| At least
22 buildings are known from Margidunum dating from the first
to the fourth century. They were revealed during excavations
carried out in the 1920s by Felix
Oswald and in the late 1960s by Malcolm Todd. Some of
these buildings were represented merely as floors of beaten
clay whereas for others wall footings survive. In contrast
to sites in Italy where Roman houses survive still standing,
very few buildings of the Romano-British period survive above
ground level. This is due to a combination of neglect, collapse,
weathering, stone robbing and plough damage. An antiquarian,
William Stukeley, visited the site of Margidunum on the 7th
of September 1722. He records in his Itinerarium Curiosum,
published in 1724, that he saw “Roman foundations
of walls and floors of houses, composed after the manner spoken
of, of stones set edgewise in clay, and liquid mortar run
upon them: there are likewise short oaken posts or piles at
proper intervals... Houses stood all along upon the Foss,
whose foundations have been dug up and carried away to the
neighbouring villages... close by has been a great building,
which they say was carried all to Newark.”
Much has clearly been lost since the
Mediaeval period. Nevertheless it is possible to piece together
a picture of what the houses may have looked like from evidence
left behind by the stone robbers such as fragments of columns
and tiny scraps of marble veneer. Although we no longer have
the superstructure intact areas of fallen daub, broken up
painted plaster, window glass, and roofing tile all help us
to reconstruct the nature of the houses. Under-floor elements
such as hypocaust systems (the Roman central heating system)
provide additional evidence for the type of building excavated.
Some further light comes from evidence from preserved building
elsewhere in the Empire, from Classical accounts of building
techniques and from the excavation of other buildings in Britain.
The curving infilled gully
found at Margidunum is thought to have been made by
dripping water off the thatched roof of an early roundhouse.
Photo: M. Todd. Nottingham University Dept Archaeology
(No B1642)
|
A curving gully found in a trench excavated
within the defensive ditch of Margidunum compares well with
the sort of archaeological remains left by the round houses,
which were so common in the Iron Age settlements of the Trent
Valley and continued to be used well into the Roman period.
Given that the fine character of the finds suggests that the
earliest Roman phase of settlement at Margidunum is likely
to be military in character, this type of gully is more likely
to belong to a very early phase before the military arrived.
Very little evidence for pre-Roman occupation
of the Margidunum site was noted by Oswald or Todd, but excavations
south of the defences of Margidunum by Trent & Peak Archaeological
Unit revealed evidence for activity near the site in the late
Iron Age.
Excavation has revealed slots and gullies
associated with occupation in the area in the first century
AD. These slots, dug through the natural clay, were identified
because of their darker fill and are all that remains of the
slots dug to hold the timber uprights. In only one case did
the “ghost” impression of one of these posts survive.
Some of these slots may have held planks into which upright
post were then inserted or, in some cases, the upright posts
were put directly into postholes and the postholes were then
packed with other material. Some gullies may represent the
effect of water run off from the roof of a building and may
not be structural at all. A rectangular building with at least
two rooms, measuring 12 ft x at least 24ft, was excavated
by Todd and there were fragments of two more possible buildings,
one of which had a post impression.
Evidence for the superstructure and some
embellishment of these early buildings survived and were recorded
by Oswald, who found “charred beams ..and many fragments
of burnt daub” underneath one of the later buildings
he excavated (building L) and post-holes associated with floors
paved with slabs of skerry
stone. Oswald suggests that the site of this last building
had been carefully prepared by levelling with quantities of
sand, probably because it was a marshy area but alternatively
an area of ground covered by natural sand drift may have been
chosen by the builders. In contrast to this, the area occupied
by building L seems to have been prone to becoming boggy and
was abandoned at the end of the second or early in the third
century when it was sealed with sterile marsh silt. Although
none of the plans of these early buildings in the area of
building G excavated by Oswald can be recovered, his account
together with the quantity of first century pottery and plaster
associated with some sections of the earliest paved surface
indicates the possibility of a stylish early house on this
site. Todd points out in his review of Oswald’s excavations
that some of the earliest paving dates to the end of the first
century or as late as 125AD but the coin he uses as dating
evidence comes from below paving south of the complex of walls
and postholes so a later date for the paving has not yet been
demonstrated.
Slots and gullies that
may have been dug to hold timber uprights and planks
in a early building
Photo: M. Todd. Nottingham University Dept Archaeology
(No B1384)
|
To see the plan of Margidunum, showing
the buildings click here
When the Romans first arrived in Britain
most buildings were made entirely of timber, with the exception
of bathhouses in which uncovered timber would constitute a
fire risk. The existing evidence from Margidunum suggests
that buildings here were timber framed with an infill of wattle
and daub. Such buildings may have been thatched or covered
with wooden shingles. They were susceptible to destruction
by fire, as seems to have happened in some cases at Margidunum,
and a Roman writer, Vitruvius, writing in the first century
BC, advises: “As to wattled walls, would they had
never been invented, for though convenient and expeditiously
made, they are conducive to great calamity from their acting
almost like torches in case of fire. It is much better, therefore,
in the first instance, to be at the expense of burnt bricks,
than from parsimony to be in perpetual risk. Walls moreover,
of this sort, that are covered with plaster are always full
of cracks, arising from the crossing of the laths; for when
the plastering is laid on wet, it swells the wood, which contracts
as the work dries, breaking the plastering. But if expedition,
or want of funds, drives us to the use of this sort of work,
or as an expedient to bring work to a square form, let it
be executed as follows. The surface of the foundation whereon
it is to stand must be somewhat raised from the ground or
pavement. Should it ever be placed below them it will rot,
settle, and bend forward, whereby the face of the plastering
will be injured. II.8.20.”
