• Keine Ergebnisse gefunden

Rebecca H. Jones

Im Dokument The Antonine Wall (Seite 110-120)

Introduction

Lawrence Keppie’s magisterial publication on The Antiquarian Rediscovery of the Antonine Wall appropriately ends with the excavations of the forts along the Wall in the late 19th and early 20th centuries (2012: 127-137). Concluding with Sir George Macdonald, Keppie notes the important excavations at Bar Hill from 1902-5 in which Macdonald was involved. These excavations uncovered a curious structure under the fort which has received various interpretations over the years, but, together with other structures along its length, may have played a critical role in the initial planning, surveying and building of the Wall.

The Bar Hill Structure

The initial excavations at Bar Hill were directed by Alexander Park, factor on the Gartshore estate, and funded by Gartshore’s owner, Alexander Whitelaw. Together with Macdonald, Park swiftly wrote up and published the results of these excavations (Macdonald and Park 1906; Keppie 2012: 133-4).

An unexpected discovery during the excavations was of a structure lying at an angle beneath the Antonine fort (Figure 7.1). This discovery led the excavators to suggest that this was an ‘Early Fort’

underlying its Antonine successor. The provision of an annexe and elaborate drainage arrangements, seemingly linked to what became the fort ditch to the south-west, were used to argue that it was something more substantial than a temporary camp and therefore the permanent home of a small garrison. This, they suggested, meant that it was one of the garrisons placed on the isthmus by the Governor Agricola in his Flavian campaigns in the AD 80s (Macdonald and Park 1906: 11-15).

Figure 7.1 Outline drawings of the two structures under Bar Hill and Croy Hill The Antonine Wall: Papers in honour of Professor Lawrence Keppie: 86–95

This view was later challenged by Steer, who argued that the enclosure was Antonine or proto-Antonine, shortly afterwards replaced by the fort (1960: 88). The recovery of turf blocks in a ‘wonderful state of preservation’ (Macdonald and Park 1906: 14) could have been from the rampart which were placed in the ditch when it was dismantled, the excellent preservation being due to them not being in the rampart for very long, therefore indicating a short-lived Antonine structure. A later suggestion was made interpreting it and its neighbour at Croy Hill (see below; Figure 7.1) as Iron Age homesteads (Feachem 1968). Re-excavation of the ditch in 1978-82 by Keppie recovered more well-preserved turf blocks, mixed in with twigs and small hawthorn branches, neatly cut and still with their bark and thorns. Finds from the subsoil below the fortlet included some of Iron Age date, but no traces of circular houses were recovered. None of the finds dated to the Flavian period, but Antonine pottery was recovered from an associated hearth (Keppie 1986: 51-8).

Croy Hill

Excavations by Macdonald in 1920 and 1931 identified the ditch of a structure under the Roman fort on Croy Hill, to which he drew parallels with that found previously under Bar Hill, continuing his interpretation to propose that both were Agricolan forts on the isthmus (Macdonald 1932: 262-6). Further excavations in 1975-77 provided more detail: the structure is in two parts, the northern part beneath the Antonine Wall fort, with a possible annexe to the south. However, these excavations concluded that this structure was not Agricolan in date. The spatial relationship between the ditch of the enclosure’s southern annexe and fort road demonstrated that the two were in use contemporaneously (Hanson 1977: 6-7); this was also confirmed by pottery found in the enclosure ditch (Hanson forthcoming). It, therefore, seemed likely that both structures (on Croy Hill and Bar Hill) were Antonine, but early in the sequence for the Wall. That at Croy Hill is about a third larger than its neighbour at Bar Hill. In addition, the rampart material in the ditch of Bar Hill suggests a short-lived occupation at that site, with its ditches deliberately filled in. At Croy, Macdonald recorded a small stretch of roadway some 18 m (60 feet) long thought to be contemporary with the enclosure rather than the overlying fort (1932: 265). This would suggest that it may have been in use for a reasonable period. Its ditches appear to have lain open to the elements for some time prior to the construction of the later fort, unlike at Bar Hill. This may suggest that the fort at Bar Hill was built before that at Croy; indeed, the likelihood that the Croy fort is later is suggested by its neighbouring fortlet (Hanson forthcoming).

Purpose

Now that an Antonine date seems likely, there has been a general assumption that both enclosures housed troops involved in the construction of the Wall or its associated forts and fortlets (e.g. Hanson and Maxwell 1986: 120). Largely due to aerial survey, some 20 or so camps are recorded along the Wall (Figures 7.2 and 7.3) and various discussions have attempted to link them to its building programme (Feachem 1958; Maxwell 1974; Hanson and Maxwell 1986; Jones 2005). The two structures under Bar Hill and Croy Hill do not fit into the general typology of camps (Figure 7.2) and are significantly smaller;

therefore, it cannot be assumed that they performed the same function. Bar Hill enclosed only some 0.28 ha and Croy Hill 0.32 ha (with annexes increasing the holding capacity of both); the majority of camps are around 2 ha. I have argued elsewhere that one plausible explanation for the structures under the forts at Bar Hill and Croy Hill is that they functioned as surveying camps for the Wall (Jones 2005; Jones 2011: 330).

