SECTION 3.4::CLAYDON PIKE SMALL FINDS by Hilary Cool

INTRODUCTION

PERSONAL ORNAMENTS
Brooches
Bracelets
Finger rings
Necklaces, beads and pendants
Hair pins and dress pins
Ear rings
Hobnails and Shoe cleats
Buckles and belt plates


TOILET AND MEDICAL EQUIPMENT
Small implements
Long-handled implements

TEXTILE EQUIPMENT

HOUSEHOLD EQUIPMENT

RECREATIONAL ITEMS

ITEMS ASSOCIATED WITH WIEGHING AND MEASURING

ITEMS ASSOCIATED WITH WRITING

ITEMS ASSOCIATED WITH TRANSPORT

STRUCTURAL ITEMS FROM BUILDINGS

TOOLS AND EVIDENCE FOR CRAFT AND INDUSTRY

FASTENINGS AND FITTINGS

AGRICULTURAL ITEMS

MILITARY EQUIPMENT

RELIGIOUS ITEMS

MISCELLANEOUS ITEMS

OVERVIEW
The middle Iron Age settlement
The late Iron Age and early Roman activity
The Roman complex
The late Roman villa
Saxon and medieval activity

BIBLIOGRAPHY

Top of Page

Introduction

This report deals with all the small finds from Claydon Pike other than the coins and the objects made from stone. In total 2652 items are considered here (see Table 1). This consists of all the material that could be assigned typologically to a medieval or earlier date and all material that was stratified in Period 4 or earlier contexts. Unstratified undiagnostic items such as featureless fragments of metal, and post medieval and modern items have been excluded from consideration.

It should be noted that approximately a quarter of the material, primarily iron, could not be personally inspected as the material was inaccessible during the period of study. Information about these items is derived from the x-radiographs and such drawings as had been made of the items during the earlier post excavation programme. In some cases it was impossible to identify the items securely by these means and reliance had to be placed on the identifications made during the 1980s. In the database such entries are indicated by having the simple name placed in brackets. Thus hobnail has been securely identified but (hobnail) indicates a degree of uncertainty. In some cases it was no longer possible to read the find number on the bags of finds. This generally affected the ironwork. Again original identification on the database had to be accepted in default of secure identification.

In this report the material will first be summarised by function. This part will provide basic date and distribution information about the objects. The second part will look at the material within the context of Claydon Pike. As with the other small finds reports the full consideration of what the finds are telling us about the site will be reserved for the overall consideration of all the finds from all the sites being considered by this project (Print Chapter 13).

Top of Page

Personal Ornaments

Brooches

Brooches were the commonest item of jewellery found (see Table 2). Many of them were types that were also found at Somerford Keynes. Given that an extensive brooch report has been written for that large assemblage, detailed typological comment is not appropriate here and the reader is referred to the Somerford Keynes report for the detailed considerations of the types. In what follows the conclusions will just be summarised, and only forms not represented there are provided with more detailed comment.

Early to mid-1st century forms (Figure 3.4.1a: Early to mid-1st century forms)

Colchester – Hull Type 90

Amongst the pre-conquest brooches the most interesting one is the single example of a Kragenfibeln (1045) from a Phase 2 context in Trench 13 which is clearly a pre-Conquest continental import (Feugére 1985, 245 type 10). In general on the continent they are known from the mid 1st century BC, but most are found in contexts belonging to the end of the 1st century BC and into the 1st century AD. This example is unusual because it appears to be hinged and this form is only encountered rarely (Hattatt 1987, 31).

The three Colchester brooches (Hull type 90 - 2802, 193, 2402) are indicative of pre-conquest occupation as the type in general had probably ceased to be made by the conquest. These were either found residually (2802 from a Phase 2D ditch fill) or were unstratified. The Nauheim Derivative brooches could also suggest early occupation though their floruit did continue post-Conquest. There are two examples of the simple wire form of Hull Type 11 (618, 3033), one of the possibly Atrebatic Hull Type 10D with a very narrow lower bow (18) and one where only the spring and pin were extant (2612). One of the wire examples (3033) had the less commonly encountered central mouldings. This brooch was the only one found stratified but was clearly residual in the Phase 3b/c ditch fill. It should be noted that there may also be an iron Nauheim Derivative in the assemblage (1893). This piece was originally identified as such but I have not been able to re-examine it and the drawn and X-radiograph evidence is ambiguous. It was found in a Phase 3 gully and again would be residual if the identification is correct.

The evidence from the Durotrigian Strip bow brooches is more ambiguous with regard to the pre-Conquest occupation. There is only a single example of Hull Type 12 with a tapering bow (635) but two of Hull Type 53 with the narrow lower bow and footknob (659, 989). The latter form was not represented at Somerford Keynes. This variant is a British imitation of the Aucissa type (Hattatt 1987, 74) and as such most likely to be post-Conquest as that is the period when the majority of the Aucissas were in use in Britain. It is possible that the context in which the fragmentary 989 was found may have been broadly contemporary with the likely floruit of the brooch (i.e. c. AD 43 to c. AD 75) as it was the fill of a Phase 2c enclosure ditch. 635 from a phase 3 context is clearly residual while 659 was unstratified. The other two brooches for which a pre-Conquest date is possible are the rosette brooch (Hull Type 27 - 2993), for which a date during the second quarter of the 1st century is appropriate and the Birdlip brooch (1279). The latter form was not found at Somerford Keynes. The form is not well dated but an early to mid 1st century date seems most likely (Hattatt 1989, 20; Mackreth 1998, 131). They are certainly not common in post-Conquest assemblages. Despite the form taking its name from a Gloucestershire site, this is a form with a widespread distribution the majority having been found in the east (Hattatt 1989, fig. 8 – though more could now be added for the Gloucestershire region now) Both this example and the rosette brooch were unstratified.

Mid to late 1st century forms (Fig. 3.4.1b: Mid 1st century forms)

The commonest mid 1st century (post-Conquest) form was represented by the Hod Hill brooch. There were four examples all of different variants (2597 – Hull Type 62; 51 – Hull type 63; 1612 – Hull Type 74; 2519 – fragmentary). All were found either residually in Phase 3 or 4 contexts or were unstratified. The only other brooch that can be assigned to this period with certainty is an unstratified Aesica (Hull Type 37 – 646), but it should be noted that several of the Colchester Derivatives discussed in the next session are also variants that are most likely to have been in use in the mid 1st century.

Mid 1st to 2nd century forms (Figs 3.4.1c: Mid 1st to 2nd century forms, 3.4.1d: Late 1st to mid 2nd century forms and 3.4.1e: 1st/2nd century AD forms)

This category is dominated by the Colchester Derivative family. The basic form is represented by six brooches. Two are examples of Hull Type 92 with the cavetto moulding on front of their bows (1194, 3031), the others belong to Hull Type 93 (2779, 1430, 2976, 2908). Interestingly both of the Type 92 brooches and two of the Type 93 (2779, 1430) have their springs held by rearward facing hooks which would suggest a mid 1st century date rather than one later in the century. The other two have the more efficient Harlow method of spring attachment and could be later. 3031 came from a phase 2/3 context and 2976 from a Phase 3 one and both could be contemporary with their contexts.

As is to be expected in this area, it is the Polden Hill form of Colchester Derivative that is most numerous here. There is a single example of the light Polden Hill form (Hull Type 97 - 674) belonging to the mid 1st century and which appears commonest in south Wales. The commonest Polden Hill variant here is the form with the stepped flanges on the upper bow (Hull Type 103) for which a date range of c. AD 65 – 125 seems most likely. Two of these have the decorated return plate which seems to be a feature of the Gloucestershire area (334, 733) but this is absent on the third (385). The form that was commonest at Somerford Keynes (Hull 98 variant – with transversely nicked crest ridge, perforated catch plate) is represented by at most two examples here (1386, 1722 – the latter obscured by corrosion and the identification not secure). 755 may also be related but lacks the normal perforated catchplate. 2200 by contrast seems to belong to the standard Polden Hill Hull Type 98. There are also two Polden Hill brooches that show elements of hybridisation with other brooch types. The heavily corroded 2682 has heavy mouldings and like Somerford Keynes 78 has a central moulding that may hint at a relationship with trumpet brooches. 183 has enamelled decoration and a sharply angled head and could either be viewed either as a variant of a Polden Hill or of the T-shaped brooch form Hull Type 110. Two of these brooches came from Phase 2 contexts (183, 755) and one (674) came from a general Phase 3 spread; the rest were either unstratified or from unphased contexts.

In addition to the brooches which could be assigned to a particular variant there were a further three (532, 782, 1510) which were too corroded for this to be possible. The six lower bow fragments (441, 951, 1103, 2833, 3036, 3074) are also likely to have come from the Colchester Derivative family. 3036 is noteworthy as it has a decorated return to the catchplate.

Late 1st to mid 2nd century forms (Fig. 3.4.1f: 2nd century AD forms)

Lower Severn T-shaped brooches are represented by three examples, all unstratified. 1766 is a Hull Type 110 and 738 is a Hull Type 122. The latter shows the typical casting flaw often seen on these brooches where the headloop is obscured by a thin film of metal indicative of the fact that though the brooches had headloops, they were probably not worn chained in pairs. 2639 has affinities to Type 122 but has the decorated return to the catchplate typical of the Gloucestershire area, and it is likely that is did not come from the same source that produced the standard Hull Type 122 which are a very homogeneous group.

The only trumpet brooch in the assemblage is an example of a Hull Type 154, the local Chester type (878) from a Phase 3 layer of cobbles. The makers of the very ungainly 707 and 722 though appear to have been influenced by the trumpet brooches as the central mouldings could be viewed as an extreme acanthus. They are examples of plate-headed T-shaped brooches (Hull Type 140). They seem very likely to have been made as a pair though they were not found together, 722 coming from the fill of a large ditch to the west of Trench 17 and 707 coming from the general cleaning level in Trench 17. The ditch fill is not closely dated so unfortunately these provide no useful information to help resolve the dating of this poorly understood type. As noted when discussing the examples from Somerford Keynes, this is a type whose main distribution has hitherto been to the south of Gloucestershire.

