SECTION 5.2: SOMERFORD KEYNES NEIGH BRIDGE ARCHAEOLOGICAL DESCRIPTION

TRENCH 5
Phase 1
Enclosures
Linear ditches
Other Features

Phase 2 (c early to mid 2nd century AD)
Phase 2a
Phase 2b

Phase 3 (Mid - ?later 2nd century AD)
Linear boundaries
‘Corn-drier’ 167

Unphased features from trench 5
Pits

TRENCHES 17 AND 8
Linear boundaries
Circular gullies/gully arcs
Pits

TRENCHES 7, 12 AND 13

FEATURES FROM OTHER TRENCHES
Trench 1
Trench 2
Trench 3
Trench 9
Trench 19

Table 1: Aisled Building Postholes

TRENCH 5 (Fig. 5.2.1: Plan of Trench 5)

Phase 1 (Figs 5.1.3: Trench 5 phase 1 and 5.2.1)

This phase was defined as all features lying beneath the phase 2 Roman linear boundaries, and is made up of many different stratigraphic sequences, presenting a composite picture rather than a single defined phase of activity. In total, 65 Phase 1 features produced diagnostic pottery groups, suggesting a date range of early/mid 1st to early 2nd century AD. A quantity of middle Iron Age sherds was also recovered, but could not be tied to any specific feature, and so must be regarded as residual (see Brown, this vol.). Of the five coins from features of this phase, three belonged to the late 3rd/4th century and must be regarded as intrusive. The remaining two comprised an Iron Age Dobunnic issue (40-20 BC) and an As of Gaius (AD 37-8).

The phase is characterised by a sequence of enclosure ditches and linear boundaries, and these are described here according to feature type.

Top of Page

Enclosures

Enclosure 1 (300, 100)

Enclosure 1 lay at the western end of Trench 5, and comprised two lengths of ditch (300, 100) enclosing an area c 22 metres in diameter, with a 2.4 m wide entrance in the west (section 160). The northern ditch section (300) ranged from 0.6 to 1.9 metres in width and 0.2 to 0.4 metres in depth, and there were at least two visible cuts. It may have cut the linear ditch 254/263 (but see below), and was cut by all other features where a relationship could be determined, indicating an early position in the phase 1 sequence. The phase 2 ditches, 318 and 101, cut through sections of the eastern side of the enclosure, but the ditch continued to the south-west to become 100. This part of the enclosure appears to have been dug in straight sections rather than a curve, and was c 1.5 metres wide, by 0.4 m deep. Only one cut was visible. It was cut by the phase 2 ditch 101, and also apparently by ditch 263/254, the feature that is recorded as being cut by the northern enclosure ditch 300. This makes little sense stratigraphically, and as the drawn section is ambiguous, it is perhaps more likely that both sections of enclosure 1 were cut by ditch 263/254, making it the earliest feature in this part of the trench. Two iron nails were recovered from the enclosure ditches, and no contemporary features were located in the interior. Pottery recovered from the ditch fills indicated a general later 1st/early 2nd century AD date.

Enclosure 2 (168, 234)

About 13 metres to the west of enclosure 1 lay enclosure 2, about 18 m in diameter, with the north-eastern side outside of the trench limits. It was generally V-shaped in profile, and 1.25 m wide by 0.5 m deep. On the western and southern sides, it was cut by the phase 2 Roman ditches 52 and 135, while to the east it cut part of enclosure 4 and another small curving gully (328). No entrance was located but it is possible that part of the north-eastern side was open. A single lead weight (Cat 879) was found in ditch section 168, and pottery indicated a mid to late 1st century AD date. The interior contained a number pits and postholes, some of which may have been contemporary, although no definite structures could be discerned. A small (c 3 m diameter) circular gully (147), possibly a ‘stack ring’ (see below), was positioned just to the north-west of centre within the enclosure, and was probably contemporary, as it also contained mid to late 1st century pottery, along with two residual middle Iron Age sherds.

Enclosure 3 (253, 275, ?335)

Enclosure 3 lay underneath the aisled building in the northern part of trench 5, and enclosed an area c 12 m in diameter. It was traced on the southern, eastern and western sides (253/275; 1.5 m wide, 0.56 m deep), but was cut by the ENE-WSW linear ditch 254 to the north. Ditch 335, running WNW-ESE, may represent a continuation of the enclosure to the north, although it was only 1 m wide and 0.22 m in depth. A shallow depression (287/8) to the north-west obscured most features in this area, aside from part of the phase 2 aisled building, and it is possible that the enclosure was open in this direction. The only finds were a single fragment of vessel glass, which was probably a modern intrusion, and a small quantity of 1st century AD pottery.

Enclosure 4 (222, 224, 270, 236)

A possible (sub)enclosure is represented by ditches 222/4 running SE-NW, continuing to curve northwards as 270, and then swinging around to the east as 236. This would create an area c 30 m across, with the south-eastern side apparently left open. Ditch section 222 appears to have been a shallow recut of 224, being 1.3 m wide and 0.4 m deep. It probably continued as 270, which was of similar dimensions (1.3 m wide, 0.35 m deep). Ditch 270 cut a number of smaller gullies (262, 271, 326, 299), some of which contained early 2nd century pottery. The remaining ditch section (236) was just slightly wider (1.4 m) and deeper (0.42 m), and was recut at least once. Stratigraphically, enclosure 4 is cut by enclosures 1, 2 and 3, along with a substantial curving ditch (136) in the west of the trench. It is thus one of the earliest major features in the trench. A small number of finds were recorded, and these included two brooches (1st - mid 2nd century AD), a bracelet, a scoop, tweezers, an intrusive late 3rd century coin, and a fired clay hearth plate. Pottery from the ditches ranged from early 1st to early 2nd century AD.

Enclosure 5 (223, 225, 175, 207)

A possible (sub)enclosure in the southern half of the trench is represented by a 16 m length of curving east-west ditch (175), with southerly extensions at the western (223/5) and eastern (207) ends. The southern side appears to have been left open. All parts were U-shaped and approximately 1.4 m wide and between 0.3 and 0.42 m in depth. The western extension was of two phases, with 225 being cut by 223, but both were seemingly part of the overall enclosure feature. There are no direct relationships with any of the other enclosures, but it appears to cut all other pre-phase 2 features in the area (340, 166, 130, 123), suggesting that it came late in the phase sequence. An Iron Age Dobunnic coin and fragment of vessel glass were the only small finds from the enclosure. Pottery indicated a late 1st to early 2nd century date.

