During training excavations at Dartington Hall, Devon in July 1996 three further trenches were excavated under the direction of C K Currie for the Gardens Archaeology Project. These included two trenches on the main lawn to further explore the area around a square-planned structured called the 'tower' by Colin Platt, following his discovery of it in 1962. A third trench continued work on the east end of the Bowling Green, where a stone wall had been found in 1994.
The excavations continued to make discoveries regarding the phasing of the southern courtyard buildings, and has suggested that conclusions drawn in previous years may need revising. It was not found possible to find the edge of a conjectured ditch suggested by Platt in his 1962 excavations. This brings up the possibility that the 'ditch' may not exist. Instead, it is possible that the original ground level was much lower here than elsewhere, although it still has to be considered that the 'ditch' may have been wider than the extent of this year's excavation.
The demolition of the southern courtyard still seems to have occurred c. 1700-20, but it is possible that there may have been other levelling episodes before this. The only dating evidence for building fragments and materials continues to lean towards a later 15th or early 16th century date. This is largely based on the discovery of a large fragment of square-headed Perpendicular Gothic window of this date found in trench 14. Two stone walls excavated south of the tower suggest that two different phases followed the building of this structure. The latest of these may have been connected with a garden wall possibly put up after c. 1700, and demolished between 1801 and 1839. The only record of this wall is a watercolour in the Hall of c. 1801 showing it to have been of considerable height. By the time of the tithe map in 1839, the wall is not shown.
Evidence was found for garden planting, possibly for espaliered plants on the west side of a subsequently demolished wall at the east end of the Bowling Green. The dating of this wall is still uncertain, but the balance of the evidence suggests it may have been built c. 1700, although there is still a chance that it is earlier.
A trench excavated north of the conjectured 'tower' failed to find the old ground surface at a depth of 1.35m when excavation was stopped for safety reasons. Much rubbly build-up was found overlying a thick layer of crushed mortar that contained a large number of fragments of medieval ridge tiles. Two distinct fabric types were recognised.
Dartington Hall stands on high ground above the River Dart, to the north of Totnes, in the County of Devon (NGR SX 798 628). The Hall has been described by Pevsner as the 'most spectacular medieval mansion of Devon' (1952, 99). It was the principal seat of John Holand, half-brother to Richard II, and has had a chequered history. Additions to the present structure were made under the ownership of the Champernowne family in the 16th century and again in Georgian times. Further remodelling of the Elizabethan living rooms were carried out in the 19th century under Henry Champernowne (Hussey 1938, 232). The estate declined as a result of the agricultural depression of the later 19th century and the Hall was much neglected when purchased by the Elmhirsts in 1925. It was under Dorothy Elmhirst's direction that the gardens were laid out in their present form.
During the Second World War (1939-45) German bombing of Exeter was thought to have caused the destruction of many of the estate archives, then deposited in the County Record Office. This has caused a number of problems for the future management of the gardens, as it was thought not possible to determine what form the earlier gardens had taken from archival sources. To enable a better understanding of the garden, and as an aid to future management decisions for the continuing evolution of the gardens, it was decided to carry out an archaeological excavation over a number of seasons.
It was decided to fund these excavations by running an annual training excavation in conjuncture with the Dartington Hall Centre. It is hoped to combine the annual excavation reports into a more definitive document at the end of the project. An outline of the proposals has been presented to the Trustees and the County Archaeological Officer in the form of a project design (Currie 1993), written to specifications for such work laid down in The management of archaeological projects (English Heritage, revised edition, 1991).
Previous work in 1991, 1993, 1994 and 1995 has deduced a number of points about the archaeological remains within the gardens.
A combination of archaeological excavation and historical research seems to confirm that the Dartington terraces are unlikely to be associated with a medieval tiltyard. All the recognised authorities now seem to agree that they were made as garden features in the post-medieval period (Platt 1962, Emery 1975, 151-2). Nevertheless, the evidence as to exactly when they were built is not conclusive, and further work is needed to resolve this problem. Historical research discovered that the terraces had once ended at the level of the first terrace up from the present Tiltyard floor. In the 19th century a formal garden existed here called the 'Dutch Garden' and the 'Rose Garden'. The Twelve Apostles can be seen on early photographs as part of these gardens (Snell 1989, 19, 22). This garden was destroyed under Dorothy Elmhirst's direction, who cut out the present floor level, reducing the lowest level of the terraces by a further 2m.