There is one example of an equal-ended plate brooch with zoomorphic terminals (2969) found residually in a Phase 4 gully. This is an example of a Hull Type 228, a basic form that is widespread throughout the western Empire with an overall date range of the later 1st to later 2nd century (Riha 1979, 191 Typ 7.16). The form is likely to have been made a numerous centres and it is not currently possibly to refine the dating for the numerous variants. This example is likely to have been very similar to one found in Norfolk (Hattatt 1985,154 no. 560) with alternating blocks of enamel in the now empty outer central cell.

Second century and later forms (Fig. 3.4.1g: 2nd century and later forms)

A small number of brooches all from unstratified contexts can be assigned to the 2nd century or later. 3039 is an example of an oval enamelled disc brooch (Hull Type 250), a 2nd century British form (Mackreth 1985, 297 no. 41). 2424 is a fragmentary lugged disc brooch (Hull type 262). In general this is a form that is of 2nd century date continuing into the 3rd century (see for example Mackreth 1995, 977 no. 104; 1998, 145). The third plate brooch (1253) is a less common form (Hull type 258). This is a relatively rare British type (Hattatt 1985, 146 no. 538). The paucity of examples makes dating difficult but the presence of one in a rubbish deposit in Turret 18B on Hadrians Wall (Woodfield 1965, 89) where there is no evidence of activity after Wall Period 1 would suggest they were in use during the second quarter of the 2nd century. 2577 is an example of a cockerel brooch (Hull Type 214). These have a wide distribution in Britain but the dating is not very exact. A 2nd century date is most likely (Crummy 1983, 15 nos. 75-6)

There is also a single example of an enamelled knee brooch (2549 - Hull Type 173). This is a British form (Hattatt 1987, 263) with a later 2nd to 3rd century floruit.

Miscellaneous brooch fragments

In addition to the brooches discussed there is a very corroded bow brooch that cannot be assigned to a type (655), numerous fragments of pins (151, 152, 169, 515, 535, 586, 748, 1051, 1505, 2149, 2598) and two copper alloy spring fragments (728, 1821) and possibly one of iron (2878). A pin from a Phase 3 context (364) is especially noteworthy as it is the only possible evidence for the presence of penannular brooches on the site.

Brooch summary

The brooches are summarised by date and period in Table 3, and by date and area they were recovered from in Table 4. These show a strong mid 1st century presence reflecting the pre-conquest origin of the occupation in Trench 13. This is missing for the brooches in Trench 17. The profile of the small brooch assemblage from Trench 29 is more similar to that from Trench 17 than from Trench 13. Interestingly the brooch assemblage from Trench 13 looks more ‘cosmopolitan’ in its earliest stages (equating to the earlier part of the Phase 2 occupation) than it does for the period when it is suggested it became an official Roman settlement complex. It is the earlier material that suggests the widest contacts containing as it does the Durotrigian strip bow brooches, the Kragenfibeln and the Birdlip brooch. The later 1st and 2nd century occupation on this site provides a curiously impoverished brooch assemblage.

Bracelets (Figs 3.4.2a: Twisted, penannular and expanding bracelets, 3.4.2b: Light bangles , 3.4.2c: Light Bangles, 3.4.2d: Miscellaneous bracelets/fragments, 3.4.2e: Shale bracelets and 3.4.2f: Glass bangle )

Bracelet wearing was a fashion of the late Roman period in Britain and most bracelets belong to the late 3rd or 4th centuries with earlier examples being relatively rare. Large quantities of bracelet fragments are not uncommon from late Roman sites reflecting the fashion for wearing several bracelets at a time (see Somerford Keynes section 5.3). Bracelets were the second commonest item of jewellery at Claydon Pike (see Table 5) and, with the exception of a single fragment of a glass bangle, all could be of late Roman date. This is reflected in the contexts they were found in as none come from an unambiguous Phase 2 context.

Cable twist bracelets are the single most common type bracelet from Roman Britain (Cool 1983, 120 Group I). They are represented here by 12 examples (9, 614, 820, 885, 1202, 1432, 1683, 1756, 1928, 2161, 2790 and 2861). They are known in early Roman contexts but by far the majority of them come from 4th century contexts. There are slight hints that those where the wires forming them are made of different alloys may be early in the sequence. 1683 appears to fall into this category judged by the differential corrosion products but was found in a Phase 4 context so an early date for it cannot be proven. Expanding bracelets (2170, 2404) have a similar date profile to the cable twist ones (ibid 130, Group III) as do the torc twisted bracelets (181, 539, 809, 839, 3037). For the latter group (ibid 135, Group IV) it is possible to refine the dating if the terminals are present. This is only the case here for 809 and 3037, both of which would have had hook and eye terminals and so can be assigned to the later 3rd to 4th century with some certainty.

As already discussed in the Somerford Keynes report, penannular bracelets occur intermittently throughout the Roman period (Cool 1983, 138 Groups 5 to 8). Here there are two examples of the plain variety (879, 1726,), and two with terminals decorated by ribs and grooves (172, 3095). There are also two fragments where the terminals are missing (2508, 2820). None can be independently dated closely.

Most 4th century bracelets belong to the light bangle tradition (Cool 1983, 152) and so it is of no surprise to see so many examples here. What is slightly unusual are the types present. These bracelets come with numerous different decorative patterns. In my 1983 survey I noted 17 major divisions, often with several sub-types in each. In the two decades or more since I completed that work many more have been published, but the pattern has not materially altered. Table 6 compares the Claydon Pike light bangles to the overall figures derived from Cool 1983. As can be seen one of the commonest types normally found with zig-zag decoration is virtually shunned and overwhelmingly those with dotted decoration were preferred. This is interesting as normally it is very difficult to pick out any regional patterns in these bracelets, but these figures might hint at a preferred design by a local workshop.

Other 4th century bracelets in the assemblage include two fragments (598, 1231) from the multiple unit style of bracelet (Cool 1983, 181 Group 31) and one fragment of a perforated bracelet (1957). The latter is from a relatively small group which has a predominantly south-western distribution (Swift 2000, 163). There is also a slender plain hook and eye bracelet (10 - ibid 1983, 196 Group 34). There are hints that the latter type may have been more popular in the later part of the century. Another later 4th century type is the bone bracelet 2819. Such bracelets (MacGregor 1985, 112) are known to have been in use earlier in the century but seem to have been particularly popular in the later 4th century and into the 5th (Cool 2000, 49).

There are also five bracelets made of black shiny material. Two are simple annular shale bracelets (1812, 2996). These were in use throughout the Roman period (Lawson 1976, 248). The three jet bracelets are late Roman. 1820 is a fragment from an elaborate chip-carved jet annular bracelet (Allason-Jones 1996, 30). 2131 is from a bracelet made of beads. These are normally elliptical (see Allason-Jones 1996, 27-8 nos 26-37) but gadrooned hemispherical ones are known (cf Lawson 1976, 244 no. 4), and have been found forming a bracelet from the East Cemetery at London (Barber and Bowsher 2000, 197 no. B452.2). It is likely that the transversely pierced pyramidal bead (3114) is also from a bracelet. Four similar beads were found in a grave belonging to the first half of the 4th century at Kelvedon (Rodwell 1988, 76, fig. 61 no. 2). There was no information about the position of the beads in relation to the body but there were no other beads with them and it seems unlikely that they were part of a necklace. Transversely pierced rectangular beads of a slightly different type were certainly used to form bracelets as can be seen from one in another London grave (Barber and Bowsher 2000, 166 no. 291.2-4)

Finally the unstratified glass bangle (103) can be considered here, though it should be noted that these annular rings were not necessarily worn as bracelets like the rest of the items discussed in this section (see Price 1988, 354). It is an example of a Kilbride-Jones (1938) type 2 bangle of a relatively rare variant with two central cords and two edge cords. Price (1988, 345) has noted that the only close parallels appear to be two from Yorkshire. A mid to late 1st century date would be most appropriate for its manufacture but fragments of these bangles do appear to have been curated as charms so it may have been lost or deposited some time after that date.

Table 7 shows the distribution of the bracelets across the different areas excavated. Given that most of them are of late Roman date, this table is a good indication of 4th century occupation which can be seen to be widespread. Attention may be drawn to the late 4th century forms present on sites 5 and 19.

Top of Page

Finger rings (Figs 3.4.3a: Phase 2-3 rings and 3.4.3b: Phase 4 rings)

The finger rings recovered at Claydon Pike span the whole Roman period. The earliest ones are the simple expanded form which was primarily a medium for carrying an intaglio, rather than a simple item of jewellery (Henig 1978, 47 types II and III; Cool 1983, 226 Type IV). In general these were in use in the 1st and 2nd century and into the 3rd century. One example (1080) is of iron and retains its glass intaglio. The device is of an eagle with wings displayed, standing on a thunderbolt (Fulmen) with a globe below. Dr Henig contributes the following note suggesting a 1st century date for the piece.

The combination of eagle, thunderbolt and globe is to be seen on an agate intaglio from Aldborough (Isurium Brigantium) in Yorkshire (Henig 1978, 105, 293). Eagles are birds of Jupiter (as the thunderbolt emphasises) and also symbols of Rome (see Tacitus, Ann. II, XVII, and as intaglios showing fulmen-holding eagles on alters ornamented with the Lupa Romana suckling Romulus and Remus make clear; Zwielein-Diehl, 1979, 82, Toynbee 1973, 240-42). The theme of world rule was appropriate both to Jupiter and to his beloved Rome (Vergil, Aeneid VI, 847-53). Close comparison may be made between the eagle on the Claydon Pike paste and the eagles on gems from Ham Hall, Somerset and Hod Hill, Dorset, where the wings are also displayed (Henig 1978, 269 No. 696, 271 No. 708). Intaglios of clear and translucent coloured glass are most common in the early days of the Empire (except for the green and blue Romano-British imitations) and the form of the ring confirms a 1st century dating.

