Silchester Town Life Project - Late Roman Insula IX Interactive Matrix Diagram
The Late Roman website is designed to give structured
access to the primary archive of the insula IX excavation
in relation to the late Roman occupation, from AD250 to
the abandonment of occupation in the insula in the fifth
or sixth century. Click on items in the Index to explore the
This website presents the evidence for the late Roman (late
third century to the sixth) century AD) occupation of
insula IX. The insula is located immediately to the north-west
of the administrative centre (forum basilica) of the civitas
capital of Calleva Atrebatum (Silchester, Hampshire). The
excavation area measures 55 by 55m and incorporates the main
east-west and the main north-south street of the town as well
as a number of buildings. The substantial stone buildings have
already been excavated once by the Victorians but modern
excavation techniques allow us to put these into context by
uncovering the associated backyards and timber buildings.
Parts of the two main north-south and east-west streets
running through the Roman town border Insula IX and appear
along the eastern and northern edges of the excavation
trench. The street grid appears to have been constructed
in the mid first century. Many of the early buildings
continued to be aligned with the earlier Iron Age town
plan and it is only the late Roman buildings that are
aligned with and face directly onto the Roman streets.
The streets were made of thick deposits of gravel and sand
and were repeatedly resurfaced both before and after a
widening in the late Roman period.
The five buildings or properties aligned to the main north-
south street are numbered, from north to south, Buildings
6, 1, 5, 7 and 8. We have retained and extended the dual
suite of numbers for Houses and Blocks from the 1893-4
excavations, only substituting the term 'Buildings' instead
of 'Blocks'. So, while House 1 remains thus, Block 1
becomes our Building 1. As a total of four 'blocks' were
excavated across the insula as a whole in 1893-4, the
adjacent building to the north of Building 1 is numbered 5.
All further structures, whether of stone or timber, are
numbered sequentially from 5.
The 'invisible' boundary defined by the two parallel, but
discontinuous east-west rows of pits which run across the
excavation area (and beyond) separates the southern area
with Buildings 1 and 5 from the remainder of the excavation
area south of the east-west street. In the backyards of
Buildings 1 and 5 (and Building 6 beneath the southern
trench edge) are a large number of cut features each
relating to different phases of occupation of the
Building 1 is a rectangular stone structure located in the
south-east of the excavation trench. It has several phases
which date from its construction in c.300AD through to its
abandonment in the 5th or 6th centuries. The evidence
suggests that Building 1 had an upper storey and that
throughout its lifetime it consisted of both residential
and working accommodation.
Building 1 was constructed c.300AD. In its first phase it
consisted of a rectangular building divided into two
approximately equal halves. The western half was further
subdivided into at least 3 rooms, and the eastern half
was a single space probably used as a workshop. The
remains of two possible hearths here suggest this. The
eastern half of the building had no southern masonry wall,
but a single post-hole here may have supported the roof and
allowed access to the yard beyond. There was no access
from the street at this stage but entrance could have been
gained via a gate giving access to the yard to the south.
It is thought that the building had an upper storey with
a roof made of stone slate or possibly thatch.
In Phase 3 Building 1 was extended to the north by the
addition of two 'tower' rooms at each end of the north
elevation of the house linked by a corridor or verandah.
The projecting rooms extended the width of Building 1 by
about 4 metres, with the linking corridor or verandah set
back from their leading north wall by about 1.5 metres.
Although there is no stratigraphic evidence to support
the suggestion that the northern extension is later than
the original 'core' of the building, the difference in
the character of the foundations allows us to suggest
In the late fourth century an accumulation of evidence
suggests that Building 1 undergoes a phase of rebuilding.
The evidence for this is a mixture of stratigraphic and
interpretive and it does suggest that at least the main
north wall was rebuilt at this late date. This later
rebuilding contained at least one large hearth,
suggesting that the building's original function as a
workshop continued into the late fourth century and
beyond. It is also apparent that as well as a sequence
of domestic occupation revealed in the eastern projecting
tower room of the extension, there were a number of pits
dug inside this late surviving building, respecting its
Building 5 is a rectangular building c. 18m long, located
to the north of Building 1. Like Building 1, its shorter
end wall fronts on to the main north-south street, and its
manner of construction is also reminiscent of that of
Building 1. It is thought to have been a single storey
building constructed in the early 300s. It was divided
internally into a larger eastern space, possibly a workroom,
and a smaller western room, possibly living accommodation.