Top of Page

Linear ditches

Ditch 263/254

In the north-western part of trench 5 lay ditch/gully 254 which ran ENE-WSW for c 28 m and then turned south-eastwards for a further 18 m as 263, before leaving the trench. It was quite variable in width (0.8 to 1.4) and depth (0.2 to 0.4), and may have formed two sides of a large sub-rectangular enclosure, although no appropriate ditches could be traced outside the trench as cropmarks. Its relationship with other features is not always that clear, but it appears to cut enclosure 3 and possibly enclosure 1 (see above) and is cut by a short length of another phase 1 ditch (308). No small finds were recorded from this ditch, and the pottery ranged from mid 1st to early 2nd century AD. The later material generally came from the southern section (263).

Ditch 165/340

Cut by all features where a relationship was determinable, was a length of ditch/gully in the south-eastern part of the trench, running 12 m NE-SW (340) before probably turning south-east for a further 9 m (165), although it is not certain that 165 was of the same feature. Both sections were approximately 0.8-1 m wide and 0.3 m deep, and may have formed part of an early phase 1 sub-enclosure, open to the south-east. Pottery was generally of later 1st century date.

Ditch 182

Approximately parallel to and just the west of Roman ditch 52 was a ditch running for 10 m, before turning north-west and disappearing into a group of ditches and pits. It may well have originally formed part of a sub-enclosure. It was 0.85 m wide and 0.23 m deep and was cut by phase 3 ditch 182. Pottery indicated a late 1st century date.

Ditches 191/194, 169, 180/196, 166, 130

In the southern part of trench 5, and cut by all features except 165, was an arrangement of approximate north-south and east-west linear ditches. Ditch 191/194 (1 m wide, 0.4 m deep) to the west was cut by ditch 169, which lay on a similar alignment (NNW-SSE), and was of similar dimensions. 169 ran for c 12 m before turning east into 180/196, which was considerably more substantial (2 m wide, 0.5 m deep). These may have formed two sides of a sub-rectangular enclosure (c 11 x 12 m?), possibly with ditch 166 (1 m wide, 0.3 m deep) forming the northern side. To the east lay the long north-south ditch 130, which may have formed the eastern side, although it clearly cut ditch 196. It ran for 27 m from the junction with phase 2 ditch 135 to exit the trench in the south, and was 0.8 m wide and 0.3 m deep. Very few finds were recovered from any of these ditches, but these included copper alloy tweezers from 169, and a brooch (mid-late 1st century AD) and Roman coin (AD 37-8) from 166. Pottery from all features except 130 indicated a mid to late 1st century date. That from 130, which was stratigraphically later, suggested an early 2nd century date. A small amount of cattle and sheep bone came from ditch 166 while much larger quantities came from ditches 169 and 130.

Ditches 123, 116/118

In the far south-eastern area of the trench lay a substantial ditch (1.5 m wide, 0.6 m deep) orientated NNE-SSW. It ran for 11 m, and was cut by ditch 175 (Enclosure 5) to the south, and phase 3 ditch 114 to the north. Its relationship with east-west ditch 116/118 was indeterminable. No finds were recovered except pottery, which gave an early 2nd century date, and a small amount of animal bone. Running perpendicular to 123 was ditch 116, which was a continuation of 118 to the east. It was cut by 114, and cut ditch 117, but its relationship with 123 was uncertain. Late 1st to early 2nd century pottery was recovered from the fills.

Ditch 142 (136, 142, 249, 324)

Parts of a substantial ditch lay in the north-east of trench 5, running approximately NW-SE, and curving eastwards out of the trench. It was cut by E 2 but cut E 4, and varied in size throughout its length, averaging 0.6 m in depth and 1.4 m in width. At least two different cuts of the northern terminal were evident, the earliest being 324, followed by 249. 324 seemed to cut through a large pit (161), although the relationship was uncertain. Most of the pottery from the feature dated from mid to late 1st century AD, with a small amount of early 2nd century material. Other finds included a copper alloy finger ring and later 1st century AD brooch, and a mid 4th century coin. A small quantity of sheep and cattle bone also came from ditch section 142.

Top of Page

Other Features

Burnt pit 273

Lying just to the north of post-pit 267 belonging to the aisled building was a small (0.4m diameter, 0.2m deep) pit containing a small amount of oak charcoal. Although not cut by 267, it is presumed to have been earlier than the aisled building, and it could have been contemporary with enclosure 3, within which it was located. No associated finds were recovered to date the feature.

Circular Gully 147

A continuous circular gully, 2 m diameter internally, was situated within enclosure 2. It was approximately 0.8m wide and 0.2m deep, had a rounded base, and was cut into the natural silt (section 63). It cut pit 148 (1.47m wide, 0.47m deep), which contained a small quantity of animal bone, and was cut by pits 150 and 153 (both c 0.95m wide, 0.3 - 0.38m deep). The only small finds from the gully fill comprised an iron knife and an iron nail, although quantities of mid to late 1st century pottery suggested contemporaneity with the surrounding enclosure. A reasonable quantity of animal bone was also recovered, including cattle, horse, sheep and pig. It is similar to features found at Fairford Thornhill Farm and Claydon Pike, which were interpreted as ‘stack rings’, used to store animal fodder. These were dated from the middle Iron Age to the early Roman periods.

‘Posthole structure’ (B 2)

A possible posthole structure was located in the north-eastern part of trench 5 within E2 (139, 140, 143, 144, 145, 153, 154, 155, 156, 158, 159, 160, 162). The ‘structure’ was approximately 8 m by 4 m in size, and lay on a SW-NE alignment. It does not relate to any of the phase 2/3 linear ditch alignments, so it’s interpretation is far from certain. Ceramic dating evidence ranged from mid 1st to 2nd century, and it is possible that it was contemporary with E 2, and the circular gully (147) lying just to the north.

Gully arc 328

An arc of a small gully, 0.5m wide and 0.26m deep, was located on the north-eastern edge of the trench. It contained no diagnostic finds but was cut by E 2, so has been assigned to phase 1.

