Evidence from monuments

One reason why insular rock art was seldom studied is that its date was uncertain, but a new consensus is emerging. This section treats decorated monuments separately from open-air sites.

Initial developments (Early to Middle Neolithic)

The oldest rock art should belong in the early or mid-4th millennium bc, when a mortuary structure at Dalladies long barrow in north-east Scotland contained a cup-marked slab; the dating evidence is imprecise (Piggott Reference Piggott1974; Sheridan & Schulting Reference Sheridan, Schulting, Gebauer, Sørensen, Teather and Valera2020). There are similar motifs on the capstones of portal tombs. New research suggests that these monuments were among the first to be built (Lynch Reference Lynch2014). Their overall distribution coincides with that of surfaces with the same motif (Jones & Kirkham Reference Jones and Kirkham2013), and Cummings and Richards (Reference Cummings and Richards2021) suggest that pieces with a special significance were introduced when megaliths were constructed.

The floruit of insular rock art (Middle to Late Neolithic)

The re-use of already decorated stones also featured in the Irish cemetery of Loughcrew where there are two sources of information (Shee Twohig et al. Reference Shee Twohig, Roughley, Shell, Reilly, Clarke and Swanton2010). It includes decorated outcrops from which pieces seem to have been detached and weathered fragments with similar decoration in the fabric of Cairn T, where they are associated with megalithic art in fresh condition. Most accounts suggest a context in the late 4th millennium bc (Eogan & Cleary Reference Eogan and Cleary2017, 331–79). That is consistent with the re-use of raw material in the passage graves of the Boyne Valley. Motifs on the back of individual kerbstones at Newgrange resemble those in the open air but, in this case, they were well preserved when the structure was built (O’Kelly Reference O’Kelly1982, 146–85).

A sequence of decorated panels is associated with Irish passage tombs. In some cases, incised designs are earlier than pecked designs, but they can also be combined with one another (Eogan Reference Eogan1997; Eogan & Shee Twohig Reference Eogan and Shee Twohig2022). Other developments concern ways of working the stone (O’Sullivan Reference O’Sullivan1986). The first images were lightly pecked (O’Sullivan’s ‘depictive’ style), but later images were bolder and moulded to the contours of the rock (his ‘plastic’ style). There is less evidence from Orkney, where incised elements predominate in passage graves. Pecked motifs are present at a few sites and employ similar techniques to the Irish monuments (Bradley et al. Reference Bradley, Phillips, Richards and Webb2001; Thomas Reference Thomas2016).

Comparable images are associated with other structures. There are pecked designs on two stone settings built around 3000 bc: Long Meg in Cumbria where the motifs recall those in Ireland (Frodsham Reference Frodsham, Hey and Frodsham2021); and possibly Temple Wood in the west of Scotland where the most conspicuous design is paralleled in Orkney tombs (Scott Reference Scott1989).

Later developments (Chalcolithic to Early Bronze Age)

Discussions of insular rock art have been dominated by accounts of decorated cist slabs. They raise many difficulties – sometimes their association with graves is debateable (Shee Twohig Reference Shee Twohig1972, 232–3) and their distribution is restricted (they are especially uncommon in Ireland). Nearly 50 years ago Simpson and Thawley (Reference Simpson and Thawley1972) argued that most of these stones were re-used. The pecked decoration was worn and sometimes truncated but this did not happen in every case (Beckensall & Frodsham Reference Beckensall and Frodsham1998; Baker & Wright Reference Baker and Wright2009). It did include motifs associated with megaliths or decorated outcrops but other panels were created specifically for inclusion in the grave. Thus, the practice of making rock art must have extended into the 2nd millennium bc.

This is shown by other structures. The decoration associated with Clava Cairns is fitted to the outlines of individual stones in the kerbs and one of the chambers (Fig. 3). Complex circular designs are absent and there is an emphasis on cup marks and cups with single rings. These monuments were used between about 2200 and 1800 bc (Bradley Reference Bradley2000). At other sites cup marks are associated with depictions of diagnostic metalwork. The decorated cists in Kilmartin Glen include drawings of axes and a halberd which occur together with an array of cups and a single ring (Watson & Bradley Reference Watson and Bradley2021). Cups also feature on the kerb of a round barrow in Wessex where they accompany carvings of axes and daggers (Piggott Reference Piggott1939).

Fig. 3. Two decorated panels in the Early Bronze Age cemetery at Clava. The upper picture shows the junction between the passage and central chamber of the SW cairn. The lower image shows a decorated kerbstone in the NE cairn. In each case the motifs have been fitted to the outline of the stone (photographs: Richard Bradley)

Evidence from open-air sites

Like those at passage graves, individual designs were made and altered over significant periods of time. There are differences of weathering between those made on the same surface and a few are superimposed or defaced (Valdez-Tullett Reference Valdez-Tullett2019, 93–6). Motifs abut one another and some panels may have increased in size or elaboration. It is not known whether that process was continuous or discontinuous.

