Department of Archaeology

Silchester Insula IX

The ‘Town Life’ Project 2011 Michael Fulford, Amanda Clarke & Sarah Lucas Our fifteenth season of excavation The late Iron Age on Insula IX in July and August 2011 saw excellent progress in uncovering c. 40 BC – AD 44 Until the ‘Town Life’ Project, the earliest evidence evidence for the occupation of Calleva for the occupation of the Calleva promontory came in the late Iron Age and earliest from the 1980s excavations beneath the forum Roman period up to the episode of basilica in the heart of the town. This produced no finds to suggest occupation before 20/10BC. destruction which we are associating However, historical evidence provides the context with the period of the Boudican for an earlier date for the origin of Calleva. The British are assumed to have originated revolt of AD 60/1. The Iron Age from a Gallic Atrebates settled in northern France. evidence is gradually being revealed, Their leader at the time of the wars between Rome as we work across from the west to and in the 50s BC was Commius. the east in the north and south of the trench. Significant structures and extensive deposits of earliest Roman occupation continue to be excavated above the Iron Age occupation across the eastern half of the trench.

Commius coin distribution in relation to Calleva

1 Once an ally of Rome, he fell out with and fled to Britain with his retinue in about 52 BC. The archaeological evidence for his presence in Britain is represented by coins carrying his name which have been found in small numbers in central southern England. These do not contradict the possi- bility that Commius was at Calleva, but to date there has been no firm evidence for settlement as early as 50 BC. However, three important developments emerged during the course of the 2011 season:

Chronology The late Iron Age ditch which we began to excavate in 2010 has been explored further. It extends north- west from the southern edge of the trench for 40m before its course is obscured in a complexity of occu- pation. It is V-shaped in profile with a depth of about 1.0 metres and there is an ‘ankle-breaker’ cut at the bottom. While there remain uncertainties about how the ditch behaves at its northern end, it is now clear that it is the earliest significant feature of the Iron Age occupation. What its purpose was, whether it represented an early division of land across the Calleva promontory, or it formed part of a defensive enclosure, is unclear, but it was a major feature of the landscape.

Finds are sparsely represented in its lower fills, but the upper deposits are rich in the pottery and other finds which are consistent with the earliest finds from the excavation beneath the forum basilica in the 1980s. These could not be dated earlier than about 20/10 BC. Since the ditch in Insula IX had largely silted up before rubbish was deposited in it, the date of its creation is sometime earlier than about 20-10 BC. How much earlier is difficult to esti- mate, but some 10-20 years, at least, seems plausible. By extrapolation this takes the date of its initial excavation back to about 40 BC.

Developments after the silting of the ditch have to be considered from the south-west and north-west quadrants of the excavation separately. While the trend of development in both areas is the same with each sharing the same broad chronology, we cannot yet impose a single, site-wide chronology until we have excavated more and linked the two areas.

Cover image: Kite photo by Geraint Jones. Left: Looking north-west across the excavation towards the end of the season. Right: Photo highlighting the Iron Age ditch (looking south-east).

2 South-west By the turn of the 1st century BC/1st century AD the N ditch seems to have lost its importance or to have been forgotten. A large timber building with late Augustan pottery was constructed across its course and at right angles to it at the southern end of the trench. Its western end is adjacent to, and at right angles to a linear slot oriented north-west/south-east and extending some 23.5 metres across the trench. There are traces of a parallel structure to the west and the pair appears to define a street or lane, some 3.5 metres in width, of the late Iron Age town. We interpret these slots as the construction trenches for fences which were in due course replaced by more robust structures employing substantial posts, but maintaining a similar width of 3.5 metres for the street. The rear wall of the building is hard against, if not actually part of the eastern row of posts.

Gravel surface Beam slot or fenceline Post hole 0 20 metres

Well Beam slot Post hole Pit with charcoal-rich fill Pit Ditch or gully Clay floor Emerging clay floor Cesspit Gravel surface 0 20 metres

Aerial photo showing NW-SE Street and Iron Age building, looking north-west.

