Catalogue of known Roman wooden artefacts from YAT Excavations,

Transcription

1 Catalogue of known Roman wooden artefacts from YAT Excavations, An Insight Report By Steven J Allen York Archaeological Trust for Excavation and Research (2017)

2 The following catalogue lists the wooden artefacts known to be of Roman date recovered from excavations in York by York Archaeological Trust since its foundation in While the list is intended to be as comprehensive as possible the research has been dependent on existing records, which are broadly complete for the sites covered in this report. Smaller projects sites that may have struck Roman deposits have not always been fully written up for the records and the dating of those artefacts is not always available. While the objects discussed and described here come from securely dated contexts there may be others in the YAT collection that still await identification and description. York was founded in AD71 as a legionary base during the conquest of northern Britain. Following the stabilisation of the Roman frontier along the line of Hadrian s Wall in the early second century York developed as the major centre of Roman administration in the north. The civil settlement associated with the fortress had acquired Colonia status before AD 237 (Ottaway 2004, 83) and served as the headquarters for Emperor Septimius Severus (died in York AD 211). The emperor Constantine 1 st was proclaimed in York in AD 306. Quite apart from the major military presence attested by the stationing of a legion in the fortress, archaeological evidence show it was a significant settlement possessing substantial high status buildings whose inhabitants had extensive links to the rest of the empire. The plan and layout of the Roman settlement is not as well known as some others of similar size. York has been almost continuously occupied since the Roman period and Roman levels are consequently buried at depth beneath the historic city core. Exposures of Roman levels are consequently very limited and while these individual sites can be well recorded it is not always possible to make direct links between these sites. In the fortress area, the layout and development can be reconstructed with some confidence as the military buildings conform to broadly similar plans arranged within a broadly similar relationship known from similar fortress sites elsewhere in the empire. Outside the fortress, especially across the Ouse in the civilian settlement around Micklegate the issue is more problematic. While the lines of some roads and locations of some buildings are known, few structures are known as complete plans and their functions cannot always be determined. Wood is a recyclable material. Larger objects can be broken up and/or reworked to make smaller ones and smaller ones still. It can also and ultimately be recycled as fuel. Therefore there are many reasons why wood should not enter the archaeological record, let alone survive in it. Its presence therefore attests a series of decisions made about the use and fate of the artefact, whether choosing to actively deposit the object through to choosing not to recover it from the place of loss. As an organic material prone to decomposition, wooden artefacts will only survive in particular burial conditions. In Roman York this means that the wood will have been carbonised- heavily burnt in a fire- or waterlogged- buried in wet anaerobic conditions that retard the processes of decay. In the former case, though the microscopic structure can be well preserved, the wood is rendered very brittle and has often distorted during the charring process. In waterlogging, much of the wood substance has been lost and the form of the object is only maintained by the water trapped within the wood structure. If the water is allowed to evaporate under uncontrolled conditions the artefact will collapse under the resulting stress, leading to severe distortion, shrinkage and in some cases, complete disintegration and loss of the artefact. 2

3 What this means for Roman wooden finds assemblages in York is that Many fewer objects entered the archaeological record than originally existed Many fewer objects will have survived burial Those objects which have survived are restricted to those which entered particular burial conditions prevailing at the time of deposition. These deposits may be highly localised. Finally, the history of urban development in York, the consequent burial of surviving Roman deposits at depth and the highly restricted nature of recent redevelopment work mean that waterlogged deposits of Roman date in York are rarely accessible for excavation. Only six such projects have taken place since 1972 and most of these have only exposed small areas of Roman occupation (Figure 1, p.23). This has had a profound effect on the quantity of material available study, an effect that one overview of wooden artefacts in Britain has failed to appreciate (Pugsley 2003, 9). Current planning guidance and the preference for preservation of the archaeological record in situ means that it is unlikely that, in the immediate future at least, York will see the sort of major development (and the resulting archaeological work) that has enriched our understanding of comparable settlements elsewhere. Catalogue: The following catalogue covers those sites which had waterlogged Roman deposits partially or fully excavated. Many more sites have been excavated in York but no wooden artefacts were recovered from those Roman deposits recorded. Only artefacts with known Roman provenance from these sites are included. Despite the limited nature of interventions which have exposed waterlogged Roman deposits, these six sites have produced some 103 registered wooden small finds, representing 292 individual artefacts or fragments thereof. Artefacts are numbered in a continuous chronological sequence by York Museum accession code, project name and small finds number within that project. Within the discussion, artefacts are referred to by catalogue number. Assigned dates for these finds (all AD) are derived from the stratigraphic sequence and summary prepared for each site by the then project officer and placed in the site archive. It is possible these dates may be refined by future analysis but any such changes are unlikely to be significant Skeldergate (Bishophill 1). An excavation in the Colonia south of Micklegate near to the river frontage. The majority of the finds are of fourth century date and are derived from the fill of a timber lined well. A summary has been published in AY 17/ Tangentially faced board with single blind sub rectangular hole (18 l, 12 w, 10 deep) let into one face. Cut from outer surface of larger roundwood timber. 138 l, 68 w, 17 th. Alnus spp. Context 2032, SF 1113 (C 4 th ). Not Illustrated Bucket, tapering towards rim, built from thirteen radially faced staves and a one-piece radially faced base board, bound with two Fe hoops. All staves have planed edges and a V section croze groove around their inner faces c. 15 above the base, with no howel or chime bevels present. Both ends of the staves are rounded off. Base board is bevelled on each face to engage with the croze groove. Bucket has a set of four fe nails driven into the outer face of staves 6 and 7 to secure the fixing point for the handle on one side. As reconstructed the opposing stave (stave 1) is badly 3

4 damaged and has no nails present. A pair of nails, of the same size and relative position as those aforementioned, are present in stave 10. Fe hoops formed by a strap with overlapping riveted ends shrunk over the staves but not nailed in place. Vessel as found was damaged and partly articulated. Most staves complete with surface damage especially at upper ends. Damage and loss to lower end of staves 1, 9 and 10 and upper ends of staves 1, 7 and 10. Base board had split along irregular grain into three refitting sections, one of which has been lost and replaced with modern wood. Base diameter c. 225, rim diameter 183. Height 210. Staves w at base, w at rim, th. Base c. 220 dia, 11 th. Hoops 28 w. All staves Abies alba Mill, base Quercus spp. Context 2068, SF 1120 (wood)/1122 (iron). (C 4 th ). Figure 2, p Offcut from peg, cut from radially faced heartwood billet. Sub rounded cross section, tapers along length 60 l, 28 w, 28 dia. Quercus spp. Context 2066, SF 1123 (C 4 th ). Not Illustrated Peg formed from reworked timber. Originally a box quartered billet with a single 24 dia through hole in face towards surviving head end. One edge split away along grain, truncating the aforesaid hole, and opposing edge has single hewn bevel, creating a sub rectangular cross section tip. 107 l, 38 w, 28 th. Fraxinus excelsior L. Context 2405, SF 1132 (C 4 th ). Figure 14, p Fragment of tangentially faced wood from outer face of roundwood shaft. One complete and two truncated oval holes (one on end, one on edge) through face. 54 l, 11 w, 05 th. Sambucus nigra L. Context 2404, SF 1144 (C 4 th ). Not Illustrated Tangentially faced shaft, sub rectangular cross section. Both ends broken away and missing. In two refitting sections. 231 l, 27 w, 22 th. Wood species not identified. Context 2405, SF 1145 (C 4 th ). Not Illustrated Tangentially faced 'bracket' or frame element cut from main stem/branch junction. Shape exploited to create angled, sub rectangular cross section element following wood grain. One end reworked with both edges cut to create bifaced tip. Other end broken away and missing. 112 l, 70 w, 20 th. Quercus spp. Context 2405, SF 1145 (C 4 th ). Figure 14, p Section of roundwood, no bark present. Single hewn facet cut at one end to create chisel tip. 80 l, 17 dia. Wood species not identified. Context 2405, SF 1145 (C 4 th ). Not Illustrated Section of roundwood, no bark present. 60 l, 17 dia. Wood species not identified. Context 2405, SF 1145 (C 4 th ). Not Illustrated Corner of tangentially faced board. Three 08 dia. through holes in face, two at the original end, one along the original surviving edge. Each houses a truncated roundwood peg, cut off flush with each face of the board. Fine knife scoring marks on one face towards broken edge. 174 l, 96 w, 20 th. Quercus spp. Context 2402, SF 1148 (C 4 th ). Figure 6, p Three pieces of thin radially faced board. Probably part of same object but no longer refitting. (i) 64 l, 34 w, 03 th; (ii) 109 l, 54 w, 04 th; (iii) 42 l, 41 w, 04 th. Quercus spp. Context 2404, SF 1153 (C 4 th ). Not Illustrated Five pieces of thin radially faced board. Probably part of same object but no longer refitting (i) 65 l, 26 w, 03 th; (ii) 60 l, 33 w, 03 th; (iii) 61 l, 51 w, 03 th; (iv) 57 l, 44 w, 03 th; (v) 55 l, 53 w, 03 th. Quercus spp. Context 2404, SF 1155 (C 4 th ). Not Illustrated Moulded shaft cut from radially faced billet. Flange along one edge tapering down to rejoin main body along length on one face, flange on opposing face eroded. One end cut away. Other end broken away and missing. In three refitting sections. 156 l, 21 w, 14 th. Fraxinus excelsior L. Context 2405, SF 1157 (C 4 th ). Not Illustrated. 4

