A46 NOTTINGHAMSHIRE The Archaeology of the Newark to Widmerpool Improvement Scheme, 2009 by Nicholas Cooke and Andrew Mudd Co tswo ld Wessex archaeology 226 A46 Nottinghamshire: The Archaeology of the Newark to Widmerpool Improvement Scheme 2009 Copper Alloy and Iron Finds Personal adornment or dress by Jörn Schuster Brooches Of the 20 Roman or Romano-British brooches and brooch fragments found, only 10 came from Romano-British contexts while the remainder were found residually in the topsoil or unstratiied contexts. Two of the earliest brooches were found in Late Iron Age/early Romano-British boundary ditches in DE3006. One is a Colchester brooch (Fig. 4.93.1) with probably oval-sectioned bow, but since its catchplate is missing it can only be assigned broadly to Mackreth types C 2e or 3e which had their floruit in the second and third quarter of the 1st century AD (cf. Mackreth 2011, 37–9, pl. 22, 211, curve of bow more like pl. 23, 9717). The other is a Langton Down brooch (Fig. 4.93.2). A further example of the type was recovered from the topsoil in DE3001. Both have the rounded head, and the complete example (of Mackreth type LD 2.b2) also has the tapering bow with straight sides, of Riha’s type 4.4.1 which starts in the Augustan-Tiberian period and with a floruit ending in the Claudio-Neronian period (Mackreth 2011, 34; Feugère 1985, 266; Schuster 2011, 198). A fragmented Hod Hill brooch with triangular foot and rudimentary knob was found in a mid-RomanoBritish layer in well 303819 in DE3002 (Fig. 4.20). It belongs to Mackreth type Hod Hill 4.d1, Riha type 5.6 or 5.16, or Bayley and Butcher Hod Hill a, indicating a date in the later 1st century AD (for a summary of the dating evidence see Schuster 2011, 208). Three brooches belong to the broad group of Colchester derivatives with hinged pins, and all come from topsoil contexts. In the example from DE3001 the original hinged pin had been replaced by a pin with remains of a spring which had been inserted into the widened slot for the original pin hinge (Fig. 4.93.3). The type dates to the second half of the 1st century AD (Mackreth 2011, 84–5). The head and beginning of the bow of a second example has completely plain wings indicative of Mackreth type CD H 5, probably of similar date. A hinged T-shaped brooch from the topsoil in DE3002 (Fig. 4.93.4; cf. Mackreth 2011, pl. 59, 2424) belongs to Mackreth’s Nene group, essentially a later 1st-century AD type (ibid., 89). An extremely corroded trumpet-headed brooch, with remains of a head loop corroded onto the rim and the head with enamelled ields, was retrieved from layer 303678 in DE3002 (Fig. 4.23). It belongs to Mackreth type TR 1.2b1/7, essentially of the 2nd century AD; a close comparison for the Margidunum brooch comes from Corbridge (ibid., 119–20, pl. 80, 4976). Two headstud brooches include the bow only of a brooch found in the topsoil in DE3002 (Fig. 4.93.5). A rectangular socket, probably for the attachment of a dog igurine, suggests that it belongs to Mackreth type HDST 2b (cf. Mackreth 2011, pl. 71, 12442). The second brooch, from Romano-British ditch 218553 on the eastern side of DE3001 (Fig. 4.16), features the decorative scheme of brooches like Mackreth type HDST 11a (cf. ibid., pl. 74, 7273), but instead of the Most copper alloy and iron inds have been x-radiographed, and conservation was carried out on selected objects. Of the 458 recorded objects, 129 were recovered from inhumation or cremation graves and these are considered in a separate section below. The distribution of the remaining 329 objects by site subdivisions and metal type is shown in Table 4.21. Of these objects, 188 (57%) were found in modern or unphased contexts, mainly topsoil and subsoil. The objects have been grouped by functional category following Crummy (1983, 5–6). Apart from a cruciform brooch (see below), a nail shank from an Anglo-Saxon sunken-featured building in DE3004 and a nail shank from a RomanoBritish/Anglo-Saxon ditch in DE3003, the remaining objects (138) were retrieved from features dated Late Iron Age/early Romano-British to late Romano-British. Table 4.21 Margidunum Hinterland: metal types per site subdivision Cu alloy Iron DE 3001 Event code 28 29 Total 57 DE 3002 99 124 223 DE 3003 2 6 8 DE 3004 5 20 25 DE 3006 4 2 6 Unstrat. 2 8 10 140 189 329 Total Table 4.22 shows a breakdown by functional category of the inds from Late Iron Age/early Romano-British to late Romano-British features. While only 12 items from Romano-British layers are intrinsically datable as Romano-British, another 14 Roman objects (mainly brooches but including a key, a vulva-shaped itting, a stylus and a T-clamp) and one late medieval strap end (cf. Egan and Pritchard 2002, 140–1 ig. 92, 650–1) are not included in Table 4.22 as they have been found residually in modern layers. However, these inds are included in the typological discussion below. Table 4.22 Margidunum Hinterland: copper alloy and iron objects from Romano-British phases (excluding graves), by functional category Functional category Personal Total 21 Toilet 3 Household 1 Fitting Agriculture Metalworking Uncertain Total 48 1 2 64 138 Margidunum Hinterland: Late Iron Age and Romano-British Settlement & Burial, & Early Anglo-Saxon Activity 227 1 3 4 2 6 5 7 enamel 0 50 mm Fig. 4.93 Metal finds (1–7). See