Figure 7.2 Outline drawings of all the temporary camps known along the Wall.

There can be no argument that the two occupy significant strategic locations on the Wall. Bar Hill is the highest fort on the Wall and a central point between the firths of Clyde and Forth, with excellent views in all directions (Figure 7.4). Next to the fort of Bar Hill is a ‘rocky peak’ known as Castle Hill with denuded fortifications believed to be of Iron Age date. This is the highest point on the isthmus, but is too small to have housed a Roman detachment. The site of the later fort, its immediate neighbour at Bar Hill, was the next best place, which is why it is likely to have been selected at an early stage to house the surveyors and planners for the Wall. To quote Macdonald and Park, ‘even the uninstructed feels instinctively that this would be a position of vital importance to any military force attempting to hold the isthmus from the south’ (1906: 2). Croy Hill lies close to the highest point on that hill, with good views along the line of the Wall including across to Bar Hill. If both enclosures were involved in the early surveying and planning of the frontier, then there must be other sites waiting to be discovered across the isthmus.

Immediately to the east of the fort of Mumrills, a small enclosure (at least 0.12 ha in area), potentially not dissimilar in size to the inner part of the enclosure at Bar Hill, has been discovered, although only one side and parts of the adjacent two have been recorded, the remainder destroyed by erosion of the scarp on which it is sited. Excavations by Anne Robertson suggested that it was Antonine but short-lived, and the suggestion was made that it may have held stores rather than troops (Steer 1961: 96); unlike Bar Hill and Croy Hill, it is not sited in a commanding position. At Inveravon, a curious small probable camp (Inveravon III) enclosing some 0.4 ha is recorded down the slope from the Wall, but again not in a strong position. It overlaps another camp possibly involved in the construction of the Wall although the chronological relationship between the two is unknown (Jones 2011: 232-3).

Neither of these two candidates appear comparable with Bar Hill and Croy Hill.

Figure 7.3 Map of the Wall showing the locations of the camps.

Figure 7.4 Viewshed analysis showing areas visible from the Bar Hill structure.

Castlehill, Bearsden

Some 11.5 miles (18.5 km) west of Bar Hill, the fort of Castlehill occupies a commanding position close to the western end of the Wall, around 118 m above sea level and noted as a ‘conspicuous landmark’

by Keppie (1980: 80). The existence of a Roman castellum of some sort on the plateau was recorded by antiquarians, but it was not until 1947 that its perimeter defences were recorded from the air (St Joseph 1951: 61). That there was also a fortlet here was suggested, particularly after John Gillam started the great fortlet hunt in the 1970s (Gillam 1975), with suggestions that a fortlet was located on the north-western side, immediately next to the fort (Keppie 1980), potentially echoing a similar situation at Duntocher to the west (Robertson 1957).

Detailed topographic and geophysical survey of the area in 2008 with further work in 2011 and 2019, has provided some details of the fort including a possible small ditched enclosure, just over 0.1 ha in area, in its north-west corner (see Hanson and Jones, this volume). The interpretation by the initial survey team was that this feature was not that of a fortlet, but they noted its similarities to the structures under Bar Hill and Croy Hill (Jones et al. 2009). Though the survey data has recently been reinterpreted as evidence for a Roman fortlet, questions remain regarding this interpretation and the suggestion of an enclosure similar to Bar Hill and Croy Hill needs more exploration.

Like Bar Hill and Croy Hill, Castlehill occupies a key position on the Wall. As the second highest fort on the line (after Bar Hill), it has an excellent outlook across the Clyde estuary to the west as well as good views north and along the Wall to the east (Figure 7.5). As well as being a key topographic location, it also marks a change in the building programme for the Wall, which has been recorded on the Distance Stones.

Distance Stones and the building of the Wall

The series of Distance Stones known from the Antonine Wall are unparalleled in the Roman World (Keppie 1998: 50-6). Recording how much of the frontier was built by each legion, Castlehill marks a change in how the lengths were subdivided and recorded. To the east of Castlehill, distances are recorded in passus (paces); to the west, in pes (feet) (cf. RIB I 2196 and 2197). Indeed, this western sector may run from just to the west of the Castlehill fort (Hanson and Maxwell 1986: 123; Macdonald 1934:

382), and therefore just to the west of the Castlehill structure.

