(In this description the bays of the nave are counted westwards from the site of the crossing. It should be noted that the nave was originally of eight bays though the easternmost bay has been almost completely rebuilt, along with the adjacent parish church of 1818-21. Nevertheless, the nave bays are numbered as if that eastern bay remains in place.)

Dunfermline, some three miles inland from the south coast of Fife, is on a southward facing slope overlooking the Firth of Forth. The ‘dun’ element of its name suggests there was an early fortification here; but, whatever its origins, there was certainly a royal residence by the later eleventh century. The abbey has its origins in a church built by St Margaret on the site of her marriage to Malcolm III c.1070, to which she brought a small group of monks from Canterbury, who formed the first Benedictine community in Scotland, and it was here that Margaret was buried. The house was re-founded as a major abbey c.1128 by David I, the sixth and youngest son of Malcolm and Margaret, with a fresh infusion of monks from Canterbury, and for over two centuries it was to be the most important mausoleum of Scotland’s royal dynasty. Until modern expansion, the town that grew up around its walls was largely to its north.

Fragmentary footings of a diminutive four-compartment church that had evidently been built in two phases (c.1070 by Margaret and c.1100 by Edgar?) were excavated below the floor of the nave in 1916. The present building, constructed of buff-coloured sandstone, was erected after the establishment of the abbey in 1128. The surviving part is seven bays of the originally eight-bay nave, which was retained in use as the parish church after the Reformation. The site of the eastern limb and transepts is now largely occupied by the church that was built there to the designs of William Burn in 1818-21 to replace the medieval nave as the place of parochial worship. The only part of the medieval church to remain in place east of the nineteenth-century building is a fragment of the feretory chapel of St Margaret, which is presumed to have been built c.1249, and of which the lower east and south walls partly survive.

No detailed record was made of the evidence for the plan of the eastern limb before the early nineteenth-century church was built, but it is thought that it was initially of four aisled bays, with an apse projecting beyond the central vessel. The church was probably the first in Scotland to have three towers, though none of them survives in Romanesque form. The central tower is altogether lost, while the north-west tower, capped by a splay-foot spire, was rebuilt by Abbot Richard de Bothwell (1444-68), and the south-west tower was rebuilt in 1811 to the designs of William Stark. An early nineteenth-century view by John Clerk of Eldin, and one made for General George Henry Hutton, both appear to show the already truncated south-west tower before its collapse and replacement as having a massive stair turret at its south-west corner.

There were initially three entrances into the nave. One is on the south side, where it opened off the cloister through what was the second bay from the east; it was therefore presumably not aligned with the east cloister walk as would be more usual. This doorway is the most finely decorated feature of the surviving part of the abbey church. It has survived well through being protected initially by the north cloister walk, and later by the burial vault of Sir Henry Wardlaw, until it was rediscovered in 1903 and exposed by the removal of one bay of the Wardlaw vault in 1905. At that time there was some renewal of masonry in the jambs, and perhaps some limited re-tooling of the inner caps; unfortunately, as a result of over a century of re-exposure, the doorway is now showing some signs of weathering. A second entrance from the cloister was cut in the seventh bay at a later date.

The entrance for the laity is on the north side, in the seventh bay, and, as part of the mid-fifteenth-century modifications carried out for Abbot Bothwell, a tierceron-vaulted porch was placed over that doorway. This doorway appears to have been of five orders, but the fifth order is now largely covered by the walls and vault of the added porch. It is set within a salient, the superstructure of which is weathered back a little below the aisle cornice, and which is now framed by the buttresses of the mid-fifteenth-century porch. The upper part of the salient, which is partly obscured by the roof of the porch is decorated with a blind arcade of seven unmoulded arches carried on en-delit shafts with cushion capitals. The details of the caps and bases of that arcade are now too eroded to describe.

The great processional entrance, which is the largest of the three, is at the centre of the west front. It is set within a salient that is weathered back below the window inserted at mid-height of the west front in the mid-fifteenth century.

Externally there are three tiers of windows along the south and north flanks, which light the aisles at the lowest level, the galleries over the aisles, and the clearstorey. All of the Romanesque aisle windows survive on the south side, apart from that in the rebuilt east bay; they were set above the cloister roof. On the north side only the aisle windows in the third and fourth bays from the east survive in their Romanesque form. They all have a plain inner order, which may have been pared back to increase the daylight opening. Framing that opening in the jambs are en-delit nook shafts (largely renewed) supporting cushion or scalloped capitals with a roll necking and plain abaci that have a quirk between the two planes. Those abaci extend outwards to form a string course that meet the flanks of the pilasters between the bays. In the arch the shafts support an order in which chevron wraps around a curved profile. Framing that order is an outer order decorated with a label moulding, around which there is a hoodmould with a chamfered soffit. The exceptions to this, which are both on the south side, are in the second bay, which is above the south-east doorway from the cloister, where the hoodmould appears to have been enriched with low-relief carving, and in the seventh bay, where there is no hoodmould.