Another glass intaglio (2058) also has a device that would seem to allude to themes popular within the Roman world. Here it depicts clasped right hands (dextrarum iunctio) depicted schematically. They clasp two cornucopiae - that on the right complete but of the left only an upper corner remains. Above them is a lighted altar. Of this piece Dr Henig writes;-

Intaglios showing fortunate symbols were common in the Roman period, especially in the 1st century BC and 1st century AD when pastes of this type were mass produced. Dr Marie-Louise Vollenwieber may well be right that their primary purpose was as symbols of political allegiance, the altar invoking not merely pietas towards the gods but the veneration of the genius of Augustus in the abundance of the Pax Augusta (Vollenweider 1979, 382-4, Nos 428-29, pl 116). The clasped hands add a symbol of good faith and harmony (ibid, 411-13, Nos 466-68, pl 121). Also note, for hands clasping cornucopiae, Maaskart-Kleibrink, 1978, 205, pl 87, no 464, and cornucopiae with a lighted altar Zwierlein-Diehl 1979, 83 pl 51. A third, exact parallel occurs at Wroxeter (G Webster pers comm).

Of these two pieces he goes on to note;-

We cannot ascribe to personal possessions like signet gems the same certainty of ownership that we must acknowledge for a military buckle, but it is most probable that two of these intaglios belonged to soldiers who owed especial veneration to his unit's standards, to Jupiter, and more generally, to Rome. Indeed, eagles, sometimes with standards are often shown on signet-gems found on fort sites.

A 1st century date is thus appropriate for each though they were found in a Phase 3 and a Phase 3 to 4 context in Trench 13 respectively. Finger rings with intaglios could, of course, remain in use for years after they were made and it seems likely that they will have arrived on the site after the change of use given the possible military overtones of the devices and the fact that in 1080 the ring itself is of iron. This shows that the owner was adhering to the sumptuary laws which forbade gold rings to people below Senatorial or Equestrian level (Henig 1978, 47). Such attention to the legal niceties would seem more appropriate to someone from a Roman rather than a native background.

Two other simple expanded rings were found unstratified. 2700 has a glass intaglio which shows a poorly preserved standing figure, and 1452 has simple moulded decoration. It is not possible to place either more closely within the 1st to 3rd century period. A fifth might be represented by 2263 from a Phase 2 context in Trench 13, but the piece is much corroded and identification is not secure.

The other rings were made in the later Roman period. There are two examples (2792, 2882) of one of the typical ring forms of the 3rd century, those with triangular shoulders (Cool 1983, 253 Group XIV). These were found unstratified on Trench 19 and from a Phase 4 context on Trench 29. There is also two unstratified examples of rings with scalloped shoulders (1072, 1914). This type (Cool 1983, 259 Group XVI) is not closely dated within the 3rd to 4th century period, but both of these rings have moulded glass intaglios which were a 3rd century form (Henig 1978, 164). 2421 is an example of a key ring. These small lever lock keys (Cool 1983, 245 Group XI) were used for small caskets (see for example Crummy 1983, 83 fig. 90) during the 3rd to 4th centuries.

In the 4th century there was a tradition of light decorative trinket rings. At Claydon Pike all the identified examples were found unstratified, There are two examples (2645 and 654) of the pronged bezel type (Cool 1983, 266 Group XIX) which appears to have been most popular in the early to mid 4th century and to have been an insular variety. The same form was used as an earring (Allason-Jones 1989, 3 Type 2b) but both of these examples are annular rings more likely to have been worn in the finger. The expanded ring with outward loops (2684) is another 4th century decorative type (Cool 1983, 224 Group IB). 2627 is very corroded but appears to be an example of the form which is ribbed all around the hoop (Cool 1983, 268 Group XXA) which probably came into use in the 3rd century.

Finally amongst the metal the silver ring 410 may be noted. It comes from a context attributed to a Phase 3 context but appears to be of relatively modern date.

There is also a fragment of a jet finger ring (1729). This was found in a context assigned to Phase 2C in Trench 13. It should be noted that most jet jewellery belongs to the late Roman period and this would be an unusually early date. A jet finger ring similar to this is known, however, from a 2nd century context at York (Allason-Jones 1996, 37 no. 170).

Top of Page

Necklaces, beads and pendants (Figs 3.4.4a: Phase 2 beads, 3.4.4b: Phase 3 beads and 3.4.4c: Phase 4 beads)

Necklaces with small glass beads are represented by two necklace fasteners (747, 2766) a wire element (32), and 27 glass beads, one of which (1840) appears to retain a fragment of copper alloy wire. On the whole small beads start to appear in Romano-British contexts from the later 2nd century but are much more numerous in later 3rd and 4th century ones, and it is clear that necklace wearing only became a common fashion in the later Roman period. Good groups of beads from contexts of the second half of the 2nd century are known from the Caerleon Baths (Brewer 1986) and the vicus at Castleford (Cool and Price 1998, 181-9). Examples of beads from Claydon Pike which can be paralleled in these groups, and so which could date from the later 2nd century onwards include the short blue biconical beads (676, 910, 1995, 2413, 2964), the cubic blue beads (1840, 2056), the small ovoid beads mainly in blue and blue/green (170, 553, 1758, 2133, 2356, 2578, 2963), and the long rectangular blue/green bead 1803. It is noticeable that segmented beads such 1678 and 2793 are missing from these assemblages suggesting they came into use later.

There are also a group of beads that appear to belong to the later 4th or 5th century. Four opaque green disc cylindrical beads were found in Trench 13. One came from a Phase 3 context (634) and three from Phase 4 ones (972, 1409, and 2586). This was a late 4th to 5th century form (Cool 2000, 50). It is possible that a similar date may be attributed to two other disc cylindrical beads in unusual colours, 15 in opaque blue/green and 47 which appears black, as it has been noted that unusual colour / shape combinations seem to be a feature of this period (op cit.). ‘Black’ beads too are rare in Roman contexts and more likely to occur in 5th century or later ones. Other unusual colour /shape combinations which might be candidates for a late date are the peacock globular bead (2483), the turquoise long cylindrical bead (2524), and the red striped green bead (732) though it may be noted that all of these three came from contexts attributed to Phase 3 in Trench 13.

There was also one jet bead (1843) that is most likely to have come from a late Roman necklace given its length though shorter cylindrical beads are known to have been used in bracelets as well (Barber and Bowsher 2000, 148 B168.4).

In addition to the beads used on the late Roman necklaces, there are also several glass beads which belong to the earlier tradition of large annular beads. 855 from a Phase 2 context in Trench 13 is an example of Guido's (1978) Class 9b which may have came into use during the first century BC and certainly died out soon after the end of the first century AD. Price (in Manning et al 1995, 105), however, has pointed out that most are found on early Roman sites so a pre-conquest origin is not certain. The occurrence of 855 at the site also adds to the concentration of Class 9 beads which is noticeable in the Somerset and Bristol Channel region which may represent a factory or entrepot in the area (Guido 1978, 77). Beads of Class 9b have cables of various colours marvered into a deep blue ground. On 855 the cable is of translucent deep blue and opaque white glass, a colour combination which also occurs on Class 9b beads from Silchester and Glastonbury (ibid 1978, 186 -87), and from Usk (Manning et al 1995, 107 no. 2).

The unstratified 48, though made after the Roman conquest, is in the pre-Roman tradition of bead making as it is a large, polychrome, annular bead. It is made from re-used Roman glass, almost certainly derived from marbled pillar moulded bowls. Pieces from two different vessels have been used to form the bead. One had a translucent purple ground with opaque white marbling and opaque yellow chips, the other had a translucent green/blue (peacock) ground with opaque white marbling. The use of these two colour combinations is interesting as both are rare amongst the polychrome marbled pillar moulded bowls known from Britain. The combination of purple and white is not uncommon, but the combination of purple, white and opaque yellow is. The use of peacock green glass as a ground colour for pillar moulded bowls is also most unusual and the only other occurrence appears to be a small fragment with white marbling from Chichester (Price and Cool 1989, 137 no. CM 3). Polychrome pillar moulded bowls are not uncommon on mid first century sites but were going out of use at that time. A mid 1st century would thus be most likely for the manufacture of this bead from considerations of both the availability of the raw materials and the tradition of bead making which it is part of.

Curiously fragments of polychrome vessel glass appeared to hold little attraction for the 1st century bead manufacturers. So it is interesting to note that two other beads using this raw material were found at Usk (Manning et al. 1995, 108 nos. 5 and 6), while a fourth was found at Birdoswald (unpublished), a fort which increasingly appears to have strong links with the Legio II Augusta in its early days. This too therefore, could be another south-western bead type.

There were three examples of plain annular beads (Guido 1978, Group 6), which are a very long-lived type. Deep blue ones such as 2130 (Group 6iva) are found in Britain in contexts ranging in date from the 5th or 6th centuries BC to at least the 8th century AD. Yellow/brown ones such as 2135 (Group 6iiia) appear in the 1st century BC and occur occasionally throughout the Roman period. Blue/green ones such as 2494 (Group 6iia), by contrast, appear to be a purely Roman form. All were found in Trench 13, 2135 in a Phase 2/3 context and the others in ones of Phase 3.

In addition to these beads made in an early native tradition, there is one example of the ubiquitous Roman frit melon bead (2128) in use during the 1st century and into the 2nd. It was found in a Phase 3 pit in Trench 13.

Two items that might have been worn as pendants may also be noted. A perforated oyster shell (5870) was found in a Phase 3/4 pit in Trench 19 and there is also a small unstratified perforated copper alloy block (299). The perforation has been worn through from long wear but it cannot be independently dated and so could be post Roman.

Top of Page

Hair pins and dress pins (Figs 3.4.5a: Hairpins, 3.4.5b: Hairpins and 3.4.5c: Dress pin, earrings, hobnails and shoe cleats)

The Roman hair pins which can be assigned to particular types have been summarized in Table 8 according to the typologies of Cool 1991 for the metal pins and Crummy 1979 for the bone ones. It may be noted that metal hair pins make up the majority of the 1st to 2nd century forms whereas the late Roman pins are predominantly of bone. As metal pins were presumably more expensive than bone ones this might hint that during the later Roman period, the women at Claydon Pike might have had less resources available to spend on their jewellery. Only a single example of a hair pin was found in a context assigned to Phase 2 (876) but as this is an example of a Crummy Type 5 which is a form more normally associated with the late Roman period, the evidence it provides for hair-pin wearing and the adoption of new hairstyles during Phase 2 is suspect. By contrast 9 of the pins are from Phase 3 contexts suggesting the adoption of the new hairstyles implied by the pins was a feature of that period. This might just reflect the original population adopting them, or it could hint at the influx of new people. It the reorganisation of the landscape in Phase 3 does reflect an official interest in the site, then one might expect the officers or officials who ran it to have been accompanied by wives and daughters more likely to have adopted current fashion.

