Analytical and historical assessment

Mingary Castle Ardnamurchan, Highland

Analytical and Historical Assessment for Ardnamurchan Estate

By Tom Addyman and Richard Oram December 2012


1        Introduction

1.1       General

2        Mingary Castle, the MacIans and the Lordship of Ardnamurchan        Richard Oram

2.1       Introduction

2.2       The Historiography of Mingary Castle

2.3       The Lordship of Ardnamurchan to c.1350

2.4       The MacIans

2.5       The MacIan Inheritance and the Rise of Campbell Power 1519-1612

2.6       Civil War to the Jacobite Era

2.7       Conclusion

3        Cartographic, early visual sources and visitors’ accounts

4        Earlier analyses of Mingary Castle

4.1       Introduction

4.2       MacGibbon and Ross

4.3       W Douglas Simpson (1938-54)

4.4       RCAHMS (1970-80)

4.5       More recent assessment

5        Description and structural analysis

5.1       Introduction – methodology

5.2       Geology, building materials and character of construction

5.3       The early castle (mid-late 13th century)

5.3.1         General

5.3.2         The early curtain

5.3.3         Entrance arrangements

5.3.4         Site of hall range

5.3.5         Other features of the interior

5.3.6         Possible well or cistern

5.3.7         The early wall heads and parapet walk

5.4       Later medieval – 16th century

5.4.1         Introduction

5.4.2         Hall range

5.4.3         Discussion

5.4.4         Garderobe tower

5.4.5         Wall head remodelling to N

5.4.6         Modifications to the wall head defences to the west, south and SE

5.4.7         Sea gate

5.4.8         West range (original build)

5.4.9         Outer works – causeway and flanking walls

5.5       17th century works

5.6       Possible siege damage – c.1644?

5.7       Later 17th – early 18th century works

5.7.1         Remodelling of the hall block

5.7.2         Remodelling of the west range

5.7.3         Causeway to NW

5.8       Later works (18th – early 19th century?)    i

5.8.1         Modifications to the north range

5.8.2         Further modification to the west range

5.8.3         East courtyard range

5.9       Recent repair works

5.10         Wall finishes

5.11         Notes on the immediate setting

6        Phased plans and elevations

7        Conclusion

8        Bibliography

8.1       Unpublished Primary Sources and Reports

8.2       Primary Sources

8.3       Secondary Works

9        Appendix A Historic Scotland – Listing description

10         Appendix B Historic Scotland – Scheduling description (29 March 1955)


The authors are grateful to Piers Dixon of the RCAHMS, and to Allan Rutherford and Aonghus Mackechnie of Historic Scotland, for their insightful comment upon the interpretation of the upstanding fabric.

Licenced material from the National Library of Scotland (NLS) has been removed from this report and replaced where possible with licence-free material. Original NLS maps can be viewed online, see

Licenced material from the Royal Commission on the Ancient and Historical Monuments of Scotland (RCAHMS) has been removed from this report and replaced where possible with licence-free material. To view original material online, visit

Revised material provided by Mingary Castle Trust to replace licenced imagery is annotated [Revised]. All other photographs and drawings in this report are by the authors unless otherwise stated.

Aerial view of the castle in its setting; Mingary House can be seen to upper left [revised]

General view from the ENE, the mouth of the Sound of Mull beyond. [revised]

Analytical and historical assessment

… on its dark cape reclined,

And listening to its own wild wind,

… where Mingarry, sternly placed,

O’erawes the woodland and the waste

Walter Scott, Lord of the Isles, 29

1       Introduction

1.1    General

Addyman Archaeology were commissioned by Ardnamurchan Estate, through Wighton Jagger Shaw Architects of Harrogate (contact Francis Shaw), to undertake an assessment of Mingary Castle, by Kilchoan, in Ardnamurchan, Highland.  The assessment was requested by Historic Scotland (contact, Allan Rutherford, Heritage Management Team Leader, Ancient Monuments North) in relation to the on-going development of architectural proposals to repair the castle and bring the structures within back into habitable use. 

This document incorporates a number of individual studies: 

Section 2, written by Professor Richard Oram, Professor of Environmental and Medieval History, University of Stirling, includes a review of historiography relating to the site; this is followed by a comprehensive study and reappraisal of historical source material, and places Mingary within wider historical, geo-political and chronological contexts.

Sections 3 and 4 respectively present an assessment of historic visual and cartographic sources that relate to Mingary, and an examination some of the analytical findings made by previous commentators who have studied the site.

Section 5, prepared following detailed on-site assessment of the upstanding fabric of the monument and some new survey work, presents an analytical account of the castle’s visible fabric.  This proposes a reassessment of the understanding of the structure and its evolutionary history, and describes the physical evidence in detail within a broadly chronological structure.  A number of new findings are discussed and reinterpretation of the site proposed.

Section 6 includes a newly revised sequence of phase plans and elevation drawings of the castle, these based in part on the existing survey and analysis of the Royal Commission for the Ancient and Historic Monuments of Scotland (RCAHMS) and partly on more recent survey work carried out in relation to the present project.

2       Mingary Castle, the MacIans and the Lordship of Ardnamurchan         Richard Oram

2.1    Introduction

Until its first emergence into the historical record in 1495, Mingary is a silent presence in the background of the long and complex history of the lordship of Ardnamurchan and, especially, its fourteenth- to early seventeenth-century MacIan owners.  Whilst it is likely to have been the venue for many of the great gatherings at which the political life of the Lordship of the Isles or the conduct of the campaigns which suppressed the rebellions mounted by those who sought to restore that Lordship after 1493 was decided, it is rarely mentioned directly in any source. 

In the following analysis, the castle remains an often shadowy presence but it should be remembered that the high political events which punctuate the narrative were mainly fought out over inheritances, land, office, and the status and influence that these things bestowed, and that Mingary constituted a physical symbol of all those things.  Although often not named specifically, it should be understood that the castle was one of the great prizes competed for, possession of it giving control of the lordship of Ardnamurchan – often conjointly with Sunart – and domination of the waters at the northern end of the Sound of Mull and between Mull, Coll, Tiree and the islands south of Skye.

Figure 2.1  A brooding presence, Mingary dominates the mouth of the Sound of Mull  [revised]

The report falls into five principal parts with a short conclusion.  Part 1 explores the historiography of the castle itself, examining the development of thinking on its date and possible builders.  Part 2 examines the evidence for the lordship of Ardnamurchan and its early owners down to c.1350.  This is followed in Part 3 by an extended exploration of the MacIan lordship down to 1519.  In Part 4, the transfer of control from MacIan to Campbell hands over the course of the sixteenth century is traced, concluding with the securing of actual possession by the Campbells in 1612.  Finally, the seventeenth- and eighteenth-century history of the building and its owners is considered, concluding with the final phase of known occupation in the early 1800s.

2.2    The Historiography of Mingary Castle

Given the remarkable level of physical preservation at Mingary, it is striking how in comparison to the other West Coast castles with which it is often linked in discussions it has been so little researched or analysed.  The summary notes with supporting bibliography on the RCAHMS Canmore database gives an indication of how relatively neglected Mingary has been in castellological or historical research in the nineteenth and twentieth centuries.[1] 

It is also evident that much of the ‘discussion’ has been simply a recycling of the interpretation of the building first offered in the late nineteenth century and that this, in turn, has fed into a remarkably circular set of arguments surrounding the whole question of date and development of a group of buildings on the western seaboard dominated by the castles of Dunstaffnage, Duart, Mingary, Tioram and Kisimul.  This is not the place to detail the development of those arguments, much of which has recently been critiqued by Geoffrey Stell in respect of Tioram, and the following section represents principally a review of the main items in the historiography of Mingary.

It was only in the late nineteenth century that Mingary first achieves a profile in architectural studies.  The relative remoteness of Mingary and its lack of a high historical profile may account for the fact that, despite the early date that they acknowledge for it, it was only in Volume III of their great five-volume analysis of Scottish castles that David MacGibbon and Thomas Ross addressed it.[2]  The first two volumes had seen the articulation of their theory of a progression in Scottish castle-development across a series of ‘Periods’, a typologising approach that was characteristic of nineteenth-century museology. 

In Vol III, they were bringing in further examples of the buildings of the different periods that had been omitted for a variety of reasons in the first two volumes.  For MacGibbon and Ross, Mingary was a castle of their ‘First Period’, which they placed broadly in the thirteenth century but with some possibly late twelfth-century stone structures in its initial phases.  They noted that the greatest level of survival – and completeness – of buildings of this period, labelled by them as ‘primitive structures’ was in the remoter Highlands and Islands, but they cautioned that this was not a true reflection of their probable former distribution across the country.  The principal cause of the loss of most castles of that period elsewhere in the mainland Scotland, they proposed, was the Wars of Independence.[3] 

In its language, MacGibbon and Ross’s account set in place a vision of primitiveness and general lack of sophistication that has characterised discussion of this group of West Coast castles down to present.  Mingary and its fellows were not only ‘ancient strengths’ but also ‘primitive fortresses’.[4]  Having given that value judgement, they then proceeded to explain that there is no record of the origins of the building and give its first historical attestation as the 1490s.  Nevertheless, they apply a broad thirteenth century construction date to it.  Their principal diagnostic dating evidence is the form of the windows in the north wall of the enclosure, which they identify as sharing characteristics with openings in the walls at Dunstaffnage and Duart.[5] 

In terms of analysis, they focus principally on the curtain wall and its character but identify that the principal buildings of the original complex are likely to have stood against the north side of the enclosure as the current main structure does.  That building they saw as essentially eighteenth-century in date.  The east and west ranges are not assigned dates by them but are interpreted as, on the west apartments with cellars below and on the east as ‘stables or byres’.  They conclude with the assessment that ‘we have here undoubtedly one of the primitive castles of the Western Islands’, an assessment based on simplicity of plan and the form of the window and door openings in the curtain, which indicated ‘a date as old as the thirteenth century’.[6]  It is an assessment that remains unquestioned in its basic details down to the present.

Although significant space is devoted to West Highland and Hebridean castles, especially Dunstaffnage and Kisimul, Mingary is not mentioned in W Mackay Mackenzie’s pioneering 1927 study of Scottish medieval castles, but neither too is Tioram.[7]  Discussion of the western castles, therefore, did not progress beyond the assessment offered in the later 1880s by MacGibbon and Ross until William Douglas Simpson began to present his own nuanced view of medieval castle-building in Scotland.  He did not differ greatly in terms of architectural discussion from the views of his predecessors, his principal divergence being in terms of the social and political circumstances that gave rise to castle-building of particular forms. 

Simpson’s view of Mingary and Tioram as part of a scattered group of West Coast and Hebridean castles of enceinte of thirteenth-century date was tempered by a view of them as of curtain-wall castles of the ‘simplest form’ and quite ‘primitive’.[8]  It was his view that their lack of donjon or mural towers was a direct consequence of the rocky nature of their sites, the result being plain multangular curtain walls of the most basic design.  In this assessment, he failed to give due notice to the remarkable double level of lancets at Mingary, which point to an internal arrangement of some sophistication.  He was still more cautious with assigning a construction date in a later assessment of Mingary, placing its erection vaguely sometime in the period after the division of the Isles between Somerled and the king of Man in 1156, in which Ardnamurchan Point was the marker of the territorial division.[9] 

For Simpson, the closest analogy was Kisimul on Barra, to which he assigned a twelfth- or thirteenth-century date based on the style of its battlements.[10]  In the absence of any other more closely datable architectural detail, the apparently early form of plan – curtain wall castles by definition being considered ‘early’ in a typological progression from motte-and-bailey sites through to complex towerhouses – coupled with the seemingly simple, and therefore also by definition ‘early’, form of the wall-head, the general trend was to assign earlier rather than later dates of construction to these buildings.  Unlike Kisimul, however, Mingary has more closely datable features in the openings in its north front.  The lancet windows in the curtain, however, were assigned vaguely to the ‘First Pointed’ period – which for Simpson extended from the close of the twelfth-century through to the early fourteenth century – but the building itself was generally characterised simply as thirteenth-century.

Figure 2.2  View of the castle from the northern approach  [revised]

Simpson’s model for the emergence of castle-building in Scotland identified the castles of the Argyll coast and Hebrides as symbols of ‘the arrival of the Anglo-Norman penetration upon our western coastline’.[11]  They were the physical evidence for the intrusion of an alien culture into this Norse-Gaelic zone and were, ultimately, the concrete reflections of the new political realities that followed the cession of sovereignty in the Isles to the Scottish crown in 1266.  Maintaining the vision offered by MacGibbon and Ross nearly a century earlier, their remarkable state of survival he saw as a consequence of their remoteness from the political heartlands of Scotland, where most early castles had been destroyed in the course of the Wars of Independence. 

Indeed, Simpson highlighted the different treatment meted out to Dunstaffnage, for example, following its capture by Robert Bruce; it was garrisoned and maintained whereas other castles captured by the king in south-eastern Scotland were destroyed.[12]  The main reason for their survival, however, Simpson saw as arising from their location in ‘a remote and backward part of the country’, where there was less likelihood that they would be recast to accommodate changing architectural fashions or defensive innovations, which were in any case made impractical by the inconvenience of their actual sites.  It was his final considered view that ‘once built, such castles [as Kisimul etc] could never be anything else than essentially a great wall of enceinte’.[13]

Although Simpson treated the west coast castles as a group in his later work, he was influenced strongly by Stewart Cruden’s architectural analysis of Scottish castles, first published in 1960, where the fully-developed thesis of a distinct group of ‘Western Seaboard’ castles of similar plan and design had been articulated.[14]  The group, represented by Kisimul on Barra, Mingary, Tioram in Moidart and Dunstaffnage near Oban, were typologised on the basis of plan:

They are all of irregular plan, and are situated, each and every one, upon abrupt rocky outcrops or on the edge of sea rock.  They follow the outline of their site, and, with one exception, they have no projecting towers.  Their siting, general characteristics, and some particular features, strongly suggest contemporaneity and the work of the same school of military architects, a supposition which is strengthened by their distribution.[15]

Cruden proposed two possible dates for the development of the group; the first in the years following Alexander III’s success over the Norwegians at the battle of Largs in 1263, the second at the end of the thirteenth century during the initial phases of the Wars of Independence.[16]  For Cruden, the unquestionably thirteenth-century date for the lancets at Dunstaffnage and Mingary establishes the date of the entire group, but he refutes any attempt to date them more closely due to a general lack of either documentary record or more closely datable features.  He also dismisses any suggestions that they may be instances of later use of a much older style, an argument that he viewed as spuriously premised on the basis of the first recording of most of them at the close of the fifteenth century.

Considering the high level of preservation of the castle, as praised by Cruden, it featured little in further discussions of castle architecture until the 1980s.  The most detailed analysis was published in 1980 by RCAHMS, in which the synopsis of the castle’s description confidently ascribed ‘the main body of the curtain-wall … to the 13th century’.[17]  The following year John Dunbar’s account of architecture in the medieval Highlands repeated that dating, noting that it was the only castle in the West Highlands to possess any features that were in the leastways datable; ‘a few lancet windows of 13th-century character’.[18] 

On the basis of the similarities in plan of the earliest phases of Mingary and nearby Tioram, Dunbar went on to suggest that the two buildings were contemporary.  This tendency to link the buildings by analogous architectural detail was already evident twenty years earlier in Cruden’s work, where the nature of the stair providing access to the parapet of the curtain wall at Mingary and Tioram was noted as rising in ‘identical fashion’ in a straight sweep from courtyard level.[19]

Despite his repetition of Dunbar’s suggestion that Mingary might have been built by a MacDonald of Islay, Tabraham in his discussion of the West Coast curtain-wall castles pointed to the general similarities in plan shared by the main examples – Dunstaffnage, Mingary, Tioram – and some smaller structures, such as Duart, Duntrune and Dunvegan, and floated the question of if that similarity might point to ‘a single hand […] involved in their design and execution’.[20]  Having raised that issue, his discussion then focuses on Dunstaffnage, the one member of the mainland group that has both datable architectural details and has also been the target of modern excavation.[21]  The lancets in the curtain walls and fish-tailed arrow-slits in the towers at Dunstaffnage are diagnostically thirteenth-century in date and the castle had in the past been dated in its entirety on the basis of those features. 

The excavations, however, raised the possibility that the towers containing the diagnostic arrow-slits were secondary additions to the curtain wall and that the lancets, too, might be secondary insertions.[22]  Dates and builders for these two phases were also postulated: Duncan MacDougall of Lorne in the 1230s for the initial construction of the curtain wall; his son Ewen MacDougall in the 1260s for the towers and lancets.[23] 

This identification raises a further question; is the commonality between the castles in the group a marker of probable MacDougall ownership of the two?  Given that Ewen’s son, Alexander MacDougall, was identified in 1293 as lord of Ardnamurchan it is most likely that Mingary’s similarity to Dunstaffnage arises from the MacDougalls’ possession of a territorial sphere that extended from Lorne north and west through Morvern, Mull and Ardnamurchan.  Confident identification of the builders of the group of thirteenth-century castles along the west coast as the MacDougalls followed, with John Gifford having embedded that view in his Buildings of Scotland discussion of Highlands and Islands castles.[24] 

In his recent assessment of the castle-building activities of the three principal lineages descended from Somerled, the twelfth-century ruler of Argyll and much of the Isles, Ian Fisher offered a further complicating suggestion; was Mingary a MacRuari-built castle.[25]  This proposal is based straightforwardly on the similarity in basic design of Mingary and Tioram, the latter of which is historically attested by charter evidence from the 1320s as a residence of the MacRuari segment of the Somerled-derived lineages.[26] 

It is, of course, entirely possible that Mingary and the lordship of Ardnamurchan may have been acquired by the MacRuari family as a consequence of the dismemberment of the MacDougall lordship in the aftermath of Robert I’s triumph over the Balliol party in the Scottish civil war, but there is no evidence to suggest that the MacRuaris’ mainland lordship included Ardnamurchan as well as Moidart.  The dating of Mingary offered by Fisher remains generally ‘thirteenth-century’, with the familiar cross-reference to the architectural details of Dunstaffnage – particularly in respect of the window openings and entrances to the mural passages and latrine chambers – but the stylistic link with Tioram and its pre-1320s existence is offered as further corroboration of Mingary’s early date.

