There were at least sixty-two objects of military, or possible military origin (Table 75). Comparisons with other large rural sites suggest that this is an unusually large number of military fittings for such a site. The assemblage principally consists of fittings from harness and armour, with just three iron projectile points (these may have been used for hunting rather than warfare). Most of the pieces are clearly Roman, although there are a few objects that could be pre-conquest. Objects of military origin occur throughout the Roman period at Elms Farm, though items of the 2nd-3rd century predominate. However, it seems very unlikely that Elms Farm ever had a large Roman military presence. However, the finds seem to indicate that there was some sort of military presence, from the conquest onwards. Bishop (1991) suggests the possibility of the billetting of, in particular, beneficiarii in towns as a source of military finds, and this could be considered a possibility at Elms Farm. Some of the items, such as the armillae, may have been votive deposits and some may well have been scrap metal for recycling.
1. Small leaf-shaped spearhead with an incomplete closed socket. Flat section. L. of blade 53mm, W. 32mm, overall L. 72mm. SF1187, Fill 4427, Pit 4487, Group 739, Area K, Period 4
2. Small socketed spearhead, with the very tip and most of the open socket missing. L. 97mm, max. W. 20.5mm. Layer 10492, Group 838, Period 5-6
3. Socketed projectile point, tip bent. It has a slight central rib, and a well-made closed flanged socket. L. 85mm, socket external diam. 9mm. Manning (175ff) discusses the type, classing them as small bolt-heads (Roman arrowheads are tanged). This example from Elms Farm is similar to some of a group of mid-1st century bolt-heads from Hod Hill (e.g. Manning V254). SF261, Machining layer 4000, Area A, unstratified
The following are copper alloy unless otherwise specified.
Loop, with an oval stud on the edge. The section is a rounded D-shape, and the stud is probably on the back, as this is the flat side. Examples have been found in both Late Iron Age and Early Roman contexts, and it has been suggested (Webster 1990, 294) that the plain examples, such as this one, evolved from the more decorative Late Iron Age loops. There is still some debate about the purpose of this type of object, but it is possibly a baldric or belt fitting, an interpretation supported by the similarity of this style of loop to the belt-hook found in situ in a late La Tène warrior burial at Owlesbury (Collis 1973, 126ff). Other possibilities include use as harness rings, or junction rings for bag or satchel straps (Hooley 2001, 96, no. 120). The loops with studs are principally from military contexts, and may have belonged to Celtic auxiliaries. In fairly good condition. Diam. 32mm, W. 6-9mm, stud 12x14mm. SF5159, Cleaning layer 5602, Area I, not dated
The two small fragments of chain mail are from associated contexts; layer 4937 overlies pit 14008. Both came from soil samples. They are very similar in condition and appearance, and it is very likely that they are both from the same object. The context of one piece is a Late Iron Age pit (14007), so this might be pre-Roman mail; however, as the fragment is so small, the possibility exists that it could have fallen through a crack in the ground, from the later context overlying the pit. Bishop and Coulston (1993, 59) note that chain mail was developed by the Celts.
5. Not illustrated. Iron. A small fragment of chain mail, in very good condition. There are parts of three interlinked rows, a total of seven links, although only one is complete. The complete link is circular, Diam. 6mm. SF8250, Layer 4937, Group 271, Area K, Period 2-3
6. Iron. Chain fragment in very good condition, comprising parts of five small links, none complete, made from wire 1.5mm diam. The original diameter of the links was c. 7mm. The links are now squashed together. The reason for the exceptionally good state of preservation is not immediately obvious. The context is Late Iron Age to mid-1st century AD. SF7362, Fill 14008, Pit 14007, Group 269, Area K, Period 2
All three button-and-loop fasteners from the site are presented here, though SF4759 and SF9523 are types not necessarily connected with the military.
