Brochs, duns and forts in early Celtic Scotland

by Angus Konstam

The background
When the Romans came to Scotland in AD 80, they knew little or nothing about the Celtic tribes who lived in the region. Faced with invasion, the tribesmen of the lowlands either submitted to Roman occupation or withdrew to what they thought was the safety of their hilltop forts. This proved a costly mistake, as the two largest forts in the area fell to the might of the Roman army and its siege artillery. The tribes who resisted (known by the Romans as the Selgovae and the Novantae) were brought to their knees by the end of the year, and the Roman Governor Agricola consolidated his northern frontier along the line of the Forth and Clyde rivers. The area was completely pacified by the start of AD 82. So much for the defensive protection of the lowland hillforts. Further north a fresh challenge awaited Agricola, as he planned to lead his armies into eastern and north-eastern Scotland, beyond the Firth of Forth. These Celtic tribesmen had used their fortified bases in what is now Stirlingshire to harry the Romans, and Agricola had had enough. In AD 83 he launched his legions in an expedition of conquest, cornering the local ‘Caledonian’ tribesmen in battle at Mons Graupius (84) and inflicting a decisive defeat on his opponents. During the advance his flanks were secured by a series of Roman auxiliary forts designed to prevent Caledonian movement out of the Highlands. His fleet sailed north as far as Orkney, forcing the submission of the coastal communities they encountered.


Although the Roman tide receded due to commitments elsewhere, the threat of punitive attacks against the Caledonian tribes continued, forcing the local Celts to maintain strong defensive positions and to ensure their near-constant readiness for war.


The Roman defensive line along the Forth–Clyde line was abandoned around AD 100, and the frontier was re-established between the River Tyne and the Solway Firth, a position which was defended during the reign of the Emperor Hadrian (AD 117–138). The Romans returned north for a time during the reign of the Emperor Antonius Pius (AD 138–161), and the Antonine Wall was built along the old Forth–Clyde line, before it too was abandoned after the death of the Emperor. From that point on Hadrian’s Wall marked the northernmost frontier of the Roman empire. Although the tribes immediately to the north of the wall were relatively peaceful, those further north were more hostile. At the start of the 3rd century AD the Emperor Septimus Severus (AD 193–211) led punitive expeditions against the Caledonians, as did the Emperor Constantius I Chlorus (AD 305–306) a century later. It was during this last expedition that we first hear of the Caledonians being referred to as the ‘Picts’, or painted people. Historians generally take this date as the mark that divides the era of the Picts from that of their Caledonian forebears, and provides a convenient finishing point for our study.


In early Celtic Scotland, there were three main types of fortifications in use during this period: the brochs, the duns and the hillforts.


Towers in the north: the brochs
The Broch of Gurness stands on the shore of a stunningly beautiful bay and sound in Orkney. It was built at some point between 500 and 200 BC, and the broch itself formed part of a defensive site that included a village and a series of encircling ramparts and ditches. The brochs of Iron Age Scotland were a virtually unique solution to the defensive requirements of their builders. Spectacular even in ruin, these structures often combined the functions of a defensive retreat with that of a communal focal point. They protected the local people from petty bandits, raiding war parties and on occasion, from full-scale invasions. As such they often formed the nucleus of small communities, or were located close to existing settlements. This means that any true study of them as fortifications needs to be combined with a look at the communities they served, and the people who built them. From there we can look at the fortifications which succeeded them, and which provided defensive strongpoints for the Picts, who inherited the land from the Iron Age broch-builders.


A broch was an imposing circular fortification built using drystone walling. This meant that no mortar was used, but the irregularly shaped stones were chosen so that they fitted roughly together. They were tall, grim, windowless structures, containing a passage within the walls which eventually led to an upper rampart. The only entrance was a small, easily-defensible doorway at ground level. Two walls were separated by passageways, stairs and galleries, which eventually led up through the walls to the circular upper parapet, where the defenders could rain missiles down on the heads of their attackers. While the Broch of Gurness is considered an early example of the genre, the Broch of Mousa in Shetland is probably the most intact example of a later (and more classical) broch structure. Precursors of the earliest brochs were probably the strong circular houses whose ruins are located in the same geographical area as the brochs.