5 014. Four fragments of radially faced thin board. In four non-refitting pieces, probably from same object. No. 1 is in tree refitting sections. (i) 58 l, 31 w, 03 th; (ii) 46 l, 44 w, 03 th; (iii) 42 l, 29 w, 03 th; (iv) 41 l, 32 w, 05 th. Quercus spp. Context 2405, SF (C 4 th ). Not Illustrated Three fragments of thin radially faced board, non-refitting. No. 2 in 4 refitting fragments, No. 3 in two refitting fragments. (i) 55 l, 45 w, 03 th; (ii) 61 l, 53 w, 03 th; (iii) 67 l, 56 w, 03 th. All Quercus spp. Context 2409, SF 1175 (C 4 th ). Not Illustrated Head of wooden mattock or similar tool cut from radially faced heartwood. Single 30 dia through hole in face on mid-line and midway along length. Both edges taper away from mid point towards ends. One end finished with a bevelled face forming a blunt sub rectangular cross section tip. Other end formed by each edge having being bevelled to form a sharp sub-triangular cross section tip. Some old damage to one of latter edges and some shrinkage distortion. 241 l, 102 w, 29 th. Quercus spp. Context 2405, SF 1181 (C 4 th ). Figure 10, P Reworked end of wheel spoke, spindle turned from radially faced billet. One end terminates in a roughly hewn blunt point. Other end terminates in a full width tenon with sawn shoulders. Centre mark from mandrel of pole lathe present on both ends. 80 l, 42 dia. Acer campestre L. Context 2451, SF 1236 (C 4 th ). Figure 11, P Rougier Street. An excavation in the Colonia, north of Micklegate with some waterlogged material coming from pit fills and dumped deposits. Not closely dated. Site written up but not yet published Offcut from tangentially faced board. One end abraded, other end sawn square to axis. One edge roughly squared, other edge split away and missing. 76 l, 86 w, 18 th. Quercus spp. Context 1205, SF 95 (C 2 nd ). Not Illustrated Fragment of withy tie. Section of roundwood with 'S' twist along length. Badly fragmented and heavily charred. 31 l, 09 dia. Salix spp. Context 1205, SF 158 (C 2 nd ). Not Illustrated Peg or loose tenon cut from boxed heart timber. Five cleft facets cut to create subpentagonal cross section with a further facet cut across junction of two faces and remaining two edges and opposing face hewn to create sub rectangular cross section tip. Other end cut square to axis of timber. 85 l, 30 w, 26 th. Populus spp. Context 1304, SF 201 (C 2 nd ). Not Illustrated Fragment of scoop cut from tangentially faced heartwood. One face backed, other face hollowed to create curving profile. Traces of dark patina on hollowed face. Upper end broken away and missing before handle. One edge split away and missing. 102 l, 17 w, 08 th. Buxus sempervirens L. Context 1326, SF 235 (C 2 nd ). Figure 8, P Double sided comb cut from tangentially faced board. Bi -convex cross section. Rounded end, slightly damaged. Coarse teeth (6 per 10mm) cut into one edge, fine teeth (8 per 10mm) cut into opposing edge. Scored marking out lines to define base of teeth. No decoration. Ends of teeth and tooth guard badly damaged. One end broken away and missing. Two detached but refitting teeth present in same bag. 46 l, 47 w, 09 th. Buxus sempervirens L. Context 1382, SF 252 (C 2 nd ). Figure 12, P Thin radially faced board, near complete. Sub rectangular cross section with taper towards one edge. Surviving corners rounded off. In 14 refitting sections. Possible box side panel. 142 l, 140 w, 06 th. Quercus spp. Context 1399, SF 253 (C 2 nd ). Figure 6, P.28. 5

6 024. Thirty-five non refitting fragments of charred heartwood. Largest piece 46 l, 53 w, 18 th. All Quercus spp. Context 1208, SF 287 (C 2 nd ). Not Illustrated Twenty-one fragments of heartwood charcoal (labelled as 9 fragments when packed). Largest piece 50 l, 31 w, 26 th.? Quercus spp. Context 1293, SF 288 (C 2 nd ). Not Illustrated Fragment of charred piece of basketry; 2-strand twining (03 dia), 5 rows (01 dia) per centimetre. 11 l, 09 w, 05 th. Plant fibre not identifiable. Context 1205, SF 402 (C 2 nd ). Not Illustrated Tanner Row (General Accident). An excavation in the Colonia north of Micklegate close to the river frontage and north of the probable abutment of the bridge across the Ouse. Extensive waterlogged deposits with timber structures present. Site written up but not yet published Thin roundwood rod, no bark present- trimmed away. One end has two opposing facets cut to create blunt bifaced tip. Tip pierced by single 2.5 dia. through hole in face carved out from each side. Abraded surfaces. Other end broken away and missing. 467 l, 10 dia. Salix spp. Context 1201, SF 868 (Late C 2 nd Early C3 rd ). Figure 10, P Fragment from end of a coopered vessel cut from tangentially faced billet, with both edges split away and missing. But end finished with a concave chime bevel. Croze groove 02 w. 02 deep across inner face. Tapers rapidly away from butt end suggesting origin in an open vessel rather than a cask. Inner face hollowed, outer face backed 116 l, 14.6w, 17.7 th. Taxus baccata L. Context 1406, SF 1147 (Mid C 2 nd ). Figure 4, P Small wedge cut from radially faced heartwood. Saw marks across butt end. One face hewn to taper from butt end towards thin end. Old compression damage at thin end. Faint incised line across cut face perpendicular to wood grain. 39 l, 31 w, 14 th. Taxus baccata L. Context 1398, SF 1306 (Mid C 2 nd ). Not Illustrated Bung or stopper cut from radially faced/sub box-quartered billet. Facets around circumference cut to create rounded plan profile with slight taper towards hewn end. One end finished by rough hewing across the grain, other end sawn through to detach from parent timber l, 49.2 w, 28.2 th. Salix spp. Context 1383, SF 1307 (C 2 nd ). Figure 4, P Wooden vessel face turned from box halved block. Flat base with distinct step between base and wall, with single turned groove around outer face just below rim. In thirteen fragments, not yet refitted. (largest piece) 36 l, 28 w, 08 th. Pomoideae spp. Context 3272, SF 1691 (Late C 2 nd ). Figure 5, P Handle cut from radially faced billet. Sub rounded cross section, slight taper away from butt end. Other end terminates in an Fe based concretion. 88 l, 37 w, 20 th. Fraxinus excelsior L. Context 3272, SF 1693 (Late C 2 nd ). Figure 9, P Part of artefact cut from radially faced heartwood. both edges and one end roughly rounded off. Remaining end broken away and missing, suggestive of a missing axial extension or handle. Rough hewing marks on one face. 133 l, 132 w, 34 th. Quercus spp. Context SF 1731 (C2nd-4 th ). Figure 15, P Handle or spike reworked into a peg. Sub rounded cross section cut from radially faced heartwood. Expanded butt end, slightly rounded with plugged axial hole in centre of end. Single 6

