Internet Archaeol 9. Hunter-Mann. Finds catalogue

Hod Hill brooch. Copper alloy with white metal (possibly silver) inlay. Very heavily corroded; distorted. Head broken and hinge missing; area below hinge has wide horizontal concavity bordered by pairs of narrower grooves; slight step in from head to tapering upper bow decorated by vertical zigzag band bordered by paired grooves; vestigial side lugs at base of upper bow. Channels on head and upper bow retain traces of white metal. Lower bow, catch plate and foot-knob heavily obscured by corrosion. Length 56mm, width of head 19mm.

The Hod Hill brooch is a very common Claudio-Neronian form, which had many variants (see Olivier 1988, 46-9). Given the number found on military sites not occupied until the Flavian period it is clear that, in the north at least, the form continued in use into the 70s and 80s. For example, Hod Hill brooches are the commonest of the Claudio-Neronian forms found at Castleford, West Yorkshire, in contexts dating to c.AD 71/4-86 (Cool and Philo 1998, 30 Table 6), and one was also found at Carlisle in a context dated to AD 71/2-78/9 (Snape 1993, 13).

No. 1 is the second example of a Hod Hill brooch to have been found at Brough. The first was recovered from a gully within the Hadrianic fort (Wacher 1969, 7; 92 (no.32); fig.39). Both were probably initially associated with the first military occupation of the site.

Penannular brooches

Penannular brooch. Copper alloy. Circular-sectioned hoop; knob terminal with diagonal grooves. Approximately half circumference and one terminal extant. Diameter 29mm, hoop section 2.5mm.

Penannular brooch? Iron. Part of circular-sectioned ring becoming rectangular-sectioned at terminal and bent back on itself; ring retains part of attachment of pin. Length 28mm, section 5mm.

No. 2 is an example of a Fowler A2 penannular brooch and no.3 is probably a Fowler C (Fowler 1960, 152). The former appears to be a common northern form in use throughout the Roman period, while the latter is more likely to belong to the 1st century BC or 1st century AD (Fowler in Crummy 1983, 19).

Hairpins

Hairpin. Bone; strip of long bone. Flat lyre-shaped head with diagonal grooves over lower part on both faces; oval-sectioned tapering shank, tip broken. Length 100mm, maximum shank section 5mm x 4mm.

Hairpin or needle shank. Bone; strip of long bone. Circular-section shank tapering to point. Length 61mm, shank section 4mm.

Bone hairpins are ubiquitous finds on most Romano-British sites where females were present, and most fall into a small number of repetitive forms with relatively simple heads (Crummy 1979; Greep 1995, 1113-20). Examples with more elaborate heads such as no.4 often tend to be unique and so are not susceptible to typological dating. No. 4 was found in the early 4th century levelling dumps and must therefore be of that date or earlier.

Bangle

Bangle. Glass. D-sectioned, blue/green bangle with central cord with ground of very dark glass appearing black and probably 3 narrow opaque white strands, right-hand twist; 2 opaque white oval dots on alternate sides. Both ends broken. Internal diameter c.45mm (17% of circumference extant). Section 9mm x 7.5mm.

Glass bangles with a central twisted cord of contrasting colours form Type 2 of the Kilbride-Jones (1938) classification. In a survey of those from East Yorkshire, Price (1988, 342) further sub-divided them according to the number of cords and no.6 clearly belongs to her variant Ai which has a single cord and may, as here, have side eyes. This was the commonest variant of the Type 2 bangles found in East Yorkshire (ibid., 359-60 nos.1-24). The dating evidence suggests the Type 2 bangles had a floruit during the Flavio-Trajanic period though many, like no.6 found in the post-medieval ploughsoil, have not been found in usefully dated contexts.

Bangles like these have close associations with military sites and, given the history of Brough during the Flavio-Trajanic period, it is very possible that the bangle from which no.6 came originally arrived at the site as a result of the 1st century military activity. It should be noted, however, that there is evidence that fragments may have been treasured long after they were broken (Price 1988, 354), and so this link may be fortuitous.

