol. Science 2, 109­28. Grigson, C., 1976. `The Craniology and Relationship of Four Species of Bos. 3. Basic Craniology: Bos taurus L. Saggital Profiles and other Non-measurable Characters', Jnl. Archaeol. Science 3, 115­36. Grigson, C., 1978. `The Craniology and Relationship of Four Species of Bos. The Relationship between Bos primigenius Boj. and B. tarus L. and its implications for the Phylogeny of the Domestic Breeds', Jnl. Archaeol. Science 5, 123­52. Hamshaw-Thomas, J., 2000. `When in Britain do as the Britons: dietary identity in early Roman Britain', in P. Rowley-Conwy (ed.), Animal Bones, Human Societies (Oxford, Oxbow), 166­9. Harman, M., 1996. `The mammal bones', in M. Shaw, `The excavations of a late 15th­17th century tanning complex at The Green, Northampton', Post-Medieval Archaeol. 30, 89­103. Hassall, M.W., and Rhodes, J.F., 1974. `Excavations at the New Market Hall, Gloucester, 1966­7', Trans. BGAS 93, 15­100. Hatting, T., 1975. `The influence of castration on sheep horns', in A.T. Clason (ed.), Archaeozoological Studies: Paper of the Archaeozoological Conference 1974, held at the Biologisch-Archaeologisch Instituut of the State University of Groningen (North Holland Publishing Company, Amsterdam), 345­51. Heighway, C.M., 1980. `Roman Cemeteries in Gloucester District', Trans. BGAS 98, 57­72. Heighway, C.M., 1983a. The East and North Gates of Gloucester (Western Archaeol. Trust Monograph 4). Heighway, C.M., 1983b. `Tanners' Hall, Gloucester', Trans. BGAS 101, 83­109. Heighway, C.M., 1984. `Anglo-Saxon Gloucestershire', in A. Saville (ed.) Archaeology in Gloucestershire (Cheltenham), 225­47. Heighway, C.M., Garrod, A.P., and Vince, A.G., 1979. `Excavations at 1, Westgate Street, Gloucester, 1975', Medieval Archaeol. 23, 159­213. Heighway, C., and Parker, A.J., 1982. `The Roman tilery at St Oswald's Priory, Gloucester', Britannia 13, 25­77. Heighway, C.M., and Bryant, R.M., 1999. The Golden Minster: The Anglo-Saxon Minster and Later Medieval Priory of St. Oswald at Gloucester (CBA Research Rep. 117, York). Herbert, N.M. (ed.), 1988. The Victoria County History of the County of Gloucester 4, The City of Gloucester. Hoyle, J.P., 1994. `Gloucester Inner Relief Road: Stage 1B, Desk Based Archaeological Assessment' (Glos. County Council Archaeol. Service typescript rep.). Hubert, J., Porcher, J., and Volbach, W.F., 1970. Carolingian Art (Thames and Hudson, London). Hunter, A.G., 1963. `Excavations at the Bon Marchй Site, Gloucester, 1958­59', Trans. BGAS 82, 25­65. Hurst, H.R.,1986. Gloucester, the Roman and Later Defences (Gloucester Archaeol. Rep. 2). Ireland, C., 1983. `The pottery', in Heighway 1983a, 96­124. Ireland, C., and Goudge, C., 1983. `The pottery', in Heighway 1983b, 102­7. Jones, G., 1984. `Animal Bones', in A. Rogerson and C. Dallas (eds.), Excavations in Thetford 1948­59 and 1973­80 (East Anglian Archaeol. Rep. 22), 187­92. Jones, G., 1994. `Mammal and Bird Bones', in B.S. Ayres, Excavations at Fishergate, Norwich, 1985 (East Anglian Archaeol. Rep. 68), 37­43. Kendrick, T.D., 1938. Anglo-Saxon Art to AD 900 (Methuen, London). King, A., 1978. `A comparative survey of bone assemblages from Roman Britain', Bulletin Institute of Archaeol. 15, 207­32. King, A., 1991. `Food Production and Consumption ­ Meat', in R. Jones (ed.), Britain in the Roman Period: Recent Trends (Department of Archaeology and Prehistory, University of Sheffield). King, J.F., 1992. `Anglo-Saxon Sculpture at St, Andrew's Church, Pershore', Trans. Worcestershire Archaeol. Soc. (3rd series 13), 129­34.