The most recent architectural discussions of the building effectively draw together all of the foregoing points in summary form.  Graham Ritchie and Mary Harman’s 1985 description is a very brief summary of the main findings of earlier more detailed analyses and avoids any attempt to attribute the building to a particular family; it simply labels the castle as thirteenth-century.[27]  The most extended discussion is offered by John Gifford, but this is basically a summary of the RCAHMS analysis of the 1970s.[28] Mary Miers cautiously gives a ‘from the thirteenth century’ date for the origin and development of the building but offers no rationale for why it is thirteenth-century beyond a sideways reference to its similarity to Tioram.[29]

2.3    The Lordship of Ardnamurchan to c.1350

Research into the medieval landholding and political structures of the West Highlands and Islands which might reveal the identity of Mingary’s owners is bedevilled by an absence of detailed record evidence for the period before c.1500.  Neither the surviving narrative nor the record sources provide a basis from which to reconstruct the political map of the region with any confidence, particularly on account of the tendency of some of the later regional clan histories to offer exaggerated views of their ancestral position as shameless propaganda or calculated claim-staking. 

An absence of hard evidence for the spheres of power of even the major families in the twelfth- to fifteenth-century period always leaves room for question in any models constructed, but there are enough fragments of evidence to permit the generalities of the pattern and the shifts in balance that occurred within it to be outlined.

Much of the difficulty revolves around the grand claims advanced in later clan histories for the status of the MacDonald lineage and its territorial sphere before their rise to power as Lords of the Isles in the fourteenth century.[30]  These contain fanciful and often chronologically impossible accounts of the pre-1300 period, typified by the suggestion that Ranald son of Somerled, whose floruit was in the late twelfth century, married ‘MacRandel’s daughter, or, as some say, to a sister of Thomas Randel (Randolph) Earl of Murray’, who died in 1332![31] 

The most serious problem, however, is the attempt in these later histories to assert MacDonald rights in the Hebrides through claims that they were ‘willed’ to them by Somerled rather than to Dugall, the eponymous founder of the MacDougall line, and that Dugall was, in case, illegitimate and had no right to an inheritance in the Isles.  Dugall’s heritage is presented as entirely mainland Argyll-based, stretching from Loch Leven in the north to Asknish in the south and his spread into Mull from there is presented as a shameless land-grab at the expense of his legitimate half-brothers and their children.[32] 

The Isles, it is asserted, were the rightful sphere of Ranald, and it was there that he attempted to find a suitable heritage for the sons of another of his brothers, Somerled.  The elder of these boys was John, whom the author of the ‘History of the MacDonalds’ identified as the founder of the lineage of the MacIans of Ardnamurchan.  Through this route, the ‘History’ seeks to establish a MacDonald interest in Ardnamurchan from at least the time of Ranald son of Somerled.  The claim is advanced further through a story that John, having achieved his majority, asked his uncle Ranald to provide him with men to seize a certain ‘Muchdanach’ who held Moidart and Ardnamurchan and was a close associate of Dugall.  John then killed rather than just captured Muchdanach, which greatly displeased Ranald, but seems to have mollified his uncle by stating that he wanted no more of his father’s lands than Ardnamurchan, Glassary in Argyll, and some property in Islay.[33]  This convoluted tale has been concocted simply to push the chronology of MacDonald/MacIan right to Ardnamurchan back by over a century.  Despite its clearly fictitious content, however, it has remained a source used by many modern popular historians, who have helped to give its mythologizing an air of credibility and acceptance.

This position has been compounded by the hesitancy over the question of the early ownership of Ardnamurchan and the identity of the likely builder of Mingary that has on occasion arisen in modern discussion.  Doubts and uncertainty expressed by modern academic or professional researchers have given rise to a series of false provenances for the construction of the castle finding their way into the academic literature.  The principal problem here is the separation of discussion of the castles from discussion of the historical record and the often greater public prominence that the former has received. 

On the basis of later MacIan-MacDonald ownership and the deep-seated strength of the ‘History of the MacDonalds’, for example, the eminent architectural historian John Dunbar suggested that the castle may have been built by one of the MacDonalds of Islay, a suggestion repeated in more recent popular literature on Scottish castles.[34]  This attribution is certainly wrong, for when Ardnamurchan appears in February 1293 as a distinct political territorial unit under the lordship of Alexander MacDougall, lord of Argyll, incorporated into the new sheriffdom of Lorne over which Alexander was to preside as sheriff,[35] it is evident that this had been MacDougall property for some considerable period. 

This latter fact is not ‘new’ evidence that came to light after Dunbar wrote; the arrangements for the new administrative structures in the west had been available in published format since the early 1800s.[36]  The difficulty lies in the academic separation of study into discrete specialist areas where lack of familiarity with the sources and the research of different fields often leads to critical documents or published analyses being omitted by specialists from different disciplines.  Study of the lordship of Ardnamurchan and Mingary Castle has long been bedevilled by that problem.

It was to separate the myth from the historically-attestable reality that in 1992 David Sellar began research that would lead to his new vision of the nature of the power structures of Argyll and the Isles in the thirteenth and early fourteenth centuries, published in 2000.[37]  Sellar started from the premise that the significance of the MacDougalls has been consistently understated in Scottish historiography through a combination of the political collapse of the family’s position in the fourteenth century and an over-emphasis on the importance of the MacDonalds.  His forensic dismantling of the surviving evidence provides the clearest discussion to date of the relationships between Somerled’s sons and reveals the certainty that it was Dugald and not Ranald who was the senior by birth.  As he rightly points out, however, ‘to lay much store by primogeniture at this period is, however, anachronistic’.[38] 

Dugald may have been the senior but his relative obscurity in the record suggests that he may well have been overshadowed by his younger brother.  Dugald’s children, however, appear to have overtaken their MacDonald and MacRuari cousins by the middle of thirteenth century, perhaps through a careful alignment of their interests with those of the Scottish crown which, under King Alexander II, was consolidating its presence in the western mainland seaboard.[39]  As Lords of Argyll, they were by the 1240s the dominant kindred in the region and already had begun a policy of intermarriage with the leading lineages of eastern and lowland Scotland. 

This regional dominance reached its apogee under Alexander MacDougall of Lorne, lord of Argyll, the man whom King John Balliol sought to establish as sheriff of Argyll in 1293.  Lord of a great complex of lands that stretched from Ardnamurchan in the north to the border of Knapdale in the south, he had the greatest territorial authority of any of his kin since the time of Somerled, an authority reinforced by his slaying of his two principal regional rivals – Alexander MacDonald of Islay and Colin Mor Campbell, lord of Lochawe – before 1299.[40]  Just over a decade later, however, all of that great assemblage of power had disintegrated and Alexander died as an impoverished exile receiving a pension from Edward II of England, his family broken through its staunch adherence to the Balliol interest in Scotland.[41] 

Although his great-grandson John Gallda MacDougall made peace with the Bruce regime in Scotland and received back a portion of his heritage in Lorne,[42] much of the great domain controlled by Alexander in the late thirteenth century had been lost forever, mostly into the hands of their kinsmen and ancient rivals the MacDonalds.

The process of formally separating the MacDougalls from their heritage and assigning it to others was probably underway well before 1314 but would have been confirmed in the general process of forfeiture of King Robert Bruce’s unreconciled enemies that commenced after Bannockburn.  A now lost charter of King Robert to Angus Og MacDonald, lord of Islay, probably datable to the period before c.1320, awarded him the lordship of Ardnamurchan, to which around the same time were added Lochaber, Duror and Glencoe.[43]

 The Ardnamurchan lordship, however, was missing from the charter of King David II issued to Angus’s son, John, in 1343.  That award confirmed him in possession of Islay, Jura, Colonsay, Mull, Tiree, Coll, Lewis, Morvern, Lochaber, Duror and Glencoe, plus the two castles of Cairn-na-burgh and the castle of Dun Chonuill.[44]  Such coming and going of these various properties in the award and counter-award of charters in this period helped to frame suggestions that Ardnamurchan may have fallen within the sphere of MacRuari power either in the thirteenth century or before the absorption of the family’s domain into the MacDonald lordship after 1346.  These arguments, however are not supported by any documentary record.[45]  Duncan and Brown, for example, argued that the failure of a list of titles afforded Alexander MacDougall’s son, John, in 1302 to mention either Ardnamurchan or Morvern, and the absence of these territories from a settlement of 1354 involving John’s grandson, John, pointed to these not having formed any part of the MacDougall heritage. 

Given that the 1302 document fails to mention other districts that were most definitely MacDougall-held at that time – such as Mull – this argument fails to stand up to scrutiny.  Their association of Ardnamurchan and Morvern with the MacRuari for no stronger reason than geographical proximity is untenable and, it must be added, would fall on the simple application of their own methodology for excluding them from the MacDougall lordship: when Ranald MacRuari formally entered the peace of King David II and had his lands confirmed to him in 1343, they were listed as Uist, Barra, Eigg, Rum and Garmoran (a mainland territory comprising Moidart, Morar, Arisaig and Knoydart) – neither Ardnamurchan nor Morvern were among them.[46]

The uncertainty over ownership of Ardnamurchan across this extended period down to the 1350s is probably the result of the complex efforts by the rivals for control of Scotland at this time – David II Bruce, Edward Balliol and Edward III of England – to win the support of the various kindreds who claimed the ability to control the region and deliver its military resources to their cause. 

Ostensibly ‘Bruce’ loyalists like John MacDonald of Islay, 1st Lord of the Isles, were not averse to cutting deals with the Balliol-English party and receiving speculative grants of land and offices which were then countered by alternative – and equally speculative – grants from David II and his supporters.[47]  Into this mix should be added John Gallda MacDougall, who was seeking to regain something of his heritage and who was prepared to sacrifice much to regain a toehold in his ancestral lands. 

It is perhaps significant that it was in the period shortly after his reconciliation with David II and around the time of his formal agreement with John MacDonald of Islay over their rival claims to the island portion of the MacDougall heritage[48] that the first hints of possible MacIan possession of Ardnamurchan come.  Was that part of the price that he was prepared to concede for a restoration to his great-grandfather’s lordship of Lorne?

2.4    The MacIans

The MacIans of Ardnamurchan were one of the senior cadets of Clan Donald.  Descended from John Sprangach (the Bold) MacDonald, a son of Angus Mor MacDonald, lord of Islay, who was active in the early fourteenth century, he appears to have been a beneficiary of his family’s adherence to the Bruce party in the Scottish civil war after 1306 and acquired significant properties from the MacDonalds’ Balliol-supporting MacDougall kinsmen.[49] 

His son, Angus, appears to have briefly achieved a leading position amongst the MacDonald kin during the renewed Scottish civil war between the supporters of the Balliol and Bruce families in the late 1330s/early 1340s, receiving in c.1341 a charter from King David II which nominally awarded to him all of Islay, Kintyre, Gigha, Jura, Colonsay, Morvern and two ouncelands in Mull.[50]  The grant was never effective, but if it had been it would have placed the MacIans in possession of what came to form the territorial heartland of the MacDonalds as Lords of the Isles.[51]  The first member of the family who has made a significant impact in surviving records is John, grandson of John Sprangach, who witnessed a charter of his kinsman, Donald lord of the Isles, issued at Rosemarkie on 16 August 1420.[52] 

He appears again in 1433, when along with his kinsmen Alexander MacDonald of Islay, lord of the Isles, and Donald Balloch MacDonald, and the island magnates Lachlan Maclean of Moy, Roderick Macleod of Lewis and John Macleod of Dunvegan, he supplicated the pope for permission to have a portable altar.[53]  This group reveals the tightness of a social and political network within the Lordship of the Isles.  He may still have been active in 1463, when a ‘John son of Alexander of the Isles’ witnessed a charter by John earl of Ross and lord of the Isles to his illegitimate brother, Celestine of Lochalsh, and his legitimate male heirs by Finvola, daughter of Lachlan Maclean of Duart, of the lands of Sleat in Skye,[54] but his son, Alexander, appears to have succeeded by 1456.[55]

The close integration of the MacIans into the political community of the northern part of the Lordship is underscored by the marriage in 1443 of John’s son, Alexander, to Anna, daughter of John Macleod of Glenelg.[56]  This marriage may have marked the emergence of Alexander as a significant figure in his own right, for his subsequent career spanned around the next fifty years (he was dead shortly before 1 August 1492).[57]  He was clearly a man of political weight and social reputation, featuring as a pledge for his Mackintosh neighbours and as a witness to charter acts issued at locations spread from Aros and Islay to Edinburgh. 

In a charter, wrongly dated to 28 June 1409 but datable to before 1475, issued by John earl of Ross and lord of the Isles at Aros Castle of Mull, Alexander witnessed the heritable settlement of a broad swathe of properties in the Isles on John’s brother, Hugh lord of Sleat in Skye and any male children of Hugh’s by his wife, Alexander’s daughter Fynvola.[58]  It was conceded that if Fynvola should die without producing a male heir, the lands would fall to any legitimate male child of a second wife, but that Alexander would be one of the close kinsmen and councillors of Earl John who would advise on the choice of a future wife. 

Alexander MacIan was clearly a major figure in the lordship of the Isles and was actively manoeuvring to further strengthen his influence in the northern territories of the lordship through closer bonds with the important new cadet lines of the MacDonald kindred that were being established in that territory.  This act was confirmed by James IV in November 1495 as part of his general effort to regularise landholding arrangements that had been set in place by John lord of the Isles before his forfeiture in 1493 and give some stability to the region.

It is not known when Alexander died, but his son John was named as a witness to a charter issued at Oronsay on 1 August 1492 in which he was styled as Lord of Ardnamurchan.[59]  Although he was also referred to in that charter as one of lords of the council of the Lord of the Isles, John MacIan was quick to recognise the way that the political tide was flowing in summer 1493 when John MacDonald, lord of the Isles, was formally forfeited in parliament and the Lordship effectively suppressed. 

When King James came west in August 1493 with a powerful following led by the principal political figures in his administration and based himself at Dunstaffnage in Lorne, MacIan and his associate Maclean of Lochbuie were amongst the first West Coast chieftains to make their formal submission to him, and received in return a confirmation of the lands and offices that they had held from the Lord of the Isles.[60]  MacIan was an influential and ambitious man, factors which had made him an attractive associate to be cultivated by the Campbells of Argyll as they extended their network through territory formerly dominated by the MacDonald lords of the Isles, but in 1493-5 his father-in-law Archibald Campbell, 2nd earl of Argyll, had been politically marginalised and was significant by his absence from the royal expedition.[61] 

As husband of one of Argyll’s daughters, MacIan could ultimately expect to – and did – benefit from his father-in-law’s political prominence in the administration of King James IV, while Argyll would use his son-in-law’s regional influence to help him in enforcing crown policy in the Isles, but in 1493 with his father-in-law temporarily eclipsed MacIan was building bridges directly with the young king and the clique of powerful Lowland lords around him.

Opportunity to prove his worth to James in the west quickly followed, when Alexander MacDonald of Lochalsh, the leading adult member of the MacDonald kindred and after the young grandson of the last Lord, John, the most obvious contender to lead a rebellion for the revival of the Lordship, rose in revolt.  Details of the event are obscure, but it appears that despite conciliatory moves towards him on the part of James and his councillors – including his knighting by James – Alexander may have been unable to risk the consequences to his personal reputation and following by simply accepting the fait accompli of 1493. 

An attack on Ross at some stage in 1494 appears to have been repulsed and apparently shortly afterwards he was killed on the island of Oronsay by MacIan.[62]  It has been suggested that MacIan, for whom there was no obvious personal advantage – indeed, more the risk of loss of status – in the restoration of the Lordship, was employed as ‘a promising hatchet-man’ in the Isles to do the government’s job.[63]  It is likely that his reward was the reconfirmation of his office of bailie of Islay that he had held from John, lord of the Isles, plus a substantial number of properties on the island, granted to him by royal charter of 14 June 1494.[64]

This confirmation of John MaIan’s position as a royal agent in Islay may have been the catalyst to drive into rebellion another MacDonald lord who had submitted to James in 1493 and been rewarded with confirmations of his possessions augmented with new lands and titles.  John MacDonald of Dunyvaig in Islay and the Glens of Antrim was the leading figure in the southern branch of the wider MacDonald kindred.  He may have hoped to build up his position to a considerable degree by filling the void left by the suppression of the Lordship, but the king’s promotion of MacIan’s interests and the decision to strengthen royal castles in Kintyre in early 1494 sent a signal to John MacDonald that the king did not trust him and was actively undermining his position.  The result was rebellion. 

Although MacDougall suggests that the rebellion might not have been suppressed until 1499, royal operations in the west in 1495 suggest that the affair was ended speedily in 1494.  Again it was MacIan who performed invaluable service in the suppression of the rebellion, capturing Sir John MacDonald of Dunyvaig and the Glens and three of his sons at the old seat of the Lordship at Finlaggan on Islay and sending them east to Edinburgh where the king eventually had them hanged together on the same gallows at Boroughmuir.[65]

It was shortly after that event that Mingary make its first appearance in the surviving historical record, when on 18 May 1495 King James IV issued a Great Seal charter to John Stirling, son of Sir William Stirling, in respect of the lands and barony of Keir in Perthshire.[66]  The charter settled the barony, which Sir William had personally resigned into the king’s hands, on John as his son and heir, resignation and bestowal apparently taking place when all parties involved were at the ‘Castle of Meware in Ardnamurchan’.  The king’s presence at Mingary was part of his wider efforts to build a new political establishment in the Isles after the 1493-4 upheavals. 

It was no random event; planning for a campaign had commenced in 1494 and over the winter and early spring of 1494-5 a fleet and the supplies needed for operations had been assembled at Dumbarton.[67]  On 5 May 1495, the king embarked at Dumbarton for the voyage round Kintyre and into the Isles, reaching Mingary by 18 May at the latest.  This venture was not just intended to ‘daunt’ opponents and would-be rebels but also to permit James to demonstrate his appreciation for the efforts of his committed supporters in the region and distribute the rewards for that support.

The stage-management of the 1495 event is unmistakable and its distinctly propagandistic significance as a clear statement of royal authority and favour-showing should not be doubted.  It was a truly glittering assembly of the most powerful men in the realm who accompanied James: Robert Blacader, archbishop of Glasgow, Archibald Douglas 5th earl of Angus (the chancellor), Alexander 2nd lord Hume (the chamberlain), Archibald 2nd earl of Argyll (Master of the Royal Household), John, lord Drummond (Justiciar), George Shaw, abbot of Paisley (Treasurer), and a group of leading gentlemen of the king’s council and household.[68] 

No documented record of the fact survives, but it is likely that many of the heads of the leading kindreds from the adjacent mainland and island districts came to make their formal submission to the king, and it has been pointed out that before the end of the year Hector Maclean of Duart had received a crown confirmation of his possession of the twin castle of Cairn-na-Burgh in the Treshnish islands and the castles of Dunchonnail and ‘Dunkerd’ in the Garvellachs, Alan Cameron, Captain of Clan Cameron, was confirmed in possession of Strome Castle at the mouth of Loch Carron in Wester Ross, and Gille-eoin McNeill received confirmation of his possession of all of Barra and other lands in the Uists.[69]  The dating of the issue of the royal charters to these chiefs in the period between mid August and mid November 1495 suggests that James and his councillors may have used the visit to Mingary as an opportunity to assess their value and reliability before committing to a course of action that increased royal dependence on them.