Bone hair-pins tend not to have regional styles but metal ones of the 1st to 2nd century are strongly regional and this is reflected here by the predominance of Cool Group 13 and the single example of a Cool Group 23B, both of which are amongst the most popular local types in Gloucestershire and the south-west.

As well as the hair-pins summarised in Table 8, there is a single unstratified shank fragment from a jet pin (666) which would have been in use during the late Roman period. A rough-out for a bone pin was found in a phase 3 /4 context in Trench 19 (2187) indicating local production. There also 53 bone shank fragments which have been included in the miscellaneous section. The may have originated from needles or spoons, but a high proportion are likely to have come from hairpins. Finally 903 from a Phase 2 / 3 context in Trench 13 may be noted. this might have been a hair pin but an equally likely identification is that is from an olivary probe made in bone.

In addition to the hair pins discussed so far there are two wrythen-headed pins in the assemblage. One (3034), in lead, was from a Phase 3 context in Trench 29, the other in copper alloy (2546) was unstratified. Despite the context of 3034 it does not seem likely that either are Roman as wrythen decoration on heads tends to be very rare on Roman hair pins. Instead they belong to a well-known late Saxon type of dress pin, see for example those from North Elmham (Wade-Martins 1980, fig. 264 no. 40) Caister by Yarmouth (Darling and Gurney 1993, 80 no. 128, fig.46).

Dress pins of Roman date are rare but one may be represented by 2127, an unstratified find from Trench 29. The identification is tentative because it is very corroded but this may be a celtic dress pin of the type discussed by Boon (1975). The dating for these is uncertain. It is possible they date as early as the 2nd century as argued by Boon but many of the heads have similarities to the terminals of Fowler Type E and F (Fowler 1964, 99 fig. 1) and so a 4th century or later date might be appropriate. They are not a local type having a distribution in Ireland and the north but have also been found in south Wales.

Top of Page

Ear rings (Fig. 3.4.5c: Dress pin, earrings, hobnails and shoe cleats)

One example (1205) of a 4th century ear-ring of Allason-Jones (1989) Type 2D was found in a Phase 4 context in Trench 13, and another (2540) of more general Roman date (Allason-jones (1989) Type 1 was found in a Phase 3 context in Trench 17. 688 from Trench 13 may also have been an ear ring of the latter type.

Hobnails and Shoe cleats (Fig. 3.4.5c: Dress pin, earrings, hobnails and shoe cleats)

As can be seen from Table 9 hobnails and shoe cleats were common finds in most of the Trenches. An interesting distribution through time can be noted by inspecting Table 2. There it can be seen that no items of iron shoe furniture were found in contexts associated with Phase 2. They only start to appear in contexts of Phase 3. It is possible to test whether this pattern arose by chance by comparing them to the distribution of the brooches as is done in Table 10. Brooches have been chosen for this comparison because they are another item worn about the body during Phase 2 and Phase 3 and logically both should have a similar distribution if both were in use contemporaneously. Table 10 is the sort of contingency table to which formal statistical test can be applied. A Fishers’ Exact test (Baxter 2003, 129) applied to this table produces a p value of 0.004 which in statistical terms is highly significant and indicates it is very unlikely that this pattern could have come about by chance. It seems very likely therefore that the inhabitants of Claydon Pike did not adopt nailed leather shoes until Phase 3. Vegetable tanned leather was one of the great introductions of the Roman period (see van Driel-Murray 2001, 185). Reliable waterproof shoes were a great advance on what had been available before, but as van-Driel-Murray has pointed out (ibid. 186) the adoption of these is ‘not simply the adoption of something novel to fulfil old functions’. There were a great range of Roman shoe types to choose from and she argues that the adoption implies the adoption of ‘a new way of using clothing in social communication’. Not all Roman shoes were nailed, and it is possible that the people who lived in Claydon Pike in Phase 2 adopted Roman foot ware but of the un-nailed variety. It seems more probable, however, that they didn’t adopt the new styles. The appearance of nailed shoes on the site in Phase 3, therefore, probably indicates a change in lifestyle as momentous as the re-organisation of the landscape.

Top of Page

Buckles and belt plates (Fig. 3.4.19: Late Saxon and Medieval)

One buckle and belt plate (122) of 13th to 15th century date (see Egan and Pritchard 2002, 31) was recovered from Trench 13. The belt plate 653 from the same trench may be contemporary. This simple form is difficult to date but is most easily paralleled amongst medieval material of 13th and 14th century date (see for example Egan and Pritchard 2002, 94 no. 422) as is the unstratified 389.

Top of Page

Toilet and Medical equipment (Fig. 3.4.6: Toilet and Medical equipment)

The items in this category can be divided into the small items suitable for personal use, and long-handled implements that would have had a role both in personal care and in medicine. No item in this category was found in Phase 2 stratified contexts.

Small implements

Most of these were found unstratified, and where their origin is known they come from Trench 13. A complete toilet set (555) was recovered from a Phase 3 context in Phase 13. It consists of the normal trio of nail cleaner, spoon and tweezers. The nail cleaner belongs to what Crummy (2001, 4) has referred to the filed bead type, a south-western type in use from the mid 2nd century until possibly the 4th century. The sheet nail cleaners (101, 638, 3086) and the tweezers (65, 1610, 1625) would also have come from toilet sets. None of these are closely dated. An unusual implement from a toilet set is the rasp 2654. Such items are normally found on elaborate set with many implements (see for example Wheeler 1930, pl. xxxix; Cool and Philo 1998, 86 no 363).

There are also two bone-headed nail cleaners from unstratified or unphased contexts (158 - Trench 13, 5871 ) which belong to a south-western form which has been found in later 2nd to 4th century contexts (Crummy 2001, 4).

Long-handled implements

Five ligulae were found, three (382, 404, 1049) from Phase 3 contexts in Trench 13, and two (2352, 2526) from Trench 17 in a Phase 3 and a Phase 3/4 context respectively. A possible long-handled scoop (498) was found unstratified.

Top of Page

Textile Equipment (Fig. 3.4.7: Textile equipment)

Textile equipment was relatively rare. The only item that may have been associated with the preparation of yarn is 2868 from a Phase 3 / 4 context in Trench 17. This is an iron item and the X-radiograph shows a central bar with three teeth on either side. This is consistent with it being a woolcomb. The identification has to be tentative because woolcombs generally have a much wider central bar (cf Wild 1970, fig. 9), but the teeth do seem to be distinct and not a product of corrosion.

Several items appear to be associated with weaving. There were four perforated small bones (metapodia etc) which are thought to have functioned as bobbins and which seem generally to have been in use during the later Iron Age and earlier Roman period (see Greep 1998, 283). Three came from Trench 13 with 846 being found in Phase 2 context and 1490 and 2592 coming from those of Phase 3. There was also an unstratified example from Trench 17 (2852). The unstratified bone implement 159 may well have been a pin beater as it has the characteristic gloss associated with working textiles (MacGregor 1985, 188).

Two needles for sewing were found in Trench 13. 2365 from an unstratified context was made of copper alloy and could have been used for both textiles and leather, whereas the bone 1570 from a phase 3 context would have been more suitable for textiles. The fragmentary copper alloy 5873 also from a Phase 3 context in Trench 13 has all the features that would normally be associated with a needle but is very large for such an item. An iron needle (1936) was also found unstratified in Trench 19.

Finally two thimbles may be noted. Thimbles came into use in the 14th century (Egan 1998, 265). Open ended ones such as 97 this continue in use until the present (Biddle and Elmhirst 1990, 805) but the individually formed pits would suggest a later medieval to post medieval date for this example, and a similar date may be suggested for the fragmentary 1547.

Top of Page

Household Equipment (Figs 3.4.8a: Miscellaneous household equipment and 3.4.8b: Spoons and bowl fragment)

The household items are summarised in Table 11.

The X-radiograph for 708 from a Phase 3 context in Trench 17 is consistent with it being an iron tripod candlestick though investigation would be needed to confirm the identification. Such candlesticks were in use during the 3rd and 4th centuries (see Eckhardt 2002, 251). A second candlestick (2129) was identified during the initial analysis but I have not been able to confirm the identification.

Another item of lighting equipment is represented by the copper alloy vine leaf (360) found in a possible floor make-up in B3 in Trench 13. During the initial post excavation analysis Francis Grew drew attention to the fact that it was similar in many respects to a well known class of first century military pendants, but noted that there were differences, and other functions should be borne in mind. He suggested that another possibility was that it could have been the reflector from the back of a metal oil lamp. This seems to be the most likely identification. The sweated on terminals at the tips of the leaves would be most unusual on a military pendant but do appear to be present on the reflector of a lamp from the Thames in London (Wheeler 1930, 62, fig. 14 no. 5). Eckardt (2002, 224) relates this to Loeshcke Type XXI which would suggest a later 1st to 2nd century date which would be consistent with the context date.

Five fragments of bone veneer or inlay were recovered from Trench 13 (102, 915, 1046, 1191 and 5687). Such material is usually thought of as decorating boxes and chests and seems primarily to have been a phenomenon of the later 3rd and 4th centuries. It was also used to decorate items, probably biers, used in cremation rituals in the 3rd century (Greep forthcoming). Generally the veneer associated with domestic contexts seems to have been quite thin (c. 2mm thickness) as is the case with 915 here, whereas that associated with the cremations is often noticeably thicker. The other fragments from Claydon Pike range from 4 to 6mm in thickness which places them alongside the cremated material.

Amongst the metal vessel fragments there two handle fragments (1690, 2076) of the type that would have been used on sheet metal jugs of Eggers Form 128 (Koster 1997, 33 no. 10). Both were found unstratified in Trench 13 at some distance from each other. They do not join but they could have come from the same handle. The type was in use from the end of the 1st century into the 3rd century and was used to heat water, suggesting that somebody on the site in the 2nd or 3rd century had quite sophisticated drinking habits. Another copper alloy vessel of broadly contemporary date is suggested by the unstratified escutcheon 3013 which probably came from a small bucket (see for example den Boesterd 1956, 47 nos. 154-5). The unstratified bowl fragment 33 could also be of Roman date but is now deformed and the original shape cannot now be reconstructed. Amongst the iron there is an escutcheon likely to have come from a cauldron (1588) found in a Phase 2 / 3 context in Trench 13. The original analysis also concluded that the iron fragments 426, 770 and 2876 were also from vessels but I have been unable to confirm this identification. Lead alloy vessels are considered below under Religion.