Beads

Bead. Opaque turquoise glass. Short cylindrical. Diameter 3.5mm, thickness 2.5mm.

Both of these beads were found in 3rd century contexts and therefore the supposition must be that they are Roman. It should be noted, however, that neither belongs to standard late Roman forms (see for example Guido 1978, 91-100). In the case of no.8 this is not surprising as tiny beads such as this are generally only found as part of sieving programmes, as this one was, and are therefore under-represented in the literature. There are more difficulties in accepting a 3rd century date for no.7 as it is made of a glossy opaque turquoise glass and the author is not aware of any beads or other glass objects in this colour from a secure Roman context. There is, therefore, the possibility that it is a relatively modern intrusion.

Hobnails

Shoe sole. Iron. Hobnails corroded together and thus retaining shape of outer edge of front of sole. 12 hobnails remaining nailed in 3 rows. 70mm x 37mm.

A total of 45 hobnails could be identified from the X-radiographs, of which 12 were corroded together and thus retain the shape of part of a shoe sole. These twelve have been catalogued as no.9; the rest are listed in the archive. No. 9 was found in the ditch fill that may have been part of the early 4th century dumping on the site, and 13 of the other hobnails came from the same context group. Of these, one group of four and two groups of two were found corroded together. It is impossible to ascertain whether all of the other hobnails in this context came from the same shoe as no.9, but the fact that groups were corroded together suggest they originated as part of one or more discarded shoes, and not as casual losses from shoes whilst they were being worn. A similar origin from discarded shoes can also be suggested for the remaining hobnails from Trench 1. They were found in two groups, and in one case (context group 1.27) were also found corroded together. The four from Trench 2 by contrast appear to be isolated examples and perhaps represent casual loss.

As far as is possible to see from the X-radiographs the hobnails have typical pyramidal heads (Manning 1985, fig.32 no.10), but there are also four studs of a similar size with flatter heads from the same context group as no.9. These might also be hobnails.

The shoe from which no.9 came was obviously very heavily nailed as three rows of hobnails can be traced around the outer edge. It has been suggested that such heavy nailing may be indicative of the caligae which was originally a soldier's shoe but which came to be adopted by the civilian population (Rhodes 1980, 113-14).

5.5.2 Military equipment

Pendant

Pendant. Copper alloy. Heart-shaped pendant with trefoil terminal; rectangular bar across top with two close-set hinge junctions. Front retains traces of white metal plating. Length 67mm, maximum width 58mm, thickness 1mm.

This pendant belongs to a small group that Oldenstein (1977, 130) includes in his heart-shaped pierced military pendants even though the piercing on them is either vestigial or non-existent as here and on a very similar pendant from Osterburken (ibid., Taf. 33 no.259). The Osterburken example has a terminus post quem of AD 150 and the hinge junctions can be compared with those on 3rd century military pendants. A late 2nd to 3rd century date, therefore, seems most appropriate for this piece.

Belt fitting

Belt plate. Copper alloy. Rectangular plate with two loops for hinge bar at one short end and two rectangular projections at corners at other (one missing); three pointed perforated lugs on back. Loop and lugs cast in one piece with plate. Groove parallel to each edge and diagonal edges diagonally nicked. Length 57mm, width 19mm.

The most likely identification of no.11 is as a buckle plate from a belt. The general shape and positioning of the hinge fittings and mounting loops are typical of the enamelled buckle plates of the 2nd century; see for example one from a Trajanic to Antonine barrack at Caerleon (Webster 1992, 123 no.88). The perforated lugs for attachment to the belt are unusual as a solid lug with a burred end is more often found. The perforated lugs and virtually plain upper face link no.11 to a group of belt fittings known in German as Waffenschildchen (Oldenstein 1977, 191-2, Taf. 60 nos. 749-55) which may be dated to the second half of the 2nd century.

5.5.3 Writing equipment

Stylus. Iron. Wedge-shaped expanded eraser without shoulders; oval-sectioned shank tapering to point. Length 92mm, width of shank 9mm.

