82 The shrine-chapel of Columba on Iona with St John s Cross (replica) in the foreground.

Transcription

1 81 The monastic landscape of Iona. The later medieval abbey is probably on the site of the main early medieval church. The shrine-chapel is immediately west of the abbey church, but is not visible here. The Romanesque chapel of St Odrán is visible at the right. The substantial earthwork in the middle ground probably served as the monastic vallum. Shrine-chapels may therefore belong to the long line of churches in the West that emulate one or more elements of the Holy Sepulchre complex. 211 Though not possible to prove, it is tempting to suggest that, as in so many other areas, Iona was where the original innovation took place: that Columba s was the first shrine-chapel to be built, perhaps when he was first translated in the mid-eighth century, most probably during the abbacy of Cilléne Droichtech (726 52; fig. 81). 212 We know little about the intellectual interests of this abbot, but can well imagine another monk of Iona, Cú Chuimne, who died in 747 (Annals of Ulster), being involved in this project. Cú Chuimne was one of the two compilers of the Collectio Canonum Hibernensis, the text that most comprehensively articulates the idea that Irish sites should be modelled on Jerusalem (see above, p. ). 213 Based on textual evidence alone, one would conclude that if ever this idea was given monumental expression it was probably on Iona in the eighth century. As Jennifer O Reilly has shown, De locis sanctis in some respects prefigures Adomnán s other major work, the Vita Columbae, which also sketches a holy land, but at the northernmost edge of the Christian world. 214 More systematically than any other Irish saint s Life, it sets out to establish connections between the local topography and that of the Holy Land. For example, on the day before his death Columba identifies Iona with Jerusalem by foretelling its future greatness among all peoples, just as Old Testament prophecies foretold the greatness of Jerusalem; 215 and at the end Columba himself is equated with Christ at the moment of his death when he breathes out his spirit. 216 Manuscript illustrations also strongly support this idea. It is not by chance that our only two early Irish manuscript depictions of the earthly Jerusalem the illustrations in De locis sanctis and the Temptation of Christ in the Book of Kells were probably produced on Iona at around this time (figs 13, 36). It is interesting to contrast these with the depiction of the Heavenly Jerusalem in the Book of Armagh dating to 807. This schematic diagram faithfully follows the description in Revelation; it is a carefully compiled explanation of the text. 217 In contrast to many depictions of this subject elsewhere in Europe, for example in apse mosaics, no attempt is made to depict the 82 The shrine-chapel of Columba on Iona with St John s Cross (replica) in the foreground. Heavenly Jerusalem as a perfect version of the earthly city in the Holy Land. Nor is there any attempt to establish visual links with Irish ecclesiastical sites to make the point that these were also modelled on the historical Jerusalem and were prefigurations of the Heavenly Jerusalem. In contrast, the Iona images are indicative of a deep interest in establishing such links. We have already considered this in relation to the Holy Sepulchre plan in De locis sanctis (see p. ), and I have also suggested that the main model for the Temple in the Book of Kells scene was probably the principal congregational church on Iona (see p. ). What clearer expression could there be of the convergence of these two sacred topographies in the minds of the monks of Iona? The particularly close association of Iona with Jerusalem may have been acknowledged outside the Columban familia, for in the Book of Armagh, opposite the passage of St Mark s Gospel in which Christ prophesies the miseries at the destruction of Jerusalem, the scribe has written the name of the abbot of Iona, Cellach, who the previous year had to flee Iona because of Viking raids. 218 This association helps to explain the unusual use of the term templum to describe the church built at Kells as a safe haven for the Iona community and for the Book of Kells a few decades later (see further below). Nor was Columba s shrine-chapel the only monumental expression of this idea on Iona. It is also evident in the eighth-century high crosses at the site. Ian Fisher s suggestion that the first of these were commissioned to coincide with the translation of Columba has much to recommend it (fig. 82). They may well be the earliest stone high crosses in the Irish/west Scottish cultural zone. Some of them are set into stepped bases designed to recall another component of the Holy Sepulchre complex, the rock of Golgotha. 219 Furthermore, Dorothy Kelly argues that Odrán s Cross in particular was inspired by the True Cross as described in De locis sanctis. She suggests that the influence of this text on the Iona crosses contributed to the contrast in form between the Latin (i.e., shafted) high crosses in Ireland and western Scotland on the one hand and those in Anglo-Saxon areas, which comprise a tall stele surmounted by a Greek (i.e., equal-armed) cross. 