assemblage 11 (2011) 47-58

Shedding Light on the Matter: An Exploration into the Regional Orientation Patterns of the Brochs and Duns of Iron Age

by THOMAS CROWTHER

The primary aim of the paper is to present, examine and interpret the full range of the available doorway orientation patterns of the Iron Age and dun structures that are scattered throughout the various regions and landscapes of Scotland. Differing orientation patterns within these regions are illustrated through distinct regional conformities to certain cardinal points of the 16 point compass and offer the potential for variation. Sectioning Iron Age Scotland into three distinct regions (The Northern Mainland and Isles; The Western Isles and Skye; and Argyll and the Inner Isles), the author attempts to combat the concept that an E/SE orientation majority existed throughout Iron Age Britain by presenting regional variation with regards to orientation conformity. The differences across Scotland also present the possibility that differing social models existed throughout Iron Age Scotland, something that is also demonstrated by the variations in architectural complexity.

Key words: Broch, Dun, Iron Age, Scotland, Orientation, Lightscapes.

Introduction large extent, their geographical location, it is their considerable variation in building style The heterogeneous character of the Iron Age and material culture which has created a web monuments scattered across the dramatic of tangled interpretations and inspired a long coastal landscapes of Atlantic Scotland have and complex source of debate and long inspired intense debate within disagreement. The unique simplicity of their archaeological circles. In particular, the monumental expression and their massive possible purpose and role of the broch and dun proportions in comparison to the smaller dun monuments has been discussed in great detail. structures found most numerously in the west In order to contribute to those discussions, the has stimulated this debate, and as such, the author has noted the available entrance broch has long dominated the scholarship of orientations from all the brochs and duns Atlantic Scottish Iron Age studies. located throughout Scotland and examined them in depth – an area of research not The term ‘broch’ is only applied to those previously fully extended this far north. The structures that show key architectural features examination of regional orientation patterns traditionally defined by MacKie (1965) as a presents the possibility of providing potential complex and developed architectural form variations, illustrated through distinct regional using the technique of hollow-wall building, in patterning and conformities to certain cardinal combination with a range of specific structural points. Rather than use the term Atlantic devices, such as intra-mural galleries and cells, , the author has chosen to use the to create a stable drystone tower while terms broch and dun in order to maintain a retaining the basic roundhouse form. As one of sense of regional disparity. The Scottish Iron the more impressive innovations of the Age is a period particularly lacking in regional Scottish Iron Age, these circular, drystone studies and thereby requires a range of new thick- and hollow-walled towers are found and original models tailored to the particular most numerously in the Northern Mainland needs of its Iron Age record. To introduce the and Isles, with fewer numbers found in the topic for discussion, a brief overview of the two Western Isles and Skye. Specific brochs such as different types of site in question is necessary. Howe, Gurness and Midhowe on and Crosskirk, Nybster and Keiss Road in , possess external buildings, perhaps Introducing the Brochs used as secondary villages with the main central tower acting as a village foci. This is a Categorised separately from the other Iron Age feature unique to northern Scotland and is not structures of Scotland (e.g. the dun and the seen in any other region. There are of course wheelhouse) by their size, architecture and to a many other well-known examples of broch

Thomas Crowther - [email protected] © Crowther 2011 © assemblage 2011 CROWTHER

structures to be found outside the Northern The structural complexity of the broch is Isles (e.g. , Lewis, therefore missing in the simple drystone walled and in Inverness-shire), however enclosure that is the dun, noted previously to be these are isolated single structure broch a permanently occupied enclosed settlement settlements with fully developed towers rather inhabited by some form of large social grouping than possessing contemporary or indeed (Alcock and Alcock 1987; Barrett 1981; Nieke secondary broch villages. 1983; 1984; 1990: 135). Although portrayed as simple, they are however a very diverse class of monument, possessing a variety of ground plans which include the possible roofed ‘dun houses’ (Harding 1984) perhaps more akin to the broch, as well as the much larger ‘dun enclosures’ (Harding 1984) which were almost certainly unroofed and were more like the much later Irish ring forts (Henderson 2000: 123; 2007: 166). Indeed, many duns actually have the same characteristic architectural features as brochs – galleries, cells and stairs – but are regarded as duns to retain a regional differentiation which is appropriate to uphold for the purposes of this paper. A small proportion of the few duns that have been excavated have produced Early Historic material (imported pottery, metalwork) (cf. Alcock and Alcock 1987; Neike 1984, 1990) and this has led to the assumption that the dun is a first millennium AD phenomenon. However, until radiocarbon dates are obtained, we should not apply this date to all drystone dun structures throughout the differing regions of Scotland, including both the smaller Figure 1 Broch and dun distribution roundhouse and the enclosure types. across Scotland illustrating the three distinct regions discussed in this paper. © Crown Copyright Ordinance Survey. A Wide Focus: Previous An Edina Digimap/JISC Supplied Service. Orientation Studies