catalogue for description spring construction of that type its pin is hinged into a slot in the head (Fig. 4.93.6). Both can broadly be dated to the late 1st and 2nd centuries. Three diferent types of plate brooch were discovered, including a tutulus brooch from pit/waterhole 301171 (Fig. 4.93.7) and a probable umbonate brooch from the topsoil (Fig. 4.94.8), both in DE3001 (Fig. 4.18). The tutulus brooch belongs to Feugère’s type 25a, on account of its cone ending in a small button (Feugère 1985, 351), which can be dated to the later 1st/early 2nd century. The identiication of the umbonate brooch is owed to the enamelled ields in the example from Margidunum, which are shaped as proper triangles rather than the customary petalled shapes usually associated with such brooches (eg, Mackreth 2011, pl. 109, 12640; Bayley and Butcher 2004, ig. 99, 379). The third plate brooch (Fig. 4.94.9), from the topsoil in DE3002, appears to have lost its decorated repoussé sheet which had been soldered onto the front. Although the animal-headed lug above the catchplate is rather worn, a good comparison comes from the north suburb of Durobrivae, Cambridgeshire (Mackreth 2011, suppl. pl. 1, 10602). The scant dating evidence for the type points towards the 2nd century and onwards (ibid., 155). The three penannular brooches found all have terminals coiled back at right angles to the plane of the ring. Two can be classed as Mackreth type PEN c2.a on account of their circular ring and proper coils (ibid., 207). One of the brooches was found in the topsoil in 228 A46 Nottinghamshire: The Archaeology of the Newark to Widmerpool Improvement Scheme 2009 8 9 12 13 14 enamel tinning 10 11 0 50 mm Fig. 4.94 Metal finds (8–14). See catalogue for description DE3004 (Fig. 4.94.10), the other is unstratiied. The type is common from the Late Iron Age into the 2nd/3rd century AD. The third brooch (Fig. 4.94.11), found in an Late Iron Age/early Romano-British layer 302287 in DE3001 (not shown in plan), is distinguished from the other two by the terminal with a recurved coil. The few dated examples belong to the mid-1st century AD with an outlier in the late 3rd/4th century (ibid., 206–7). A triangular catchplate fragment from the topsoil in DE3002 and the corroded head of a hinged brooch from a late Romano-British layer (303677) are likely to derive from early Roman brooches. Two brooch pin fragments from mid- and late Romano-British postholes in DE3002 provide no close dating. Hobnails Approximately 22 hobnails were recovered from settlement contexts, including nine from pit 301178 in DE3001 (Fig. 4.18) and at least six with mineral-preserved organic and leather remains from well 303819, associated with Structure 16 in DE3002 (Fig. 4.20). Possible personal objects A rectangular-sectioned rod fragment with lat lanceolate end was collected from the overburden in DE3002 (Fig. 4.94.12). The object appears to be uninished and – if Roman – may have been intended to be either a snakehead bracelet or perhaps a spoon probe. Two small wire fragments may have been fragments of chains or necklaces. A small U-curved fragment was found in early Romano-British Enclosure H ditch 218522 (Fig. 4.11) and a very ine wire – coiled into an eye at both ends, probably from jewellery, perhaps a necklace or earring – came from ditch recut 218524 of late Romano-British Enclosure K (Fig. 4.33). Toilet equipment Three items of toiletry were all found in DE3002: a ligula with lat spoon (Fig. 4.94.13) from mid-Romano-British layer 303677 can be compared to a similar example from Corbridge (Allason-Jones 1988, 169 ig. 80, 84, 86); a rectangular-sectioned strip with a fragmented loop at one end may be a toilet spoon or nail cleaner (cf. Crummy 1983, 58 ig. 62, 1869 or 60 ig. 64, 1900); and a bent strip with parallel sides and fragmented end loop may be what remains of a small pair of tweezers. The latter two came from mid-Romano-British layers. Household utensils A bucket handle from DE3002 (layer 303509, a mixed layer sealing Roman road surface 303505) is very similar to one found during earlier excavations at Margidunum (Oswald 1927b, pl. 11). A rectangular hinge plate with two square holes (Fig. 4.95.16) from mid-RomanoBritish layer 303678 may have been part of a box. Margidunum Hinterland: Late Iron Age and Romano-British Settlement & Burial, & Early Anglo-Saxon Activity 229 15 16 17 18 0 50 mm Fig. 4.95 Metal finds (15–18). See catalogue for description Writing and written communication A copper alloy stylus from the overburden at DE3002 is the only object in this category (Fig. 4.94.14). It can be compared to a stylus from Springhead, Kent (Schuster 2011, 251 SF 15898), which may be a copper alloy version of iron styli like Manning’s type 1. During earlier excavations at Margidunum a square seal-box with an enamelled swash-N was found in ditch 9 (Oswald 1927b, ig. 2; 1952, pl. 5, 1B). Fasteners and fittings Of the 103 items in this category, 68 are either nails or nail shanks, 38 of which were retrieved from RomanoBritish contexts, including one small copper alloy nail with globular head (cf. Crummy 1983, 115 ig. 116, 2991) from layer 302016 (the upper layer of trackway 218547 in DE3001; Fig. 4.30). Two T-staples with relatively wide stems were recovered from mid-RomanoBritish layers 303678 (Fig. 