220 The evidence outlined here raises the possibility that this text was also a factor in the contrast between the architectural settings of the cult of relics in these two areas. Fisher also suggested that the Book of Kells was commissioned to coincide with the translation of Columba s relics. 221 Neumann de Vegvar has shown that the columns and arches that frame its canon tables (and those of other Insular Gospel books) are probably representations of the colonnade of the Anastasis Rotunda that provided access to the Edicule, just as the Gospels provided access to Christ. 222 Just like the Edicule, Columba s shrine-chapel stands at the west end of the later medieval abbey complex, which presumably occupies the site of the main complex of early medieval churches. While its position was probably determined by a pre-extant grave (see above), it may be significant that, when Columban clerics came to build a successor to it at Kells, they also located this west of the main group of 78 79

2 142 (right) The doorway of the Scattery Island damliac from the interior. The projecting part of the lintel has two square holes for the stationary uprights of the wooden door frame, like those in the ex situ lintel at Kilfenora (cf. fig. 104). 143 (below) Brian Boru was involved in the establishment of the office of steward of St Patrick s in Munster. This individual was based at Ardpatrick, Co. Limerick, and it is possible that Brian had a role in commissioning the round tower and early church at this site. who commissioned them. Around 1005, however, Brian apparently developed a strong interest in a site just 18 kilometres south-west of Emly, namely Ardpatrick, Co. Limerick; and in this case his involvement in the construction of its mortared stone buildings seems more likely. This site is dramatically situated on the summit of a conical hill with commanding views over Limerick and into Tipperary. At its centre is a multi-period church and the stump of a round tower (fig. 143). Contrary to previous accounts, 202 the church does not have antae; in fact, apart perhaps from the north end of the west wall, none of its fabric seems to be early medieval. It does incorporate large blocks that probably come from a pre-romanesque church, and the former presence of such a building is confirmed by a previously unrecognised ex situ lintel with plain-relief architrave, which, given its lack of curvature, cannot belong to the round tower (fig. 144). Ardpatrick is not mentioned in the seventh-century Patrician documents; but the Tripartite Life indicates that, probably by the ninth or tenth century, Armagh had invented a tradition that it had been founded by St Patrick. It states that Patrick desired an ecclesiastical establishment (congbail) in Ard-Pátraic; and Patrick s flagstone (lec) is there and the boundary of his church site (torainn achilli). 203 This may be a reference to the impressive ecclesiastical enclosure that still delimits the 141 (facing page) The damliac and round tower of Scattery Island, Co. Clare, may have been built with the involvement of Brian Boru or his son Donnchad. The church was extended eastward in the twelfth century. 131

3 149 Killulta, Co. Limerick, is an example of a minor site within the Limerick cluster of churches. It did not become a parish centre in the high medieval period. 148 Temple Benan is a satellite church of St Enda s monastery on Aran, dedicated to a disciple of St Patrick. Its exposed position may explain its unconventional north south orientation. The great density of eleventh- and early twelfth-century churches on Aran may express the development of these islands as a place of pilgrimage. that is not in a geologically conducive area. This concentration of stone churches, however, is primarily due to competition between different ecclesiastical powers in this strategically important area (see p. ). The churches on the Aran Islands three of which, as we have seen, feature the only dedicatory inscriptions in the country (see p. ) are culturally distinct because, to a greater or lesser extent, these churches must all have been linked to St Enda s great monastery on the largest of the three islands. It seems possible that rebuilding them in stone was part of a coordinated effort to promote Aran as an important ecclesiastical centre and pilgrimage destination, as has been argued in the case of the circuit of outdoor altars, probably of eleventh-century date, on the perimeter of Inishmurray. 230 After all, this was the great age of pilgrimage throughout Europe (fig. 148). 