Two large stone-built conduits were located under the terraces. These were associated with water control and supply. The uppermost of these was associated with a stone-built conduit house at the top of the southern terraces. It is possible it once supplied the Hall with drinking water, but the original conduit was cut off by the building of the terraces. Later a pipe was inserted to take water to the feature known as the Swans, and the old entrance to the stone conduit was covered over.
The second conduit took a small stream, fed by springs at the head of the valley, under the former Dutch Garden. When the new floor of the Tiltyard was cut out after 1927, this feature was left at a higher level.
Of the non-garden features found on the private lawn, it can be concluded that there is more than one phase of building in the area known as the south court. Generally the 1993 excavations confirmed Platt's earlier dating (Platt 1962), but showed the situation to be more complex than he envisaged. The exact dating for the construction of the south court can only be conjectured at present, but it is possible that a building on the site of part of the earlier court was still in existence in 1839, possibly as a garden building. Until 1995 this author considered that this structure coincided with Platt's 'tower', a building found in 1962, but recently a more careful study of the evidence has suggested that this might now be questioned.
In 1994, part of the plan of the 'tower' was revealed. This appears to have been a late medieval or early post-medieval building. Further excavation is needed to clarify points about this building, but it may have had a basement or cellar beneath it, only part of which was excavated.
A substantial stone wall was discovered at the east end of the present Bowling Green, but it was not possible to date it in 1994 beyond the fact that it seemed to cut across medieval horizons. A piece of local coarseware pottery was found mortared into the wall in 1995, suggesting a post-medieval date, although this needs to be confirmed.
Further excavation on the first terrace of the Tiltyard failed to locate traces of the Dutch Garden. Although this is clearly shown on earlier photographs, it must now be assumed that all evidence for it was removed during the earth-moving activities required to create the present layout of the Tiltyard.
The trenches excavated in 1996 continue the numbered sequence began in 1991. Three trenches were excavated this year beginning with trench 14, and finishing with trench 16. The strategy was based on the continuing need to resolve unanswered questions concerning the evolution of the garden.
Two main themes continued to be explored. The first of these was to recover further data on the structure known as the 'tower' (trench 14). The 1996 excavation concentrated on trying to complete the excavation of a feature thought to be the tower 'ditch' by Platt (1962). This year attempted to find the southern edge of this proposed feature. A new avenue of enquiry opened in 1994 tried to determine the date of origin and evolution of the feature known as the Bowling Green. This was continued in 1995 with the excavation of trench 12, and was further examined in trench 15 in 1996.
Another small trench (16) was opened nearer to the Georgian wing than the 'tower' to test the archaeological survival in this area.
This was an L-shaped trench, the main trench being 4m N-S by 2m E-W, with an eastwards extension at the southern end measuring 2.5m E-W by 1.5m N-S. It was excavated adjacent to the east side of trench 11 (excavated in 1995). This partly re-excavated trench 3 from 1993. As well as attempting to locate the edge of Platt's conjectured 'tower ditch', this trench explored the relationship between the 'tower' and an adjoining wall on the south (given number 018 in 1993).
A number of interesting and unexpected findings occurred during this season's excavation here, causing some re-interpretation of earlier work. The first of these to be noted was the failure to recover a cut for the postulated 'tower ditch'. Instead the soils recovered were a series of rubbly dumps (224, 227, 241) overlying a possible earlier ground surface of brown clay (243). Context 224 was considered at least partly contaminated as this level had been partly removed over this area in 1993. It seemed to be approximately equivalent to possible 18th-century levelling layers to make the present lawn terrace. Contexts 227 and 241 may have also been related to this levelling, but were apparently different levels.
What was uncertain was whether this levelling had all been undertaken in one episode. It is possible it was not as the type of finds in 227 and 241 were different. Context 227 contained more fine wares, including 17th and early 18th-century tin-glazed wares, whereas 241 contained mainly local coarsewares of late medieval or post-medieval date. These dumps were excavated for 2.4m south of the south 'tower' wall (244), but no cut for a ditch was found. This led to the suspicion that a ditch may not exist on this side of the tower, and that the artefact rich layers butting against the tower were simply levelling dumps that extended all the way to the terrace revetment wall, some 6m further south. Platt (1962, 210) had excavated less that a metre south of the tower to draw his conclusion that there had been a ditch around it. On retrospect, this might seem to be an insufficient sample to enable complete reliability of the inference.
