Site Report

WoSAS Pin 57999
NM 4967 5406

A substantial sub-rectangular building within a rectangular enclosure, possibly representing the remains of a chapel and burial ground, stands on a ridge about 700m to the SSW of Baliscate House. There is a second rectangular enclosure to the south which maybe related. The site stands in a forest clearing on the slopes of Coille Creag A'Chait' at a height of approximately 108m. The building is situated on an east-west axis, with an entrance in the north wall. It measures 8.7m in length and 6.8m in breadth, over walls, some 0.9m thick. The enclosure around it occupies and east-west axis, with an entrance in the east wall. It measures 18.7m by 15.7m over walls some 1.2m thick. The walls on both structures stand to a maximum height of 0.7m and appear to be of turf and stone construction. The second enclosure is approximately square with an entrance in the south wall. It measures 12m by 12m over drystone walls some 0.9m thick, which stand to a maximum height of 0.6m. During recording some limited uncovering revealed a stone slab some 0.47m x 0.36m lying in the second enclosure. The slab has the appearance of a straight cut along one of its long axis. In the first enclosure there is a smaller stone with a dished appearance . The site is shown on ariel photographs in particular Ordnance Survey photograph OS72-080-292 which is held in the RCAHMS library. The recollections of Hector MacKinnon, carpenter Baliscate, state that according to local tradition there was a church on the hill behind the crofts at Baliscate and a cemetery near the Tobermory Poorhouse - The Poor House stood on the south bank of the Tobermory river, some 250m north west of this site (Source Isle of Mull Museum, Tobermory, Mull)
Marsh, H. & Langhorn, B., Scotland's Rural Past Project 2008
Entered WoSAS (MO'H) 11/02/2009

Resistance Survey
Resistance data were collected at 0.5m intervals along traverses spaced 0.5m apart in order to gain a higher resolution of the surveyed area. The data have also been subjected to a high pass filter in order to suppress the background geology.

Areas of high resistance (A) correspond to the northern, western and parts of the enclosure surrounding the chapel. The southern section of the enclosure is not shown within the data, possibly indicating that this site comprised an outer bank with no stone revetting.

The chapel can be seen within the data as areas of high resistance (B). There is a slight difference in the data along the northern section which corresponds to the break in the earthworks.

High resistance responses (C) correspond to the extant wall remains of the smaller square enclosure. A mound of stones has caused the response in the very south eastern limits of the data.

GPR Survey
A very small area immediately north of the chapel was surveyed with radar in order to identify any abutting structures or possible burials. Given the age of interments in this area, the detection of the latter would, in reality, only be possible if they were stone-lined or buried in solid caskets. (as opposed to wood or a simple shroud).

The pattern of response across the survey area is characterised by sporadic zones of increased amplitude, such as (D), and areas of diminished response. These are assumed to be a facet of varying soil composition within the material used to level the site and natural features below. The low amplitude linear trends, for example (E), may be shallow drainage cuts running away from the main structure toward the boundary wall.

An exploratory trench was put in to investigate the seemingly more coherent reflector (F) which, given the depth of approximately 1.40m+, could have indicated a collapsed, stone-lined grave or other solid feature. The excavated material contained a number of large stones, suggesting an origin for the random high amplitude responses recorded across the site. The trench bottomed out on a former ground surface at around 0.8m with dating evidence recovered suggesting this was a prehistoric level. Anomalies beyond this are therefore natural reflectors.

Conclusions
High resistance responses relate to the earthworks of the chapel and that of the smaller square enclosure; they indicate stone revetments on three sides of the chapel. GPR data have potentially discovered a prehistoric buried surface but due to the difficult ground conditions and natural features no definite features associated with the chapel were discovered.
Gater, J., 'Geophysical Survey Report: Baliscate, Tobermory, Isle of Mull', GSB Prospection Ltd, (2009)

The earliest phase of activity observed on the Site may be prehistoric in date, comprising evidence of agriculture in the form of the ard marks recorded in Trench 2, and a possible stone cairn in Trench 5. The ard marks were distinctively narrow (less than 0.05m wide), which is not particularly diagnostic of medieval or later ploughing, and these are therefore perhaps Neolithic or Bronze Age in date. The function of the cairn is unclear; it may be a burial cairn or it may be just a clearance cairn. The structure is slightly irregular as the possible kerb does not encompass the whole structure, and the position of stones may be purely coincidental, but it is possible that material was taken from the cairn to use in the construction of the buildings of the monastic complex. Without further excavation, a date or function for the cairn is unclear.

