UJA2012.qxd:UJA 2005
22/4/12
10:56
Page 17
Ulster Journal of Archaeology, Vol 69, 2010
TRANSHUMANT STRUCTURES AT DEER’S MEADOW HUT GROUP C, STANG, COUNTY DOWN STUART RATHBONE
[email protected]
Three groups of booley houses at Deer’s Meadow, Co Down, came to prominence in the late 1950s when one of the buildings in Hut Group A was excavated in advance of the creation of the Spelga Dam reservoir. This paper presents a survey of the larger Hut Group C. The site is discussed in relation to developments in our understanding of booley houses since the 1950s.
INTRODUCTION Located in the heart of the Mourne Mountains, Deer’s Meadow is important to our understanding of transhumant farming in Ireland. It is one of the few locations that has been positively identified as an area of transhumant settlement through historical records (in an account of 1744), and that has also been archaeologically investigated (Evans & Proudfoot 1958). The creation of the Spelga Dam reservoir caused an area of the meadow to be flooded, and in advance of this a rescue excavation was organized. Three groups of huts were identified, with groups A and B being small clusters of just a few buildings and group C consisting of a larger number of buildings running along the banks of a small stream. Group A was located in the area that was to be flooded and so was selected for a small scale archaeological investigation in 1957 (Evans & Proudfoot 1958). One hut site was excavated and found to be square in plan, measuring 3.6m by 3.6m internally, and with a substantial area of intense burning in the interior. The walls were constructed of layers of sod and gravelly soil. There was evidence that this hut overlay a previous construction but the plan of this could not be recovered (Ibid, 127–29). Small grassy mounds and a series of large stones were thought to represent the remains of five further structures in the immediate area. A large scale plan of the wider area shows the three groups in relation to the reservoir, with simple points denoting that there were two buildings at Hut Group B and eight buildings at Hut Group C (Ibid, fig 1). Sketch plans of two buildings from Hut Group C had previously been published by Evans in 1951 (Evans 1951, figs c, d). On 16 December 2001 the author undertook a field survey of Deer’s Meadow Hut Group C as part of an undergraduate research project (Rathbone 2002). After locating the site a brief inspection revealed the presence of nine extant hut remains,
which appeared as very low earthen banks sited on small artificial terraces, and three empty platforms of similar shape and size which might have supported structures at some point in time. The structures are situated on sloping ground in the townland of Stang (IGR J 2732 2648; SMR DOW 048:033) on either side of a small stream that runs from east to west, and are about 500m east of the reservoir (Fig 1). A small stone-built ‘fodder stand’ is located approximately 30m to the north of the buildings. During the survey it was noted that Hut Group B is now located in an area of forestry plantation and the current condition of the buildings in that group is unknown. Eight of the buildings, two of the platforms and the fodder stand were planned at a scale of 1:10. Where the line of the walls became unclear due to vegetation cover or damage, the position was identified by probing with the end of a ranging pole; when this did not reveal anything the line was estimated in relation to adjacent extant segments of wall. Due to failing light structures C10 and C12 could not be planned in detail but quick measurements were taken and sketch plans were recorded which have been included in Fig 1.
17
THE SURVEY The buildings at Deer’s Meadow Hut Group C have been classified according to the author’s recently published typology of transhumant structures around the northern basin of the Irish Sea (Rathbone 2009). This typology is based on the designs of shielings from north-western England and the Isle of Man, hafods from North Wales and Booleys from the Mourne Mountains and the Antrim plateau. A total of eight styles of simple structure were found to be present across these areas: circular one-roomed buildings; oval one-roomed, two-roomed and threeroomed buildings; rectangular one-roomed, two-roomed and three-roomed buildings; and square one-roomed buildings (Ibid, fig 5.2). It was
UJA2012.qxd:UJA 2005
22/4/12
10:56
Page 18
Fig 1 Plan of Deer’s Meadow Hut Group C. The outlines of the buildings’ walls are shown in black; those of the empty platforms are shown in grey.