Three sites provide more direct evidence of sequence and, in this respect, they compare with the information from chambered tombs. At Hunterheugh in Northumberland excavation of a decorated outcrop identified two phases of decoration (Waddington Reference Waddington2005). The first group consisted of cup marks and cups enclosed by one or two rings. They were like the simple motifs behind the kerbstones at Newgrange. During a second period of activity at Hunterheugh large sections of the outcrop were removed and new designs were created on the freshly exposed surfaces. Now they were larger and more prominent and took a different form. One of the cups was enclosed by three concentric rings and other motifs were joined together. A new panel featured a wavy line or ‘serpentiform’ with parallels in megalithic art (Eogan & Shee Twohig Reference Eogan and Shee Twohig2022, 115).

There may have been similar developments at Auchentorlie above the Clyde Estuary in western Scotland. Here there were at least two phases of stone working separated by an interval in which parts of the surface were removed (MacKie & Davis Reference MacKie and Davis1989). The sequence is difficult to work out as the interpretation was based on photographs taken before the outcrop was destroyed but, in one case, part of a panel of cup marks and cups with single rings was truncated and a pair of eight concentric circles took its place. Their size and execution seem to be shared between open sites and passage graves.

A third sequence was at Achnabreck at Kilmartin in western Scotland and was identified by differences of weathering between images on the same surface (Fig. 4; RCAHMS 1988, 87–98; Watson Reference Watson, Frodsham and Sharpe2022). This approach cannot be taken too far but Aaron Watson’s recent work identified obvious contrasts between the condition of adjacent motifs, some of which were superimposed. The first phase featured three groups of linked spirals, as well as single and concentric circles; there were no cup marks. The distinctive horned spirals compare with those at monuments in Orkney and with the decorated flint macehead from Tomb 1B at Knowth in the Boyne Valley (Eogan & Cleary Reference Eogan and Cleary2017, 440–2). A second phase at Achnabreck is represented by very different images: cup marks, cups surrounded by several concentric rings, cups and rings joined to radial lines, and a large enclosure. Again, there is a striking difference between the techniques used to make them. Watson compares the execution of the first designs with the earlier style of tomb decoration in Ireland. In the second phase at the Scottish site the motifs were bolder and in this respect they have more in common with later panels in Irish passage graves.

Fig. 4. The decorated surface at Achnabreck, Kilmartin Glen (Western Scotland), where successive designs have been identified by differences of weathering and style (photograph: Aaron Watson)

These comparisons cannot be taken too far but there is more direct dating evidence. Two decorated outcrops on the Ben Lawers estate in the southern Highlands of Scotland produced finds of Arran pitchstone which should date from the Early or Middle Neolithic period (Bradley et al. Reference Bradley, Watson and Anderson-Whymark2013). The same happened at Torbhlaren near Kilmartin on the west coast where concentrations of rock art are also associated with radiocarbon dates in the Late Neolithic and Chalcolithic periods: 2920–2860 cal bc and 2580–2340 bc (Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 57–8). Similarly, there was a deposit containing Grooved Ware and flint artefacts beside panels of rock art on Rombalds Moor in the southern Pennines. In this case there are radiocarbon dates of 3090–2630 cal bc and 2960–2610 cal bc (Edwards & Bradley Reference Edwards, Bradley, Cleal and MacSween1999). A few decorated outcrops in north-east England were overlain by circular cairns attributed to the Bronze Age. One of those sites is Hunterheugh (Waddington Reference Waddington2005).

Incised angular designs

These are a particular feature of Orkney passage graves (Bradley et al. Reference Bradley, Phillips, Richards and Webb2001; Thomas Reference Thomas2016). It is not clear how they were used, and it seems possible that they were cut through areas of pigment using a sgraffito technique (Bueno-Ramírez et al. Reference Bueno-Ramírez, de Balbín Behrmann, Barroso-Bermejo, Laporte, Gouezin, Cousseau, Müller, Hinz and Wunderlich2019). The same motifs were employed in other buildings, including houses like those at Pool or Skara Brae and the specialised structures on the Ness of Brodgar (Thomas Reference Thomas2016). This distinctive style extends to Irish passage tombs where it represents one of the earliest elements, although some panels are embellished with a combination of incised and pecked motifs (Eogan Reference Eogan1997). Despite that connection, angular designs are usually employed on portable objects and connected with the domestic domain (Jones & Díaz-Guardamino Reference Jones and Díaz-Guardamino2019). Like Grooved Ware which shares similar elements, they probably originated in Orkney (Bayliss et al. Reference Bayliss, Marshall, Richards and Whittle2017).