A line of post-holes suggest that Buildings 1 and 5 may have
been linked by a fence-line next to the north-south street,
incorporating a gate leading via a gravel path or yard in
to the insula beyond.
Pot and coin evidence suggests that Buildings 1 and 5 may
have continued to be occupied in some form into the 5th
century or beyond. Pits dug through the foundation trenches
of the buildings mark their final abandonment. Coin evidence
suggests that the eastern projecting room of Building 1
remained in use into the 4th century and a similar case can
be made for Building 5. It is thought that the well in the
shared backyard of the buildings, in which the discarded
ogham stone was buried, was not backfilled until the 5th or
6th centuries. The differential pattern of robbing of the
foundations of Building 1 suggests that its extension may
have been robbed first.
In the northern area, the concentration and disposition
of a range of features is consistent with at least two
timber-framed buildings continuing the built-up frontage
from Building 5 northwards along the north-south street
to its intersection with the east-west street. Other
features, including pits, post-holes and possible wall
trenches, suggest the possibility of up to four further
properties fronting on to the east-west street.
Building 7 is a small, rectangular timber-framed building
immediately adjacent to Building 5, with its narrow end
facing the street. It was constructed after Building 5 and
dates to the early to mid 4th century. The eastern limit of
this building is marked by a parallel shallow cut edging
the north-south street, but there is no clear evidence for
the opposite end wall. Two stone post pads define its
northern wall, and its southern wall may have been shared
in part with the north wall of Building 5. The latest
occupation deposits associated with Building 7 consist of
dark soils containing a spread of oyster shells.
The evidence for Building 8 is both confusing and slight.
A hearth with adjacent cuts and post-holes (both earlier
and later than the hearth) suggest that a building or
buildings, possibly of several phases, occupied the
north-east corner of the insula during the fourth and
into the fifth-sixth century. There is evidence to
support a structural phase later than 364-78. The
function of later post-pits cutting the streets at
their intersection and a ditch cut through the surface
of the street may have been as much to drain water away
from a building occupying the street surface as one
within the angle of the streets.
House 1 was cut by several pits both single and in groups,
including two interrupted rows running east-west. Five of
these, including the pit with the ogham stone, had been
excavated in 1893. The original digging of these pits
provides a terminus ante quem for the final abandonment of
House 1. The two rows of east-west pits extend the line
projected by the north wall of Building 5 and divide the
excavation into two halves. Two further pits excavated
outside our trench in 1893 extend the line further across
the whole of the insula. The majority of the pits contained
pottery (and coins) indicative of a fourth century date.
Within both halves of the excavation area, it has been
possible to identify properties within which each of the
An 'invisible' boundary defined by two parallel but
discontinuous east-west rows of pits separate the
excavation into two areas south of the east-west street.
In the southern half we have identified two major buildings,
Building 1 and 5. The evidence shows that each of these
buildings fronted directly onto the north-south street and
that the area between them was fenced off from the street
with a gate leading on to a gravelled yard or path extending
to the body of the insula beyond. When first constructed it
is possible that each building had its own backyard, however
at a later stage we argue that they were amalgamated into
a single property.
The northern properties and their associated divisions are
very indistinct and have been the subject of some
reinterpretation. The evidence suggests two structures
fronting on to the main North-South street in the Late
Roman period. A large property associated with Building 8
included both Building 7 and a well. The rest of the
northern zone was occupied by a large plot associated with
the building identified from aerial photography in the
north-west corner of the insula. Dark soils in this area
suggest open ground, possibly used as paddock or orchard,
or for cultivation of some kind.
In the backyards of the late Roman houses were a whole
series of rubbish pits and wells. Wells are generally
deeper than rubbish pits but towards the end of their
use they would also often have been filled with rubbish.
The finds from these features can tell us a great deal
about what life was like in Insula IX.
Some of the rubbish and cess produced by the inhabitants
of Silchester may have been put out on the surrounding
fields as night soil, but a lot of material, including
animal bones, broken pottery and building rubble, appears
to have been dumped into pits in the immediate vicinity
of the houses. Environmental analysis of the pits has
revealed imported plants and fruits, and weeds. One pit
contained rat droppings. Some pit alignments may relate
to the ways in which the backyards were laid out. For
example, two rows of parallel pits probably respect a
fence line which originally ran between them. Some groups
of pits are located adjacent to a building, while others
do not appear to be associated with any structure.