The two aisles are now braced by arched buttresses along much of their length, those on the south side being of 1620, and those on the north side of 1625. The roofs over the galleries were rebuilt to steeper pitches, possibly at the same time that the buttresses were built. On the south side this evidently meant that the upper part of the outer gallery walls was cut down, truncating or possibly completely destroying the window at that level, while the clearstorey windows were blocked in their lower part. On the north side the gallery windows appear not to have been altered, and the new roof instead rose to mid-height of the clearstorey windows. These changes were reversed during restorations of 1845-55, when some of the details of the gallery and clearstorey windows may have been modified. Evidence for the sequence of changes is to be found both in early views and in masonry changes, since the later work is marked by the use of a fine-grained grey stone, rather than the buff-coloured stone of the medieval work.

The nave gallery windows that may preserve their Romanesque appearance are in the second to the seventh bays on the south side, and in the second and third bays on the north side. All those on the south side, however, were renewed in the 1840s, and are now blind; the renewal is identifiable from the use of grey stone. The Romanesque windows at this level have triangular heads composed of two stones set at an angle to each other, which are carried on en-delit nook shafts with cushion capitals. Within that frame is a pair of triangular-headed openings with a lozenge shaped opening in the spandrel, all those openings having chamfered arrises to the reveals. In the fourth, fifth and sixth bays on the north side these windows have been replaced by small single-light windows, probably of thirteenth-century date, though the ghosts of the Romanesque jambs are still detectable in parts. It may be speculated that the triangular window heads could be a pointer to the possibility that the gallery storey was originally finished externally by a series of lateral gables, as is known to have been the case at Durham Cathedral, though at Dunfermline there is no other evidence to support this possibility.

The nave clearstorey windows are all of the simplest kind, having semicircular arches, with the reveals of both jambs and arch being relieved by no more than a narrow rebate. They have no hoodmould.

Internally the division between the monastic choir and the nave is known to have occupied the easternmost bay of the nave. The pulpitum, which is only known from the evidence of early views, and whose date is unknown, was in the west crossing arch; its single round-arched opening is shown on a view by Henry Cave drawn before construction of the parish church of 1818-21. The lower courses of the choir screen, with its two doorways, were located in excavations by Peter Macgregor Chalmers in 1916 and remain in view; traces of the screens across the aisles are also visible within the masonry at the internal side entrances to the early nineteenth-century church.

Within the nave there is a striking contrast between the decorative treatment of the arcade storey and that of the gallery and clearstorey levels, the former being treated with considerable richness and the latter with extreme plainness. It is thus likely that, while the arcade storey of the nave was built as a continuation of the campaign on the eastern limb and transepts, the upper storeys were built following a change of design. However, views drawn for Francis Grose and General Hutton in the later eighteenth and earlier nineteenth centuries respectively indicate that the gallery level in the easternmost nave bay was treated more richly than the bays to its west, evidently with four sub-arches within each opening, suggesting that as part of the first campaign an abuttal for the central crossing was provided at gallery level. There may then have been a slight hiatus before construction of the upper storeys of the nave, when work was re-started in a very much more frugal manner. Nevertheless, it is assumed that work on the nave was largely complete by the time of a dedication recorded in 1150.

The nave arcade piers are generally of cylindrical form, the two eastern ones on each side being decorated with incised patterns. The south-west tower pier is composed of elements that reflect the form of the wall responds along the aisles, except on its north face, towards the central vessel, where it has a plain face with no engaged shafts or pilasters. Thus, on each of the east, south and west sides there is the combination of a half-shaft on the face of a pilaster, flanked on each side by a set-back half roll. On the north side the west arcade bay beneath the north-west tower was replaced by a solid wall terminating in an arcade respond as part of Abbot Bothwell’s work in the mid-fifteenth century, and the pier to its east was rebuilt to octofoil clustered shaft form.

Along the aisle walls the bays are articulated by major responds. The lower walls below the windows of the north and south aisle, and across the west front, were initially decorated throughout with blind arcading, except in those bays where there are doorways. However, the arcading has been removed in those parts that were rebuilt by Abbot Bothwell in the mid-fifteenth century. It has also been almost entirely lost below the south-west tower, which was rebuilt in the early nineteenth century, and in the seventh bay of the south nave aisle, where a later medieval doorway was inserted. In addition, it has been removed or partly obscured in the fifth bay from the east of the south aisle, and in the second and fourth bays of the north aisle, due to the insertion of post-medieval memorials.