There is one example of a round bowled spoon (2542) from a Phase 3 / 4 context in Trench 17. These were a common 1st to 2nd century form (Crummy 1983, 69 Type 1). The unstratified copper alloy spoon with a mandolin-shaped bowl from Trench 19 (2818) is also of Roman date (ibid type 3). A Roman date, however, cannot be advanced with certainty for the large oval-bowled spoon 1842, also found unstratified. A fragment of a lead ladle may also have been a household utensil, but again it cannot be dated independently and it was found unstratified.

Possible iron fittings from buckets (178, 1583) and a barrel (802) were also noted in the earlier analysis.

Top of Page

Recreational Items (Fig. 3.4.9: Recreational Items)

Items in this category are curiously few in number (see Table 1), and all were recovered from Trench 13. Two counters are of standard and common forms. The glass counter (911) is of a type that was commonest in the 1st to mid 2nd centuries (Cool et al 1995, 1555), whilst one of the bone counters (198) was in use throughout the Roman period (Greep 1986, 202 Type 3). the other bone ‘counter’ (1411) is a curious piece. It is decorated on both the faces and the sides and was clearly intended to be seen in the round, and so has been termed a counter. Its’ Roman date seems certain because it was found in a Phase 3 context but I have not been able to locate a similar piece in either British or continental assemblages.

Two dice with the typical Roman arrangement of dots where opposite faces sum to 7 were found. 434 from a Phase 3 context is of particular interest as it is made of ivory. Ivory was a luxury item in the Roman world and so hints at a degree of affluence on the site during Phase 3. The other (347) lacks the typical radial lines of ivory and is more likely to be of bone.

Top of Page

Items Associated with Weighing and Measuring (Fig. 3.4.10: Weighing and measuring)

There are four typical Roman biconical steelyard weights made of lead with a central iron suspension wire (620, 1065, 1254 and 2499) from the unstratified material. A fifth (391) was cylindrical like an example from Somerford Keynes 614. The evidence for the use of equal-armed balances is more equivocal. There are items of lead in the assemblage that might have functioned as weights but none have the typical shapes or marks denoting their weight that might be expected, and so they have been relegated to the miscellaneous category.

Of particular interest given the major re-organisation of the landscape in Phase 3 which would presumably have required a degree of survey activity, is the number of plumb bobs in the assemblage. Two were found in Trench 13, one in a Phase 3 – 4 context (2605) and the other in one of Phase 4 (462). A third was found unstratified (3012). Plumb bobs were an essential part of the groma, the principle tool used by Roman surveyors.

Top of Page


Items associated with writing (Fig. 3.4.11:Writing)

Four styli can be recognised with certainty. Two are of iron, a Manning (1985) type 1 came from a Phase 3 context in Trench 13 (1040), and a Manning (1985) Type was found unstratified (108). There were also two of copper alloy, both from unphased contexts in Trench 13. 377 is another example of the decorated type which appears to be a local variant of the 2nd century (see Somerford Keynes 1145 for discussion), while 129 is undecorated. In addition 1997 from a Phase 3 / 4 context in Trench 13 may be a third iron stylus but would need investigative conservation to confirm the identification.

Given the supposition that in Phase 3 the occupation at Claydon Pike became ‘official’ in some way, it should be noted that the presence of styli on the site cannot be used as evidence for this as the evidence from other sites is that styli are regularly found on rural sites of no great status (Cool and Baxter 2002, 375-6).

Top of Page

Items associated with Transport (Fig. 3.4.12: Transport )

The items attributed to the transport category in Table 1 provide us with some problems as four of the items are horseshoes or their nails. As Manning (in Manning et al 1995, 42) has shown Roman horseshoes do exist and horses may even have been shod in the late pre-Roman Iron Age in Britain. It cannot be said, though that they are at all common from Roman contexts. It is particularly unfortunate that it has not been possible to re-examine 621 from a Phase 3 context in Trench 13, as the original analysis identified it as a horse-shoe, but there is neither an X-radiograph nor a drawing extant to confirm this. There are also three horse-shoe nails, 5880 from a Phase 3 context in Trench 13, 5886 from a contemporary context in Trench 17 and 5875 from a Phase 4 context in Trench 19. Given the presence of horseshoes in Roman contexts, there must naturally have been horseshoe nails as well, but it has to be said that 5880 and 5886 are typical fiddle key nails which are generally thought to have been associated with the types of horse shoes in use during the 11th to 14th centuries (Clark 1995, 86-7)

The four linch pins are all spatulate-headed forms (Manning 1985, 74 type 2). Where the precise form can be identified they have turned over loops (337, 338). In the case of 90 and 311 it is not clear whether it was a turned over or peg loop, but it is likely to have been the former as these are the commonest Roman form. Three came from Trench 13 and all were unstratified. Trench 13 also produced an enamelled strap loop (789) probably of 2nd century date from a Phase 3 context and two possible snaffle bit fragments (1047; 1240) from Phase 4 contexts, though these would need investigative conservation to confirm the identification.

2299 belongs to the post Roman activity on the site. It is one of the junction loops used to fasten the straps of a bridle together and to fix the reins to the bit. It is a form that was in use in the late Saxon period (Goodall 1990, 1043 - see especially 1044 no. 3880 from a late 9th to early 10th century context at Winchester).

Top of Page

Structural items from Buildings (Fig. 3.4.13: Structural items from buildings)

The structural items are summarised in Table 12. as structural ironwork is not intrinsically dateable only those items stratified in Phase 4 or earlier contexts have been included.

As can be seen the category is dominated by iron nails. The quantification of the nails is based on the nail heads as this is the most reliable form of quantification given the propensity of the nails to disintegrate into many fragments. The nails, including those from the later and unstratified contexts, are tabulated by area in Table 13. The inhabitants of the Phase 2 settlement were clearly using nails in their buildings but to nothing like the extent that was to happen in Phase 3. This may just reflect an intensification of the occupation but seems more likely to reflect a change in building practice. It is often noted that timber clad building require considerable numbers of short nails (Manning 1985, 134). They were also vital for fixing tiles and shingles for roofs. Shirley (2001, 143 Table G), for example, has calculated that a wooden legionary fortress such as Inchtuthil would have required over one and a half million nails weighing 12 tons. The concentration within Trench 13 clearly reflects the fact that this was the main focus of building activity.

The other category of material well represented in this category is window glass. This is summarised in Table 14 according to whether it was made by the cast process which results in relatively thick glass with distinctive matt / glossy surface or by blowing which results in thinner glass that is more transparent. Cast glass was the dominant type of the 1st to 3rd centuries with blown glass only appearing in quantity in the 4th century. This normally observed date breakdown casts some doubt on the presence of glazing during Phase 2 as Table 14 suggests as one of the fragments (1647) is typical of a 4th century blown fragment. It was found in a ditch fill and possible contamination might be suspected. The other fragment from a Phase 2 context (759) is cast and so could have been contemporary with the occupation, but it might be questioned whether one fragment is sufficient to posit glazing in Phase 2 especially as it was found in a ditch fill associated with E16 where it is known that the tops of the ditches associated with this enclosure remained open well into Phase 3. On balance, therefore, glazing seems most likely to have first occurred on the site during Phase 3. As can be seen from Table 15 it was concentrated within Trench 13. The quantity of 4th century blown glass present would suggest that the cottage style villa of Phase 4 also had some glazed windows.

Top of Page

Tools and Evidence for Craft and Industry (Figs 3.4.14a: Punch, chisels, awl and saw, 3.4.14b: Coopers crozes and knives , 3.4.14c: Knives and handles , 3.4.14d: Miscellaneous tools and industrial debris )

It has to be noted at the outset that the problems associated with the study of the iron described in the introduction disproportionately affect this category, as does the lack of investigative conservation on the material. The difference between whether an item is a set (a smiths tool) or a chisel (a carpenters tool) is often the fact that the smiths’ tool will show a degree of battering because it will have repeatedly been struck by a metal hammer. Such evidence is generally more apparent after conservation. For convenience knives have been included in this section though it should be noted that though some are likely to have been craft tools, others would have been used in the kitchen and as small items of personal equipment.

Table 16 summarises the tools according to craft and phase. The only smiths tool identified with certainty is the punch 2178 from a Phase 3 context in Trench 17, an appropriate location given the amount of smithing and smelting debris recovered from the area. As noted the chisel or set 2252 from a Phase 3 context in Trench 13 could be either a smiths tool or have been used by a carpenter. Woodworking activities are definitely represented by the paring chisel 704 and the chisel of less certain form 2885, both from Phase 3 contexts in Trench 17. In the original analysis 485, 1830, 2022, 2160 and 2331 were also identified as chisels but I have been unable to confirm these identifications. A remarkable find is the virtually complete blade from a one-handled saw with a triangular blade (157). This has been bent through 45o presumably deliberately. The discovery of a complete item like this is unusual and deserves comment. Generally casual loss is not an appropriate explanation in these circumstances. A similarly complete one from Verulamium (Manning 1972, 166 no. 12), for example, was found because a building had been burnt down. 157 was found in a shallow gully not closely dated within Phases 2 and 3 in Trench 13. Gullies and ditches are often places where deliberately structured deposition can be identified, often formally marking the end of use of the feature. There is the possibility that the deposition of this item marks something similar with the folding perhaps ritually killing the artefact.

Two very distinctive iron tools were recovered from Trench 29, one unstratified (2084) and one from a Phase 3 / 4 context (2084). These were identified in the original analysis as cooper’s crozes which are a type of specialised plane used to make the groove inside the barrel staves so that the barrel head would fit (Blandford 1974, 102). I have not been able to confirm this identification.

The working of leather is indicated by six awls. One (320) is of bone; the others are made of iron. 258 is an example of a Manning Type 3, 366 and 2877 are Manning Type 4a, and Type 4b is represented by 79 and 2756 (Manning 1985, 39-41). The only stratified examples are 366 and 377 from Trench 13 in a Phase 3 and Phase 3 / 4 context respectively.