The recent focus of scholarly research into Introducing the Duns regional studies has thus far revealed three distinct cultural zones within Atlantic Throughout Argyll, the Inner and Western Scotland: the Northern Mainland and Isles, Isles and elsewhere across Scotland, there are Argyll and the Inner Isles, and the Western hundreds of drystone sites that simply do not Isles and Skye (Henderson 2007: 150), and in have the full range of architectural attributes this paper, the author will focus upon these that would qualify them to be classed as three distinct zones (Fig.1). Each region brochs. Too small to be considered as forts, presents significant differences in the material they are classed under the very loose term culture and architecture of the brochs and ‘dun’, regardless of their shape or unique other Iron Age structures and even within features. The Argyll inventories of the Royal smaller geographical areas, distinctions in Commission (RCAHMS 1971: 18, quoted by architecture and material culture occur Harding 1997: 123) classify these structures as: (Henderson 2007).

A comparatively small defensive Most of the questions regarding the structure with a regionality, the chronology and the disproportionately thick dry stone contemporaneity of the brochs and duns wall, usually but not always sub depend largely on the evidence gathered circular or oval on plan, and through excavation (Ballin Smith & Banks enclosing an area not exceeding 2002: 219). However, it is well known that few about 375 sq.m (4000sq.ft.) brochs or duns have ever been excavated as far as their original foundations – an exception to this is Kiess broch, Caithness (Heald and

-48- assemblage

Jackson 2001) – and this fact demands that Age Scotland. By taking the advice of Armit et archaeologists discuss these structures with the al. (2001: 14) and Pope (2007: 223), who call limited evidence acquired from the beginnings for further research on regional differences, a of broch studies a hundred years ago or more. brief investigation and interpretation regarding It is this lack of evidence from excavation for the regional orientation patterns of the brochs the Scottish Iron Age which has led scholars to and duns follows and illustrates distinct explore other approaches to the archaeological regional patterning and conformities to certain record. One such method which has the cardinal points. potential of revealing regional differentiation is the study of doorway orientation patterning. Recently, Rachel Pope’s (2007) research has The Big Picture pointed out the general weaknesses in contemporary orientation research, by which Out of the 1,276 sites recorded in the author’s this study was influenced. dataset, only 588 had discernable entrances and this is believed to be a sufficient number for examination (Fig.2). Out of these, 190 are brochs and 398 are duns. A general eastern preference is noted, with a total 58 percent of sites orientated between NNE and SSE (the eastern arc). The western arc accounts for 32 percent with the additional 10 percent shared between due N and S. With regard to Oswald’s statement of 'a strong bi-modal tendency (to the E/SE’ (1997: 89), it is a surprising discovery to find that the orientations of the E, ESE and SE (the E/SE arc) only account for 36 percent of the entire set. This essentially leaves a high proportion (64 percent) of the set to be Figure 2 Compass graph illustrating the orientated outside of the E/SE arc. Obviously available orientations of 588 brochs and therefore, a single cosmological model based duns across Scotland. on a supposed bi-modal tendency towards

E/SE stretching across Britain – thereby including Scotland – is unfounded and there is Focusing particularly on Oswald’s (1997) and indeed far more variation than both Parker Parker Pearson’s (1996, 1999) orientation Pearson’s (1996, 1999) and Oswald’s (1997) studies, Pope notes how previous orientation studies suggest, as Pope (2007) has previously research within Britain has attempted to stated. extend single cosmological models based on a supposed tendency to orientate within the E/SE arc, across Iron Age Britain – thereby including all the diverse regions of Iron Age Scotland. In particular referring to Oswald’s statement that there is ‘a strong bi-modal tendency (to the E/SE), which is repeated with little variation throughout the Iron Age, with only rare local difference’ (1997: 89), Pope (2007) believes that such studies have resulted in a general acceptance of an E/SE orientation tradition existing throughout the diverse regions of Iron Age Britain’ (Armit 2003: 63; Fitzpatrick 1997: 77; Faust 2001: 130; Hill 1995; Parker Pearson 1996; Parker Pearson Figure 3 Compass graph illustrating the 1999: 43). orientations of all available Scottish broch orientations (190 sites). By utilising information acquired from Canmore within the Royal Commission on Ancient and Historical Monuments of Scotland database inventories, this paper attempts to present, examine and interpret the range of doorway orientation patterns of the brochs and duns throughout the various regions of Iron