This simple stylus is an example of a Manning type 1 stylus (Manning 1985, 84), a common form which is not closely datable within the Roman period. Styli have been relatively common finds during excavations at Brough though no significance should be read into this as styli are frequently common finds on late Romano-British sites, even rural ones. There are two simple styli very similar to no.12 (Corder 1935, 33 no.14, fig.8; Wacher 1969, 99 no.17) and three more elaborate iron examples of Manning type 4 (Corder and Romans 1938, 47 no.3, fig.10; Wacher 1969, 99 nos.20 & 36, figs.41, 43). A copper alloy stylus was also found during 1933 (Corder 1934, 32 no.4, fig.7).

5.5.4 Household

Glass drinking vessels

Cylindrical cup. Colourless glass. Vertical rim, edge fire-thickened; straight side. RD 100mm. WT 1.5mm, PH 34mm. (EVE 0.4)

Cylindrical cup. Colourless glass; some small bubbles. 2 joining vertical rim fragments, edge fire-thickened; straight side. RD c.80mm, WT 1mm, PH 11mm. (EVE 0.2)

Hemispherical cup. Green-tinged colourless glass; many small bubbles. Convex-curved side curving into concave base; 2 pulled-up blobs on lower body. BD c.30mm, WT 1.5mm, PH 15mm. (EVE 0.4)

Two types of glass drinking vessels are represented. There are two examples (nos.13-14) of the colourless cylindrical cup with double base ring (Isings 1957, form 85b; Cool and Price 1995, 82) which were very common in Roman Britain during the later 2nd and early to mid 3rd centuries. Sites often produce large numbers of these cups and so it is not surprising that the form has already been recorded from Brough. Two examples of the plain variant like nos.13 and 14 (Corder and Romans 1938, 42 nos.2-3, fig.8) and two examples of the trailed variety (Corder 1934, 34 no.14, fig.7; Corder and Romans 1938, 42 no.1, fig 8) were found during the 1930s excavations.

The second cup type (no.15) is hemispherical with pinched-up decoration (Cool and Price 1995, 86). It overlaps in date with the cylindrical cups as it is the commonest cup type of the mid to late 3rd century.

Glass jugs

Conical jug. Blue/green glass. Straight side with tip of extension trail from lower handle attachment retaining 3 pinched projections. 24mm x 22mm.

Conical jug. Blue/green glass. Carinated lower body fragment retaining 3 light ribs. 78mm x 28mm, WT 1mm. (EVE 0.14).

Body fragment. Blue/green glass. Straight-sided fragment with vertical rib. Probably from same vessel as no.17. 42mm x 23mm, WT 1mm.

Nos. 16 to 18 all come from conical jugs of Isings (1957) form 55, a very common type in use during the later 1st century and the first half of the 2nd (Cool and Price 1995, 120). A yellow/green example was recorded from the earlier excavations (Charlesworth 1969, 106), and it is possible that two other fragments from these excavations are from the same form. Both of those fragments, however, just retain part of the pushed in base and this is found on other contemporary forms such as globular jugs (Isings 1957, form 52) and collared jars (ibid., Form 67c).

Glass bottles

Square bottle. Blue/green glass; many bubbles and impurities. Concave base fragment. Base design — ‘L'-shaped angle mouldings, 1 complete and 1 fragmentary surviving. Base width c.55mm. (EVE 0.28)

Square bottle. Blue/green glass. Lower body and base fragment. Base design — square moulding parallel to edge. PH 18mm. (EVE 0.28)

Square bottle. Blue/green glass. Corner of base retaining parts of 2 sides. PH 24mm. (EVE 0.28)

Cylindrical bottle? Colourless glass. Narrow horizontal shoulder curving up to neck and down to side. 20mm x 16mm, WT 4mm. (EVE 0.14)

Blue/green square bottles (Isings 1957 form 50) were the commonest glass container in Roman Britain from the mid 1st to early 3rd century (Cool and Price 1995). Fragments are found on most sites that show even the slightest tendency towards Romanisation, and on military and urban sites they are often found in very large numbers. Here the form is represented by nos.19-21, together with five fragments noted in the archive. Cylindrical blue/green bottles (Isings 1957, form 51) such as no.22 were as common as the square bottles in the later 1st century but went out of use early in the 2nd.

