Roman Fort (roman)
- Council Edinburgh, City Of
- Parish Edinburgh (edinburgh, City Of)
- Former Region Lothian
- Former District City Of Edinburgh
- Former County Midlothian
NT17NE 3.00 1895 7683
(Centred NT 1895 7683) Roman Fort (NR) (site of)
OS 6" map, (1968).
NT17NE 3.01 NT 1899 7679 and NT 1896 7684 Architectural Fragments
NT17NE 3.02 NT 1899 7698 Flints; Pits; Bead
NT17NE 3.03 NT 1894 7673 Roman Road; Pottery
See also (trenching at NT 189 769).
For trial excavations in Cramond Inn car park (NT 1897 7705), see
For Cramond 'lioness' (found at NT 1890 7702, but first published as being from the fort), see
For (2005) watching brief (NT 193 767 to NT 193 770), see
Cramond Roman Fort (see RCAHMS 1929 plan, fig.3) encloses just under 5 acres; a civil settlement lay outside the E rampart. Excavations 1954-66 showed that it was first constructed about 140 AD, with a second Antonine occupation. There was a third occupation under Severus, with extensive reconstruction along existing alignments. Some use of the fort (probably civilian) with minor building in a Roman manner also took place in the post-Severan period which may be associated with a little 4th c pottery. The civil settlement was occupied during all four phases and yielded some 1st century coins, but if there was an Agricolan fort at Cramond, no trace of this has been found.
Inscribed stones which may be attributed to the fort are a centurial stone of Legio II Augusta (RIB 2137): an altar "To the Alatervan Mothers and the Mothers of the Parade-ground" erected by a cohort of Tungrians (either first or second) at the instance of a centurion of the Twentieth Legion (RIB 2135): and an altar "To Jupiter, Best and Greatest" erected by the fifth cohort of Gauls (RIB 2134). Rae gives a full description of the pottery, coins, etc found during the excavations. Material from the fort is in the NMAS and Huntly House Museum, Edinburgh.
The part of the fort now owned by the City Corporation has been laid out as a public open space in which the outline of buildings, rampart and gate appear against a background of grass, and in which a stretch of Severan wall has been conserved by the DoE.
A Rae and V Rae 1974; RCAHMS 1929; G Macdonald 1918.
An emergency excavation immediately S of the fort revealed traces of masonry and cobbling in an area where a civil settlement would be expected. Tiles, Roman pottery, and a Severan coin were also found.
M Walker 1971.
Excavations were carried out in three areas in the vicinity of the Roman fort:
NT 191 767. Two areas of the southern walled garden of the manse were investigated. A quantity of Roman tile, amphora and pottery fragments were recovered, mostly in an abraded condition, as well as one Roman coin, but there were no identifiable structural remains datable to the Roman period. It seems clear, therefore, that the vicus did not extend this far to the S.
NT 190 769. An attempt was made to locate the position of a N ditch of the fort by means of a machine cutting. No trace was found.
NT 190 770. Part of a Roman building was revealed during ground clearance for the construction of a car park S of Cramond Inn. Excavation showed the structure to be fairly extensive, and further work will be required to determine its complete ground plan. The walls at the W end survive to a height of over 1.5m from foundation level but elsewhere some stretches of the wall have been completely destroyed. The area so far excavated contains the remains of at least two hypocausted rooms, and the enormous number of broken flue-tiles removed suggests that the building was a bath-house, but so far no distinctive features have been discovered. The structural remains are clearly of at least two phases, and the small amount of stratified Roman material so far recovered suggests that these date from the Antonine period only, but one coin from the lowest of the disturbed levels is of Severus.
Information from Discovery and Excavation, Scotland Ms 1975.
The part of the fort marked out on the ground is as described but there are no surface indications of the remaining extent. No further information was obtained regarding the finds to the S of the fort. The excavation S of Cramond Inn is centred at NT 1901 7701.
Visited by OS (BS) 11 December 1975.
Excavations previously reported were completed in 1986. In addition to the road leading SE from the East Gate, a foundation-trench and post-holes for timber buildings of Antonine and Severan date were noted.
S S Frere 1988.
(NT 189 769) Exploratory trenching continued to the N of the 1988 excavation. Traces were noted of a late trench, possibly one dug by A and V Rae during the fort excavations in the late 1950's.