-49-

CROWTHER

large proportion. Conversely, only 33.7 percent of brochs are orientated within the western arc, with 12.6 percent facing due-west. These percentages dispute Parker Pearson and Sharples’ (1999: 17) account that at least half the brochs are oriented west whilst also disputing Parker Pearson (1999: 45), who states that more than half face west, as Pope (2007: 212) has previously noted. When comparing type, 59.5 percent of broch sites are orientated within the eastern arc as well as 56.8 percent of duns. The western arc is equally similar with the duns’ 33.9 percent Figure 4 Compass graph illustrating the compared with the brochs’ 33.7 percent. orientations of all available Scottish dun Although this suggests little distinction, the orientations (398 sites). duns actually form a much wider range across the sixteen cardinal points and their weighted majority towards the overall NE (84 percent) and SW (74 percent) orientations distinguish them from the brochs and suggest a possible difference in function and purpose, but more importantly, it provides the potential for regional variation.

Considering functionalist reasoning behind differing orientation, the most obvious concern could be assumed to be the provision of light and the avoidance of the wind. Regarding light, the sun cycle is, during all seasons, from the east through the south to the west, thus a northern doorway would never admit any Figure 5 General seasonal location of the direct light and this accounts for their small rising and setting sun. Measurements percentage of the overall set (4.8 percent). A taken by author. southward doorway would be expected to maximise daylight (Oswald 1997), yet due south is strictly avoided accounting for 3.8 percent of the dataset. The reasoning for this N/S deficiency is argued by Parker Pearson (1999: 43) and Parker Pearson and Richards (1994: 47) to be the result of a focus on the rising and setting sun of the equinoxes and solstices, yet this need not be the case. Figure 5 illustrates the seasonal location of the rising and setting sun. At the time of the midsummer solstice in June, the sun rises in the NE and sets in the NW. Every day after that, the sun rises and sets a little further south and by the Figure 6 Compass graph illustrating the time of the autumn equinox, the sun rises due orientations of the available Argyll and east and sets due west as it also does in spring. Bute duns. It continues to rise and set ever southward until, by the winter solstice in December, it rises in the SE before setting in the SW. After Narrowing Down: Comparing the that day, the sun rises and sets a little further Brochs and Duns north until the summer solstice when the cycle begins again. The rising and setting sun of the When compared by structure type, the brochs solstices and equinoxes thereby account for and duns present striking dissimilarities (Figs. any direction in the eight point compass which 3 and 4). The brochs appear to be orientated in is not due N/S. Yet if the brochs and duns had an uneven arc from east through the south to no openings (e.g. windows) other than the west, with a tendency on the E/SE arc, which single doorway to allow light to directly enter make up 45.8 percent of broch orientations, a the house (Hingley 1992: 14), it would be

-50- assemblage functionally illogical to orientate a doorway dataset in comparison to the 7.9 percent from due south as the sun is always highest in the the broch sites of Sutherland, Caithness, southern sky and would not directly illuminate Orkney and . The very few brochs that the interior as effectively as a doorway do exist in the Argyll region seem to be orientated between the other points of the east- generally orientated within the eastern arc. By south-west arc. Many sites are thus merely taking a functionalist interpretation, a NE orientated toward the low sun either in the orientated house is somewhat illogical as the morning or afternoon – depending on the time interior would receive direct sunlight through a of year and the location and latitude of the very short period of the year – midsummer. A given site, as will be explored. due east/west orientation on the other hand would receive direct sunlight throughout Any orientation within the east-south-west arc, spring, summer and autumn, whereas a excluding due south, would admit direct SE/SW orientated site would receive light sunlight into a roundhouse throughout the throughout the year. majority of the year. With this in mind, the bulk of broch orientations (76.8 percent) are found within this arc, thereby suggesting that rather than equinoxes or solstices, it is more likely that the majority of broch inhabitants favoured direct light from the low sun throughout the majority of the year. Illustrating a striking contrast, only 60.3 percent of duns are to be found within that same arc and when considering points outside that arc, 18.1 percent of the entire dun dataset is orientated NE compared to 7.4 percent from the overall broch dataset. The reasoning Figure 7 Compass graph illustrating the behind this is that there is a clustering of NE orientations of those Argyll duns which are orientated duns located within the Argyll below 15m in diameter. region of Scotland. Many sites in Argyll are focussed toward distinctly illogical but cosmologically significant referents; that being With regard to climatic considerations; by the midsummer sunrise of the NE and the facing directly opposite the prevailing SW midwinter sunset of the SW of which the duns winds (Pope 2007: 213), it may be that a NE also have a higher ratio of 3:1 over the brochs. orientation is avoiding such a wind, after all, as Lamb (1981: 55) contends, the Iron Age may With regards to settlement type, I now follow have witnessed ‘an extraordinarily great the orientation analysis into three distinct prominence of westerly winds, unequalled in areas of Iron Age Scotland; the Northern any other period we know about’ and its effects Settlement Sequence (including Shetland, were emphasised by the general openness of Orkney, Caithness and Sutherland), the the landscape (Armit 1996: 19). Considering Western Isles Settlement Sequence, and the this, a NE orientation is not as functional as a Argyll and Bute Settlement Sequence. It is SE which would admit more light and would notable that each of these regions present not be affected by the northern arctic winds. significant differences in orientation. I will Thereby, we may consider the possibility that first explore the significance of a NE majority there is a more cosmological interpretation to in Argyll before examining the broch be found. orientations of Northern and Western Scotland. Roofed Dun Houses and Unroofed Dun Enclosures: A The Orientation Patterns of the Difference in Orientation? Duns of Argyll and the Inner Isles Considering the size range within the category Oswald (1997: 93) notes that there are weak of duns in Argyll, Harding (1984) has trends towards the midsummer sunrise across previously noted that any dun up to 15m or less Britain. Dominated by the dun structures, the in diameter could have been roofed, thus Argyll region of Scotland contains the vast making a compact dwelling house. This he majority of NE orientated sites within Scotland categorises as a dun house, as distinct from the (Fig.6) and NE orientated sites account for larger, potentially unroofed dun-enclosures as 20.4 percent of the entire Argyll and Bute dun noted before. To the strictly circular dun