Almost all brochs are located in the north and west of Scotland in Caithness, Orkney, Shetland and Skye, while a few others were built further south. The majority of them are concentrated in Orkney, Shetland and Caithness. Dating evidence suggests that most were built between the start of the 1st century BC, and the late 1st century AD, although this has been questioned due to the equivocal dating material so far discovered. Further evidence suggests that ‘proto-brochs’ or precursors to the classical broch structure could have been built as early as the 6th century BC, while we know that some remained in use until at least the early 3rd century AD, if not later. Although we know a lot about the structures themselves, and we can analyse their defensive qualities, we know very little about who exactly built them, and why. Obviously numerous theories have been proposed, and it was only recently that archaeologists reached a general consensus on what may have happened.


Clearly, they were designed for defence. The Broch of Mousa alone stands to a height of some 13 metres (40 feet), and would have been a proof against all but the most determined assault, unless the attacker had Roman-style siege artillery. The low narrow entranceway made it difficult to batter down the door, and the walls of Mousa were too high for ladders. The hollow interior was probably roofed over, and was large enough to house livestock, provisions and people until the threat had passed. We know little about who the threatening attackers might be, but Celtic, Roman or German raiding parties might have come to these areas in search of slaves. Although not impregnable, smaller brochs would have guaranteed that an attack against them would have been costly, and thus they acted as a form of deterrent against any potential aggressor.


Until comparatively recently, brochs were sometimes referred to as ‘Pictish towers’, or even associated with the Norsemen (vikings). While these links have been disproved, the terms indicate a general lack of understanding of the brochs and the broch builders. We know a certain amount about the late prehistoric people who lived in what is now Scotland from their archaeological legacy. They were not Scots, as that political entity post-dated the broch builders by a millennium, but we have no alternative name to identify them by, as no written records survive from this culture and period. The term ‘Celtic’ has been widely used to describe all the iron age people of this period who inhabited most of Europe, including Scotland, but some archaeologists baulk at using such a widely applied appellation. As for the term ‘Pictish’, their time came later, and the Picts have usually been identified with the inhabitants of north-east and east central Scotland from the early 4th century, when the name first appeared in Roman written records. The broch builders had been long gone by then, and while the Picts may well have been the descendants of these broch builders, archaeological information is unable to prove a clear descent from one group to the other. Various theories have been proposed, including ones where the Picts reached Scotland from overseas, and similarly that the broch people were somehow different from the pre-Celtic people who inhabited the rest of Scotland.


It is probably true that the pre-Celtic people of Scotland intermingled with later waves of Celtic migrants, but there is no direct Celtic broch building tradition. It has been suggested that while the rest of Scotland was overrun by the Celts, the broch builders retained their independence, and fortified their settlements. Whoever built them, their appearance coincided with the arrival of the Celts, and their disuse began following the arrival of the Romans in Scotland. Some archaeologists have given the broch builders the clumsy appellation of proto-Picts, but this does the earlier people a disservice. The broch builders displayed certain qualities which were absent elsewhere in the Pictish homeland (which included Orkney and Shetland), so although there are many theories, there are few answers to the mystery of who these enigmatic people were. It is possible that by the time the Pictish era, the local population had effectively become as Celtic as the rest of Scotland. Certainly we know that most brochs were abandoned at some point during the 3rd century AD, which is close enough to the appearance of the Picts as a distinctive people to suggest some link between the two dates.


Blockhouses in the west: the duns
The term ‘dun’ is used to identify a particular type of small fort which was built extensively all across south-western and western Scotland, with the greatest concentration found in Argyll. These circular or oval dry-stone structures were similar to brochs, but were much smaller. While some were built on flat ground, most were constructed on rocky outcrops or natural defensive positions to enhance their defensive properties. Their walls were usually built using two thick dry-stone walls, with a solid core of rubble used as infill between them. Some used timber to lace the structures together (as was the case with the first hillforts), but most had a smooth outer face, devoid of timber reinforcements. In some examples the wall was reinforced at the base to allow the construction of higher or heavier structures. Like the brochs, the entrance was small and protected by chambers to discourage battering attempts. A particularly impressive example (the Dun of Leccamore, on Luig) even boasts an internal stairway, and other design features suggest some form of correlation between the broch builders and the defensive properties of these smaller dun structures.