Further building remains were found, on a similar alignment to those previously reported and separated from them by rubbish pits. Below quantities of demolition rubble and pantile, traces of three floors or surfaces were noted; two of these were of a yellow clay base showing patches of wear which exposed an unerlying layer of coal dust and
shell, and probable clay and turf spread from the Fort rampart lay below the S surface. The tird (central) had an upper beaten earth surface above a floor made from tiny pebbles and fragments of crushed pottery set in clay; this area had been laid over a substantially-built drain, running N-S. The visible length of this drain was 6m, with two construction phases; the first was of square-cut stones, revetted on the W with very large boulders. A spread of cobbles and oyster
shells extended across the N end, after which the drain's sides had been crudely rebuilt with rounded boulders. The capstones were missing at this point, and the drain was overlain by the N clay surface. Glass and pottery from the trench indicate an 18th century date for the building remains, and iron slag may indicate some minor industrial activity. Vast quantities of shells of many varieties indicate the scale of local consumption.
V E Dean 1989.
(NT 189 769). Excavation of the post-medieval building remains and rubbish pits, revealed in 1988, was completed. A possible grave lay beneath the floor slabs. Indications are that the fort rampart had been removed; the underlying clay was found to be natural. Further work was done on the second building and drain (noted in 1989). More immense boulders were revealed, revetting the W side of the drain. These were overlain by dumped stone and shell, apparently as levelling for the drain's later phase; this contained pottery ranging from Roman to post-medieval green-glazed ware.
V E Dean 1990.
(NT 189 769). Work continued on the exploratory trench (noted in 1988, 1989 and 1990) which ran parallel and 7m W of the still-visible Roman gutter on the W side of the road through the N gateway of the fort.
To date, the trench has extended to 23m N of the outer face of the N rampart as shown by A and V Rae. No trace has been found of a fort ditch which might have been expected to the N of the fort. An as of Domitian (81-96 AD) was retrieved.
The trench was excavated into the subsoil and showed that the group of huge boulders noted in 1990 was embedded in glacial till and their crevices packed with boulders and clay. The drain and floor of the robbed-out post-medieval building had been constructed through and over this group. Four conjoining pieces of a larger stone mortar of
unknown date had been built into the drain's sides.
Sponsors: Edinburgh City Museums and Galleries, Edinburgh Archaeological Field Society.
V E Dean 1991.
(NT c. 190 769). Three trial trenches were opened within the perimeter of the fort's defences in advance of the installation of a new service pipe trench to Cramond House.
Trench 1 (NT 1894 7687) was located on the northern edge of Kirk Cramond roadway, on the projected line of the western wall of the granary building, previously excavated to the south of the modern road. A rubble layer was encountered, probably the levelling for the road. Below this was a surface of compact mortar set with stones, suggesting either a arbbled surface, or perhaps the core of a rubble-filled wall, possibly of Roman date. This lay below the depth of the new service trench and would not be damaged by it.
Trenches 2 (NT 1897 7688) and 3 (NT 1902 7691) both contained material deposited as levelling for the modern road.
Sponsor: DM Hall (Chartered Surveyors).
H Smith 1991.
NT 1895 7685. Excavation of new footings for the old mortuary house in the NW corner of the kirkyard uncovered a 5m length of Roman granary wall, a continuation of the structure excavated immediately to the N by Alan Rae (Britannia 1974). The wall had been robbed to the course above foundation level in most places. It was constructed of a mortared rubble core with squared sandstone facing blocks. The main body of the wall was 0.9m wide with a single offset foundation course projecting on both sides. Two buttresses on its E side were recognised: both had been badly damaged by robbing. A vent ran through the wall width between the buttresses. After recording the wall was reburied.
Above the wall a depth of 1.3m of dark humic loam to modern ground contained many disarticulated human bones. No intact skeletons were observed.
Sponsor: City of Edinburgh District Council.
M Collard 1992.
NT 1899 7696. The exploratory trench in the area N of the Roman fort is now 38m N of the outer face of the N rampart and still without trace of a fort ditch.
The subsoil comprised glacial till, giving way to yellow clay with an overlying deposit of water-lain sand, into which had been cut a number of small pits. Several of these contained Roman pottery, but also a little post-medieval pottery.
Overlying the pits was one of two spreads of dumped stone, shell and bone, on top of which were the remnants of clay and cobble surfaces and mortar patches. Pottery within the dumps was predominantly early post-medieval green glaze, whereas that in the demolition layer above was largely 18th century material.