It has been possible to assign some of the knives to a range or well-known Roman types (Manning 1985, 108-19). Type 10 is represented by 1761, Type 11 by 549 and 844 with the latter having a loop handle, Type 19 by 128, and Type 24 by 593. The commonest form is Type 21 (587, 710, 1020), a late Roman type. There is also a fragment from a knife with a perforated tang (2848) which could have been from a Type 1, 5 or 7. 456 and 2968 are fragments from knives with triangular blades. Two cleavers can also be identified with certainty, 435 from one of Manning Type 2A and 1868 which is a Manning Type 3 (Manning 1985, 120-2). The cleavers are both from Trench 13 and would have been associated with butchery. Their presence is also indicative of a ‘Roman’ rather than native approach to life as the native butchery practices involved the use of knives and cleaver butchery tends to be more associated with urban and military sites (Maltby 1985, 20).

The final knife that can be identified with certainty is a small folding knife with a copper alloy handle depicting a dog chasing a hare (1605). This is a widespread form in the Roman world (see Lloyd-Morgan 1995, 1035 no. 452) but not often found in well-dated contexts which would allow a closer dating. Many, like this one, have been found unstratified. Folding knives like this remind the modern viewer of pen-knives and the intriguing possibility has recently opened up that this may have been one of their functions in the Roman period as well. Feugère (2003) has recently drawn attention to the fact that some knife forms normally thought of as razors (e.g. Manning 1985, 108 Type 1) are regularly found in association with writing equipment, and were probably used to sharpen the nibs used for writing on ink tablets. A small pocket knife like this would have been ideal for such a purpose.

In addition 21, 91, 790, 973, 1385, 1488, 2046, 2145, 2539, 2557 and 2877 were originally identified as knives or blade fragments, but the identification cannot currently be confirmed. This is especially unfortunate in the case of 21 as it is one of the very few stratified small finds from the Phase 1 middle Iron Age occupation.

The tools and knives are summarised according to Trench in Table 17 which suggests that whereas carpentry was widespread, leather working was concentrated in Trench 13.

In addition to the tools there are also some debris indicative of craft working (see Table 1). The bone pin rough out 2187 has already been noted. Another rod-shaped rough out came from Trench 13 in a Phase 4 context (5690). There are also fragments of sawn antler fragments (2872, 5693) and two long bone fragments with saw marks that are the result of deliberate working (380, 419).

There is also evidence for both the hot and cold working of copper alloys. Unfortunately all casting waste (980, 1261, 1664 and 2762) was unstratified and unlocated, but off-cuts of sheet were found in Trench 13 (615, 1581) and Trench 19 (1897) as well as unstratified (1909). The two blue glass tesserae from Trench 13 (954 – Phase 3) and Trench 17 (2337 – Phase 4) are probably also indicative of high temperature craft activity, either enamelling or bead-making. Though such tesserae could be used in mosaics, in Roman Britain they seem to occur far more frequently in contexts that suggest they were small pieces of raw glass (Price in Manning et al 1995, 307; Cool and Price 1998, 193).

Top of Page

Fasteners and Fittings (Figs 3.4.15a:Phase 2 fasteners and fittings, 3.4.15b: Phase3 fasteners and fittings, 3.4.15c: Phase 4 fasteners and fittings , 3.4.15d: Undated fasteners and fittings , 3.4.15e: Pottery repairs (clamps) , 3.4.15f: Pottery repairs (Plugs) , 3.4.15g: Security phase 3, 3.4.15h: Security phase 4 and 3.4.15i: Security unphased )

The fasteners and fittings are summarised in Table 18. Amongst the stud and mount category there are several that would have been used to decorate boxes and chests (Figs K1 - 4). There is, for example, an example of a lion-headed stud (661) from a Phase 4 context in Trench 13. This was residual as the form was in use during the later 1st century Flavian period (Borrill 1981, 315; Cool and Philo 1998, 104 nos. 513 and 527). Composite studs and bosses made of copper alloy, lead alloy and iron such as 63, 301, 1161, 1363, 1587, 1607, 1790, 2169, 2496, 2737, 2772, 2796, and 2894 were also frequently used to decorate boxes throughout the Roman period as can be seen see for example on a casket used in a mid to late 2nd century burial at Skeleton Green (Borrill 1981, 305), and on one from a 4th century burial at Colchester (Crummy 1983, 85, nos 2179-82). Here those that were stratified generally came from Phase 4 contexts. Only 2496 came from a Phase 3 context in Trench 17, whereas 1587, 2796 and 2894 came from Phase 4 contexts in Trench 13 and 2169 came from a contemporary context in Trench 19.

A major component of this category consists of lead pottery repairs (Figs K5 - 6) They have been summarised in Table 19 according to the various categories identified in the Somerford Keynes report. Rivetting obviously started during Phase 2 but the relative rates during the period of occupation cannot really be ascertained as so much of the assemblage is unstratified or from unphased contexts. Nor is it possible to explore any inter-site variability because over three quarters of the pieces are unlocated as to Trench. In all cases where pottery fragments are retained they belong to coarse pottery vessels. The pottery report also notes two mortaria that are riveted but there is no indication of whether any samian pottery was also repaired. The concentration on riveting coarse pottery is similar to that seen at Somerford Keynes but here the rate of riveting appears to be lower judged against the extant pottery assemblage which has been estimated to be c. 100,000 sherds. A rate of 0.1% is indicated which is in line with the 0.05 to 0.2% recorded by Evans (in Booth et al 2001, 382) for a variety of Lowland sites and draws attention again to the anomalous Somerford Keynes rate.

The assemblage is rich in fittings associated with security (Figs K7 - K9). Latch lifters can be regarded as practical items not particularly indicating a concern with locking things securely but rather exhibiting an anxiety for closing doors. There are four in this assemblage, 2960 and 2962 from Phase 3 contexts in Trench 17, one (1146) from a Phase 3 context in Trench 13 and one (794) unstratified. Tumbler locks (see Manning 1985, 89-93) do suggest a definite aim of locking things. There are seven lift keys (112, 541, 898, 956 Trench 13; 1889, 2752 Trench 19; 1557 unstratified), two slide keys (827 – Trench 13, 50 - unstratified) and a lock bolt (495 - unstratified) from this sort of lock. Lever locks and padlocks provided the highest level of security in the Roman period. There is one lever lock key (956) from a Phase 4 context in Trench 13 and four padlock fittings have been identified with certainty, one (754) from a Phase 3 context in Trench 17 and three from Trench 13 (72 – U/S, 627, 854 Phase 3). Four other padlock fittings were identified in the original analysis (22, 255, 451 and 779) but I have not been able to confirm the identification and they will not be considered further here.

If these fittings are divided between the low level security latch lifter and the higher security tumbler and lever locks and padlocks an interesting pattern emerges (see Table 20). The table suggests that most of the things that needed to be secured were in Trench 13 and that there was a higher concern for security in Phase 4 than in Phase 3.

Of the other fittings the most noteworthy is the fragment 2775 (Fig. K1) which has an infilled basket work pattern. This decoration is very similar to that on celtic mirrors (see for example Fox 1958, 95, fig 60) but the original diameter indicates it cannot have come from such an item. This decorative pattern is very rare on post conquest metalwork and so a late pre-Roman Iron Age date would be appropriate.

Top of Page

Agricultural Items (Fig. 3.4.16: Agriculture)

A small number of agricultural items can be recognised. These included two ox goads, one unstratified in Trench 13 (341) and the second from a Phase 3 context in Trench 17 (2555). There was also a pruning hook from a Phase 4 context in Trench 13 (585), and a another small hook that could also have been used in pruning from one of Phase 3 in Trench 17 (2495). A blade fragment from a contemporary context (696) might be from a scythe. There appears to be a ridge along the back and this is a feature of scythes (Manning 1985, 49). Finally a spud (509) from an unstratified context may be noted.

Top of Page

Military Equipment (Fig. 3.4.17: Military)

The assemblage contains a small quantity of items that suggest an intermittent military presence on the site, strongest in the later 2nd to 3rd century and concentrated in Trench 13.

There is one piece of 1st century cavalry equipment (124) found unstratified in Trench 13 about which Mr Francis Grew writes:-

This is a fragment of a cavalry harness clip for attachment to a junction ring. The lower end (as illustrated) is broken and would originally have had two pairs of perforations through which the object would have been rivetted to the end of a harness strap. The upper end could be clipped to a ring. More complete examples are known from Wroxeter and Gloucester (Webster 1960, nos. 247 and 101). This is an Early Roman types and does not seem to have been in use much beyond the end of the first century

Later 2nd to 3rd century material is more numerous. Mr Francis Grew contributes the following note on the 5493.

This is the pendant half of a hinged decorative baldric terminal as identified and illustrated by Oldenstein (1977 abb 11.12, taf 83). Such objects have been commonly found in the Rhine and Danube forts (including Zugmantel and Saalburg) though examples are also known from British sites such as Silchester (Boon 1974, fig.8.4). The Claydon Pike piece was unstratified but they are generally thought to belong to the second or third centuries.

Other military items of this date include a vulva mount (912) from a Phase 4 context in Trench 13 and one certain (1674) and one possible (1918) caterpillar mounts, both of which are unstratified. These two forms were also found at Somerford Keynes and have been discussed there. There is a strap end (1506) from a Phase 4 context in Trench 13 which can be compared to one from the Saalburg (Oldenstein 1977, 144, 249 no. 297), and a pelta mount (1937) from a Phase 3 / 4 context in Trench 13 which seems most likely to be related to the simple pelta-shaped mounts often found in later 2nd and 3rd century military assemblages (ibid, 178). There are also two unstratified barrel beads (2630, 2699). As discussed in the Somerford Keynes report the status of these is open to question, but they do appear to occur more often on military sites than civilian ones. If soldiers were present on the site in the later 2nd to 3rd century, this would provide a context for the one certain (146) and one possible (418) spear heads found in Phase 3 contexts in Trench 13.