Top of Page

Phase 2 (c early to mid 2nd century AD) (Figs 5.1.4: Trench 5 phase 2 and 5.2.1: Plan of Trench 5 )

At some point in the early 2nd century AD, the phase 1 features in trench 5 were replaced by a more regular layout of east-west and north-south linear ditches forming rectilinear enclosures and trackways. A substantial aisled building was also erected in this phase. Although much of the pottery was quite mixed, the general date range for features of this phase falls within the 2nd century, with a slight preponderance of early to mid 2nd century material. There is some stratigraphic basis for dividing the phase into 2a and 2b, although certain features (e.g. the aisled building) undoubtedly existed in both. It is not possible to date these sub-phases more accurately than the chronology given to the whole phase. The linear boundaries from both sub-phases will be described initially, followed by an account of the aisled building.

Top of Page

Phase 2a

Robber trench/beam slot 70 and gully 305

In the western part of trench 5 lay a north-south robber trench or beam slot (70) running for 22 metres. Throughout most of its length, it had steep sides with a roughly flat base, and was 0.8m wide and 0.16m deep. The nature of the fill appeared to suggest stone robbing or the removal of a sleeper beam, and so it may represent part of a substantial palisade. Stratigraphically, it cut all phase 1 features, and was cut by the east-west trackway ditches 318 and 101. It may also have been cut by a curving east-west gully (305) running down from the north-west, although the relationship here is uncertain. This gully was cut by another similar feature (306) on approximately the same alignment, lying just to the north. This gully also cut beam slot 70. A copper alloy sheet and mid 2nd century pottery were recovered from 70.

Ditch 135

A substantial east-west ditch entered the trench from the east and ran for 24m before being cut by one of the north-south trackway ditches (52). It reappeared on the west side of 52, continuing for 2.5m into the trackway, thus indicating that this feature was not contemporary. The ditch was between 1.35 and 1.7 m wide and 0.5 to 0.6m deep and only a single cut (section 45). It cut all features except 52 and north-south ditch 114 to the east, and is probably contemporary with ditch 122. A single fragment of fired clay and small quantity of late 1st to 2nd century pottery were the only finds recovered.

Ditches 122 and 112

Ditch 122 ran south from 135 for 25m before existing the trench. It varied from 0.92 to 1.54m in width and 0.4 to 0.56m in depth, and appeared to have been cut by 135, although on a spatial basis, it is likely that they were broadly contemporary. It is also probably contemporary with east west ditch 112, which joined perpendicular and ran parallel to 135, 5m to the south. This ditch was 0.72m wide and 0.47m deep, and was cut by the phase 3 ditch 114. No finds were recorded from these ditches aside from early to mid 2nd century pottery.

Top of Page

Phase 2b

North-south ‘trackway’ ditches (52, 172; sections 119, 101)

Two north-south ditches (52, 172) ran for 56m through the middle of trench 5, c 3 to 4m apart, and it is suggested that they defined a possible trackway. They were between 1.4 and 1.7m wide and 0.42 to 0.5m deep, and had one visible cut. The eastern ditch (52) cut through east-west ditch 135, along with many phase 1 features, and was cut along its length by three short ditch sections (181, 216, 244), a major ditched boundary (164) and a possible corn drier (167). Associated pottery was virtually all 2nd century in date, with fragments of early (AD 90-110) and later (AD 150 +) samian suggesting that it remained as an open feature at least until the later half of that century. The western ditch (172) was cut only by 164 and 181 to the south, and most of it was probably open for longer than 52, although the date range of the pottery was similar (see discussion). Considering the length of the trackway ditches, finds were quite scarce. Two Roman coins (late 3rd and 4th century AD) and a copper alloy tube were recovered from the upper fills of ditch 52, while six nails, a bone pin, iron cleat, lead weight and a Denarius of Antoninus Pius (138-61) came from ditch 172.

East-west ‘trackway’ ditches (318, 277/8, 101)

Joining perpendicular to ditch 172 were two parallel east-west ditches, c 5m apart, which may have formed another trackway. They ran for 34-9m from the western end of the trench to apparently terminate at the north-south trackway, although no clear relationships were recorded. The northern trackway ditch had two cuts at its western end, with 278 cutting 277. Both were 0.8-9m wide and 0.2m deep. Towards the east only a single cut was visible (318), which was 1.5m wide and 0.3 m deep. The southern trackway ditch (101) was more substantial, being up to 1.9 m wide and 0.5 m deep, although these measurements did vary along its length. Both ditches cut beam slot 70, while a short length of north-south ditch (320) appeared to cut 318, but terminate just to the north of 101. It is quite possible that ditch 101 was still in use into phase 3. Six iron nails and a copper alloy fitting were the only small finds recorded from ditch 318, while 101 contained melted lead, an iron nail and a 2nd century AD copper alloy stylus in its fill. Most of the pottery from the trackway ditches was (mid) 2nd century in date.

Gully 306

Gully 306 ran approximately east-west for c 30 m from the western part of trench 5, terminating at what must have undoubtedly been the outer wall of the aisled building, although no traces remain of this. The gully was 0.94m in width and 0.14m in depth, and cut phase 2a features 305 and gully/beam slot 70. Finds from the gully comprised tweezers, a fragment of window glass and early 2nd century pottery.

Northern parallel ditches

In the northernmost part of trench 5 ran 4 parallel east-west ditches, lying between 2 and 2.5 m apart. They were 1 to 2 m in width, but were not excavated in this trench. From the soil marks over the site, these ditches appear to run from the D-shaped enclosure, sectioned in trench 13, and are probably 2nd century in date (see trench 13). In trench 5 they appear to form the northern boundary of an enclosure around the aisled building.

Aisled Building (Figs 5.2.2: Aisled building and 5.2.3: Aisled building Sections; Table 1)

Located to the north of ditch 318 and parallel with ditch 172 was a very regular alignment of postholes on a north-south orientation (see Table 1). All of the postholes were between 0.35 and 0.55 m deep, measured 0.75 to 1.75 m in diameter (Fig. 5.2.3), and formed a substantial aisled building. The outer walls were presumably constructed purely of timber, and a reasonable reconstruction based upon calculations by Mackreth (1996 66) would give overall outside dimensions of 10-12 m wide and 27m long. Most of aisle posts on the south, east and west sides were about 1.5 m apart, while the north - where the entrance presumably lay - remained open. Many of the postholes seem to have been replaced at some point, especially those along the western side. The postholes cut all other related features with the exception of shallow pit 260, which appears to have cut the north-east edge of posthole 257. Two stone lined post-pits (310, 311) in the central interior of the building may have been part of the structure, though form no easily definable pattern, and 311 contained early to mid 1st century pottery, so could well belong to phase 1. Finds from within the posthole fills comprised two iron nails, a piece of glass, and fragments of a quernstone and whetstone. All postholes contained a small amount of pottery, most of it dating to the 2nd century AD and presumably part of the post packing. This assemblage included sherds of samian indicating a terminus post quem of AD 100 - 125 for the construction of this building.