231 However this may be, the other western church clusters are unlikely to be a reflex of important new developments in social or ecclesiastical organisation because, in contrast to the majority of the new local churches in England, 232 most of them are in politically as well as geographically and economically marginal areas. The clusters in Limerick, north-west Clare / south-west Galway and to some extent Aran are in subsidiary kingdoms of the Dál Cais, while the Lough Corrib / Lough Mask cluster is within the overkingdom of the Uí Briúin Aí kings of Connacht: it must be significant that none of them is in the home territories of these important groups. The construction of mortared stone churches was probably intended to bolster the position of these sites in the context of evolving ecclesiastical and secular power structures (see pp. and ), but most of their counterparts elsewhere in the country were accomplishing this with churches of wood. Initially, prestige may have been a factor in the choice of mortared stone, though it cannot have remained so when even minor sites within these clusters acquired a mortared church. Defence may also have been a factor, though probably less so at minor sites than at politically important ones (see p. ). If the wages of the wright as set out in the eleventh- and twelfth-century glosses on the law tracts are any indication, dairthechs and damliacs were considered to be on a par in terms of labour investment. 233 When the cost of materials is taken into account, however, dairthechs were probably usually cheaper to build. There is evidence, though, both from dendrochronology and from later material relating to the law tracts, that mature oak timbers were not as readily available in the tenth and eleventh centuries as they had been earlier. 234 Unfortunately, there are problems with using palynology to confirm this, because it cannot distinguish between young and mature trees. 235 One eleventh-century commentary on a law tract even refers to the excavation of bog oak by builders. 236 In some western areas the relative scarcity of oak may have tipped the economic balance in favour of mortared stone. 237 Even so, the situation was clearly not critical enough to prompt the development of a system of quarrying and transportation of stone to ecclesiastical sites in other areas, as was developing in England (see p. ). Indeed, the modest numbers of Romanesque and Transitional churches in Ireland, and their uneven distribution, suggest that woodlands were still commonly being exploited for the construction of churches during the high medieval period, though not usually at sites that get extensive coverage in the documentary sources (see p. ). 238 This section finishes with a more detailed look at the cluster of late antae-less churches in Limerick, to illustrate the fact that these churches were built in very different circumstances from their predecessors (fig. 135). 239 Some of these are within the Hiberno-Norse hinterland of Limerick, but others, such as Killulta, Clonshire and Kilfinny, are outside it (cf. p. ). 240 With the notable exception of Kilfinny, they are all built in a distinctive, regularly coursed masonry style that lacks the very careful fitting of blocks that one finds in some of the other clusters. They are found at a heterogeneous group of sites, and were apparently built to fulfil a variety of functions. This is illustrated by the fact that they vary considerably in size, though none of them comes close to the size of earlier churches at major sites such as Lorrha and Scattery. Furthermore, in the high medieval period the sites at which they occur varied considerably in function and importance. Some of them are at quite minor sites, and three did not become parish churches in the high medieval period: Kilrush (31 m 2 ), 241 the small subsidiary church at Mungret (26 m 2 ) and Killulta (16 m 2 ; fig. 149). 242 It is therefore unlikely that the principal purpose of this cluster of churches (or the other late clusters) was the creation of a network of parish churches (see further, pp. ), 243 but some of those at locally important sites, which were later converted into parish churches, were probably built to serve a pastoral role (see p. ). These include the possible (though much rebuilt) examples at Cloncagh (49 m 2 ), Kilfinny (>41 m 2 ) and Clonshire (38 m 2 ; fig. 150). For instance, in the later

4 189 The drystone church of Inishvickillane is unusual in that, for local topographical reasons, its door is in the east wall. This explains why the window is to one side and the altar comprises a simple projecting slab directly under it. This is the only example of this arrangement in Ireland, but there are parallels for it elsewhere. 187 (left) The masonry altar and aumbry at Kilgobnet, Aran. The small number of churches with stone altars also tend to be the ones with wall aumbries, hinting perhaps that in some other small churches there may have been wooden chest altars that doubled as aumbries. 188 (right) Like Kilgobnet, High Island has both a stone altar and a wall aumbry. Excavations showed that this church is late eleventh or twelfth century. The window and altar were conserved in the 2000s. Full-size masonry altars are found against the east walls of ten pre-romanesque churches, all of them small churches in the west of the country (fig. 187). 144 Early medieval masonry altars (as distinct from stone table altars and chest altars) are occasionally found abroad, 145 but we cannot rule out the possibility that a significant number of the Irish examples are high or post-medieval additions. Three of them are in drystone churches that have been cleared of rubble since the collapse of their roofs, and some of these may have been built using the material from this collapse (fig. 222). 