3 In addition to the major building the area to the east of the street contains a number of rubbish pits filled with pottery and other finds of Augusto-Tiberian date. While some to the north respect the line of the construction trench, others clearly lie within the pre- sumed footprint of the building. However, the latter group also appears to respect the (shallower) con- struction trench of a possible second timber building constructed on a slightly different alignment. While more work needs to be done on the relationships between the two buildings and with the rubbish pits, and on their chronology, the 2011 season revealed evidence of a further post-built structure which ran, more or less at right angles, across the line of the more prominent of the two buildings. Further excava- tion will clarify whether this is part of a building or a post line. Thus, in the south-west quadrant, in addi- tion to the complexities associated with at least two distinct phases of development of the street, there is a sequence of three timber structures, of which two at least appear to be the remains of buildings. Further work on the finds from the rubbish pits may help to refine the chronological sequence, but it is clear that the vast majority of the features excavated in the south-west area in 2010 and 2011 are pre-AD 43 in date.

North-west The complexities of the development of the site in the late Iron Age are further borne out by discoveries in the north-west quadrant of the trench where excava- tion beneath the north-east/south-west-oriented street or lane has revealed evidence of another, substantial rectangular timber building, as well as other struc- tures. Although we do not yet have dating evidence for this structure, its presence indicates that the setting out of the north-east/south-west lane was secondary to a phase of occupation and, very prob- ably, secondary to the north-west/south-east street, where we have not yet found evidence of any underly- ing buildings. This is supported by the discovery in 2010 of a cluster of pits containing Augusto-Tiberian pottery right at the point where street and lane are projected to intersect. The 2011 season also revealed a spur, defined by shallow, parallel ditches, 3.5 metres apart, of the north-east/south-west lane. It ran off at right angles towards the centre of the excavated area. Whether it is contemporary with or secondary to the creation of the north-east/south-west lane is yet to be established. Thus in the north-west we have evidence of an Iron Age building, possibly associated with the cluster of early rubbish pits, before the lane and its associated spur, which strikes towards the Photo showing NW lane, looking north-east.

4 1 north-east, was laid out. While evidence is emerging of at least two phases of ditches defining this lane, Claudio-Neronian (c. AD 40-60) pottery has been identified in some of the later ditch fills which flank the north-east trending lane. The overall pattern is of increasing density of occupation reflected in the subdivision into smaller properties by the creation of new lanes and boundaries.

The makings of a sequence While the course of the north-west/south-east street runs slightly obliquely to that of the earlier ditch it seems to have continued as a key structure of the late Iron Age settlement, respected by the earliest build- ings and associated rubbish pits of Augusto-Tiberian Well Beam slot Post hole Pit with charcoal-rich fill Pit Ditch or gully Clay floor Emerging clay floor date in both the north-west and south-west areas. The Cesspit Gravel surface 0 202 creation of the second lane running to the north-east metres and its associated spur indicate an intensification of occupation within the excavated area in the second quarter of the 1st century AD.

1. c. 40-20 BC: ‘Early’ defensive or boundary ditch 2. c. 20-10 BC: Establishment of north-west/south-east street 3. c. 20 BC – AD 20: Rectangular buildings and associ- ated rubbish pits in the south-west and north-west 4. c. AD 20-44: North-east trending lane established at right angles to ‘axial’ north-west/south-east street

N 4 Gravel surface Beam slot or fenceline Post hole 0 20 metres 3

Well Beam slot Post hole Pit with charcoal-rich fill Pit Ditch or gully Clay floor Emerging clay floor Cesspit Gravel surface 0 20 metres