7 decorative incised groove around shaft just above butt end. Artefact tapers evenly towards tip which is reworked with a single hewn facet cut to create a blunt point. Significant surface damage. 128 l, 42 w, 32 th. Fraxinus excelsior L. Context 3343, SF 1760 (Mid C 2 nd ). Figure 9, P Double sided comb cut from tangentially faced heartwood. Biconvex cross section with coarse teeth (4 per 10mm) sawn into one edge and fine teeth (7 per 10mm) sawn into opposite edge. Marking out line scored across faces to define plain central zone from sawn teeth. Surviving end with tooth guard slightly rounded off. Opposing end broken away and missing. Ends of teeth damaged especially on fine tooth edge. Ends of teeth damaged, tooth guard on fine tooth edge badly damaged. In two refitting sections, same context number but separate small find no s- SF 1774 and SF l, 54 w, 7.5 th. Buxus sempervirens L. Context 3343, SF 1774 and 1856 (Mid C 2 nd ). Figure 12, P Peg or thick pin, sub rounded cross section cut from radially faced billet. Knife trimmed butt end, tapering continuously to sub rounded cross section tip with fine shaving marks along length. 116 l, 17 dia. Alnus spp. Context 3362, SF 1835 (C 2 nd ). Not Illustrated Double sided comb cut from tangentially faced heartwood. Biconvex cross section with coarse teeth (5 per 10mm) sawn into one edge and indications of fine teeth (10 per 10mm) sawn into opposite edge. Marking out line scored across faces to define plain central zone from sawn teeth. Surviving end with tooth guard slightly rounded off. Fine teeth deliberately cut/trimmed away. Opposing end broken away and missing. Ends of coarse teeth damaged. Minor surface damage. In three refitting sections, same context number but separate small find no s- SF 1960 (2 refitting sections) and SF l, 103 w, 10 th. Buxus sempervirens L. Context 2361, SF 1960 (Late C 2 nd Early C 3 rd ). Figure 12, P Bung cut from radially faced board, sub-circular plan.. Edge/ends hewn to continuous bevel towards inner face. Slight charring on inner face. Charred patch on outer face. 68 l, 63 w, 26 th. Abies alba Mill. Context 2361, SF 2033 (Early C 3 rd ). Figure 4, P small pieces; largest has a remnant flat face with narrow groove along one side; another has smooth surface; other is entirely torn; all have blue staining, possibly from vivianite in surrounding matrix. Largest piece 25 l, 11 w, 06 th. All pieces Fraxinus excelsior L. Context 3372, SF 2101 (Early C 2 nd ). Not Illustrated Turned box lid cut from tangentially faced heartwood. Main disc has biconvex cross section with lower face hollowed out to create continuous rim around circumference to fit over and around container rim, with small flange. Inner face also exhibits centre mark and four pin marks from mandrel of pole lathe. Outer (upper) face decorated with pair of shallow turned grooves towards the edge and a single shallow turned groove about the centre point. Centre mark of mandrel also present on this face. 103 dia, 19 high. Fraxinus excelsior L. Context 4155, SF 2134 (Late C 2 nd Early C 3 rd ). Figure 5, P Fragment of oval basketry, now in three pieces; in cross-section shallow cone-shape, central point missing and outer edges broken. Stakes approx. 4mm wide radiate outwards; thinner canes approx. 3mm wide are twined (2-stand twining) around these: approx. 9 rows per 100 mm, but closer set towards the centre. All stakes reach into the middle, but near the centre the twining canes pass over two or more stakes at a time. All elements Salix spp. (identified by A.R.Hall). The object has been lined with yellow clay which has been pressed through the interstices of the basketry. A circular object such as a pot-base has been pressed into the clay. This clay has been impregnated with a pine resin pitch (identified by J.Evans). 190 l, 110 w, 40 th. Context 4155, SF 2135 (Late C 2 nd Early C 3 rd ). Not Illustrated. 7

8 042. Fragment of wall from a globular vessel spindle turned from a boxed heart blank. Faint turning marks around outer face towards surviving lower end. Base and rim broken away and missing. Dark patina on inner face only. 110 dia, 64 high, 07 th. Buxus sempervirens L. Context 2420, SF 2143 (Mid C 2 nd ). Figure 5, P Surviving half of the blade/scoop from spatula cut from tangentially faced heartwood. One face hollowed to create scoop, outer face trimmed to create flat back with bevelled edge. Scoop end rounded, opposing end has part of stopped shoulder and scar of broken away axial handle. Dark patina on inner (scoop) surface. Split longitudinally on mid- line. 176 l, 38 w, 15 th. Buxus sempervirens L. Context 4155, SF 2164 (Late C 2 nd Early C 3 rd ). Figure 8, P Double sided comb cut from tangentially faced heartwood. Biconvex cross section with coarse teeth (4 per 10mm) sawn into one edge and fine teeth (12 per 10mm) sawn into opposite edge. Marking out line scored across faces to define central zone from sawn teeth. Central zone decorated with convex moulded ridge. Both ends and tooth guards broken away and missing. Ends of teeth damaged especially on fine tooth edge. Ends of teeth damaged. In two refitting sections, same context number but separate small find no s- SF 2174 and SF l, 62 w, 08 th. Buxus sempervirens L. Context 4155, SF 2174, 2799 (Late C 2 nd Early C 3 rd). Figure 12, P Perforated block or fitting, cut from tangentially faced board. Single 20 dia through hole in face. both edges and one end rounded off to meet at distinct point midway along end. Mid-point defined by pair of incised grooves across thickness. Other end cut square to axis. One face finished, opposing face mostly split away. Some surface damage. 55 l, 63 w, 27 th. Fraxinus excelsior L. Context u/s but Roman, SF Figure 15, P Section of thin radially faced board. Both ends and both edges broken and missing. One corner detached but refitting. 225 l, 62 w, 05 th. Quercus spp. Context 4240, SF 2334 (C 2 nd ). Not Illustrated Stave from bucket or pail, cut from radially faced heartwood. Shallow 'V' section croze groove 23 mm from lower end. No howel or chime bevel. Slightly backed and hollowed. Both ends present, but both edges split away and missing. 309 l, 34 w, 12 th. Quercus spp. Context 4240, SF 2334 (C 2 nd ). Figure 4, P Base of bucket or pail. Single piece board cut from radially faced heartwood. Trimmed to sub circular plan. Edges/ends hewn to create a single continuous bevel on one face only except where lost through later damage. Some surface damage, abraded surfaces. In two refitting sections. 190 l, 190 w, 11 th. Quercus spp. Context 4240, SF 2334 (C 2 nd ). Figure 4, P Small shallow vessel face turned from halved block. Flanged rim decorated with pairs of parallel grooves on upper face and edge. Further pair of parallel grooves around outer face just below flange. Further pair of parallel grooves around inner face level with base of flange. Base almost flat but with 60 dia foot ring present, inside which is yet another pair of concentric turned decorative grooves. c. 30% estimate vessel equivalent. In seven refitting fragments. 100 dia, 20 high, 05 th. Juglans regia L. Context 2455, SF 2376 (Mid C 2 nd ). Figure 5, P Spindle whorl. Hemispherical cross section spindle turned from radially faced billet. Slight turned rim around each end of an 06 dia axial perforation. Pair of turned grooves around edge of artefact just above flat face. Flat face decorated with single turned groove parallel to edge. Faint turning marks on all surfaces. 29 dia, 14 high. Buxus sempervirens L. Context 1387, SF 2893 (Mid C 2 nd ). Figure 9, P.31. 8