While some earlier timber-laced duns have been dated to the 6th or 5th century BC, the majority appear to have been built during the period after the Romans came to Scotland, during the 2nd and 3rd centuries AD. Some show evidence of occupation, abandonment and re-occupation, suggesting they were used when the situation warranted it, and in more peaceful times they may have been abandoned for more spacious and convenient settlements nearby. They also show signs of a far longer occupation than the brochs to the north or the hillforts to the south and east. Dun Cuier on Barra was occupied until around AD 500, while Kildalloig in Argyll appears to have remained in use as late as the 8th century. Unlike the brochs or hillforts, most duns appear to have been little more than fortified homesteads or farms, but they remained a feature of the Scottish landscape for over a thousand years, and outlived both other forms of early Celtic fortification.


Strongholds in the south: the hillforts
Nobody knows how or exactly when the Celts reached Scotland. Towards the end of the Bronze Age (around 700 BC), these newcomers began to arrive, bringing the new technology of the Iron Age with them. These Celts also introduced a new feature to the Scottish landscape. Over the next eight centuries, hillforts would appear in various sizes, from small fortified farms to full-scale fortified hilltop townships. They provided refuge for the local Celtic communities who faced attacks and raids from their neighbours. While well-designed to protect the Celtic tribespeople from their own kind, they proved less effective against the Romans.


Although the early Bronze Age defensive ring at Meldon Bridge in Lothian is probably the earliest fortified site in Scotland, the first hilltop fortifications appeared around 600 BC or slightly earlier. These took the form of timber-laced fortified circles. In some cases the ramparts were fire damaged during their period of use, which allowed the sites to be carbon dated. While the dating range is wide most appear to have been actively built or expanded during the 6th century BC or later. These timber-laced structures continued to be built in Scotland until the coming of the Romans in the late 1st century AD, although the style of the fortifications became more elaborate with time. Timber-lacing was a technique used to stabilise both earthen ramparts, stone walls or rubble infill by laying horizontal wooden beams across the structure, binding it together. In other words, the timber provided a massive framework which was filled with stones and rubble, then faced with solid stone. A wooden walkway and palisade were then built on top of this defensive perimeter. Thick wooden gates protected the entrances to these hilltop enclosures.


Surviving examples such as the stone and earthen wall of the fortification at Abernethy in Perthshire (occupied during the 1st century BC) show the surviving slots in the walls where these beams were placed, and had rotted away. In cases where the forts were destroyed by fire (probably during an assault), fire damage caused by the burning timber has left its mark on the surviving stonework, which has sometimes been fused together. In rare instances, the remains of timber-lacing survives, such as at Kaimes Hill in Midlothian. Timber-laced forts were built throughout central and eastern Scotland and around the Moray Firth to the north, and this distribution matches that of early Celtic finds such as axe heads dating from the 7th century BC and later. This proves that the early Celtic people who occupied central and eastern Scotland relied on these types of fortifications for their protection.


The nature of these hillforts changed over time. In some cases, the original timber-laced structures were replaced or rebuilt in later periods. At Kaimes Hill a series of stone-faced ramparts replaced these earlier defences, and a series of ditches were dug around the perimeter to strengthen the position. One additional refinement was the setting of a ring of pointed stones around the outside of the wall, creating a disruptive obstacle which would hinder any attackers. The trouble with timber-lacing was that the timbers were difficult to replace once they rotted, or they could be destroyed by fire with relative ease. Archaeological evidence suggests that while timber-lacing continued to be used in Scotland during the early Celtic period, the weakness of the design was apparent to the builders. Consequently when the local Celtic tribes of lowland Scotland were faced with the prospect of Roman invasion in the late 1st century AD, many forts were strengthened and improved by the addition of all-stone walls and the digging of ditches outside the walls. In addition to their walls or earthen ramparts, most of these defensive positions were topped by wooden palisades.


These forts were almost exclusively built on hilltops to impove their defensive capabilities, and in many cases the walls enclosed some form of interior settlement. Of some 1,500 fortified sites in Scotland, the majority of these forts were located in lowland Scotland, below the Forth–Clyde line. This surprisingly high figure includes small fortified farmsteads and isolated stone structures from the same early Celtic period. As some of these were built over 700 years before the Romans appeared, it comes as no surprise that many had been abandoned for centuries by the 1st century AD, although a handful remained in continuous use throughout their history. Unlike the sprawling hillforts such as Maiden Castle in England, these Scottish fortifications were small, and probably only served small local communities. The two exceptions were Traprain Law and Eildon Hill, both of which were substantial defensive positions, and the latter containing over 300 roundhouses. This meant that in times of danger, an entire tribe could seek refuge within its walls. One of the problems with Scottish hillforts is the lack of available information about their history. We rarely know how long they were occupied or when, and what function they served apart from a defensive one. It does seem that at least at certain periods, the hillforts which enclosed settlements tended to be under continuous occupation when the Romans appeared.