The colourless fragment no.23 is likely to have come from the form of cylindrical bottle in use during the later 2nd and early 3rd century (Cool and Price 1995, 200). Whilst not an uncommon form, these bottles are never found in the volume of the blue/green bottles and do not seem to have been utilitarian containers as the blue/green ones undoubtedly were.

Iron bowl

Bowl; iron. Flat flange rim sloping up slightly; convex-curved upper body. RD c.140-150mm. PH c.35mm.

It is rare to find bowls made of iron such as no.24. This is not so much because they were rare during the Roman period but rather, as Manning has pointed out in discussing a similar one from Verulamium (Manning 1972, 176 no.47, fig. 65), because the thin sheet they are made of is so easily corroded. The Verulamium example is very closely dated to the period AD 155-160 because it was engulfed by the Antonine fire that swept through Insula 14, and given that no.24 was found in the dumped material over the site it must have a terminus ante quem of the early 4th century.

Bucket and suspension fittings

Bucket mount? Iron. Round-ended plate with perforation in rounded end and broken across second perforation. Length 55mm, width 19mm.

Fitting? Iron. Carefully made thin circular terminal with small central perforation; square-sectioned bar becoming circular at broken end. Length 96mm, width of terminal 20mm, maximum section 11mm.

The identification of no.25 as a bucket mount needs to be viewed with some caution. From the X-radiograph no.25 has the perforated rounded end that could be intended for articulation with the bucket handle (see for example Scott 1990, fig.118), but no investigative conservation has been carried out to confirm the identification. The identification of no.26 is even more tentative. It is clearly a very carefully made element of a composite item. Possible identifications include elements of a lamp hanger (Manning 1985, 99-100 no.P6, fig.26, pl. 45) or a cauldron suspension chain (ibid., 100-101 no.P9, fig.27).

Knives and handles

Knife blade. Iron. Broken thin triangular blade tapering to point; straight back and curved cutting edge. Length 88mm, maximum width 31mm.

Knife blade. Iron. Straight cutting edge; convex-curved back; rectangular-sectioned handle plate with rivet. Area of brass plating around rivet on one side. Point of knife broken. Length 86mm; maximum width of blade 29mm.

Handle; section of long bone. Fragment from lower edge of handle. 3 horizontal grooves around bottom of handle; cross-hatched grooved design above. Traces of iron corrosion on exterior and one broken edge. 28mm x 21mm.

Two knives can be identified with certainty. The blade fragment no.27 probably came from the form that Manning (1985, 115, fig.28) has classified as a type 15 and described as ‘the commonest form of all'. Given that the tang end of the knife is missing, this identification cannot be totally secure and the blade pattern of a straight back and curved cutting edge would also fit his types 11 and 12. Another possible blade fragment (identified only from the X-radiograph) is noted in the archive. In contrast to no.27, no.28 has a straight cutting edge. It is possible that it is the blade from a folding clasp knife, similar to examples from a later 2nd or 3rd century context at Canterbury (Greep 1995, 1144 no.1005, fig.503), and a late 4th century one at Baldock (Manning and Scott 1986, 162 no.609, fig.69).

The bone handle no.29 could have been used with either a knife or some other implement. The fragment comes from a one-piece handle decorated with incised cross-hatched grooving and when complete may have been similar to the less carefully decorated example from Colchester (Crummy 1983, 109 no.2929, fig.110). Such decorated one-piece handles are generally of 3rd or 4th century date (op. cit.), though this example came from a 2nd century context.

Quernstones

Quern; upper stone. Spilsby sandstone; greyish white, fine- to coarse-grained with mainly sub-rounded to (less commonly) rounded grains, poorly sorted, moderately compacted, calcareous, with some 2-4mm pebbles of quartz, chert and ‘lydite'. Terminal Jurassic to lowest Cretaceous. Skirt fragment with face ground smooth; upper surface concave with flat margin to sloping edge. Edge and upper face roughly dressed. Diameter c.450-500mm, c.6% circumference preserved, height at outer edge 75mm.

