Castles are the best-known type of medieval elite residence, both in terms of popular awareness and academic study. Around 110 possible and actual castles are recorded in Perth and Kinross. This figure includes tower houses, but excludes post-1700 mock-castles. When discussing this type of site, it should be borne in mind that many residences we now call castles are not referred to as such in contemporary records, often just being named as the place. Castle is a loaded term and there is considerable debate about its use. In this research framework it is used to encompass most forms of fortified residence used by medieval elites.
While long recognised as of historic interest by the Ordnance Survey, RCAHMS and Historic Environment Scotland (and its various predecessors), most castles in Perth and Kinross have never been excavated. As a result, many important questions about the form, function and evolution of castles in the region remain unanswered. Among the gaps in our knowledge is a lack of understanding about when and how castles were introduced to the region. In the past, the adoption of castles in Scotland was often attributed to growing royal authority in the 12th and 13th centuries, and an associated Normanisation of the Scottish elites. Yet this model does not necessarily work for Perth and Kinross, or indeed most of Scotland outside the Borders. Much of what is now Perth and Kinross was part of the earldom of Strathearn, which in the 12th and 13th centuries was held by a native family with established roots in the area. It has been suggested that the earls of Strathearn, and their officials and vassals, showed little interest in castle building prior to the 1290s (Watson 2005). This theory is largely based on information from charters, partly because of the small proportion of excavated medieval elite residences. The relationship between castles, other forms of medieval noble residences, and earlier high-status sites is a topic which needs more research.
It is often assumed that early Scottish castles were mainly built of earth and timber, and typically consisted of a mound, or motte, and an associated enclosure or bailey. Yet other, perhaps less easily recognisable, styles of timber castle probably existed. A low-level cropmark enclosure at Green of Invermay (MPK1910) has been interpreted as a medieval ring-work – a type of early castle with circular earthworks but lacking a motte. Recent excavation revealed that the site was a ‘significant residence’ in the 13th century, and may have been occupied as early as the 11th century (Campbell and Driscoll 2020, 187). Green of Invermay potentially provides vital evidence for shifting expressions of lordship and the early evolution of castles in this region.
At least 13 possible mottes have been identified in Perth and Kinross (see Table 3). Several of these sites have been significantly altered or damaged over the last two hundred years. In the 19th century the motte at Mote Hill (MPK3747), Blairgowrie was removed during building work, while the possible motte at Law Knowe (MPK5379), Errol had a burial vault placed within it. Meanwhile in the 1970s the motte at Castle of Rattray (MPK5364) was destroyed by quarrying, and the motte at Barton Hill (MPK4671), Kinnaird was significantly altered by the construction of a new house. Fortunately, several other probable mottes remain relatively undamaged by later development, and could be priorities for archaeological investigation (R Oram pers comm). Only two surviving mottes have been excavated; there was a limited rescue excavation of Barton Hill in the early 1970s and the possible motte in Glen Devon Reservoir (MPK1305) was investigated in the early 2000s. The dig at Barton Hill revealed a number of features which were not visible above ground; these included the remains of a ditch around the perimeter of the castle. This indicates the need for excavation (Tabraham 1997). The mottes at Carnbane Castle (MPK187), Invervar, and Gleneagles Castle (MPK1272), Blackford have the remains of later stone castles built on top of them. Similarly, the possible motte at Cleish Castle (MPK5584) and the motte at Meikleour House (MPK5398) have later tower houses nearby. It is likely that a number of post-1400 castles and great houses are on the sites of earlier castles and residences, and this possibility should be considered when such properties are studied.