7. Fragment of a boss-and-petal motif, almost certainly from a button-and-loop fastener of Wild's Class III. The type is later 1st century-early 2nd century, and has a principally northern British distribution, occurring mainly on military sites. The motif is discussed by Bishop (1998, 63-4) with reference to finds from Castleford, noting a possible early example from Sheepen, Colchester (Niblett 1985, fig. 64.41). The latter fragment is somewhat dubious, but occurs in a Neronian context together with a number of items of military equipment. Surviving L. 15mm, W. 15mm. SF9523, 3999, Spoil-heap
8. Button-and-loop fastener, with a mushroom-shaped button with a ridge round the edge, and a wire loop. The top has white metal and niello decoration, in poor condition. Little of the pattern can be made out, but it probably consists of a star of 'holly leaves', similar to a fastener from Vindonissa (Unz and Deschler-Erb 1997, Taf. 71, 2060). Head diam. 14mm. SF4136, Spread 4665, Group 2030, Area K, Period 2-3
This is Wild's Class VIII (Wild 1970b), which he named the 'Vindonissa' type, as so many were found there. The date is pre-Flavian to Flavian, probably out of production by the end of the 1st century; and the source is probably Mediterranean. They are particularly associated with early military sites, both in Britain and on the continent
9. Button-and-loop fastener, with a triangular loop, and part of a ring head. The junction of the ring and loop has a plain moulding, with transverse ribs either side. In fairly poor condition, with a grey patina where the surface survives. Margaret Brooks notes that the surface patches of silvery appearance may indicate a high tin bronze. This is a member of Wild Class II (Wild 1970), which is very rare outside northern Britain. The date of the type in general s 1st to ?early 2nd century. One of the few from southern Britain is from a context dated c. 75-85 at Verulamium (Waugh and Goodburn 1972, 122, no. 55), although the detail is somewhat different to the Heybridge example. A large proportion of the known examples come from the Stanwick hoard (MacGregor 1962), which consists predominantly of horse trappings and chariot fittings. Some of the Stanwick button-and-loops bear decoration which is very similar to that on definite items of horse harness (nos 33 and 45. for example), and for this class at least, it is probable that the fasteners were part of horse harness. Loop W. 17mm, surviving L. 28mm. SF4759, Layer 13576, Group 600, Area I, Period 3B
Pendants such as the following example are usually cited as decorations for military horse harness, and they are certainly particularly associated with military sites. However, it should be noted that they also occur in other contexts (e.g. at the Casa del Menandro, Pompeii) and with other associations that suggest they could be for either animal or human adornment (Allison 1997, 80).
10. A heart-shaped plate with damaged edges, a terminal knob and a central ring boss, decorated with stamped dots (Figure 563). The suspension loop is made from wire, with the end coiled back on itself. It is bent over and incomplete. The form of the suspension loop is paralleled at Aldborough (Bishop 1996, 74, no. 458), citing unpublished material from Vindonissa. L. 32mm, W. 18mm. This is a 1st-century type, and is from a late 1st-century dump over a gravel surface. SF3396, Layer 6418, Group 506, Area H, Period 3
Two of the fragments from machining layer 11000 (SF5837/6931 and SF6941) are extremely similar, and could possibly be part of the same armlet; at the least, they are likely to be from the same workshop. Their positions were not recorded, so it is not possible to say whether they were found in proximity to each other (Figure 563).
11. A fragment in fairly poor condition, with damaged edges. The squared terminal has a plain panel, then four shallow longitudinal grooves with punched dots. L. 45mm, W. 12mm. SF8132, 3999, Spoil-heap
12. Terminal, surface rather pitted. The only visible decoration consists of two transverse lines across the end, and lines down the edges, one on one side and two on the other. L. 30mm, W. 12mm. SF9530, 3999, Spoil-heap
13. Two joining fragments from the terminal of an armilla (fresh break). The end panel has two groups of transverse lines with a ?plain panel between them; the adjacent panel has a longitudinal panel down each edge infilled with transverse grooves, and a central line. The surface is in fairly poor condition. L. 20mm, W. 12mm. SF5837/6931, Machining layer 11000, Area A, unstratified