Possibly related to the use of the building, was a substantial regular arrangement definite tegulae and a limited amount of box flue tile. These looked to have been stacked against the exterior southern wall, and presumably relate to the later history of the building, in phase 3 (see discussion). In addition to this stack, very large quantities (c100kg) of mixed tile were found within a general layer (25) under the ploughsoil in trench 5 immediately to the east of the aisled building.

Top of Page

Phase 3 (Mid - ?later 2nd century AD) (Figs 5.1.5: Trench 5 phase 3 and 5.2.1: Plan of Trench 5 )

Around the middle of the second century AD, many of the earlier features were overlain by a series of north-south, east-west oriented linear ditches (163, 164, 114, 181, 216, 244, 320), which probably represent a conscious revision and redefinition of the phase 2 boundaries. It is likely that at least part of the east-west and north-south trackways went out of use at this time, although the aisled building seems to have continued in use. New trackways appear to have been constructed. The phase is not well dated as most of the pottery was quite mixed. However, there is nothing in the stratified ceramic record that need be dated much beyond the later 2nd century AD, and so it is presumed that the ditches and building were abandoned by this point. The quantity of later 3rd and 4th century coins and small finds probably relate to a separate phase of activity, although no actual archaeological features can be dated to this period.

Top of Page

Linear boundaries

Ditch 114

Running north-south for 22m in the eastern corner of trench 5 was V-shaped ditch 114. It was approximately 1.22m wide and 0.54m deep, and was the stratigraphically the latest feature in that area, cutting phase 2 ditches 135 and 112. Finds comprised a piece of limestone masonry, an iron sheet and nail, a fragment of glass, and early to mid 2nd century pottery. This ditch may have formed the eastern part of a north-south trackway (8 m wide) with ditch 163.

Ditches 163 and 164

Ditch 163 lay 8m to the west of 114, on the same north-south alignment, and probably part of the same trackway. Upon entering the trench from the south, it ran for 24m before curving to the west and terminating at what may have been the entrance into a sub-rectangular enclosure (21 x 9m) containing the possible corn drier (see below). The ditch measured 1.4m in width and 0.52m depth and its fill contained two brooches, a bone pin, an iron fitting and a piece of slag, which is slight evidence for some light industrial activity in the area. It was joined perpendicular with east-west ditch 164 (0.92 - 1.5m wide, 0.36 - 0.47m deep), which ran for 32m before existing the trench in the west. This ditch cut through both of the north-south trackway ditches (52, 172), and formed the southern boundary to the ‘corn drier’ enclosure. It also formed part of an east-west trackway (5-6 m wide), probably replacing the one further to the north. There is evidence for at least one re-cut. Finds included a fragment of vessel and window glass, two 2nd century AD brooches, five iron nails and a spindlewhorl. Pottery of mid to late 2nd century date was recovered from the fills of both these features. A reasonable quantity of animal bone was recovered from these features, mostly sheep and cattle.

Ditches 181, 216, 244

Three relatively short (10-13 m) lengths of east-west ditch were cut along the line the central north-south trackway. To the south, ditch 181 (0.9 m wide, 0.44 m deep) cut both of the trackway ditches and continued out of the trench to the west. Its eastern terminal lay upon the line of ditch 52, and it probably formed part of an east-west trackway with ditch 164/173. Further to the north, ditch 216 (1.15 m wide, 0.48 m deep) ran eastwards from ditch 172, cut ditch 52, and formed part of the boundary to the ‘corn drier’ enclosure. 25 m north of 216 lay 244 (1.5 m wide, 0.46 m deep), which ran ESE from 172 for 10 m – again cutting ditch 52 - before exiting the trench. The finds from these ditches comprised a copper alloy clothes fitting from 216, and a small quantity of later 1st to mid 2nd century pottery, most of which is presumed to be residual. The stratigraphy of these ditches suggests that part of the western north-south trackway ditch (172) was still in use (possibly recut?) during phase 3, while the eastern ditch (52) was probably filled in.

Gully 320

Less than 2 m to the west of ditch 172 and immediately south of the aisled building lay north-south gully 320 (0.8 m wide, 0.5 m deep). It cut through the northern of the east-west trackway ditches (318) and terminated immediately to the north of the southern ditch (101), possibly suggesting that this feature was still open at this time. The function of this short (5.6 m) length of gully is uncertain, but its proximity and shared alignment with the aisled building suggests some association. Aside from a few sherds of 2nd century pottery, the only find recovered was a single iron nail. Overlying ditch 229/101 just to the east of this gully was an area (0.8 x 0.6 m) of six limestone slabs (261), some with evidence for burning. It is unsure what this feature represents, though it was probably contemporary with phase 3 activity.

Top of Page

‘Corn-drier’ (167; Fig. 5.2.4: ‘corn drier’)

Cutting through ditches 52 and 166 was a horseshoe shaped channel lined with several large flat pieces of limestone (c 0.25-0.3 m wide, 0.04 m thick). It was c 4.2 x 3 m in size. The slabs had traces of burning and there was a layer of burnt material on the base of the channel. Collapsed slabs in the east (167/C) suggest that the feature was originally covered over. The probable stoke hole, which lay to the south east (section 167/B), was 0.9 m diameter and 0.37 m deep, and this led into the lined flue channel that was initially 0.32 m wide, but then broadened to c 0.8 m. The slabs at the side of the channel were at an angle of 80 to 85 degrees to the north, but only 50 to 70 degrees to the south. The lowest fill within the flue lay underneath the slabs, suggesting that they represented a re-lining of the flue. Lining slabs were not present in all sections, having presumably been removed after the use of the structure. To the west (167/F), there is the clearest indication that the structure had more than one phase, as at least one later cut is visible. A pit in the north-western side may well have been an earlier stoke hole, to be eventually replaced by the pit in 167/B, although this must remain uncertain. There were generally three fills throughout most sections of the feature, consisting of silty clay materiel with charcoal. The only small find recovered was a single iron nail. Pottery was quite mixed and ranged from late 1st to 2nd century in date.