Early castles can also lie under urban development, as at Perth, where there is written evidence for a castle in about 1160. Perth Castle is often thought to have been abandoned after severe flood damage in 1209, although new historical research suggests that it may have been occupied after this (R Oram pers comm). In the early 2000s excavations in the Horse Cross area, where Perth Concert Hall now stands, revealed a medieval ditch (about 15m wide and 1.4m deep) and postholes from a possible palisade (Cox et al 2007). These discoveries have been tentatively linked to Perth Castle. Unlike some of the rural sites previously discussed, much of the Horse Cross area has been densely occupied for centuries, potentially destroying the remains of an earth and timber castle.
| HER/ Canmore ID | Place (Site and Locality) | National Grid Reference | State of Remains | Archaeological Investigation |
| MPK4671/ 30437 | Barton Hill, Kinnaird | NO 2437 2864 | Surviving motte with modern house built on it. | Rescue excavation (1971). Watching brief (2007). |
| MPK3747/ 28727 | Mote Hill, Blairgowrie | NO 1785 4544 | Motte removed in 19th century. | Antiquarian investigation during building work (1838). |
| MPK187/ 24516 | Carnbane Castle, Invervar | NN 6771 4788 | Rise in ground which has been tentatively interpreted as a motte. Remains of 16th-century castle on same site. | No detailed investigation beyond field visits. |
| MPK5364/ 30764 | Castle Of Rattray, Rattray | NO 2099 4539 | Motte largely destroyed by quarrying in 1970s. | Archaeological evaluation (2013). Watching brief (2014). |
| MPK5584/ 49626 | Cleish Castle, Cleish | NT 0820 9790 | Traces of possible motte near 16th-century castle. | No detailed investigation. |
| MPK301/ 24825 | Coney Hill, Comrie | NN 7755 2238 | Surviving ditch and motte (significantly altered by later landscaping). | No detailed investigation beyond field visits. |
| MPK1094/ 25713 | Edradynate Castle, Edradynate | NN 8808 5215 | Earthwork which has been tentatively interpreted as a motte. Stone foundations of uncertain date on summit. | No detailed investigation beyond field visits. |
| MPK1305/ 25949 | Glen Devon, Upper Glendevon Reservoir | NN 90407 04675 | Covered by Upper Glendevon Reservoir. Possible motte visible when water levels are low. | Excavated 2003 during low water levels (Cowley and Strachan 2006). |
| MPK1272/ 25906 | Gleneagles Castle, Blackford | NN 9288 0924 | Possible motte still exists. Remains of 15th-century tower on same site. | No detailed investigation beyond field visits. |
| MPK65/ 24233 | Kerrowmore, Innerwick | NN 5889 4669 | Remains of terraced earthwork which has been tentatively interpreted as a motte. Also footings from seven former buildings of uncertain date. | No detailed investigation beyond field visits. |
| MPK5379/ 30496 | Law Knowe, Errol | NO 2315 2238 | Possible motte still exists. 19th-century burial vault inserted into mound. | No detailed investigation beyond field visits. |
| MPK5398/ 28558 | Meikleour House Policies, Meikleour | NO 1536 3869 | Motte still exists. Dip on summit where a dovecot formerly stood. | No detailed investigation beyond field visits. |
| MPK1177/ 25807 | Tom An Tigh Mhoir, Struan | NN 8073 6536 | Motte still exists, also a ditch (tentatively associated with a bailey). Some damage to earthworks from ploughing. | No detailed investigation beyond field visits. |
| HER/ Canmore ID | Place (Site and Locality) | National Grid Reference | State of Remains | Archaeological Investigation |
| MPK5483/MPK3899/ 28875/288886 | Ardblair Castle, Blairgowrie | NO 1635 4453 | Earthwork platform, possibly from a medieval fortification, near to a largely 16th and 17th-century tower house. | No detailed investigation beyond field visits and photographic survey. |
| MPK5453/ 28488 | Castle Hill, Cargill | NO 1579 3743 | Remains of probable 12th-century earthworks and a small section of ditch. | No detailed investigation beyond field visits. |
| MPK1910/ 26588 | Green of Invermay, Forteviot | NO 0526 1621 | Cropmark enclosure excavated by SERF project to reveal a ring-work castle with timber buildings, perhaps including a tower. | Excavation (2009). |
| MPK3485/EPK266/ 28392/269281 | Perth Castle, Perth | NO 118 238 | Ditch and timber postholes discovered during excavation. | Excavation (2003). |
The date at which stone became the preferred building material for elite residences in Perth and Kinross is unclear. It is possible that the construction of mottes and timber palisades continued relatively late, while some elements of stone construction may have been included from quite early on. We should not necessarily assume that timber gave way to stone – the use of both materials may have co-existed.