Pit Beam-slot/fenceline Gravel surface Post hole Ditch/gully 0 20 metres

Pit with charcoal-rich fill Well Beam slot Post hole Pit with charcoal-rich fill Emerging clay floor Pit Previous phase Clay floor Emerging clay floor 5 Cesspit Gravel surface 0 20 metres for buildings is that much work remains to be done Earliest Roman in excavating the remains of the succeeding, civilian phase of the town, our Period 1. However, work began c. AD 44 – c. AD 60/1 on the excavation of a sub-rectangular pit adjacent to the Roman north-south street. The character of this Chronological framework pit is reminiscent of that excavated adjacent to the east-west street in 2009 and provisionally interpreted The period up to the destruction of our first Roman as a military-style latrine pit. A further, sub-rectangu- phase (Period 1) is provisionally divided into a mili- lar subsidence parallel to the north-south street and tary period, whose chronology is shaped by the about 10m south of the pit under excavation is indica- general perception of the progress of the Roman tive of a yet larger pit. It awaits excavation following conquest of Britain, followed by civil development. completion of work on the overlying timber building. Significant developments occurred under the second governor Ostorius Scapula (AD 47-52) represented by the withdrawal of the XX legion from Colchester to lead campaigning against the Silures of south-east Wales and the establishment of the colonia around AD 49. Although the Roman army invaded south- east Britain in AD 43, we put the beginnings of intervention at Calleva in the following year. Den- drochronological dating of the construction of a fort at Alchester, some 35 miles due north of Calleva, to AD 44 shows that the Roman army was reaching into the heart of lowland Britain within a year of the initial landings. The end of the military phase at Calleva is provisionally put at c. AD 49 in line with the redeployment westwards of the legion from Colchester and the start of the civil town there, and the assumption of a general shift westwards of Plan showing the Iron Age earthworks of Calleva, the north-south street, Iron military resource around this time. However there Age lanes aligned north-west/south-east and north-east/south-west found in Insula IX and under the forum basilica and the location of the possible Claudian is no independent evidence for the date of either military structure under the basilica. the beginning or the end of the military phase at Calleva. It cannot have been for very long because it Military equipment is becoming clear from Insula IX and elsewhere that the town was densely populated and prosperous by If we so far lack military buildings in Insula IX the 50s of the 1st century AD. to compare with the large, possible headquarters building beneath the basilica, the number of pieces of military equipment continues to grow. The season added a further seven items of arms and armour Conquest period to bring the total to about 25 altogether from the c. AD 44 – c. AD 49 excavation. Some of our distinctive, earliest Roman brooches are also likely to have been worn by soldiers. What did the military phase look like? While the season added more finds with a military association to the assemblage from Insula IX, no further evidence of buildings emerged to complement that already reported, namely post-holes aligned at right angles to the line of the north-south street in the north-east quadrant of the trench. This is important because it shows that the origin of the Roman grid oriented on the cardinal points, north-south, east-west, is very 2011 Military finds. Left to right: Bagendon brooch; hinge bar from Lorica Segmentata strap buckle; early. The principal reason for the lack of evidence stud; strap mount.

6 Iron Age buildings

The remains of at least three rectangular or sub- N rectangular buildings and one D-shaped structure have so far been identified: 1

1. South-west: a clear linear slot, some 0.48m in width by 0.20m in depth, extending some 11.2m in length seems to define one side of a building. A shorter length to the north may represent its other side with a further slot acting as an internal divi- sion of the building. If all these elements are part of the same, rectangular building it measures at least 5m by 11.2m, but further work is necessary to determine whether its limits have been reached.Well Beam slot Post hole Pit with charcoal-rich fill 2 Pit Ditch or gully Clay floor Emerging clay floor 2. South-west: the substantial construction trenchCesspit Gravel surface 0 20 of a sub-rectangular building with a rounded end metres was identified. The trench, 0.4m in width and 0.4m in depth, is neatly cut. The building extends at least 11m to the north-east but neither its other end nor its southern side have been identified. The Remains of dog articulated remains of a miniature dog were found buried, presumably as a foundation deposit, at the

one ‘corner’ of the building so far to be identified.Well Beam slot Post hole Pit with charcoal-rich fill Pit Ditch or gully Clay floor3 Emerging clay floor 3. North-west: remains of a rectangular timber Cesspit Gravel surface 0 20 building, of which the construction trenches of one side wall and one end wall appear complete, metres were found beneath the lane. The northern end and side walls were disturbed by later cuts and a further slot, not quite at right angles to the side wall, and also cut by later features, may have divided the building into two rooms. The building measured 7.5m by 4.5m. 4

4. North-west: adjacent to the above was the con- struction trench of a D-shaped structure with a maximum width of 2.5m. There was no evidence to help with its interpretation.