Although not of the conventional T-shape, this feature was initially interpreted as a corn drying oven. Physically, it can most easily be equated with Morris’s ‘rectangular’ type drier found at sites such as Longthorpe in Cambrideshire (Morris 1979, 101, fig.11). A more local parallel may possibly be found at Birdlip quarry in Gloucestershire (Mudd et al 1999, p.191), where an unusual elongated sub-rectangular pit with limestone blocks and a charcoal spread resembles one side of the Somerford Keynes structure. However, the interpretation of the Birdlip Quarry structure remains uncertain (ibid. p.192). The principal argument against the Somerford Keynes structure being interpreted as a corn drier is that the environmental samples only contained a single charred grain. It may therefore have been that this oven was utilised for a different purpose.

Structure 167 was positioned within the western side of sub-rectangular enclosure, with which it was undoubtedly contemporary.

Top of Page

Unphased features from trench 5 (Fig. 5.2.1: Plan of Trench 5)

There were a number of features within the trench that could not be assigned to any phase, either because of a lack of stratigraphic relationships and datable material, or else because they were not excavated. These features comprised ditches, gullies, pits and postholes, and have not been described in detail here (see database).

Pits

39 pits came from trench 5, ranging from 0.12 to 0.6 m in depth, and from 0.4 to 2.6 m in diameter. Very few had any associated finds, and so their function and chronology remain uncertain. An exception was a small circular burnt pit (273) at the north end of the aisled building, belonging to phase 1 (see above).

Top of Page

TRENCHES 17 AND 8 (Fig. 5.2.5: Trenches 17 and 8)

Trench 17, located c 25m east of trench 5, was the second largest excavated area on site (235.4 m˛), while the much smaller trench 8 led off from its north-western corner. Like trench 5, trench 8 contained a series of intercutting ditches, gullies and pits, although many of the relationships were unclear due a combination of shallow disturbed stratigraphy and problems of standing water. The chronological range of the pottery was quite similar to that of trench 5 (1st - 2nd century AD), with very little to suggest activity beyond the 2nd century AD. Only three stratified coins were recovered from features within trench 17, two with a date of 1st-2nd century AD and the other belonging to the late 3rd century. Of the 52 small finds recovered from the trench, 12 were brooches with a general 1st to mid 2nd century AD date range. However, there were reasonable quantities of late 3rd and 4th century small finds recovered by metal detecting in the vicinity of trench 17 (see Fig. 5.1.6: Distribution of finds from metal detecting), which does point to late Roman activity of some kind in this area.

The spatial arrangement of ditches and gullies appears less regular than in trench 5, and is difficult to reconstruct in a meaningful way. This, and the fact that the pottery appears to have been very mixed also ensures that accurate phasing of the features is not possible. Various layers of alluvial silt and gravel lay across the site, particularly obscuring features in the far eastern area. In the west was a layer of mid grey brown silty loam (34) containing much occupation debris (pottery, bone etc), including a small number of 1st to 4th century coins. A stone spread (427) was associated with this layer, being particularly concentrated in the tops of ditches.

The following archaeological description is arranged by feature type.

Top of Page

Linear boundaries

Ditch 46/65

In the western part of trench 8 lay a substantial ditch (65), measuring 4.6m across and 0.3m deep, which would seem to have been a continuation of the main curving enclosure ditch (16) seen in trench 13. Ditch 46 (2.17m wide, 0.23m deep) was probably a recut of this boundary. Late 1st and 2nd century pottery came from these features, along with many iron nails and a mid 1st to 2nd century AD Polden Hill brooch.

Gullies 404, 406

Possibly forming the western side of a rectangular enclosure in the northeast of the trench, gully 404 ran south from ditch 409 until cut by pit 400, and was 0.65m wide and 0.19m deep. It cut pit complex 405 and gully 406, and contained a possible lead seal and 2nd century AD pottery. A probable continuation of this feature was traced to the south, but was not excavated. Gully 406 appeared as a terminal extending eastwards from 404, and possibly continuing through pit complex 405 to the west, although this was not certain.

Ditch 407

A ditch ran east-west in the northern half of trench 17, measuring 0.85m wide and 0.42m in depth. It was cut by a large well (400) to the west, and by the major north-south ditch to the east (410), although it may have been roughly contemporary with the latter feature, forming part of a rectangular enclosure. Finds from the silty grey loam fill comprised a 2nd century AD pottery, an iron hook and a small quantity of animal bone.

Ditch 408

Between ditch 410 and pit 400 lay narrow V-shaped ditch 408, which ran north-south for 7.5m before turning eastwards for another 5 m where it was cut by 410. It was 0.6m wide and 0.37m deep, and was cut by ditch 407. No finds were recorded.

Ditch 409

Ditch 409 ran east-west in the northern corner of trench 17, at right angles to ditch 410, with which it may have been contemporary. There were indications of 3 cuts, with the latest being 1.3m wide and 0.13m in depth. It was filled with alluvial material, which contained a small amount of 2nd century pottery and animal bone.

Ditch 410

Running north-south through the central northern part of the trench, ditch 410 is stratigraphically the latest feature in this area, cutting through ditches 408, 407, 420, 447 and gully 414. It was 1.30m wide and 0.3m deep, and may have formed the east side of a rectangular enclosure with ditches 407, 409 and gully 404. Its fill of silty loam contained pottery of later 1st to mid 2nd century date, along with a lead weight, two pieces of slag and a small quantity of animal bone.

Ditches 415, 417, 418

In the eastern part of trench 17 were two major north-south ditches (415, 418), with an east-west ditch (417) lying between them, and cut by both. Ditch 415 to the east measured 2.2m wide and 0.45m deep, and contained a quantity of mixed finds in its fill, including late 1st century pottery, vessel glass and a post-Roman coin. Lying c 10m to the west, ditch 418 ran north-south, but veering slightly towards 410 in the south. It varied in width from 0.85 to 1.3m, and in depth from 0.34 to 0.43m. and contained 2nd century pottery. It was cut by ditch 421 to the north, and appeared to terminate 1. m to the north of this. Running east-west between these boundaries was ditch 417, 1.3m wide and 0.26m deep. A very small quantity of animal bone came from the silty loam fills of ditches 417 and 418.