146 Significantly, none of the excavated drystone churches has produced evidence for a stone altar, or for that matter of a wooden one; perhaps some of them had wooden altars that were not earthfast and therefore left no trace. Of the ten masonry altars in pre-romanesque churches, only the one on High Island has been excavated (fig. 188). It was found to replace an earlier stone altar, but firm date ranges were not attained for either one. 147 The excavations showed that the church was relatively late late eleventh or twelfth century and so this example cannot be taken as evidence that stone altars were common in the early medieval period. A similar masonry altar covered in plaster was excavated in the early Romanesque church of St Peter s, Waterford (see p. ). While a few of the masonry altars in the west of Ireland are covered with one or more slabs, none of these is sculpted; thus they need not be mensae of early medieval date. Nor, tellingly, do we have any early ex situ mensae from elsewhere in the country. The only probable example of an early medieval mensa is in a drystone church on Inishvickillane. This building is unusual in that, because of the steep rise in ground level immediately to the west, its doorway as well as its altar and window are in the east wall (fig. 189). Probably because of these constraints, its altar comprises a small mensa that projects from the east wall, south of the doorway and immediately under the window. 148 Elsewhere in Europe, some chapels within large churches were arranged in this way; for example, the late tenth-century first-floor chapel over the western porch at Deerhurst, Gloucestershire, was accessed via an off-centre doorway in the east wall, beside which was an altar and a small window that allowed the celebrant to see into the main body of the church. 149 A much larger slab ( m) in the relatively late (eleventh-century?) drystone church of Temple Geal is incised on one face with a u-shaped groove running parallel to each of the long sides, like those on the altar frontal from Flotta in Orkney (fig. 190). 150 On Inishmurray there is a rectangular slab with a five-cross motif, but it is too small ( m) to have served as an altar frontal on its own (fig. 191). This site is famous for its circuit of outdoor masonry altars or leachta, which have been rebuilt several times since they were first constructed in the tenth or eleventh centuries (see p. ); it is therefore possible that the slab originally formed part of one of these. Altar frontals, probably of early medieval date, are still incorporated into outdoor leachta at Tobernacroisseneeve and Mason Island, Galway. 151 As we shall see, there is much less equivocal evidence for indoor stone altars in the first half of the twelfth century, but even then they seem

5 age is illustrated by the contrast between Glendalough and the much more modest architecture of Mungret, another ancient establishment competing with a Hiberno-Norse port (Limerick), but one that Muirchertach did not favour (see p. ). A number of sources, some discussed below, point to the direct involvement of Muirchertach at Glendalough during this period, and one in particular hints at the involvement of an important individual in Muirchertach s circle in the construction of at least one of its churches: Máeltrena, a noble priest and learned superior of Cró Coemgen [St Kevin s House], the beloved fosterling of Ua Dúnáin, noble senior of Ireland, died, as became an ecclesiastic (Annals of the Four Masters, 1125). Bishop Máel Muire Ua Dúnáin gravitated to Muirchertach s circle soon after 1095, around the time that the king changed his policy towards Canterbury, Dublin and Glendalough (see p. ). Though we do not know for certain, Máel Muire may well have influenced this change. It has been argued that the name of the first bishop of Dublin, Donatus ( ), is a Latinised version of the family name Uí Dúnáin. 77 If so, he was a close relative of Máel Muire and would surely have known him well. As we have seen (see p. ), he seems to have founded Dublin from Cologne rather than Canterbury, and may therefore have had a different vision for the development of the port as an ecclesiastical centre than that which prevailed after his death. We can well imagine that Máel Muire did not approve of the alternative strategy pursued by Donatus successors, and that as a result he was not particularly well disposed towards the involvement of Canterbury in Irish ecclesiastical affairs. In 1096 he signed the letter seeking Canterbury approval for the appointment of the bishop of Waterford, but thereafter, when the ambitions of Dublin and Canterbury became clear, he played a pivotal role in the establishment of ecclesiastical structures independent of Canterbury. It is unlikely to be mere coincidence that Máel Muire is mentioned in one of only two annalistic references to Cró Coemgen, which is identified below as the double-vaulted