Quern; lower stone. Sandstone, pale brownish grey, mainly medium-grained, fairly well sorted, moderately compacted, with a few small flattish and other voids, and traces of slightly undulating laminar bedding. Middle or Upper Jurassic, almost certainly Crinoid Grit. Eye and breast fragment with face ground smooth and base roughly dressed flat. Eye diameter at face c.50mm, thickness at eye 55mm, maximum length 180mm.

Both of the identifiable quernstones come from flat rotary stones and are obviously Romano-British products. The production of such stones was already underway in northern Britain by the late 1st century (Welfare 1985, 157) and continued throughout the Roman period. Insufficient of nos.30 or 31 remains to suggest a more precise date. At least one imported quernstone is suggested by a quantity of lava fragments found in the early 4th century dumps. The fragments are now very friable and retain no recognisable dressed faces but they almost certainly came from a flat lava quern imported from Germany. Lava quernstones are commonest on military sites in the north of Britain and overall the trade in them had declined by the 3rd century (Buckley and Major 1998, 245), so it is possible that the fragments from Brough could relate to the early military activity at the site. It may be noted, however, that approximately one-third of a lava quern was found in occupation contexts of the late 2nd to early 3rd century in Building G1 at Manor House (Wacher 1969, 102 no.8, fig.45) so the fragments, like so many of the other finds, could relate to the later 2nd to 3rd century occupation.

5.5.5 Fasteners and fittings

Locks and keys

Key handle. Copper alloy. Fleur-de-lys handle with keyhole-shaped perforations; rectangular-sectioned base with 3 cordons from which projects smaller rectangular socket with three cordons. Corroded iron within socket from broken stem of key; socket broken. Length 52mm, maximum width 33mm.

Lift key. Iron. L-shaped key with bit of 3 teeth. Outer tooth and majority of stem missing. Dimensions 30mm x 28mm.

Slide key. Iron. Handle with pointed loop; bit with two triangular teeth and one square tooth. Length 91mm, length of bit 25mm.

Barb spring padlock bolt. Iron. Rectangular-sectioned terminal and plug; at least three barbs welded to tip of central bolt. Length 108mm, terminal section 22 x 19mm.

Items associated with security are well represented in this assemblage. There is one example (no.32) of the common fleur-de-lys key handle (see Allason-Jones and Miket 1984, 144 no.3.347 for references), two keys (nos.33-4) and an element of a lock (no.35). The lift key no.33 would have been used with a simple door lock (Manning 1985, 90). Only the bit survives on no.33 but two complete examples have been found previously (Corder and Romans 1938, 47 nos.1-2, fig.10). The slide key no.34 was designed for use with a more complex tumbler lock. The relatively small size of such slide keys with a straight bit (Manning 1985, 93 Type 2), together with the fact that they are as often made in copper alloy as iron, could suggest that they were intended for use in locks that secured chests. No. 35 is the bolt from a barb spring padlock and the plug-like end shows that it came from the variant with a looped hasp (ibid., 96 type 2). An iron implement found in 1936 is very likely to have been a key used with a similar padlock (Corder and Romans 1938, 48 no.11, fig.10).

All of these items could have been in use at any time during the Roman period. Crummy (1983, 126 no.4161) has suggested that the fleur-de-lys key handles date to the period after AD 150. However, they were clearly in use earlier as a copper alloy lever lock key with a fleur-de-lys handle was found in a context that pre-dated AD 90 at Richborough (Henderson 1949, 125 no.86, pl. XXXIV).

Chain

Chain. Iron. Parts of 3 links remaining articulated. Each link made of rectangular-sectioned oval compressed into figure-of-eight pattern. Length of complete link 40mm; section of ring 5mm x 3mm.

Only one fragment of chain can be identified with certainty in this assemblage, though it is possible that some of the ring fragments (see nos.65-8) could have come from chains. The extant fragment no.36 came from a chain with figure-of-eight loops (Manning 1985, 139).

Box

12 complete or fragmentary iron nails from chest, all but one retaining mineralised wood grain in corrosion products.