A later item of possible military equipment is the buckle fragment 450 also from a Phase 3 context in Trench 13. This would appear to be an example of a Simpson (1976) Group I buckle. Buckles of this sort are rarely found in Britain (see for example Swift 2000, fig 234) though an example was recovered from the Lankhills School cemetery in Winchester (Clarke 1979, 270, fig. 34 no 70). If this identification is correct the dating is in conflict with the phase of the context it has been found in as they are of later 4th century date. As has been discussed in the Somerford Keynes report, there is some debate as to whether these late 4th century belt fittings can be regarded as providing good evidence of the presence of soldiers, or whether in this area they are merely a civilian fashion.

Top of Page

Religious items (Fig. 3.4.18: Religion)

A small number of items may have had a religious or ritual significance. A single miniature axe 2814 was located in an unstratified context in the area of the shrine (Trench 27) and is most likely associated with the use of this structure. A fragment from a lead vessel (388) was found unstratified in Trench 13, while the location of another (3045) is completely unknown. The religious associations of items like these have been discussed in the Somerford Keynes report. The best preserved of the two lead vessels (388) was a trullum similar to one found in the sacred spring at Bath (Sunter and Brown 1988, 20 no. 28, fig. 10). That one had dolphins on the handle, and the mouldings on the face of this handle might have been similar but are too obscured for the decoration to be identified. A third lead vessel fragment may also be represented by 1938 from a Phase 3/4 context in Trench 13.

The small bell 107 is of particular interest. It was found in a post-hole of B3 in Trench 13. It should be noted that it is complete with the iron clapper still in place. It could of course have got into the post-hole by accident, but the possibility that this was a deliberate foundation deposit is a distinct possibility. Bells were used as musical instruments and probably also as cowbells but it should not be forgotten that in antiquity they also served an important role in scaring away evil spirits (Johns 1982, 67-70). A bell has been found apparently deliberately placed below the threshold of a building in the small Roman town of Scole (Seeley 1995) and it is tempting to see this as playing a protective role. Could this bell have been part of similar rituals for B 3?

Top of Page

Miscellaneous items (Fig. 3.4.19: Late Saxon and Medieval)

In the archive there are 466 items which fall into the miscellaneous category either because they are fragments of wire, sheet metal etc or because they might have a number of different functions, metal rings are a good example of these. Table 21 summarises all of those that are stratified in Roman or earlier contexts (approximately 70%) of the total together with a few items of intrinsic interest. Leads whorls have been included in the table as though none were found in stratified contexts the possibility that some were Roman is likely as examples were found stratified in early Roman contexts at Somerford Keynes. Attention may also be drawn to the large number of shank fragments, 52 of which were made of bone with the rest being of copper alloy. These could have come from hair pins, needles, long-handled toilet implements or been the handles of spoons. In discussing the hair pins, it was suggested that new hairstyles were a feature of Phase 3. Two bone shank fragments were found in Phase 2 contexts (924, 1854) which might hint that hair pins were present during that time but given the range of items these shanks could come from, that suggestion has to remain unproven.

Amongst the material summarised as ‘object’ there is a bone toggle 5692 from an unstratified context. These were in use during the late Iron Age and early Roman period. Their function is unknown (Greep 1998, 283). Another item of worked bone 5694 from a Phase 3 / 4 context in Trench 13 might have come from either a hinge or a handle. A broken copper alloy spike (664) was recovered from a Phase 3 context in Trench 17. The final ‘object’ (2011) is an iron item from a Phase 4 ditch fill was identified as a caltrop in the original analysis. These were multi-pointed spikes used against cavalry by the Roman army. I have not been able to inspect this piece and it would need further investigation before such identification could be proven. From the X-ray the identification is unconvincing.

Top of Page

Overview

Having considered the finds from a functional point of view, it is now possible to explore what they are telling us about the site. This part of the report will be structured by site period. As can be seen from Table 22 more than half the finds came from Trench 13 and it is about the activity in this trench and a lesser extent Trenches 17 and 19 that the finds have most to say.

The middle Iron Age settlement

The material associated with this phase of the site is negligible as is to be expected and the small finds can cast little light on the nature of the occupation. Only five items were found stratified and no items belonging to this period were identified typologically. Of the stratified items, two were structural iron finds 24 and 3966, the latter being an iron nail and the possibility that this was intrusive must be strong. The other items consisted of featureless fragments of copper alloy (20) and iron (296) and a possible iron blade fragment (21).

The late Iron Age and early Roman activity

Personal ornaments can be a very good indicator of how people present themselves to the world. Those which typologically belong to this phase suggest quite a conservative rural society . The brooches are on the whole typical of what is to be expected in a native population in this region as many are local forms. There are occasional examples of brooches from more distant sources such as the Kragenfibeln 1045 and the Birdlip brooch 1279 but these are in a distinct minority. Following the Conquest the community was happy to adopt certain new types but only where those types filled pre-existing needs. Thus Hod Hill brooches were adopted but the women did not adopt the new hairstyles nor did the community apparently embrace the new styles of footwear. On the whole the material culture that can be associated with this phase of activity is relatively modest as can be seen both in Table 1 and Table 23 where the material from Trench 13 from stratified contexts excluding the building material and miscellaneous items is summarised. Apart from personal ornaments the only functional categories represented are fasteners and textile working. It is possible that the population were using more objects than these tables would suggest as it should not be forgotten that 15% of the objects considered in this report were unstratified. There is nothing in that material, however, that considered from a typological point of view must belong to this phase of activity.

In Table 24 those stratified items which can have typological spot dates attributed to them are summarised according to site Phase. As is to be expected Trench 13 has the largest number of 1st century items but the figures for Trench 17 should also be noted. They suggest activity here prior to the Phase 3 re-organisation of the landscape.

Top of Page

The Roman complex

It is clear from the finds that it was not just the landscape that changed at Claydon Pike in Phase 3. There are major changes in the finds record as well. As can be seen from Table 24, a very wide range of activities appear in this phase and even in the categories present in Phase 2 there are new departures. The personal ornaments suggest women started wearing their hair in different styles and as a whole the population started wearing Romanised shoes. As has been noted above building techniques changed. Windows were glazed, and the buildings were probably timber clad. Doors were designed to be closed. Inside the buildings the furnishings changed. There was for example at least one exotic copper alloy oil lamp (360) and new ways of preparing drink are suggested by the handle fragments 1690 and 2076. The iron cleavers may indicate changes in butchery practise.

The finds are telling us of very deep-seated lifestyle changes. It is almost as if the population adopted the Roman cultural package wholesale. The question needs to be asked whether this was affecting the same population who had lived on the site in Phase 2 or was this the result of new people with different customs moving in. Would a woman who had spent her girlhood on the site have lived to see her grand-daughter adopting new fashions, or would she have looked on from the side-lines whilst ‘foreigners’ lived there in alien buildings with outlandish lifestyles?

In as far as can be seen if the Phase 3 population was new, it was certainly not ‘foreign’. The metal hairpins and the brooches they were using are still local forms. If the re-organisation was ‘official’ then it has left no trace in the finds record. There is, for example, no Hadrianic military equipment. Instead there is a noticeable presence of later Antonine / Severan equipment. Quite what the status of this is, though, is open to question as similar material was found as Somerford Keynes. As discussed there, this might just be part of a pattern of dispersed deployment where soldiers were involved in policing duty and not be indicative of official involvement or ownership of the Claydon Pike estate. Against the hypothesis of continuity of the population the finger ring 1080 may be considered. It is likely that this was an heirloom by the time it was lost, but it seems unlikely that the 1st century inhabitants at Claydon Pike would have observed the sumptuary laws as the original owner must have. There are also more hints of luxury in the assemblage in Phase 3 than is suggested by the Phase 2 finds. The copper alloy lamp, the ivory dice, the use of metal rather than bone hairpins all hint at a degree of affluence.

In Table 25 the types of objects stratified in Phase 3 contexts are summarised. As can be seen the widest range of functional categories is seen in Trench 13. It has a more domestic flavour than the assemblage in Trench 17. In the tools category, for example, most of the material from Trench 13 consists of knives and blade fragments whereas in Trench 17 the category is dominated by carpentry tools. Trench 17 also has agricultural implements whereas the category is missing in Trench 13. This together with the smithing activity suggests this was much more a service area than a domestic one.

Top of Page

The late Roman villa

As may be noted from later 3rd and 4th century material is quite widely distributed across the excavated area even in areas such as Trench 17 where it is not believed there was any formal Phase 4 occupation. Later 4th or early 5th century activity is clearly indicated in Trenches 5, 13 and 19 but it should be noted that it takes the form of a handful of items of personal ornament and so the degree of activity or occupation is unknown. This might just reflect casual loss.

If the building material is excluded, and if allowance is made for the smaller number of contexts belonging to Phase 4, this final Roman phase is much more productive of finds. In phase 3 there are 533 contexts and a ratio of 0.6 finds per context. In Phase 4 there are 206 contexts and a ratio of 1.1 finds per context. This should be born in mind when looking at Table 26 which summarises the material stratified in Phase 4 contexts as superficially it might be thought there was a decrease in finds. Again Trench 13 provides the greatest range of finds categories, and again the character of the occupation is decidedly domestic with the tool category still dominated by knives and blade fragments. Attention has already been drawn to the fact that one interesting feature of the fasteners is the much greater concern for security they show in Phase 4 than they had in Phase 3.

Saxon and medieval activity

A small number of late Saxon and late medieval items were recovered. They consist of an item of harness and dress and belt fittings. As with the evidence provided by the very late 4th century items, their significance is questionable. They show that people passed over Claydon Pike during that time but probably cannot be taken as evidence of sustained occupation.

Top of Page

Bibliography

Allason-Jones, L., 1989. Ear-rings in Roman Britain BAR Brit.Series 20, (Oxford).

Allason-Jones, L., 1996. Roman Jet in the Yorkshire Museum (York).

Barber, B. and Bowsher, D., 2000. The Eastern Cemetery of Roman London Excavations 1983-1990 MoLAS monograph 4 (London).

Baxter, M., 2003. Statistics in Archaeology (London)

Biddle, M., 1990. Object and Economy in Medieval Winchester, Winchester Studies 7.ii, (Oxford),

Biddle, M. and Elmhirst, L., 1990. ‘Sewing Equipment’ in Biddle 804-17.