A further short stretch of stone-lined drain was noted, similar to that previously reported. Its projected line would join the other at approximately 125 degrees. Many of the side and packing stones were re-used, with considerable amounts of mortar adhering to them; post-medieval pottery was contained in its fill.
Sponsors: City of Edinburgh District Council, Museums and Galleries; Edinburgh Archaeological Field Society.
V E Dean 1992.
NT 1899 7698. Northward extension of the exploratory trench between the Roman fort and bathhouse revealed a possible entranceway to the W side of the mid 19th-century kennels building of Cramond House. The building has been cut into the edge of the 15m raised beach, necessitating steeply ramped access routes to the S and W. The above path consisted of a compacted surface of mortar and sandstone chips overlying steep tips of mortar and stones, not yet bottomed. This was bounded on the N by a length of curved walling and parallel drain, which overlay a very substantial clay-bonded wall, 1.1m wide, running E-W, whose large, squared, re-used stones bore traces of mortar. North of this, another wall, running N-S, and consisting of at least three courses of mortared stones of assorted shapes, appeared to have been deliberately collapsed to the E into a trench dug, through the clay-bonded wall, to receive it. This building had an outer rendering of ochre-yellow mortar, with an inner blue-painted plastered wall. No dating for these walls is yet available.
Sponsors: City of Edinburgh District Council, Archaeology Service; Edinburgh Archaeological Field Society.
V E Dean 1994
Excavation revealed the poorly preserved metalling of a road leading S from the S gate, and the W side of a ditch 2m wide and 0.95m deep, orientated N-S. This ditch, not aligned on the known fort defences, could belong to an earlier (Flavian?) installation.
L J F Keppie 1996
NT 1899 7698 The clay-bonded wall previously reported (Dean 1994) and running W-E was found to turn sharply to the N; traces of it extended for 1m. It had been badly damaged by later activity and most of its stones removed.
In the trench to the S, the compacted layer of mortar and sandstone chips was found to be the upper fill of a large circular pit, 3m in diameter and at least 1.3m deep, which was not bottomed. It was filled with successive layers of mortar and stone. At the N and S sides, it had cut post-medieval midden pits; these had apparently then been revetted with banks of clay and boulders. As the lower sides revealed the sand of the 15m Late Glacial raised beach, the pit may have been a sand extraction quarry related to the building of the adjacent mid-19th-century kennels.
Sponsors: City of Edinburgh Council, Edinburgh Archaeological Field Society.
V E Dean 1996.
NT 1908 7684 During October 1998 Headland Archaeology Ltd were commissioned by Smith Scott Mullan & Associates on behalf of Cramond Kirk to investigate the survival of archaeological remains within the proposed site of the extension to Cramond Kirk Hall. The site lies within the Scheduled Ancient Monument area of Cramond Roman Fort. According to previous excavations the proposed development site lies outside the E gate of the fort within the civilian settlement area.
Three archaeological trenches were dug. Trench 1 across the development area on the E Side of the Kirk Hall and trench 2 & 3 close to the W elevation of the Hall. Two smaller trenches were also dug to allow the architects to locale service pipes and cables.
Trench 1 uncovered a N/S aligned ditch, which is not parallel to the forts eastern Defences. This supports evidence found in 1995 suggesting the existence of a previously unknown Flavian Fort or possible Roman marching camp. Trench 2 revealed Roman road cobbles, which the report suggests, belong to the road leading from the E gate of the fort. Evidence shows that the road was reused many times, however further examination is needed to fully understand its history. The trenches also revealed some evidence of more recent building work related to an old stable, which was previously on the site of the Hall.
The report argues that these remains are of some importance and Historic Scotland, prior to any development will insist a mitigation strategy for their avoidance or excavation.
Sponsors: Smith Scott Mullan & Associates on behalf of Cramond Kirk
NMRS MS/899/107 (Headland Archaeology Ltd, October 1998)
NT 1899 7698 Two further studies have been made to the W of the trench in which a length of mortared stone wall was found (DES 1997, 32); these have extended to the W and S of the survey reported previously (DES 1999, 36).
The first was an area survey that showed higher resistance to the W of the previous survey and cast doubt on the likely continuation of the wall suggested in DES 1999. Seven linear array measurements (three E-W and four N-S) show a broad high-resistance ridge some 3m N of the wall line. At a depth of 1.5m, the ridge is at least 14m long and is most likely to relate to debris of a house in Old Street demolished in 1826.