N

0 20 metres

Beam slot 0 20 Well Beam slot Post hole Pit with charcoal-rich fill metres Ditch or gully Clay floor Emerging clay floor 7 Pit Cesspit Gravel surface 0 20 metres

Well Beam slot Post hole Pit with charcoal-rich fill Pit Ditch or gully Clay floor Emerging clay floor Cesspit Gravel surface 0 20 metres Northern buildings (looking east). Southern buildings (looking south-west).

Close-up of southern buildings (looking south-east).

8 Calleva of the Client Kingdom c. AD 49 – c. AD 60/1

Some eight or nine timber buildings have now been identified within the excavation trench which can be associated with the first phase of the Roman-period town. All but one of the buildings has been identi- fied on the basis of their distinctive clay floors. More difficult has been the recognition of clear traces of structural elements such as beam slots or post-holes; thus all the buildings appear to have been relatively slight in their construction and of single storey. Three rectangular buildings are at right angles to, and adja- cent to the line of the north-south street. Two of these with remains of hearths re-using Roman tile underlie Period 2 buildings 5 and 8. To the north, remains of a third building, close to the line of the north- south street, have been identified. More central to the excavated area is a group of three buildings, one circular, one definitely rectangular and the largest of the group, the other probably rectangular, but significantly reduced by later pitting. The orientation of these buildings has been hard to establish but, as its excavation has developed, the largest appears to

Early Roman timber buildings beside north-south street.

have an orientation more north-east/south-west than east-west as originally perceived. It is from the latest contexts associated with this building that the figure of Harpocrates, detached from a fancy, bronze Cam- panian drink-warmer (below and above right), was recovered. It is to these buildings that the spur heads from the north-east trend- ing lane. To the north lies a further, sub-rectangular clay-floored building and, adjacent to it, is a penan- nular ring gully, a possible indication of a second round-house. Traces of a further, clay-floored building have been preserved in the slumped fills of a rubbish pit close to the line of the east-west street and adjacent to a very

Well Beam slot or fenceline Post hole Pit with charcoal-rich fill large rubbish pit dated to this period. Pit Ditch or gully Clay floor Cesspit Gravel surface 0 20 metres Simplified plan of Period 1

9 In addition to the traces of buildings, isolated wells and adjacent hearths are also indicative of further buildings whose floors and structural remains have been destroyed by later activity. Thus two wells suggest further buildings in the north-west of the excavated area, taking the total up to about ten. The remaining three wells, shown on the plan, cluster around the central group of buildings, of which the largest, with its associated finds, appears to be rela- tively high status. Although these buildings do not directly underlie the Period 2 town house and associ- ated buildings, their character does suggest that they are, in essence, their predecessors.

Hearth Pre-ERTB 8 Harpocrates (drawn by Brian Williams)

Hearth Pre-ERTB 5

Campanian drink-warmer, showing position of the Harpocrates figure. Oven Pre-ERTB 8

10 Although the location of buildings suggest that the the north-east trending lane in the north-west corner north-south street remained a thoroughfare during of the trench. The pits in both Insula IX groups are the life of the town in this period, it is also clear of an order of magnitude greater in size compared that the Iron Age streets and lanes remained in use. with those which flanked the equivalent Iron Age Their additional role as boundaries between plots streets discovered beneath the basilica in the 1980s. and properties within the town is emphasised by the Depths over 1.0m were exceptional among these pits. excavation of large and, sometimes, deep circular pits While the latter clearly functioned as rubbish pits, this is not so obviously the case with the Insula IX pits whose fills were generally much less rich in finds. Whether their purpose was more to reinforce bound- aries, or, assuming that they were all excavated at the same time, even, in the first instance, to act as simple defensive features, is unclear.