-51-

CROWTHER

houses which he illustrates, we might add the the entirety, of the year. The orientations of the possibility that an oval dun measuring 15m on unroofed enclosures on the other hand would its short axis might have been covered by a be expected to form a random pattern without hipped ridge roof rather than by a conical one consideration of either light or climate, as we (Alcock and Alcock 1987: 133, quoted in would expect for a house. Yet beyond the 15m Harding 1997: 123). On this basis, Alcock and diameter range, there remains a marked Alcock (1987: 133) note that only 66 percent of tendency to orientate NE (Figs.7 and 8) with a Argyll duns could have been roofed dwellings. notable avoidance of due N, S and W. Even From the author’s dataset, around 62 percent when we extend the diameter range to 17m of duns in Argyll are actually below 15m in there remains a trend to orientate NE (Fig.9). diameter (184 sites), leaving the remaining duns (104 sites) outside the 15m diameter Two thirds of the NW facing duns in Argyll are range as possible dun enclosures; perhaps dun enclosures (with a diameter over 15m), originally surrounding some form of internal and they demonstrates a tendency to orientate dwelling. this specific structure in such a direction. The NW is also a functionally illogical orientation for any house, receiving little light and affected by northerly and westerly winds, yet it is a significant point of focus for solstices as the midsummer sun sets in this direction.

Figure 8 Compass graph illustrating the orientations of those Argyll duns which exceed 15m in diameter.

Figure 10 Compass graph illustrating the available orientations of the duns of the northern mainland (Caithness and Sutherland) and isles.

Figure 9 Compass graph illustrating the orientations of those Argyll duns which exceed 17m in diameter.

If purely functional considerations, such as the provision of light or the avoidance of prevailing winds, were taken into account by the builders Figure 11 Compass graph for the brochs of of the duns in Argyll, then we would expect the northern mainland and the isles. dun houses (less than 15m in diameter) to orientate themselves accordingly; preferably within Pope’s (2007) E/SE orientation The orientations of ‘dun houses’ and ‘dun optimum in which both light and shelter are enclosures’ toward distinctive cosmological taken into account. The E/SE is indeed the referents suggests that in Iron Age Argyll and most preferable direction for doorway the Inner Isles, there was a degree of focus – orientation as it avoids the prevailing SW cosmologically based or not – on the NE, as winds whilst also maintaining direct sunlight well as the NW to a lesser extent. From this to enter the roundhouse for the majority, if not evidence, it could be argued that the solstices

-52- assemblage were potentially marked as significant. This focus on the NE could however also be the result of a general orientation towards an as yet unknown landscape focus. The exploration of this is beyond the remit of this paper, yet it is certainly a phenomenon that warrants further investigation in the future.

Considering the duns in Northern Scotland (those within Caithness, Sutherland, Orkney, and Shetland), there is an easterly focus (Fig. 10), with a slight tendency to orientate NE and SW. It is tempting to suggest the reasoning behind such orientation extended from Argyll to northern Scotland, but the numbers of confirmed duns in the North are far too low to conclusively demonstrate a true extension. Conversely, the orientation patterns of the brochs of the Northern Mainland and Isles show little conformity toward distinctive cosmological referents such as the NE/NW and thus present the possibility of a different model altogether, as will become apparent.