192
JO VALLENDER
Lauwerier, R.C.G.M., 1988. Animals in Roman Times in the Dutch Eastern River Area (Nederlandse Oudheden 12). Lang, J., 1991. Corpus of Anglo-Saxon Stone Sculpture III: York and Eastern Yorkshire (Oxford University Press). Langton, B., Ings, M., Walker, G., and Oakey, N., 2000. `Medieval field systems at Tinker's Close, Moreton-in-Marsh, Gloucestershire. Excavations in 1995­6', in N Oakey (ed.), Three Medieval Sites in Gloucestershire (Cotswold Archaeol. Trust Occasional Paper 1), 15­23. Levitan, B., 1985. `The Animal Bones', in A. Saville, `Salvage recording of Romano-British, Saxon, Medieval and Post-medieval remains at North Street, Winchcombe, Gloucestershire', Trans. BGAS 103, 130­5. Levitan, B., 1987. `Bone Analysis and Urban Economies: Examples of Selectivity and a Case for Comparison' (Ancient Monuments Laboratory Rep. 224/87). Maltby, M., 1979. `The Animal Bones', in Heighway, Garrod and Vince, 182­5. Moore, D.T., 1978. `The Petrography and Archaeology of English Honestones', Jnl. Archaeol. Science 5, 61­73. Moore, P.D., Webb, J.A., and Collinson, M.E., 1991. pollen analysis (Blackwell Scientific Publications). Noddle, B., 1975. `A comparison of the animal bones from 8 Medieval sites in southern Britain', in A.T. Clason (ed.), Archaeozoological Studies. (North Holland Publishing Company, Amsterdam), 248­60. O'Connor, T.P., 1982. The animal bones for Flaxengate, Lincoln c.870­1500 (CBA 18/I). O' Connor, T.P., 1984. Selected Groups of Bones from Skeldergate and Walmgate (Archaeol. of York: The Animal Bones 15/1. CBA). O'Connor, T.P., 1989. Bones from the Anglo-Scandinavian levels at 16­22 Coppergate (Archaeol. of York: The Animal Bones 15/3. CBA). O'Connor, T.P., 1993. `Bone assemblages from monastic sites: many questions but few data', in R. Gilchrist and H. Mytum (eds.), Advances in monastic archaeology (BAR British Series 227. Oxford), 107­11. OS 1884­86. Ordnance Survey Map 1:2,500. Glos XXV.15 (1st edn.). OS 1902. Ordnance Survey Map 1:2,500. Glos XXV.15 (2nd edn.). OS 1923. Ordnance Survey Map 1:2,500. Glos XXV.15 (3rd edn.). OS 1936. Ordnance Survey Map 1:2,500. Glos XXV.15 (4th edn.). OS 1955. Ordnance Survey Map 1:2,500. Sheet SO8318. Powell, A., and Clark, K.M., 1996. `The animal bones from Denmark Road High School, Gloucester, Phase III' (Centre for Human Ecology and Environment unpublished rep.). Price, E., 2000. Frocester: A Romano-British Settlement, its Antecedents and Successors I (Gloucester and District Archaeol. Research Group). Prummel, W., 1978. `Animal bones from tannery pits of `s-Hertogenbosch', Berichten van de Rijksdienst voor het Oudheidkundig Bodemonderzoek 28, 399­422. Pyper. A., 1999, `Tanners' Hall, Gloucester, Building Recording' (Glos. County Council Archaeol. Service typescript rep.). Pyper, A., and Vallender, J., 2001. `Tanners' Hall, Gloucester', Glevensis 34, 59­66. Raimbault, M., 1973. `La cйramique gallo-romaine dite "а l'йponge" dans l'ouest de la France', Gallia 31, 185­206. Rawes, B., 1972. `The Roman pottery kilns at Gloucester', Trans. BGAS 91, 18­59. Rawes, B., 1991. `A Prehistoric and Romano-British Settlement at Vineyards Farm, Charlton Kings, Gloucestershire', Trans. BGAS 109, 25­89. Richardson, B., 1986. `Pottery', in L. Miller, J. Schofield and M. Rhodes, The