Ditches 420 and 447

Feature 420 may represent a number of ditch cuts, running WNW-ESE at right angles to the southern part of ditch 410, which cuts through it. It was 1.2 m wide and 0.4 m in depth. To the west of 410 lay 447, which may well be a continuation of 420, as it lay on the same alignment, and had similar dimensions (1.1m width, 0.48m depth). It was traced WNW for just over 2m. None of the ditch sections contained any finds.

Ditches 421, 423

Ditch 421 ran east-west, terminating just east of ditch 410. It was 0.9m wide, 0.35m deep, and contained no finds. It was cut by the large posthole 422 (0.6 m wide, 0.23m deep), and cut ditches 418 and 423. Ditch/gully 423 ran from 421 to the north-east and was 0.6m wide and 0.32m deep. The eastern extent of both ditches was not able to be located, and neither contained any finds.

Ditch 428

Running NW-SE in the western half of the trench, ditch 428 was seemingly associated with rubble spread 427, which in part formed the upper fill. It was cut by ditches 430, 433 and pit 431, and cut ditch 429 and pit 435. Three or possibly four cuts were visible, with the average width being 1.8m and depth 0.48m. Early to mid 2nd century pottery and a quantity of animal bone was recovered. Lying 2-3m to the west of the ditches northern section were two postholes (445, 446), underneath the rubble spread.

Ditch/gullies 429 and 430

Ditch/gully 429 ran north-eastwards for 12.5 m from the western edge of the trench to the edge of ditch 428, which partially cut through it. It was just 0.5 m in width and 0.3 m depth and contained a small amount of 2nd century pottery. c 5 m to the north another ditch/gully (430) ran north-east, but on a different alignment. It was of similar dimensions (0.5 m width, 0.23 depth) and cut through ditch 428, although it was not traced on the other side. Late 1st - early 2nd century pottery was recovered.

Ditches 432, 433, 434, 438

At least three separate ditch cuts were seen running SW-NE from the south-western edge of trench 17. Ditch 433 (1.25m wide, 0.4m deep) ran parallel with 432 (0.7m wide, 0.45m deep) for 7 m before they converged. Another cut, 434 (0.4m depth), was also seen in one of the sections. 433 seems to have been the latest feature stratigraphically, cutting both 434 and 432, along with ditches 438 and 428. Only the western terminus of ditch 438 was located, and it followed a similar line to the other ditches. It was 1 m wide but was not bottomed due to too much standing water. Later 1st to 2nd century pottery was recovered from all ditch cuts, along with two worn Roman coins. An illegible 1st-2nd century coin came from 432, while another illegible late 3rd century coin came from 433. Animal bone was recovered from all ditches, with the small assemblage from 433 comprising mostly horse bones.

Ditch 439

Ditch 439 ran west-east near the southern corner of the trench, cutting through all other features (433, 432, 440). It was 0.8m wide and 0.35m in depth, and seemed to continue westwards, appearing in trenches 9 and 1. If this correct however, it would be the only one of the parallel ditches to continue eastwards from the main curving ditch (16). Unfortunately, the relationship of this feature with the curving ditch was not recovered. 2nd century pottery came from the silty clay fill.

Top of Page

Circular gullies/gully arcs

Gully 419

A circular gully in the north-eastern part of trench 17 measured 0.26m wide, 0.15m deep, and 8.5m in diameter. Its western half was obscured by alluvial material, and the eastern side was cut by ditches 423 and 418, making this feature the earliest stratigraphically in this part of the site. 2nd century pottery was recovered from the fill.

Gully 440

A 3.5m length of curving gully in the south of the trench ran NW-SE, cut by ditch 439. It was 0.9m wide and 0.29m in depth, and its northern terminal was probably cut by shallow pit 441 (1.6m wide, 0.15m deep), although the relationship could not be determined.

Top of Page

Pits

Fourteen pits were located within trench 17 of which eight contained pottery dating from the 1st to mid 2nd century AD. Many of the pits were not fully excavated.

North-western pit group (400, 401, 402, 405)

A number of very large pits lay in the north-western part of the trench, including pit 400, which was 4.7m across and 0.56m deep, and cut through gully 404. Many finds were recovered from this pit, including 2nd century pottery, three copper alloy brooches with a 1st to mid 2nd century date range, and a reasonable quantity of mixed animal bone. To the west of this lay a large oval pit/watering hole (402), 3.3 x 4.15m across and 0.53m in depth, which was cut by another large oval pit 401 (2.45 x 1.9m across). Finds from these features included two brooches (1st-2nd century), vessel glass, a worn Roman coin (66-81 AD) and 1st to 2nd century AD pottery. Lying 1.5m to the north of 400 was what appeared to be a series of intercutting pits (405, 424, 425) cut by gully 404. This was not fully excavated but produced some 2nd century pottery.

Pits 411, 412

Positioned just to the east of the upper part of ditch 410 were two intercutting oval pits. Pit 411 was 1.5m wide, with steep sides and flat bottom, 0.34m in depth. It cut through gully terminal 414, and was cut by pit 412, which was 2.1m long and 0.34m deep. Late 1st century pottery came from the first earlier pit (411), while pit 412 produced pottery of early to mid 2nd century date.

Pit 426

An oval pit was excavated, lying just within the area of alluvium in the east of the trench. It was 1.7 x 1.55m across, 0.3m deep, and was filled with large pieces of limestone and roof tile. Early to mid 2nd century pottery was also recovered.

Burnt pits (413, 416)

Two small burnt pits were located in the northern half of the trench. 413 (0.52 m dia, 0.16 m depth) was located south of pit 400, while 416 (0.38 m dia, 0.12 m depth) lay south-east of pit 411. Both features contained considerable quantities of Quercus sp. charcoal. No associated pottery was recovered to date the pits, but they presumably relate to the 1st - 2nd century activity in the area.