Figure 14 Map with accompanying compass graph illustrating the site orientations of the brochs of the northern mainland only (Caithness and Sutherland). © Crown Copyright Ordinance Survey. An Edina Digimap/JISC Supplied Service.

Figure 12 Compass graph illustrating the Western and Northern Broch and orientations of the brochs of the Western Dun Doorway Orientation Isles and Skye. Comparison

As a set, the brochs present a very different orientation model to the vast majority of duns that are located in Argyll. Most brochs have a single doorway (Armit 2003: 63) and out of the 523 brochs recorded in the author’s dataset, only 190 had discernable entrances. As already noted, a northern doorway would never admit any direct light, thereby accounting for their small percentage within the broch dataset (3.1 percent). A southward doorway would be expected to maximise daylight (Oswald 1997), yet due south is strictly avoided, representing Figure 13 Compass graph illustrating the only 3.7 percent of the dataset. With the usual orientations of the duns of the Western Isles exception of due south, orientation within the and Skye. E-S-W arc would admit direct sunlight into a

broch throughout the majority of the year and, as noted, the bulk of broch orientations (76.8 percent) are found within this arc. Considering

-53-

CROWTHER

the northern brochs as a set (only brochs possibility of an E/SE broch tradition across within Sutherland, Caithness, Orkney and Scotland, the data has already shown that this Shetland), there is certainly a high tendency to it is non-existent, but there are stark contrasts orientate within the E-S-W arc which accounts between regions. Although the northern broch for 80.3 percent of orientations in this zone builders generally preferred E/SE orientations (Fig.11). Within that percentage, 41.7 percent which account for 47.7 percent, only 40 are orientated within the more southerly ESE- percent of the western brochs are found within S-WSW arc. In comparison however, we find the same arc, and a northerly E/NE orientation that the broch builders of the Western Isles was preferred which account for 44 percent. In and Skye (Fig.12) tended not to favour a comparison however, the duns within the southerly orientation, with a much lower 28 Western Isles and Skye preferred to orientate percent orientated within the southerly ESE-S- themselves toward the E/SE (36.6 percent) WSW arc in comparison to the 40 percent rather than the E/NE (31 percent). This is in found within the more northerly ENE-N-WNW striking contrast to the duns found within the arc. The duns that are also distributed Argyll and Bute region of Scotland just south of throughout the Western Isles and Skye also the Western Isles in which 41.8 percent of duns tended to favour a northerly orientation were to be found orientated E/NE in (Fig.13) with 36.6 of sites orientated within the comparison to the E/SE which only made up ENE-N-WNW arc in comparison to the 26.8 27.6 percent of the Argyll and Bute dun found within the southerly ESE-S-WSW arc. dataset. This difference may be in part due to the latitude of northern Scotland and its islands in As direct sunlight only enters the doorways of which the hours of daylight shrink to a brief houses orientated within the ENE-N-WNW arc midday window in the middle of winter, whilst within the summer months alone, these results midsummer daylight can last until almost also suggest that the broch (and duns to a midnight (Richards 1990: 121). We can lesser extent) builders of the Western Isles therefore assume that savouring the midwinter were much keener to orientate their homes for sunlight over the summer ambience was a the summer months only. In contrast, the primary concern to a typical northern broch Northern broch builders largely orientated and a doorway orientated within the more their homes southerly to maximise available southerly orientated SE-S-SW arc would have winter daylight. This confirms Pope’s (2007) been a more logical choice as it provides conclusions that the E/SE majority presented protection against the harsh northern winter by Oswald (1997) should not have been conditions. A SE-S-SW orientation somewhat extended into all the diverse regions of Iron protects against the northerly winds whilst Age Scotland without consideration of the providing the broch interior with maximum differing settlement trends found throughout winter daylight; however the south can largely Scotland. be affected by the SW winds and this fact accounts for the lack of SWW-SW-WSW (the As 83 percent of all broch structures are SW arc) orientations both in the Western Isles located in the Northern Mainland and Isles and Skye (8 percent for the brochs; 7 percent (including Sutherland, Caithness and the for the duns), and in the Northern Mainland Orkney and Shetland Islands), the and Isles (7.6 percent for the brochs). An E/SE interpretation and significance of the orientation actually provides the optimum orientations within this area alone will be the protection from the northerly and prevailing main focus of attention for the remainder of SW wind whilst allowing some degree of direct this paper. light to enter the broch or dun throughout the year. If both light and climate have a role in influencing orientation, we could thus expect The Northern Broch Analysis that this might have led to an E/SE tradition, especially on the Northern Mainland and Isles The Northern Mainland (only Caithness and where the lack of sheltering forests and winter Sutherland) contains a large majority of broch daylight would have accentuated the sites orientated within the E-S-W arc (74.2 requirement for maximising both shelter and percent), with a tendency to orientate within light. When we actually consider the brochs as the E/SE arc, accounting for 42 percent a single group however, we find that an actual (Fig.14) compared with the minority of 31.2 majority of sites (54.2 percent) are orientated percent which lie within the southerly SE-S- outside the E/SE arc. This means that only SW arc. 45.8 percent – less than half of the broch dataset – is orientated towards the E, ESE or SE (the E/SE arc). When considering the