In the south nave aisle the string course below the windows survives in the second to the seventh bays, though in the second bay there are only short stretches on each side of the doorway rear-arch in that bay. In the north nave aisle the string course below the windows has survived only in the third and fourth bays, though much of that in the third bay is restored. The rich detailing of the Romanesque aisle window rear-arches that rest on the string course is one of the clearest indicators of the high aspirations behind the first phase of construction at Dunfermline. They have survived in the second to the seventh bays of the south nave aisle, and in the third and fourth bays from the east of the north nave aisle.

The aisles are vaulted throughout, except within the early nineteenth-century south-west tower. The only area where the vaulting appears to be essentially in its Romanesque form, however, is in the second to fifth bays of the north nave aisle, where the webbing is of heavily plastered rubble. Here there are broad transverse arches with a flat soffit and sunk angle rolls. The diagonal ribs have a soffit roll flanked by segmental hollows. Diagonal ribs of this kind are also to be seen down the length of the south aisle, though the form of the transverse ribs, together with the ashlar construction of the webbing, makes clear that the vaults on that side have been reconstructed at some stage, retaining only the diagonal ribs. A combination of the facts that the diagonal ribs in the second to fifth north aisle bays inter-relate rather awkwardly with the moulded second arcade order, and that in the south arcade the second arcade order towards the aisle is of simple rectangular profile in the third to sixth bays, might suggest that it had been initially intended that the vaults should be groined and not ribbed. But there can be no certainty on this. Beneath the east side of the south-west tower, where there is also a surviving Romanesque transverse arch, there is a second order to that arch, with a pair of rolls to the soffit and a segmental hollow to the face.

The gallery openings, which rest on a string course, are simple round arches with plain arrises, and those that are still in their Romanesque form are relieved by no more than a single arch order at the centre of the wall thickness. It cannot be ruled out, however, that the original intention had been for each arch to contain a number of sub-arches, as appears to have been the case in the first bay on the north side on the evidence of early views. In the sixth to eighth bays on the north side the gallery arches were reconstructed for Abbot Bothwell in the mid-fifteenth century. In the sixth and seventh bays simple unmoulded round arches were built, while the eighth bay was left blank,

The clearstorey windows also rest on a string course, and have no more enrichment than a single nook shaft on each side in the jambs. There is a wall passage at clearstorey level, which appears to have continued across the west front; there was also a lower wall passage across the west front a little below the level of the gallery. The clearstorey of the sixth to the eighth bays on the north side were rebuilt as part of Abbot Bothwell’s reconstruction of the north-west corner of the nave, and are even more simply detailed than their Romanesque counterparts.

The central vessel of the nave is now covered by a flat panelled ceiling braced by transverse timber arches dating from the mid-nineteenth century. A view of 1805 by the Rev’d John Sime shows that when in use as the parish church a ceiling had been inserted above the level of the arcade arches, with two levels of timber lofts within the arcades and further lofts at each end. The upper level of lofts in the aisles was lit by a number of windows cut through the aisle walls, some of which involved paring back of the vault webbing, and although much of this was reversed in 1845-55, traces of some of those openings are still discernible.

There are extensive remains of the ranges on the east and south sides of the site of the cloister, and of the guest house/palace to the south-west, though none of these incorporate detectable Romanesque fabric and need not be discussed here.

Dunfermline Abbey was a foundation of particular significance for the history of Scotland, for its royal dynasty and for its medieval Church. It was the first foundation for a Benedictine community in Scotland; it was the shrine church of the founder of that community, St Margaret; and it was the most favoured mausoleum of the Scottish royal dynasty from St Margaret in 1093 to Robert I (the Bruce) in 1329.

Dunfermline was evidently an important royal residence by the eleventh century, and it was here that King Malcolm III married (St) Margaret c.1070. Turgot, Margaret’s biographer, says she built a church in the place of her marriage in honour of the Holy Trinity, to which Archbishop Lanfranc of Canterbury sent Goldwine and two other monks. This nucleus of a monastic community was dispersed in the troubles following Malcolm III’s death in 1093, but a further group was sent 1100x07 by Archbishop (St) Anselm at the request of King Edgar. Excavations beneath the east end of the nave in 1916 located the footings of an early church of four compartments that had evidently been constructed in two phases. The first phase appears to have consisted of a small square tower/nave (?), with a narrower but longer rectangular cell to its east; to the east of that was added a longer and wider cell that was interpreted as having had an apse on its east side. It seems likely that the earlier part represents the church built by Margaret and the later part an extension by Edgar.

