There can only be a handful of Irish church sites that have not been visited and described, however briefly, over the past 150 years. Until recently, Bishop’s Island belonged to this dwindling group.
Situated off the coast of County Clare, about 3km south-west of Kilkee (NGR 0B5B 1595), it is essentially a large sea stack (180m x 90m) with sheer sides rising 40m from the sea’ In 1839 Eugene O ‘Curry (OS Letters) wrote that those ‘of the neighbouring landsmen as have nerve enough’ used it for grazing, presumably by hoisting their sheep up the cliffs on ropes. Strong winds and ‘the frightful steepness of the cliffs’ forced him and his companion, John O’Donovan, to abandon their own attempt to ascend, and their brief description of the remains was based on what they could see from the mainland. In his Handbook of Irish antiquities (1858) William Wakeman warned that visiting the island is ‘only to be effected by a skilful climber, and after a long continuance of calm weather’. He must have fitted the bill, for his description and engravings of the church and clochaun could only have been produced by someone who had mad it there. All subsequent accounts of the site have been based on Wakeman’s.
In an important article entitled ‘How old is Gallarus Oratory?’Peter Harbison (1970) plucked Bishop’s Island from relative obscurity and gave it a small niche in the history of Irish architecture. His thesis was that corbelled drystone churches, which are mainly confined to peninsular Kerry, might have been built as late as the twelfth century. He was, in his own words, acting as devil’s advocate and has since moderated his views considerably, especially in the light of Jenny White Marshall and Claire Walsh’s excavation of an island church site on Illaunloughan, Co. Kerry (see Archeaology Ireland Heritage Guide No. 24). Evidence from that excavation proves that examples of the type can be of eighth-century date. Probably they are not much older than that, as excavated sites in the area reveal that the oldest churches were of wood or sod or some other combination of organic materials, pre-dating stone churches (e.g. Church Island, Illaunloughan, Caherlehillan and possibly Reask). But were they still being built at the end of the early medieval period or even beyond? Harbison argued that Bishop’s Island is important to this question. A south-facing doorway would be untypical of pre-twelfth-century churches in Ireland, yet Wakeman recorded a church apparently of Gallarus type, with a south-facing doorway, on this tiny island.
A Flying Visit
Last October, Brian Cullen Save us the opportunity to visit the island by helicopter. The few hours available did not allow a full-scale survey, but we were able at least to assess the visible remains, including some features not mentioned by Wakeman. Apart from the church, there is evidence for three or four other buildings, though substantial remains only survive for a clochaun. Despite some collapse, this is in much the same condition as it was in the nineteenth century. It features a series of four offsets or annuli, an entrance at the east and, opposite this, a second, much-ruined aperture. A dry stone wall extends from its south side to the cliff edge. The clochaun is slightly flattened at its junction with the wall terminal, suggesting that the wall is also an early feature,’respected’ by the clochaun. Apart from this stretch of walling there is nothing to indicate that the main group of monuments was enclosed. Immediately north of the clochaun are the 1ow footings of a small rectangular building of east/west orientation, with a gap in the eastwa11 that may represent a doorway. Its dry stone walls are relatively thin, probably incapable of supporting corbelling, so the roof was probably of some organic material such as thatch. A second, short stretch of remnant walling occurs about 8m south-west of the church. About 25m north of the clochaun is a circular grassy mound, about 5m in diameter and with a central hollow, which may be the remains of a second clochaun. About 16m north of the church a single uninscribed pillar stone stands on the verge of the eroding cliff edge. Further north again are two enigmatic groups of low, edge-set stones, forming linear settings, one aligned east/west, the other north/south. Some of these may be grave-markers. Finally, adjacent to the east wall of the church is a large recumbent slab (at least 1m x 1.55m), which, though uninscribed, might mark a special grave.
Inspection in the field leaves no doubt that this is a drystone church of Gallarus type. Its roof was already partly collapsed when Wakeman visited, and more has fallen since then, but its corbelled south wall and much of the east and west walls are extant, while the footings of the north wall are visible under at least two phases of modern rebuilding that probably relate to its reuse as a sheepfold. It is also clear that the trabeate doorway at the west end of the south wall was part of its original design and that it never had a west door. Two edge set slabs at the interior represent the footings of a drystone altar that abutted the east wall. With an area of 9m2 (3.66m x 2.46m) it is somewhat smaller than the average drystone church and was obviously designed to accommodate only the small resident community. The masonry in the interior is of well-fitted, roughly coursed slabs, and is closely comparable to that of the Kerry drystone churches. The exterior has been more extensively rebuilt but, in contrast to the interior, even the original sections are of poorly fitted, uncoursed blocks. Five projecting slabs serve as steps allowing access to a narrow offset or scarcement (0.35m wide) high on the outer face of the western end wall. At least one of these ‘steps’ is in a rebuilt section, and it may be that they were inserted in modern times so that shepherds could mount to roof level to signal to the mainland.
The east window of the church is highly unusual in being smaller on the inside than on the outside. Its stepped sill is characteristic of pre-Romanesque architecture, but here, in a reverse of normal practice, the step is on the exterior. Similarly, the jambs also have an outward splay. Double-splay windows (i.e. splaying both outwards and inwards from the middle of the wall) are common in late Saxon England, and there is one Irish example, in the drystone church at Kilmalkedar, but the Bishop’s Island window is unique, at least among the Irish pre-Romanesque group, in lacking any sort of inward splay or embrasure. This obviously limited the amount of light getting into the church, but it also meant that the thickness of the wall acted as a shield against the weather. Perhaps this was a priority in such an exposed location, especially given the lack of evidence for shutter fittings and the fact that the altar, with its sacred offerings, stood below the window against this same east wall.