4.95.15) and 303677 at DE3002 (cf. Crummy 1983, 120 ig. 128, 4073; Hawkes and Hull 1947, pl. 105, 20; Manning 1976, 58 ig. 25, 163). From the topsoil in the same area came a T-clamp with anchor-shaped head, comparable to similar objects from Aldborough (Bishop 1996, 86–7 ig. 46, 539–40), as well as a lever lock key with a proiled bow ending in a loop surmounted by a knob (Fig. 4.95.17; cf. Birley 1997, 19 ig. 6). Objects associated with agriculture and horticulture Late Romano-British posthole 303589 (Fig. 4.46) contained fragments of a U-proiled strip which might have 230 A46 Nottinghamshire: The Archaeology of the Newark to Widmerpool Improvement Scheme 2009 been part of a spade sheath, and the fragments of what might once have been a reaping hook were recovered from a later Romano-British ill of well 303715 (Structure 21) (Fig. 4.20). Military equipment The only object which may belong in this category is a vulva-shaped itting from the overburden in DE3002 (Fig. 4.95.18), probably for attachment on a belt or horse harness strap where it would have served as a good-luck charm. While mainly found on late 2nd- to mid-3rd-century military sites (cf. Oldenstein 1976, 137–9, Taf. 34, 269; Allason-Jones and Miket 1984, 239 no. 870–1), these objects occasionally turn up in civilian contexts, for instance Brougham, grave 301 (Cool 2004, 242 ig. 4.245, 1). Objects and waste material associated with metalworking Evidence for both copper alloy and ironworking was mainly found in DE3002; a piece of copper alloy casting spill from the topsoil in DE3001 has to remain undated, whereas a trapezoidal-shaped block of iron (52 x 30 x 8 mm) from late Romano-British ditch 304521 (layer 304336, intervention 304335) in DE3004 is clearly a piece of bar iron for smithing. Nine of the 10 metal- working items found in DE3002 were recovered from the topsoil, including a copper alloy casting sprue and a sheet fragment with a repair attached by tubular rivets. An iron bar fragment weighing 329 g was retrieved from mid-Romano-British layer 303677 (Fig. 4.23), and a further four bars or billets, weighing between 12 g and 109 g, were found in the topsoil. Discussion The small number of inds noted in the various categories does not permit much further analysis beyond indicating the possibility that certain activities had been carried out, or certain types of equipment may have been used, at Margidunum. It is notable that no knives or any other tools have been recorded, even though there is clear evidence, in the form of both metalworking debris and slag (see Starley, below), that ironworking/smithing and – probably on a much smaller scale – casting of copper alloy objects took place at Margidunum Hinterland, especially at DE3002. In contrast to this, it is possible to compare the brooches from the recent excavation, together with those found during earlier investigations at Margidunum, with the average brooch assemblage for the Midlands (Fig. 4.96). The groups shown in Figure 4.96 follow the deinition by Plouviez, who did not include the various Margidunum – 39 brooches Brooch group Midlands – 546 brooches Earlier exc. (n = 25) B B C C D D E E F F G G H H I I J J Brooch group (after Plouviez 2008) Langton Down etc Birdlip etc Colchester Nauheim derivative Colchester derivative (Harlow) Colchester derivative (Dolphin) Aesica etc Hod Hill etc Early plate Polden Hill T-shaped (early types) Colchester derivative (hinged) Trumpet Headstud Dragonesque etc Developed T-shaped Trumpet derived, Alcester etc Trumpet/headstud related Enamelled hinged bow Equal-ended hinged Disc (Continental types) Umbonate disc Flat disc (British types) Applied repoussé Representational Knee Gilded plate P-profile Crossbow Brooch group (after Plouviez 2008) MGM09 (n = 14) K L M N O P Q R S T U K L M N O P Q R S T U V V W W X X Y Y Z Z ZA ZA ZB ZB ZC ZC ZD ZD 0 5 10 Percentage 15 20 0 5 10 15 20 Percentage Fig. 4.96 Comparison of brooches from Margidunum Hinterland with those from earlier Margidunum excavations by type (after Plouviez 2008) Margidunum Hinterland: Late Iron Age and Romano-British Settlement & Burial, & Early Anglo-Saxon Activity 231 penannular types in her comparison because of the diiculty of separating earlier from later types (Plouviez 2008, 171–3, table 4.1.1 and ig. 4.1.1). While this task has now been rendered somewhat easier with the publication of Mackreth’s book on Late Iron Age and Roman brooches (Mackreth 2011), it would have been beyond the scope of this study. Of the 20 brooches from the Margidunum Hinterland excavation, 14 could be grouped following Plouviez’s scheme, and this could be augmented by a further 25 brooches found during earlier excavations. What immediately becomes apparent is the under-representation of early types like Colchesters, Colchester derivatives of Harlow construction, and Nauheim derivatives, whereas Langton Down and Dolphin brooches are found in quantities comparable to the regional and national averages. The quantity of the ubiquitous, mainly post-conquest, Hod Hill types (group I) is slightly lower at Margidunum: with just over 10% it is comparable to the quantity in the neighbouring region of East Anglia, which also provides a better comparison for the number of hinged Colchester derivatives (group M). The amount of early T-shaped brooches, on