Roman Quay at St Magnus House, London (London and Middlesex Archaeol. Soc. Special Paper 8), 96­138. Richardson, L. (revised R.D. Beckinsale), 1972. A Handbook to the Geology of Cheltenham and Neighbourhood (Chichley, Buckinghamshire). Roe, F., 1999. `Quernstones', in C. Parry, `Iron-Age, Romano-British and Medieval Occupation at Bishop's Cleeve, Gloucestershire: excavations at Gilder's Paddock 1989 and 1990­1', Trans. BGAS 117, 109­10. Roe, F., unpublished a. `The Worked Stone', in `Gloucester Archaeological Unit rep. on excavations at Upper Quay Street' (GCEU ref 8/89). Roe, F., unpublished b. `The Worked Stone', in `Gloucester Archaeological Unit rep. on excavations at Lower Quay Street' (GCEU ref. 16/93)
EXCAVATION AT TANNERS' HALL, GLOUCESTER, 1997 AND 1998 193 Saunders, G., 1977. `A 16th century tannery at St Albans', Hertfordshire's Past 3, 9­12. Schmid, E., 1973. `Ziegenhцrner als Gerberei-Abfall', Scheiz Volkskunde 63, 65­6. Serjeantson, D., 1989. `Animal Remains and the Tanning Trade', in D. Serjeantson and T. Waldron (eds.), Diet and Crafts in Towns: The evidence of animal remains the Roman to the Post-medieval periods (BAR British Series 199), 129­47. Serjeantson, D., 1996. `The animal bones', in S.R. Needham and A. Spence (eds.), Refuse and Disposal at Area 16 East Runnymede (Runnymede Bridge Research Excavations 2. London, British Museum Press), 194­223. Shaw, M., 1996. `The excavations of a late 15th­17th century tanning complex at The Green, Northampton', Post-Medieval Archaeol. 30, 89­103. Stace, C., 1991. New Flora of the British Isles (Cambridge University Press). Sykes, N., n.d. `The Norman Conquest: A Zooarchaeological Perspective' (Ph.D. thesis, University of Southampton). Talbot-Rice, D., 1952. `The Gloucester Christ', Trans. BGAS 71, 98­100. Timby, J.R., 1991. `The Berkeley Street pottery kiln, Gloucester', Roman Pottery Studies 4, 19­32. Tomber, R., and Dore, J., 1998. The National Roman Fabric Reference Collection: a Handbook (London). Tyers, P.A., 1996. Roman Pottery in Britain (Batsford, London). Ussher, W.A.E., 1906. The Geology of the Country between Wellington and Chard (Memoir for Sheet 311, Geological Survey of Great Britain, HMSO London). Vallender, J., 1995. `An archaeological evaluation at Gloucester Inner Relief Road stage 1B (Tanners Hall)' (Glos. County Council Archaeol. Service typescript rep.). Vallender, J., and Jamfrey, C., 1998. `Buildings recording on the site of stage 1B of the Gloucester Inner Relief Road, Gloucester' (Glos. County Council Archaeol. Service typescript rep.). Vince, A.G., 1979. `The Pottery', in Heighway et al., 170­81. Vince, A.G., 1982. `Grass-tempered pottery', in Heighway and Parker, 50­1. Vince, A.G., 1983. `The medieval and post-medieval pottery', in Heighway 1983a, 125­61. Vince, A.G., 1984. `Appendix 1: Grass-tempered pottery', in Heighway, 240­1. Webster, L., and Backhouse, J., 1991. The Making of England: Anglo-Saxon Art and Culture AD 600­900 (British Museum, London).

JO VALLENDER

File: archaeological-excavation-adjacent-to-the-tanners-hall-gloucester.pdf
Title: 03 BP - The Expedition copy
Author: JO VALLENDER
Published: Mon Oct 17 01:54:27 2011
Pages: 64
File size: 1.46 Mb


ABOUT THE PRODUCTION, 27 pages, 0.53 Mb

, pages, 0 Mb
Copyright © 2018 doc.uments.com