Another variant of the hillfort was the promontory fort, which was found at various points along the east coast of Scotland, such as St. Abb’s Head, Dunnotar and Urquhart (the last was actually on the shores of Loch Ness, not the North Sea). All but the last were most probably established as fortified sites well before 300 AD, but all three were developed into major fortifications during the Pictish period, and the last two were actual Pictish fortifications. Similarly, the promontory at Burghead on the Moray Firth was developed as a Pictish stronghold. In all three sites, elements of the old hilltop fort designs were used, as the headland was cut off from the mainland by a series of defensive walls and ditches. Again, the Burghead fortification may have predated the start of the historic Pictish period, but the lack of hard dating evidence makes it impossible to say so with any certainty. Certainly the system of three lines of earth and rubble defences and intervening ditches is similar to that found in hilltop forts from 300 BC on, and we know the Picts added an inner citadel to the fortified point at Burghead. The spot was also a good anchorage, and it has been suggested that Burghead was used as a Pictish base from which maritime raids were launched down the coast into Roman Britain. Certainly there seems to be a legacy of construction methods which linked the known Pictish fortifications (Inverness, Dunadd, Dundurn, Dunottar, Dunkeld, Clunie, Scone, Inveralmond and Forteviot) to the earlier hilltop forts in the same area (Tayside, Moray and Grampian).


Summary
To sum up, although the Iron Age landscape of Scotland is littered with fortifications, these can be divided into three groups. The brochs of the northern and western isles are virtually unique, and their design displays a high level of architectural and military appreciation. To the south-west the duns were smaller counterparts, and less likely to be situated in coastal locations. These remained in use until well after the arrival of the Scots from Ireland, and outlasted all but a handful of coastal fortifications which were probably used by both the Celts and their Pictish descendants in eastern Scotland. As for the rash of hillforts in southern Scotland, the majority fell from use following the Roman invasion of the late 1st century AD. Despite this, their methods of construction were adapted for use by the Picts as well as the Scottish peoples who inhabited the southern lowlands when the Romans withdrew. Scotland is unique in that so many of its monuments are still extant, and have been saved from centuries of development. Although the region produced methods of early Celtic fortification that were unique, any study of these defensive sites helps us understand both the people who built them, and their Pictish or Scottish descendants.


Further Reading
Armit, Ian, Celtic Scotland, Historic Scotland Publication – Batsford Press, London, 1997
Breeze, David J., Roman Scotland, Historic Scotland Publication – Batsford Press, 1996
Ritchie, Anna, and Breeze, David, Invaders of Scotland, Historic Scotland Publication – HMSO, 1990
Ritchie, Graham and Anna, Scotland: Archaeology and Early History, Edinburgh University Press, 1991
Ritchie. J. N. G., Brochs of Scotland, Shire Publications, 1988
Sutherland, Elizabeth, In Search of the Picts, Constable and Co., 1994
Wagner, Paul, Warrior 50: PictishWarrior AD 297–849, Osprey, 2002

broch plus

Outline plans of five Iron Age sites in Scotland and one (West Brandon) in northeast England. All illustrate the importance of the round house in its different forms. Dun Troddan is a broch, and the defences at Clickhimin enclose another structure of this type. Kilphedir is a wheelhouse. Glenachan, West Brandon, and Hayhope Knowe are bounded by single or double palisades. At Hayhope Knoll there is also an earthwork enclosure. Information from D. Harding (2004).

A form of defended homestead peculiar to Scotland is the broch. Nearly 600 are known, almost all of them lying beyond the Highland Line, with over a hundred in Caithness. A broch is a tower, of drystone construction, the walls about 15 feet thick and the interior area anything from 30 to 40 feet in diameter. Few are more than 20 feet high, although the one at Dun Telve, Inverness-shire, is, even now, 33 feet high, and this in spite of its having been robbed over the centuries for building stone, a fate which has overtaken many of the others. The most famous, the largest and the best-preserved broch is that at Mousa, in the Shetlands, and this is 40 feet high. There was but a single door into a broch, and a staircase and other chambers were constructed in the thickness of the walls. They are often difficult to date individually, but in general terms they were being built in the last centuries BC and occupation in some cases, as at Dun Cuier, on Barra, continued on and off into the seventh century AD.