14. Part of an armilla, with longitudinal mouldings, the outer ones stamped with cross grooves. L. 17mm, W. 13mm. SF6941, Machining layer 11000, Area A, unstratified
15. Buckle with incomplete integral openwork plate, with two rivets. In fair condition, broken in antiquity. L. 35mm, W. 31mm. SF269, Machining layer 4000, Area A, unstratified
16. Just over two sides of a large rectangular buckle, with no surviving cross bar, although there is a trace of a scar where the crossbar has detached. There are traces of white metal coating. The form looks more post-medieval than Roman, especially with the white metal coating, but it is from a phase 4 pit. Rectangular buckles with cross bars are a rare Roman form. It could perhaps be something like the 3rd-century military buckles illustrated in Bishop and Coulston 1993, 152, fig. 108.2-3. 58x50mm. SF4156, Fill 4870, Pit 4913, Group 4016, Area K, Period 4
17. Strap end, bifurcated, and broken across a loop. There is a single iron rivet. There are three notches on each edge below the loop, on both faces. This is similar to the 3rd-century military types illustrated by Bishop and Coulston (1993, 152, fig. 108, nos 13-15), which either have a ring terminal, or a ring in the middle of the strap end. Examples of the latter form have been found at Brancaster (Sparey Green and Hinchliffe 1985, 48, no. 16), where a 2nd-3rd century date was suggested, and Cramond (Rae and Rae 1974, 195, no. 5), with parallels cited from Newstead. The surface is in poor condition, and the metal very mineralised. L. 41mm. SF5418, Machining layer 11000, Area A, unstratified
18. Copper alloy. Scabbard slider, incomplete. The shape is very square, and similar to an unstratified example from Watercrook (Potter 1979, 214, no. 37), but with an incomplete, distorted terminal, and a small hole at the base of the slide (Figure 563). The terminal would probably have been a plain or openwork shaped plate. The top end is missing. In fair condition, surface partly obscured by earth. L. 73mm. SF5698, Layer 7683, Group 1354, Area G, Period 3-5
19. Copper alloy. The foot of a scabbard slider (Figure 563). It has a faceted D-shaped section with a low rectangular projection on the top, 14mm from the complete end. The other end has an offset flat bar with oblique lines on the top. There is a slight suggestion that an attachment rivet has broken off the back at the base of the offset, which would make this similar to Crummy 1983, 138, no. 4923. The cross-moulding is paralleled on an iron scabbard slider from Zugmantel (Oldenstein 1976, Taf. 17.98), and the diagonal line decoration on a copper alloy example from the same site (Oldenstein 1976, Taf. 13. 60). L. 60mm. SF6508, Machining layer 11000, Area A, unstratified
20. Iron. One end is rolled over, the other is pointed. There is possibly a central perforation, though this may be a result of corrosion. Cf. Jackson 1985b, 132, no. 5. W. 8mm, L. 115mm. SF5432, Fill 4994, Pit 4989, Group 1147, not phased
21. Iron. Scabbard strap slide fragment, of the form illustrated by Stephenson (1999, fig. 29), which is 3rd century in date. L. 59mm, W. 4-14mm. SF5992, Cleaning layer 20000, Area L, not phased
22. Not illustrated. D-shaped ring from the rear face of a baldric phalera. The ring is broken. The object was made in two parts, with the shank fixed to the D-ring by a short pin on its back. There is a very similar example from Aldborough (Bishop 1996, 68, no. 423). L. 22mm, Ht 13mm. SF4905, Layer 9427, Group 1301, Area D, not dated
23. Fragment of an openwork mount with an integral fungiform back rivet (Figure 564). It is probably a military mount, with a 2nd-3rd century date; cf. Bishop 1996, 73, nos 443-4. SF2202, Layer 10126, Group 8014, Area F, not dated
24. Two joining fragments from a rectangular openwork mount with one incomplete surviving rivet on the back. The pattern is incomplete, but appears similar to Crummy 1983, 137 no. 4240; the central motif was probably different. The type is 2nd-3rd century. Broken in antiquity, and in fairly poor condition. This is one of two military objects from the context, the other being a double-headed rivet, SF5232. Surviving dimensions 30x26mm. SF5234, Fill 10621, Pit 10622, Group 828, Area F, Period 4