suggested tentatively that the circular buildings were the oldest form, followed by the oval and rectangular forms, with the square forms being the latest in the sequence. It was also noted that in many cases there may be very little difference between the oval and rectangular buildings other than the building materials that were available (Ibid, 114–21). Details of the classifications of the Deer’s Meadow Hut Group C buildings, and their dimensions, are presented in Table 1. Building C1 is a rectangular two-roomed structure, C2 and C3 are small, one-roomed square structures and C4, C6, C8, C9, and C11 are all small, one-roomed oval structures (Fig 2). Although part of the structure had been removed by the undercutting action of the stream, C10 is clearly rectangular or square and was probably of a similar design to C1. If it had been a square structure, it would have been noticeably larger than either C2 or C3. In all cases where walls survived they consisted of a low bank of sod with occasional large stones included. In no instance was there a substantial build up of collapse or slumped material, and it has to be imagined that the walls were simply low foundations for a slight timber superstructure. The numbers C5, C7, and C12 were represented only as levelled platforms or small terraces, where no traces of actual structure were visible. The use of artificial terraces or enhanced natural platforms is a common feature at this site and C3, C8, C10, and C11 are all built upon such features. The larger platforms had apparently been built of stone and earth or of earth revetted by stone, whilst the smaller platforms were apparently created by cutting into the hillside and pilling the spoil up slightly down slope. In the case of C3 the height of the platform was quite substantial, being around 1.5m at its southern end. The platforms C5 and C7 are so small that only very small buildings similar in size to C8 could have stood on them, but a slightly larger building could have been built on C12. The absence of surviving walls is problematic because there would be little need to rob the walls for building material; both sod and stone would always have been in plentiful supply in this area. It is possible that these three platforms held buildings that were of a different design
Table 1 Details of Deer’s Meadow Hut Group C structures. 18
UJA2012.qxd:UJA 2005
22/4/12
10:56
Page 19
Fig 2 Transhumant structures recorded at Deer’s Meadow Hut Group C.
that did not require the use of an earth and stone foundation wall. On the other hand, they may simply have been levelled areas where various outdoor tasks, such as milking cattle or making butter, could have been performed. Further fieldwork, in particular excavation or geophysical survey, would be required to provide an answer to this question. The two-roomed structure C1 is quite possibly the building Evans published in 1951 as Deer’s
19
Meadow D, although potentially C10 could have been complete at that time and so is also a candidate (Evans 1951, fig 51, d). The identity of his other plan is more elusive, not least because the length shown in the plan, 15 feet (4.6m), differs from the measurement of 11 feet (3.4m) given in the accompanying text. It could match any of the buildings C4, C6, C8 or C9, although none of these exactly match the plan as Evans drew it. It also has
UJA2012.qxd:UJA 2005
22/4/12
10:56
Page 20
to be noted that the 1958 plan of the sites indicates a total of eight buildings, but with six buildings in a line to the north of the stream and only two buildings to the south (Evans & Proudfoot 1958, fig 1). Ignoring the empty platforms, the recent survey found five buildings north of the stream and four buildings to the south. Taking the Evans and Proudfoot plan at face value, it would seem that they may have missed C10 and C11 to the south of the stream and that they included one of the empty platforms north of the stream, or that a building north of the stream was missed during the more recent survey. It is of course possible that a building that was close to the northern edge of the stream has subsequently been washed away. The small circular stone-built fodder stand was situated to the north of the group of structures (Figs 1, 2). It was well constructed of small boulders (c 0.2m) positioned in interlocking courses. It measured 2.75m in diameter and in places survived to 0.7m in height, which may have been close to its original height judging by the lack of collapsed material around it. That this structure was so well preserved suggests that it is relatively recent and may have been in use during the final phase of this settlement’s occupation, or possibly that it post-dates the settlement entirely. There is a distinct possibility that other small structures lie hidden beneath the heavy vegetation surrounding this settlement.