There was no question of the loyalty to the crown and significance for royal plans for strengthening the king’s authority in the west of the MacIan lineage, yet there is no indication that James displayed his personal gratitude in any way other than choosing to base himself at Mingary.  MacIan had performed invaluable service to the king in the capture of John and his sons at Finlaggan, but it could be argued that he had already received reward for his service through the June 1494 charter; it was over four years more before he received further formal rewards from King James for that conspicuous support at a time of crisis. 

Instead, in the immediate aftermath of this king’s personal display of favour to MacIan at Mingary and the charters to Maclean, Cameron and McNeill, the Lords of Council led by Archibald earl of Argyll, whose star was in the ascendant again after his eclipse in 1493-5, passed legislation designed to make clan chiefs responsible for enforcing legal summonses issued against their dependents, or face legal proceedings themselves for failure to do so.  It was an act intended to impose the rule of law and curb disorder – Argyll reporting that five leading chiefs (including MacIan of Ardnamurchan) had already given financial pledges of their good behaviour[70] – but it had the opposite effect in that it made it difficult for as yet uncommitted chiefs to persuade their yet-more-unconvinced clansmen to accept the reality of Scottish royal power. 

This more coercive policy on the part of the government was followed up in 1498 when the king conducted a second campaign in the west, this time focussed on Kintyre, where he was building a new castle at Loch Kilkerran (Campbelltown) and extending the old royal castle at Tarbert.[71]  James and a major gathering of his council waited at Kilkerran for the arrival of large numbers of Islesmen to make their formal submissions to him, as may have happened at Mingary three years earlier, but few came.  MacIan, however, may again have been one of those who did, for when the king made a further very brief foray into the outer Firth of Clyde and Kintyre district in March 1499,[72] he was rewarded by James for having made a formal submission to him of various properties and offices in the peninsula. 

At Tarbert Castle on 29 March 1499 he was finally given recognition of the services that he had performed in the suppression of the 1494 rising and for his surrender of the lands of mid Kintyre and the office of steward there, receiving a crown charter of lands in Sunart, Jura, and an extensive group of properties in the Middleward of Islay.[73]  On the same occasion he received a second charter granting him the lands of Ardnamurchan with the ‘castle and fortalice of Mingary’, to be held by him of the king for the same – unspecified – service as he had held them of John lord of the Isles before 1493.[74] 

The opposition to the new royal policy of assertiveness which the poor attendance of clan chiefs at Kilkerran in early summer 1498 had revealed may have convinced James and his councillors – including Argyll – of the need to give another clear signal of their support for one of the crown’s most consistent friends in the region; it was not a distinction which endeared MacIan to many of his fellows in the Gaelic west.

Over the following years, MacIan proved very to be the most loyal and able lieutenant of Argyll, to whom James entrusted the task of imposing strong government on the Isles.  Communication between MacIan and the king increased notably from 1500 onwards, possibly reflecting the new rapprochement that had been achieved in March 1499.  A ‘priest of Makaanis’ received generous cash gifts from James in February 1500/1, and May and June 1501,[75] in May being described as coming from MacIan, presumably with information from the Isles.  In June 1501, the king sent a man west ‘with writingis of the Kingis, and to bring ansueir agane’.[76] 

In April and July 1502 and August 1503 it was ‘a man of Makanis’ that brought news,[77] while on 16 April 1503 a cash gift was given to a man who ‘brocht lettrez fra Makane furth of the Ilis to the king’.[78]  While the real beneficiary in this process was Argyll, MacIan power also spread, largely at the expense of neighbours like the Macleans of Duart and Lochbuie.  Resentment at the seemingly inexorable rise of the MacIans was one factor that drove the Macleans to join the 1504 rebellion in the Isles in support of the bid by Donald Dubh MacDonald, grandson of the last Lord, to restore the Lordship of the Isles.  MacIan rose higher still in royal favour as a result of his activity on the king’s behalf in the suppression of Donald Dubh’s rebellion between 1504 and 1506.  James was clearly receiving alarming news of Donald Dubh’s plans in February 1504 when he sent letters by ‘ane man to pas in the Ilis to Makane’, and again in April in the hands of a man called ‘Stratoun’.[79] 

It was probably in one of these exchanges that formal instructions were sent to MacIan commissioning him to harry the lands of the rebel lords and to capture, if possible, the forfeited men.[80]  Letters (contents unspecified) sent by him to James and messages returned to him in June, July and November 1504 and March 1505, for example, almost certainly contained intelligence concerning the rebellion.[81]  The post-1499 relationship was clearly paying dividends for the king.

There were several signs of favour from James to MacIan in this period.  In November 1505, the Treasurer accounted for payments totalling £23 16s 6d spent on buying damask and cloth for a gown, red hose and new shoes for the loyal MacIan, and a further £4 10s on cloth for his priest, who served as a messenger between the west and the king’s court.[82]  The greatest sign of his standing with the king, however, came on 19 November 1506, when John MacIan was confirmed ‘for his good service’ in possession of various blocks of property in Islay and Jura that his grandfather John, son of Alexander MacIan, had held in the mid-fifteenth century of Alexander earl of Ross and lord of the Isles or Donald MacDonald, lord of Dunyvaig in Islay and the Glens of Antrim, but which had been taken into the king’s hands following the forfeiture of John lord of the Isles in 1493 and the treason and death of Sir John MacDonald of Dunyvaig, son of Donald.[83] 

A week later, MacIan possession of Mingary and the lands of Ardnamurchan was confirmed heritably by James IV on 25 November 1505 as part of general confirmation and ratification of all previous royal charters and grants in favour of John MacIan.[84]  At that time, MacIan lordship stretched from Islay (where John held the office of bailie for the king’s lands there) and Jura, where his landed heritage had just been restored by royal charter, to Ardnamurchan and Sunart, with as the centrepieces of these properties ‘the houses and fortalices of Castlemeary in Ardnamurchan and Donavagane (Dunyvaig) in Islay’.

After the end of Donald Dubh’s rising in 1506 the level of communication between MacIan and the king fell away to a considerable extent.  Only two letters to MacIan are explicitly recorded in the last seven years of the reign, one sent in October 1507 and the other in April 1508, but payments to MacIan’s priest in June 1512 and February 1513, and to one of his servants in July 1513, were probably in respect of personally-delivered messages.[85]  Gift exchange between the two men, however, suggests that the personal bonds between them were being maintained.  In December 1507, MacIan sent James two horses, with a third delivered in March 1508, while in 1512 the king sent him four saddles as a gift.[86]

Weak and distracted central government in the immediate aftermath of the death of James IV and so many of his leading councillors and magnates at Flodden in 1513 provided an opportunity for MacIan’s enemies in the west to seek to turn the tables on him.  In 1515, John MacIan was one of the principal targets of a rebellion raised by Donald MacDonald of Lochalsh and Alexander MacDonald of Dunyvaig and the Glens, who were seeking revenge for John’s part in the slaying of their fathers.  Alexander of Dunyvaig had ambitions to restore the Lordship of the Isles but Donald of Lochalsh’s principal motivation appears to have been straightforward personal animosity towards MacIan.  Mingary was apparently besieged on this occasion but does not appear to have fallen, although the surrounding district was systematically ravaged. 

A fresh rising in 1517 was even more explicitly directed against MacIan and his lands and a complaint laid before the Lords of Council in Public Affairs narrated how Mingary had been ‘destroyed’ and the surrounding district laid waste by Alexander of Dunyvaig and his associates.[87]  John survived that attack, but two years later he and two of his sons, John and Angus, were slain at Creag an Airgid.[88]  A violent opportunist who had built a personal empire from the ruins of the Lordship of the Isles, capitalising on James IV’s need for a willing and effective local agent who was not averse to spilling blood, his end was perhaps seen as fitting by those of his rivals who still lived.  His passing, however, represented the beginning of the end for the MacIans, whose heritage was to become a coveted prize over the next three decades.

2.5    The MacIan Inheritance and the Rise of Campbell Power 1519-1612

Following John MacIan’s slaying in 1519, Colin Campbell, 3rd earl of Argyll, John’s brother-in-law, moved swiftly to secure a royal grant of the wardship of the lands and castles of Ardnamurchan.[89]  Alongside this, the Earl received the gift of the ward of his late brother-in-law’s other offices including the captaincy of Dunyvaig Castle, together with the right to arrange the marriage of his young nephew, Alexander MacIan.[90]  The succession of a minor to the MacIan lordship came at a critical time in the politics of the western seaboard, where a series of powerful clan chieftains and would-be royal agents were jostling for dominance.  The eldest of John’s sons by his Campbell wife, Donald, had apparently been slain in Ireland in c.1499, leaving a young son, Alexander.[91] 

He, however, appears to have been set to one side in favour of his equally young uncle, also called Alexander.  One of this Alexander’s sisters, Catherine, was married to Alexander MacDonald of Islay, who subsequently became embroiled with Argyll in a dispute that appears to have had the MacIan inheritance at its heart and that may have been a factor in triggering the MacDonald-led rebellion in the Isles of 1528-31.[92]  Although MacDonald appears to have gained the advantage in the initial round of the struggle to control the MacIan inheritance, following Earl Colin’s death in 1529 his son Archibald, 4th earl of Argyll, showed equal determination to extend his influence within MacIan territory. 

Although the new Earl of Argyll was effectively removed from his leading role as crown agent in the western seaboard districts – in large part replaced by the erstwhile rebel and Argyll’s rival, Alexander MacDonald of Islay[93] – in 1534 he managed to secure a marriage contract for his cousin, Mariota MacIan, the sister of Alexander and Catherine, to John MacDonald of Moidart, Captain of Clanranald.  Four years later Alexander MacDonald of Islay, Alexander MacIan and John MacDonald of Moidart were all dead and Mariota had become heiress to Islay, Ardnamurchan and Sunart.[94]

Mariota’s inheritance of the MacIan lordship triggered a crisis.  The non-entries (fine for the failure of an heir to properly take up their inheritance and pay the entry-fine due to the crown) stood at eighteen times the value of her heritage, being calculated for the twelve years of her brother’s wardship after the death of their father and the six years that had elapsed since then, totalling £7747.[95]  With no hope of being able to raise this sum, Mariota, with consent of her new husband Robert Robertson of Struan, resigned her heritage into the king’s hands with the intention that it was to be settled on her cousin, Argyll.[96]  He, however, seems equally to have been unable to pay any composition for this massive non-entry charge and received no royal confirmation of possession. 

With a royal garrison established in Dunyvaig since the Isles’ rebellion, the MacDonald of Islay heir in ward in Dunbar Castle, and the only other possible MacIan heir Alexander Mac Donald MacIan passed over and apparently in exile in Ireland, the king was able in August 1541 to buy out Argyll’s interest in the MacIan lordship for £5000.[97]  The effect was to convert the lordships of Ardnamurchan and Sunart, together with the MacIan heritage in the islands, into a further component of the crown lands in Argyll and the Isles that James V had been building up.  The king’s sudden death in December 1542 brought to an end any plans to develop on that holding or to exploit it more rigorously.  As the regency government for his infant daughter Queen Mary struggled to secure the support of key provincial nobles, Argyll was soon poised to pocket the £5000 received from the late king and reacquire physical possession of the MacIan heritage.

A first attempt to regain control of the MacIan lordship by a member of the family appears to have come in 1545 when Alexander MacDonald MacIan emerged as a significant member of the council clustered around Donald Dubh MacDonald in Ireland.  Donald Dubh had escaped from custody in 1543 and was rallying support for a bid to recover the Lordship of the Isles when he died suddenly in 1545.  With hopes for restoration through a MacDonald-led rebellion against the royal settlement of the Isles dashed, Alexander MacDonald MacIan seems to have opted instead for an attempt to gain his heritage through entering into Queen Mary’s peace.  In 1548, he was admitted to the Queen’s peace and received a remission for his past crimes.[98] 

It was, however, not he who profited from the regency government’s need to curry favour with powerful regional interests, but his kinsman Argyll.  On 20 September 1550, the regency government of the young Queen Mary formally yielded up to Archibald, 4th earl of Argyll, royal interest in ‘the lands of Ardnamurchan with the castle and fortalice of Castell-Mayerie, sheriffdom of Inverness’, which had been resigned twelve years earlier into the hands of the queen’s father, King James V, by Mariota MacIan, the daughter and heiress of John MacIan.[99]  No formal record of that resignation survives in the records of James V’s reign.  This act came a week after instructions were issued under the Privy Seal for the drawing up of a royal charter of confirmation of a charter of alienation that had been given by Archibald, earl of Argyll, to James McConnell of Dunyvaig.  Argyll’s charter conveyed the eighty merklands of Ardnamurchan with the castle of Mingary and all pertinents including woods, fisheries, mills and multures.[100] 

Despite the hostility between the two MacDonald cadet lines in the early 1500s, ties between the MacIans and MacDonalds of Dunyvaig had been restored by the marriage of Catherine MacIan into the Dunyvaig line and strengthened during the period of MacDonald of Dunyvaig dominance of the MacIan heritage in the 1530s.  The strength of the relationship was revealed in the course of the feud between the MacDonalds of Dunyvaig and the Macleans of Duart in the 1580s, when the MacIans aligned with their Islay kinsmen.

It is somewhat ironic that in the midst of the feud both John MacIan and Lachlan Maclean of Duart were named recipients of the directive issued by James VI’s government on 29 July 1587 that set out arrangements for ‘the quieting and keeping in obedience of the disordered subjects, inhabitants of the borders, highlands and islands’,[101] for in 1588 John MacIan was seized under trust by the Macleans who then proceeded to besiege Mingary with the aid of a force of Spanish seamen and marines from the Armada galleon the Florencia, which was anchored in Tobermory Bay.  Despite this professional assistance and the superior military technology at their disposal, after three days’ of siege Maclean was forced to abandon his attack in the face of a muster of local clansmen.[102]  In the end, when it was clear that no advantage lay in continuing with the feud, the hostages and prisoners held by both sides were exchanged and MacIan gained his liberty before the end of 1588.[103]

After the 1588 attack on Mingary, there is a lull in record information relating to MacIan and Ardnamurchan until 1593, when John MacIan appears as one of the figures involved in the widespread disturbances that had rocked the Highlands revolving around the interlinked murders of the Earl of Moray and John Campbell of Calder.[104]  On 16 March 1593, Privy Council issued instructions lifting charges against a long list of men who had been put to the horn and arranging for them to be readmitted to the king’s peace. 

It is a diverse list ranging from the leaders of the party responsible for Moray’s murder in February 1592 through to notable ‘recusants’ suspected of conspiracies against the king and the Protestant political-religious establishment, and various Highland chieftains whose men had been involved in feuds or raids that were deemed to be linked to the Moray’s highly political slaying. 

Amongst the men pardoned was John MacIan of Ardnamurchan.[105]  No sooner had he been readmitted to the king’s peace than in June 1593 John ‘MacLan’ of Ardnamurchan, along with Angus MacConnell of Dunyvaig and the Glens, Donald MacDonald of Sleat, and their associates, received a summons for treason and lese majeste against the king, initiating a process of forfeiture, apparently related to suspicion of involvement in the plots led by Francis Stewart, earl of Bothwell.[106]  The final resolution of this episode in MacIan history is obscure.

Across this period from 1550, the MacIans were supposedly de jue occupants of Mingary and Ardnamurchan under the superior lordship of the Earl of Argyll, but there is little sign that the award of superiority to the 4th Earl of Argyll had been effective.  It was still ostensibly as a MacIan-held property that in 1608 Lord Ochiltree – enforcing James VI’s policy of pacification of the western Highlands and Islands – took assurances that Mingary would be surrendered on demand to the officers of the crown.[107] 

In the early 1600s, however, Argyll began moves to convert what may have been nominal superiority into actual possession, commencing in 1602 with a contract between himself and John mac Alasdair MacIan which involved the resignation to Argyll of all the Ardnamurchan properties.[108]  In return, Argyll restored MacIan as his vassal for annual payment of one merk feu-duty.  It seems from the terms of the contract that despite the resignation of the crown lands to the 4th Earl and James V’s earlier buying out of the Campbell interest in 1541 that the MacIans had remained as direct crown tenants holding the feuferme of the royal lands in Ardnamurchan.  Although the 1602 arrangement seemed to offer MacIan security of possession of the castle and lands, it quickly became clear that the Earl had other ideas as he secured royal confirmation and ratification of his legal position with respect to Ardnamurchan. 

Confirmation of the feuferme settlement of crown lands in Argyll to the earl was given on 30 March 1610, with Ardnamurchan included in that act and with a breakdown of the various components of the lordship given.  Amongst these were the 3½ merklands of ‘Meare’ or Mingary, although there is no specific mention of the castle.[109]  It was a short step from this point to extinguishing MacIan possession and Argyll’s ambition to consolidate his hold on the MacIan territories led in 1612 to the 7th Earl of Argyll’s issuing of a commission to his brother-in-law Donald Campbell of Barbreck-Lochow to take the castle of Mingary into his hands and place keepers in it at the earl’s expense.[110]  After some 250 years as the centrepiece of the MacIan lordship and the symbol of their power and status, the castle had been prised from their hands.

2.6    Civil War to the Jacobite Era

The 1612 commission was, of course, resisted by the MacIans, who sought to regain the castle by force.  The result was a brief episode of violence leading to the effective elimination of the remaining leadership of the family by a group of Campbell lairds, who enjoyed the backing of Privy Council.[111]  The commission was converted into tenure by a charter dated at Inverary on 30 December 1615 and confirmed by the king in July 1619, which confirmed Donald Campbell in possession of Mingary, Kilchoan and Clash, together with the keepership of the castle.[112] 

On 9 March 1625, Archibald Campbell, Lord Lorne, son of the 7th Earl of Argyll (who had converted to Roman Catholicism in 1617 and was living in exile on the Continent), received a charter from James VI confirming him in possession of the barony of Ardnamurchan.  The following year, Donald Campbell received a charter from Lord Lorne which granted him the lands of Ardnamurchan, again converting his tenure under commission into effective possession under Lorne’s superiority rather than that of his disgraced and exiled father.[113]  On 28 June 1633, Lorne received an act from King Charles I which set out the bounds of his authority in the West Highlands, Ardnamurchan being among the lands confirmed as forming part of his heritable jurisdiction.[114]

In August 1643 when the Covenanter government established the Committees of War in the shires, which were to be responsible for organising the military resources of the districts, Sir Donald Campbell of Ardnamurchan was one of the leading figures in the Argyllshire committee.[115]  His position, however, was soon to be challenged by the representative of the McConnell/MacDonald of Dunyvaig line, to whom Ardnamurchan had been awarded by the Campbells before 1550. 