25. Strap mount in the shape of a fly, with an integral T-shaped rivet. Probably military. There is a parallel from Brancaster (Sparey Green and Gregory 1985, 209, no. 35). L. 21mm, W. 15mm. SF4287, Machining layer 11000, Area A, unstratified
26. Horse harness pendant, military, with the loop terminal and central rivet missing, and damage to the edges (Figure 564). It has four crescentic cells, two with red enamel surviving, a central cross with a pale green, glassy appearance, and incised linear and punched dot decoration . It is virtually identical to an unstratified pendant from Harlow Temple (Gobel 1985, 90, no. 117), where it is noted that the cells may contain niello. It was claimed as pre-Flavian (Claudian?) in date, though it should be noted that both it and the Gallic example cited are unstratified. There is another example from Moulsham St, Chelmsford (Major in prep), from a medieval context. There is a close parallel from Vindonissa (Unz and Deschler-Erb 1997, Taf. 56, no. 1556). L. 37mm, W. 21mm. SF3881, Machining layer 11000, Area A, unstratified
27. Openwork fitting, probably a 3rd-century cavalry pendant. Roughly heart-shaped, with the loop broken. It is rather poorly finished, and in fairly poor condition, with traces of white metal coating. L. 34mm, W. 16mm. SF4670, Machining layer 11000, Area A, unstratified
28. Sheet pendant. The edges are damaged and distorted, and the original shape is uncertain. The centre is bossed, with a teardrop-shaped cutout below. It is decorated with punched dots and incised lines. L. 65mm, max. surviving W. 32mm. SF3069, Machining layer 11000, Area A, unstratified
29. Heart-shaped pendant, openwork, terminal missing. There is a notch either side at the base of the loop, a common feature of the type. This is a 3rd-century harness pendant, occurring either as a solid or an openwork heart (Bishop and Coulston 1993, fig. 112), often with a phallic terminal, attached to a riveted hanger. The suspension loops were cut in order to fit them onto the hanger, a feature clearly seen on examples from Winterton (Stead 1976b, 213, no. 111) and a larger version from Chester (Thompson 1976, 190, no. 11). The Heybridge example appears to have had the loop forced open, presumably to detach it from its hanger. The date of the object is later than the date of its context, so it is possibly intrusive. W. 22mm, surviving L. 31mm. SF3510, Prepared surface 13045, Group 381, Area H, Period 3
30. Not illustrated. Conical stud with integral shank and washer. The type is typically military, dating from the 2nd to 3rd centuries. Cf. an example from Aldborough (Bishop 1996, 74, no. 452). In fair condition, with recent damage to the head. Head diam. 26mm, Ht 15.5mm, rove diam. 13mm. SF5559, Machining layer 11000
31. Not illustrated. Conical stud with integral shank and washer, as no. 30. Recent damage to the head, in fair condition, surface obscured by earth. Original diam. c. 30mm, Ht 13.5mm, washer diam. 12mm. SF6076, Fill 8747, Pit 19172, Group 678, Area P, Period 6
32. Oval vulvate stud with two integral rivets, hollow backed (Figure 563). The edges are damaged. This is a fairly common military type of later 2nd-3rd century date. Among the parallels are examples from Brough-on-Humber (Wacher 1969, 89, no. 19), and one from a Late Roman context at the Duckend Carpark Site, Stansted (Major 2004d, 277. no.18). L. 20mm, W. 11mm, Ht 9mm. SF5315, Machining layer 11000, Area A, unstratified
33. Vulvate stud with two integral rivets on the back. Unusually for the type, part of the edge has survived, showing that it had bowed sides and almost straight ends, with an oval dome. The surface is in rather poor condition. L. 31mm, W. 18mm. SF5491, Machining layer 17000, Area Q, unstratified
34. Peltate stud, with two integral rivets on the back; in good condition. The type is military, and probably associated with auxiliary cavalry. The type is quite common, with examples from many sites including Gestingthorpe (Henig 1985, 36, nos 98-99). L. 25mm, W. 25mm. SF2989, Machining layer 11000, Area A, unstratified
35. Stud, with nine moulded petals. Possibly military. In good condition, slightly squashed on one side. Diam. 16mm, Ht 10mm. SF5314, Machining layer 11000, Area A, unstratified