the other hand, is closer to the trend in the south-western graph, and that of the Headstud brooches (group O) is more in keeping with the north (Plouviez 2008, 173, ig. 4.1.1). This pattern may be the product of the location of Margidunum on the Fosse Way and close to Ermine Street. The small numbers of brooches in groups P and below show the general decline in the use of brooches from the later 2nd century onwards and are thus in keeping with the national trend, but the small numbers (most of the bars in the Margidunum diagram merely represent a single brooch) caution against an overinterpretation of these trends. Catalogue of illustrated metal finds All copper alloy, unless otherwise stated. (Fig. 4.93) 1 2 3 4 Colchester brooch; six spring coils, those on right more pronounced and extending beyond undecorated wings; hook with long triangular end; bow oval-sectioned; catchplate missing. Plouviez D. Mackreth C 2 or 3. ON 1008, context 306080, cut 306076, ditch 218766, DE3006. Langton Down brooch with broad rounded head. Plain spring cover with remains of spring; damaged edges of bow tapering in gentle curve, bow with central groove accompanied by two longitudinal grooves either side, central and outer grooves with dark grey covering from ?tinning. Base of open catchplate remaining. Plouviez B. Mackreth LD 2.b2. Riha Typ 4.4.1. Cf. Springhead (Schuster 2011, ig. 84, 23; for tinning see ig. 84, 27). ON 1033, context 306636, cut 306637, ditch 218794, DE3006. Colchester derivative brooch. Plain hinge tube with end reels; original hinged pin replaced by extending pin slot and insertion of spring fragment of which pin is missing. Proiled bow with central groove. Foot broken of just below beginning of catchplate wing. Plouviez M. Mackreth CD H 3.cd. ON 525, topsoil 301000, DE3001. Colchester derivative brooch with hinged pin and iron 5 6 7 axis. Hinge tube with groove near ends. Bow with short beaded crest, foot with half-proiled knob. Pin missing. Plouviez L. Mackreth CD H 6ad [Nene group]. ON 1382, topsoil 303500, DE3002. Headstud brooch. Part of bow and foot of brooch with saw-tooth edges and stud at foot tip; only base of catchplate preserved. Remains of rectangular socket (?for dog igurine) at top of bow. Very corroded. Plouviez O. Mackreth HDST 2.b. ON 1321, topsoil 303500, DE3002. Headstud brooch with hinged pin, remains of head loop acting as axis visible at side of hinge cover. Wings decorated with ive reels and grooves, outer and innermost reels beaded, central with zigzag pattern. Hook continues into enamelled head stud; both appear to be cast on. High-rectangular sectioned bow prob. with central groove originally illed with white metal. Stud at foot end with three beaded reels, end with two concentric enamelled zones. Plouviez O. Mackreth HDST 11.a but lacks spring (cf. Mackreth 2011, pl. 74, 7273). ON 765, Context 301986, cut 301985, ditch 218553 (Fig. 4.16), DE3001. Plate brooch with central hollow cone ending in tiny button, surrounded by lat zone margin which may contain residues of now corroded enamel; chamfered rim with small notches around edge preserved above double pin lug; trapezoidal catchplate, pin rest missing. Plouviez V. Mackreth PL CONT 16, Feugère 25a, Riha 7.11.1. ON 682, context 301176, waterhole 301171, DE3001. (Fig. 4.94) 8 ?Umbonate brooch. Fragment of central cone. No remains of enamel. Central, slightly raised cone with eight triangular ields pointing to small central knob; separated by pronounced groove from wide lange with in- and outward facing triangles. Outer edge missing apart from one small projection. Cf. Bayley and Butcher 2004, ig. 99, 379 but this has petalled boss. Plouviez W. Mackreth Plate 7.a (Petalled boss) (eg, pl. 109, 12640). ON 505, topsoil 301000, DE3001. 9 Circular plate brooch. Disc with remains of small head loop and possibly an animal-headed or (more likely) a stepped lug extending above remains of catchplate on opposite side. Left side of plain disc damaged. There may have been a circular groove around the perimeter. Remains of white-metal coating on front and back. Double pin lug. Pin missing, only base of catchplate remains. Plouviez Y. Mackreth Plate 1x (cf. Mackreth 2011, suppl. pl. 1, 10602). ON 1266, topsoil 303500, DE3002. 10 Penannular brooch, terminals coiled back at right angles to plain, subcircular sectioned ring. Circular sectioned pin, bent down, lattened end coiled around ring. Mackreth PEN c2.a. Bayley and Butcher 2004 P3, Fowler 1960 type C. ON 1509, subsoil 304001, DE3004. 11 Penannular brooch, terminals coiled back at right angles to plain, circular sectioned ring (D 2.0–2.2mm) and ending in recurved coil. Pin, bent down, lattened end coiled around ring. Mackreth PEN c1.c (cf. Mackreth 23011, pl. 143, 13037). ON 767, layer 302287, DE3001. 12 Bracelet? Rectangular-sectioned rod fragment with lat, lanceolate end. ON 1448, topsoil 303500, DE3002. 13 Ligula. Small, lat prob. circular spoon with circularsectioned shaft; opposite end missing. ON 2102, spread 303677, DE3002. 14 Stylus. Scriber separated from plain body by slight step, parallel-sided eraser (W. 7mm) grows out of body. Slight bend at either end. ON 1456, topsoil 303500, DE3002. 232 A46 Nottinghamshire: The Archaeology of the Newark to Widmerpool Improvement Scheme 2009 (Fig. 4.95) 15 Iron T-staple. Flat stem with curved arms. ON 2068, spread 303678, DE3002. 