In 1644, the Supreme Council of Confederate Ireland – the leadership of the mainly Roman Catholic party that had rebelled against Protestant English rule in Ireland – decided to intervene in Scotland in support of the Royalist adherents of King Charles I.  To head this intervention, the Confederate Council chose Alasdair MacColla MacDonald.  He brought an army of around 2000 Irish Macdonnells across from Ulster with the combined aim of supporting the Royalist forces commanded by the Marquis of Montrose and recovering property to which his lineage laid claim from the hands of the Marquis of Argyll, who was the principal military figure in the Covenanter opposition to Charles I in West Highland Scotland. 

MacColla’s force landed in Ardnamurchan, making a bold statement of intent to recover possession of lands that he regarded as rightfully the heritage of his line, as the first stage in a truly remarkable military campaign that saw them criss-cross Scotland winning a string of victories over Argyll and other Covenanting leaders, but which ended in defeat in 1645 and withdrawal first into Argyll and subsequently back to Ireland (where he was eventually defeated and killed in 1647).

The initial landing in Ardnamurchan had been intended to provide him with a bridgehead and a secure base was obtained when his men quickly took possession of Mingary.[116]  An account of the castle’s capture was recorded by Patrick Gordon of Ruthven in his history of the Civil War in Scotland, A Short Abridgement of Britane’s Distemper.  This account has been reproduced in RCAHMS Argyll vol 3[117] but merits repeating here for the detail that it provides:

Hee forced the stronge castell of Migarie by a desperate assailt with meir resolution, for they had nether cannon to batter nor pittard to blow up, nor scalled ladders to ascend the walles, and yet, notwithstanding the incessant shoures of muscatel balles that came from the walles, with the continuall playeing of there other ordinance, they regarded it no more then if they had been snowe balles; mairching and adwanceing speedily till they ware at the foote of the wall, then fyreing the gates and heaping on all sorts of combustible stuffe round about, they set fyre to the castell, maintaineing the fewell till they within were almost chockit; and to add more malour to those confounded defenders, the continuall thundering of muscate and cannon did so shaike the rocke as thair wall [well] went dry, and haueing some punchiones of aill, they ar forced to poure them doune aboue the gate where they were most infested with fyre; at length, finding no intermission, nor end of there enemies assailt night nor day, thrist, watching and wiriness forced them to yeeld.  This castell he maned [manned] and wictualled for a yeare from the prices [prizes] he had takin.[118]

A Latin account of the castle’s fall was produced by Giovanni Battista Rinuccini, Cardinal-Archbishop of Fermo, papal nuncio to the Irish Confederates from 1645 until 1649, but it offers no additional detail.[119]

This seizure of the castle did not pass uncontested and there followed a naval battle with five Campbell-owned warships, a conflict to which a seventeenth-century warship wreck recently discovered close to the castle probably belongs.[120]  On MacColla’s departure for the campaign in the Lowlands, he left a garrison in Mingary and the castle remained in Royalist hands until 1647 when it was retaken by the Covenanting general, David Leslie.[121] 

Amongst prisoners whom MacColla left in the castle were three Church of Scotland ministers and their wives who had been captured in a vessel heading for Ireland that the Ulster force had encountered whilst sailing north.  The account of their imprisonment suggests that few comforts had remained in the castle after the MacDonnell assault on Mingary, a position worsened during a seven-week siege by a Campbell force, during which the only drinking-water available was from rainwater that gathered on parts of the wall-head.  In September 1644, the garrison released the ministers’ wives and some other prisoners, but retained the ministers whom they hoped to exchange for MacColla’s father, Coll Ciotach, and brothers, who were prisoners of Argyll.  Before the end of the siege, unfortunately, two of the ministers had died, and the third remained a prisoner in the castle until released by negotiation in May 1645.[122]

As Scotland was again put into a posture for war in 1648, and then again in early 1649 following the execution of Charles I by the English parliamentarians, Sir Donald was again nominated to the committee for organising the military resources of Argyllshire.[123]  Given that MacColla’s garrison had only been expelled from Mingary in 1647, it is unknown what the condition of the castle and the Ardnamurchan lands generally was at that date.  Given reports of the condition of other Campbell territories following the defeat of MacColla’s army, it is unlikely that Sir Donald would have been in much position to make an effective or substantial contribution to the war effort.[124]  Sir Donald Campbell died in 1651 and it appears that the castle and lands reverted at that time to the Campbells of Argyll.

Considering Mingary’s still not inconsequential strategic significance, demonstrated as recently as the 1640s, it is striking that the castle does not appear to have been considered of any importance during the Cromwellian occupation of Scotland in the 1650s.  With major English garrisons established in massive new fortifications at Ayr and Inverness from which to ‘daunt’ the western Highlands and Atlantic seaboard, and government squadrons patrolling western waters, there may have been no significant threat perceived from the owners of the medieval fortifications that still studded the coastline and islands.  Like many former places of strength, Mingary appeared to be facing a future of quiet insignificance as the somewhat old-fashioned estate-centre of a property held by a largely absentee landlord.

There was a flurry of activity concerning Mingary in the aftermath of the overthrow of the regime of King James VII and II in April 1689.  Although the main Jacobite opposition to the new regime of King William III and Queen Mary II had been stalled with the death of Viscount Dundee at Killiecrankie on 27 July and then repulsed at the battle of Dunkeld on 21 August, a significant Jacobite force remained at large in the Highlands until its final defeat at Cromdale on Speyside in May 1690.  Until May 1690, therefore, there was a possibility that the counter-revolutionary forces in the Highlands might successfully consolidate their position and provide a base from which the exiled James VII might recover his position. 

It was in this context that on 21 December 1689, the staunchly Jacobite Robert Stewart of Appin wrote to James Seton, 4th earl of Dunfermline, who had been a leading member of Dundee’s rising and had fought at Killiecrankie, concerning an order from one Brigadier Cannon to Campbell of Lochnell to ‘render up’ the house of Mingary.  Appin requested that Dunfermline advise the Brigadier against pursuing this action so as ‘not to make ennimies … and no particular business should make men doe that quhich is not just for Lochneill swears to me that he will be as loyal as anny of our selves’.[125]  Whether or not Lochnell had any Jacobite sympathies is unclear, but there is no evidence that Mingary became a Jacobite outpost on the west coast.  Indeed, after the emergency had subsided, in 1696 Archibald, 10th earl of Argyll confirmed Lochnell possession of Mingary by a formal infeftment that granted Ardnamurchan and the castle to Alexander Campbell, 6th of Lochnell, and his son Duncan, unlikely had their loyalty been in any way suspect. 

It is RCAHMS’ view that one of those men built – or at least remodelled – the present north courtyard range, and it is certainly the view of Geoffrey Stell that the north range as it stands belongs to a pre-1715 development and, in his words, ‘undoubtedly was a barrack-block and in this case may have been Campbell-built’.[126]

Ardnamurchan was purchased from the Campbells in 1723 by Alexander Murray, son of Sir David Murray of Stanhope, who embarked on widespread improvements of the Ardnamurchan estate.[127] As summarised by RCAHMS, on taking over the property he undertook repairs mainly on the roof of the castle, where he appears to have lived on occasion.  His ambitious investments, however, yielded little profit and by 1743 the estate was heavily burdened with debt.  It was possibly on account of charges on the property that the Campbells appear to have reacquired an interest in Mingary, and in August 1745 when Charles Edward Stuart landed in Moidart at the start of the Jacobite rebellion of that year, the castle was garrisoned by Donald Campbell of Auchindoun, the factor of the Duke of Argyll’s Ardnamurchan properties.[128] 

An initially small garrison was strengthened to fifty-nine officers and men in January 1746, some of whom were involved in harrying of the properties of Jacobite landowner in Morvern and district in the months before Culloden.[129]  Just over twenty years later, possession passed from the Campbells to the Riddell family, but they do not appear to have been regularly resident at Mingary.[130]  Nevertheless, the north range of the castle, described as of ‘three stories in the modern style, the broadest part of the old wall forming the back’, with floors, rafters and boards of the roof [ ] all of massive oak, the slates of an extraordinary size, fastened by oak pins’, appears to have been largely intact in 1838 when the New Statistical Account for Ardnamurchan was written.[131]

2.7    Conclusion

From the foregoing discussion, it emerges that there is far more to say about Mingary from the historical record than has appeared in earlier analyses.  It remains the case, however, that the fundamental questions of original construction date and identity of the builders remains conjectural.  From the late thirteenth-century ownership of the lordship of Ardnamurchan by Alexander MacDougall of Lorne, it can be suggested that the castle formed a northern extension of MacDougall power reflecting the progressive expansion of their domain north and west from their mainland base from the last quarter of the twelfth century and at an accelerating pace in the first half of the thirteenth century. 

The survival of thirteenth-century architectural details in the curtain wall permits some confidence in assigning a general thirteenth-century date for the construction of the main enclosure and the buildings that originally were ranged round its courtyard and, given the excavation results from Dunstaffnage, that work should probably be seen as mid- to later-thirteenth century rather than earlier.

There is no historical record that casts any light on phases of building in the lifetime of the castle, the long period of MacIan ownership in particular lacking in evidence for developments to address issues created by the changing status and political aspirations of that family.  We have records of physical attacks on the building from the early 1500s onwards but nothing concrete to enable any of the phases of alteration to the building to be linked to post-attack reconstruction until the 1640s.  Even the dates of construction of the existing courtyard buildings cannot be fixed with certainty.  One detail which does emerge from the seventeenth-century records, however, is that the castle suffered problems of water shortage.  The description of the 1644 siege and its aftermath refers to the draining of the well through the repeated shocks of the bombardment, possibly indicating that the water-supply came from a cistern rather than from a spring.

3       Cartographic, early visual sources and visitors’ accounts

Not unexpectedly, as a major site and prominent landmark, Mingary appears on most early maps that provide coverage of the western seaboard.  However before the 18th century it usually simply appears as a symbol, variously named – Castell Megarie (Robert Gordon, c.1636-52); Chastel Megarie and Castel Megary (Joannes Bleau, 1654); etc.


Figures 3.1 and 3.2  Details from ‘A plan and perspective of the improved lands of Mingary’, of c.1734, surveyed by John Cowley and engraved by Emanuel Bowen – reproduced in ‘The True Interest of Great Britain, by Sir Alexander Murray in 1740 (©National Records of Scotland. Licensor 000-000-184-613-C)

John Cowley’s survey of c.1734, which illustrates the existing topography and proposed land and harbour improvements between the area of the castle and the settlement of Kilchoan to the west, is the first that shows the castle and associated small clachan in any detail.  Of the latter the small group of three structures to the left – probably at about the site of the existing Mingary House – seems a more formally arranged trio, perhaps also suggestive of improvement.  Improvements are shown in the vicinity – enclosures immediately north of the castle, and associated dykes running off (there seems to be a slightly different convention for denoting ditches on the plan).

While the representation of the castle in plan provides little detail (which may account for the absence of depiction of ancillary ranges – the west one must have been in existence at this date) the drawing also includes an important perspective view from the south in the form of a vignette.  This, the earliest significant representation of the structure, is demonstrably accurate and provides a number of important details, such as the dormers of the principal range, and the roof detailed with diminishing tile courses.  The landing arrangements are also clearly shown, with an access skirting the east side of the castle and rising up the brae beyond.

Roy’s map, delineated only 10 – 15 years later, adds no more detail of the castle itself.  However it does appear to confirm the existence of enclosures and improved land in the immediate vicinity of the castle, and the presence of ancillary structures.

Figure 3.3  Detail of Mingary Castle and immediate hinterland, correctly oriented, produced in 1806 by Bald for Sir James Riddell [revised]

There are few topographic accounts that add significantly to the understanding of the former appearance of Mingary, particularly from the period before the middle of the 19th century when parts of the structure still retained its roofing, as suggested by the 1838 New Statistical Account, quoted in the preceding section.  The outer reaches of the Sound of Mull seem largely to have been bypassed by the highland tourist and the castle only seems to have come to wider public notice in the second decade of the 19th century, when visited and described by the likes of Sir Walter Scott and the artist William Daniell.    

Figure 3.4  Plate from A Voyage round Great Britain (1814-25) by William Daniell and Richard Ayton

William Daniell, who visited in 1813, illustrated the exterior of the monument in its setting.  In the accompanying text he included the following note,

The ruins have not the magnitude which we usually associate with the idea of a castle.  The structure is of a triangular form, with the corners rounded off; and a house, with a few windows in front, occupies nearly its whole breadth.  It is now the property of Sir James Riddell.  There are a few huts in the vicinity.

Daniell, 1818, 70

Much of the remainder of his description and historical note simply paraphrases that of Sir Walter Scott, in an appendix to his epic poem, The Lord of the Isles (281-2), written following a voyage to the west in 1814, published in 1815.  Between them Scott and Daniell seem to have brought this remote castle to wider public notice.  However even JMW Turner, who compiled a sketchbook of the castles lining the Sound of Mull in 1831, omitted Mingary – although some of his views are very sketchy and are not firmly attributed.

4       Earlier analyses of Mingary Castle

4.1    Introduction

There have been three major commentaries upon the understanding and analysis of Mingary Castle. Unsurprisingly the first is that of MacGibbon and Ross (1889); the others are those of William Douglas Simpson (1954) and that published by the RCAHMS (1980).  Most recently Geoffrey Stell (2006) presented significant discussion of Mingary though this was in the light of comparison with Castle Tioram.  Indeed most commentators compare Mingary to Tioram rather than the other way round.  This is an important subtlety, as is generally felt in the light of the present study, that Tioram is likely to be the later construction and that some of the points of comparison are perhaps more apparent than real. 

Following the general discussion of the historiography of Mingary (section 2.2, above), which examines these sources in more general terms, it is also useful to note some of the detailed analytical observations that have been made.

4.2    MacGibbon and Ross

In their Castellated and Domestic Architecture of Scotland (Volume III, 1889, 43-46) David MacGibbon and Dr Thomas Ross presented the first significant analytical account of Mingary Castle.  They neatly encapsulated the strategic significance of the castle,

The situation is very conspicuous, and the castle is well seen from the Sound and from Tobermory, which is about six miles distant.  Occupying, with its irregular outline, the whole of the top of an isolated rock from 20 to 30 feet high, close to the shore, this fortress guards the entrance from the open sea, both to the Sound of mull and to Loch Sunart, while it is so placed as to command a view down the greater part of the Sound.  The Point of Ardnamurchan, which is only a short distance off to the westward, formed a division between the Norderies and the Suderies [actually between the northern isles and southern isles].  Mingarry thus possessed the gateway to the southern division of the islands.

Indeed most coastal surveys and maritime charts feature the castle prominently.  The structure, when retaining more of its exterior lime-work, must have been a highly visible navigational marker.

Figure 4.1  View from the NW, MacGibbon and Ross (1889)

MacGibbon and Ross plan

Figure 4.2  Plan, MacGibbon and Ross (1889)            

Figure 4.3  Plan WD Simpson (1954) – see below

MacGibbon and Ross’ plan and view of the castle from the NW are useful in a number of individual details they present.  Most significant is the evidence that the existing causeway across the rock-cut ditch to the landward entrance had been detailed with parapet walls on either side, parts of which then still survived closer to the entrance itself.  Their view suggests that dressings still defined the entrance at that stage – jambs and a lintel stone seem to be represented.  It is possible that the plan indicates the stair within the principal range was more intact in the later 19th century; the stair si shown with an elongated newel pier.  The cross wall within the western range shows more detail of its eastern end than now survives, suggesting there had existed a press on that side of the fireplace and no intercommunication between the chambers. 

Their text provides further information, that the seawards entrance was defended by an iron gate, which still exists – this evidently a yett, and the observation that it was the inner door that had been wooden (i.e. the yett must have occupied the outer of the two rebates that exist – also see section 5.iv.f, below).  The plan also includes some significant inaccuracies – particularly that the north and east walls are shown at considerably lesser thickness than is the case; the projections of the eastern perimeter are also missing.

In terms of their analysis of the castle they recognised its antiquity and articulated a number of principal analytical observations to be made of the fabric,

We have here undoubtedly one of the primitive castles of the Western Islands.  The general plan of the enceinte – the small doorways – the narrow windows with their pointed arches, similar to those at Duart and Dunstaffnage, all indicate a date as old as the thirteenth century.  The parapet also seems to be of the same date, and the angle turrets have a very simple and original appearance.  It is possible, however, that the parapet may have been reconstructed and the turrets added in the fourteenth century.  They have been further altered in later times to suit firearms.

The existing principal range they note as probably an erection of the last [i.e. 18th] century … and … the offices and outhouses, apparently of the same date.  Of the latter it is observed that the range to the west had contained apartments on the upper floor and cellars beneath, and that to the east seem to have been stables or byres.  It is not clear whether this was supposition or based on evidence that still survived at that stage.

4.3    W Douglas Simpson (1938-54)

W Douglas Simpson’s account was published in the Transactions of the Glasgow Archaeological Society in 1954 was based on inspections carried out in 1938.  His plan, figure 4.3 above, is a re-drawing of MacGibbon and Ross’, with some corrections of detail.

4.4    RCAHMS (1970-80)

Much the most authoritative existing account of the castle is that of the RCAHMS, published in their Argyll Inventory volume 3, Mull, Tiree, Coll and Northern Argyll,in 1980 but based on survey work carried out in 1970 and 1971.  This presents a comprehensive description and analysis of the building, a brief review of historical source material, and new survey drawings that included phased plans at five levels, an elevation of the principal block internally, and individual details of early openings and features of the landward-facing parapet.


The RCAHMS’ original survey drawings of 1970-1 (©RCAHMS DP102198 and DP102199 Licensor

Many of the RCAHMS’ analytical observations of Mingary are discussed in relation to the individual area or feature within the account of the castle presented in section 5, below, it is thus unnecessary to reiterate these here.

4.5    More recent assessment

In his entry in The Buildings of Scotland : Highlands and Islands (1992) John Gifford provides an excellent and succinct account of the understanding of the castle to date, this primarily derived from the RCAHMS’ interpretation.

Recent study that makes reference to Mingary Castle is wholly coloured by comparison with Castle Tioram – this has been particularly the case since the purchase in 1997 of the latter by new owners, and the ensuing commission of an extensive series of detailed assessments, carried out by GUARD and others.  Of these significant new architectural commentary had been provided by Caitlin Evans and Allan Rutherford (GUARD, 1998), and more recently by Tom MacNeill (2004), and Geoffrey Stell in his Statement of Significance (2006).  Stell in particular includes an important essay on the evolution of understanding of the early castles of the western seaboard and more detailed comparison of Tioram with other sites, in particular Dunstaffnage and Mingary.  For Castle Tioram, he proposes a later dating than has generally been accepted, to the early-mid 14th century; he also proposes that this revised dating be extended to Mingary Castle as well.