Pecked circular designs

These designs were common in the open landscape where angular motifs are very rare. Curvilinear designs are found in at least 14 regional concentrations across Britain and Ireland (Fig. 2) and extend from simple circular devices to sets of concentric rings that extend right across the decorated surface. They are often integrated with radial lines leading to (or from) the centre of the circle where there is a cup mark. In the most elaborate panels, like those in south-west Ireland, western Scotland, and Northumberland, these elements may be linked together to form complex panels. Many designs are conspicuous and, in certain cases, the images were renewed. Occasionally there are indications that part of the decorated surface was removed; Hunterheugh provides a well-documented example (Waddington Reference Waddington2005). The curvilinear style of rock art is rare on buildings or portable artefacts and is more common in the open landscape (Bradley Reference Bradley1997). In Ireland the main concentrations complement the distribution of passage graves (Loughcrew is an exception) and in Orkney, where there are similar monuments, there is little to suggest that natural exposures of bedrock were embellished.

Megalithic art

This represents a third tradition but it is different from the others because it could combine angular and curvilinear designs in the same panels (Robin Reference Robin2009; Eogan & Shee Twohig Reference Eogan and Shee Twohig2022). Apart from some early motifs they were formed by pecking. Individual elements might be changed several times and some were eventually erased (Eogan Reference Eogan1997; Cochrane Reference Cochrane and Russell2006; Robin Reference Robin2009; Eogan & Shee Twohig Reference Eogan and Shee Twohig2022). The curvilinear designs are best known in the Boyne Valley and are quite different from those in the open-air rock art of other regions. Cup marks occur less often, the rings are seldom breached by a radial line and spirals are represented as well as circular motifs. The same applies to a few panels associated with passage tombs in Orkney (Thomas Reference Thomas2020). Scottish and Irish monuments share some elements with places in the wider landscape, but they are rare (Freedman in Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 284–311). They include: simple or more elaborate spirals; C-shaped or ‘horned’ spirals; rosettes; unbroken concentric circles; stars; ‘boxed U motifs’; and ‘serpentiforms’ (Robin Reference Robin2009; Eogan & Shee Twohig Reference Eogan and Shee Twohig2022). The best known examples are in Inishowen in the north of Ireland (O’Connor Reference O’Connor2006,118–19), at Ballochmyle (Stevenson Reference Stevenson1993) and Ormaig near the west coast of Scotland (Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 204–10), Morwick in Northumberland (Beckensall Reference Beckensall1991, 111–16), and Copt Howe in Cumbria (Bradley et al. Reference Bradley, Watson and Style2019). They stand out from most of the other panels in the same regions.

In a sense, passage tomb art in Ireland provided a synthesis of the styles associated with the domestic domain and the landscape respectively (Fig. 5; Bradley Reference Bradley2009, 120–2). The comparison goes even further as there are panels of open-air rock art that combine the design elements associated with natural places in a similar manner to the decoration of Irish monuments. They include examples at Achnabreck, Ballochmyle, and Morwick (Fig. 6).

Fig. 5. The styles of insular rock art. The left-hand image illustrates an angular design from the Late Neolithic settlement at Pool, Orkney, and the right-hand image shows part of a curvilinear design in the open landscape at Hunterheugh, Northumberland. Passage grave art is represented by two designs at Knowth which combine these elements. Adapted from Bradley (Reference Bradley2009). Drawing: Courtney Nimura

Fig. 6. Conjoined circular motifs in the Irish passage graves of Newgrange and Loughcrew compared with images in the wider landscape at Achnabreck, Ballochmyle (Western Scotland) and Morwick (Northumberland) (drawing: Courtney Nimura)

Summary

These relationships shed further light on the age of insular rock art. If the decoration of Irish passage tombs drew on two sources of information – an angular tradition that originated in Orkney and curvilinear images with a wider distribution – at some stage all three styles would have co-existed, although their overall chronology is not clear. Radiocarbon dating suggests a period beginning in the late 4th millennium bc and continuing long afterwards. On the other hand, there is no evidence that that they remained significant for the same lengths of time. As O’Connor argued in 2006, megalithic art and open-air art represent a continuum (cf. Johnston Reference Johnston1993; Erlander Reference Erlander2015), but the decoration of natural surfaces retained its importance when chambered tombs were no longer built; Clava passage graves provide the obvious exception (Bradley Reference Bradley2000).