Sf number	Description	Woodgrain	Length
1010	Complete	Horizontal	45mm
1011	Lower part of shank missing; tip of second nail corroded to shank near break	Horizontal	41mm
1012	Lower part of shank missing	Horizontal	29mm
1013	Shank fragment	Horizontal near break; vertical towards point	31mm
1014	Complete, small head?	Horizontal for c.6mm below head; rest vertical	54mm
1015	Lower shank missing	Horizontal	29mm
1016	Complete	Horizontal	36mm
1017	Lower shank missing	-	32mm
1018	Tip missing	3^o or 4^o from horizontal	48mm
1019	Lower shank missing	Horizontal	38mm
1020	Tip missing	Horizontal	53mm
1021	Lower shank missing	Horizontal	22mm

Table 16: Nails from Box

These nails were associated with the box in grave 124 and retain indications of the grain of the wood in the corrosion products, though it has not been possible to identify the species of wood used. On most of the nails only horizontal wood grain impressions are present but on two there are a combination of horizontal and vertical from which the thickness of the wood could be ascertained. On the complete nail Sf1014 the change between horizontal and vertical comes only 6mm below the head suggesting a thin sheet of wood. On Sf1013 there are hints that the timber may have been thicker. The upper shank and head are missing but the junction appears to be 25 to 30mm from the tip of the nail. Five of the nails are either complete or lack only their tip (Sfs1010, 1014, 1016, 1018, 1020) and the mean length of these is 47mm. If it is assumed that Sf1013 was originally 45 to 50mm long, then the wood is likely to have been 10 to 20mm thick.

Wooden chests used as the containers for cremation burials of the 1st and 2nd centuries can be very elaborate as can be seen in a group from the south-east of England (Borrill 1981). They were decorated with copper alloy fittings in the shape of elaborate plates and studs and often had other iron fittings such as angle braces, hinges and lock mechanisms. No. 37 by contrast simply appears to be a box nailed together.

5.5.6 Recreation

Counter. Re-used pottery body fragment; grey ware. Fragment chipped to circular shape. Diameter 35mm, thickness 9mm.

Counter. Re-used pottery body fragment; grey ware. Fragment chipped to oval shape. Diameter 40mm x 33mm, thickness 9mm.

Counter. Re-used pottery body fragment; oxidised ware. Fragment chipped to circular shape. Diameter 44mm, thickness 6mm.

Counter. Re-used pottery fragment, heavily reduced fabric. Edges ground to circular shape and chipped in 2 places. Diameter 46mm, thickness 7mm.

Counter. Re-used pottery body fragment; grey ware. Fragment chipped to circular shape. Diameter 49mm, thickness 8mm.

Counter? Jurassic minutely bioclastic limestone. Asymmetrical oval disc with chipped and ground edges. Length 70mm, width 64mm, thickness 6mm.

During these excavations no custom-made gaming counters of bone or glass were found though such items were found during the 1930s excavations (Corder and Romans 1938, 44 no.8, 47 nos.10-11, 18-19, figs.8-9). Instead there are six counters made of pottery (nos.38-43) and a possible example from stone (no.44). Whether such re-used pottery roundels were indeed counters has been the subject of some debate (Crummy 1983, 93-4), and it may be noted that the examples catalogued as counters here are all of a diameter greater than that normally seen in custom-made bone and glass counters.

5.5.7 Craft and industry

Blacksmithing

Punch. Iron. Square-sectioned bar tapering to point with battered head. Length 71mm, head section 5mm.

Punch? Iron. Square-sectioned bar tapering to point. Head broken. Length 96mm, maximum section 12mm.

The battered head of no.45 strongly suggests that it was a blacksmith's punch used to make holes in hot metal (Manning 1985, 9). Short examples such as this would have been held by a wire handle. No.46 could also be a punch but as the head is broken it is not possible to be certain.

Carpentry

Bit head. Iron. Sub-rectangular-sectioned head tapering to asymmetric point; square-sectioned broken stem. Length 51mm.