9 051. Bucket with at least eight staves of pine; each is backed and hollowed, with planed edges, with croze groove present. and groove at lower end. All have slightly curved transverse section. Faint traces of fe hoops (SF 1964) on outer face, one towards rim, other towards base. Largest stave 260 l, 61 w, 11 th.? Pinus spp. Context 3268, SF 3085 (C 2 nd ). Figure 3, P Fragment of writing surface from writing tablet. Cut from radially faced heartwood. one edge and on end meet at right angle- knife score across face at end suggests this is from the corner of the writing surface with the 'frame' broken away along the defining cuts used to create the recessed surface. Other end and edge broken away and missing. No surviving inscriptions. 24 l, 07 w, 02 th. Abies alba Mill. Context 2260, SF 3117 (C 2 nd ). Figure 7, P Offcut from radially faced thin board. One end sawn square to axis of grain, long edge split away and missing, other edge and end eroded. 98 l, 75 w, 10 th. Quercus spp. Context 4153, SF 3122 (Late C 2 nd Early C 3 rd ). Not Illustrated Fragment from edge of writing tablet or small frame. Cut from radially faced board. Sub rectangular cross section. One edge has two opposing steep bevels along its length where centre recessed panel (writing surface?) has been cut away by later reworking. Both ends cut square to axis- narrower end through reworking. Single 02 dia. through hinge hole in face midway along length for binding cord. 9 l, 14 w, 06 th. Abies alba Mill. Context 4155, SF 3231 (Late C 2 nd Early C 3 rd ). Figure 7, P Wellington Row. An excavation in the Colonia north of Micklegate close to the river frontage and north of the probable abutment of the bridge across the Ouse. Extensive waterlogged deposits with timber structures present. Site written up but not yet published Labelled as a 'handle'. Packed in 10 fragments, not currently refitted. Roundwood or boxed heart conversion. Possible axial hole at one end. 43 l, 11 w (when recovered). Fraxinus excelsior L. Context 3113, SF 214 (C 3 rd ). Not Illustrated Section of halved roundwood, no bark present. In two refitting sections. 56 l, 21 w. 12 th. Acer campestre L. Context 2063, SF 315 (C 2 nd ). Not Illustrated Fragment of basketry worked in? 2/1 twining. Stakes are thin slivers of tangentially faced Quercus spp mm wide, 2-3 per100mm; twining elements are Salix spp., 6-7 dia (wood identification by A.R.Hall). Context 3143, SF 434 (C 2 nd ). Not Illustrated Tangentially chipping from outer surface of roundwood. In two refitting sections. 30 l, 31 w, 15 th. Salix spp. Context 3109, SF 439 (C 4 th or later). Not Illustrated Three non-refitting fragments of wood. (1) Radially faced, 74 l, 35 w, 26 th; (2) Radially faced 107 l, 19 w, 12 th; (3) Tangentially faced 59 l, 16 w, 08 th. All Quercus spp. Context 3153, SF 448 (C 2 nd ). Not Illustrated assorted wood fragments: (1) - (8) Roundwood, no bark present. No refitting sections, largest 44 l, 07 dia; (9) - (21) Radially faced thin board fragments, no refitting sections, largest 80 l, 23 w, 04 th; (22) - (26) Chippings, no refitting sections, largest 44 l, 10 w, 05 th. All Quercus spp. Context 3153, SF 458 (C 2 nd ). Not Illustrated Splinter torn away from parent timber. Tangentially faced conversion. In two refitting sections. 311 l, 14 w, 07 th. Quercus spp. Context 3156, SF 479 (C 2 nd ). Not Illustrated. 9

10 062. Bung or stopper cut from radially faced billet, spindle turned. Rounded terminal knob at one end with shoulders sloping outwards and downwards towards a circumferential ridge. Shaft below ridge has eight hewn facets forming polygonal cross section tapering towards other end. End of shaft sawn away. 34 l, 38 dia. Fraxinus excelsior L. Context 6228, SF 1538 (C 4 th or later). Figure 5, P Shallow dish or platter, face turned from tangentially faced board. Decorative groove around rim. Faint dark patina on inner face. Broad rim with steep drop to base. Foot ring around base. In 14 refitted sections, but badly warped and distorted during burial and treatment. 220 l, 260 w, 06 th, 35 high, c. 344 dia. Acer campestre L. Context 4231, SF (C 2 nd ). Figure 5, P Offcut from radially faced board. One edge cut at angle to create wedge shaped plan. 51 l, 47 w, 26 th. Quercus spp. Context 4230, SF (C 2 nd ). Not Illustrated Offcut or wedge. Cut from radially faced board. One face bevelled to create chisel tip. Other end cut roughly square to axis of wood grain. 53 l, 34 w, 11 th. Quercus spp. Context 4230, SF (C 2 nd ). Not Illustrated Offcut, sawn from end of tangentially faced board. 16 l, 61 w, 30 th. Quercus spp. Context 71885, SF (Later 1 st - early 2 nd ). Not Illustrated Two non-refitting offcuts: (i) Tangentially faced, possibly sawn from parent timber. 28 l, 50 w, 37 th. Quercus spp.; (ii) Radially faced, very eroded. 86 l, 20 w, 18 th. Quercus spp. Context 72123, SF (C 2 nd ). Not Illustrated Offcut from radially faced board. Original end hewn away, opposing end and edge broken away and missing. 105 l, 135 w, 20 th. Quercus spp. Context 72177, SF (C 2 nd ). Not Illustrated Three non-refitting offcuts: (i) Radially faced, ends broken away and missing. 77 l, 30 w, 19 th; (ii) Tangentially faced, ends broken away and missing. 77 l, 30 w, 19 th; (iii) Radially faced, ends broken away and missing. 77 l, 30 w, 19 th. All Quercus spp. Context 72209, SF21253 (C 2 nd ). Not Illustrated Section of roundwood, no bark present. Both ends broken away and missing. 48 l, 15 dia. Fraxinus excelsior L. Context 72217, SF (C 2 nd ). Not Illustrated Part of blade from spatula or flat shovel, Cut from radially faced heartwood. Both ends broken away and missing. 191 l, 70 w, 09 th. Quercus spp. Context 72225, SF (C 2 nd ). Figure 10, P Bung from coopered vessel. Disc cut from radially faced heartwood. Steep bevel cut around edge, except where lost to later damage. 60 l, 45 w, 22 th, 60 dia. Abies alba Mill. Context 72225, SF (C 2 nd ). Figure 4, P Fragment of radially faced board. In two refitting sections. Both ends and one edge broken away and missing. 182 l, 61 w, 03 th. Quercus spp. Context 72250, SF (Later 1 st - early 2 nd ). Not Illustrated Two non refitting fragments of same tool blade- flat spade or shovel. Cut from tangentially faced board with one face having hewn bevel at one end creating chisel tip. 2x fe nail holes through face (one in each piece). One has single 25 dia through hole in face on mid-line truncated by broken edge. Other has single 34 dia27 deep blind hole cut into face on mid- line. First piece 315 l, 88 w, 37 th; second piece 280 l, 62 w, 32 th. Quercus spp. Context 72250, SF (Later 1 st - early 2 nd ). Figure 10, P