The present study suggests that the traditionally held dating of Mingary to the mid-late 13th century is more likely to be correct, this on the grounds of architectural comparison and historical context.  If Mingary is of considerably earlier date than Tioram, then the comparison should take the detailed understanding of Mingary as the starting point.  It is also suggested that direct comparisons with Castle Tioram should be subject to more caution than is usually the case, its superficial similarity and geographical proximity perhaps belying some more fundamental differences.

5       Description and structural analysis

5.1    Introduction – methodology

The present study involved two detailed site inspections carried out on 13-14 September and 4-5 October 2012, these involving Tom Addyman, Richard Oram, Kenneth Macfadyen and Ross Cameron.  All accessible areas of the site were considered in detail, including some brief assessment of the immediate surrounding landscape, notably the fields to the NE of the castle and parts of the adjacent shoreline.

Existing available survey imagery of the castle was marked up on site with analytical information.  The imagery comprised a combination of the RCAHMS survey records and Wighton Jagger Shaw Architects’ as-existing plans and outline elevation drawings.  These latter were partly based upon a recently undertaken laser-scanning exercise, and upon new topographic survey work.  Existing un-rectified photographic records were also used as a basis for marking up.  Some additional survey work was undertaken by hand measurement in individual areas, particularly in the north and west courtyard ranges.

The exercise was intended to map phasing information visible in the upstanding fabric of the castle, this as a key part of a general review of the understanding of its evolutionary history.  As well as the major phase and construction breaks this mapping exercise extended to the plotting of the individual construction lifts that are such a feature of the appearance of the structure, and in particular its external walling.  As is often the case variation in the character and method of construction proved to be a principal indicator of period of construction at Mingary.  The majority of secondary phases of construction, both later medieval and post-medieval, demonstrably cut across or otherwise disrupt the pronounced horizontal banding of the early masonry.

The drawing set presented in this report is included as an interim reappraisal of the phasing and interpretation of the castle – see section 6, below.  Many areas remained inaccessible for close inspection at the time of survey, while other areas remain obscured by debris.  It is anticipated that considerable refinement will be possible as additional evidence becomes available.

In the following written account it was felt unnecessary to reproduce the detail of the RCAHMS’ systematic and lengthy description of Mingary except where it seemed beneficial to develop their analysis of the structure or to provide additional observation of the fabric.  In many respects the RCAHMS’ analysis is demonstrably correct though in some instances an alternative interpretation is offered and additional commentary made.  The structure in this report is broadly chronological.

Photography was taken of much of the castle though this primarily in support of the present report and was not intended as a formal or systematic record.

It is hoped that the result of this reassessment of the architectural history of the castle has been to refine its understanding.  In a number of areas significant new evidence for additional complexity is proposed and in some respects this represents a major reinterpretation of the standing fabric.

5.2    Geology, building materials and character of construction

W Douglas Simpson provides an excellent synopsis of the geological formations that underlie the castle,

Mingary stands on a rocky promontory, overlooking the sea, and rising to a height of some 25feet above the foreshore.  … The rock is a sill of craignurite, greenish-grey in colour, in roughly formed vertical columns, resting upon a lower basaltic intrusion of a more basic composition – the whole over-lying shales of lower Liassic date (Broadford beds), deposited under shallow water conditions.  Out of these beds some fossils may be picked, notably the characteristic zonal Ammonite Arnioceras.  Unfortunately owing to the baking and hardening of the shales caused by the volcanic intrusions, these fossil shells are poorly preserved

At least four quartz-dolerite dykes traverse the sill upon which the castle stands, and are continued eastward through a broad foreshore outcrop of limestone, grey and rather earthy in composition, also belonging to the Broadford series, and heavily charged, though in scattered patches, with good specimens of Gryphaea arcuata.  In spite of its impurities, this limestone has been quarried for burning, and doubtless it supplied the lime used to build the castle.  The whole coastal exposure here, forming as it does a section across the skirts of the great Tertiary volcanic complex of Ardnamurchan, is of the highest geological interest.

Figure 5.1  Photographic view of the south and west curtain showing the character of masonry construction and the solid geology below [revised]

Unsurprisingly Mingary Castle is for the most part built from these stones available in the immediate vicinity.  Clearly much of this material was quarried from the rock-cut ditch that extends along the north perimeter.  It is likely that other material was extracted from elsewhere around the foot of the castle along the seaward-facing sides where there is some evidence for the enhancement by quarrying to a vertical face.  It is possible that the smooth gently sloping expanse of bedrock framed by the crag on the seaward side was in part deliberately exposed and enhanced as a place to draw up vessels in the protective lee of the crag.  Quarrying of the overlying material will also have generated substantial additional quantities of building stone.

Of considerable additional interest is the variety of freestone – i.e. sandstone – employed for dressings.  In the absence of immediately local sources these were necessarily imported, doubtless by boat.  The RCAHMS described the stone employed for the dressings of the earliest phase, a coarse-grained buff-coloured sandstone probably deriving from the Lochaline area.  They also noted dressings that appear to be composed of a finer-grained sandstone of pale green colour more akin to Carsaig stone, from the southern shore of Mull, and that the dressings of the sea gate were partly of buff coloured sandstone and partly of dressed blocks of shale and schist.  The variation in the type of freestone employed bears a close relationship to the individual phases of construction – and the quarry sources that must have been available at each period.

5.3    The early castle (mid-late 13th century)                         

5.3.1   General                                                                          

The early perimeter wall of Mingary is notable for its horizontal coursing, as described by Simpson,

Random rubble, brought to course, showing a markedly striated texture and [a] tendency to introduce rows of large blocks set with their broad faces outwards.

This type of mass masonry construction is particularly distinctive at Mingary on account of the variation of stone types that can be seen.  Castle Tioram also displays such coursing, but there it is less obvious because of the greater uniformity of stone employed; it is also a feature of other early fortresses of the western seaboard, such as Dunstaffnage and Sween.  It seems to have involved the formation of lifts by means of the construction of inner and outer faces, generally between 0.5m and 1m in height (though occasionally taller) and the infilling of the void between by what appears to have been a layering of packed rubble stone and hot lime mortar mix laid over.  The character of the interior of this early walling is best seen when looking up at the ceilings of the two early garderobe chambers on the east and west perimeters, where the timber lintels have subsequently perished – the pouring of the lime mix over dry-laid stones is quite apparent.

Figure 5.2  Photographic view of the east and SE curtain showing the character of masonry construction and the solid geology below [revised]

The construction lifts provide much important information about the process and progress of construction.  While many minor lifts are apparent there are a number of major lifts that may indicate seasonal breaks.  These latter are most clearly seen on the landwards-facing sides where construction seems to have been planned and executed in three or four stages – the lower walling rising from the bedrock to about 1.5m – 2.0m in height (the establishment of a level construction platform), the ground floor up to just below the first floor windows (just over 5m in height), and the walling above (3m to the parapet base; about 2m above that).  There is clear correlation between some of the lift boundaries and features of the build – eg. a build line at parapet walk level – that reflect natural stages of the construction process.

There are two major stepped breaks visible on the west perimeter, the steps corresponding to individual courses, see figure 6.9.  This may be a feature of the construction process. Suggesting the the curtains on the landward side and principal seaward approach were built first.  However it is possible that the masonry between, which is less finely coursed than the remainder of the perimeter and perhaps slightly cruder in execution, is of a secondary build (this possibility is indicated on the plans and elevations).  This work is nonetheless likely to be of medieval date.  One possibility may be a failure of the bedrock outcrop below the west perimeter (where even today the fissuring is notably unstable), perhaps requiring an extensive rebuild.  The well-developed coursing provides a key indicator of the surviving extent of the original fabric of the castle; in subsequent phases it is far less pronounced or wholly absent.

5.3.2   The early curtain

The surviving fabric of the original castle principally consists of the exceptionally well preserved curtain wall.  If an early dating is accepted for the slight anomaly of the western side, its circuit is essentially complete.  The external simplicity of the main body of its walls is only punctuated by a shallow garderobe projection to the SE, which seems to be integral to the fabric of the curtain, and by a more prominent tower-like garderobe projection at the junction of the east and SE aides.  The latter is now recognised to be a construction of considerably later date.

Figure 5.3  The landward-facing north curtain

A notable feature of the build is the provision of rounded corners, which may be a consequence of the unavailability of freestone for quoining, and is a feature of a number of other fortresses of the western seaboard, most prominently Castle Tioram, and other geographical areas where freestone is harder to come by (e.g. Aberdeenshire). 

Figure 5.4  Plan of the castle at ground floor level showing surviving (red and darker pink) and possible projected extent (pale pink) of medieval fabric.  The shadows of the projected courtyard structures are superimposed [revised]

The curtain is massively thicker on the landward approaches (the northern part of the west wall, the NE, north and east walls), being up to 2.7m in thickness.  By contrast the remainder of the perimeter walling is only of about 1.8m in thickness.  On the landward approaches the early curtain rose to about 12m in height; to seaward this decreased to about 9.0m though in most areas early fabric now only upstands to about 7.5m.

5.3.3   Entrance arrangements

There are two entrances to the castle, one to landward to the NW and the other seawards-facing to the south.  The former is certainly part of the original build but has seen considerable reworking of its exterior parts; the latter has been largely reworked, and all of its dressings replaced.  However it is also likely this had been a feature of the early castle and the draw-bar socket may be a relict of the original arrangement. 

The landward entrance retains dressings internally; of pale sandstone, a number of these preserve the fine diagonal tooling that is a common characteristic of medieval workmanship.  Externally it is possible that the threshold stone may form part of the original entrance and beneath this there are the remains of three corbels.  The better preserved of these, to the west, retains a socket in its upper surface, figure x.  The corbels seem to have formed part of the original construction and were evidently the pivot point for a retractable bridge, the socketed corbel tops intended to hold an horizontal timber member. 

Figures 5.5 and 5.6  Entrance area interior and, right, detail of corbelling at threshold, one socketed

Evidence for the mechanism for raising the drawbridge remains obscure.  An existing opening at the wall-face above the entrance is certainly of secondary date.  The RCAHMS’ account, published in 1980, notes evidence for slots above and to either side of the entrance,

Another early, and possibly original, feature is a pair of slots set at slightly different levels on each side of the segmental-headed arch of rubble that now spans the opening.  These slots may have housed chains for raising the drawbridge.[132]

It proved difficult to identify these features in 2012 as its seems that the evidence is now obscured by repair work carried out in the intervening period; such evidence is even difficult to identify in photographs taken before these works were carried out.  The implication of the RCAHMS’ observation is that the slots may have been rainures, apertures through which a winding mechanism was connected to the bridge, presumably by chains.  If this had actually been the case at Mingary it could either imply the former presence of a mural chamber above the entrance, or that the chains were fed down again into the interior of the entrance passage.  Perhaps on account of the various episodes of rebuilding in the area of the entrance little additional evidence was seen to suggest there had been such a chamber or, indeed, any evidence for an access to it internally.  The purpose of the existing secondary aperture seen externally immediately above the entrance also remains obscure, the RCAHMS suggest that the horizontal iron stanchion embedded between the dressings that define the feature may have been designed to secure a pulley block.[133]  There does seem to have been some form of void within the walling that is behind this feature, but is now largely in-filled with rubble masonry.


Figure 5.7  Entrance in 1974                                      

 Figure 5.8  Entrance in 2012

(©RCAHMS sc 05 581258S


The existing masonry of the causeway overlies an earlier abutment at its NW end.  This may well have formed part of the original arrangements, defining a bridge span of about 3.75m.

At the exterior angle that forms the junction between the SW and south sides of the curtain there is evidence that a flanking wall formerly ran out to the south, the wall face patched following its removal.  With no obvious evidence that this was a secondary insertion it is possible that this is a feature of the original build.

5.3.4    Site of hall range                                       

That the principal accommodation of the castle lay on the site of the existing north range has never been disputed.  That this must have been the case is demonstrated by the presence externally of a series of windows, six in total (four in the north curtain, all partly or wholly blocked, and two to the east), that are clearly wholly integral to the surrounding rubble fabric of the original build, formed of pale sandstone dressings and of diagnostically early character, the narrow lancet form invariably compared to the windows of Dunstaffnage.  The three at first floor level had clearly been of generous proportions internally, with broad ingoes defined by dressings at their inner angles but with arched heads of rubble, and the remains of projecting stone benches on either side – clearly relate to a hall interior and, possibly, an associated chamber. 

Figure 5.9  Early windows within the north curtain         

Figure 5.10  Remains of early window visible

internally truncation of bench seats can be seen

Large sockets in the interior face of the north curtain demonstrate the positions of the common joists of the hall floor.  Four sockets remain open further to the east; additional probable sockets, now blocked, were identified further westwards, within the western compartment of the existing range.  Curiously the two groups of sockets seem set at differing centres, those further west more close-set.  The reason for this is not apparent and the whole wall area awaits more detailed scrutiny.

The fenestration helps to define the former extent of the range, the westernmost window lying a little short of the entrance area – the west wall of the range must therefore have lain in approximately the same position as the existing.  There is no visible evidence upon the internal wall face of the curtain to suggest tying-in points (whether for timber or masonry) elsewhere.  While certainly the range must have incorporated a hall, and that there is little doubt that the west wall of the hall itself lay on about the line of the existing W wall, it is just possible that the range extended over the entrance area and ran fully to the west curtain, though again there is no evidence at the internal wall face of the curtain to demonstrate this. 

A further feature that must relate to the interior of the original hall range is the existing first floor entrance within the east wall that now leads to a mural passage and externally projecting garderobe tower.  Though modified the lower parts of this entrance are of the original build, employing pale dressings and detailed with a chamfer; those of the northern jamb are preserved to higher level, the uppermost suggesting the beginnings of an arched head.  It is clear from the present analysis that in its original arrangement this entrance only provided access to a mural stair that rose up to the north.  The lower parts of the stair survive including steps and some of the original stone lintelling of its ceiling, this of large unworked slabs.  The lower part of the stair was later broken through and the existing southwards-running passage formed to provide access to the garderobe tower (this secondary work is described below).  The stair itself was lost at an even later date when the flues of the existing east wall fireplaces were inserted.

Figure 5.11 Remains of early entrance in the east wall, first floor

The stair rose to the NE angle of the curtain at second floor level; here it gave access west to an arched entrance and then beyond to a mural chamber within the north curtain, this served by the single and double-lancet windows seen externally.  The interior of the chamber was subsequently in-filled with masonry and is inaccessible.  Its interior arrangement thus remains uncertain  – whether it provided access to an upper level of the hall range or was simply a mural chamber has not been determined (its projected outline in figure 6.4, below, follows that suggested by the RCAHMS).  However there is little obvious evidence for an entrance into an upper storey although the interior wall face of the north curtain is obscured at points by later wall plaster, and by the existing cross walls.  As commentators have noted, it is a feature without clear parallel although the possibility it functioned as an oratory, which might explain the particular embellishment of the paired lancet, is suggested.

The presence of the first floor entrance to the east suggests the line of the south wall of the early hall block must have lain further to the south.  That it was approximately upon the line of or just to the south of the south wall of the existing block is likely.  This is further suggested by an interior discontinuity in the thickness of the curtain wall masonry at the angle between the eastern and SE curtain sections.  Unfortunately this area has been disturbed by later construction relating to the formation of the later garderobe tower.  It is at this junction that the wall thickness decreases and there is now a slight but uncomfortable dog’s leg in the alignment of the interior wall face.  It is possible that this area marks the point of junction of the south wall of the hall block.  It could be hypothesised that for there to be such a dog’s-leg at the internal wall face that the walling running in to the junction may have been of masonry rather than framed.

The siting of the early hall range took best advantage of the limited space available, erected on the highest part of the interior and with its main courtyard elevation south-facing which, in spite of the surrounding curtain, would have taken maximum advantage of available daylight.  For this reason and because of the greater security afforded by the interior it is probable that the elevation contained more generous fenestration than was provided within the north and west walls.  It is also likely that the hall level was accessed directly by means of an external stair.  The lower level had most likely formed a timber-ceilinged undercroft accessed by separate entrance off the courtyard.  The height of the structure remains uncertain. This in part relating to the possibility there had been an upper storey, or not.  If this had not been the case the hall would typically have had an open roof structure; again no evidence was seen to suggest whether this had been double-pitched or a lean-to structure rising against the north perimeter.  Detailed access to the upper parts of the north perimeter may reveal more information.


Reassessment of its upstanding fabric led to the recognition that the existing principal accommodation block, which has usually been described as an addition dating to the later 17th or early 18th centuries, actually represents a pre-existing structure that saw extensive remodelling at about that time.  In fact much of the masonry of the earlier building remains intact, the existing fenestration simply having been broken through its south frontage.  The possibility is raised here that this surviving earlier fabric may incorporate parts of the south and west walls of the original hall range.  However at the time of writing, pending more detailed examination, it is suggested that this may be a secondary, later medieval or early post-medieval construction as it stands, and is described as such in section 5.iv, below.  The character of its masonry construction and the absence of tying-in of its walls to the curtain together suggest it to be a secondary construction.

5.3.5   Other features of the interior

The remainder of the interior – specifically the interior walling of the early enceinte – retains comparatively few individual features and no readily recognisable evidence for other courtyard ranges that may have once abutted them at the early period. 

Two garderobes survive though both have seen subsequent modification.  In the western perimeter the garderobe seems to have been access by an entrance, no longer existing, that was reached part-way up the stair to the parapet.  This entrance was later moved slightly northwards, the southern internal splay now the only visible remnant of the original.  After entering the original entrance a recessed area immediately adjacent on the north side would have contained a hatch below to access a lower mural chamber, a possible pit-prison, that still survives though partly rubble-choked.  The garderobe itself ran off to the south and is still in a reasonable state of preservation.  The ceiling of the passage had evidently been lintelled over in timber, all now gone.

The eastern garderobe partly occupies an external projection and is also reasonably well preserved internally.  However the dressings of the entrance no longer remain and its south side is partly in-filled with later masonry.  The latter extends to the east side of the chamber where part of it in-fills an original light.  This chamber had also had a timber-lintelled ceiling.  There is no indication that this garderobe may have been accessed from within a structure within the courtyard.

The other principal feature of the interior walled circuit is the forestair giving access to the parapet-walk.  This rises steeply in two flights at the NE angle.  The lower, southernmost part of the stair was subsequently dismantled; its remaining footing can be seen where it extended southwards, into the walling of the north end of the existing kitchen range.