Following earlier discussions by Macdonald (1934) and Keppie (1975), Hassall proposed that the Wall was constructed in three phases: the central section from Castlehill to somewhere just east of Castlecary; the short western section from Castlehill to the Clyde; and the eastern section from near Castlecary to the eastern end at Bridgeness (1983). The curious distances in the central section, with its

‘distinctly odd division’ (Keppie 1975: 154) of around 3666 paces on each Distance Stone was explained by Hassall as being a fairly even division of labour in the central sector between the three legions (II, VI and XX) deployed (1983: 263). He further suggested that this central sector may have been laid out and constructed first. If this were correct, then it places additional emphasis on any ancillary structures recorded between Castlehill and Castlecary. Given that Castlehill was clearly important in the measuring and building of the Wall, this places additional emphasis on the significance of the structure revealed through geophysics.

Figure 7.5 Viewshed analysis showing areas visible from Castlehill.

Regardless of which sectors were built in which order, it is likely that the route of the Wall and its structures was established before construction began. Poulter has argued that the location of all the forts on the Wall was established at an early stage (2009: 123), regardless of the speed with which they were built. Gillam’s thesis (1975) that there was a change in the plan of the Wall, with ‘primary’ and

‘secondary’ forts, rather than modifications and sequencing in fort construction, has been challenged recently (Graafstal et al. 2015; Graafstal, this volume), and metrical analysis of the Distance Stones based on LiDAR survey has raised further doubts about it (Hannon et al. 2017). What these discussions serve to underline is the importance of understanding the curious structures at Bar Hill, Croy Hill and Castlehill.

Construction camps

In addition to these possible surveying camps, the suite of camps along the Wall has already been noted (Figures 7.2 and 7.3). The majority are around 2 to 2.5 ha in size but there is one group of four camps which are over 4 ha in size, around double the size of the rest. The location of these four is of interest: Dullatur I, Garnhall I, Balmuildy and Wester Carmuirs (note that Garnhall II on Figure 7.2 may post-date the Wall (Woolliscroft 2008: 167-8; Jones 2005)). Two (Balmuildy and Wester Carmuirs) lie to the north of the Wall. Had the Wall been in an advanced stage of construction, the act of crossing it from the north could have proved an obstacle. It is therefore assumed that these two, if they do indeed relate to the construction of the Wall, are early in the building sequence. In addition, that at Balmuildy lies only a short distance from the fort of the same name, itself considered to be early in the building sequence due to the provision of stone wing-walls constructed before the Antonine Wall rampart reached the fort (Miller 1922). Therefore, it is perhaps no surprise to propose an early camp at this location. Wester Carmuirs, also to the north of the Wall, lies close to the cluster of camps at Lochlands, an existing gathering ground where some of the camps are likely to be of first century date (Jones 2011:

257-62). It is a key nodal point on the Wall for a number of reasons, including proximity both to the road through the Wall at Watling Lodge and to the River Carron. The camp at Garnhall I, whilst situated south of the Wall, lies close to the fort of Castlecary, another fort with stone walls constructed prior to the arrival of the linear barrier (Christison et al. 1903). It is noteworthy that Balmuildy and Garnhall lie close to forts (Balmuildy and Castlecary) that have been proposed as very early in the building sequence for the Wall (Graafstal, this volume). Located south of the Wall midway between the forts of Croy Hill and Westerwood, the reason why the camp at Dullatur should relate to an early stage in the construction process is less obvious.

Conclusions

When considering the construction sequence for the Wall, all the evidence available should be deployed to tell the story, including the Distance Stones, camps, evidence for sequencing, topography and varied structures. Whilst it is not currently possible to prove that the curious incident of the structure at Bar Hill is a surveying camp, I hope to have demonstrated that it, together with its relatives on Croy Hill and possibly Castlehill, have a significant role in our understanding of the Wall and deserve further analysis. Whilst the enclosures at Mumrills and Inveravon III do not appear comparable, they may yet relate to the building and occupation of the Wall.

The evidence from Bar Hill, Croy Hill, possibly Castlehill and the likely early camps, supports Hassall’s suggestion of an emphasis on the central sector (1983) and Graafstal’s proposal of the importance of Balmuildy and Castlecary early in the building sequence (this volume). Taken together, these aid our interpretations of the planning and building sequence for the Antonine Wall.

Acknowledgements

I would like to thank Kevin Macleod for preparing Figures 7.1-7.3 and Frank Thomas at HES for undertaking the viewshed analysis and providing Figures 7.4 and 7.5. I would also like to thank Erik Graafstal and David Breeze for stimulating discussions on the building sequence for the Antonine Wall.

Bibliography

Christison, D., M. Buchanan and J. Anderson 1903.

Excavation of Castlecary fort on the Antonine Vallum. Proceedings of the Society of Antiquaries of Scotland 37: 271-346.

Feachem, R.W. 1968. Medionemeton on the limes of Antoninus Pius, Scotland, in J. Bibauw (ed.) Hommages à Marcel Renard vol. 3: 210-18.