Blandford, P.W., 1974. Country Craft Tools (Newton Abbot).

den Boesterd, M, H.P. 1956, Description of the Collections in the Rijksmuseum G.M. Kam at Nijmegen V: The Bronze Vessels (Nijmegen).

Blockley, K., Blockley, M., Blockley, P., Frere, S. S. and Stow, S., 1995. Excavations in the Marlowe car park and surrounding areas, Archaeology of Canterbury V, (Canterbury

Boon, G. C., 1974, Silchester the Roman town of Calleva, (Newton Abbot & London)

Boon, G.C., 1975. ‘Two Celtic pins from Margam Beach, West Glamorgan’. Antiquaries Journal 55, 400-4.

Booth, P.M., Evans, J. and Hiller, J. 2001. Excavations in the Extramural Settlement of Roman Alchester, Oxfordshire, 1991. Oxford Archaeology Monograph 1, (Oxford).

Borrill, H., 1981. 'Casket burials' in Skeleton Green: a late iron age and Romano-British site, Britannia Monograph 2 (London), 304-21

Brewer, R. J., 1986, 'The beads and glass counters, in Zienkiewicz, J. D., The legionary fortress baths at Caerleon. Volume II: the finds, (Cardiff), 146-56.

Clark, J., 1995. The Medieval Horse and its Equipment Medieval Finds from Excavations in London 5 (London).

Clarke, G., 1979. The Roman Cemetery at Lankhills, Winchester Studies 3 Pre-Roman and Roman Winchester Part II (Oxford)

Cool, H.E.M., 1983. A study of the Roman Personal Ornaments made of metal, excluding brooches, from southern Britain (Phd Thesis University of Wales).

Cool, H. E. M., 1991. 'Roman metal hair pins from southern Britain', Archaeol J. 147 (1990), 148-82

Cool, H.E.M., 2000. ‘The parts left over; material culture in the 5th century’, in Wilmott, T. and Wilson, P. (eds) The Late Roman Transition in the North, BAR British Series 299, 47–65.

Cool, H.E.M. and Baxter, M.J., 2002. ‘Exploring Romano-British finds assemblages’, Oxford Journal of Archaeology 21, 365-80.

Cool, H. E. M., Lloyd-Morgan, G. and Hooley, A. D., 1995. Finds from the Fortress, Archaeology of York 17/10, (York).

Cool , H.E.M. and Philo, C. (eds.), 1998. Roman Castleford Excavations 1974-85. Volume I: the small finds Yorkshire Archaeology 4 (Wakefield).

Cool, H.E.M. and Price, J., 1998. 'The vessels and objects of glass' in Cool and Philo (eds.), 141-94.

Crummy, N., 1979. 'A chronology of Romano-British bone pins', Britannia 10, 157-64

Crummy, N., 1983, The Roman Small Finds from Excavations in Colchester 1971-9, Colchester Archaeol. Rep. 2, (Colchester).

Crummy, N., 2001. ‘Nail-cleaners: regionality at the clean edge of Empire’, Lucerna Roman Finds Group Newsletter 22, 2-6.

Darling, M.J. and Gurney, D., 1993. Caister-on-Sea. Excavations by Charles Green 1951-1955. East Anglian Archaeology 60 (Dereham).

van Driel-Murray, C., 2001. ‘Vindolanda and the dating of Roman Footwear’, Britannia 32, 185-97.

Eckardt, H., 2002. Illuminating Roman Britain Monographies instrumentum 23 (Montagnac)

Egan, G., 1998. The Medieval Household Daily Living c. 1150 – c. 1450 Medieval Finds from Excavations in London: 6 (London).

Egan, G. and Pritchard, F., 2002. Dress Accessories c. 1150-c. 1450 Medieval Finds from Excavations in London : 3 (Woodbridge, 2nd edition).

Feugère, M., 1985. Les Fibules en Gaule Méridionale de la conquête à la fin du Ve siècle après J.-C. Revue Archéologique de Narbonnaise Supplément 12 (Paris).

Feugère, M., 2003. ‘Penknives from Newstead: writing accessories’, Lucerna Roman Finds Group Newsletter 26, 9-12.

Fowler, E., 1964. ‘Celtic metalwork of the fifth and sixth centuries AD’. Archaeological Journal 120, 98-160.

Fox, C., 1958. Pattern and Purpose (Cardiff).

Goodall, I.H, 1990. ‘Bridle bits and associated strap fittings’ in Biddle, 1043-6.

Greep, S. J., 1986. 'The objects of worked bone', in Zienkiewicz, J. D., The legionary fortress baths at Caerleon. Volume II the finds (Cardiff), 197-212

Greep, S., 1998. ‘The bone, antler and ivory artefacts’, in Cool and Philo (eds.), 267-85.

Greep, S., forthcoming. ‘Bone and antler veneer’ in Cool, H.E.M., The Roman Cemetery at Brougham, Cumbria: Excavations 1966-67, Britannia Monograph 21.

Guido, M., 1978. The Glass Beads of the Prehistoric and Roman Periods in Britain and Ireland, Rep. Res. Comm. Soc. Antiq. Lond. 35, (London).

Hattatt, R., 1985. Iron Age and Roman Brooches, (Oxford).

Hattatt, R., 1987. Brooches of Antiquity, (Oxford)

Hattatt, R., 1989. Ancient brooches and other Artefacts, (Oxford).

Henig, M., 1978. A Ccorpus of Roman Engraved Gemstones from British sites, Br. Archaeol. Rep. 8, (2nd ed. Oxford).

Johns, S., 1982. Sex or Symbol (London).

Kilbride-Jones, H. E., 1937-38. 'Glass armlets in Britain', Proc. Soc. Antiq. Scotland. 72, 366-95.

Koster, A., 1997. The Bronze Vessels 2 Description of the Collections in the Provincial Museum G.M. Kam at Nijmegen (Nijmegan).

Lawson, A.J., 1976. ‘Shale and jet objects from Silchester’, Archaeologia 105, 241-75.

Lloyd-Morgan, G., 1995. ‘Knife handle’ in Blockley et al. , 1035-6

MacGregor, A., 1985. Bone, Antler, Ivory and Horn. The Technology of Skeletal Materials since the Roman Period, (London and Sydney)

Mackreth, D., 1985. ‘Brooches from Roman Derby’, Derbyshire Archaeological Journal 105, 281-99.

Mackreth, D.F., 1995. ‘Pre Roman and Roman brooches’ in Blockley et al , 955-82.

Mackreth, D, 1998, ‘Copper-alloy brooches (Site 1, Site 2 and field-walking)’ in Timby, J.R., Excavations at Kingscote and Wycomb, Gloucestershire, (Cirencester), 113-49.

Maltby, J.M. 1985. 'Assessing variations in Iron Age and Roman butchery practices: the need for quantification', in Fieller, N.R.J., Gilbertson, D.D. and Ralph, N.G.A. (eds.) Paleaobiological investigations: research design methods and data analysis BAR IS 266, 19-30.

Manning, W. H., 1972. 'The iron objects', in Frere, S. S., Verulamium Excavations Volume I, Rep. Res. Comm. Soc. Antiq. Lon. 27, (Oxford), 163-95

Manning, W. H., 1985. Catalogue of the Romano-British Iron Tools, Fittings and Weapons in the British Museum, (London)

Manning, W.H., Price, J. and Webster, J., 1995. Report on the Excavations at Usk 1965-1976. The Roman Small Finds (Cardiff).

Price, J., 1988. 'Romano-British glass bangles from East Yorkshire', in Price, J. and Wilson, P. R. (eds), Recent Research in Roman Yorkshire, Br. Archaeol. Rep. British Series 193, (Oxford), 339-66.

Price, J. and Cool, H.E.M. Cool, 1989. ‘Report on the Roman glass found at the Cattlemarket, County Hall and East Pallant House sites, Chichester’ in Down, A. Chichester Excavations VI (Chichester), 132-42.

Oldenstein, J., 1977. 'Zur Ausrstung römischer Auxiliareinheiten. Studien zu Beschlägen und Zierat an der Ausrüstung der römischen Auxiliareinheiten des obergermanisch-raetischen Limesgebietes aus dem zweiten und dritten Jahrhundert n. Chr.', Bericht der Römisch-Germanischen Kommission 57 (1976), 49-366

Riha, E., 1979. Die römischen Fibeln aus Augst und Kaiseraugst, Forschungen in Augst Band 3, (Augst).

Rodwell, K.A., 1988. The Prehistoric and Roman Settlement at Kelvedon, Essex CBA Research Report 63 (London).

Seeley, F., 1995. ‘Roman doorbells’, Roman Finds Group Newsletter IX, 5-6

Shirley, E., 2001. Building a Roman Legionary Fortress (Stroud)

Simpson, C.J., 1976. ‘Belt-buckles and strap-ends of the later Roman empire: a preliminary survey of several new groups’, Britannia 7, 192-223.

Sunter, N. and Brown, D., 1988. ‘Metal vessels’ in Cunliffe, B. (ed.) The Temple of Sulis Minerva at Bath. Volume 2: The finds from the Sacred Spring, Oxford University Committee for Archaeology Monograph 16, (Oxford), 9-21.

Swift, E., 2000. Regionality in Dress Accessories in the late Roman West Monographies Instrumentum 11 (Millau).

Toynbee, J M C, 1973 Animals in Roman life and art, London

Vollenweider, M-L., 1979. Musée d'Art et d'Histoire de Genève. Catalogue raisonée des sceaux, cylindres, intailles et camées II, (Geneva)

Wade-Martins 1980. Excavations in North Elmham Park 1967-1975 East Anglian Archaeology 9 (Dereham).

Webster. G., 1960. 'The Roman military advance under Ostorius Scapula', Archaeol. J. 105 (1958), 49-98

Wheeler, R. E. M., 1930. London in Roman Times, London Museum Catalogue 3, (London)

Wild, J. P., 1970. Textile Manufacture in the Northern Roman Provinces, (Cambridge)

Woodfield, C., 1965. ‘Six turrets on Hadrian’s wall’ Archaeologia Aeliana Series 4 43, 87-200.

Zwierlein-Diehl, E., 1979. Die Antiken Gemmen des Kunsthistorischen Museums in Wien II, (Munich)

Top of page