DISCUSSION There are clearly three different forms of structure in the Deer’s Meadow Hut Group C: a rectangular form, a square form and a small oval form. The implication here may be of an extended chronology. The presence of ‘empty’ platforms may be consistent with the activity of stone robbing during a prolonged period of occupation, as has been recorded at other transhumant settlements, such as Nant Criafolen in North Wales, although it is far from certain that the empty platforms here ever held buildings (Allen 1979, 52–4). On typological grounds, the square structures C2 and C3 are consistent with an 18th- or 19th-century date, in line with the historical account and the excavation at Deer’s Meadow Hut Group A (Evans & Proudfoot 1958, 134–35; Rathbone 2009, 120). It is extremely difficult to date the other structures accurately on typological grounds, although they may be earlier than the square examples. A tentative date range from the 14th to the 17th century is proposed on the basis of known dates from other transhumant structures in Northern Ireland, but excavation would be required to confirm or revise this estimate (Rathbone 2010, 121). Transhumant settlements in Northern Ireland have been subject to important but sporadic
20
investigations, including sites in Co Antrim at Craigs and Glenmakeeran and the disputed site at Goodland, as well as Deer’s Meadow, Co Down, which was the subject of earlier investigations (Williams & Robinson 1983; Williams 1988; Case et al 1969; Horning 2004; Evans & Proudfoot 1958). A recently published survey of transhumant sites in the Mourne Mountains by Mark Gardiner is a major contribution to the study of this settlement type in Ireland, and for the first time we have an Irish equivalent of the detailed surveys that have been published for several areas of Britain, such as Perthshire in Scotland or the Black Mountains in Wales (Gardiner 2008; Stewart 1990; Ward 1991). Gardiner’s study identifies 24 separate transhumant settlements located on the northern side of the hills and includes the three settlements at Deer’s Meadow, with Groups A, B and C relabelled as Sites 13, 14 and 15 respectively (Gardiner 2008, 150). Gardiner divides the Mourne booley houses into two types. Type A is an oval or almost circular type often divided into two or more cells, whilst Type B is a small square building. Gardiner’s plans show numerous examples of rectangular multi-roomed structures, but he considers each room to be an individual building of his Type B form that has simply been built in a conjoined row. This is an interesting idea and in certain instances such as the three-roomed structure at Castle Bog River or the two adjacent square buildings here, C2 and C3, it is a convincing explanation. However interconnecting rooms may be difficult to identify prior to excavation and certainly other regions demonstrate numerous examples of rectangular structures with two or three interconnecting rooms, such as Nant Criafolen and Aber Valley in North Wales, Dean Moor in Devon, and numerous examples from the Lake District in Cumbria (Allen 1979; Butler 1962; Fox 1958; Ramm et al 1970). This issue could be resolved through excavation of several different examples of these rows of ‘cabins’ or multi-roomed structures. Such work could focus on determining the presence or absence of breaks in the dividing walls, determining whether external entrances were provided for each compartment, whether hearths were present within each compartment, and whether artefacts differed significantly from compartment to compartment or represented a repeating assemblage. The difference between Gardiner’s simple typology, where all of the Mourne booleys fall into one of two categories, and the more complex scheme used by the author is very noticeable. Of the five regions that informed the development of the more complex typology, the Mourne Mountains were seen to be the most diverse area, with six out of the eight types being present (Rathbone 2009, table 5.1). This may
UJA2012.qxd:UJA 2005
22/4/12
10:56
Page 21