The excavation of these levels enabled a profile of the wall 225 to be seen. This butted against the southern 'tower' wall (244), and was not apparently cut through by it as was previously thought. The north end was not bonded into wall 244, but had been built up against it, following the slight banter of 244 outwards towards the bottom. The wall survived to a depth of 0.8m (1.35m) below the present ground surface), whereas the foundations of 244 extended to a greater depth.
Wall 225 was 2.95m in length before it turned eastwards for a further 0.6m. This eastward wall was given context number 228, although it was to all intents the same wall as 225, being bonded into the latter. Wall 228 ended in a manner that suggested interruption. There was no sign of a clean end, the eastern terminal being of an irregular nature.
0.86m further to the east was another wall (233). Between 228 and 233 there seemed to be a level containing what appeared to be demolition materials (234). This deposit contained a number of post-medieval coarseware pottery sherds, and a large fragment of high status carved stone. This appears to have been part of a late Perpendicular flat-headed window, dating from the later 15th or early 16th century.
The wall (233) was 1.1m thick, but had relatively shallow foundations. It aligned approximately with the western edge of the original Georgian wing. This alignment matches up to a high wall shown on a watercolour of the hall dating from c. 1801. This wall had apparently gone by the time of the tithe map in 1839. Wall 233 was overlain by a lighter brown clay loam (255), suggesting a levelling of the lawn after the early 1800s.
A trench 4m by 2m was excavated to further explore a distinctive parch mark at the east end of the Bowling Green further. In 1994 this had proved to be a substantial wall (101) of local stone 0.82m thick. Further work was undertaken to try to obtain better dating for this feature. It was also decided to explore the area on either side of the wall, work that had been begun in 1995, but had been left unresolved.
Again the wall (231) was located as expected directly below the parch mark in the soil about 0.1m below the present surface. A well-defined construction cut (235) was observed parallel with the wall on the east side. This was 0.3m wide, and appeared to cut through a mortar layer (237) parallel with the wall. However, as this layer existed on both sides of the wall, it could be trample created by the workmen building the wall. Although it is not impossible that it existed before the wall, a reason for its earlier existence would need to be found. The mortar was, in places, up to 0.1m thick, and it overlay a brownish clayey soil (238) characteristic of sub-soil levels in the vicinity. The construction cut was about 0.6m deep. It came down on to a step in the wall, making for a wider foundation near its base.
On the west side of the wall, another foundation cut was located (247). This was much narrower than that on the west, being only 0.05m wide in places. The step in the wall to make a wider foundation was also much higher in the soil profile. Here it was approximately 0.2m higher than on the east side. A mortar layer was also located on this side of the wall at a depth of about 0.55m. The edge (249) of a 1994 season trench (trench 8) was also located within 0.1m of the northern edge of the trench.
Other cuts within the mortar layer (251, 253) were thought to be genuinely historic. 251 was a semi-circular cut butting up against the wall (231). It was not particularly deep, being cut only just over 0.1m into the mortar. However, it is thought that the cut derived from the original ground surface. It may have been a planting pit for an espaliered plant against the west facing side of the wall. The second cut (253) was more linear, being apparently at right-angles to the wall. This did not seem to have been particularly deep either, just cutting through the mortar level. However, it was probably cut from a higher level like 251, and may have been for a similar purpose.
A trench 3m by 1m was excavated north of the site of the tower to investigate the survival of archaeological features here.
Topsoil was about 0.2m deep in this trench. This was followed by a deep layer of rubbly material, mainly slatey material mixed with some stone and fragments of ceramic ridge tile (239). This layer was up to 1.05m deep. It became progressively deeper towards the west end. At the east end of the trench it was only about 0.5m deep. This overlay another thick layer (240). This time it was composed mainly of crushed mortar. The bottom of this layer was not reached at a depth of 1.35m, and excavation was stopped for safety reasons. As with layer 239, it dropped off sharply to the west, as if following some drop in old ground level here. Layer 240 contained much broken ridge tile of two different types. A thicker coarse tile, with incised designs, a green glaze and pointed peaks was found together with a similar tile in a finer fabric with pointed peaks along the ridge. Platt (1962, 217) noted these tiles, but concluded that they came from different horizons. This may not have been the case in this trench, as both types seemed to be mixed together in the same contexts.