An aerial view of the broch on Mousa, in the Shetlands. The ruin in the rectangular enclosure is a house, the old Haa.

Round dry-stone defensive structure of the 1st millennium BC, examples of which are concentrated in Shetland, Orkney, Caithness and the Western Isles in Scotland. Traditionally the term `broch' has been restricted to the more elaborate circular examples of a diverse family of compact fortified buildings erected in the Scottish Iron Age; the famous broch of MOUSA forms, in effect, a type site. Mackie has attempted to regularize the term by selecting the technically sophisticated high hollow-built wall, which often contains chambers and a staircase to an upper floor, as a principal defining feature (Mackie 1965). However, Hedges (Hedges and Bell 1980) and others have argued that simpler round fortified structures such as the early (c. 600 BC) Bu Broch, Stomness, are also true brochs. A solution to this, partly semantic, problem is provided by Armit who argues that Mackie's elaborate brochs are best classified as `tower brochs' within a wider category of round houses, thus freeing the term `broch' to be used more loosely (Armit 1990). 

Traditionally, brochs have been consigned to a relatively brief interval after the 1st century BC. However, modern excavations have made this late dating seem unlikely, and even the developed `tower brochs' may have been emerging a century or so earlier. Recent explanations of the origins of the broch (e. g. Armit 1990) have discarded the idea of migrants from the south, preferring to stress local prototypes and the broch's functional and symbolic importance in the negotiation of power.

#
 

In the early centuries of the first millennium, native society was undergoing profound political and social change. We see in the heightened development of social inequality and hierarchy the emergence of what anthropologists might term a `chiefdom society'. The appearance of souterrains (large underground stores) in the eastern mainland is but one reflection of ongoing attempts to maximize the extraction of wealth from the land and concentrate it in the hands of the few. All over Scotland, small-scale power structures founded on face-to-face relations were being superseded by far-reaching systems of control, distant authority delegated to local leaders in return for a share of the tribute. The rise and fall of the famous brochs are an architectural manifestation of the beginning of this trend away from the intensive and towards the extensive exercise of power, as hierarchies of space within a settlement (internally differentiated sites of similar form throughout the landscape) were replaced by new hierarchies between settlements (major centres controlling dependent sites). Political units, however, remained comparatively small. Identity was vested at the level of the tribe whose members might have numbered only a few thousand.

#

The tendency to build self-contained circular enclosures reached its apogee in western and northern Scotland, and especially in the Hebrides, Orkney, and Shetland (Fig. 5.12). Unfortunately some of the strongest patterning has been obscured by disagreements about terminology and chronology (Armit ed. 1990; Armit 1992; Parker Pearson and Sharples 1996: chapter 12; D. Harding 2004: chapter 5).

Again the circular archetype was very important and extended from individual dwellings to more monumental walled enclosures. All these features were conceived on an impressive scale. They vary from the crannogs built in open water to small circular compounds, and from relatively insubstantial dwellings to massive domestic buildings, the most impressive of which - the brochs of the Scottish mainland, the Western and Northern Isles - resemble towers (Armit 2003b). Some of these structures are isolated but quite densely distributed and were surely designed to impress, whilst others can be found inside defended enclosures which contain a variety of other buildings. Many of them were distributed along the coast where there were a number of promontory forts (Armit 1992).

Here is another case in which large round houses may have been an important settlement form from an early stage of the Iron Age. Ian Armit (2003b) has argued that structures ancestral to brochs were built as early as 600 BC and that during the Iron Age stone houses in Atlantic Scotland became increasingly complex in design. True brochs seem to have emerged about 200 BC. They are interpreted as defended high-status dwellings, characterised by such features as internal staircases and guard cells. They must have had more than one storey, and there can no longer be any doubt that they were roofed. 

Some of the structural principles that characterise the brochs extend to other forms of defensive architecture: to some of the circular-walled enclosures known as duns and even to the monumental gateways of a number of the promontory forts in Shetland (Hamilton 1968; D. Harding 2004: 137-50). Another key element is the way in which domestic structures were organised. In some brochs there seems to have been a communal space with a hearth located in the centre of the building. It was ringed by a range of compartments which were divided from one another by partitions projecting from the interior wall. In Shetland and the Hebrides, this principle was expressed on a smaller scale in the distinctive dwellings known as wheelhouses (Parker Pearson and Sharples 1999: chapter 12). A number of these were built after the broch themselves, and it has yet to be worked out how far their histories overlapped. In any case there are important contrasts between them. 