5.3.6   Possible well or cistern

Occupying the NE angle of the curtain at ground floor level is a small chamber defined by post-medieval walls whose eastern part is occupied by an approximately square masonry-walled shaft extending downwards.  Though the feature is now mostly rubble-filled it is just apparent about 1m down that the masonry sides of the feature abruptly under-cut and give way to hewn bedrock, whose sides where visible to the NW, W and SW, suggest a circular or sub-circular shaft.  The feature seems likely to have been water-related and the chamber above thus a well-house or cistern-house.  It was judged more likely that the feature had been a cistern of comparatively shallow depth rather than a well of considerably greater depth.  This is likely to have been a feature of the original castle.

Figure 5.12  Vertical view of the interior of the ‘well-house’, showing upper parts of shaft to right and, left, access from the under-stair area

If the identification of the feature as a water-related facility is correct it still remains unclear exactly how it had functioned.  There would seem to be three possibilities (1) that it is simply a well excavated down to the level of underlying water-bearing geological strata; (2) that it is a cistern that collected rainwater from the roof of the hall block and adjacent parapets; or (3) that water may have been channelled to it from beyond the perimeter wall to the north.  If the latter instance the water channel would have to run across the base of the rock-cut ditch – or perhaps just cut in to its base.

As already outlined in section the account of the siege of 1644 makes important mention of issues of water supply, specifically the description of … the continuall thundering of muscate and cannon did so shaike the rocke as thair wall [well] went dry … The position of the probable well, at the NE angle of the castle, increases the likelihood that it is this feature being referred to.  At its closest point the perimeter of the well is only about two metres from the landwards-facing south face of the rock-cut moat.  Both well and moat are cut through the broad brand of igneous rock that is the upper geological stratum upon which the castle is built; this stratum is notable for its vertical fissuring.  One can imagine that ballistics impacts in the vicinity could well cause opening of these fissures, enough to drain the well.

5.3.7   The early wall heads and parapet walk

It has long been recognised that there are extensive remains of the original parapet and crennellations surviving at Mingary, embedded within later masonry on the north, east and NW perimeter.  This is a rare survival at this period – other comparable examples include embedded parapets at Rothesay Castle, and at Skipness (NE tower), parts of Dunstaffnage’s parapet, which remained in use, are also considered by the RCAHMS to be of early date.  At Mingary the RCAHMS analysed the evolution of the wall head on the north side, showing how the early parapet had been later extended, figure x, below.  At the same time they mapped the extent of survival of the original fabric. 


Figure 5.13  RCAHMS’s survey of 1970-1 showing their analysis of the landwards-facing parapet, (later redrawn for their 1980 publication); here the extent of survival of the original parapet can be seen in shaded pencil on the lower plan; coloured areas represent later phases of construction  (©RCAHMS DP102200 Licensor

Figure 5.14  The wall head of the east curtain showing the embedded merlons of the original parapet walk

The parapet was evidently drained by means of occasional sandstone spouts (two are visible on the east side, and two to the north); shedding to the exterior may have been a particular necessity in those areas, where the hall block existed within.

The formation of the individual merlons is notable for the employment of vertically-set stones defining their sides.  The copings of the early merlons have mostly been removed.  However the original detail may yet be preserved at the west end of the north side, where a section of the parapet masonry rises to greater height and returns to the south.  This southwards return, now embedded within later builds, retains a peaked top, formed up of smaller pieces of rubble stone.

In addition to the areas described it is possible that the northern part of the west parapet is also of early construction.  The former is of somewhat thicker build and different character externally to the parapets beyond – the remainder of the west parapet, and those to the south and SE were extensively or wholly rebuilt at a later period (see below). 

The wall head to the NW, above the landwards entrance, is an area of particular structural complexity that incorporates at least three phases of work, see figure 5.25, below.  Evidence for the original arrangement still seems to exist but is difficult to decipher without close access.  The southwards return of the north parapet probably marks the position of one side of a projecting bretache or hoarding.  Though less well preserved there are corresponding vertical alignments of stones visible both externally and internally, to suggest a corresponding return, the space defined between measuring 3.5m in width.  Within the upper extent of the early walling between there appear to be the remains of sockets that may have held over-sailing timber members.

5.4    Later medieval – 16th century

5.4.1   Introduction

There are various elements of the existing complex that on stratigraphic evidence are certainly secondary constructions though still of comparatively early date; many of these have been recognised by previous commentators.  However between these elements there are few direct physical relationships to demonstrate whether they are coeval or whether they represent a number of separate interventions, phases or sub-phases.  It they can be associated, together they would represent a major refurbishment of the castle, both an upgrading of the accommodation within and a modernisation of its defensive capacity.

In summary these works include the erection of the existing hall block, the addition of the projecting latrine tower at the east curtain, the remodelling and extension of the landward-facing parapet, the erection of the kitchen range (west courtyard range), the remodelling of the west, south and SE parapet walks, modifications to the two entrances, the addition of features in the area of the rock-cut ditch, and apparent evidence for repair – possibly following siege damage.

5.4.2   Hall range

The principal accommodation block occupies most of the width of the north side of the castle, excepting the entrance area to the NW, which remained as an uncomfortable and constrained introduction to the castle interior.  Previous commentators have suggested this range to date to the later 17th or early 18th century.[134]

Reassessment of the structure demonstrated it to embody considerable structural complexity, and that the features upon which earlier dating was based – such as the existing fenestration, fireplaces, gable details, the cross walls and internal planning generally – were all secondary works.  The existing windows and many features of the interior are demonstrably later insertions into pre-existing fabric and in some instances cut through earlier features or, in the case of the internal cross walls, they abut.

The surviving earlier fabric constitutes much of the principal walls that define the range – the south elevation and west gable wall.  On the south elevation can be seen the construction breaks that define the insertion of the existing openings.  Elsewhere there survives evidence for a number of relict features within the pre-existing fabric.  These include parts of the jambs and relieving arch over an opening at courtyard level, this visible both externally and internally, and cut on its west side by the existing entrance.  Part of a further relieving arch is visible internally at first floor level within the eastern compartment of the interior at first floor level; this is unambiguously cut by the existing window.  There are also probable locations of early joist sockets visible within the interior face of the south wall that do not correspond to subsequent floor levels.

Figure 5.15  The courtyard-facing frontage of the north range – [revised]

Figure 5.16  The courtyard-facing frontage of the north range

Figures 5.17 Interim elevation survey showing indicative phasing,

horizontal lift breaks, etc.

To further suggest this is a pre-existing structure is the notable thickness of the south and west walls, at 1.4m and 1.0m respectively; this would be exceptionally substantial for an early 18th century building of this scale.

Figure 5.18  Hall block, principal entrance, showing evidence for the eastern side a pre-existing opening

Figure 5.19  Part of the truncated relieving arch of the early entrance is visible internally

Figure 5.20  Springing of arch head of an early opening visible internally at first floor level, eastern chamber

That the range has not been suggested to be of medieval or early post-medieval date before is understandable.  With one exception there are barely any obviously diagnostic features to indicate an early date.  Below parapet level there are no surviving freestone dressings within the earlier fabric that may have preserved diagnostic carved detail.  The jambs of the lower level entrance appear to have been formed of angular blocks of whin rather than of dressed freestone.  Similarly the quoining at the SW angle is of carefully chosen unworked squared blocks of whin. 

In contrast to the prominent banding and variation in the locally-derived stone types employed in the early curtain, the rubble stone of the walls of the range is of a mostly homogenous construction of larger blocks of the same dark stone (the natural planes in the rock often display an orangey hue) with smaller pinnings, often laddered, between.  The whin (basalt) is a stone that effectively does not weather, the appearance of the wall surfaces perhaps belying an earlier dating for this reason. 

Only three main construction lift boundaries are readily visible.  The masonry of the lower two is very closely matched while there is slightly more variation at the upper level, but perhaps not enough to suggest a different phase of work.  The masonry of the west gable wall is also notably homogenous.

An important feature whose possible diagnostic significance has been little remarked upon is the parapeted wall head of the range.  Along the south frontage of the range there are, very prominently, multiple parapet-draining spouts.  All (21 one of an estimated 24 still survive, though some are broken) are of pale sandstone of a slightly lighter hue than that of the dressings of the existing windows below, and considerably more weathered.  These spouts had drained what must have been a saddle and trough arrangement of parapet-walk stones, these doubtless formed of un-worked slabs as employed elsewhere, now obscured by overlying turf.  WD Simpson suggested the spouts had been reused from an earlier building.[135]  However on close inspection they appear to be in situ (accepting some possible disturbance above the lintels of the secondary windows).

Figure 5.21  Multiple spouts draining the hall block parapet

That this type of wall head arrangement is diagnostically early can be suggested by comparison to other sites.  What is immediately striking is the preponderance of the detail in the earlier castles on the western seaboard and its general absence in much of mainland Scotland.[136]  Mingary is comparatively unusual in that the spouts are of freestone, where many other sites employ slabs of unworked or coarsely worked locally available stone.  While this parapet walk detail is long-lived and not closely datable its use had effectively ended by the close of the 16th century – for it to have been employed in a construction of c1700 would have been highly anachronistic. 

5.4.3   Discussion

There are a number of stratigraphic relationships between the walls of the hall block and the areas of masonry they adjoin.  As most commentators have recorded there is little evidence for the bonding of the walls of the range with the curtain wall.  The north end of the west wall clearly abuts the north curtain wall, with little apparent attempt to tie in.  This also seems to be the case where the south wall runs up to the east perimeter, at least internally.  Externally the situation at this junction is less clear – here there has been extensive rebuilding of the internal wall face of the curtain at the angle of its east and SE walls, works associated with the formation of the secondary garderobe tower, described below.  Here the masonry does appear to tie across, this apparently suggesting the hall block and garderobe works are coeval (given that the garderobe tower contains broadly diagnostic features this apparent association may be critical for the understanding of the hall).

A further relationship is apparent at this point above parapet level, where the SE angle of the hall block is free standing.  Here there is a construction break apparent where the walling of the hall block meets that of the upper part of the original wall of enclosure.  The later masonry rises over the level of the embedded medieval parapet and appears to run in to the masonry of the raised wall head, which seems to be of similar character – perhaps demonstrating the two to be coeval.

In summary it was concluded that the south and west walls that define the north courtyard range considerably pre-date the generally proposed dating of the structure to the late 17th or early 18th century.  Rather it is judged to be no later than c.1600 and possibly considerably earlier.  Evidence for only two original openings was recorded in the south frontage though doubtless there had been more.  The opening at ground floor had been an entrance; the nature of the upper level opening, towards the east end of the south frontage is unclear – a further entrance accessed by an external stair being an alternative possibility to a window. 

If the range is of later medieval – later 16th century date then it is unusual for the absence of vaulted cellarage.  It is perhaps possible that vaulting had once existed but was then removed during the subsequent major remodelling; however some evidence for its former presence might be expected and there are no obvious indications.  Similarly there are few obvious indicators of internal arrangements within the surviving fabric and it seems likely that the pre-existing features of the north and east walls, the window openings, simply continued in use.  It is possible that the joist sockets visible in the north wall also relate to this phase. The lower part of the eastern internal cross wall is slightly thicker than the walling above and may be relict pre-existing fabric.

The remains little evidence for the roof form of the range as first built.  It is possible that the roof structure was retained when the range was remodelled in c1700 (see below), and that the existing gables perpetuate the profile of the earlier structure.  The brief description given by the New Statistical Account of 1838 of the rafters and boards of the roof [ ] all of massive oak, the slates of an extraordinary size, fastened by oak pins may well indicate the retention and survival of an earlier roof structure, or the recycling of its timber.  An oak roof structure would be highly unusual in Scotland at the end of the 18th century.

5.4.4   Garderobe tower

Occupying the junction of the east and SE walls the projecting semi-circular garderobe tower appears to be an addition to the original enceinte.  Just beyond its eastern side there is a major construction break where associated masonry was let in, and evidence for a similar break on its south side. 

Intramural access to the tower was newly formed within the thickness of the eastern curtain; here a passage was broken through from the base of the existing mural stair, and similarly accessed by the small first floor entrance at the east end of the hall block.  The head of the entrance may have been raised at this time.  The formation of the passage required the remodelling of the lower part of the stair – here its ceiling was raised and it and that of the passage lintelled over in timber. 

The passage ran southwards before angling to the SE into the tower.  Formation of this passage angle required the rebuilding of part of the interior wall face of the curtain at its junction with the hall block, new masonry canted across the re-entrant.  Here a small light was formed, subsequently blocked. 

The tower itself was evidently intended to be of dual function, both as a latrine and as an additional defence.  Projecting beyond the eastern flank of the main curtain it was provided with two shot holes to allow additional defensive coverage, giving a combined arc of fire extending from north round to the ESE – covering the landward approach to the NE and the eastern foreshore.  The tower chamber was lit by a narrow light to the SE.  The shot holes provide the main evidence for the dating of the tower, presumably to the second half of the 16th century or earlier 17th.      

5.4.5   Wall head remodelling to N

The remodelling of the northern wall head, and of the parapets to the SE, south and SW has long been recognised.  The RCAHMS, in particular, undertook a detailed analysis of the complexities of the north wall head, figure 5.21; and it is unnecessary to reiterate the details here.  They and most other commentators suggest these works date to the later 16th century and reflect a desire to strengthen the defensibility landward-facing side, by raising it substantially and providing for the use of firearms. 


Figure 5.21 The RCAHMS’ analysis of the modifications to the landward-facing (north) parapet (©RCAHMS DP102200 Licensor

5.4.6   Modifications to the wall head defences to the west, south and SE

It is also generally recognised that the parapet-walk sections of the curtain to the west, south and SW has been extensively rebuilt, again in the later 16th century, figure 5.22.  The principal criterion suggesting this date is the form of the bartizans at the SW and SE angles of the curtain whose form and details are typical for that period.  The principal variation on this interpretation is that, as noted above, the northern part of the west curtain may retain its medieval parapet.  Elsewhere the masonry of the new work is very clearly visible to the exterior, this notable for its homogeneity and absence of major coursing, and better preservation of its lime-work.  The variation in the form of the embrasures is a notable feature; particularly curious are the tall and narrow loops.  The bartizans had evidently been roofed and parts of the jambs of entrances survive within the parapet walling. 

Figure 5.22  The curtain to south and west, showing evidence for the re-built wall head; note sea gate and wall protecting its approach

5.4.7   Sea gate

The existing sea gate appears largely to be a secondary rebuild, presumably of a pre-existing opening, figure 5.22.  Its dressings are of an unusual mix of sandstone and blue-grey schist – the only instance of its use at the site.  The dressings are detailed with a rounded outer aris.  Details of the yett, or iron openwork gate, within this entrance were recorded by Christison in 1887-8,

Mingary, Argyllshire.—Mr J. J. Dalgleish, the proprietor of this remote castle, informs me that the yett is in an entrance of the courtyard wall, at the water port, in a very exposed place, above and within 20 feet ‘of the sea, which probably accounts for its corroded and mutilated condition, one upright and two cross bars having entirely disappeared, as well as all trace of bolts and hinges, except a mere fragment of one of the latter. Mr Armstrong, factor on the estate, has forwarded an accurate plan of the yett and its doorway. The latter, as shown in fig. 4, h, gets narrower outwards, has shallow holes in the walls close behind the original situation of the yett, as if for some kind of bar; and 7 inches in rear of these, the usual deep tunnel and hole for a bar, the object of the position of which, at least a foot behind the yett, is not easy to explain. The yett, to suit its reduced dimensions, has been transferred to the narrower space intended for the wooden door, and it has been turned upside down. The main or landward entrance to the castle courtyard has no yett.                   

Figure 5.23  Plan of the sea gate, with yett (Christison, 1887-8)

The yett has since disappeared.

5.4.8   West range (original build)

The west range, generally referred to as the kitchen range, is a structure that has usually been assigned a dating similar to that of the principal block, namely the later 17th – early 18th century.  This is on account of similarities of construction that are apparent in its masonry – particularly the general use of the dark whin and the employment of larger square blocks of whin for quoining.

Close study of the range reveals it to be a structure of some complexity and of more than one phase.  In its ruinous and rubble-choked state there are aspects of its interpretation that remain obscure.  It was originally built as a lean-to structure against the west curtain, extending as far as the south wall of the curtain.  The range’s north wall remains almost intact however much of its frontage, except to the south, has fallen.  The roof of the range had been of a single pitch rising up to the just beneath the parapet walk, where corresponding sockets and fixing points are still to be seen.  The range may always have been of two compartments, divided roughly into 2/3 to the north and 1/3 to the south, although the upper parts of the existing cross wall seem to be of a secondary build.  With the downwards slope of the interior a lower level chamber had been formed within the southern compartment, this accessed by a separate entrance.  The lower parts of an entrance to the northern compartment still survive, and there may have been a further entrance into the upper level of the southern compartment, possibly accessed by an exterior stair.

Features of the interior of the main compartment include evidence for a major chimneybreast within the north wall, this likely serving a broad kitchen fireplace at the lower level.  The kitchen chamber had also been provided the additional facility of a deep in-cutting to the west curtain wall; within this there are further aumbry-like recesses.  There had also been an upper level chamber, this served by a smaller fireplace with projecting chimneybreast above, both cut into the face of the west curtain wall, with chimney rising at the parapet-walk.  The early openings within the range – the entrances and narrow windows further south – were not defined with freestone dressings, rather their jambs were formed with angular pieces of rubble stone.

With an alternative earlier dating for the hall range proposed, it is probable on account of its similar character of construction, that the west range is also of earlier date – possibly later 16th century.  Its surviving details, such as the large kitchen chimneybreast and smaller first floor fireplace, are not inconsistent with this dating.

Figure 5.24 General view of the western (kitchen) range, looking west

5.4.9   Outer works – causeway and flanking walls

Both ends of the rock-cut ditch were closed off with masonry flanking walls detailed with a coped top; as the RCAHMS suggests these may be of later 16th or 17th century date.  The dating of the existing walling flanking the west side of the steep stair to the sea gate remains uncertain, whether of the 16th century or later, figure 5.22.

5.5    17th century works                                                                   

The upper parts of the NW curtain section saw repeated modification.  The latest of these include much of the upper walling, infilling between the earlier work to either side within which was formed an entrance to a bretache or box-machicolation.  The masonry is of notably different character to that employed elsewhere, using a combination of slab-like stone of orange hue and recycled slates of west highland character, the latter particularly used to form up the jambs of the openings. As noted above the detailed understanding of this area requires better access.  The dating of this work, suggested by the RCAHMS to be of the 17th century, rests partly on its physical relationship to the surrounding work, being the most recent intervention.  That it is not broadly coeval with the rebuilding of the parapet immediately to the east is suggested by the contrast in masonry character.