Why is Bishop’s Island Different?
Among the most striking characteristics of Irish pre-Romanesque churches are their simplicity and homogeneity, and one of the fundamental tenets of this architecture is the positioning of the doorway in the west wall. This is so rigidly observed that one suspects it was introduced at an early date, most likely from sub-Roman Britain, and quickly became essential in identifying a building as a church. South doorways have been postulated for some pre-Romanesque churches, such as Temple Brecan on Aran and Killinny on mainland Galway, but invariably it turns out that their original west door has been largely robbed out, making it difficult to identify. The only single-cell mortared church of pre-Romanesque q?e with a south doorway that appears to be original is St Daithlean’s on Kerry Head, It is possible that it was moved there from the west wall, but that wall is ruined now, and apart from some resetting at lintel level there is no evidence for this in the fabric surrounding the door itself.
Both St Daithlean’s and Bishop’s Island are minor sites in peripheral areas and are unlikely to have been at the forefront of architectural or liturgical innovation. So why are they different? A likely explanation is that these are relatively late buildings, responding to changes in church layout, but doing so in an architecturally conservative manner. If so, then, as Peter Harbison pointed out, Bishop’s Island can be taken as evidence that Gallarus-t,?e construction was still current at the end of the early medieval period. Indeed, following this logic, we might easily assign both churches a post’1200 date, given that the principal doorways of Hiberno-Romanesque churches are still usually positioned in the west wall.
An alternative explanation should also be considered. Perhaps the marginal locations of the two churches in question allowed them to deviate ftom normal practice, at least in part for practical reasons. At St Daithlean’s the ground level drops away steeply immediately west of the church owing to erosion by a stream, and it may simply have been considered impractical to position a doorway, and thus the approach to the building on that side. While Bishop’s Island is flat, we have already noted that its exposed position resulted in an east window that is, to all intents and purposes, inside out. Similarly, perhaps the doorway was oddly positioned in the south wall to avoid the violent buffets of Atlantic winds via an entrance in a more orthodox position in the west wall. Certainly, protection from weather is the best explanation for the east doorway in the putative drystone pilgrimage church that was excavated by Gerry Walsh (1994,7) on the summit of Croagh Patrick, Co. Mayo, and for the north/south orientation of Temple Benan, a tiny mortared church on an exposed ridge on Aran, recorded by Con Manning (1985). In contrast, the off-centre east doorway in the drystone church on the remote island of Inishvickillane, Co. Kerry, may have more to do with its unusual position at the west end of the main terrace. Whatever the specific motivations in each case, it is interesting that these unusual features tend to occur in isolated churches in atypical settings.
Obviously, limited resources, learning and building expertise meant that local churches were not as carefully designed or as exactly positioned as those at major sites. This is borne out by a general survey (6 Carrag6in, forthcoming) that shows that local churches are often less carefully oriented than those at major church sites and are much less likely to have symbolically significant length-to-breadth ratios, such as 1:1.414 (√2), or 1:1.62 (the Golden Section). Nonetheless, the builders or patrons of most local churches, whether primarily monastic or pastoral, would have been anxious to emulate the major establishments, and so marked departures from the architectural norm are rare. However, the residents of truly eremitic sites like Bishop’s Island may not have been so concerned with impressing a lay or clerical audience. The written sources show that, while ascetics like these commanded great respect, there was unease about their independence from mainstream monastic authority and a suspicion that some of their practices were fundamentally unorthodox. Perhaps such individuals were more willing to dispense with convention, especially if it made eking out an existence in their ‘deserts in the ocean’ more tolerable.
None of this helps to establish whether the Bishop’s Island church is as early as the eighth-century example that was excavated on Illaunloughan. Indeed, although hard evidence is lacking, it does seem likely that drystone churches were still being built until the twelfth century and some may even have remained in use into the high medieval period. In his Gallarus paper Harbison did the subiect a great service by dismissing the idea that the vaulted churches at important sites like Glendalough developed from the corbelled drystone churches at minor sites in peninsular Kerry. The argument here reinforces this crucial point. We should be cautious about using churches at obscure sites to make sense of broader architectural patterns, not least because unconventional buildings are sometimes more common on the margins.
We owe a big thank-you to Brian Cullen of Killaloe for so generously agreeing to fly us to the island and for helping us to gather as much information as possible in the few hours available.
Harbison, P. 1970 How old is Gallarus Oratory? Medieval Archaeology 14,34-59.
Manning, C. 1985 Archaeological excavations at two church sites on Inishmore, Aran Islands. Joumal of the Royal Society of Antiquaies of Ireland 115,96-120.
O’Carragain, T. (forthcoming) Pre-Romanesque churches in lreland. Dublin, Four Courts Press.
O’Donovan, J. and O’Curry, E. 1997 The antiquities of Co. Clare. Ennis.
Walsh, G. 1994 Preliminary report on the archaeological excavations on the summit of Croagh Patrick. Cathair na Mart 14, 1-10.
White Marshall, J. and Walsh, C. 2005 lllaunloughan Island: an early medieyal monastery in County Kerry. Bray. Wordwell.