The community was re-established by David I on a far more ambitiously endowed scale, and in 1128 Prior Geoffrey of Canterbury was sent as the first abbot of the enlarged community. Construction of a major new church was evidently started at the same time, and there was a dedication in 1150, by which time it is assumed that the main body was largely complete. A master Ailric the mason was granted a portion of Dunfermline’s lands c.1153, though we do not know if this was because he had been involved in the design and construction of the new abbey church. By the time of David I’s death in 1153 the abbey was the burial place of St Margaret, Malcolm III, Edgar and Alexander I, and it was here that David was himself entombed. Subsequent kings to be buried there were Malcolm IV, Alexander III and Robert I, along with other members of the royal family.

Some further construction was in progress in 1226, when Pope Honorius III referred to the making of noble structures, and a bull of 1231 mentions great expenses in building works. The thirteenth-century works on the church culminated in 1249, when there was a papal ruling that a fresh dedication was not required. Since St Margaret was translated to her new feretory chapel in the following year, it is likely that the construction of that chapel had been the main focus of the later phase of those activities. The work appears to have involved the squaring off of the east end of the eastern limb, perhaps to provide an ambulatory beyond the presbytery, with the construction of the low rectangular feretory chapel beyond; the lower east and south walls of that feretory partly survive.

Subsequent medieval works of which we have some knowledge include the construction of a large chapel on the north side of the east limb, which in that position is assumed to have been a Lady Chapel. It is thought that structural difficulties were the reason for major reconstruction at the west end of the nave, which can be attributed to Abbot Richard de Bothwell (1444-68) since his arms are located on the boss of the rebuilt vault in the sixth bay from the east of the north nave aisle, on the vault of the north porch and on a buttress tabernacle of that porch. Bothwell’s work entailed rebuilding much of the sixth, seventh and eighth bays of the north nave aisle and arcade wall, the reconstruction of the north-west tower, and the addition of a porch over the north nave doorway. It probably also extended to the reconstruction of much of the upper parts of the west front. All of this had a significant impact on the Romanesque work. Other later medieval works that have had an impact on the Romanesque work include the insertion of a doorway, in the seventh bay of the south nave aisle, the insertion of traceried windows in the second, fifth and sixth bays of the north nave aisle, and the reconstruction of much of the aisle vaulting.

A number of post-Reformation events or operations have had a further impact on the Romanesque fabric. A two-bay burial vault was built against the south side of the second and third bays by Queen Anne of Denmark, possibly as her own intended burial place, but in 1616 it was granted to Sir Henry Wardlaw. In 1620 and 1625 arched buttresses were built up against some of the bays of the south and north nave aisle respectively, and it was probably at the same time that the roofs over the nave galleries were rebuilt to a steeper pitch. In 1672 much of the eastern limb collapsed. In 1753 the central tower collapsed and concern was expressed about the condition of what remained of the south-west tower, which eventually collapsed in 1807. In 1810 the south-west tower was rebuilt to a simplified design by William Stark, though as early as 1809 it had been decided the parish church in the nave was inadequate, and the local presbytery ordered the heritors to provide a new one. In 1818-21 that new church was built over the site of the eastern limb to the designs of William Burn; he also made proposals for restoration of the nave, involving the enrichment of its upper storeys, which were not acted upon. In 1829 Robert Reid, Master of the Works in Scotland, made more conservative proposals for dealing with the nave on the assumption that it was crown property, though it was only from 1845 that the state began to assume active responsibility. In 1848 three of the south nave arcade piers were rebuilt, the outer walls of the south gallery and the roofs over the galleries were rebuilt to their assumed original height and pitch, and a flat ceiling was installed over the central vessel of the nave, all under the direction of William Nixon, Clerk of the Works in Scotland. In 1905 the Wardlaw burial vault was reduced by one bay to reveal the doorway in the second bay from the east of the south nave aisle, which had been rediscovered two years earlier. In 1916 Peter Macgregor Chalmers excavated below the east end of the nave to reveal the earlier church, which also incidentally exposed the lower courses of the choir screen, and in 1917 the outline plan of the early church was marked out on the nave floor at the expense of the Carnegie Trust.