36. Keel-shaped rivet with an integral shank and back disc. It is slightly hollow underneath. There is a similar rivet from the temple at Lamyatt Beacon, thought to be possibly military (Leech 1986, 321, no. 17). They occur as 2nd-3rd century auxiliary fittings in the Rhineland, though more typically with two integral rivets (Oldenstein 1976, Taf. 58). L. 27.5mm, Ht 12mm. SF7010, Machining layer 11000, Area Q, unstratified
37. Cast double-headed rivet or fastener. Round-sectioned shank with a circular head either end, one with a circumferential line and a central dimple. The type is typically military, dating from the second half of the 2nd century to the 3rd century. There are examples from many sites in Britain, including an almost identical, though larger, one from Vindolanda (Bidwell 1985, 122, no. 34). SF672 is on the small size for the type, with a distance between the heads of only 6mm. Head diam. 14-15mm. SF672, Machining layer 4000, Area A, unstratified
38. Double-headed rivet with circular ends, cast. The top has a raised edge, and a small, shallow hole in the middle. In fair condition, with slight damage to the edges. This is one of two military objects from the context, the other being a fragment of an openwork mount, SF5234. Ht 14mm, head diams 16mm and 15mm. SF5232, Fill 10621, Pit 10622, Group 828, Area F, Period 4
39. Rivet with an integral shank and circular rove, and a cockleshell-shaped head. The surface is in poor condition, and no detail survives. Possibly military. Parallels include one from Cambridgeshire (Taylor 1987, 31, no. 188). SF6722, Machining layer 11000, Area A, unstratified
40. Belt buckle plate, with integral propeller stiffener. The axis bar is iron, and two of the four solid copper-alloy rivets are present. There is no decoration apart from notches along the edges, and the piece is not particularly well finished. The back has been left fairly rough. L. 45mm, W. 42mm. SF3576, Machining layer 11000, Area A, unstratified
Propeller belt-stiffeners are commonly occurring elements of 4th-century military belt fitting sets, but this is a very rare example of a combination plate and stiffener. Bishop and Coulston (1993, 173) noted that they were found only on Rhenish and Danubian sites, and that none had been found in Britain. However, there has been at least one found in Britain since then, besides this example. The type can be as late as the early 5th century, but the width of this example is at the lower end of the range, suggesting a date in the first half of the 4th century.
41. Military buckle fragment, comprising part of a loop with a scrolled terminal, with a bar below. There is an integral plate with two notches on the edge, and part of a circular cell with a central hole, presumably for the fixing stud. The edges of the plate were cut in antiquity. The back is flat. L. 35mm, W. 12mm. SF7197, Machining layer 11000, Area A
This object is unusual, as the type normally has a separate plate. The normal date for such a buckle with plain scrolls would be earlier Roman (see, for example, Bishop and Coulston 1993, fig. 59); on later examples the scrolls form the tails of opposed dolphins (Bishop and Coulston 1993, fig. 126). The integral plate and notched edge suggest a later Roman date for this piece, despite there being no clear indication that this was a dolphin buckle. There is also one style of very Late Roman/Early Saxon buckle with involuted loops and an integral plate, Evison's quoit brooch style buckles (Evison 1968), a type found in Early Saxon graves at Orpington, Mitcham and Bishopstone. These buckles are rather more ornate than this example, with less crisply modelled scrolls, and while not claiming that the Elms Farm buckle is the same type, it could be of similar or slightly earlier date.
42. Openwork buckle plate, of 4th-century military type. It is incomplete, having apparently been cut up in antiquity. Cf. Bishop and Coulston 1993, 175, no 3. The two holes at the buckle end are of different sizes, suggesting that this was originally likely to have been twice as wide. One hole has iron staining round it, probably from an iron stud or rivet, and there are the stubs of two integral rivets on the back. One corner has a ring-and-dot motif; the other corner is obscured by iron corrosion. The back has casting burrs. 26x24mm. SF5710, Machining layer 11000, Area A, unstratified
The following four ribbed fittings are probably military, with 4th-century parallels (Figure 565). One (probably two) are slotted fittings, and two are studs, but all are similar in size and profile.