In Britain wattle and daub houses are
thought to be by far the most common type, probably because
of the ready availability and cheapness of timber and the
relatively high cost of quarrying and preparing suitable stone.
Fragments of building debris associated
with pottery of this early period included a fragment of a
millstone grit arch or window and window
glass. Some complete roof
tiles in Nottingham University Museum are recorded as
coming from a well dating to the mid to late first century.
These fragments allow us to reconstruct at least one building
with stone elements, a tiled roof and window glass. This is
very likely to be a bathhouse since these commonly had glazed
windows to keep in the heat and tiled roofs to lower the risk
of a conflagration of the roofing material.
Click here
to see a reconstruction drawing of Roman bathhouse at Slack
Fort, Yorkshire.
Click here
for a reconstruction of how bathhouses worked.
Rather more is known of the buildings
dating to the second century. At least nine buildings were
erected during this time along the Fosse Way (itself dating
from the late first century) and all of these had stone footings,
made from the local skerry stones. Buildings with stone footings
are very common in Roman Britain and it is often difficult
to reconstruct the superstructure. Guy de la Bedoyere concludes
that most of these had timber-framed superstructures. The
bathhouses are likely to have stone superstructure because
of the fire risk in timber structures.
Five or six methods of construction were
used to build the stone-footed buildings:
1. In buildings G, H, I and J wall trenches filled with skerry
rubble loosely packed.
2. In building
C wall trenches filled with closely packed skerry lumps
between two rows of facing block of skerry.
3. In building D wall trenches were filled with pitched skerry
slabs. This technique was used to consolidate part of
the wall of building J where it ran over an earlier pit. A
second example of pitched footings with no mortar was found
beneath the town rampart.
4. In the undated building F a section drawn by Oswald records
a wall trench filled with stone slabs set on edge with mortar
at the bottom.
5. In building L Oswald found pitched skerry stones set into
clay with two courses of flat stone slabs remaining above
and in building M Oswald records that the wall trenches still
contained several horizontal courses. These may be more intact
examples of type 3.
6. The east wall of building L had been constructed slightly
differently with a thicker wall (c1foot wider) set in concrete.
This may have been constructed to take a heavier superstructure
such as a tall facade.
To see a plan of Margidunum showing the
buildings click here.
Although the Romans were skilled builders
and architects with instruction books, such as Vitruvius’s
manual mentioned above, and tools for surveying, building
work in small towns and rural settlements was sometimes rather
haphazard and even “bodged”. At Margidunum the
builders of at least two of these houses were forced to make
provision for possible subsidence because of encountering
the unstable infill of earlier pits. In both cases the wall
foundation trenches were deepened to bottom the pits and the
wall trenches were filled with carefully alternately pitched
skerry slabs
to prevent later subsidence and instability.
Rather more evidence for the appearance
of these second century buildings has been excavated. Some
buildings, such as those fronting the Fosse Way (buildings
H, I and J), had simple floors of beaten clay whereas other
(building G) had areas of skerry paving. These were probably
simple workshops and traders stalls with resident blocks at
the back or, possibly, in an upper story. These are common
in Roman roadside settlements. The occupants of Building C
aspired to rather more durable flooring: paving in some rooms
and elsewhere a mixture of mortar, clay and earth. This embellishment,
together with the neatly
faced wall footings may reflect a higher social status.
Remains of the floor of
Building C
Photo: M. Todd. Nottingham University Dept Archaeology
(No B1638) |
The building also had three rooms, contrasting
with the single apartment life style of the pre-Roman roundhouse
dwellers and some of the other inhabitants, and was set back
from the road. This house also has evidence for a back yard
and the convenience of a large water
tank situated just outside. Evidence from this building
is rather better than from some of the other houses but it
does seem to stand out as being rather more affluent.
The most elaborate buildings are L and
M (the bathhouse). Neither of these is likely to be a private,
domestic building, but they may have been used by travelling
Imperial officials as well as the townsfolk. Building L is
very large with no signs of subdivisions. Oswald recovered
evidence of a concrete surface covering the outside area to
the north and overlapping the foundations suggesting it abutted
a timber-framed wall of some sort. The interior was covered
with several phases of gravel floors. Wall plaster fragments
were found interleaved with these gravel floors indicating
this building was probably plastered. Painted fragments of
wall plaster from under the lowest floor identified by Oswald
tells us that this level of sophistication was present at
or near the site of this building from an early date. Oswald
identified an area of concrete surmounted by a plaster of
Paris base which he though could be a base for a statue. This
base still had impressions of the wooden framework used to
make it. It is uncertain what this building is but it is likely
to have had a public function, perhaps with a religious element.