A metal-detector survey prior to the resistance measurements produced a section of a ?lead seal.
Sponsors: Edinburgh Archaeological Field Society, University of Edinburgh - Dept of Geology and Geophysics.
H M D Jones 2000
NT 1899 7698 The function of the stone-lined pit previously noted (DES 1999, 36) cannot be confirmed as a kiln, due to the lack of a stoke-hole, although signs of burning were evident. Its base comprised a worn, square-socketed millstone in a stone setting; a coin tentatively identified as a turner of Charles I (1642-50) was incorporated in the stone lining. The cobbled Old Street of Cramond Village (DES 1997, 32) overlay a substantial build-up of stones and midden material. It was not possible to date the massive clay-bonded wall (DES 1996, 38). With the exception of a small pit which yielded Roman boot studs, no feature appears to be earlier than the later medieval period.
Sponsors: City of Edinburgh Council Archaeology Service, Edinburgh Archaeological Field Society.
V E Dean 2000
NT 1900 7701 The series of trial trenches, begun in 1988 (DES 1988, 18) to provide a stratigraphic link between the N rampart of Cramond Roman fort (NMRS ) and its bathhouse located some 100m to the N, was concluded in 2001 when part of the SW corner of the clay-bonded furnace room was briefly exposed. The visible section was in good condition and appeared to be stable.
The final length of trench had entered an area of backlands lying to the NE of the former village, destroyed in 1826. The only further feature noted was a trench containing a 2" cast-iron pipe, probably for water.
NT 1894 7698 A brief examination was made of a previously noted well (DES 1990, 29), and the upper stonework was recorded. In plan, the upper part of the well is D-shaped with the straight section on the northern side. A lintel-type stone bridges across the well, below which the wall appears to be circular in plan; this suggests that a modification had been made to support a pump. Nearby, a 1m long stone was found which has a semi-circular notch cut in one long edge and the remains of four iron brackets on its upper surface. A pump could have been mounted on this stone, which would have fitted comfortably across the top of the well. The OS 25" map of 1895 shows a pump at this location. Probing indicated that the depth of the well is in excess of 2.5m; the lower 0.5m appeared to be damp.
Sponsors: City of Edinburgh Council Archaeology Service, Edinburgh Archaeological Field Society.
V E Dean 2001
NT 1899 7698 An area ground-resistance survey was made approximately E-W from the side of previously excavated trench G across to the wall abutting onto Glebe Road. Undergrowth and trees limited the survey area to 384m2. High resistance appeared on the N side of the survey, which appears to relate to buildings shown on a vennel running W off Old Street on the 1815 map. High resistance at the end of the vennel equates to paving round the still open well. This area is to the W of the linear array survey previously reported (DES 2000, 33).
Sponsors: HS, Edinburgh Archaeological Field Society, City of Edinburgh Council Archaeology Service.
H M D Jones 2001
NT 1898 7702 An archaeological watching brief on the machine-excavation of a 35.6m long cable trench located immediately within the northern boundary of the Scheduled area of Cramond Roman fort failed to penetrate beyond modern made ground, and therefore failed to encounter either significant archaeological features or artefacts.
Sponsor: Scottish Power plc.
J Gooder 2001
NT 1908 7685 An archaeological excavation was carried out in advance of the construction of an extension to the kirk hall. The site lies within the Scheduled area around Cramond Roman fort. The excavation uncovered a previously unknown third defensive ditch around the Antonine fort and a series of pits, drains and other linear features probably associated with the Severan reoccupation of the fort in the early 3rd century AD. In addition to the Roman features, a large pit of medieval date was uncovered in the N end of the site.
Full details lodged with the NMRS.
Sponsor: Smith Scott Mullen & Associates for Cramond Kirk.
M Dalland 2001
NT 1910 7675 An area ground resistance survey was made within the N walled garden and on a small grassed area between the garden and the church hall. This is the area on which some of the fort vicus is assumed to have existed.
A linear low resistance about 4m wide runs adjacent to the W wall of the garden, continuing on to the area to the N; this aligns approximately with a path shown on the 1893 25 inch OS map. A linear high resistance runs parallel up the middle of the garden and another, less clearly defined, alongside the E wall; these align with smaller paths on the same map. A curved high resistance line on the grassed area could also align with an 1893 path.
The high/low differences between these paths cannot presently be explained, nor can the smaller high resistance strips to the E side of the walled garden.