-54- assemblage

space, at least in Orkney. The fact that a high majority of sites on Orkney are located within the E/SE arc also suggests that the further north the brochs were built, the more important environmental concerns became. As an E/SE orientation permits sunlight to directly illuminate the interior of the broch throughout the year, the E/SE majority on Orkney suggests that the Orkney brochs were potentially occupied for the entirety of the year. In comparison, a high proportion of brochs within the Western Isles and Skye were northerly orientated and this suggests a summer habitation as an orientation within the ENE-N-WNW only permits summer sunlight to illuminate the broch interior. Either that or the Western brochs were indeed inhabited by those who dwelt within upper storeys without the requirement of direct sunlight to illuminate the lower storeys and who could therefore orientate their dwellings in a more random manner.

Orientation patterning on Orkney may of course also vary due to topographical necessities (Bourdieu 1977; Oliver 1987: 161- 163) and it is Orkney’s flat topography (and Caithness to a lesser extent) which would have allowed the optimum uniformity of E/SE orientations in which the light from the low rising sun of winter, spring and autumn was not compromised by high terrain and would have entered the low doorways of the brochs. Figure 15 Map with accompanying Although we would expect the same conformity compass graph illustrating the site from the Shetland Islands, they offer a orientations and patterns of the brochs of distinctly dissimilar pattern of orientation the Orkney Islands. © Crown Copyright (Fig.16) as only 37 percent of Shetland’s brochs Ordinance Survey. An Edina are orientated within the E/SE arc. However, Digimap/JISC Supplied Service. direct sunlight is still a concern as a very high

Only the Orkney Islands present what could be majority of brochs are focussed within the E-S- described as a standard pattern of orientation W arc (88.9 percent) and a high proportion of (Fig.15). The conformity towards the E/SE may these (40.7 percent) are found orientated be linked with Pope’s (2007) general within the more southerly SE-S-SW arc, shelter/light optimum for structure indicative perhaps of the need to maximise orientation, with E/SE orientations making up winter sunlight in such extreme northern 73.9 percent of the Orkney set, a very high latitudes. Shetland’s brochs also present a proportion indeed. Of course, orientation may couple of sites orientated due south – a also depend on whether the ground floor of a direction of no obvious solar cosmological broch was a utilised space or not – an issue referent. In consideration of low light however, which may have differed across Scotland and a due south orientation is made possible by requires further investigation that is currently Shetland’s extreme northern latitude and outside the scope of this paper. The would have allowed low southern sunlight to commonality of a scarcement ledge in most directly enter the domestic interior in the well-preserved brochs has led to the theory depths of winter. that the main living area was located above ground level on a raised timber floor (Sharples 1998: 38-40) although some scholars disagree (e.g. MacKie 2010). The E/SE orientation conformity towards the light/shelter optimum on Orkney does however suggest that the ground floor was indeed utilised as a living

-55-

CROWTHER

explore whether there was indeed an E/SE doorway orientation majority that could be extended into all the multifarious regions and landscapes of Iron Age Scotland. Out of the 1,276 duns and brochs located throughout Scotland, 588 had discernable entrances towards specific cardinal points and a high proportion of these were actually found to be orientated outside of the E/SE arc. By then narrowing down this dataset into structural types, the duns and brochs appear to be strikingly divergent from one another. Although the dun dataset forms a wide pattern across the entirety of the 16 point compass, a high majority of brochs are found to be located within the more southerly E-S-W arc, suggesting that the broch inhabitants favoured an orientation that sought direct sunlight throughout the majority of the year. This is because the sun cycle is, during all seasons, from the east through the south to the west and therefore an orientation within the E-S-W arc maximises the amount of daylight granted throughout the year.