There are considerable difficulties associated with dating castles in Perth and Kinross. Many of these structures were occupied for long periods of time. Indeed, over 30 castles in the area are still occupied today for residential and commercial use, while one, Castle Huntly (MPK5112), Longforgan, is a prison. Their ongoing use can reflect continuities between medieval and modern landscapes of power. However, the combination of centuries of habitation and relatively limited archaeological investigation creates significant challenges when attempting to establish construction dates. Written sources provide some clues, although before the 17th century detailed building accounts are rare, even for elite residences. At present the presumed chronology of many castles in Perth and Kinross still relies on the work of a limited number of early and mid-20th-century architectural historians, and their assumptions about the date at which certain stylistic features were introduced to the region.
Relatively few castles in Perth and Kinross seem to have significant standing remains from before 1400. Lochleven Castle (MPK8538) is one of the best-preserved, having a substantial late 14th-century stone tower house and a curtain wall with foundations from the early 1300s, although much of the rest of the wall probably dates from the late 14th or early 15th centuries. Lochleven has seen limited archaeological investigation, with small-scale excavation in the 1980s and 1990s. However, there is considerable potential for further research (Historic Environment Scotland 2005a). Substantial stone keeps, tentatively dated to the 14th century, also survive at Balthayock Castle (MPK3334), Kinfauns, and Garth Castle (MPK573), Keltneyburn. Both of these have seen major restoration, yet they may still repay detailed study.
The remains of Kinclaven Castle (MPK3575), Murthly, are more fragmentary. Written records suggest Kinclaven was built in the early 13th century and largely demolished in the late 1330s. Today the site is relatively overgrown by trees, but several sections of wall survive from what was a square enclosure castle. A survey in the 2010s showed that the masonry work is likely to date from one phase of construction (Historic Environment Scotland 2015). As Kinclaven was probably not used after the mid-14th century it provides an exceptional site for exploring the design of Scottish stone castles before and during the Wars of Independence. The relationship between the castle and other local features of importance at Kinclaven, such as the parish church and ferry, is another worthwhile topic for research.
Limited 14th-century remains might also be present at Caisteal Dubh (MPK1613), Moulin and Castle Cluggy (MPK900), Crieff. In both cases only small sections of walls still survive above ground. Caisteal Dubh was once surrounded by a loch, seemingly drained in the 1700s. Evaluation in 2005 to the south-west of the castle revealed 18th-century field drains, though little else was discovered (Clapperton 2005). Castle Cluggy is now limited to a small square tower, but appears to have once had a wider enclosure. The tower was evaluated in 2001 prior to consolidation of the remains. Inside the tower researchers uncovered what may have been the original floor, but unfortunately found no evidence to date its construction (Roy 2001).
It is likely that a number of 13th and 14th-century castles were built on or beside earlier elite residences. This appears to have been the case at Castle Hill (MPK3954), Clunie. There is written evidence that Clunie was already a place of importance by the 9th century (Alcock 1981, 161). In the 12th century Clunie is recorded as a royal residence, while there is a reference to a castle here in 1215 (O’Driscoll and Noble 2019). Much of the stonework from Castle Hill was reportedly removed in the early 1500s when the tower house on the island in Loch Clunie was built (MPK5255; Lock and Ralston 2017). However, some fragments of stonework from a large building and a possible curtain wall remain. There are also extensive earthworks, including a remarkable series of terraces on the hill on which the castle stood. Nearby is the site of an early church and an open area which may relate to further medieval settlement. Clunie appears to be a site of considerable significance both in relation to the structures surrounding a castle, and the evolution of an elite residence across the medieval period. Geophysical survey and some small-scale excavations took place in 2018 and 2019, but there is potential for much more extensive investigation (O’Driscoll and Noble 2019).