43. A cast, rectangular-sectioned fitting, with a slot through the length. The top has moulded longitudinal ribbing; the sides and base are plain. There is a possible parallel from Gorhambury (Wardle 1990, 126, no, 175), from a 4th-century context. It is described as a mount, has three broad ribs on the top, and is exactly the same size as the Elms Farm example. The Gorhambury mount is clearly incomplete, but it is unclear where the breaks are, and it is impossible to confirm that they are the same type of object without seeing the Gorhambury mount. In good condition. Fibres visible in the slot are probably recent plant material. 23x22x9mm, slot W. 19mm. SF309, Machining layer 4000, Area A, unstratified
44. The top of a fitting with a double-humped section, similar to 4000, SF309. There is no sign of a shank, and two of the edges are broken at right-angles, so this may be a slide. 20.5x17mm. SF5419, Machining layer 11000, Area A, unstratified
45. Stud, shank broken. Square, with a double humped profile, and a lower median rib, hollow on the back. There are parallels from Verulamium (Goodburn 1984, 47, no. 145), from a context dated AD 330-360, Wickford, Essex (Bishop 1991, fig. 5.3) and a metal detector find from Magdalen Laver, Essex (unpublished). The condition appears good, but it is extensively covered in earth and ?mineralised organic material. 22x22mm. SF1190, Machining layer 4000, Area A, unstratified
46. Not illustrated. Half of a rectangular double humped stud, shank broken. L. 21mm, original W. 20mm. SF9529, context 3999, Spoil-heap
47. Small fragment from a flat plate object, damaged and incomplete (Figure 565). It comprises a broken loop on the end of a strip or plate, broken on one side. There is a pierced hole adjacent to the loop. This is probably part of a Late Roman amphora-shaped strap end. Bishop and Coulston (1993, figs 125 and 126) illustrate a number of strap ends of this form. L. 22mm. SF7053, Cleaning layer 5597, Area I, not dated
48. Strap end (Figure 565). It has a narrow, rectangular attachment loop, with an asymmetric shaped terminal, possibly incomplete. There is some grooved detailing, but the condition of the piece has rendered the pattern somewhat unclear. Both faces were probably the same, with a rounded edge. A strap end of the same general form, and with a rectangular loop of exactly the same size, came from Richborough (Cunliffe 1968, 95, no. 120). The latter was possibly 4th century, although a cited parallel was possibly 2nd century. The form of this example is very unusual; asymmetric strap fittings are rare. This object appears to be representational, but it is unclear what it is meant to be. L. 38mm, max. W. 10mm. SF738, Machining layer 4000, Area A
49. Strap junction loop, damaged, incomplete, and bent. One integral rivet survives on the back. It appears to be plain, bar low transverse mouldings, which is quite unusual. Cf. Unz and Deschler-Erb 1998, Taf. 61, no. 1684. The present example differs in that the loop is narrower than the plate. The surface is poor in places, and powdery. L. 40mm, max. W. 10mm. SF930, Machining layer 4000, Area A
50. Oval ring with a variable section, oval to lentoid (Figure 565). It has a flat oval 'bezel' on top, and a 'skirt' at the bottom, with moulding at the junction. This is probably part of a terret, with the bottom loop missing. This 'skirted' form is particularly associated with the military, although the bezel and junction moulding are unusual. The general shape is illustrated by an example from Verulamium (Waugh and Goodburn 1972, 130, no. 126). In fair condition. Loop W. 43mm, Ht 31mm, overall ht 47mm. SF7172, Machining layer 11000, Area A
Some studs of possible military origin have been included with the non-military studs in Function Category 11.
51. Strap fitting (Figure 565). A plain disc with a central knob and a rectangular back loop. In good condition, some earth on the surface but details clear. Edge damaged. Presumably a military harness mount, although the single loop is unusual. Diam. 23mm, Ht 14mm. SF669, Machining layer 4000, Area A
52. Strap fitting (Figure 565). A disc with a rectangular loop on one side and a small circular loop on the other. The edge was damaged in antiquity. This is probably a harness fitting. A disc of the same size with a back loop came from a 1st to 2nd century level at Nettleton (Wedlake 1982, 210, no. 30). The condition is fair. Diam. 19mm, Ht 15mm. SF1455, Fill 10001, Pit 10018, Group 837, Period 5
53. Bipartite barrel-shaped fitting, probably a strap slider with a rectangular back loop, now broken (Figure 565). A comparable piece came from a 2nd-century context at Lullingstone (Meates 1987, 74, no. 162). L. 18mm, W. 7mm. SF9532, Context 3999, Spoil-heap
54. Rectangular mount with four integral rivets on the back (Figure 565). The central rectangular hole has a ridge round the edge, and lines at either end. The surface is in poor condition, with patches of white metal coating surviving at one end only. 45x24mm, hole 14x6mm. It is possibly allied to the hinged or rivetted openwork plates such as those illustrated by Oldenstein (1976, Taf. 62.783, Taf. 63.815). A similar plate from Dragonby, also unstratified, has millefiore inlay (Knowles and May 1996, 273, no. 26). SF955, Machining layer 4000, Area A