11 075. Ten non-refitting radially faced board fragments. Largest 131 l, 25 w, 07 th. Quercus spp. Context 72250, SF (Later 1 st - early 2 nd ). Not Illustrated Tangentially faced board fragment, 56 l, 45 w, 09 th. Quercus spp. Context 72250, SF (Later 1 st - early 2 nd ). Not Illustrated Eight chippings, assorted conversions. Largest piece 96 l, 48 w, 20 th. Quercus spp. Context 72250, SF (Later 1 st - early 2 nd ). Not Illustrated Large bark chipping, fragmenting and delaminating in storage. 137 l, 65 w, 14 th. Wood species not identifiable. Context 72258, SF (Later 1 st - early 2 nd ). Not Illustrated Two non-refitting chippings: (i) radially faced, 73 l, 34 w, 16 th; (ii) tangentially faced, 30 l, 14 w, 09 th., both Quercus spp. Context 72231, SF (C 2 nd ). Not Illustrated Nine assorted non-refitting wood fragments in same bag.: (i) halved, 81 l, 28 w, 12 th; (ii) tangentially faced, 78 l, 49 w, 05 th; (iii) quartered, 39 l, 20 w, 10 th; (iv) radially faced, 118 l, 48 w, 08 th; (v) radially faced, 50 l, 20 w, 06 th; (vi) tangentially faced, 36 l, 21 w, 12 th; (vii) radially faced, 33 l, 13 w, 12 th; (viii) radially faced, 41 l, 11 w, 03 th; (ix) radially faced, 19 l, 05 w, 04 th. All Quercus spp. Context 72233, SF (C 2 nd ). Not Illustrated Four assorted fragments in same bag: (i) Fragment of radially faced board. One end and one edge broken away and missing. 82 l, 35 w, 05 th. Quercus spp; (ii) Fragment of box quartered heartwood. Both ends broken away and missing. 50 l, 42 w, 32 th. Fraxinus excelsior L; (iii) Section of roundwood in two refitting fragments. Bark present, both ends broken away and missing. 89 l, 09 dia. Salix spp; (iv) Fragment of curved roundwood, no bark present 46 l, 08 dia. Acer campestre L. Context 72225, SF (C 2 nd ). Not Illustrated Four non-refitting fragments of thin radially faced board: (i) 108 l, 75 w, 03 th; (ii) 62 l, 34 w, 03 th; (iii) in two refitting sections, 137 l, 70 w, 06 th; (iv) 78 l, 28 w, 04 th. All Quercus spp. Context 72224, SF (C 2 nd ). Not Illustrated Ten assorted wood fragments, non-refitting, in same bag: (i) Section of Roundwood, no bark present, both ends broken away and missing. 40 l, 16 dia. Salix spp; (ii) Section of Roundwood, no bark present, both ends broken away and missing. 41 l, 13 dia. Salix spp; (iii) Section of Roundwood, no bark present, both ends broken away and missing. 27 l, 09 dia. Salix spp.; (iv) Box quartered chipping. 38 l, 22 w, 18 th, Quercus spp; (v) Tangentially faced chipping. 39 l, 19 w, 13 th, Quercus spp; (vi) Quartered chipping. 26 l, 16 w, 12 th, Quercus spp; (vii) Quartered chipping. 33 l, 15 w, 11 th, Quercus spp; (viii) Radially faced board fragment. 35 l, 19 w, 03 th. Quercus spp; (ix) Tangentially faced chipping, 44 l, 09 w, 05 th, Quercus spp; (x) Tangentially faced chipping. 19 l, 06 w, 03 th, Quercus spp. Context 72224, SF21666 (Later 1 st - early 2 nd ). Not Illustrated Stave from small coopered vessel. 'V' section croze groove on inner face towards one end. No chime bevel, or howel. Badly warped during burial/treatment. Upper end broken away and missing. 183 l, 43 w, 05 th. Abies alba Mill. Context 72225, SF (C 2 nd ). Figure 4, P Offcut from box quartered timber. Sawn ends. 85 l, 77 w, 55 th. Quercus spp. Context 72245, SF (C 2 nd ). Not Illustrated Fragment of radially faced board, both ends and both edges eroded and missing. 130 l, 77 w, 14 th. Quercus spp. Context 72177, SF (C 2 nd ). Not Illustrated. 11

12 087. Eight non refitting fragments of wood in same bag: (i) Radially faced board fragment, 135 l, 39 w, 16 th.; (ii) Radially faced board fragment, 112 l, 32 w, 02 th; (iii) Radially faced chipping, 61 l, 15 w, 07 th; (iv) Box quartered chipping, 52 l, 14 w, 14 th; (v) Radially faced board chipping, 26 l, 22 w, 02 th; (vi) Radially faced board fragment, 28 l, 21 w, 02 th; (vii) Tangentially faced chipping, 32 l, 11 w, 09 th; (viii) Section of roundwood, no bark present, both ends broken away and missing, 61 l, 10 dia. All Quercus spp. Context 72209, SF (C 2 nd ). Not Illustrated Offcut, from radially faced fragment, both ends and one edge broken away and missing. 123 l, 30 w, 20 th. Quercus spp. Context 72245, SF (C 2 nd ). Not Illustrated Part (c. 25%) of a box lid spindle turned from a boxed heart billet. Faint turning marks present. Slight raised rim, pair of turned grooves around outside of rim. 50 dia, 07 high. Juglans regia L. Context 72245, SF (C 2 nd ). Figure 5, P Six non- refitting chippings: (i) radially faced, 33 l, 28 w, 09 th; (ii) tangentially faced, 53 l, 15 w, 06 th; (iii) radially faced, 46 l, 32 w, 04 th; (iv) radially faced, 50 l, 21 w, 04 th; (v) tangentially faced, 57 l, 13 w, 06 th; (vi) radially faced, 70 l, 10 w, 10 th. All Quercus spp. Context 72217, SF (C 2 nd ). Not Illustrated Five non-refitting pieces in same bag: (i) Peg, sub rectangular cross section, cut from box quartered heartwood. One end has single hewn facet cut to create chisel tip. 61 l, 14 w, 11 th. Abies alba Mill. ; (ii) Peg, cut from halved roundwood. One end has single hewn facet cut to create chisel tip. 37 l, 20 w, 12 th. Taxus baccata L.; (iii) Chipping. Radially faced conversion. 36 l, 11 w, 05 th. Quercus spp.; (iv) Chipping, radially faced conversion. 28 l, 06 w, 03 th. Quercus spp.; (v) Chipping, radially faced conversion. 16 l, 05 w, 02 th. Quercus spp. Context 72515, SF (C 2 nd ). Not Illustrated Three non-refitting offcuts in same bag: (i) Radially faced chipping, 20 l, 13 w, 04 th, Taxus baccata L.; (ii) Radially faced chipping, 32 l, 15 w, 05 th, Quercus spp.; (iii) Thin slice sawn from end of boxed radial timber, Quercus spp. Context 72515, SF (C 2 nd ). Not Illustrated Micklegate (Queens Hotel). An excavation in the Colonia north of Micklegate close to the river frontage and the probable western abutment of the Roman bridge across the river Ouse. Extensive waterlogged deposits with timber structures were present. Site written up but not yet published Wooden vessel, Object lost. Context 5017, SF 1016 (C 4 th ). Not Illustrated Roundwood stake, no bark present. Trimming facets cut to start tip. Tip and upper end broken away and missing. Very eroded with irregular cross section. 269 l, 30 dia. Sambucus nigra L. Context 5189, SF 2171 (C 3 rd ). Not Illustrated Hollow cylinder cut from boxed heart heartwood. Axial perforation roughly cut out with narrow gouge. Outer face follows natural surface of roundwood. Pair of circumferential grooves carved, not turned, around object either side of mid-point. Zone between these grooves has seven 04 dia through holes equally spaced around the circumference each with scored marking out lines to establish the centre point of each hole. One end partially charred. 47 l, 30 dia. Pomoideae spp. Context 5201, SF 2173 (C 2 nd ). Figure 15, P