Three groups of pits, 19 in total including at least two
cess-pits, have been identified in the backyard of
Buildings 1 and 5. They are all on an east-west
alignment and define the southern side of the boundary
dividing the southern half from the northern half of
the excavation area. One group of four pits cut through
the foundations of House 1. All the pits were filled by
the mid 4th century and contained assemblages of pottery
and animal bone as well as specially placed deposits.
Three groups of pits have been identified in the
northern zone. The first comprised the pits that
extend along the proposed east-west boundary between
the northern and southern areas. The second is a small
group of mostly deeper pits which extend across the
northern zone. To the north of these, the third group
includes shallow pits and the late fills in the tops
of earlier pits where subsidence has taken place.
There are a total of four late Roman wells excavated within
Insula IX ranging in depth from 2.35m to 2.72m. Wells not
only provide essential clues about property boundaries,
layout and sharing, they also help date the construction
and destruction of the buildings to which they relate.
One group of pits appears to have been cut at the very
end of the Roman occupation in Insula IX. These pits are
cut into the Roman roads and through the late Roman
buildings, suggesting that both the roads and some of the
buildings had ceased to be used. Some of these pits
contained coins dating to the very end of the
4th century AD.
The associated finds pages introduce the viewer to the
catalogues of the various categories of finds stored in
the database which are associated with the Late Roman
occupation of insula IX. Some of these are illustrated
Small finds comprise all those finds which were individually
registered at the time of the excavation. Apart from
coins which are introduced separately these include objects
in a variety of materials: Amber, antler, bone, copper alloy,
gold, glass, iron, jet, lead, shale and stone.
The coins struck after AD 250 and recovered from the
excavations of insula IX are entirely of copper alloy. Such
coins represent the small change of the period. The latest
issues are of the period of 388-402. The database records
244 coins struck after 250 from the area we have excavated,
while a further 227 were found in periodic searches of the
topsoil after it was removed by mechanical excavator.
In addition 66 coins struck from Republican times to the
early third century were also found and are recorded in
More than 13,500 sherds weighing some 180kg have been
recorded from the late Roman occupation of insula IX
including a few complete, or near-complete vessels. Pottery
produced in the period between c. AD 250 and 400/25 is
represented by: vessels for the table (including cups,
beakers, bowls, dishes and flagons) decorated in red
(oxidised) and black (reduced) slips, and vessels for
cooking (including jars, bowls and dishes) made in reduced,
grey or black sandy fabrics with similar colouring to the
A total of 8,480 fragments of animal bone were recovered
by hand collection from the late Roman pits of which 2,601
(31%) are identifiable to order or species. A further 10,500
bone fragments from the remaining late Roman contexts
were also assessed.
Several human neonate bones, representing perhaps seven
individuals, were identified from four late Roman pits
and one other context. Most of the bone, probably
representing just one individual, was found in a single
pit 3251, close to Building 1. All the bones belonged
either to premature or stillborn babies of less than 37+/-
weeks, or to neonatal or full-term babies of 38+/- weeks.
Silchester is very fortunate to have environmental
preservation by waterlogging, charring and mineralisation.
Examination of the seed and insect assemblages has
provided detailed profiles of the surrounding vegetation,
diet and ultimately the level of urbanisation Silchester
had reached by the late Roman period.
As the most durable residue of the metalworking process,
slag in all its various forms is essential to our
understanding of industrial processes that may or may not
have been occurring in late Roman Insula IX. Quantification
and analysis of the slag has given us important insights
into both the types of metalworking that were carried out
and also to what extent.
The late Roman hearths of Silchester vary greatly in size,
composition and function. Analysis of the structures and
associated contexts implementing chemical,
micromorphological and mineralogical methods has
significantly aided our understanding of these features.
The hearths are fundamental in helping us interpret the
buildings to which they relate.
The worked stone from Late Roman Insula IX includes a
range of objects typical of assemblages from this period
including rotary querns and whetstones; roof, floor and
other building stone.
The majority of the building material recovered from the
late Roman contexts was either ceramic or flint. There was
also a small quantity of stone in lithologies other than
flint. There was no window glass of characteristic late
type (see Glass ). Nails and other structural objects are
reported under Small Finds (Structural Materials).
Altogether fragments belonging to perhaps 135 different
glass vessels were recovered from late Roman contexts,
but only some 12 fragments show characteristics which
might suggest a later third of fourth century date. The
majority of the identifiable vessel fragments are of 1st
and 2nd century date and mostly from bottles. These are
likely to be residual, re-deposited in late Roman contexts.