16 Iron rectangular hinge plate with two square holes. Spread 303678, DE3002. 17 Iron lever lock key with square-sectioned handle with transverse mouldings before shank and loop of bow. Bow surmounted by knob. Circular-sectioned shank, only base of bit remaining. ON 1434, topsoil 303500, DE3002. 18 Vulva-shaped itting with hexagonal base. ON 1452, topsoil 303500, DE3002. Metal finds from graves Metal objects were recovered from 12 graves in DE3001, predominantly within Enclosure K (Fig. 4.35), and one late Romano-British cremation grave, and a further two graves in DE3002 (see Grave Catalogue above). Apart from a copper alloy inger-ring from grave 301132 and an unidentiiable small copper alloy lump from grave 303854, all other metal objects are made of iron. Where identiiable the iron objects are either nails or hobnails. Finger-ring The inger-ring (ON 66; Fig. 4.97.1) was found at the disturbed south-western end of grave 301132 in the area between the (missing) feet and the edge of the grave cut. The ring is very corroded and its original bezel missing, but its hoop had been bent and broken, probably by or at the time of deposition, prior to some more recent breaks. The ring is a variation of Guiraud’s type 3a, on account of its slight shoulders and the setting for the bezel, but it is unusual in that the axis of the bezel is orthogonal to the axis of the hoop, a detail found in the more proiled type 3g. A Guiraud type 2a ring with this bezel orientation comes from Springhead (Schuster 2011, 237 ig. 103, 158). Shouldered inger-rings of Guiraud’s type 3 mainly belong to the later 2nd to later 3rd centuries AD (Guiraud 1989, 185). 1 0 2 50 mm Fig. 4.97 Metal objects from graves (1–2) Hobnails Table 4.23 shows the minimum number of nails, nail shanks and hobnails identiiable from each grave. The numbers can only be an approximation as the preservation of most nails and hobnails was poor or very poor; only the four hobnails from cremation grave Table 4.23 Margidunum Hinterland: minimum number of nails/shanks and hobnails per RomanoBritish grave Grave Nails/shanks Lengths of nails (mm) Hobnails Inhumation 301131 26 301132 8 301133 26 301134 11 (n=4) >32–>94 - 301135 2 (n=2) >38–64 - 301245 4 (n=1) 28 41 (n=1) 62 64 301510 6 301572 11 301850 303921 (n=22) >52–115 (n=1) >21 (n=13) 21–>84 52 1 136 (n=4) 51–56 - 6 - - 1 (n=1) 25 - 2 - 4 Cremation 302289 302289 (Fig. 4.30) were in a better condition (Fig. 4.97.2), probably because they had been exposed to the heat of the funerary pyre (cf. Cool 2004, 41; 2006, 4). Six graves contained hobnails. One was cremation grave 302289 in the north of DE3001. The ive inhumation graves with hobnails were all located along the inside of the ditch of Enclosure K. On the basis of the number of hobnails it is probably safe to assume that nailed footwear was present in four graves (Table 4.23). The situation is more ambiguous in grave 301132 where the single recorded nail may have ended up in its position near the right femur during the disturbance of the south-western end of the grave, which removed the feet and – had it been present – any footwear. The clearest evidence for shoes worn on the feet comes from grave 301133 (Fig. 4.36) where the hobnails were found in a compact area around the feet at the southern end of the grave. The hobnails in grave 301245 lay in a conined area at the eastern end of the grave, close to the edge of the grave cut (Fig. 4.37, ON 685). As in most graves at Margidunum Hinterland the bone preservation in grave 301245 was poor, and no bones were recorded below the lower tibia. The position of the hobnails is perhaps more indicative of the shoes having been placed next to the feet, unless the feet had been propped up against the edge of the grave. In grave 301131 the shoes had been placed outside the coin for skeleton 301180, and it is uncertain whether they belong to this burial or to skeleton 301183 whose bones had been moved to the eastern end of the grave cut (Fig. 4.36, ON 621). A clear case of unworn footwear is present in grave 301510: the 64 hobnails, recorded in the shape of a shoe, were found outside the right tibia on the north-western side of the grave (Fig. 4.37, ON 709). It was not possible to block-lift any of the concentrations of hobnails, which would have allowed x-radiography in order to analyse the nailing patterns of the soles. Equally, Margidunum Hinterland: Late Iron Age and Romano-British Settlement & Burial, & Early Anglo-Saxon Activity 233 it is not always clear whether a single shoe, a pair or more shoes had been deposited, as the numbers of nails per shoe varied according to the type of shoe, ranging from as few as 25 to 160 (cf. Cool 2004, 4–5 table 1). For Margidunum Hinterland this could mean that graves 301131 and 301245, with at least 52 and 41 hobnails respectively, could have contained pairs of very lightly nailed shoes or a single, more heavily nailed shoe. The outline of a shoe from grave 301510 with 64 hobnails appears to belong to a shoe with a similar quantity of nails per sole as the pair worn by the individual in grave 301133 where at least 136 hobnails were found. A tile bearing the mark of a hobnailed boot was also recovered (context 304023, late Romano-British ditch 304527, DE3004 – see