13 096. Section of twig, partial bark present. No evidence of modification. 71 l, 05 dia. Too small for species identification. Context 5201, SF 2191 (C 2 nd ). Not Illustrated Offcut. Boxed heart conversion with two opposing faces and two waney edges (no bark present). One end sawn square to axis of timber, other end has four hewn facets cut to create sub rectangular cross section blunt tip. 143 l, 51 w, 34 th. Fraxinus excelsior L. Context 5201, SF 2240 (C 2 nd ). Not Illustrated Eight assorted non-refitting fragments in same bag: (i) Section of roundwood, no bark present. Both ends broken away and missing. 62 l, 27 dia; (ii) Offcut from thin radially faced board. Ends cut. 39 l, 50 w, 04 th; (iii) Fragment from radially faced board. One end cut, other broken away and missing. 73 l, 38 w, 07 th; (iv) Fragment of radially faced thin board. Both ends broken away and missing.32 l, 27 w, 03 th; (v) Fragment of tangentially faced board. Truncated tenon with sawn shoulders at one end. Other end broken away and missing. 48 l, 35 w, 15 th; (vi) Fragment of radially faced board. Both ends broken away and missing.78 l, 18 w, 07 th; (vii) Fragment of tangentially faced board. One edge end split away along grain tapering towards end. 66 l, 14 w, 08 th. (viii) Fragment of tangentially faced board. Both ends broken away and missing. 17 l, 15 w, 03 th. All Quercus spp. Context 5201, SF 2464 (C 2 nd ). Not Illustrated Offcut from radially faced thin board. Both ends sawn away. 31 l, 20 w, 05 th. Quercus spp. Context 5201, SF 2467 (C 2 nd ). Not Illustrated St Leonards Hospital. An excavation within the western corner of the fortress with dumped deposits and the foundation trench of the stone fortress wall. Site in course of being written up but not yet published Twenty non-refitting fragments in same sample. All Quercus spp. heartwood chippings except (xix) and (xx). (i) Tangentially faced, 168 l, 60 w, 18 th; (ii) Tangentially faced, 127 l, 22 w, 09 th; (iii) Tangentially faced, 61 l, 23 w, 12 th; (iv) Tangentially faced, 66 l, 22 w, 10 th; (v) Radially faced, 52 l, 30 w, 15 th; (vi) Radially faced, 53 l, 26 w, 20 th; (vii) Tangentially faced, 54 l, 26 w, 20 th; (viii) Tangentially faced, 40 l, 13 w, 13 th; (ix) Tangentially faced, 44 l, 18 w, 12 th; (x) Radially faced, 65 l, 17 w, 15 th; (xi) Tangentially faced, 62 l, 19 w, 07 th; (xii) Radially faced, 73 l, 18 w, 05 th; (xiii) Radially faced, 61 l, 12 w, 08 th; (xiv) Radially faced, 67 l, 16 w, 03 th; (xv) Tangentially faced, 43 l, 12 w, 08 th; (xvi) Tangentially faced axe chipping, 42 l, 32 w, 02 th; (xvii) Tangentially faced axe chipping, 53 l,30 w, 08 th; (xviii) Radially faced axe chipping, 47 l, 32 w, 02 th; (xix) Section of roundwood, no bark present. 46l, 14 dia. Corylus avellana L; (xx) Section of roundwood, no bark present. 60 l, 11 dia. Corylus avellana L. Context 6040, BF 4471 (Late C 1 st ). Not Illustrated Six non-refitting chippings in same bag: (i) radially faced heartwood chipping. 92 l, 80 w, 18 th; (ii) radially faced heartwood chipping. 61 l, 37 w, 24 th; (iii) Tangentially faced heartwood chipping. 93 l, 34 w, 20 th; (iv) Radially faced heartwood chipping. 76 l, 43 w, 12 th; (v) Radially faced heartwood chipping. 56 l, 30 w, 22 th; (vi) Tangentially faced offcut from board with large knot. 238 l, 40 w, 20 th. All Quercus spp. Context 6046, BF 4474 (Late C 1 st ). Not Illustrated Tent peg, complete. Cut from tangentially faced wood with sub triangular cross section. Both adjacent edges hewn to create even taper towards tip. Sub triangular notch cut into edge towards thicker end with wear marks. 268 l, 35 w, 30 th. Quercus spp. Context 6046, SF 4535 (Late C 1 st ). Figure 14, P

14 103. Four radially faced heartwood chippings in same bag. All Quercus spp. (i) 123 l, 44 w, 39 th; (ii) 147 l, 33 w, 10 th; (iii) 144 l, 27 w, 20 th; (iv) 93 l, 20 w, 15 th. Context 6044, BF 4475 (Late C 1 st ). Not Illustrated. Botanical Name Abies alba Mill. Acer campestre L. Alnus spp. Buxus sempervirens L. Corylus avellana L. Fraxinus excelsior L. Juglans regia L. Pinus spp. Pomoideae spp. Populus spp. Quercus spp. Salix spp. Sambucus nigra L. Taxus baccata L. Common English Name Silver Fir Field Maple Alders, exact species not determinable Box Hazel Ash Walnut Pines, exact species not determinable Fruit wood (Apple/Pear, exact species not determinable) Poplars, exact species not determinable Oaks, exact species not determinable Willows, exact species not determinable Elder Yew 14

15 Discussion: Patterns of deposition- Most of the sites where waterlogged Roman material has been excavated are on the west bank of the River Ouse, across the river from the fortress in the civil settlement. Roman waterlogged deposits do exist on the fortress side but those excavated do not appear to have been as extensive as those in the Colonia. To some extent this may be because the large scale redevelopment work that has taken place close to the Colonia river frontage in York has not been mirrored by similar development on the fortress side. Yet extensive archaeological work has taken place within and around the fortress at, for example, the Aldwark, Back Swinegate, Bedern, Coppergate, County Hospital and Hungate sites, not to mention the smaller interventions and observations made over the years within and immediately outside the fortress. These have revealed waterlogged deposits and features of Medieval and Anglo- Scandinavian date, so the conditions under which such deposits could form are present- but not during the Roman period. At the time of writing, it appears that either extensive waterlogged deposits north east of the Ouse are present, but just have not been reached in the areas which have been excavated, or that very different patterns of land use and waste disposal were taking place to either side of the river- a reflection of direct or indirect military control? No assemblages of wood have yet been recovered in the colonia prior to the early second century and no assemblages of wood from the fortress have been found after the early second century. The temporal distribution of artefacts is thus skewed and it is difficult to make direct comparisons. While the military assemblage from St Leonards site does not include any of the everyday items seen across the river, the reason for this may lie in the fact that deposits comparable to those seen so far in the Colonia have not yet been excavated in the fortress. Wood species exploitation- At least fourteen species of wood are attested in the assemblage. Native hardwoods predominate, though most of this material consists of chippings or offcuts from structural woodworking and the pegs associated with fastening of joints. These should really be classified and discussed with the structural timber component of each assemblage rather than with the portable artefacts considered here. None the less the chippings and offcuts indicate that some woodworking was taking place locally. No large bark chippings or waste from primary conversion and shaping is present but clearly wood was being brought into the settlement area and final working/trimming taking place if not actually on the site then fairly close by. None of the woodworking waste appears to be associated with making or finishing portable objects and it would appear these were being brought to York as finished artefacts. Although Roman masonry walls are among the most spectacular structural remains that survive, the wood is a reminder that much of early Roman York would have been of timber and wattle construction and that these methods would continue to be used alongside and as part of masonry buildings throughout the life of the fortress and colonia. Other uses for native hardwoods include tools and vessels. Tools include parts of solid wooden tools and wooden handles which may have belonged to all wood or partly metal tools. Vessels include examples of stave built containers and associated bungs and lathe turned containers or lids. Other wooden artefacts include combs, cut and carved from solid wood, and writing tablets. Unclassifiable are those pieces which have clear signs of worked joints or fixings but which have either been reworked, or so severely damaged that their original function can no longer be ascertained. A small but significant component of the assemblage is made from softwoods, which may be native (yew,) or imported (silver fir). The pine found may be native, as occasional finds of this wood are 15

16 present in late prehistoric contexts in the Yorkshire region, but it is possible some may have been imported. One hardwood known to have been introduced by the Romans is walnut. Two artefacts of made from walnut have been identified, both from second century contexts. Walnut is slow growing and at the limit of its climatic range and it is unlikely that a tree planted in Britain immediately after the conquest would have sufficient time to grow to sufficient size to produce wood used for the bowl. This artefact would have arrived as a finished artefact from overseas. To summarise, most of the wood used in Roman York appears to have been obtained locally and been supplemented by a small amount of imported wood. These imports are of finished objects or have been reworked from imported objects, rather than raw timber and presumably represent items which were of value to the owners or objects for which locally produced equivalents were somehow unsuitable. Types of artefacts: 1. Stave built vessels. Catalogue No 002, 028, 030, 038, 047, 048, 051, 072, and 084. Figures 2 (P.24), 3 (P.25) and 4 (P.26). Coopered vessels (excluding casks, which at YAT are classified as Structural Timbers) occur as near complete and as fragmentary artefacts. A number of bungs from casks have also been identifiedthese are not associated with their parent vessels. Small coopered drinking vessels or tankards were introduced in the century or so immediately before the Roman invasion and appear in richly furnished graves in the south east. While it is uncertain whether these are of local manufacture or were imported as finished goods, the widespread appearance of coopered technology is coincident with the Roman occupation. Coopered vessels may be open containers, such as buckets or closed containers such as casks. Casks were the bulk goods transportation containers of the day. While amphorae are frequent finds on sites of the period they were fragile and largely reserved for the carriage of liquids such as wine or oil. Casks are significantly more robust, could contain liquids or dry goods and could be repaired and reused many times- indeed most cask finds are from contexts where the cask has been reused for another purpose, such as lining a well or a pit. To date, the only certain cask fittings recovered from Roman York are bungs, used to stopper openings in casks that had been prepared to fill or empty the vessel. Examples appear in second through to fourth century contexts. Some are cut from native hardwoods- ash, willow- while others are cut from an imported softwood- silver fir. In the former cases, the bungs could derive from imported vessels or ones made within the province. In the latter case we can be confident that the object is imported, attached to the cask it sealed. Open vessels are better represented in York, though for the most part they only survive as fragments- part of a stave and a base from Tanner Row cut from oak (047 and 048), whose narrow v section croze groove suggests a container for liquid. Almost all of the other open vessels are cut from softwoods. Yew was used for an isolated mid- second century stave from Tanner Row (028) which could be locally made or imported. The remaining finds are made from imported wood such as the isolated silver fir stave from Wellington Row (084), the Tanner Row bucket with at least eight pine staves (051) and the near complete bucket from Skeldergate (002) with silver fir staves and an oak base. This latter is of interest because the bucket has an inverted profile- the rim is narrower in diameter than the base. While the Skeldergate bucket may have been deliberately deposited in the well as part of a closure act, it is difficult to avoid the conclusion that it was originally associated with drawing water. Both 16