Perth and Kinross has a large number of 15th and 16th-century castles, often relatively well-preserved. Outstanding examples include Balvaird Castle (MPK3221), Glenfarg, Elcho Castle (MPK3313), Rhynd, and the two linked tower houses at Huntingtower Castle (MPK5534), Ruthvenfield. However, many other castles of this period survive, often still in use as family homes. Several 15th and 16th-century castles from Perth and Kinross have elaborate carved stone detailing, providing insights into the visual language employed by the Scottish elites. Historic Environment Scotland has undertaken some research into the carved stones at properties in their care. However, more study in this area could be undertaken. A number of elaborate carved stone panels and lintels have also been removed from 16th-century castles and laird’s houses and reused in other buildings in Perth and Kinross. Decorated architectural fragments, most probably from nearby castles or other elite residences, can be found built into later structures at Castlehill (MPK4820), Longforgan; Kilspindie Farmhouse and Manse (MPK5361), Kilspindie; Pitkindie Castle (MPK4828), Abernyte; and Pitroddie Farm (MPK6700), Kilspindie. Many of these carved stone fragments have received limited study or recording.
Some examples of interior decoration also survive from this period. At Huntingtower Castle there is an impressive painted timber ceiling, probably from around 1540. Elements of 16th-century interior decoration may survive at other properties, possibly concealed by later decorative schemes. Invaluable information on age and construction methods can also be obtained from historic timbers preserved within these buildings. It is important that recording and conservation efforts at elite residences include both exterior and interior features.
Towards the end of the Middle Ages the always rather mutable distinction between a castle and a grand house seems to have become even more blurred. It is debatable to what extent many of the castles built in Perth and Kinross in the 15th and 16th centuries were intended as serious fortifications. Scholars such as Charles McKean have argued that the majority of Scottish castles at this time were essentially country houses with symbolic martial decorations (McKean 2001). Certainly, most 16th-century castles in Perth and Kinross were not designed to resist heavy artillery. Instead, the fashionable form appears to have been variants of the tower house, often with an associated walled enclosure or barmkin. Given the current state of academic debate, it seems sensible to consider both aesthetic and defensive explanations when examining the architecture of castles built at this time.
Castles and tower houses have traditionally dominated our understanding of medieval elite residences. Yet other forms of aristocratic dwellings clearly existed in Perth and Kinross. At the start of the 15th century the bishop’s palace at Dunkeld (MPK2453) consisted of ‘great houses built upon the ground’ (Oram 2005). Perhaps significantly, Bishop Robert de Cardeny replaced this low-level collection of buildings with a now demolished, tower house – arguably a sign of changing fashions in aristocratic housing. The extraordinary complex of medieval buildings at Stobhall (MPK3714), Guildtown also appears to be an example of a largely unfortified noble residence which subsequently acquired a tower house. If likely locations of medieval halls, or other forms of unfortified elite dwellings, are identified they should be a priority for archaeological investigation.
Castles and other high-status residences must not be viewed in isolation. They typically had a number of ancillary buildings and were often placed within carefully planned landscapes. The ancillary buildings associated with medieval elite residences have frequently been demolished, even where core structures such as stone tower houses have been preserved. Yet they are fundamental to understanding the appearance and function of a site. Excavation in the 1980s led to a more detailed understanding of the ancillary buildings in the inner courtyard at Balvaird Castle (MPK3221), although further research could be undertaken on the outer courtyard and former walled gardens (Historic Environment Scotland 2013; Lewis 1992).
The designed landscapes surrounding medieval castles served practical purposes and formed a key part of how elite residences were publicly presented (Creighton 2009). Recent research by Brown has significantly advanced our understanding of medieval Scottish gardens (Brown 2012). However, there is potential for far more survey work and excavation of the remains of medieval designed landscapes in Perth and Kinross. It should perhaps be noted that medieval gardens were sometimes located a little distance from elite residences – cropmarks from a possible walled garden near Rossie Priory (MPK4640) may be an example of this phenomenon (Brown 2012, 61–2).