In spite of the loss of its eastern limb and transepts, Dunfermline Abbey church is one of the most complete Romanesque buildings in Scotland. It dates from the earlier phases of the revitalisation of the Scottish church that took place during the reigns of the sons of Malcolm III and St Margaret, a revitalisation that was pursued especially energetically under David I (1124-53), who was the re-founder of the community here. As the youngest son of Malcolm and Margaret, there had been no expectation of his succeeding to the throne, and he had spent much of his earlier life in England, where his brother-in-law, Henry I, provided him with vast estates through marriage to the richest heiress then available, Matilda, daughter of the earl of Northampton and Huntingdon. He was a profoundly pious individual, and before his accession was able to build up a close understanding of the wave of renewal that was taking place across the European Church, and of the buildings being erected for that Church in England. Even before his accession he took active steps to modernise the church in south-western Scotland, the area over which his brother, Alexander I, allowed him to exercise authority. On succeeding to the throne, he founded large numbers of houses for the religious orders, and actively fostered the establishment of an effective network of dioceses and parishes. The consequent requirement for an unprecedented wave of new building across the lowland areas necessitated the importation of large numbers of masons from England, and the indications that many of these masons came from areas with which David had personal connections strongly suggest that he played a significant role in their selection.

In the design of Dunfermline Abbey there appears to be an awareness of a number of English buildings, from the abbeys of Selby and Southwell to that of Waltham, suggesting that its designer - and perhaps also its patron – was closely involved in the exchange of ideas taking place between all these centres of creativity. But the building with which Dunfermline shows the closest relationships is Durham, and here it should be remembered that Scottish monarchs since Malcolm III had been leading patrons of the see of Cuthbert; indeed, the Rites of Durham records that there were images of Edgar, Alexander I and David I on the pulpitum. It should also be remembered that Durham was one of the northern English counties which David considered to be part of his kingdom, and over which he was able to exercise effective control during the civil war between Stephen and Matilda.

The many borrowings from Durham are seen in features such as the label moulding around the nave arcade arches, a motif that was introduced in the course of building Durham’s nave, and that was soon afterward also to be copied at Kirkwall. Is it also a possibility that the triangular gallery windows of Dunfermline might point to the middle storey having been designed for lateral gables to each bay as are known to have existed in the nave at Durham? But the feature that shows the most specific debts to Durham is Dunfermline’s south-east nave doorway. In the detailing of its capitals there are clear parallels with the rear-arch capitals of the south-west nave doorway and with the chapter house corbels at Durham Cathedral. There are also parallels with a number of manuscripts produced in the Durham scriptorium, such as a twelfth-century copy of Jerome’s commentary on Isaiah (B.II.8).

The west doorway is of particular interest for its relationships with a number of buildings that appear to have followed its lead, and which therefore highlight Dunfermline’s role as a channel for the introduction and dissemination of new ideas. Of particular significance in this is the richly detailed fifth order, which has heads carved in mid-relief alternating with motifs that include interlace, foliage, a triquetra and a bird with outspread wings. Partial parallels for such forms may be found in the south nave door at Dalmeny Church and a destroyed door at Edinburgh St Giles that is now known only through drawings.

If much of the arcade level of Dunfermline appears to show the adaptation of ideas developed at Durham for the needs of a church of about half its size, and with no requirement for high vaulting, the internal treatment of the gallery and clearstorey levels of Dunfermline’s nave shows virtually no parallels with Durham. It seems likely that the stripped-down austerity of those upper levels must have been necessitated by financial difficulties, at a time when David was involved in the dauntingly costly construction of a number of other major buildings. At some of those other buildings, including Augustinian Jedburgh and Tironensian Kelso, it may be that a need for financial restraint was a factor in a pause in construction after the eastern limb and some of the lower parts of the nave had been completed. However, at Dunfermline, which was of outstanding importance for David because of its role as a royal shrine and mausoleum, it may be that it was felt better to complete the building to a less costly design than to leave it in a half-finished state. If the design of the upper storeys was the consequence of more than simply a rejection of costly detailing, is it possible that its patron was looking to the north-west rather than the north-east of England? In that area the nave of Carlisle Cathedral was being built in a rather stripped-down idiom that shows some similarities with Dunfermline in the rejection of vertical articulation and in the unmoulded arrises of the openings. The diocese of Carlisle had been established by Henry I in 1133, much to David’s annoyance since, like Durham, it was within the northern English counties that he considered to be part of Scotland. However, David was able to re-establish control over Cumbria for a period after Henry’s death, and it is known that he undertook major construction works on Carlisle Castle, where he was eventually to die in 1153. It would therefore not be surprising for masons to be drawn from Carlisle, and indeed, David might not have been entirely unhappy to be the creator of some difficulties at a cathedral whose foundation he had not supported.