The duns of Argyll and the Inner Isles, on the other hand, did not reflect the orientation pattern of the brochs, as both types of dun structure – dun houses and dun enclosures – focus orientation toward specific cosmological referents (NE, NW). This feature is notably lacking within the orientations of the Northern broch set of Caithness, Sutherland, Orkney and Shetland. The fact that a large proportion of dun enclosures – a type of site that may have enclosed an internal dwelling – were also focussed toward the NE also suggests a more Figure 16 Map with accompanying ritualistic rather than functional purpose to compass graph illustrating the site orientation. This conformity of orientation orientations and patterns of the brochs of towards the NE, and its possible cosmological the Shetland Islands. © Crown Copyright associations, is also something dramatically Ordinance Survey. An Edina lacking anywhere else in Scotland. Digimap/JISC Supplied Service.

With regards to an E/SE majority, only the Conclusion brochs upon Orkney present E/SE orientation conformity; an optimal orientation that permits year round sunlight to enter the Within this paper the author has attempted to roundhouse and protects against the SW wind, present, examine and interpret a range of the thereby suggesting that the brochs on Orkney available doorway orientation patterns of the were utilised throughout the year. The brochs and duns, focusing particularly on three Northern brochs in general prefer a southerly cultural zones of Iron Age Atlantic Scotland; orientation, perhaps due to the extreme Argyll, the Western Isles and Skye, and the northern latitude of Caithness, Orkney and Northern Mainland and the . especially Shetland. In comparison however, Pope’s (2007) research, which explored the the brochs of the Western Isles tended to weaknesses in recent orientation studies that favour northerly orientation. We could attempted to extend single cosmological interpret this by suggesting that the broch models based on a tendency to orientate within builders of the Western Isles and Skye sought the E/SE arc across Iron Age Britain, acted as a to inhabit their homes only during the summer platform on which the author sought to explore months when direct sunlight was able to enter regional orientation variance across Iron Age the doorway whereas the northern brochs Scotland. First and foremost, the aim was to

-56- assemblage sought sunlight throughout the year, especially Ballin Smith, B. and Banks, I. 2002. during the winter months when low southerly Conclusions. In: Ballin Smith, B. and Banks, I. sunlight direct illuminated the interior. It may (eds.) In the Shadow of the Brochs. Stroud: also suggest however that those inhabiting the Tempus, 217-221. Western brochs lived above ground level and therefore the requirement for direct sunlight to Barrett, J. 1981. Aspects of the Iron Age in enter the entrance was minimized, whereas the Atlantic Scotland: A Case Study in the Northern broch builders may have sought to Problems of Archaeological Interpretation. utilise their ground level as a living space Proceedings of the Society of Antiquaries of which is demonstrated in their logical Scotland 111: 205-19. southward orientations. Bourdieu, P. 1977. The Kabyle House or the It is by conducting the above analysis that this World Reversed. Cambridge: Cambridge paper has attempted to shed new light on University Press. Scottish orientation research, and by which it is suggested that instead of a uniform E/SE Faust, A. 2001. Doorway Orientation, orientation majority that spread throughout Settlement Planning and Cosmology in Ancient Iron Age Scottish society, there was actually Israel during Iron Age II. Oxford Journal of wide spread variation that was in accordance Archaeology 20: 129-155. with architectural distinctions. The author hopes to instigate further scholarly research Fitzpatrick, A. 1997. Everyday Life in Iron Age into the orientation patterns of Iron Age Wessex. In: Gwilt, A. and Haselgrove, C. (eds.) Scotland. Reconstructing Iron Age Societies. Oxford: Oxbow Monograph 9, 73-86.

Acknowledgements Harding, D. 1984. The Function and Classification of Brochs and Duns. In Burgess, First and foremost, I am very grateful to Dr C. and Miket, R. (eds.) Between and Beyond Meggen Gondek at the University of Chester the Walls: Essays on the and for advising me to write and publish this paper. History of North Britain. : Donald, I would also like to thank all the staff in the 206-20. History and Archaeology department at the University of Chester for their support whilst I Harding, D. 1997. Forts, Duns, Brochs and studied there. Thanks also to my parents and : Iron Age Settlements in Argyll. In: my friends Jonathan Coley and Penny Lane for Ritchie, G. (ed.) The Archaeology of Argyll. their encouragement. Edinburgh: Edinburgh University Press, 118- 140.