55. Hook or strap fitting made from a sheet strip (Figure 565). The squared end is notched, and there is a single rivet through this end, with most of the shank missing. The strip is cracked and bent. The other, tapering, end is bent over and possibly incomplete. There are traces of white metal coating and hints of decoration, but the surface is in very poor condition. This is possibly a military fitting, such as a hinged strap fitting (e.g. Bishop 1998, fig. 67, no. 222). Military pieces are frequently tinned, but rarely have only a single rivet. Original L. 27mm, W. 8mm. SF4142, Fill 4774, Pit 4774, Group 1147, Area K, not dated
56. Loop from a cavalry harness strap junction (Figure 565). There is a transverse moulding at the base of the loop and the front strap is slightly waisted, with two rivet holes. The back strap is missing. This is very similar to Unz and Deschler-Erb 1997, Taf. 62, no. 1751. The object is somewhat distorted, and the surface is in poor condition. L. 43mm. SF3003, Machining layer 11000, Area A
57. Strip fragment in fairly poor condition, one end broken in antiquity, the other recently. It has a central, slightly domed, element flanked on either side by two transverse lines. This is probably part of a military strap junction; cf Unz and Deschler-Erb 1997, Taf. 62, nos 1726-30. L. 21mm, W.12mm. SF5616, Machining layer 11000, Area A
58. Small scallop-shell shaped terminal from a mount or pendant, with a flat back and moulded upper surface. This is a common motif in military pendants and mounts; cf. Unz and Dechler-Erb 1997, Taf. 50, nos 1369-72, Taf. 59, nos 1658 and 1662 etc. The context is late 1st century BC to mid-1st century AD; if the object is not intrusive, then it must belong at the very end of this date range, since it is a Roman form. L. 8mm, W. 9mm. SF7521, Fill 20304, Pit 20335, Group 43, Area L, Period 2A
59. Rectangular plate in poor condition, split and cracked in antiquity. There is a low flange on the back along one long side, and a very slight flange on the other. One end has a rectangular bar projecting at right-angles to the back, with a short stub in the middle, possibly a broken shank. The other end of the plate is broken. The top has a longitudinal panel of enamelled lozenges and triangles, some of the lozenges retaining blue enamel, with another, unknown colour (now brownish, possibly originally red) in the triangles. One of the lozenges has a small blob of copper alloy in it instead of enamel. The detail of the frame is poorly preserved. There is probably a groove along each long edge, and the end with the bar on the back has at least one transverse line, and possibly a ring-and-dot at the corner of the enamelled panel, although there is no trace of any others. The end bar has a small curved line in relief. This may be allied to the enamelled belt-plates of Bateson's Group 1 (Bateson 1981, 54), which are generally enamelled in red and blue (as this piece may have been), and some of which have similar patterns. Most are from military sites, and they may be of 2nd century date. L. 60mm, W. 29mm, end bar 16x8mm. SF4674, Machining layer 11000, Area A
60. Pelta-shaped mount, incomplete, with two projections on the back. The concave edges are outlined by ridges. The surface is in poor condition, and powdery. Probably military. W. 29mm, L. 22mm. SF19, Machining layer 400, Area W
61. Part of a mount, possibly leaf-shaped with a terminal knob, and one surviving integral rivet. There is a line down the middle, and the back is rough. Possibly military. In good condition. L. 28mm, max. W. 20mm. SF692, Machining layer 4000
62. Dumb-bell toggle. In fair condition, with some damage to the surface. L. 26.5mm, head diam. 12mm. SF5421, Machining layer 11000
This is a well-known type that can be made from either bone or copper alloy, and occurs in both Late Iron Age and Roman contexts. Some are flat backed, others modelled in the round, as is this example. Their distribution is principally in northern Britain, and mainly from military sites. There is a very similar example from Dragonby (Knowles and May 1996, 273, no. 34). Their purpose is obscure; a bone example from an Arras burial tradition grave at Acklam was probably from the area of the lower right leg, and possibly associated with the sword from the burial, leading to the conjecture that it was attached to the harness supporting the scabbard (Dent 1983).
Internet Archaeology is an open access journal. Except where otherwise noted, content from this work may be used under the terms of the Creative Commons Attribution 3.0 (CC BY) Unported licence, which permits unrestricted use, distribution, and reproduction in any medium, provided that attribution to the author(s), the title of the work, the Internet Archaeology journal and the relevant URL/DOI are given.