17 this intact bucket and the fragmentary staves from Tanner Row were well made with iron hoops and fittings- though they had been stripped off the Tanner Row bucket before disposal the impressions made by the metal are still evident. Isolated softwood staves are known from several sites including London (Chapman 1977, 67 no. 489; Weeks et.al no. 10.3). Isolated fragments of yew staves have been recovered from a well in Welshpool (Boon 1961, 26). Bungs cut from silver fir are known from first century contexts at Caernarvon (Boon 1986, 99 no. 23), second century contexts at Billingsgate, London (Chapman no s 674 and 675), third century contexts at New Fresh Wharf London (Weeks et.al. 1986, 230). It is worth noting that though usually associated with coopered vessels, some of these bungs or stoppers may have come from Amphorae. Use of different wood species in buckets has been observed at Newstead (Curle 1911, 310 and plate LXIX) where three separate buckets each combined a pine base with oak staves 2. Turned vessels, lids and stoppers. Catalogue No 031, 040, 042, 049, 062, 063, 089, and 093. Figure 5 (P27). Whilst pottery vessels are the containers and tableware most associated with Roman sites, it is clear that wooden equivalents were available and used. Many imitate the forms of ceramic or metal equivalents, particularly in the use of flanged rims and foot rings, but this does not necessarily mean that they were cheap alternatives used only by those who could not afford the ceramic or metal. Walnut was used for a small lid (No 89) and a small decorated bowl (No. 49). As mentioned above, this wood is not a native species and would have arrived as imports- probably as finished artefacts. Walnut has historically been valued for its decorative qualities and this bowl is a relatively high status object. Certain woods are traditionally valued for their appearance and this includes the fruit wood used for no. 31. Ash and boxwood are frequently used as raw material for Roman tuned objects and examples are present as nos. 40, 62 and 42. While it is difficult to assign precise uses for the objects recovered, the dark patina on the inner face of no. 42 strongly suggests a use associated with food preparation or consumption. The large Field Maple platter from Wellington Row (No. 63) follows this trend, being reminiscent of a metal plate. The findspot of this object- immediately below the earliest metalled surface of the road leading to the river crossing suggests a foundation deposit and presumably therefore, this platter was a valued object rather than a poor substitute for a metal version. Most of the vessels are face turned- a log was split in half, roughly trimmed to shape with an axe and mounted on the pole lathe with the axis of the wood grain at right angles to the axis of rotation of the lathe. Spindle turning- with the axis of the wood grain parallel to the axis of rotation- was occasionally used. There is one wall fragment (No. 42) from a globular box created this way and an ash stopper (No. 62) spindle turned, but with the lower end trimmed and faceted to fit within what is likely to have been a bottle or amphora. This may suggest it was supplied as a turned object and modified later to stopper a particular vessel. Not all turned objects are vessels or stoppers. One tool handle (no. 55, below) was made by spindle turning a radially faced billet. There are also two lids, both circular in plan and made to close up a cylindrical hollow boxes. No 89, turned from walnut like no. 49 would have come from a button lid while no. 40, though turned from a native wood species, has a complex profile hinting at more than 17

18 a utilitarian function. Unusually, this was partially produced on a lathe with four mandrel spikes on one face rather than the usual one. Consideration of a logical sequence of events shows that the object must have been initially turned on a normal single spike mandrel and hollowed before being mounted on a four point mandrel for final finishing and decoration of the outer face. Far from being made by someone who had never turned a pyxis before (Pugsley 2003, 72) it demonstrates the activity of a woodworker with the command of a range of specialised tools and skills suitable for what was needed. The form of this object is reflected in an example from the Valkenburg in the Netherlands (van Rijn 1993, 189 no 55, fig 28) 3. Other containers. Catalogue No 010, 023, 026, 041 and 057. Figure 6 (P.28). There are three fragments of basket work (no s 26, 41 and 57), one heavily charred and all are too small to be certain how large the objects were from which they are derived, or indeed what form they may have taken. None the less these small fragments demonstrate the existence of a class of woven wooden containers that, owing to their fragility, rarely survive in the archaeological record. Similar material has been recovered from Newstead (Curle 1911, 108) and London (Pritchard et.al. 1986, 233) Much of the fragmentary board material from York is too thin to be derived from structural features and may well be derived from box-like containers. No 23 would make a near complete end panel from a small box, perhaps slotted into grooves in the side as there are no surviving nail or peg holes present. No 10 does have peg holes- with three in situ pegs- that would allow it to be fastened to the rest of the box, while the knife scoring across one of the faces could indicate either reworking or damage accumulated in use. 4. Writing tablets. Catalogue No 052, 054. Figure 7 (P.29). Writing tablets are thin slices of wooden board and belong to one of two main types. Some are plain thin wooden boards with smooth surfaces upon which text was written with ink. Others are slightly thicker boards with a sub-rectangular recess let into one or both faces, leaving a raised edge around the object. This recess would be filled with a thin layer of wax and texts could be inscribed into that waxed surface, corrected and/or erased with a metal stylus. Both types might be pierced for cords that could be used to fasten several tablets together much like a book or a notepad. Similar tablet fragments have been recovered from Carlisle (McCarthy et al 1982, 83; Caruana 1992, 69), London (Chapman 1977, 67 no. 486; Chapman 1980, ; Chapman et.al. 1986, 227), Newstead (Curle 1911, 308 and plate LXXX), Scole (Rogerson 1977, 206 and figure 87) and Vindolanda. Fragments broken in a very similar manner to those from York have been identified at The Valkenburg in the Netherlands (Van Rijn 1993, 185, no s 52 and 53, fig 27). among other locations and the information they contain has revolutionised our view of Romano-British society. The two fragments from York are of the waxed variety of tablet- part of the edge and a corner of a writing surface. Unfortunately neither contain any text. None the less they are significant finds. They are a reminder that the inhabitants of Roman York were part of a literate society and that the waterlogged deposits of Roman York have the potential to preserve this type of document. 18