The international scale

The international context of insular rock art is the subject of new research by Valdez-Tullett (Reference Valdez-Tullett2019). Instead of focusing exclusively on the pecked motifs, she combines her analysis of the images with an investigation of other features: the selection of particular surfaces for treatment; their positions in the local terrain; and the techniques by which the designs were made. She accepts the reality of a single tradition of rock art extending from Portugal to Ireland but recognises the existence of regional variations between her five study areas: Monte Faro (northern Portugal), Barbanza (north-west Spain), Iveragh (south-west Ireland), Galloway (south-west Scotland), and Rombalds Moor (in the southern Pennines). Chronology remains a problem as Iberian scholars disagree on when Atlantic rock art originated. Although there are depictions of Early Bronze Age weapons in Spain and Portugal, they may not date from its first phase, and some researchers favour an earlier origin, basing their case on comparisons with the decoration of passage graves (Fábregas Valcarce Reference Fábregas Valcarce, de Balbín Behrmann, Bueno Ramirez, González Antón and del Arco Aguilar2009; Santos Estévez Reference Santos Estévez, Cruz Berrocal, García Sanjuan and Gilman2013). That is plausible since it is accepted that megalithic art shares elements between Iberia, Brittany, and Ireland. They were reviewed in a recent study (Eogan & Shee Twohig Reference Eogan and Shee Twohig2022, 227–57).

The regional scale

Valdez-Tullett’s analysis suggests connections between northern Portugal and south-west Ireland. She also identifies close links between the rock art of Iveragh and Rombalds Moor (in the southern Pennine group). Her findings complement accounts of long distance connections based on more familiar sources: the forms of chambered tombs, henge monuments, and stone circles. As Carlin recognised, the presence of Grooved Ware is important on both sides of the Irish Sea (Carlin Reference Carlin2017). But such studies suffer from inevitable limitations. They depend on the existence of specialised kinds of monuments and the survival of decorated pottery; other kinds of information play a limited role. The distribution of rock art extends into regions in which this evidence is lacking. Can it shed light on the relationship between places that have not been considered before? The presence of pecked designs depends on the availability of suitable bedrock and certain areas have little to show, but discoveries of objects with comparable decoration provide some indication of what might once have been present. Their distributions extend from the east coast of Ireland to the Isle of Man. In Britain they are known as far north as Orkney and the Great Glen and as far south as Wessex and Cornwall. Decorated plaques were often buried in pits. They have been analysed by Jones and Díaz-Guardamino (Reference Jones and Díaz-Guardamino2019) whose book Making a Mark identifies a series of connections between Ireland and Britain.

The local scale

In some ways most progress has been made by research on a local scale. Although decorated surfaces must have been lost as land was cleared, the distribution of open-air rock art retains some salient features and, occasionally, it can be investigated in lowland areas. Decorated outcrops on the valley floor near Kilmartin in western Scotland are in an area where there is evidence that crops were grown (Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 167). Similarly, in Strathtay in the southern Highlands, there are concentrations of worked quartz on river terraces associated with cup marked stones; more complex decorated panels overlook these concentrations from higher ground (Bradley Reference Bradley1995). In upland areas their distribution focuses on springs and they can be found in areas with worked flints; Barningham Moor and Rombalds Moor are two examples in the Pennines (Beckensall & Laurie Reference Beckensall and Laurie1998; Deacon Reference Deacon2020). The rock art on Ben Lawers is associated with a series of shallow basins which are sheltered from the prevailing winds (Bradley et al. Reference Bradley, Watson and Anderson-Whymark2013); during the historical period they were used as seasonal grazing land and the remains of shielings have been identified there (Atkinson Reference Atkinson2016). Pollen analysis shows that the local landscape had already been cleared when the prehistoric designs were made. Again, this area would have been suitable as summer pasture. Insular rock art might have been integrated into the pattern of settlement.

Discussion

In some ways these approaches are inconsistent with one another. Valdez-Tullett’s study considers the relationship between areas that must have been connected by sea. That is emphasised by the term ‘Atlantic’ rock art (Valdez-Tullett Reference Valdez-Tullett2019). But most regional studies have employed the methods of landscape archaeology, investigating the relationship between the decorated surfaces and other parts of the terrain: viewpoints, routeways, and valleys (Bradley Reference Bradley1997; Fairén Jiménez Reference Fairén Jiménez2007). The sea plays little part in this research. It can be mentioned in passing – the visibility of the shoreline features in accounts of rock art in Galloway (Morris Reference Morris1979; Bradley et al. Reference Bradley, Harding and Matthews1993) – but there are two important exceptions. In south-west Ireland it has played a central role in research. In 2005 Westlake published a new study of the rock art of the Dingle peninsula and showed how its siting was related to the positions of estuaries as well as terrestrial resources. Similarly, the use of GIS has demonstrated that the main panels of rock art at Kilmartin in western Scotland had extended views out to sea (Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 227–36); this approach can be informative, but it can never be a substitute for dedicated fieldwork and must be supplemented by observations made at the sites themselves. It is interesting that work on comparable rock art in north-west Spain has undergone the same transformation. Again, it is apparent that it was related to the inlets known as rías (Santos Estévez & Gϋimil Fariña Reference Santos Estévez, Gϋimil Fariña, Skoglund, Ling and Bertilsson2015).