Wheelhouses were sometimes set into the ground, whereas brochs were conspicuous monuments, and on certain sites the domestic accommodation was probably at first floor level (Armit 2003b: chapter 3). Wheelhouses were sometimes associated with souterrains, but the connection between storage structures and individual dwellings is entirely different from the more centralised system illustrated by hillforts in southern Britain.

BROCH AT BURROUGHSTON, SHAPINSAY, ORKNEY

Situated on the NE coast of Shapinsay, the walls of the Iron Age broch at Burroughston still stand above first-floor level in places. Its entrance and adjacent 'guard' cell are similarly well preserved, while a wall, ditch and rampart, which probably once encircled the broch, are evident around the structure. These extra defences may have been constructed because of the broch's low-lying position in the landscape.

A broch, the base of which is fairly complete, stands close to the sea, some 32 yds from the low rocky beach. The site was excavated by Petrie and planned by Dryden about 1862 and the plan incorporates some features taken from Dryden's plan which are no longer to be seen.

Except on the E towards the sea, the broch is surrounded by a rampart rising 6'- 7' above the bottom of a ditch nearly 20' wide. The ditch is bounded on its outer edge by a parapet, parts of which still stand from 2'6" to 3' high and Dryden found traces of a stone wall on the inner side of the rampart some 9' from the broch wall. The broch has an average diameter of 33'6" within a wall 10' - 14' thick. The inner wall-face, showing a scarcement, stands approximately 12' high but the outer face was not laid bare in the 1862 excavations. The roofless entrance passage in the E has door-checks, bar-hole and guard chamber and the interior contains a 10' deep well and a later work now hidden by debris, but traces of radiating walls and compartments can still be seen. In the sloping area in front of the entrance the 1862 excavations revealed traces of "out-buildings" connected to the broch by an extension of the entrance seawards. These structures have now become covered with turf.


A broch with outworks . The outer face has been exposed in two or three places giving a wall thickness of 3.7m except in the N where it is only 3.0m but here the coarser masonry suggests a later reconstruction. The outworks probably once completely encircled the broch, but have been destroyed in the E by the later buildings. The name, though still known, is no longer commonly applied to the broch.


About 150m ENE from Easthouse, on the NE shore of Shapinsay, stands one of the most visually interesting brochs of Orkney. It was excavated c.1862 by Colonel D Balfour's estate-workers under the direction of George Petrie, and was subsequently meticulously recorded by Sir Henry Dryden. Excavation was confined to the interior of the broch and to a small area immediately outside the entrance; the interior is now choked with rubble which obscures the architectural features below the level of the scarcement. The entrance-passage is well preserved, and there is an impressive cell within a length of walling that still stands to a height of 3.5m above the rubble. Grazing has recently ceased on the site, which has thus become overgrown; the wet outer ditch and bank are still clearly visible.

THE BROCHS

To the north of the dun territory, in Orkney and Shetland, Caithness and Skye, appeared a series of tall, tapering, stone towers, clearly designed as fortifications and known as brochs. As Euan MacKie has observed, these were the only advanced architectural buildings ever to be created entirely within prehistoric Britain, apart from Stonehenge. About 500 of them are known in northern Scotland where some of them once stood 9 m. (10 yd) or more in height and over 20 m. (22 yd) in diameter (plate 51). Each circular tower consists of an inner and outer layer of dry-stone walling tied together with a series of horizontal lintels which bridge the gap between them. In this space, which is about 1 m. (1 yd) wide, are a series of galleries superimposed one above the other and a slab-built staircase which climbs clock-wise to the top of the tower. The galleries were probably intended only to lighten the weight of the tower wall, thus allowing it to be built higher. The total thickness of the wall at its base was about 4.5 m. (5 yd).