19 5. Tools and handles. Catalogue No 016, 021, 027, 032, 034, 043, 050, 055, 071 and 074. Figures 8 (P.30), 9 (P.31) and 10 (P.32). Two scoops are present in the assemblage, both have part of the blade or scoop but only one with has part of the handle surviving. Both are cut from straight-grained boxwood. The blades have a hollowed inner face and an angled back face. The end is rounded off and while the blade expands widthways towards the end it tapers in thickness towards the tip. More complete examples have lobed protrusions, much like the pommel and guard of a sword hilt, at each end of the handle; only part of a guard at the junction of handle and blade survives and any decorative carving that may once have been present has been lost. There are two pieces of textile working equipment in wood. One is the boxwood spindle whorl from Tanner Row (no. 50). This has been spindle turned from a billet of radially faced heartwood, has an axial perforation and limited decoration. A thin roundwood rod (no. 27), with bark trimmed away, terminating in a crude bifaced tip which in turn is pierced by a single small hole is a potential spindle or needle. A disc shaped spindle whorl in boxwood was recovered from a late third century well at Chew Park, Somerset (Rahtz and Greenfield 1977, 277 no. 8). These objects are usually made from bone or antler (cf an example of similar hemispherical form from The Valkenburg; Verhagen 1993, 347 no 1, fig 5). Wooden tool handles are represented by an example (no. 32) with the corroded remains of the iron tool itself still embedded in the end. One handle (no. 55) also cut from ash has disintegrated in storage and its form cannot now be reconstructed. A third handle (no. 34) has a rounded butt end and tapering cylindrical body. An inserted plug is present in the butt end but the other end has been cut away when the object was reworked into a peg. All three handles are cut from ash. Pieces of two flat bladed oak tools were found at Wellington Row but none survive well enough to be certain of their original form. No. 71 consists of fragments from a blade, no.74 could be part of the handle and blade of a shovel. One tool is readily identifiable- a mattock head from Skeldergate (no. 16). This piece is cut from oak and tapers towards a blunt point at each end. A centrally placed hole through the face is the attachment point for the handle. The object had suffered some damage before disposal but as with similar objects, shows no sign of having had a metal shoe fitted. 6. Reworked spoke. Catalogue No 017. Figure 11. One of the objects (no. 17) from the well at Skeldergate is the truncated stump of a spindle turned cylinder cut from a radially faced billet of Field Maple. One end terminates in a full width tenon, the other has been roughly hewn away to create a short blunt point. The object could originally have been part of a rail from a piece of furniture, or part of a ladder rung. It would be unusual to have a ladder rung prepared from turned wood, but not impossible. The absence of a hole for a peg or nail suggests the object was held in place by compression and in that context, it is quite likely that this piece is a fragment from the spoke of a wheel. Such spokes are known =from late iron age contexts as at Glastonbury Lake village (Bulleid and Grey 1911, 341 X67)- wood species not specified. More complete examples from Roman contexts turned from willow have been recovered at Newstead (Curle 1911, 298 and plate LXIX). However a better parallel has 19

20 been found at The Lanes Carlisle (McCarthy et. al ) the spokes of a wheel also cut from field maple have terminals at the nave end of very similar form to the fragment described here. 7. Combs. Catalogue No 022, 035, 037 and 044. Figure 12 (P.34). Four wooden combs are known, all from the second through to early third centuries AD. All are cut from tangentially faced slabs of boxwood (Buxus sempervirens L.), a fine grained, dense and hard wearing wood. As with similar Roman wooden combs, they are double sided, with a set of coarse teeth on one side and a set of fine teeth on the opposing side. The spacing of the teeth varies there are generally between 4 and 5 teeth per 10mm on the coarse side and 8-10 teeth per 10mm on the fine side. Exceptionally, one of the Tanner Row combs (no 44) has an extremely fine 12 teeth per 10mm. A pair of marking out lines is generally scored across the width of the comb and these define the base of the teeth on each side, a point at which the cutting of the teeth stopped to leave a central zone or spine separating the two sets. Marks on the exposed sides of the teeth show that they were cut with a saw. To achieve the evident spacing, this must have been a very thin saw blade. It is unclear whether the sawing was done with a single bladed saw or if a tool with several parallel saw blades was employed- minor irregularities in the spacing could result if the teeth were cut individually or in the setting (and repositioning) of a multiple bladed saw. This zone would seem an obvious place for any decoration and several decorated examples are known from Roman Britain but little is evident on these four from York. A moulded ridge is present on no. 44, which is the comb with exceptionally close tooth spacing The longitudinal cross section is bi-convex- the thickness of the object reduces towards the tip of the teeth on each side and the ends of the teeth are rounded off, rather than finishing in a sharp point. The teeth do not continue to the ends of the artefact, where an area of unsawn wood is left to guard the vulnerable teeth from lateral damage. These terminals may be shaped, typically rounded off to create a curved end to the comb, though sometimes simply cut off square and the sharp corners rounded off. There is no elaborate decoration to the examples described here. These combs were made for personal grooming and might have belonged to men, women or children. Though well made they are not perhaps the most expensive or elaborate artefacts but are good examples of everyday personal artefacts. Boxwood combs are relatively common finds from Roman sites. Simple undecorated examples have been found in first century contexts at Caernarvon (Boon 1986, 98 no s 1 and 2),second century deposits at Carlisle (McCarthy 1991, 34), third century contexts at Chew Park, Somerset (Rahtz and Greenfield 1977, 277 no s 9 and 11), London (Weeks et. al. 1986, 230 no. 10.2) 8. Tent Peg. Catalogue No 012. Figure 13 (P.35). The one identifiable artefact from a military context at St Leonards is a complete tent peg (no. 102) cut from radially faced oak. Cut to a sub triangular cross section and tapering towards a sharp tip, there is a notch in one edge to engage with a guy rope. This example has a flat head- others are known that have pointed head ends (Caruana 1992, 70) 20

21 Tent pegs are relatively common finds from military sites and are a reminder that units of the Roman Army were expected to take the field against an enemy rather than await them in fortified locations. Outside of wartime activities, units might also be tasked with building camps and living under campaign conditions as part of their training. Accommodation for the troops was provided by leather tents, fragments of which are also common finds from waterlogged deposits in Roman military sites. Near contemporary pegs have been recovered from Annetwell Street Carlisle (Tullie House Museum online catalogue object W 12048) and Melandra Castle, Derbyshire (Wild, JP ) 9. Fastenings and fittings. Catalogue No 003, 004, 007, 019, 020, 029, 036 and 065. Figure 14 (P.36). Twisted lengths of wood rope were used as ties and there is one fragment from Rougier Street (No. 19) representing what may have been a common feature. More common are pegs, cut from pieces of timber rather than small diameter roundwood, and often carefully finished, as befits objects used to attach one object to another. Two are reworked from other artefacts (Nos. 34 which is the reworked handle referred to above and no. 04 which comes from a timber which had at least one peg hole itself). Two wedges are also present-typically used to make a timber joint fit more tightly together (such as no 65) or for a variety of securing purposes. The choice of yew for no 29 suggests a more specialist application as this very tough softwood is not a species casually chosen. One object is an oddity. No. 07 is a bracket or frame element, cut from a stem/branch junction that was exploited to help form the right angle while retaining the direction of the wood grain. The identification is not certain as the object has been reworked and broken before disposal to create a blunt tip, making the object into a wooden staple. However a possible suggestion for the original form of the artefact is a find from the Valkenburg (van Rijn 1993, 184 no 45, fig 25). This object has a carved return where the York example has been reworked and is interpreted as the handle for a beaker. 10. Unidentified artefacts. Catalogue No 005, 033, 045 and 095. Figure 15 (P.37). Four artefacts cannot be identified at present. No. 5 is a small chipping from the outer face of a section of roundwood, but the chipping has parts of three holes cut into or through the parent timber. This piece is derived from the reworking of a finished timber but what that timber was cannot be known at present. No. 33 is the end of an object or tool. Superficially similar to a very thick and roughly finished lid or bung, one edge has the scar of a broken away extension, shaft or handle along the axis of the grain. Again, the parent object cannot currently be identified, but a very similar smaller version of the same object, broken in the same manner, was retrieved from Bar Hill (Robertson et. al. 1975, 54 fig 17). No. 45 is a perforated block. One face is damaged but the object is otherwise complete. Marking out lines are present and the object has clearly been carefully shaped but its function cannot be 21

22 identified. The object was recovered from Roman levels but the exact context, and consequently its date, is also unknown. Finally, no. 95 is a hollow cylinder, with a spindle turned outer face and the axial perforation finished with a narrow gouge. A pair of turned grooves define a band around the circumference at the midpoint. Seven through holes are cut through the wall of the cylinder in a row around the middle of this band and marking out lines show these were intentionally placed. These holes do not align across the diameter of the object. The cylinder is reminiscent of a hinge, of which good examples in bone are often found on Roman sites. However there are more holes present than on hinges. The absence of alignment and slightly irregular spacing of the holes would seem to exclude a function as the nave of a small wheel, though a function as a seven spoked wheel nave cannot be absolutely ruled out. One end is charred, though whether this was incurred by the object while it was being used or whether it was in the process of being recycled as fuel cannot be determined. 22

23 23