To conclude, Insula IX indicates a community of relatively small, but densely packed, single-story, timber buildings during the 50s of the 1st century AD. While Roman ceramic building material is in evi- Section through one of the northern pits. dence, mostly re-used in hearths, the tile derived from buildings elsewhere in the town, our buildings were along one side, that is the outside from the perspec- probably all thatched. The finds, notably the ceramics, tive of the property, which appears to occupy most continue to show a wealth of links across the Channel of the excavated area. The last pit of the row within to Gaul (France), Spain and the Mediterranean. As well the excavation trench and flanking the north-west/ as wine, olive oil and fish sauce imported in ceramic south-east street was excavated. Although it had been amphoras, remains of large wooden barrels, which excavated to a depth of 3.10m, such that it might have could carry 8-900 litres, attest to the bulk importa- originally served as a well, it produced no evidence tion of liquids, particularly wine, as well. However, of a lining or of any waterlogged remains from the despite the strong commercial feel of the town from rapid sediments at the bottom. Indeed there were the material finds, preliminary work on waterlogged very few finds altogether from the lower fills. Match- and charred plant remains shows that there was also ing the group in the south-west corner of the trench, a strong agricultural component to town life with further examples of what constitutes a second row of animals stalled within the town and grain brought in large pits were excavated along the northern side of for processing from nearby fields.

Northern pits from above.

11 Iron Age foundation burial in Germany and we have found remains of a very rare donkey-mill which we illustrated in our report for 2009-10. In 2011 we recovered two examples of another rare type made of Puddingstone, a very hard flint-pebble conglomerate, a rock extremely difficult to shape and drill through, which appears at the surface fringing the London and Hampshire basins. Identified by Chris Green, these are the first examples ever to have been recorded from Calleva and the better preserved is made of the well known Hertfordshire Puddingstone. The smaller fragment probably originated from outcrops around Swans- Dog burial under excavation. combe in north-west Kent. Apart from examples in Buried at one end of the large Iron Age house on the north Kent, the majority of finds of Puddingstone south side of the excavation trench was the skeleton querns are from East Anglia, but especially from of a dog, the remains placed to replicate a normal, Hertfordshire and Essex. While the Kent fragment is relaxed, resting posture of the animal. Dr Kate Clark of late Iron Age date, the Hertfordshire quern dates estimates the dog was in its second or third year. to around the mid-1st century AD. With an estimated shoulder height of 29 cm, what is most unusual about it is its small size. Characteris- A rare brooch tics of the bones indicate similarity with modern toy breeds, such as the poodle, rather than short-limbed types such as basset hounds and dachshunds. With only half a dozen or so examples so far recorded across Britain as a whole, very small dogs of this period are extremely rare finds.

Puddingstone querns Rock types suitable to make rotary querns for grinding corn into flour are rare in south-east Britain. Most of the Iron Age and early Roman querns we Line drawing of complete brooch, Nertomarus brooch fragments & the have found were made from a distinctive type of penannular brooch. Lower Greensand quarried at Lodsworth in West . However the Romans introduced querns Brooches of copper alloy or iron are among the and mills made of lava from the Eiffel Mountains commonest finds of metalwork that we recover from late Iron Age and early Roman occupation in Insula IX. The condition varies considerably from an almost ‘as new’ example as represented by the illus- trated penannular brooch to one which is in very poor condition, where the best preserved part is its ‘head’. However, appearances are deceptive, as Nina Crummy points out that these fragments are of an extremely rare find from Britain. They are from a brooch which was made in Eastern Gaul (France) and imported into Britain just before the Roman conquest of AD 43. The drawing shows a complete example with the maker’s name, Nertomarus, stamped on the cover of the spring. That this would have been a prestigious piece for the Callevan wearer is indicated by the find of two of these brooches in the richly fur- nished grave of a Catuvellaunian warrior.