Figure 5.25  Upper parts of the NW curtain showing evidence for wall head arrangements above the landward entrance

Figure 5.26 Interior of the NW curtain wall head

5.6    Possible siege damage – c.1644?

At points around the structure there exist areas of patching of the wall-face that may relate to repair of damage caused at time of siege.  Most notable are crudely executed areas of reinstatement of the face-work on the north and NW sections of the curtain, one major area concentrating lower down at their junction.  With little obvious reason for failure at these points, and the particular concentration of patchings at the landward-facing approach, these may well represent conflict damage.  Another area that may have had to be renewed following damage is the principal entrance, where it is clear that there is much secondary work.  As noted above the sea-facing entrance was also renewed, though this may have been a more deliberate up-grading.

Figure 5.27 Prominent patching of the wall face at the NW angle

An obvious contender for such damage is the recorded siege of 1644, the accounts of which describe an attack from the landward approach.  However one cannot exclude the possibility that the damage related to one of the earlier documented actions involving the castle.

5.7    Later 17th – early 18th century works

5.7.1   Remodelling of the hall block                      

At a secondary stage the hall range was substantially remodelled.  Externally the principal effect of these works on the pre-existing structure was the insertion in to its south frontage of an entrance and substantial windows at each level, these regularly arranged into four bays.  On account of their size and lack of moulded detail a later dating has generally been proposed for the range, this usually in the early 18th century.  The regularity and austerity of the frontage suggested to many the block may have been erected at that stage as a barracks.  However detailed study of the internal arrangements perhaps indicates the structure had been remodelled with domestic occupation in mind, apparently a compact mansion-house, this notwithstanding the later documented garrisoning of the site.

The existing interior space was subdivided into three by means to two masonry cross walls.  New gable walls were erected upon the wall heads to the east and west, and the flue-bearing western cross wall was also brought up to a chimney (the non-flue bearing eastern cross wall terminated at loft floor level).  The interior was thereby formed into larger areas of accommodation to the west, of two bay’s width, and a stairwell and less substantial chamber to the east, each of a single bay’s width.

The works also involved blocking up the existing medieval first floor windows in the north wall and at second floor level in the same wall the infilling of the mural chamber with its windows.  It is unclear whether the mural chamber was simply regarded as an inconvenience, not required for the modernisation of the range, or whether the infilling in part addressed structural concerns.

Figure 5.28  North range showing phasing and projected internal arrangements at c.1700

(it should be noted that the ground floor plan also shows subsequent blockings – in green)

Much evidence of the detail of the internal arrangements of this phase still survives, preserved in the form of sockets where features ran in, timber sections built into the wall face, or silhouetted by the wall plaster that still adheres to many surfaces.  Figure 5.28 presents in plan form an initial interpretation of this evidence at each level.             Entrance and stairwell

The principal entrance led into the stairwell, the foot of the first flight lying directly ahead.  The entrance was detailed with an over-lintel light and was flanked on its east side by a small aperture, now partly damaged, but apparently a loop to permit an additional degree of security (figures 5.18 and 5.19, above).  From the entrance area further entrances provided direct access to the ground floor chambers to east and west, an arrangement repeated on each level of the stairwell.

There is extensive evidence for the form and details of the generously proportioned stair.  This had mostly been of timber, of scale-and-platt form, rising clockwise.  A masonry baulk, still visible, evidently supported (or formed) the lower flights of the stair.  The MacGibbon and Ross plan (figure 4.2) suggests there had been an elongated masonry newel.  Sockets for the stair treads remain within the side walls, indicating flights of gentle inclination between ground and first and first and second floor levels.  The loft stair, rising against the eastern cross wall where it was supported by an off-set rather than sockets, seems to have been steeper and of a single flight.             Ground floor rooms

At ground floor level there had evidently existed service arrangements.  The larger chamber to the west preserves evidence in the interior plasterwork of the north wall that suggests there had been a north/south aligned partition, or perhaps a sub-division occupying the NW angle of the room. 

The smaller ground floor area to the east of the stairwell was itself further subdivided by an east/west aligned masonry cross-wall through which there was no access.  Though its dressings are robbed; the window lighting the interior of the larger southern space was, unlike the others of the elevation, evidently of much smaller dimension (it is possible that this was in part a pre-existing feature).  In the absence of a fireplace it seems probable that this room accommodated storage and/or cellarage.             Back-room arrangements

At the NE corner of the western room there had been a further entrance that led through the western cross-wall into a back-stairs area that was evidently partitioned off from the main stairwell interior.  This area permitted access to what may have been a narrow back-stair, and to a further entrance through the northern part of the eastern cross wall.  This latter entrance provided access to a short straight stair flight and platform above a large rectangular cavity or short shaft occupying the extreme NE angle of the early enclosure wall circuit; this evidently formed a well-head above a circular rock-cut well or cistern.

The back-stair rose to the first stair landing level; this seems to have been partitioned off from the stairwell proper but doubtless gave access to it.  The remainder of the partitioned part of the landing may have been accessed separately from the landing and used as a press, this area encompasses a large wall recess to the north.  This proposed back-stair arrangement suggests the larger ground floor room to the west had functioned as a main service area/service accommodation.  That it may not have been an actual kitchen is suggested by the small proportion of the chimneybreast in the east wall, and by the fact that the west courtyard range seems to have performed this function (though perhaps not necessarily at the later stages of occupation).             First floor

The stair provided access to what appear to have been principal first floor rooms to either side, each furnished with a fireplace in its eastern wall.  The larger room to the west seems to have been a single interior space.  A large recess inserted into the centre of the west wall may have been for a buffet, which in turn suggests dining had been a primary function of the space.  A press was formed on the north side of the fireplace, and the former hall window in the north wall was partly blocked, broad its ingo being formed into a further recess following the removal of the stone bench seats; here dook holes cut into the former window jambs indicate an architrave.  The smaller eastern room may have had a more intimate function (a principal bed-chamber?); the medieval window at the north end of the east wall was retained though its benches were clawed back.             Second floor

At second floor level the chambers on either side of the stairwell were provided with fireplaces in similar positions to those on the floor below.  Within the principal chamber to the west there is evidence for a north/south aligned partition preserved in the plasterwork of the north wall (box beds to the NW?).  On the west side of the partition line there is a blocked rectangular opening of uncertain former function and dating. 

The eastern chamber made use of parts of the medieval passage between the stairwell and the mural chamber in the north wall, a new entrance was broken through to this at the north end of the east wall (this following the blocking-off of the mural stair).  The early window in the east wall of the passage was also remodelled.  With the blocking off of much of the main mural chamber, the much-reduced remaining space may have functioned as a stool closet.

Rising from the second floor level was a further short flight of four steps which gave access to a further half-landing, the northern part of which had evidently been partitioned off to form a chamber, the evidence for which is a fireplace in the west wall and a large recess in the north wall (formed within the area of the pre-existing mural chamber), this likely intended for a bed.  The stair itself most likely immediately doubled-back, rising to loft level.             Loft level

The eastern cross wall terminated at loft floor level whereas the flue-bearing western cross wall must have risen to the roof apex and chimney above.  This wall contains parts of a fireplace serving the western chamber at this level.  The side walls of the range at this level culminated in gables rising from within the parapet.  The western gable survives intact, there is a window within and a dummy chimney at the apex.  The eastern area may have been subdivided by timber partitioning though no evidence now remains; there had been a fireplace at this level within the eastern gable.

John Cowley’s view of 1734 indicates the former presence of four dormer windows on the south side (report cover); why these should be curiously bunched to the east is not clear – perhaps a peculiarity of the drafting (an attempt to fit the dormers accordingly is shown in figure 5.28d).  Cowley also correctly depicts the arrangement of three chimneys.             Discussion

The re-planning of the former hall range resulted in the creation of a compact but intelligently arranged modern mansion house that must have provided some considerable degree of comfort.  The new arrangement of rooms, service areas and access is consistent with planning of the later 17th – early 18th century.  At Mingary these very much suggest a domestic context.

5.7.2   Remodelling of the west range 

The interior of the west range also saw extensive remodelling that may be coeval with the works to the hall block.  The existing cross-wall is evidently a secondary construction; this contained a chimneybreast whose flue angled up to the west culminating at a chimney rising at parapet level.  A somewhat confusing series horizontal alignments of sockets in the west curtain indicate modification of floor levels that do not relate to those of the pre-existing interiors.  A new entrance was formed within the north wall.

5.7.3   Causeway to NW

The existing causeway providing access to the landward entrance overlies the masonry of the abutment of the earlier retractable bridge.  It is possible that this work was carried out at the time of the later 17th – early 18th century refurbishment of the castle or, as the RCAHMS suggest, even later in the 18th century.  Certainly the parapet walls that had existed on either side, visible in MacGibbon and Ross’ 1889 view and plan, figures 4.1 and 4.2, are of more non-military character than may have been appropriate at an earlier date.  The construction of the causeway seems to have necessitated the removal of the southern parts of the flanking wall closing off the west end of the rock-cut ditch.

Figure 5.29  Vertical view of the causeway and damaged wall

5.8    Later works (18th – early 19th century?)                                     i

5.8.1   Modifications to the north range

There is some evidence for secondary modification of the remodelled principal block.  The access points through the two cross walls into the under-stair area were blocked with mortared rubble masonry.  This included closing off the then-existing access to the well house – perhaps the feature had become redundant or was accessed by an alternative means (from the floor above?).  The entrance at the west end of the western ground floor room was also blocked off.  It could be speculated that provision of such service arrangements may have been less of a necessity when the site became a garrison – perhaps these modifications relate to one of the later episodes of military occupation.

5.8.2   Further modification to the west range

A fireplace and canted chimneybreast above was inserted at the NW angle of the southern chamber of the west range, at the upper level.  This is clearly secondary to the existing cross-wall and has been let into both its fabric and that of the west curtain.  Its flue was run in to the existing chimney.  Clearly the room, which seems not to have been heated previously, was converted for domestic use.  Its floor levels do not correspond to those of the chambers in the northern part of the range.  Access to this southern chamber at the upper level was most likely by an external stair and separate entrance (MacGibbon and Ross’ plan shows no internal intercommunication between the rooms, figure 4.2).

5.8.3   East courtyard range

The existing east range has generally been accepted as the most recent major addition to the fabric of the castle, this perhaps dating to the later 18th century.  Though the lean-to range is generally without closely datable diagnostic detail there is no reason to suppose this dating is incorrect.  The character of its masonry construction is clearly different to that of the other two courtyard ranges.   

5.9    Recent repair works                                              

Two areas of recent masonry repair exist at Mingary.  The sockets left by the robbing of the exterior dressings of the NW entrance were made good with new mortar-bedded rubble construction in c.2002 by the Ardnamurchan Estate.  Internally the lower parts of the SE curtain were repaired where the walling had been undercut to provide additional space for the chambers of the east range.  Here new cast-concrete lintels were inserted and the masonry of the wall face above rebuilt with mortar-bedded rubblework.

5.10 Wall finishes

Although a more systematic assessment is required as better access permits, it was noted that there survives extensive evidence for external finishes, particularly on the exterior wall faces of the curtain.  Notable areas of survival include the upper parts of the south side of the curtain and the mid-level and upper parts of the east curtain.  Without closer scrutiny it is difficult to determine to which phase these belong or, indeed, whether more than one phase is represented.  Exterior harl remains do extend over the rebuilt areas of wall head to the SE, south and W, which suggests the harl in those areas was applied no earlier than the 16th century. 

There is some evidence for an external lime finish on the south wall of the hall block though this may relate to the original build rather than the remodelling of c.1700.  Given its sheltered location the range is perhaps notable for the absence of well-preserved areas of lime finish, though the general dampness of the interior space is certainly a major contributing factor in relation to decay.

As has been described in section vii.a, above, there are many areas of well-preserved interior wall plaster still adhering within the principal range.  These generally date to the c.1700 phase of works though there also survive areas of secondary work, and evidence for modifications and patchings.

5.11 Notes on the immediate setting

Some cursory assessment of the immediate setting of the castle was undertaken.  Upon the area of raised beach to its NE were seen a number of archaeological features that are likely to be associated with the period of active occupation of the castle but for which more precise dating remains uncertain in the absence of detailed investigation. 

These features include a possible rectilinear platform upon the small promontory above the foreshore immediately east of the castle; a well-built section of walling of notable thickness and extremely straight, across a dip of the shoreline scarp – remains of the feature are visible on either side of the small burn within the dip (a defensive wall? / a retaining wall (for a dam?) or other); a quite pronounced circular structure possibly the site of a doocot (probably too small for a hut circle); possible walling or boundary feature running between two low outcrops of bedrock; the northern of these outcrops is surmounted .by a further comparatively well-defined rectangular structure / platform.

It is possible that these features lay within a more or less formally defined and/or protected perimeter that extended eastwards just above the shore-line before angling up-slope to the north, utilising the more prominent outcrops.

Immediately to the west of the castle entrance area and causeway there exists what appears to have been an early access down to the shingle beach bounding the castle on that side.  This appears to have been deliberately formed in part, with some in-cutting of the whin outcrop in that area to create a sloped ascent rising to the east.

That there had been enhancement of the sides of the crag upon which the castle rests has already been suggested, as has the possibility that there had been some improvement to form a more effective boat noost immediately below the castle to the SW.  John Cowley’s view of 1734, figure 3.2, shows an access around the east side of the castle crag that then rises to the upper level, a feature that can still be recognised.

6       Phased plans and elevations

The following presents a graphic visualisation of the phasing and analytical information described in the preceding section.  The plans are based upon those of the RCAHMS, combined with additional topographic data supplied by Wighton Jagger Shaw Architects, and enhanced with additional phasing and analytical data resulting from field survey.  The elevation drawings are based upon Wighton Jagger Shaw’s as-existing survey that in turn incorporates data from a laser-scanning exercise.

This visualisation presents the interim results of the analysis of the upstanding fabric of the castle.  It is anticipated that considerable refinement will be possible with more comprehensive access to the fabric, augmented survey data, and clearance of the structure.

Figures :

6.1     Lower level plan

6.2     Ground floor plan

6.3     First floor plan

6.4     Second floor plan

6,5     Lower parapet level plan

6.6     Upper parapet level plan

6.7     North curtain

6.8     NW curtain

6.9     West curtain

6.10    South curtain

6.11    SE curtain

6.12    East curtain

Figure 6.1 superimposed on RCAHMS survey (©RCAHMS DP102202 Licensor

Figure 6.2 superimposed on plan [revised]

Figure 6.3

Figure 6.4

Figure 6.5 and 6.6

Figures 6.7 and 6.8

west curtain

Figure 6.9

Figures 6.10, 6.11 and 6.12

7       Conclusion

The production of this report has provided a rare opportunity to reappraise the analytical and historical understanding of Mingary Castle, architecturally one of the most significant fortified sites of the western seaboard.  While this is a required stage in the development of proposals for the repair and reoccupation of the site it nonetheless stands as a discreet study.  The work of previous commentators, most significantly that of the RCAHMS, published in 1980, has been built upon and extended, and a number of significant new observations made and hypotheses proposed.

It is also certain that as the current project progresses it will be possible to refine this analysis and understanding further, as better access permits closer scrutiny of inaccessible areas and as clearance further reveals part of the fabric.

Perhaps most significant development in relation to the proposed project is the better understanding of the north and west courtyard ranges, both of which appear to be of considerably earlier origin than has been generally reported to date.  The complexities of the interior arrangements of the two ranges are now better understood although important questions still remain to be resolved.

An assessment of the significance of the site, made in the light of this study, is addressed in a separate report, Mingary Castle, Ardnamurchan : Assessment of Significance and Impact Assessment (Addyman Archaeology, January 2013).

8       Bibliography

8.1    Unpublished Primary Sources and Reports

·         Evans, C, and Rutherford, A, Castle Tioram : 6 Building Analysis and Interpretation (GUARD Project 519, 1998)

·         McNeill, T E, Tioram Castle, an account of its cultural significance, (Historic Scotland, Edinburgh, 2004)

·         National Records of Scotland, Papers of the Leslie Family, Earls of Leven and Melville: Jacobite Papers GD26/8/64.

·         Stell, GP, Castle Tioram: A Statement of Cultural Significance, (2006)

·         Wessex Archaeology, Mingary, Ardnamurchan, The Sound of Mull, unpublished Designated Site Assessment for Historic Scotland, 2007.

8.2    Primary Sources

·         Accounts of the Lord High Treasurer of Scotland, eds T Dickson and J B Paul (Edinburgh, 1877-1916).

·         Acts of the Lords of Council in Civil Causes, 1496-1501, eds G Neilson and H Paton (Edinburgh, 1918).

·         Acts of the Lords of Council in Public Affairs 1501-1554, (Edinburgh, 1932).

·         The Acts of the Parliaments of Scotland, vol 1, ed T Thomson (London, 1814).

·         Campbell, H, (ed), Abstracts of the Particular and General Registers of Sasines for Argyll, Bute and Dumbarton, otherwise known as the Argyll Sasines (Edinburgh, 1933-4).

·         Dunlop, A I, and MacLauchlan, D, (eds), Calendar of Scottish Supplications to Rome, volume IV, 1433-1447 (Glasgow, 1983).

·         Fotheringham, J, (ed), The Diplomatic Correspondence of Jean de Montereul, 1645-8, volume 2 (Scottish History Society, 1899).

·         Gordon, P, A Short Abridgement of Britane’s Distemper (Spalding Club, 1844).

·         Munro, J, and Munro, R W, (eds), Acts of the Lords of the Isles, (Scottish History Society, 1986).

·         Ninth Report of the Royal Commission on Historical Manuscripts.  Part II, Appendix and Index (London, 1884).

·         The Records of the Parliaments of Scotland to 1707, K M Brown et al eds (St Andrews, 2007-2012).

·         Regesta Regum Scottorum, vi, The Acts of David II, ed B Webster (Edinburgh, 1982).

·         The Register of the Privy Seal of Scotland, vol 4, 1548-1556, ed J Beveridge (Edinburgh, 1952).

·         Registrum Magni Sigilli Regum Scotorum, eds J M Thomson and others (Edinburgh, 1882-1914).

·         Registrum Secreti Sigilli Regum Scotorum, ed M Livingtsone, vol I (Edinburgh, 1908).

8.3    Secondary Works

·         Cameron, J, James V: The Personal Rule, 1528-1542 (East Linton, 1998).

·         Campbell, A, A History of Clan Campbell, Vol 2: From Flodden to the Restoration, (Edinburgh, 2002).

·         Christison, D, ‘Additional Notices of Yetts, or Grated Iron Doors, of Scottish Castles and Towers’, Proceedings of the Society of Antiquaries of Scotland, XXII (1887-88), 286-320

·         Cruden, S, The Scottish Castle, 1st edition (Edinburgh, 1960), 3rd edition (Edinburgh, 1981). 