Little could be said about this trench other than the fact that the ground surface seemed to be rather deep here. The abundance of roof tile might be thought to demonstrate that a roof was demolished here, but the ridge tiles in the mortary layer were found spread right across the site by Platt (ibid).
The information found this year was more enlightening than in previous years in that it caused a number of hypotheses to be reinterpreted. Firstly it would seem a possibility that a 'ditch' may not exist for the tower. No cut was found 2.45m south of the tower wall where excavation stopped. It is therefore possible that one did not exist, and that for some reason the old ground surface is much lower here than elsewhere. This is odd because on would expect the old ground surface to rise the further west one goes.
Another point that needed reinterpretation is that wall 225 appears to be later than the tower, and not earlier as thought previously (Currie 1993). In fact there is some reason to think that wall 225 was built against the 'tower' wall after the ground level nearby had risen somewhat. This would appear to have been before the demolition phase of c. 1700. If the 'tower' really is late medieval or early post-medieval, as suggested by Platt (1962, 216-17), when did this build-up occur?
The newly discovered wall this year (233) was built at a different level again, possibly after the c. 1700 build-up levels had been deposited. Pictorial and cartographic evidence seems to suggest that the wall was demolished between 1801 and 1839.
Trench 15 on the Bowling Green proved to be more enlightening than in previous years. Construction cuts for the stone wall here were found on both sides. Also found on both sides was a layer of mortar, possibly at the level from which the original wall was built. On the west side features were found cut into the mortar against the wall, suggesting that they were planting pits for espaliered plants, possibly fruit trees, against the wall. The mortar level seemed to be more consistent on the west side of the wall, a point observed in 1994 (Currie 1994). This indicates that it would be a useful exercise in another year to try to find out how extensive this mortar layer is, and to try to recover some sort of planting pattern on the west side of the wall.
Trench 16 was not particularly forthcoming about the archaeological survival north of the 'tower'. One thing it did reveal was the great depth of the old ground surface in this part of the garden. This still has not been explained satisfactorily. Platt suggested that there might be a ditch around the 'tower' to explain the great depth of soil, but this might be called into question. In particular the apparent continuation of this unusual depth in the vicinity of trench 16 suggests that the discrepancy extends well beyond the confines on any conjectured ditch around the tower.
The excavations continued to make discoveries regarding the phasing of the southern courtyard buildings, and has suggested that conclusions drawn in previous years may need revising. It was not found possible to find the edge of a conjectured ditch suggested by Platt in his 1962 excavations. This brings up the possibility that the original ground level was much lower here once, although it still has to be considered that the 'ditch' may have been wider than the extent of this year's excavation.
The demolition of the southern courtyard still seems to have occurred c. 1700-20, but it is possible that there may have been other levelling episodes before this. The only dating evidence for building fragments and materials still seems to lean towards a later 15th or early 16th century date. This is based on the large fragment of square-headed Perpendicular Gothic window of this date found in trench 14. Two stone walls excavated south of the tower suggest that two different phases followed the building of this structure. The latest of these may have been connected with a garden wall put up after c. 1700, and demolished between 1801 and 1839. The only record of this wall is a watercolour in the Hall of c. 1801 by Archdeacon Froude showing it to be of considerable height. By the time of the tithe map in 1839, the wall is not shown.
Evidence was found for garden planting, possibly of espaliered plants on the west side of a subsequently demolished wall at the east end of the Bowling Green. The dating of this wall is still uncertain, but the balance of the evidence suggests it may have been built c. 1700, although there is still a chance that it is earlier.
The discoveries this year offer exciting new possibilities for future years in that features associated with the old gardens are beginning to be found. Furthermore the excavations are beginning to come to terms with the possibility of obtaining dating for the older buildings of the southern courtyard.