A number of prehistoric sites are known from Mull, including the Baliscate standing stones (NM499541) c. 280m to the east of the Site and a flint arrowhead recovered from near the stones (WoSASPIN 566), with several other standing stones and stones circles located around the island, including Ardalanish (NM378189), Ardnacross (NM 542491), Dervaig (NM439520) and Lochbuie (NM618252, NM617255).

Early Christian
Time Team were invited to Baliscate to investigate what would turn out to be just a small part of a much larger monastic complex. The identification of the surrounding monastic vallum; a physical and spiritual barrier separating the ecclesiastical from the secular, the approaching hollow-way, and entrance through the vallum, put the chapel in the middle of an enclosure measuring approximately 210m by 100m and covering an area of 1.37 hectares. The enclosure is tiny when compared to such as that on Iona, where by the 8th century the vallum surrounded an area of approximately 8 hectares (Fisher 1997, 185-6).

It is clear that Baliscate was never a large religious community, but considerable effort went into the preparation of the Site, to create a level platform for the chapel, by the construction of a c. 1m high revetment and the dumping of tons of earth to level the Site, and surround it with stone and turf bank for over half of its 500m perimeter.

The construction of the terrace provided not only a level construction platform but an elevated position, providing a vista across the Sound of Mull to mainland Scotland and Ardnamurchan, where St. Adomnan recalled St. Columba baptising the inhabitants (Smyth 1989, 113).

The earliest phase of the chapel comprised a timber sill beam structure, with one associated inhumation burial, radiocarbon dated to 610-690 cal. AD; this potentially places the burial within the lifetime of St. Adomnan (c. AD 627- 704). The full extent of the timber chapel is unclear, as it was overlain by later construction. The timber chapel was perhaps smaller than the stone chapel that replaced it, or perhaps on a slightly different alignment or position - the associated grave was clearly sealed beneath the later walls.

Information regarding early Celtic Christian timber chapels is sparse - few have been excavated because they are generally found below existing stone structures; however, one has been excavated beneath the oratory at Illaunloughan in Munster.

On the Brough of Deerness off mainland Orkney, excavation revealed a pre- Norse timber-built chapel clad on two sides in stone, which as the site developed was rebuilt or consolidated in stone on all sides (Morris and Emery 1986, 301-74). It would appear that the nature of the stone structures was dependent on the acquisition of usable stone. At Deerness, stone that could be shaped was available, whereas at Speke Keeill on the Isle of Man, a probable 9th century chapel excavated in 2006 (Wessex Archaeology 2007) revealed use of unshaped stones, revetted with turf, and it is likely that a similar construction technique was utilised at Baliscate. The walls of the chapel appeared to have an earthen core, but it is also possible that the structure was partially revetted in turf as well.

The stone chapel appears to have been an entirely new build, sealing the earlier grave, with the construction of the leacht at the east end and a series of graves against the southern wall. No date was obtained for this phase of building although the stone cross fragment has been tentatively dated to the 8th century. The identification of the leacht is extremely rare in Scotland; most information on these structures comes from Ireland. This is not unexpected as the chapel probably has a probable association with St. Columba. However, on Iona, excavations on top of Torr an Aba to the west of the Abbey revealed a cross base not dissimilar in appearance to the leacht at Baliscate (Fowler and Fowler 1988, 18).

A number of leachta have been identified throughout Ireland, for example at Illaunloughan (Munster), Skelling Michael (Kerry), and Innishmurray (Sligo). The function of the leacht, however, is still not entirely understood and a number of uses are possible, from external altar to a founder's shrine or reliquary (Monk and Sheehan 1998, 105; O'Sullivan and O Carragain 2008). The identification of the mortise hole to take an upstanding stone cross within the fabric of the leacht at Baliscate implies that they also acted as foundation bases for such crosses. This is confirmed by the association between leachta and stone crosses at Skellig Michael in Co. Kerry (http://www.environ.ie/en/Publications/Heritage/ManagementPlans/FileDown Load,14828,en.pdf).

Like Baliscate, the leacht at Illaunloughan was covered with white quartz pebbles left as offerings (Monk and Sheehan 1998, 105). At Baliscate it was clear that the water worn pebbles, not all of them white quartz, had been left on top of the leacht and had been later incorporated into the overlying rubble layer as the structure collapsed or was demolished. No stones were found within the material of the structure; they were clearly brought in after its construction. The deposition of white pebbles is known from prehistoric 16 contexts although they are also intrinsically linked with St. Columba and the Celtic church throughout Dal Riata, the Isle of Man and the Western and Northern Isles of Scotland; St. Adomnan recounted the teachings of the founder and first Abbot of Iona and also recorded his travels, and in one particular chapter the use of a sacred white stone for healing the sick.