Jones, Ceramic Building Material, above). Of the inhumation burials recorded during earlier excavations at Margidunum Southern Cemetery, outside the southern defences of the town, only one of the c. 16 inhumations had been buried wearing hobnailed footwear (grave 9; Todd 1969, 76), compared with ive out of 13 burials in Enclosure K. The strengthening of shoe soles with hobnails was a Roman introduction to Britain (Mould 2004, 392), but this need not imply that the custom of placing shoes in graves did not exist before when it was merely invisible due to use of unshod footwear. Furthermore, the scarcity of the rite in the Mediterranean ‘makes it more likely to be a Celtic concept acquiring a Roman material expression’ (Philpott 1991, 165, 171). The inclusion of hobnailed shoes in cremation burials is irst found in the south-east of England in the mid-1st century AD, but had only become well established as a rite in the 2nd century, especially in rural settlements and small towns (ibid., 165, 416–7). The practice of placing hobnailed shoes in inhumation graves had become widespread especially in south central England by the 4th century, and Philpott suggests that the occurrence of footwear in some contemporary cremation burials may by that time have been inluenced by the common practice in inhumation burials. While earlier cremation burials containing hobnails tended to be rather richly furnished, the decline in the provision of grave goods in cremation burials during the 3rd and 4th centuries also extends to graves with hobnails (ibid., 166, 458 ig. 28). The radiocarbon date of cal AD 140–330 (at 95% probability) (SUERC-39054) for cremation grave 302289 its the outlined picture and, considered together with the lack of any other grave goods, may point to a date at the later end of the range. The number of inhumation burials with hobnails from Roman Britain increased during the late 2nd and 3rd centuries, but most graves thus furnished date to the 4th century, although Philpott points out that the proportion of burials with hobnails is higher in cemeteries which started in the 2nd or 3rd centuries than in those not established until the 4th (ibid., 167–8; 226). The practice appears especially prevalent among the middle ranks of the rural population but is scarce with richer burials, in lead or stone coins or mausolea. This picture is supported by the evidence from Margidunum, where the two lead coins discovered during earlier road works were devoid of any grave goods (Todd 1969, 77–8). Nails The number of nails recovered from graves at Margidunum ranged from none to 26 nails per grave (Table 4.23). Where identiiable, the nails are predominantly of the lat-headed Manning (1985) type 1b, although it is possible that ON 603 from grave 301133 is of the T-headed type 3 and ON 707 from grave 301510 an L-headed type 4. Due to the poor condition of the nails only a selection were measured and, although the emerging image is not very clear, it can be said that there is a tendency for the use of larger, sturdier nails in graves that contain a greater number of nails. Only grave 301572 appears to be an exception, in that the four measurable nails consistently range between 51 mm and 56 mm. In the two graves with 26 nails these were found in groups across the head and foot area, not the ends, and in grave 301133 there is an additional row of three nails across the middle of the grave. While no trace of the coin had been observed in either case, the coins appear to have been constructed out of individual boards held together by rafters nailed across. A second type of construction is represented by – sometimes very short – rows of nails along both sides of the grave as in graves 301132, 301134, 301510 and 301572, or only one side in the case of grave 301245. The small number of nails from grave 301135 may merely indicate that the lid had been nailed shut, while the remainder of the coin had been constructed using wooden joints (O’Brien 1999, 13). It is equally possible, especially when considering the shallowness of the graves at Margidunum, that no coin had been present and the nails found their way into the graves by some later disturbance. Another possibility is that an individual nail may have been deposited in a grave for ritual or magic purposes, eg, as an apotropaic amulet, in order to protect the deceased (or the living?) from evil (Black 1986, 223; Dungworth 1998, 153). Such a use is diicult to prove, but could be considered if a single nail is deposited in a conspicuous position in the grave, as found in infant burials at West Thurrock, Essex, burial 17062, and Springhead, Kent, grave 12222 (Schuster 2009, ig. 10, 34; 2011, 278). The nail from feature 303921 at Margidunum Hinterland may be another example of this rite, but unfortunately the nail has clearly been disturbed. Anglo-Saxon cruciform brooch Catalogue description ON 1500 (Fig. 4.98) Copper alloy and iron. Cruciform brooch. Proiled knob (lattened underside) at top cast in one with head plate which has slightly downsloping wings set of from slightly thicker, mildly trapezoidal head plate. Lines along inner and lower edge of wing on near complete left wing, right wing almost completely broken of; single semi-circular lug to hold iron axis of iron spring, remains of which only preserved on