The outer wall of the broch has no windows and is unbroken, except for an entrance, and rises upwards for two storeys or more, with a gentle, inward batter, like an electricity cooling-tower. The inner wall face is sometimes broken by vertical openings called voids which allowed light and air into the staircase and relieved some of the weight of the walls, especially above the entrance. The long entrance passages needed to be strongly built to withstand attack. They averaged 1.5 m. (1.6 yd) high and 0.75 m. (0.8 yd) wide. About half-way along the passage were stone door jambs and a slot into which a bar could be placed to secure a heavy wooden door, to which was attached a strong iron ring handle, like the one found at Dun Ardtreck (Skye). (Confusingly duns and brochs are both called duns or duins in Gaelic.) Behind the door was a corbelled guard chamber in the thickness of the wall.

Around the inside of the broch ran a ledge or scarcement, varying between 1.5 m. and 3.5 m. (1.6 yd–3.8 yd) above the floor, which seems to have supported a wooden gallery which jutted forward on to a ring of posts. Higher up, a second scarcement probably carried a roof over the gallery. It seems likely that the central part of the tower was open to the sky which must have provided problems in bad weather for the central cooking hearths that were often present, though storage tanks in the same area could have been filled with rain water from the gallery roof. Drinking water was often provided by a well.

Some brochs have outer defensive walls or ditches added, like the bailey of a Norman castle, which contain small villages of domestic buildings. At the Broch of Gurness (Orkney) there were sufficient houses for 30 to 40 families enclosed within three stone-faced ramparts and ditches. Enclosed buildings at the broch of Howe (Orkney) would have accommodated about 250 people.

Although there is considerable uniformity in broch design, on one point there is a major variation between those in the Hebrides and those in north Scotland and the Northern Isles. This shows in the ground plans which are usually ground-galleried in the Western Isles and normally solid-based in the rest of Scotland. As the names imply the ground-galleried brochs have cells and galleries starting at ground level, whilst the solid-based examples have an almost continuous solid wall of masonry at ground level, broken only by the entrance passage and two or three corbelled cells.

The classic broch is Mousa on Shetland, standing on a rocky headland guarding Mousa Sound (fig. 61). The tower is almost complete, standing to a height of 13 m. (14.2 yd). Although its external diameter is 15 m. (16.4 yd) at the base, the thick walls leave room for an interior only 5.5 m. (6 yd) across. The entrance passage leads directly into the interior with no guard chambers. Inside are three large cells with tiny wall-cupboards all set in the thickness of the wall. The continuous staircase passes up through five floor levels or galleries in the wall before reaching the walk-way at the top of the tower. The galleries were lit by three voids. Two ledges or scarcements projecting from the inner wall face at heights of 2.1 m. (2.3 yd) and 3.7 m. (4 yd) above the floor of the broch probably represent the floor and roof height of a wooden gallery, which could be reached from an entrance on the stairway. The interior of Mousa is now partially blocked by a wheelhouse- type structure of perhaps the third or fourth centuries AD, though recent work in Orkney may suggest that this is contemporary with the main broch occupation.

Many brochs, though not all, stand on the seashore, and this has led to speculation that they were erected against seaborne attack by people united by maritime interests. It is likely that over many centuries they guarded the small tracts of good farming land that were dispersed in an inhospitable terrain, hemmed in by mountains and moorland. It is clear that in the closing centuries BC life in northern Britain was ever more turbulent and attack was always a possibility. The loss of livestock or personal belongings or capture for a southern slave market must have been a constant fear. The brochs probably acted as temporary refuges into which the local community could retreat to withstand a short attack or siege. It would be difficult for an enemy to overcome the great height of the broch tower which would prove impregnable to spear or slingstone. Scaling would be incredibly dangerous and battering of the door difficult due to the narrowness of the passage and the presence of guard chambers and armed guards. The excavation of Dun Mor Vaul (Tiree) showed that although the broch was in use for about 250 years, occupation was only sporadic and for short periods.

There have been years of debate regarding the origin of the brochs. It now seems likely that they developed in the Orkneys as prestigious fortified towers. Recent study has begun to show various round houses that could be claimed as ancestral to them there. That they can have originated in the mid-first millennium BC is shown by a radiocarbon date around 600 BC that has been obtained for the demolished broch site at Bu, near Stromness (Orkney). To the south-west, in the Hebrides, Dun Ardtreck on Skye has given a radiocarbon date of 115 BC. As has already been suggested, the great similarity between the brochs reflects a certain unity amongst the people of north-western Scotland, which may have led to a specialist group of local architects responsible for their design, who travelled from one community to another and passed on their knowledge from one generation to another. It is clear, however, that not all of them were built with the same degree of sophistication.