Iron bits from carpenters' drills are not infrequently found broken at the stem as on no.47. As Manning (1985, 27) has pointed out, the soft iron stem would have been a point of weakness if the bit jammed in use and force was exerted to free it.

Antler working

Antler tine fragment. Widest end showing cuts from two sides at 90^o to each other and cancellous core partially hollowed out; narrower end more roughly snapped off with cut lines below. Rough longitudinal faceting. Length 63mm, maximum diameter 20mm.

Antler tine fragment. One end smoothly cut, other end snapped. Length 12mm, section 10mm.

These two fragments of antler tine indicate that antler working was carried out in the vicinity. No.48 might have been an unfinished one-piece handle (see no.29)

Hones

Hone. Sandstone, greyish white, fine- to mainly medium-grained, moderately sorted and fairly well compacted, calcareous with sparse muscovite and glauconite. Probably ‘Kentish Rag', i.e. Hythe Beds in Lower Greensand (Lower Cretaceous) of Kent and Sussex. Rectangular with long sides concave from wear. One long edge has remains of groove from initial manufacture. Length 93mm, maximum width 20mm, thickness 14mm.

Hone. Sandstone, pale greyish brown, fine-grained with subangular to subrounded grains, fairly well sorted, well compacted, slightly calcareous, with sparse minute muscovite. Conceivably local Middle Jurassic Kellaways Sand, but more probably from Middle or Upper Jurassic of north-eastern Yorkshire. Short rectangular bar with rounded ends. Length 56mm, width 33mm, thickness 17mm.

Hones have a place in both the workshop and the kitchen so whether these were originally domestic or craft items is impossible to tell. Roman honestones are generally rectangular as here but are not closely datable within the Roman period. The Kentish Rag from which no.50 was made seems to have been used as early as the 1st century for a honestone industry as an example from London has been found in a context with a terminus ante quem of AD 100 (Rhodes 1980, 132 no.685, fig.77), and thereafter this area seems to have exported hones to many parts of Roman Britain, with the trade well established by the 3rd century (MacGregor 1976, 5).

Polishing tool?

Burnisher. Black, extremely fine-grained (no granularity visible), soft (c.2-3 on Moh's Scale), fissile and splitting slightly along narrow edges, low specific gravity, with black steak. Either carbonaceous/bituminous mudstone (if so, probably from the ‘bituminous paper shales' in the upper part of the local Lower Jurassic sequence), or impure caneloid coal (if so, probably from the Thorncroft Sands in the local Middle Jurassic sequence although an origin in the Coal Measures cannot be precluded).
The geology of this precludes the item being a hone, despite its shape. It is possible that is may have been used as a burnishing tool. These were used in a variety of non-ferrous metal-working trades such as jewellery (Johns 1997, 119 no. 351), and also had a role in other crafts where surfaces needed to be brought to a shine. Rectangular bar with rounded ends and angles. Length 44mm, width 9mm, thickness 6mm.

5.5.8 Structural finds

Windows

These excavations produced only one fragment of cast window glass but such glass was recorded during the Wacher excavations (Charlesworth 1969). This type of glass was used in buildings from the 1st to 3rd centuries.

Wall hooks

Wall hook or pivot. Iron. Square-sectioned L-shaped bar with longer arm tapering. Length 65mm, width 49mm, maximum section 9mm.

These L-shaped bars could either be wall hooks (Manning 1985, 129) or the L-shaped staple for use with a drop hinge (ibid., 12, fig.31 no.1). In the case of no.55, its size suggests the latter identification is most likely.

Double-spiked loop

Staple. Iron. U-shaped staple with curved pointed ends. One tip missing. Length 39mm, width 33mm, maximum section 13mm x 3mm.

Split pins such as no.56 were designed to be driven into wood or masonry to provide anchor points for other fittings such as rings (Manning 1985, 130, see for example Pl. 61 nos.R34-8), and it is possible that the staples nos.57-8 served the same purpose. An alternative interpretation for the latter might be as some form of joiner's dog.

Joiner's dogs

Joiner's dog. Iron. Square-sectioned bar tapering to D-sectioned pointed arm. Part of bar and one arm missing. Length of arm 72mm, maximum section 9mm.