12 Traces of buildings or fills of rubbish pits?

With some of our poorly preserved buildings the only remains of their floors occur where they have subsided into underlying rubbish pits. We also find that such pits are sometimes sealed with clay to help conceal the smell of the decaying rubbish.

How then can we distinguish between the two very different purposes to which the clay was being put? One approach is to examine microscopically the profile of the soils which fill the pit. To do this a very thin (30 microns) slice of the soil profile has to be prepared, mounted on a glass slide. This can then be examined under a microscope. 5297 Top Units identified in thin section, and the contexts/ interpretation identified in the field

P

O Above : pit section shortly after micro- morph sampling, and the sample after being embedded with resin and cut. Below: The resulting slides after N sanding to 30 microns, with interpretive annotation. M – C10148 Levelling/ fill of pit Left: Cindy examining one of the slides 5297 Base under the microscope. L Units identified in thin section, and the contexts/ K – C11744 Occupation interpretation identified in the field deposit slumped into pit J – C11745 Occupation deposit slumped into pit I

H I

G

F – C10169 H G

E F – C10169 Clay capping D - C11960 Occupation Deposit E

D – C11960 C - C11959 Occupation deposit Occupation Deposit

In the example shown here, prepared by Cindy van B - Zwieten, you can see a sequence of fine, laminated C11962 - Pit floor surfaces with clear boundaries between them. A lack of materials, such as bone, charcoal or pottery fragments, indicates that the floors were kept clean, but regularly re-surfaced. Had the clay been dumped A - C11671 - Pit simply to seal the pit then the inclusions in the clay would appear unsorted and unorientated.

13 Iron Age food imports Dobunnic coin As part of a wider study of waterlogged plant remains A silver Iron Age coin of the western or Dobunnic from Late Iron Age and earliest Roman wells by Lisa series dating to about 20BC – AD10 was found in the Lodwick, several important identifications have Iron Age boundary ditch. One side shows a stylised been made of Late Iron Age plant head, the other a triple-tailed horse with an animal, foods. Samples from the base of perhaps a bird, above the horse. These circulated one well produced an olive stone mostly in territory to the north-west of Calleva, (below) and seeds of celery (left) centring on modern Gloucestershire and west Oxford- and coriander, all dating to pre shire. This is the first coin of this type from the town.

AD 43. A second well produced a Below: Excavating the Iron Age ditch, with the coin (inset). celery seed from a context dated to before AD 43 and several dill seeds dated to AD 40-50. Celery, coriander, dill and olive were all cultivated in the Medi- terranean region and literary evidence shows they were part of Roman cuisine. Whilst the import of olive oil and wine during the Late Iron Age is evidenced at Silches- ter and elsewhere throughout southern Britain by the remains of large pottery containers (amphoras), we were previously unaware that olive fruits and season- ings were also being imported. These new records from Insula IX suggest the Late Iron Age occupants of Silchester were selectively adopting foods and seasonings from the Mediterranean world prior to the Claudian invasion.

Above: Celery seed and olive stone.

Below: Removal of the layer covering the ditch.

14 Acknowledgements All staff and excavation participants 2011. The University of Reading and the Department of Archaeology, School of Human and Environmental Sciences. Hampshire County Council, Basingstoke and Deane Borough Council. The Calleva Foundation, The Englefield Charitable Trust, The Friends of Silchester, GML Ltd, The Headley Trust, The Horne Foundation, The Leathersellers’ Company, Nick and Biddy West.

The Silchester ‘Town Life’ Project

For information about the Field School, please contact: Amanda Clarke Field School Director University of Reading Whiteknights PO Box 227 Reading, RG6 6AB [email protected] Tel +44 (0)118 378 6255 www.reading.ac.uk/silchester

For more information about Archaeology at the University of Reading, please contact: Mrs. M.D.McGuire School Undergraduate Administrator [email protected] Tel +44 (0)118 378 8733 General Office [email protected] Tel +44 (0)118 378 8132 www.reading.ac.uk/archaeology