left. Both fully proiled 234 A46 Nottinghamshire: The Archaeology of the Newark to Widmerpool Improvement Scheme 2009 0 50 mm Fig. 4.98 Anglo-Saxon cruciform brooch from sunkenfeatured building (Structure 27) side knobs preserved, left still in place; both knobs with cuts on inside where they would have been slid onto wings. Heavily worn bow (W 9 mm) with two transverse double lines at neck, upper double line set between heavily worn facets either side of it; the same repeated 3x at transition of D-shaped bow to foot. Bow has remains of heavily worn central line/groove accompanied by lateral lines. Foot with broad facets separated from zoomorphic terminal by two cast transverse ribs; groups of two lines above and below each eye, groups parallel to each other; pronounced nostrils, heavily worn; back of terminal hollow. Context 304060, sunken-featured building 304064 (Structure 27), DE3004. Cruciform brooches are found around the entire North Sea basin, from northern Scandinavia, via Jutland and northern Germany to England, chronologically spanning the late 4th/5th and 6th centuries, and possibly extending to around AD 600 in Scandinavia and England (Bode 1998, 67–8; Sherlock and Welch 1992, 37–8; Hirst and Clark 2009, 489). Summaries of earlier studies of the typological and chronological development of cruciform brooches have, for instance, been provided by MacGregor and Bolick (1993, 95–6) and Scheschkewitz (2006, 95–7). Important typological studies of the type include those by Reichstein (1975), Bode (1998) and Mortimer’s (1990) unpublished thesis which deals speciically with the English material (Mortimer’s study was unfortunately not available to me during the study of the Margidunum Hinterland cruciform brooch). While it was not possible to ind a close comparison for brooch ON 1500 which combined all elements in a similar way, the details of headplate, bow and foot can be compared to some of the forms deined by Bode (1998, Abb. 7–10). Thus the headplate with its slightly raised, trapezoidal centre is Kopfplattenform 10, the bow with lines along the centre and the sides is Bügelform 2. The foot, however, is more diicult to place within her scheme but may perhaps best be described as an amalgam of Fußformen 38 and 48, although there are no raised eyes or ears, which have instead been indicated by the rare combination of double chevron lines angled towards the nose above and below the eyes as well as above the nostrils. There is as yet no direct comparison for the arrangement of the lines; however, on a brooch of Reichstein’s Typ Trumpington from the eponymous site in Cambridgeshire (Reichstein 1975, Taf. 116, 9) the lines above and below the eyes appear as on the brooch from Margidunum Hinterland – as open chevrons pointing towards the tip – whereas the lines above the nostrils are part of a double saltire across the nostrils. Typ Trumpington is one of the types which Reichstein could not date precisely, and apart from the double lines their zoomorphic terminals are additionally distinguished from the Margidunum Hinterland brooch by pendants suspended from the tip of the nostrils (ibid., 46). Other brooches with double lines pointing towards the nostrils can be found among Reichstein’s Typ Midlum from Rudstone, Yorkshire (ibid., Taf. 87, 3–4). However, the overall combination of the elements of the Margidunum Hinterland brooch most closely resemble a brooch of Reichstein’s Typ Stratford from Girton, Cambridgeshire (Bode 1998, 179 Liste 1; Reichstein 1975, Taf. 92, 7). Reichstein’s Typen Stratford and Midlum belong to his late cruciform brooches, which he dates to the 5th century and especially its second half (ibid., 95). This date range is not contradicted by Bode (1998, 68), whose Kopfplattenform 10 and Bügelform 2 are part of her third element group with a date range from the second half of 5th century to the beginning of the 6th, while Fußform 48 belongs to an even younger stage of element group 2 of the middle third and later half of the 5th century. In their recent study of Anglo-Saxon inhumation burials based on the four cemeteries of Morning Thorpe, Spong Hill, Bergh Apton and Westgarth Gardens in East Anglia, Penn and Brugmann (2007) devised a classiication of cruciform brooches into three types. According to this scheme the Margidunum Hinterland brooch belongs to type X1 which ‘covers all cruciform brooches of “classic” design with animal-shaped foot-terminals and without lappets or any other additional design to the head- or foot-plate’ (ibid., 24). Their correspondence analysis shows that these brooches belong to female combination phase FA1, dated to an earlier part of the second half of the 5th century (ibid., 69–71 igs 5.20–21). Apart from the Margidunum Hinterland brooch, other brooches of type X1 found regionally include examples from Besthorpe (Nottingham Univ. Mus., G. Kinsley, pers. comm.) and Brough (Alvey 1980, 84) as well as an example of Reichstein’s Typ Foldvik-Empingham with trapezoidal footplate from an unknown ind spot at Margidunum, but presumed to come from a female grave (BHTA 2012). The cruciform brooch from DE3004 was recovered Margidunum Hinterland: Late Iron Age and Romano-British Settlement & Burial, & Early Anglo-Saxon Activity 235 from the ill of a sunken-featured building. Considering that the brooch