A box flue tile. These ducted
hot air from the underfloor heating system up the walls
to let the steam out the eaves. They were used in bath
houses to heat up the warm rooms. This example has traces
of wall plaster still adhering.
Nottingham University Museum Photo: Robin Aldworth
|
The neighbouring
building, M, is a bathhouse. This had two types of under-floor
heating: a floor raised on brick pillars and a floor
with channels lined with tile built into them for the
hot air to travel along. The floors were made of a special
Roman concrete, opus signinum, made from concrete
mixed with crushed brick and tile. This was a durable
choice in the steamy atmosphere of the baths and reduced
the risk of fire. Debris from the collapse of the bath
building included box flue tiles, roofing
tiles and wall
plaster. A white mortar floor was found by Oswald
below the opus signinum floor of the bath suite and
this, together with the painted plaster in the early
levels of building L, suggests there was an elaborate
building on or near this site before either building
L or M were built, perhaps an earlier bathhouse. We
can suggest that building M had a stone wall, rather
than a timber frame, to hold up a tiled roof and not
catch fire. The walls would have been painted and the
windows were probably glazed.
A well in the area of building G contained a fragment
of a small column of Lincolnshire
limestone. This had a flattened back and a rectangular
notch cut in the upper side suggesting it was used as
a pilaster (column fitted against a wall) decorating
an elegant house. This fragment was found in a rubble
layer between groups of pottery, which would indicate
a date in the earlier Roman period, perhaps as late
as the 2nd century. We do not know which building this
pilaster may have belonged to but it does indicate the
adoption of classical styles by one of the inhabitants.
|
The Roman name for Margidunum means marly
fort and there is much evidence in the area of the excavations
and in later records that the site was situated on a small
rise in an area of flat marshland. The area occupied by buildings
L and M was close to a damp depression within this area and
the buildings seem to have been flooded during the third century.
Sometime around the middle of the third century, large stones
were used to consolidate the area above the silted-up remains
of buildings L and M and a gravel layer was laid down. A large,
well-appointed stone-footed building (N), surrounded by this
gravel layer, was found by Oswald to the south of buildings
L and M. It had floors of opus signinum (Roman concrete) and
plastered
walls decorated with grey panels and bands of red, orange
and green. Oswald noted a plaster quarter moulding had been
made at the junction of the walls and floors. He also found
the plaster in situ on the walls. A large number of Charnwood
roof
slates were recovered from this building with nail holes
in them, which would result in a diamond pattern when laid.
This building was aligned with buildings L and M and although
Oswald dates it to the fourth century, the complexity of the
plan suggests alterations and rebuilding over a longer period
of time, perhaps overlapping in use with these other buildings.
A more modest building (B) east of the Fosse Way was found
above the first century timber houses. This was built using
the dry rubble type of stone foundation and had a beaten clay
floor.
Remains of a beaten clay
floor and dry rubble-type of stone foundation in Building
B
Photo: M. Todd. Nottingham University Dept Archaeology
(No B1402) |
Sherds of colour-coated ware dating from
150 AD at the earliest was found below this floor but many
sherds of late third to fourth century pottery overlay it.
The date of this building is therefore sometime after the
mid-2nd century and possibly as late as the late 3rd-4th.
In style the house resembles the buildings that are dated
to the second century. We know that a building of higher status
existed near to this building as a thick layer of broken opus
signinum was recovered from just south of it. This may be
the source of the late 3rd-4th century pottery found above
the floor of building B.
Other elaborate buildings of the late
3rd-4th centuries are known only from building debris found
above earlier features. These layers of painted plaster and
opus signinum with late 3rd-4th century pottery, were recorded
by Oswald in a layer above the consolidated silts overlying
buildings L and M, but below the gravelled area around building
N and the rampart ditches. In the backfill of a well, which
fell out of use and was used as a handy rubbish pit, roof
tiles were found in a layer above debris containing late 3rd
century coins. Although these tiles may be from a building
first constructed in the second century, it does not seem
likely that the tiles from first century buildings would be
dumped in a well in the late 3rd century. Pryce, a co-worker
of Oswald’s, mentions finding “isolated tessera”
(small, square piece of a tile or other material used for
making mosaic floors) and these are most likely to belong
to a building with at least one mosaic or tesselated pavement.
Skerry paving floor
of a simple house, built late in the Roman period
or after it.
Photo: M. Todd. Nottingham University Dept Archaeology
(No B1627)
|
The latest structure discovered at Margidunum
is represented by an oval patch of skerry paving associated
with six large postholes. This was found to overlie the upcast
bank of the town defences and was dated by pottery to after
the mid-4th century. There is, however, no record of pottery
associated with its occupation and such a simple hut may date
to the very end of the Roman period or even the early Saxon
period. It represents a real fall in living standards and
its position over the rampart implies that the civil authorities
were no longer able to take care of such public works.
If you want
to know more detail about the excavations click
here to find the books and reports with it in.
|