The chapel at Baliscate is likely to have had an association with Iona and was founded by followers of the teachings of St. Columba. One of the inhumation graves showed evidence of the Iona connection - the head end was marked by a headstone of Iona marble, and from the grave backfill came a fragment of red granite, similar to that found at Fionphort, on the south-west coast of Mull, facing Iona.

It is possible that there would have been more ancillary buildings within the monastic complex, but only a single building was observed. Constructed in a similar manner to the chapel with stone and earthen walls, the building in Trench 4 is likely to have been contemporaneous, but the lack of datable material makes this difficult to prove.

Medieval to Post-medieval
The abandonment and subsequent demolition of the chapel is not firmly dated, but activity in the medieval period (late 12th to 15th century) is implied by the pottery and the coin of Edward II. There is a possibility that the chapel was reused as a domestic structure, but the paucity of finds implies only temporary use of the Site, most probably only during demolition.

The chapel was extensively robbed of usable material to form the sheep fank and shieling to the west. No dating for this structure was recovered.
Wessex Archaeology Ltd, 'Baliscate Chapel, Isle of Mull: Archaeological Evaluation and Assessment of Results', (2010)

There is undoubtedly a chapel and leacht at the Baliscate site, and the argument for the presence of a monastic enclosure / monastery is also fairly convincing. However, a lot more field investigation and excavation would be required to prove this.

Baker holds a different opinion in relation to the interpretation of the later phases of the site. Time Team / Wessex have interpreted the site as demolished, which implies deliberate and concerted destruction rather than more gentle decay, which may be more likely, though there is no doubt that the chapel has been extensively robbed. Under the Time Team / Wessex interpretation, Layer 110 in the interior of the building is interpreted as representing demolition, rather than occupation; however, Baker suggests that Layer 110 probably does indicate occupation. She also suggests that the stone structure we see and interpret as a chapel is in fact a remodelled version of the chapel, and in its latest phase was in use as a dwelling rather than as a chapel. Wessex argue for non-occupation based on the paucity of finds, but Baker suggests that 34 pot sherds and a coin is a fairly rich assemblage for a rural site on the West Coast of Scotland.

The presence of oats and barley in layers below the artefact-rich deposit 110 and above the gravel layers interpreted as floors may also suggest agricultural use / occupation rather than simply wall core collapse and spread. Although there is no reason why this could not be religious occupation and use as opposed to secular, the presence of oats and barley is notable. However, very few soil samples were collected and processed, so the picture is unclear.

Baker is also not convinced that the charcoal spreads can be interpreted as evidence of a timber chapel. They might be the base of coffins, as there are examples of charred timbers being used for coffins in the early Christian period (DES 2009 p150-151 includes a report of a charred log coffin founf at Forteviot). There may well be an earlier timber chapel, though the evidence may be slightly stretched to assert this.

None of the archaeologists from FIRAT Archaeological Services dug the leacht or the interior of the chapel as these were excavated solely by the Time Team, so Baker is unable to comment in detail of the interpretation of the stratigraphy in these areas.

In general, Baker does not disagree with the Time Team / Wessex report, although she believes that there are elements that have been omitted – for example, she notes the presence of a possible construction cut for the E wall of the stone chapel, which does not appear in the report. However, there will be no further work on the site by Time Team / Wessex, and as FIRAT Archaeological Services were also present on the site, she has provided comments and annotations on the interpretations contained in the project report in the hope that the site may be taken up as a research project.

For full comments on the Time Team / Wessex report, see WoSAS Archive 3754 (annotated version of original Time Team / Wessex report)
Baker, F., FIRAT Archaeological Services (2010)

Site scheduled on the 19th of July 2011 as 'Baliscate, chapel 655m WSW of.' Monument comprises the remains of a small, early historic chapel and burial ground, set within a rectangular enclosure. A second square enclosure, of unknown date, lies immediately to the WSW. The remains are visible as low earthworks. The chapel and burial ground were in use in around the 7th to 8th centuries AD, and the chapel was later re-occupied, probably between the 12th and 15th centuries AD. The site lies about 1km south-west of Tobermory, occupying a level terrace on NE facing slopes overlooking the N end of the Sound of Mull. It stands at 100m above sea level, and offers views north-eastto the Ardnamurchan peninsula.
Text from Historic Scotland scheduling document
Entered WoSAS (MO'H) 01/08/2011