Relationships with estuaries, bays, and landing places

Westlake’s analysis of the Dingle peninsula emphasised the importance of estuaries (Westlake Reference Westlake2005). It applies to other complexes in south-west Ireland and to Inishowen in the north, which included important concentrations of petroglyphs (O’Connor Reference O’Connor2006, 112–13; Busher O’Sullivan Reference Busher O’Sullivan2020). The same approach extends to a series of Scottish sites, especially those in Galloway where the distribution of rock art focuses on a series of estuaries and bays (Van Hoek Reference Van Hoek1995). This example is especially persuasive as the region contains an unusual proportion of images like those at Irish monuments. Visits to these sites in 2020 showed that most of them were near to places where it would be safe to land small boats; the relationship is not affected by subsequent changes of sea level. On the other hand, the wider distributions of artefacts and monuments in the same area include sections of the coast with cliffs and rocky shorelines.

Other groups of decorated rocks illustrate the point. Those in Kilmartin Glen were accessible by sea. Again, they were near sheltered landing places, and the repertoire of pecked designs shares features with distant monuments (Fig. 7; Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 314–17). In this case there were connections with both Orkney and Ireland. The same approach might explain the concentration of decorated outcrops on Clydebank which are harder to investigate because evidence has been lost to modern development. This is another part of western Scotland with occasional motifs like those in megalithic art (Morris Reference Morris1981). There is similar information from two other regions. In England, the decorated panels at Fylingdales on the North York Moors are not far from a potential landing place in Robin Hood’s Bay (Vyner Reference Vyner2011; Brown & Chappell Reference Brown and Chappell2012). In Ireland, Dundalk Bay provided access to the concentration of decorated surfaces investigated by O’Connor (Reference O’Connor2006) in Counties Monaghan and Louth. It is known that cobbles collected on beaches in this area were taken to passage tombs in the Boyne Valley (Mitchell Reference Mitchell1992).

Fig. 7. A decorated surface at Ormaig (Western Scotland) overlooking the sea and a potential landing place (photograph: Aaron Watson)

Relationships with rivers, valleys, and the shoreline

Other concentrations of rock art are further inland but could have been accessed from the coast (Bradley Reference Bradley, Frodsham and Sharpe2022a). Three sites are particularly relevant here. In two instances – Ballochmyle and Copt Howe – they were about 20 km (a day’s travel) from the shores of western Scotland and Cumbria respectively. Ballochmyle could be reached by following the River Ayr and Copt Howe along the valleys or lakes leading north from Morecambe Bay. The decorated cliff at Morwick overlooked the Coquet a shorter distance upriver from the North Sea (Beckensall Reference Beckensall1991 111–16). All three places were readily accessible. They had other characteristics in common. The rock art of Copt Howe resembles the decoration of passage graves in Ireland and the panels at Ballochmyle share certain elements with the megalithic art of Loughcrew, especially the presence of concentric circles and stars. On the other hand, the panels at Morwick feature linked spirals like those in Orkney. In every case the decoration is confined to vertical surfaces that resemble the walls of tombs (Bradley Reference Bradley, Frodsham and Sharpe2022a). There are less dramatic concentrations of rock art along other rivers. They include the Eden in Cumbria, which runs inland from the Solway Firth (Sharpe Reference Sharpe, Mazel and Nash2022), and the Till, Breamish, and Coquet, which lead to concentrations of petroglyphs in Northumberland (Beckensall Reference Beckensall1991).

Relationships with land routes

There are important groups of rock art even further from the sea. Some of the sites are scattered but many illustrate a wider pattern. It is no accident that their distributions emphasise three long distance paths linking the North Sea and the Irish Sea. In England, petroglyphs are well represented in valleys east of the Pennines (Beckensall & Laurie Reference Beckensall and Laurie1998; Boughey & Vickerman Reference Boughey and Vickerman2003; Brown & Brown Reference Brown and Brown2008) and, in some cases, they followed (or more often overlooked) two of these trails. That would account for the exceptional number of sites around Rombalds Moor (Cowling Reference Cowling1946; Boughey & Vickerman Reference Boughey and Vickerman2003) and for similar groups near Barnard Castle (Beckensall & Laurie Reference Beckensall and Laurie1998; Brown & Brown Reference Brown and Brown2008). Both cross country routes featured in The Personality of Britain, where one of Fox’s maps showed how they could have connected important places on the coast (Fox Reference Fox1932, map B). There is similar evidence from the southern Highlands of Scotland where the concentrations of decorated surface images in Strathtay and further west below Ben Lawers follow part of a route leading between opposite sides of the country (Stewart Reference Stewart1958; RCAHMS 2005).

