178 



THE WILTSHIRE ARCHAEOLOGICAL AND NATURAL HISTORY MAGAZINE 



recovered from the northern end of the site. 

 Hipposandals are not a common find in Wiltshire, 

 another complete example from Cunetio is in 

 Devizes museum (from Annable's excavations at 

 Cunetio, cutting Z). They are usually thought to 

 have been used when horses or mules were taken 

 onto metalled roads. 



Possible iron-working waste/bar iron 

 Despite the small quantity of iron-working slags 

 recovered from the site three fragments of possible 

 bar iron (nos 20-22) have been identified one of 

 which is illustrated (Figure 16.2). 



Structural ironwork 



A total of six items in this category are strip or sheet 

 fragments which may be fragments of bindings or 

 similar (nos. 25-30). Timber dogs/clamps account 

 for a further two objects (nos. 23, 24). A single small 

 corner bracket, perhaps from a box or casket (Figure 

 16.3), and a large strap hinge or bracket fragment 

 were also recorded (no. 32). 



Miscellaneous and unidentified objects 

 A total of ten objects fall into this category and 

 include single examples of shoe cleat, ring, washer 

 or rove, featureless fragment, a possible lock 

 fragment (Figure 15.6) and a possible padlock key 

 fragment (Figure 15.05). The remaining three 

 objects (nos. 39-41) are long narrow rods the largest 

 two are over 200mm long, a single example is 

 illustrated (Figure 16.4). It is possible that these 

 are iron-working evidence, but may have a specific 

 function which has not been identified. 



Few of the objects can be dated with any 

 certainty to the Romano-British period although 

 styli and hipposandals are obvious exceptions. 

 Several objects do have close parallels from dated 

 Romano-British sites. These include many of the 

 tools including the hammer and paring chisels and 

 the other tools with socketed handles. The shoe cleat 

 and large strap hinge fragment also have dated 

 parallels. The condition of the ironwork, heavily 

 corroded with a thick concreted layer of sand/soil 

 surrounding them has hampered identification in 

 some instances as has the lack of solid iron core 

 noted in many objects which were cleaned. Some 

 objects, the possible lock fragment (Figure 15.6) 

 for example are too fragmentary for certain 

 identification. The range of object type and the 

 quality of many of the objects, the paring chisels 

 (Figures 17.01 and 17.2) for example are 

 noteworthy, and the presence of tools, styli, the 



hipposandal and possible lock fragment and key 

 fragment indicate a relatively sophisticated site. 



Catalogue of Iron objects from the 

 midden 



Fig. 15.5. (Object 44). Area 2, midden seg. 4087, SF327. 

 Strip object fragment. Tapering strip incomplete at both 

 ends, possibly part of a barrel padlock key. Length 

 c.70mm. 



Fig. 15.6. (43). Area 2, midden seg. 4039, SF378. 

 Object fragment comprising a strip or sheet of iron with 

 a curving element at 90a. Reminiscent of lock cases. 

 Compare with tumbler lock from Verulamium (Manning 

 1972, fig. 67, 66). 



Fig. 15.7. (18). Area 2, midden seg. 4017, SF389. 

 Brooch/buckle pin? Small pin with hooked end. Length 

 43mm. 



Fig. 15.8. (17). Area 2, midden seg. 4033, SF351. A 

 fairly ornate stylus, possibly of Manning's Type 2 (1985, 

 85-7). The tip is missing, but the point is clearly separated 

 from the stem by a distinct shoulder. The radiograph 

 indicates that there is a simple moulding at the junction 

 of tip and stem, and if this is the case this example should 

 be assigned to Type 4. The eraser is a short, broad triangle, 

 and appears to be ornamented. Extant length 58mm. 



Fig. 15.9. (16). Area 2, midden seg. 4017, SF225. 

 Incomplete stylus with broad triangular eraser as 17 

 (below). Shank appears to be simple tapering, circular in 

 cross-section. Compare with Manning's Type la (1985, 

 fig. 24) which Manning suggests may be of first or early 

 second century AD date. Whether this example has an 

 eraser broad enough to be classed as Type la rather than 

 Type 1 is debatable. Extant length 83mm. 



Fig. 16.2. (22). Area 2, midden seg. 4139, SF345. 

 Waisted rod or bar. Rectangular-sectioned the bar tapers 

 along its length . Each end appears complete, although 

 there is some uncertainty. Furthermore, each end appears 

 to taper into a blunt point when viewed from the side. 

 The radiograph, however suggests that the ends are 

 irregular and split, and shows the central part of the bar 

 to be narrowed or waisted. The irregularities, splits etc 

 suggest that this may be a scrap of bar iron . Length 

 101mm. 



Fig. 16.3. (31). Area 2, midden seg. 4233, SF239. 

 Corner bracket with one rounded terminal and one plain. 

 Length of arms 33 and 51mm, width 29mm. Probably 

 from a small box or casket. 



Fig. 16.4. (41). Area 2, midden seg. 4017. Long rod 

 of almost even thickness, square in x-section. Length 

 210mm, section c. 7 x 6mm. 



Fig. 16.5. (19). Area 2, midden seg. 4031, SF219. 

 Complete hipposandal. A Type 1 hipposandal as defined 

 by Aubert (1929), a classification continued by Manning 

 (1985, 63-65). There is no evidence for grooving on the 

 sole plate. This type of hipposandal is the most commonly 

 found in Britain, ranging in date from the AD 1 st to 4th 