Sponsors: HS, City of Edinburgh Council, Edinburgh Archaeological Field Society.
H M D Jones 2002
NT 1910 7675 It was thought that the lack of Roman features found in the previous survey (DES 2002, 47) could have been due to the ground having been made up to a depth that the survey could not detect, i.e. exceeding 0.75m. Two linear array measurements were later made that detected to a depth of 4m. The first was E-W across the walled garden and showed only the geology. The second ran NW from the grassed area N of the walled garden and recorded low and high resistances in positions that suggest the detection of the two ditches, the wall and building within the fort.
Report lodged with Edinburgh SMR and the NMRS.
Sponsors: HS, City of Edinburgh Council, University of Edinburgh - Dept Geology & Geophysics, Edinburgh Archaeological Field Society.
H M D Jones 2003.
NT 192769 An area ground resistance survey was conducted on the parkland to the E of Cramond House. It comprised 23 squares of 20 x 20m, and extended 120m to the E and 100m at its widest point. The most northerly row of six squares extended just over the edge of the old raised beach level.
High and low resistance lines, running parallel to the tree avenue that frames the view of North Berwick Law from the house, suggest that a formal parterre garden was associated with the original house, built in 1680. Lines and rectangles that cross the parallel lines appear to be at right angles to the Roman road out of Cramond fort.
Resistive linear array measurements across a low resistance line that runs along the edge of the raised beach on the area survey appear to show a ditch 2m wide and 2.5m deep, with possible upcast on the seaward side. This is very similar to that shown on a previous section across the inner ditch of the fort. Magnetometry results are still being analysed but do not show major features associated with resistive anomalies.
Sponsors: HS, Cramond and Harthill Estates, University of Edinburgh ¿ Dept Geosciences, Edinburgh Archaeological Field Society.
H M D Jones 2004
NT 193 767; NT 193 770 Due to the proximity of Cramond Roman fort (NT17NE 3.00), a programme of archaeological
works was undertaken between April and June 2005 on linear groundbreaking works associated with proposed pipe-laying connected with residential and leisure development. The works consisted of three elements: topsoil stripping and excavation, a watching brief, and the monitoring of the Scheduled area to the NE of the fort.
The topsoil stripping and subsequent excavation revealed the continuation of three large Roman ditch features recorded during recent excavations, together with a further outlying ditch feature. A number of linear cut features and a pit of unknown date were also recorded, as were two boundary walls of probable post-medieval date.
The watching brief phase uncovered a 23m long section of Roman road, aligned NW¿SE, which also appeared to show continued use throughout the medieval period.
A number of small finds, consisting of Roman and medieval ceramics and several iron and copper objects, were recovered from the ditch features and road.
Archive to be deposited in NMRS.
Sponsor: AMA Construction.
E Engl 2005
NT 192 769 Ground resistance and magnetometry survey (DES 2004, 53) was augmented by further magnetometry, with readings taken at closer intervals. Lines of magnetic anomaly correlated both with the resistive linear `garden' features aligned on North Berwick Law, and with those on the Roman alignment. Some isolated significant magnetic anomalies, c 2m square, could represent kiln and metalworking sites in this NE vicus area of the Roman fort.
Sponsors: HS, Cramond and Harthill Estates, University of Edinburgh, Dept Geosciences, Edinburgh Archaeological Field Society.
H M D Jones 2005
NT 1906 7695; NT 1896 7695 A ground resistance survey of three 20 x 20m squares was made in the area to the E of Cramond Tower to see if the ditch, detected as a low resistance along the edge of the raised beach (DES 2004, 53), continued to the W and linked with the known Roman ditch at the tower pond. A low resistance, c 3m wide, was found to curve to the S to encompass the pond. The curvature was such that the ditch could be seen as a continuation of the N¿S ditch on the E side of the fort rather than forming a T-junction onto the side of it.
The second survey, 100m to the W of the first, comprised a 40 x 40m area, with its N edge parallel to and 15m N of the fort wall and set to cover the N exit from the fort. The Roman gutter on the W side of the exit was recognised, and a 5m linear high resistance continued this line to the N of the fort wall. The survey found no indication of the wall. Low resistance lines are likely to relate to previous excavations.
Sponsors: Mr Eric Jamieson, Cramond Tower, City of Edinburgh Council, Edinburgh Archaeological Field Society.
H M D Jones 2005.