Relationships with megalithic art

The distribution of pecked spirals is particularly informative as they were a prominent feature of megalithic tombs, although they should not be treated in isolation. It was studied by Van Hoek 30 years ago (Van Hoek Reference Van Hoek1993). There have been new discoveries but the basic picture has not changed (Frodsham Reference Frodsham and Frodsham1996; Brown & Brown Reference Brown and Brown2008, 128–31; Freedman in Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 284–311). The main concentrations are near the Scottish coast with others in Northumberland. There are possible examples in north Wales and more are distributed around the shores of Ireland (apart from those in the south-west they rarely face the Atlantic). Few are far inland. The motifs themselves take two forms. One group resembles those associated with Irish passage graves but the other compares with designs in monuments of the same kind in Orkney. In each case their relationship with the sea must have been significant.

Spirals are not the only distinctive designs and there are more panels that combine features of passage grave art with the normal repertoire of open-air petroglyphs (Freedman in Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 284–311). Taken together, their distributions identify two networks. One extended along the Irish Sea and is evidenced by specialised motifs shared with chambered tombs on both sides of the water. The other followed the North Sea and was connected to Orkney. Motifs associated with Orcadian monuments can be found on the west coast as far south as Anglesey and down the east coast to the North York Moors. These axes were connected by overland routes associated with concentrations of rock art. How should these relationships be understood? By making specialised designs, local communities might have acknowledged a shared system of belief. Otherwise, travellers could have displayed their wider connections, employing unusual motifs associated with monuments in distant areas. There are useful analogies with the Bronze Age rock art of south Scandinavia. Some of the drawings of boats that cluster along the seashore are associated with depictions of artefacts found in other parts of Europe (Ling & Uhnér Reference Ling and Uhnér2014).

Activities on the open-air sites

Were monuments used in the same ways as concentrations of petroglyphs or was there a clear distinction between them? This section discusses the information from open-air sites, followed by an account of elements they shared with decorated tombs.

One feature which has been recognised at sites with rock art is the special importance of quartz. It was used to make artefacts, including hammers for working the stone, but it was also represented by broken fragments which were never used (Bradley & Watson Reference Bradley and Watson2019). People could have exploited the special properties of this material – it emits sparks when it is broken and glows when pieces are rubbed together (Whitley et al. Reference Whitley, Dorn, Simon, Rechtman and Whitley1999). It helped to give these places the animated quality emphasised by Jones and his colleagues in their study of Kilmartin. There were concentrations of pieces of quartz around decorated surfaces at Torbhlaren in western Scotland (Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011) and at Urlar in the southern Highlands (Bradley & Watson Reference Bradley and Watson2019). Similar deposits have been found by excavation beside the rock art on other sites in Scotland: Allt Coire Phadairlidh on the Ben Lawers estate (Bradley et al. Reference Bradley, Watson and Anderson-Whymark2013) and Upper Gaskan in the same region (Ballin Reference Ballin2019). Unmodified quartz pebbles were also discovered at Hunterheugh in Northumberland but were interpreted as natural deposits (Waddington Reference Waddington2005). At Urlar some of the quartz flakes had complex histories. They were quarried from veins exposed in the surface of the outcrop and returned to the ground without any indications of use (Bradley & Watson Reference Bradley and Watson2019). A similar procedure was followed at Copt Howe in Cumbria but here it involved fragments of tuff. They were removed by flaking one face of a prominent boulder and set vertically within a rubble platform beneath a decorated panel (Bradley et al. Reference Bradley, Watson and Style2019).