Joiner's dog. Iron. Rectangular-section bar tapering to arms; one arm broken, other straight-edged. Length 72mm, length of complete arm 25mm.

Joiner's dogs such as these were used to join timbers or sometimes masonry (Manning 1985, 131).

Holdfasts and nails

As ever on a Roman site, the commonest items of ironwork are the nails. In addition to nos.61-2 catalogued above, a further 185 complete or fragmentary nails were recovered from Roman contexts. Roman nails can be divided according to length and the types of heads depend on how long the nail was. The longest (Manning 1985, 134 Type 1A) have conical heads to withstand the force used to drive them into timber, and the shorter ones (ibid., Type 1B) have flat heads. In this assemblage there is one example (no.61) of a type 1A nail with conical head and all of the others appear to be shorter nails with flat heads (see no.62). There are 68 nails that are either complete or lack only their tips. These range in length from an example lacking its tip and measuring only 25mm to a complete example 124mm long. The average length is 58mm and 50% of the complete or near complete nails fall between 45 and 66mm in length. They are therefore the all-purpose Roman nails suitable both for some structural work such as timber cladding and for other joinery purposes such as making the box found in grave 124 (see no.37 above). Holdfasts were used to guarantee a strong join between two pieces of timber, such as was required in boat-building (Manning 1985, 132). Only one example (no.63) could be identified with certainty in this assemblage.

All these finds seemingly relate to wooden structures, though some objects such as the joiner's dogs may potentially have been used on masonry. The site has also revealed some ceramic building material in the form of fragments of brick and tile, whilst some of the stone found on site may also have had a structural function.

5.5.9 Miscellaneous

The assemblage includes the normal range of items for which no certain function can be suggested either because such items can have a large number of functions such as the rings (nos.65-8) or because they are now too fragmentary. The more diagnostic items are catalogued below (nos.64-75) and the less diagnostic ones from Roman contexts are listed in the archive. Of special interest is no.72 where the ghost of a decorated wooden rod is preserved in iron corrosion products.

Collar. Iron. In two joining fragments with approximately one-third of circumference missing. Diameter 40mm, width 20mm, thickness 5mm.

Ring. Iron. Approximately one-third of circumference extant. Corroded onto separate and unconnected fragment of iron. Diameter c.40mm.

Spike or bar. Iron. Square-sectioned bar tapering to point, bent into ‘L'-shape.

Riveted strip. Iron. Rectangular section; one end broken. One perforation in complete end; one perforation 50mm from end retaining small flat-headed nail. Length 77mm, section 14mm x 1.5mm.

Rod. Originally wood, now preserved as a ghost hollow inside iron corrosion products. Circular section. Rod originally decorated by N-shaped incised line. Length 46mm, diameter 13mm.

Tessera. Fired clay. Re-used fragment from grey ware pot with edges rubbed smooth to form a regular rectangle. Length 18mm, width 14mm, thickness 7mm.

Disc. Jurassic oolitic limestone. Chipped to circular shape with uneven upper and lower faces. Diameter 146mm, thickness c.30mm.

Worked oyster shell. Fragment of oyster shell with circular perforation close to one edge. 44mm x 38mm.

5.5.10 Coins

Identified by Craig Barclay, Yorkshire Museum

Copper. Irregular denarius of Elagabalus, c.AD 218+. RIC 46.
Obv: draped and diademed bust right. Legend IMP […]INVS PIVS AVG
Rev: Emperor standing left, sacrificing over lighted altar, holding patera and branch. In left field, star. 18mm; 2.08g.

Silver. Penny of Edward I/II, class 10 ff. (c.1301-27).
Obv: Facing bust with bifoliate crown. Legend […]GL DnS[…]
Rev: Legend […]/[ ]AS/DV[ ]/[…]. 17mm (clipped); 0.65g.

RD	-	rim diameter
RT	-	rim thickness
BD	-	base diameter
WT	-	wall thickness
PH	-	present height
EVE	-	estimated vessel equivalent

5.5 Small Finds Discussion

5.5.1 Personal ornaments

Bow Brooch