Again, a number of interesting finds were made, adding to the extremely fine assemblage collected to date. As there is a possibility that excavations might continue in future years, it has been decided to leave a full assessment of the finds until the end of the project.
Following the Coroner's decision on July 25th 1996 at Totnes Court, it is now possible to report on the exceptionally fine silver with gold gilt pin that was found in context 204 in 1995. This was declared by the court not to be Treasure Trove, and its ownership was vested in the Trustees with the recommendation that it should be given to Exeter Museum on permanent loan for public display. This should not prevent the Trustees from having the pin back at Dartington for display at the Hall on special occasions if they desired it.
The find has been in the British Museum since August 1995 where it had been cleaned and conserved. The pin was reported on by John Cherry as follows:
'The silver gilt pin found at Dartington Hall Devon was submitted to me on August 15th 1995. It has been cleaned in our Department of Conservation and analysed in our Department of Scientific Research. This analysis shows that the pin contains approximately 95% silver. It weighs 14.35 grams which corresponds to a silver bullion value of £1-60. The present length of the pin which has a bent shank is 93mm.'
The pin consists of two parts - the gilded head and the ungilded shank. The head of the pin is hollow and consists of two hemispheres soldered together. Each hemisphere has six raised bosses surrounded by a double wire strand. On the upper hemisphere there is an applied knob of silver above, and three smaller beads between each boss surrounding a smaller wire circle between the larger circles. This is reversed on the lower hemisphere. The top of the head is surmounted by a knob. Between the head and the shank is a protruding loop. The silver shank is bent at present though how far this reflects the original arrangement is uncertain.
The pin was used for the adornment of dress in some form. The style of manufacture of the spherical head in two parts and its decoration with knobs and wire appears to link this pin with three pins found in post-medieval contexts in Norwich (see Sue Margeson Norwich Households East Anglian Archaeology 58 (1993) p.10). The Dartington pin is likely to be of late medieval/early post-medieval date i.e. dating from the fifteenth or sixteenth century.'
The Trustees of Dartington Hall are gratefully acknowledged for their permission to excavate and continuing support. The author would like to thank the following for their continued assistance, and for the cheerful welcome that is always given to the team: Graham Gammin and his garden staff, particularly Graham himself, Neil and Alan who frequently discuss the findings with the author and freely offer their personal knowledge of the garden; the staff of Dartington Centre who provide the food and accommodation; the staff of the White Hart for the midday refreshments, and the many other staff and friends of the Hall too numerous to mention separately, whose good company makes the occasion so enjoyable.
Finally the author would like to thank the County Archaeologist, Simon Timms, for his continuing support and advice.
At Dartington Hall:
Ordnance survey map 1:2500 scale 1904 edition.
Tithe Map of Dartington, 1839 in archive room.
Photograph of Dutch Garden in Tiltyard, undated but with pencil notes on back suggesting c. 1864, in toilet on first floor of private wing.
Photograph of Dutch Garden in Tiltyard, undated but probably c. 1900, in bar of White Hart; reproduced on page 19 of Snell 1989.
C K Currie, Excavations in the gardens of Dartington Hall, Devon, 1993, unpublished report to the Trustees of Dartington Hall gardens, 1993
C K Currie, Excavations in the gardens of Dartington Hall, Devon: second interim report, 1994, unpublished report to the Trustees of Dartington Hall gardens, 1994
C K Currie, Excavations in the gardens of Dartington Hall, Devon: second interim report, 1995, unpublished report to the Trustees of Dartington Hall gardens, 1995
A Emery, 'Dartington Hall, Devonshire', M J Swanton (ed.), Studies in medieval domestic architecture, London, 1975, 134-152.
C Hussey, 'Dartington Hall, Devon part 1',Country Life, 27th August 1938
C Hussey, 'Dartington Hall, Devon part 2', Country Life, 3rd September 1938.
S Margeson Norwich Households, East Anglian Archaeology 58 (1993)
N Pevsner, The buildings of England: South Devon, Harmondsworth, 1952.
C Platt, 'Excavations at Dartington Hall, 1962', Archaeological Journal, 119, (1962), 208-224.
R Snell, From the bare stem. Making Dorothy Elmhirst's garden at Dartington Hall, Exeter, 1989.