References Heald, A. and Jackson, A. 2001. Towards a new understanding of Iron Age Caithness. Alcock, E., and Alcock, L. 1987. Proceedings of the Society of Antiquaries of Reconnaissance Excavations on Early Historic Scotland 131: 129-148. and other Royal Sites in Scotland, 1978-84, Excavations at Dunollie Henderson, J. 2000. Shared Traditions? The , Oban, Argyll, 1978. Proceedings of the Drystone Settlement Records of Atlantic Antiquaries of Scotland 117: 119-47. Scotland and 700 BC-AD 200. In: Henderson, J. (ed.) The Prehistory and Early Armit, I. 1996. The Archaeology of Skye and History of Atlantic Europe. Oxford: British the Western Isles. Edinburgh: Edinburgh Archaeological Reports: International Series, University Press. 861, 117-154.

Armit, I. 2003. Towers of the North: The Henderson, J. 2007. The Atlantic Iron Age: Brochs of Scotland. Stroud: Tempus Settlement and Identity in the First Publishing. Millennium BC. Oxon: Routledge.

Armit, I., Champion, T., Crieghton, J., Gwilt, Hill, J. 1995. Ritual and Rubbish in the Iron A., Haselgrove, C., Hill, J., Hunter, F., and Age of Wessex. Oxford: Archaeopress, BAR Woodward, A. 2001. Understanding the British Series 242. British Iron Age: An Agenda for Action. Salisbury: Wessex Archaeology.

-57-

CROWTHER

Hingley, R. 1992. Society in Scotland from Reconstructing Iron Age Societies. Oxford: 700BC to AD 200. Proceedings of the Society Oxbow Monograph 9, 87-95. of Antiquaries of Scotland 122: 7-53. Parker Pearson, M. 1996. Food, Fertility and Lamb, H. 1981. Climate from 1000BC to Front Doors in the First Millennium BC. In: 1000AD. In: Dumbleby, G. and Jones, M. Champion, T. and Collis, J. (eds.) The Iron Age (eds.) The Environment of Man: The Iron Age in Britain and Ireland: Recent Trends. to the Anglo-Saxon Period. Oxford: British Sheffield: J.R. Collis Publications, 117-132. Archaeological Report 130. Parker Pearson, M. 1999. Food, Sex and Death: MacKie, E. 1965. The Origin and Development Cosmologies in the British Iron Age with of the Broch and Wheelhouse Building Particular Reference to East . Cultures of the Scottish Iron Age. Proceedings Cambridge Archaeological Journal 9: 43-69. of the Prehistoric Society 30: 93-146. Parker Pearson, M., & Richards, C. 1994. Mackie, E. 2010. The Broch Cultures of Architecture and Order: Spatial Atlantic Scotland. Part 2. The Middle Iron Age: Representation and Archaeology. In: Parker High Noon and Decline c. 200 BC-AD 550. Pearson, M. and Richards, C. (eds.) Oxford Journal of Archaeology 29: 89-117. Architecture and Order: Approaches to Social Space. : Routledge, 38-72. Nieke, M. 1983. Settlement Patterns in the First Millennium AD: A Case Study of the Parker Pearson, M. and Sharples, N. 1999. Island of Islay. In: Chapman, J. and Mytum, H. Between Land and Sea: Excavations at Dun (eds.) Settlement in North Britain 1000 BC-AD Vulan, South . Sheffield: Sheffield 1000. Oxford: British Archaeological Report Academic Press. 116, 299-326. Pope, R. 2007. Ritual and the Roundhouse: A Nieke, M. 1984. Settlement Patterns in the Critique of Recent Ideas on the use of Domestic Atlantic Province of Scotland in the 1st Space in Later British Prehistory. In: Millennium AD: A Study of Argyll. Haselgrove, C. and Pope, R. (eds.) The Earlier Unpublished PhD thesis, University of Iron Age in Britain and the Near Continent. Glasgow. Oxford: Oxbow Books, 204-228.

Nieke, M. 1990. Fortifications in Argyll: RCAHMS. 1971. Argyll: An Inventory of the Retrospect and Future Prospect. In: Armit, I. Ancient Monuments: Vol. 1, . (ed.) Beyond the Brochs: Changing Edinburgh: Her Majesty’s Stationary Office. Perspectives on the Later Iron Age in Atlantic Scotland. Edinburgh: Edinburgh University Richards, C. 1990. The Late Neolithic House in Press, 131-42. Orkney. In: Samson, R. (ed.) The Social Archaeology of Houses. Edinburgh: Edinburgh Oliver, P. 1987. Dwellings: The House across University Press, 111-124. the World. Austin: University of Texas Press. Sharples, N. 1998. Scalloway: A Broch, Late Oswald, A. 1997. A Doorway on the Past: Iron Age Settlement and Medieval Cemetery Practical and Mystic Concerns in the in Shetland. Oxford: Oxbow Monograph 82. Orientation of Roundhouse Doorways. In: Gwilt, A. and Haselgrove, C. (eds.)

-58-