church known as St Kevin s House (see p. ). It is tempting to conclude that Máel Muire played some part in commissioning this building, and that he installed Máeltrena, his beloved fosterling, as its first superior. If so, then this may have taken place between 1096 and 1100, by which time, as we shall see, Muirchertach and Máel Muire had probably commissioned the typologically later church of St Flannan s, Killaloe. Whether the other new buildings at Glendalough were constructed incrementally over the full period circa , or were mostly built immediately before and/or after Ráith Bressail, is a matter of speculation. It remains for us to consider what exactly Muirchertach and his allies were trying to achieve through this remarkable programme of building. Rebuilding the City of Angels In the very years when the new churches were being built, Glendalough experienced a revival in scholarship, especially during the abbacy of Tauthal Ua Cathail ( ), who is described in his obit of 1106 in the Annals of Inishfallen as a man of learning, fer léginn. 78 The curriculum was wide-ranging and up to date; it included, for example, works by Gerbert d Aurillac and probably Macrobius, as well as Chalcidius translation of Plato s Timaeus. 79 Significantly, Lebor na Cert, a work of propaganda commissioned by Muirchertach circa 1101, was transmitted via the Book of Glendalough. 80 As Candon has pointed out, this text fabricates Patrician authority for the idea that the king of Munster should be king of Ireland, and also emphasises that Armagh, rather than Canterbury or Dublin, is the rightful head of the Irish Church. 81 As we have seen, the late eleventh century was a time of unprecedented church building in Dublin. Muirchertach was probably directly involved in some of this during his decade as king of the port ( ). 82 By 1096 Dublin had far more churches than any other site in Ireland, many of them probably of mortared stone. 83 There is an acknowledgement of this in Lebor na Cert which includes the line: I bestow upon all Dublin supremacy [...] of reverence in its churches, supremacy for dwellings and sacred heights. 84 From around 1096 the policy of Muirchertach and his allies with regard to Glendalough was to set it up as a legitimate alternative to Dublin. It seems likely, then, that competitive emulation was one of the motivations behind their building programme. 85 A number of these churches, including St Kieran s, Temple-na-Skellig and Reefert/Dísert Coemgen, probably replaced earlier, more traditional, ones of wood. St Kevin s probably replaced one of stone, for it incorporates a door from an earlier church (see p. ). It is possible, however, that some of the others were built on virgin sites. One possible example is Trinity, which may have been a baptismal chapel (see p. ). St Mary s is first recorded as Lady Church in the Down Survey (fig. 243; also figs 39, 92, 227). 86 In its own enclosure some distance from the main complex, it recalls the position of nunneries at sites like Clonmacnoise and Lemanaghan (see p. ). In the Irish Life of Kevin, which is probably thirteenth 243 St Mary s, Glendalough, is probably the church of a nunnery that may have been established or rebuilt at around this time, possibly with the involvement of Muirchertach Ua Briain s mother, Derborgaill. century, the saint resurrects two women who had been beheaded while walking in the termon [i.e. church] land and made black [i.e., Benedictine] nuns of them in his own church. 87 In contrast, the earlier Latin Life seems to portray a civitas without a nunnery. There the episode takes place in the countryside away from Glendalough and afterwards the women go back to their families; the Latin Life also states that women and cows were far away from the monastery. 88 This text apparently took its present form in the late twelfth century; it therefore raises the possibility that a nunnery was not established at Glendalough until after circa On balance, however, it seems unlikely that St Mary s was originally built for some other purpose and converted into a nunnery a century later. Instead the version in the Latin Life may be pre-1100 in origin, for this text seems to incorporate earlier material (see p. ). There are hints in the annals that Muirchertach may have had a particular interest in establishing, or rebuilding, a nunnery at Glendalough. They record that his mother, Derborgaill, daughter of Tadg mac Gilla Pátric, King of Osraige, died there in 1098 (Annals of the Four Masters, Annals of Tigernach). Muirchertach s parents had divorced when he was young, and Derborgaill had then gone on to marry a succession of other kings. 89 One of her husbands was Augaire Ua Lorcáin, the king of North Leinster; this marriage may have been arranged to strengthen the links between the Uí Briain and the Uí Lorcáin. Augaire, who died in 1112 (Annals of the Four Masters), was in power when Muirchertach s strategic interest in Glendalough was at its height, and he must therefore have been an important partner in the rebuilding project. Derborgaill s death notices do not state that she was in monastic retirement when she