·         Daniell, W, A Voyage round Great Britain undertaken in the year 1813, Vol 3 (London, 1818)

·         Donaldson, G, Scotland: James V to James VII (Edinburgh, 1965).

·         Dunbar, J, ‘The medieval architecture of the Scottish Highlands’, in L Maclean of Dochgarroch (ed), The Middle Ages in the Highlands (Inverness, 1981), 38-70.

·         Duncan, A A M, and Brown, A L, ‘Argyll and the Isles in the Earlier Middle Ages’, Proceedings of the Society of Antiquaries of Scotland, XC (1956-7), 192-220.

·         Fergusson, J, Argyll in the Forty-Five (Edinburgh, 1951).

·         Fisher, I, ‘The Heirs of Somerled’, in R Oram and G Stell (eds), Lordship and Architecture in Medieval and Renaissance Scotland (Edinburgh, 2005), 85-95.

·         Gifford, J, The Buildings of Scotland: Highland and Islands (Harmondsworth, 1992).

·         Gregory, D, The History of the Western Highlands and Isles of Scotland, from AD 1493 to AD 1625, 2nd edition, (1881).

·         Lewis, J, and others, ‘Dunstaffnage Castle, Argyll & Bute: excavations in the north tower and east range, 1987-94’, Proceedings of the Society of Antiquaries of Scotland, 126 (1996), 559-604.

·         McDonald, R A, The Kingdom of the Isles: Scotland’s Western Seaboard c.1100-c.1336 (East Linton, 1997).

·         McKerral, A, Kintyre in the 17th Century (republished, Campbeltown, 2001).

·         MacDougall, N, James IV (East Linton, 1997).

·         MacDougall, N, ‘Achilles Heel?  The Earldom of Ross, the Lordship of the Isles, and the Stewart Kings, 1449-1507’, in E J Cowan and R A McDonald (eds), Alba: Celtic Scotland in the Middle Ages (East Linton, 2000), 248-75.

·         MacGibbon, D, and Ross, T, The Castellated and Domestic Architecture of Scotland, vol 3 (Edinburgh, 1889)

·         Mackay Mackenzie, W, The Medieval Castle in Scotland (London, 1927).

·         Macphail, J R N, (ed), Highland Papers, vol 1 (Scottish History Society, 1914).

·         Miers, M, The Western Seaboard: An Illustrated Architectural Guide (Edinburgh, 2008).

·         Oram, R D, Alexander II King of Scots 1214-1249 (Edinburgh, 2012).

·         RCAHMS, Argyll: an Inventory of the Monuments, volume 3, Mull, Tiree, Coll and Northern Argyll (excluding the Early Medieval and later monuments of Iona) (Edinburgh, 1980).

·         Ritchie, G, and Harman, M, Exploring Scotland’s Heritage: Argyll and the Western Isles (Edinburgh, 1985).

·         Sellar, W D H, ‘Hebridean Sea Kings: The Successors of Somerled, 1164-1316’, in E J Cowan and R A McDonald (eds), Alba: Celtic Scotland in the Middle Ages (East Linton, 2000), 187-218.

·         Simpson, W D, Scottish Castles (HMSO, 1959).

·         Simpson, W D, The Ancient Stones of Scotland, 2nd edition (London, 1968).

·         Stell, G P, ‘Castle Tioram and the MacDonalds of Clanranald: a western seaboard castle in context’, in R D Oram (ed), The Lordship of the Isles (forthcoming, Leiden, 2013).

·         Stevenson, D, Highland Warrior: Alasdair MacColla and the Civil Wars, (Edinburgh, 2003).

·         Tabraham, C, Scotland’s Castles (London, 1997).

9       Appendix A   Historic Scotland – Listing description


This building is in the Highland Council and the Ardnamurchan Parish. It is a category A building and was listed on 20/07/1971


13th century, with late 16th and 17th century alterations and with further early 18th century additions within courtyard. Rubble with ashlar dressings. Irregular 6-sided curtain wall, following outline of rocky outcrop by shore, crenellated wallhead and wallwalk, which underwent some alterations in late 16th century and additions of 2 corbelled bartizans at SW and SE. Small seaward (S) and landward (N) entrances, the latter approached by causeway crossing defensive ditch. Various small pointed-headed lancet windows survive from 13th century, with paired example in north wall. Within courtyard a 3-storey and attic, 4-bay range has been constructed against the north wall during the 18th century, with associated single storey range (probably kitchens) against the west wall. Roofless 3-storey range has 2 rooms of unequal size in each floor, divided by scale and plat staircase.


Scheduled Monument. Commanding site overlooking the Sound of Mull. Obscure early history, subsequently fought over by MacDonald factions, besieged and taken by Alasdair MacDonald in 1644, but in 1651 appears to have reverted to Campbells of Argyll. Small force of Government troops sent to hold castle in 1745, and increased in 1746 to 59 officers and men. Probably the 18th century blocks were constructed as barracks to house troops. The attic storey of this block is illustrated with dormers in drawing of 1734.


MacGibbon and Ross, THE CASTELLATED AND DOMESTIC ARCHITECTURE OF SCOTLAND, iii (1889), pp.42-46. The Royal Commission on the Ancient and Historical Monuments of Scotland, ARGYLL iii (1980), pp 209-217. Stewart Cruden THE SCOTTISH CASTLE (1960, 2nd ed. 1981) pp 38-40, 45-7.

10   Appendix B   Historic Scotland – Scheduling description (29 March 1955)

[1] See

[2] D MacGibbon and T Ross, The Castellated and Domestic Architecture of Scotland, vol 3 (Edinburgh, 1889), 42-6.

[3] Ibid, 41-2.

[4] Ibid, 42.

[5] Ibid, 45.

[6] Ibid, 46.

[7] W Mackay Mackenzie, The Medieval Castle in Scotland (London, 1927).

[8] W D Simpson, Scottish Castles (HMSO, 1959), 8.

[9] W D Simpson, The Ancient Stones of Scotland, 2nd edition (London, 1968), 133.

[10] Ibid, 132.

[11] Ibid, 136.

[12] Ibid, 136-7.

[13] Ibid, 137.

[14] S Cruden, The Scottish Castle, 1st edition (Edinburgh, 1960), 3rd edition (Edinburgh, 1981).  For ‘Castles of the Western Seaboard’, see 3rd edition, 38-49.

[15] Ibid, 38.

[16] Ibid, 39.

[17] RCAHMS, Argyll: an Inventory of the Monuments, volume 3, Mull, Tiree, Coll and Northern Argyll (excluding the Early Medieval and later monuments of Iona) (Edinburgh, 1980), no 345, p 209.

[18] Dunbar, ‘Medieval architecture’, 46.

[19] Cruden, Scottish Castles, 24.

[20] Tabraham, Scotland’s Castles, 36.

[21] J Lewis and others, ‘Dunstaffnage Castle, Argyll & Bute: excavations in the north tower and east range, 1987-94’, Proceedings of the Society of Antiquaries of Scotland, 126 (1996), 559-604.

[22] Tabraham, Scotland’s Castles, 36.

[23] Ibid, 37; Lewis, ‘Dunstaffnage Castle’, 600 (where the date for Duncan’s work is given wrongly as ‘third quarter’ of the thirteenth century rather than second quarter).

[24] J Gifford, The Buildings of Scotland: Highland and Islands (Harmondsworth, 1992), 47.

[25] I Fisher, ‘The Heirs of Somerled’, in R Oram and G Stell (eds), Lordship and Architecture in Medieval and Renaissance Scotland (Edinburgh, 2005), 85-95 at 91.

[26] R A McDonald, The Kingdom of the Isles: Scotland’s Western Seaboard c.1100-c.1336 (East Linton, 1997), 189 n 120.

[27] G Ritchie and M Harman, Exploring Scotland’s Heritage: Argyll and the Western Isles (Edinburgh, 1985), 84.

[28] Gifford, Highland and Islands, 257-60.

[29] M Miers, The Western Seaboard: An Illustrated Architectural Guide (Edinburgh, 2008), 100-101.

[30] See, for example, ‘History of the MacDonalds’ in J R N Macphail (ed), Highland Papers, vol 1 (Scottish History Society, 1914), 5-72.

[31] Ibid, 13.

[32] Ibid, 12.

[33] Ibid, 12-13.

[34] J Dunbar, ‘The medieval architecture of the Scottish Highlands’, in L Maclean of Dochgarroch (ed), The Middle Ages in the Highlands (Inverness, 1981), 38-70 at 46.  See also C Tabraham, Scotland’s Castles (London, 1997), 35.

[35] The Records of the Parliaments of Scotland to 1707, K M Brown et al eds (St Andrews, 2007-2012), 1293/2/17.  Date accessed: 25 November 2012.

[36] The Acts of the Parliaments of Scotland, vol 1, ed T Thomson (London, 1814).

[37] W D H Sellar, ‘Hebridean Sea Kings: The Successors of Somerled, 1164-1316’, in E J Cowan and R A McDonald (eds), Alba: Celtic Scotland in the Middle Ages (East Linton, 2000), 187-218.

[38] Ibid, 199.

[39] Ibid, 201.  For the most recent discussion of this see R D Oram, Alexander II King of Scots 1214-1249 (Edinburgh, 2012).

[40] Sellar, ‘Hebridean Sea Kings’, 212-3.

[41] Ibid, 213-5.

[42] Ibid, 217-8.

[43] Registrum Magni Sigilli Regum Scotorum [RMS], vol 1, ed J M Thomson (Edinburgh, 1882), appendix ii, nos 56, 57 and 58.  Contrary to statements in McDonald, Kingdom of the Isles, 184, Morvern was not part of this group of awards.

[44] Regesta Regum Scottorum, vi, The Acts of David II [RRS, vi], ed B Webster (Edinburgh, 1982), no 72.

[45] A A M Duncan and A L Brown, ‘Argyll and the Isles in the Earlier Middle Ages’, Proceedings of the Society of Antiquaries of Scotland, XC (1956-7), 192-220 at 204.

[46]RRS, vi, no 73.

[47] For example, J Munro and R W Munro (eds), Acts of the Lords of the Isles, (Scottish History Society, 1986), no 1.

[48] Acts of the Lords of the Isles, no 5.

[49] For the MacIans and their place in the MacDonald genealogy, see Acts of the Lords of the Isles, Appendix D.

[50] RMS, vol 1, appendix i, no 114.

[51] N MacDougall, ‘Achilles Heel?  The Earldom of Ross, the Lordship of the Isles, and the Stewart Kings, 1449-1507’, in Cowan and McDonald (eds), Alba, 248-75 at 267 n 77.

[52] Acts of the Lords of the Isles, no 20, Appendix D, 284.

[53] A I Dunlop and D MacLauchlan (eds), Calendar of Scottish Supplications to Rome, volume IV, 1433-1447 (Glasgow, 1983), no 106.

[54] Acts of the Lords of the Isles, no 80.

[55] Ibid, 259.

[56] Acts of the Lords of the Isles, no B33; CSSR, IV, no 965, where the dispensation is erroneously stated as being for John son of Alexander.

[57] Ibid, 259.

[58] RMS, ii, no 2286.

[59] Acts of the Lords of the Isles, no 123.

[60] RMS, ii, nos 2200, 2201, 2202.

[61] N MacDougall, James IV (East Linton, 1997), 102-103,

[62] Highland Papers, vol 1, 56.

[63] MacDougall, James IV, 104.

[64] RMS, ii, no 2216.

[65] MacDougall, James IV, 105.

[66] RMS, ii, no 2253.

[67] This section summarises the account in MacDougall, James IV, 115-6.

[68] MacDougall, James IV, 116.

[69] RMS, ii, nos 2264, 2281, 2287.

[70] Acts of the Lords of Council in Civil Causes, 1496-1501, eds G Neilson and H Paton (Edinburgh, 1918), 41.

[71] MacDougall, James IV, 176-7.  For James at Loch Kilkerran, see RMS, ii, nos 2406, 2424, 2425.

[72] MacDougall completely misses this expedition in his analysis of the king’s reign.  On 17 March 1498/9, James was at Ayr (RMS, ii, no 2484) and on 1 April 1499 was at Rothesay (ibid, no 2485), with his presence at Tarbert being attested on 29 March.

[73] Acts of the Lords of the Isles, no A58.

[74] Ibid, no A57.

[75] Accounts of the Lord High Treasurer of Scotland [TA], ed J B Paul, vol ii (Edinburgh, 1900), 96, 110, 112.

[76] TA, ii, 113.

[77] TA, ii, 142, 154, 389.

[78] TA, ii, 367.

[79] TA, ii, 419, 427.

[80] Records of the Parliaments of Scotland, A1505/3/82.  Date accessed: 25 November 2012.

[81] TA, ii, 442, 445, 449, 467; ibid, iii, 128.

[82] TA, iii, 103.

[83] RMS, ii, no 3001; Acts of the Lords of the Isles, no A64.

[84] RMS, ii, no 2895.

[85] TA, iv, 80, 350, 405.

[86] TA, iv, 86, 104, 308.

[87] Acts of the Lords of Council in Public Affairs 1501-1554 (Edinburgh, 1932), 87; Gregory, Western Highlands, 118.

[88] Acts of the Lords of the Isles, 285; Gregory, Western Highlands, 125.

[89] Registrum Secreti Sigilli Regum Scotorum, ed M Livingtsone, vol I (Edinburgh, 1908), no 3048.

[90] APS, ii, 361, 405.

[91] Acts of the Lords of the Isles, 285.

[92] J Cameron, James V: The Personal Rule, 1528-1542 (East Linton, 1998), 240-241.

[93] For the marginalisation of Archibald, 4th earl of Argyll, in the 1530s, see Cameron, James V, 232-9.

[94] Cameron, James V, 241.

[95] ER, xvii, 750-1.

[96] Cameron, James V, 241; Acts of the Lords of the Isles, 285.

[97] Cameron, James V, 241.

[98] RSS, iii, no 2924.

[99] RMS, iv, no 527.

[100] The Register of the Privy Seal of Scotland, vol 4, 1548-1556, ed J Beveridge (Edinburgh, 1952), no 908.

[101] Records of the Parliaments of Scotland, 1587/7/70.  Date accessed: 25 November 2012.

[102] D Gregory, The History of the Western Highlands and Isles of Scotland, from AD 1493 to AD 1625, 2nd edition, (1881), 238-40.

[103] Gregory, Western Highlands, 240.

[104] Gregory, Western Highlands, 241-56.

[105] Register of the Privy Council of Scotland, vol 5, ed D Masson (Edinburgh, 1882), 54.

[106] Records of the Parliaments of Scotland, A1593/4/7.  Accessed: 25 November 2012.

[107] Gregory, Western Highlands, 324.

[108] Gregory, Western Highlands, 406.

[109] RMS, vii, no 272.

[110] Gregory, Western Highlands, 407.

[111] Ibid, 410-11; G Donaldson, Scotland: James V to James VII (Edinburgh,1965), 228.

[112] RMS, vii, no 2057.

[113] H Campbell (ed), Abstracts of the Particular and General Registers of Sasines for Argyll, Bute and Dumbarton, otherwise known as the Argyll Sasines, vol 2 (Edinburgh, 1934), no 202.

[114] Records of the Parliaments of Scotland, 1633/6/90.  Date accessed: 25 November 2012.

[115] Ibid, 1643/6/91.

[116] D Stevenson, Highland Warrior: Alasdair MacColla and the Civil Wars, (Edinburgh, 2003), 111-14; A Campbell, A History of Clan Campbell, Vol 2: From Flodden to the Restoration, (Edinburgh, 2002), 217.

[117] RCAHMS, Argyll, vol 3, 217.

[118] P Gordon, A Short Abridgement of Britane’s Distemper (Spalding Club, 1844), 65.

[119] Ninth Report of the Royal Commission on Historical Manuscripts.  Part II, Appendix and Index (London, 1884), ‘The Rinuccini MS’, 354.

[120] Wessex Archaeology, Mingary, Ardnamurchan, The Sound of Mull, unpublished Designated Site Assessment for Historic Scotland, 2007.

[121] J G Fotheringham (ed), The Diplomatic Correspondence of Jean de Montereul, 1645-8, volume 2 (Scottish History Society, 1899), 245.

[122] The New Statistical Account of Scotland, 1834-45 [NSA], vol 7, Parish of Ardnamurchan, County of Argyle, 146-7

[123] Records of the Parliaments of Scotland, 1648/3/79; 1649/1/133.  Date accessed: 25 November 2012.

[124] For discussion of the ravaging of Argyll properties by Alasdair MacColla’s forces in the 1640s, see A McKerral, Kintyre in the 17th Century (republished, Campbelltown, 2001), chapters 5-9.

[125] National Records of Scotland, Papers of the Leslie Family, Earls of Leven and Melville: Jacobite Papers GD26/8/64.

[126] RCAHMS, Argyll, volume 3, 217; G P Stell, ‘Castle Tioram and the MacDonalds of Clanranald: a western seaboard castle in context’, in R D Oram (ed), The Lordship of the Isles (forthcoming, Leiden, 2013).

[127] NSA, vol 7, 141-2.

[128] RCAHMS, Argyll, vol 3, 217.

[129] J Fergusson, Argyll in the Forty-Five (Edinburgh, 1951), 14, 21, 52, 58, 82, 118-9, 179, 200.

[130] RCAHMS, Argyll, vol3, 217.

[131] NSA, vol 7, 145-6.

[132] RCAHMS, 1980, 212

[133] Ibid.

[134] MacGibbon and Ross (1889) – 18th century; WD Simpson (1954) – none of the courtyard buildings earlier than the latter part of the 17th century; RCAHMS (1971/80) – early 18th century; Gifford (1992) – c.1696-1723

[135] 1954, 84

[136] Western examples include Dunstaffnage – 13th century – north and central section of the east perimeter; E Carrick Castle – later 14th century; Old Castle Lachlan – an early 15th century structure whose wall head arrangements were attributed by the RCAHMS to c1500, though they did not rule out being part of the original build (1992, 239); Moy Castle – dated to the first half of the 15th century, with parapet details very similar to the range at Mingary.- multiple spouts, though of slate.

2 thoughts on “Analytical and historical assessment

  1. Pingback: Mingary Castle - neues Leben einer alten Burg

  2. Stephen Moreton

    Sir Alexander Murray of Stanhope, Peebleshire, purchased the Ardnamurchan estate in 1722 and set up residence in Mingary Castle. He bought the estate with a view to working lead mines at Strontian (National Library of Scotland, ABS.3.83.26(1). Likely it was he who carried out the renovations in the early 18th century.

Leave a Reply

Your email address will not be published. Required fields are marked *