Top of Page

TRENCHES 7 AND 13 (Fig. 5.2.6: Trenches 7 and 13)

Trench 13 was located c 40m north of trench 17, being connected to it via trench 7, and c 40m east of trench 5. It was the third largest excavated area at Somerford Keynes (113.45 m²), and was positioned in order to investigate the large curving ditch (16) and its relationship with the parallel linear ditches running off to the west (21, 20, 2917). These ditches were located in trench 12 just to the west (Fig. 5.1.2: Site plan showing trench locations), where they were given a single group context number (36). Pottery from the major ditched features indicated a 2nd century date, although 1st century material was recovered from some features and general spreads. Very few other finds were recovered from features within this trench, although these did include the sculptural fragments of an eagle and shield. Significant quantities of small finds dating from 1st to 4th century AD were recovered from metal detecting just to the southeast of trench 13 (Fig. 5.1.6).
The following archaeological description is arranged by feature type.

Enclosure ditch (16)
Ditch 16 formed a major D-shaped enclosure, which was traced running for just over 120 m within the eastern part of the salvage area (Fig. 5.1.2). Within trench 13 it was sectioned in five areas, which indicated that there were at least two major cuts with total dimensions being c 4.3m wide and on average 0.4m in depth (see section 148). Five layers of fill indicated that the ditch gradually silted up, although pottery recovered from all layers was generally 2nd century AD in date, suggesting that its entire period of use and abandonment lay within this period. Aside from pottery, the only finds comprised four pieces of stonework, two of which (sf 281-4) were sculptural fragments of the shield and eagle (see Henig, this vol.). The other two comprised part of a limestone mortar and an unidentifiable piece of worked stone. These were found on top of the north-western cut, just outside of the ditch. A quantity of animal bone was also recovered from the ditch.

Parallel ditches (291, 20, 21)
Contemporary with at least one cut of ditch 16 were four parallel ditches running to the west, the southern of which was not excavated. Ditch 291 to the north was 1m wide and 0.26m deep, although it did shallow considerably towards its junction with 16. Ditch 20, 3m further south, was of similar dimensions (1m wide, 0.2m deep), and was shown to abut ditch 16. Lying 5m to the south of this was the more substantial ditch 21, being 1.4m wide and 0.28m deep. A small gully connected this to the unexcavated fourth ditch, c 2m to the south. None of these ditches had any associated finds to indicate date or function, although their relationship with ditch 16 indicates contemporaneity (i.e. 2nd century AD). However, it remains uncertain as to whether any or all were open at the same time, or if they represent a succession of northern boundaries for the site. The ditches were traced further west in trenches 12, 5 and 3.

Ditch 297, gullies 329, 298
Running east-west from the central eastern part of trench 13 was ditch 297. It was 1.34m wide, 0.38m deep and was cut by ditch 16 to the east. On the same alignment 13m to the north was gully 298 (0.8m wide, 0.1m deep), but this was only traced for 6.5m before exiting on both sides of the trench. These features may have formed part of an enclosure leading off from ditch 16, and both contained late 1st to 2nd century pottery. Running parallel 1m to the east of ditch 16 was gully 329 (0.75m wide, 0.16m deep). This cut through ditch 297 and may have formed a new western boundary for the enclosure.

Ditch/gullies 313-317
A series of approximately east-west ditches were excavated within the northern part of narrow (1.7m wide) trench 7, which linked trenches 13 and 17. Gully 313 (0.7m wide, 0.39m deep) cut through both a layer of silting (312) overlying part of southern trench 13, and ditch 314 just to the south of it. It contained late 1st - early 2nd century pottery. Ditch 314 (1.4m wide, 0.34m deep) was stratigraphically the earliest of the sequence, and contained mid 1st century AD pottery. To the south it was cut by gully 317 (0.9m wide, 0.32m deep), containing mid to late 1st century AD pottery. Finally, lying further south in trench 7 were gullies 315 (0.6m wide, 0.3m deep) and 316 (0.58m wide, 0.28m deep), from which was recovered pottery of early to mid 1st century AD date. All of these features were quite consistent in date (early to late 1st century AD), and probably relate to the earliest phase of activity on the site (phase 1 in trench 5).

Top of Page

FEATURES FROM OTHER TRENCHES (Fig. 5.1.2: Site plan showing trench locations)

Trench 1

Located in the far south-west of the site, trench 1 (31 x 1.8m) cut through a number of linear east-west boundaries, including the three parallel ditches (3-5) running from circular enclosure 16 across the southern part of the site. No finds were recorded from any of the features.
Ditches 3-5
Running from 2 to 2.4m apart in the northern part of trench 1 were three parallel ditches, all approximately 2m wide and 0.3m deep where excavated. Ditch 3 was located in trench 5 as ditch 30, while ditch 4 was probably the same as ditch 63 in trench 9 and possibly also 439 in trench 17. It is also possible that ditch 5 may well be the same as gully 62 in trench 9.
Other features (42-45, 47)
In the southern half of the trench lay four further ditch/gullies, aligned east-west. They ranged from 0.56 to 1.2m in width and 0.12 and 0.24m in depth, and most did not show up outside of the trench area. Gully 45 cut pit 47 (2.4m wide, 0.36m deep), while ditch 43 may well have cut gully 42, but the relationship could not be determined.

Top of Page

Trench 2

Another narrow linear trench (19 x 1.8m), lying 11m to the north of trench 1, revealed two ditches and a gully. Ditch 2 was 2m wide and 0.42m deep. It was traced westwards out of the trench and eastwards into trench 5, as ditch 164. Lying 6m further south, ditch 1 was of similar dimensions (2.2m wide, 0.42m deep) and was shown to continue westwards parallel to ditch 2. At 10m further east, it turned sharply southwards, parallel to ditch 172 coming from trench 5. These ditches undoubtedly represent an east-west trackway, which turned to the south along the same alignment as the aisled building with which it was presumably contemporary. No finds were recorded. Just to the north of ditch 2 was a small section of gully (8) with early to mid 2nd century pottery from its fill.

Trench 3

A 22 x 1.8m trench in the north-west of the site was dug to investigate the northern parallel east-west boundaries, also revealed in trenches 5, 12 and 13. The northern most ditch (19) was 1.3m wide and 0.4m deep, and to the south of this lay ditches 18, 15, 14 and 13, which were between 1.5 and 1.96m wide and 0.3 to 0.38m deep. No finds were recovered from these features, other than a small quantity of animal bone.