Excavated features around the decorated rocks

In Scotland excavation has focused on the areas around the decorated surfaces. At Ormaig, in the Kilmartin complex, there were few artefacts and structures were entirely absent (Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 204–19), but that did not apply to two outcrops associated with simpler motifs on the nearby site of Torbhlaren; similar deposits and features were identified beside the petroglyphs at Allt Coire Phadairlidh in the southern Highlands (Bradley et al. Reference Bradley, Watson and Anderson-Whymark2013). The clearest evidence comes from Tiger Rock at Torbhlaren. Here there was a small timber structure: a roughly circular setting of stakes or small posts with a diameter of 1.5 m, associated with a radiocarbon date of 2580–2340 cal bc (Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 52–6). Jones and his co-authors suggest that it was replaced by a clay platform bounded by a stone wall but this claim is more controversial. The wall itself is undated and a study based on soil micromorphology did not support the excavators’ interpretation. On the other hand, it identified a trampled surface at the base of the outcrop where there were quantities of worked and broken quartz. The evidence from Allt Coire Phadairlidh on the Ben Lawers estate is rather similar. Again, there were concentrations of quartz around a series of decorated outcrops and, in one case, an area of boggy ground beside a complex petroglyph was consolidated by a layer of cobbles. This cannot be described as a platform but it was in the best position for people to view the pecked design (Bradley et al. Reference Bradley, Watson and Anderson-Whymark2013).

The excavated evidence from England is even more limited, but at Backstone Beck on Rombalds Moor four decorated surfaces were identified during the fieldwork on a later prehistoric enclosure; their presence on the same site seems to have been coincidental (Edwards & Bradley Reference Edwards, Bradley, Cleal and MacSween1999). The pecked designs were very simple and, in two cases, areas of burning, possibly the sites of hearths, were found beside them. Both were associated with notable concentrations of artefacts. In one instance pottery was still preserved. Apart from a sherd of Beaker, it consisted of Durrington Walls style Grooved Ware (Manby Reference Manby, Cleal and MacSween1999, 73). Both areas contained remarkable collections of worked flint, much of which was introduced from the North Sea coast 90 km away. It included oblique arrowheads, Levallois cores, and fragments of polished knives. The composition of these groups recalls the contents of Late Neolithic features on the Yorkshire Wolds, some of which include the remains of feasts (Manby Reference Manby1974; Rowley-Conwy & Owen Reference Rowley-Conwy and Owen2011).

Deposits within the decorated rock

The parts of the stone selected for carving are equally significant. Artefacts were inserted in natural features on top of the decorated outcrops. This must have been intentional as fieldwork at Kilmartin in western Scotland showed that the people who created the petroglyphs were reluctant to work smooth surfaces and preferred those with basins, cracks, and fissures (Jones et al. Reference Jones, Freedman, O’Connor and Lamdin-Whymark2011, 18–32). The artefacts deposited inside these features included tools that could have been used to make the pecked designs. That also happened at Urlar in Strathtay (Bradley & Watson Reference Bradley and Watson2019). The platform at Copt Howe in Cumberland contained a cache of artefacts of this kind (Bradley et al. Reference Bradley, Watson and Style2019). Perhaps the quarrying of decorated panels at sites like Hunterheugh and Auchentorlie followed similar principles. Although it refreshed the decorated surfaces, it suggests a particular concern with the interior of the stone.

The same idea could account for the distinctive character of open-air rock art where the commonest elements include a circular design with a radial line extending to its centre. It could represent a tunnel leading into the rock itself. In that case its outer surface was only a membrane through which it was possible to pass. There are obvious dangers in comparing insular motifs with those in other parts of the world but two features do seem to be especially common. One is the idea that people or other beings can travel between the outer world and the interior of the stone. This idea is well documented in regions as far apart as Finland (Lahelma Reference Lahelma2008) and southern Africa (Lewis-Williams & Dowson Reference Lewis-Williams and Dowson1990). At the same time openings in the decorated surface can be associated with deposits with special properties. An account of Californian rock art describes how pieces of quartz were placed in fissures and cracks like those associated with rock art in Scotland (Whitley et al. Reference Whitley, Dorn, Simon, Rechtman and Whitley1999).

If the interior of the stone was important, so was the space above it. Quartz reflects sunlight and moonlight in a way that few materials can and the siting of rock art in the landscape echoes this concern. There was a preference for horizontal surfaces that faced the sky and a few decorated outcrops were directed towards the rising or setting sun. In most regions there were views towards the solar arc rather than the north (Barnett et al. Reference Barnett, Valdez-Tullett, Bjerketvedt, Frodsham and Sharpe2022a; Reference Barnett, Valdez-Tullett, Hoole, Jeffrey, Robin, Bjerketvedt, Alexander, Mazel and Nash2022b). The importance of the sun is most obvious at Copt Howe and Ballochmyle where long distance visibility is limited by a vertical surface and the panels themselves are orientated on the midsummer and midwinter sunsets respectively (Bradley Reference Bradley, Frodsham and Sharpe2022a). In the same way at Allt Coire Phadairlidh in the southern Highlands a decorated outcrop commands a direct view of the east end of Loch Tay (Fig. 8). It is where the midsummer sun rises (Scott & McHardy Reference Scott and McHardy2020, 122). Similarly, the rock art at Achnabreck on the west coast of Scotland was ideally located for watching the passage of the winter sun over the island of Arran 45 km away (Watson Reference Watson, Frodsham and Sharpe2022). At a more general level, sites were selected that faced south-east or north-west: the directions of the rising or setting sun at the turning points of the year (Barnett et al. Reference Barnett, Valdez-Tullett, Bjerketvedt, Frodsham and Sharpe2022a; Reference Barnett, Valdez-Tullett, Hoole, Jeffrey, Robin, Bjerketvedt, Alexander, Mazel and Nash2022b). Those directions seem to have influenced the placing of rock art above Loch Tay on Ben Lawers, and in Kilmartin Glen (Bradley et al. Reference Bradley, Watson and Anderson-Whymark2013; Watson Reference Watson, Frodsham and Sharpe2022).