Brussels: Latomus.

Gillam, J.P. 1975. Possible changes in plan during the construction of the Antonine Wall. Scottish Archaeological Forum 7: 51-65.

Graafstal, E., D.J. Breeze, R.H. Jones and M.

Symonds 2015. Sacred cows in the landscape:

rethinking the planning of the Antonine Wall, in D.J. Breeze, R.H. Jones and I.A. Oltean (eds) 2015. Understanding Roman frontiers:

a celebration for Professor Bill Hanson: 54-69.

Edinburgh: John Donald.

Hannon, N., D.J. Rohl and L. Wilson 2017. The Antonine Wall’s distance slabs: LiDAR as metric survey. Journal of Roman Archaeology 30:

447-68.

Hanson, W.S. 1977. Corbridge and Croy Hill:

recent work on Agricola’s third and fourth campaigns, in J. Fitz (ed.) Limes: Akten des XI Internationalen Limeskongresses: 1-9. Budapest:

Akadémiai Kiadó.

Hanson, W.S. forthcoming. Excavations outside the Roman fort on the Antonine Wall at Croy Hill, 1975-78. Edinburgh: Scottish Archaeological Internet Reports.

Hanson, W.S. and G.S. Maxwell 1986. Rome’s North-West Frontier. The Antonine Wall (revised edn).

Edinburgh: Edinburgh University Press.

Hassall, M.W.C. 1983. The building of the Antonine Wall. Britannia 14: 262-64.

Jones, R.H. 2005. Temporary Camps on the Antonine Wall, in Zs. Visy (ed.) Limes XIX.

Proceedings of the XIXth International Congress of Roman Frontier Studies held in Pécs, Hungary, September 2003: 551-60. Pécs: University of Pécs.

Jones, R.H. 2011. Roman Camps in Scotland.

Edinburgh: Society of Antiquaries of Scotland.

Jones, R., J. Huggett and A. Leslie 2009.

Geophysical and topographic survey at Castlehill Roman Fort, Bearsden: a preliminary report. Unpublished report, Department of Archaeology, University of Glasgow.

Keppie, L.J.F. 1975. The building of the Antonine Wall: archaeological and epigraphic evidence.

Proceedings of the Society of Antiquaries of Scotland 105: 151-65.

Keppie, L.J.F. 1980. The Roman fort(s) on Castlehill, Bearsden. Glasgow Archaeological Journal 7: 80-84.

Keppie, L.J.F. 1986. Excavations at the Roman fort on Bar Hill, 1978-82. Glasgow Archaeological Journal 12: 49-81.

Keppie, L.J.F. 1998. Roman inscribed and sculptured stones in the Hunterian Museum, University of Glasgow (Britannia Monograph Series 13).

Glasgow: Society for the Promotion of Roman Studies.

Keppie, L.J.F. 2012. The Antiquarian rediscovery of the Antonine Wall. Edinburgh: Society of Antiquaries of Scotland.

Macdonald, G. 1932. Notes on the Roman forts at Old Kilpatrick and Croy Hill. Proceedings of the Society of Antiquaries of Scotland 66: 243-96.

Macdonald. G. 1934. The Roman Wall in Scotland.

Oxford: Oxford University Press.

Macdonald, G. and A. Park 1906. The Roman Forts on the Bar Hill, Dumbartonshire. Glasgow:

Maclehose.

Maxwell, G.S. 1974. The building of the Antonine Wall, in D.M. Pippidi (ed.) Actes du IXe Congrès International d’Études sur les Frontières Romaines:

327-32. Bucharest and Köln-Wien: Editura Academiei/Böhlau Verlag

Miller, S.N. 1922. The Roman fort at Balmuildy on the Antonine Wall. Glasgow: Maclehose and Jackson.

Poulter, J. 2009. Surveying military landscapes across northern Britain: the planning of Roman Dere Street, Hadrian’s Wall and the Vallum,

and the Antonine Wall in Scotland (British Archaeological Reports British Series 492).

Oxford: Archaeopress.

Robertson, A.S. 1957. An Antonine fort: Golden Hill, Duntocher. Edinburgh: Oliver and Boyd.

St Joseph, J.K.S. 1961. Air reconnaissance in Britain. Journal of Roman Studies 41: 52-65.

Steer, K.A. 1960. The Antonine Wall, 1934-1959.

Journal of Roman Studies 50: 84-93.

Steer, K.A. 1961. Excavations at Mumrills fort, 1958-60. Proceedings of the Society of Antiquaries of Scotland 94: 86-132.

Woolliscroft, D.J. 2008. Excavations at Garnhall on the line of the Antonine Wall. Proceedings of the Society of Antiquaries of Scotland 138: 129-76.

Im Dokument The Antonine Wall (Seite 110-120)