Top of Page

Trench 9

This small linear trench (25 x 1.8 m) lay between trenches 5 and 17, and a revealed number of east-west ditches and gullies, along with 7 pits. Very few finds were recovered, with just one feature providing any dating evidence (2nd century AD).
Linear boundaries
Running east-west in the north of trench 9 was a ditch (63) measuring 1.26m wide and 0.32m deep. It was traced in the salvage area running further east and west, and may well be the same as ditch 439 in trench 17 and ditch 4 in trench 1. The only finds comprised an iron nail and modern vessel glass. At 4m to the south of 63 lay gully 62 (0.5m wide, 0.16m deep), which lay on the same alignment as ditch 5 in trench 1 and may represent a continuation. The gully cut pit 61. Just 2m further south was ditch 60 (1.36m wide, 0.41m deep), which was shown to continue westwards before turning north and entering trench 5 as ditch 128. This was the only feature to produce pottery, which indicated a 2nd century AD date, and a possible quern stone was also recovered from this ditch. In the south of trench 9 were two further gullies (67, 56) running parallel 0.4m apart. Both were of similar dimensions (c 0.6m wide, 0.2m deep) and neither contained any finds.
Pits
7 pits or probable pits were revealed with trench 9 (57-59, 61, 66, 68, 72), ranging from 0.56 to 1.86m in width and 0.08 to 0.2m in depth. None of these features contained any finds, and it is possible that some of them (e.g. 58) represent tree throws.

Top of Page

Trench 19 (Fig. 5.2.7: Trench 19)

Lying in the south-eastern corner of the site, trench 19 (72.7 m²) was largely obscured by volumes of standing water, but did reveal a number of ditches, gullies and pits, with virtually all datable materials confirming a 2nd century AD date.

Ditch 9/80
At the far south-western corner of trench 19, a short section of ditch (9/80) was partially excavated, although precise dimensions were difficult to ascertain due to waterlogging. The ditch was seen in the salvage area to continue curving round to the northeast towards the river and seemed to form part of a large ‘inner’ enclosure. It is uncertain how it would have related to ditch 16 further west, but the recovery of late1st-early 2nd century pottery suggested that the two features could have been contemporary. Large quantities of 1st to 4th century AD finds were recovered by metal detecting from topsoil in areas to the east of this ‘inner’ enclosure.

Ditches 32, 73, 74, 31
Lying just to the east of ditch 9/80 was ditch 32, orientated NW-SE. It was quite substantial in size, measuring 5.1m across, although the depth could not be discerned. To the east of this were three further ditches (73, 74, 31), all aligned north-south. Ditch 31 furthest to the east was 2.8m wide, 0.48m deep and cut through ditch 74 (1.22m wide, 0.48m deep). The only finds from these features comprised a small amount of 2nd century pottery from 32 and 31.

Ditches 79, 81
In the north-western part of trench 19 were two ditch sections running NE-SW. Ditch 79 (1.3m wide, 0.3m deep) contained 2nd century pottery, including samian dated after AD 150, and would seem to be a continuation of the ‘inner’ enclosure located further south (9/80). Ditch 81 to the east (1.6m wide, 0.06m deep) was very shallow and contained late 1st to early 2nd century AD pottery. It is uncertain if this feature was a ditch, or a broad shallow pit. Both features contained rubble and tile in the top fill.

Rubble layers (75 - 77)
In the centre of trench 19 lay three areas of rubble lying over a boggy depression (85). The only associated feature was a posthole (84), found under layer 75 and cutting the depression. Late 1st to 2nd century pottery was found within the rubble layers, along with a copper alloy loop and modern glass. To the north of the rubble layers were a series of ditches, but these were obscured by silting and were not excavated.

Top of Page

Aisled Building Postholes

Cxt Number

Measurements

Cut

Re-cut?

Fills

104

D: 0.35m, E-W: 1.25m, N-S: 1.75m

square, slightly rounded base

Yes

Good limestone packing

105

D: 0.37m, E-W: 0.78m, N-S: 1.35m

curved base

Yes

Limestone packing

106

D: 0.39m, E-W: 0.7m, N-S: 1.4m

square, rounded base

Yes

Limestone packing

107

D: 0.4m, E-W: 0.8m, N-S: 1.2m

rounded sides, flat base

Yes

Limestone packing

108

D: 0.35m, E-W: 0.5m, N-S: 1.35m

rounded sides, flat base

Yes

Limestone packing

109

D: 0.34m, E-W:?, N-S: 0.3m

rounded sides, flat base

Yes

Limestone packing

281

D: 0.38m, E-W: 1.1m, N-S: 0.95m

Rounded sides, flat base

No

Limestone packing

110

D: 0.3m, E-W: 0.8m, N-S: 1m

Rounded sides, curved base

No?

Limestone packing

268

D: 0.33m, E-W: 0.6m, N-S: 0.6m

Near vertical sides

No

No limestone packing

111

D: 0.27 m, E-W: 0.7m, N-S: 0.7m

Near vertical sides

No

Limestone packing

266

D: 0.45m, E-W: 0.9m, N-S: 1.13m

Steep sides, flat base

Yes

Limestone packing

283

D: 0.45m, E-W:0.75m, N-S: 0.9m

Nearly vertical sides

No

Limestone packing

289

D: 0.37m, E-W: 0.55m, N-S: 0.55m

Steep sides, flat bottomed

No

No limestone packing

264

D: 0.35m, E-W: 0.78m, N-S: 0.97m

Nearly vertical sides

Yes

Limestone packing

256

D: 0.32m, E-W: 0.95m

Vertical sides

Yes

Limestone packing

282

D: 0.33m, E-W: 0.75m, N-S: 0.75m

Rounded base

No

Limestone packing

274

D: 0.49m, E-W: 0.8m, N-S: 0.9m

Vertical sides, flat base

No

Limestone packing

257

D: 0.57m, E-W: 1.06m, N-S: 0.99m

vertical sides, curving base

No

Limestone packing

247

D: 0.59m, E-W: 1.1m, N-S: 1.05m

near vertical sides, rounded base

Yes

Limestone packing

246

D: 0.48m, E-W: 0.9m, N-S: 1m

near vertical sides

Yes?

Limestone packing

267

D:0.34m, E-W: 0.8m, N-S: 1m

rounded sides,

Yes

Limestone packing

235

D: 0.46m, E-W: 1m, N-S: 1m

rounded sides, rounded base

Yes?

Limestone packing


Top of page