do more to highlight the arbitrary nature of the typologies than to illuminate a significant pattern in the data. As more information regarding the dates and styles of different buildings becomes available it may be possible to better understand these differences in the recorded levels diversity amongst transhumant buildings in this region. The fodder stand to the north of the settlement is an interesting and unusual feature. A whole range of small additional structures associated with transhumant settlements have been recorded elsewhere, and they are found to be a regular but far from universal occurrence (Rathbone 2009, table 5.2). The range of functions that these additional features provided for is very poorly understood (Rathbone 2010, 112–21). An almost identical structure to the fodder stand was found next to two hafods at Ynys Ettws in North Wales and was described as a stack stand by the investigator (Smith 1997, 11). However the term ‘stack stand’ is more properly used to describe small platforms where hay stacks were built surrounded by drainage channels or raised above the surrounding ground, which are common in the uplands of Northern England (Ramm et al 1970, 54). Despite the disagreement over the terminology, the function suggested by Smith may well be correct, that hay or braken was placed into them for tethered animals, possibly ponies or young calves, to feed on. An appropriate analogy is with the galvanized steel roundabout feeders which are a regular feature of the modern rural landscape. More work is clearly needed to understand the function of the array of small features found in association with transhumant settlements, with soil phosphate analysis being one potentially revealing avenue of investigation. It is worth mentioning that in the pioneering study of transhumant structures in Cumbria any building associated with additional features of any form, such as stack stands, fodder stands, out buildings, enclosures or agricultural plots, was automatically classified as a small farmstead and left out of the discussion (Ramm et al 1970, 2). This seems to have been an overly harsh distinction and a whole range of features relating to diverse agricultural practices have subsequently been shown to be associated with transhumant structures from different periods and locations (Rathbone 2009, 121). The presence of the fodder stand, located at some remove from the other buildings, should therefore not be used to suggest a more permanent style of occupation of this site. It is hoped that the publication of this detailed account of Deer’s Meadow Hut Group C will make a small but useful contribution to the expanding knowledge of this important but often neglected farming practice.
REFERENCES Allen, D 1979 ‘Excavations at Hafod y Nant Criafolen, Brenig Valley, Clwyd, 1973–74’, Post- Medieval Archaeology 13 (1979), 1–59. Butler, L A S 1962 ‘A long hut group in the Aber Valley’, Trans Caernarvonshire Historical Soc 13 (1962), 1–59. Case, H J, Dimbleby, G W, Mitchell, G F, Morrison, M E S & Proudfoot, V B 1969 ‘Land use in Goodland townland, Co. Antrim, from Neolithic times until today’, JRSAI 99 (1969), 39–53. Evans, E E 1951 Mourne Country, Landscape and Life in South Down. Dundalk. Evans, E E & Proudfoot, B 1958 ‘Excavations at the Deer’s Meadow’, UJA 21 (1958), 127–31. Fox, A 1958 ‘A monastic homestead on Dean Moor, S. Devon’, Medieval Archaeology 2 (1958), 141–57. Horning, A J 2004 ‘Archaeological explorations of cultural identity and rural economy in the north of Ireland: Goodland, County Antrim’, International J Historical Archaeol 8.3 (2004), 199–215. Ramm, H G, McDowall, R W & Mercer, E 1970 Shielings and Bastles. London. Rathbone, S 2002 Transhumant Britain: A Comparative Study of Transhumant Settlement From Around the Northern Basin of the Irish Sea. Unpublished BSc thesis, Bournemouth University. Rathbone, S 2010 ‘Booley houses, hafods and sheilings: a comparative study of transhumant settlements in and around the northern basin of the Irish Sea’, in Horning, A & Brannon, N (ed), Ireland and Britain in the Atlantic World, 111–29. Bray. Smith, G 1997 Three Medieval Rural Settlements in Gwynedd, Excavations in 1997. Unpublished report prepared for Gwynedd Archaeological Trust. Stewart, J 1990 Settlements of Western Perthshire: Land and Society North of the Highland Line 1480–1851. Edinburgh. Ward, A H 1991 ‘Transhumant or permanent settlement?’, in James, H (ed), Sír Gar: Studies in Carmarthenshire History, 1–23. Camarthen. Williams, B B 1988 ‘A late medieval rural settlement at Craigs, County Antrim’, UJA 58 (1988), 91–102. Williams, B B & Robinson, P S 1983 ‘The excavation of Bronze Age cists and a medieval booley house at Glenmakeeran, County Antrim, and a discussion of booleying in north Antrim’, UJA 46 (1983), 29–40.
21