was heavily worn, the brooch would have been used for at least 30 to 40 years (cf. Richthofen 1998, 256; 2000), suggesting a deposition no earlier than the late 5th or the irst half of the 6th century. Lead Objects by Grace Perpetua Jones A total of 166 pieces of lead was recorded from Margidunum Hinterland. Few were recorded from archaeological contexts, with only 15 from features of Romano-British date and two from Anglo-Saxon contexts. The remainder were recovered during metal detecting of the topsoil and subsoil. Each object was classiied according to type and a summary description entered on the project database (MSAccess). Summary of the assemblage Romano-British Romano-British spreads 303677 and 303678 (DE3002; Fig. 4.23) produced ive waste/of-cut fragments (ONs 2056, 2065, 2089 and one unregistered); a lump of building dross (unregistered) and a folded sheet fragment, stamped XX (ON 2046; Fig. 4.99). The stamped piece is now broadly of rectangular shape, and appears to have been created by folding a larger sheet fragment. It may have had some loops or straps at the back, now lattened. The characters are set within a recessed rectangle. Other pieces of lead sheet are known with an XX stamp or similar (Frere and Tomlin 1991, 2436.8 and 11) but their purpose is not known. They may have been decorative, or perhaps represent the trial stamping of a punch (Roger Tomlin, pers. comm.). The example from spread 303678 probably served a utilitarian function of some description. Late Romano-British ditches 218660 (Structure 25) and 304508 (north of Structure 22) produced three sheet/strip waste fragments (ONs 739, 744, 796, 1251), a further four were associated with trackways 218547 in DE3001 (Fig. 4.30) (ONs 739, 744, 796) and 303505 in DE3002 (Fig. 4.46) (ON 1251). A rivet (ON 2069), probably used to mend a ceramic vessel (ON 2069), was recovered from posthole 303629 (Structure 18) in DE3002. Anglo-Saxon A disc-shaped fragment, perforated close to the edge (ON 1502), was recovered from sunken-featured building 304064. A disc of plano-convex section (ON 1501) was present in Anglo-Saxon feature 304061; measuring 14 mm in diameter and 4 mm thick; it may have been part of a late medieval/post-medieval cloth seal (and therefore intrusive), but the surfaces appear plain, with the exception of a possible central depression in the convex side, in which case it is not chronologically distinctive. Unstratified The metal-detected lead from the topsoil and subsoil, and those pieces that were unstratiied, were dominated by waste fragments and ofcuts, including folded sheet and twisted strip fragments (122 pieces). These could not be closely dated but many are likely relate to activity in the post-medieval and modern periods (details in the project archive). Only two objects can be identiied as being earlier in date. A lead plug, 30 mm x 23 mm x 13 mm (ON 1358), would have been used to repair a ceramic vessel during the Romano-British period. Part of a cloth seal (ON 1366) is of late medieval or postmedieval date; it is circular, 16 mm in diameter, the mark on the seal is no longer clear, but was contained within two concentric incised rings. Lead seals were used as part of a system to regulate the production of commercially produced cloths (Egan 2001, 43). Metalworking Debris by David Starley 0 50 mm Fig. 4.99 Stamped lead object from spread 303678 The material examined in this report derived from trial trenching, strip, map and sample excavation and detailed excavation along the entire length of the scheme. All bulk debris encountered was saved and the report is based on the interpretation of this debris. Methodology A total of 26.2 kg of bulk slag was visually examined; approximately 20 kg of this total came from Margidunum Hinterland. This material was classiied into standard categories based on those used by the former English Heritage Ancient Monuments Laboratory.Visual observation of the exterior was backed up by examination of fresh fracture surfaces, the use of a geological streak plate and magnet. The A46 trunk road in Nottinghamshire has its origins as the Roman Fosse Way, and archaeological work ahead of road improvements in 2009 between Newark and Widmerpool has shed new light on both Roman and pre-Roman use of this transect of land. A number of significant sites were revealed, including evidence for Iron Age and Roman settlement in the hinterland around the Roman small town of Margidunum near Bingham. Further to the south-west near Saxondale, Roman roadside enclosures became the location of early Anglo-Saxon cremation burials and perhaps also a ‘tumulus’, as recorded by William Stukeley in 1722 in the middle of the Fosse Way. The prehistory of the landscape included further Iron Age settlements and land boundaries on the higher land of the Wolds. Earlier still was a Beaker-period ringditch and inhumation burials at the foot of the Wolds near Stragglethorpe. The story of human occupation revealed during fieldwork goes back much further, with the discovery of Late Upper Palaeolithic flintwork at Farndon Fields on the gravel terrace south of Newark. This nationally important site comprised scatters of debris left in situ by flint-knappers of the Creswellian and Federmesser hunter-gather cultural traditions. ISBN 978-0-9553534-6-8 9 780955 353468