Fig. 8. Allt Coire Phadairlidh in the Southern Highlands of Scotland where a decorated outcrop commands a direct view of the east end of Loch Tay (photograph: Aaron Watson)

Activities shared with chambered tombs

The repertoire of pecked motifs is not the only connection between open-air rock art and passage graves. There was a subtle interplay between features that were concealed from view and those with a more public character. That applies to three distinct elements: the architecture of the monuments; the placing of megalithic art; and the deployment of quartz. The tombs themselves could admit few people but, during subsequent phases, additional structures were erected beyond the limits of the mounds or cairns. They included earthworks, monoliths, and post circles. There were also arrangements of small stones similar in size and outline to the timber setting at Torbhlaren (Eogan & Cleary Reference Eogan and Cleary2017, 216–46). The structural sequence at Knowth is particularly informative as the decorated kerb was among the last components to be built there. It was a massive undertaking and featured a series of conspicuous curvilinear designs which have more in common with open-air rock art than the other motifs on that site (Eogan & Shee Twohig Reference Eogan and Shee Twohig2022). There are even more connections between these media. Although some scholars question O’Kelly’s reconstruction of the façade of Newgrange (Cooney Reference Cooney2006), there is no doubt that there were significant deposits of quartz outside the passage graves at Knockroe, Knowth, and Loughcrew (Hensey & Shee Twohig Reference Hensey and Shee Twohig2017). In the Boyne Valley it featured in some of the stone settings around the principal structures (O’Kelly Reference O’Kelly1982, 75–8; Eogan & Cleary Reference Eogan and Cleary2017, 216–46).

Robin (Reference Robin2009) considers that the larger monuments had a distinctive layout (cf. Jackson Reference Jackson, Helskog and Olsen1995). They were organised as a series of concentric rings, only the latest of which was visible in the exterior of the monument. There was a chamber towards the centre, accessed by a long straight passage (Fig. 9). He interprets the rear of that chamber as the portal to another world where participants in ceremonies might seem to pass through the stone. The whole structure resembles one of the commonest designs in Neolithic rock art. But there is a most important distinction – nearly all the circular motifs associated with passage tombs lack the radial line found on open-air sites. There may be an explanation for this contrast. Where the design was executed in three dimensions and an actual passage was provided, rock art seldom referred to these features. On natural outcrops, however, passages leading into the stone were evoked by a schematic outline.

Fig. 9. The basic components of Irish and Orcadian passage graves in the scheme devised by Robin (Reference Robin2009) (drawing: Courtney Nimura)

In this way Irish tombs enshrined the same ideas as the panels of pecked decoration and both illustrate a concern with what was hidden underground. They also acknowledged the importance of the sky. This evidence is much better known. Nearly a fifth of the well-preserved passage tombs in Ireland were orientated towards the rising or setting sun at the solstices (Prendergast Reference Prendergast and Doyle2020). The same applies to examples in Orkney (Downes Reference Downes and Doyle2020) and Anglesey (Burrow Reference Burrow2010). Because they were solid constructions the evidence is much clearer than it was with open-air rock art but the concerns it illustrates seem to be the same. Certain directions were significant; indeed, the excavated tomb of Knockroe was directed towards both the midwinter sunrise and the midsummer sunset (O’ Sullivan Reference O’Sullivan1993; Prendergast Reference Prendergast and Doyle2020). These ideas are consistent with an interpretation suggested by Lewis-Williams and Pearce (Reference Lewis-Williams and Pearce2005) in their book Inside the Neolithic Mind. Perhaps these features referred to the widely shared idea that the cosmos comprised three layers: an upper world associated with the sun, the moon, and the sky; the familiar world of human affairs; and an underworld connected with the dead and the supernatural. It was represented in solid form by the structure of the passage tomb and in outline by open-air rock art. Both expressed the same concepts and chambered tombs and decorated outcrops might have been where people could communicate between the different spheres. Such practices continued after chambered tombs went out of fashion.