The Monumental Cemeteries of Northern Pictland

THE EMERGENCE OF FORMAL CEMETERIES is one of the most significant transformations in the landscapes of 1st millennium ad Scotland. In eastern and northern Scotland, in the lands of the Picts, square and circular burial monuments were constructed to commemorate a small proportion of the population — perhaps a newly emerging elite in the post-Roman centuries. This paper presents the results of a project that has consolidated and reviewed the evidence for monumental cemeteries of the northern Picts from Aberdeenshire to Inverness-shire, transcribing the aerial evidence of many sites for the first time. In addition, the landscape location of the cemeteries is assessed, along with their relation to Pictish symbol stones, fortified sites and settlement landscapes of the 1st millennium ad. Two particular elements of the burial architecture of northern Pictland are highlighted — barrow enlargement, and the linking of barrows through the sharing of barrow/cairn ditches. Both of these practices are suggested here to be implicated in the creation of genealogies of the living and the dead during an important transitional period in northern Europe when hereditary aristocracies became more prominent.


INTRODUCTION
The archaeological evidence for early medieval burial traditions in Scotland has increased dramatically over the past few decades, with newly excavated sites adding to the corpus of upstanding and previously excavated sites and those revealed by aerial photography. 3 Recent reviews have critically examined the timings and tempos of burial practices in the post-Roman period and added significantly to debates on Christianisation and the extent to which burial architecture was implicated in the important social transformations that occurred during the 1st millennium ad. 4 Despite this, a large proportion of analyses have remained at the site or synthetic level. As a result, the regional evidence largely remains uncharted, although there are exceptions, 5 and previous studies have not taken into account the rich aerial photograph archives available. This is true particularly in northern Scotland, where many early medieval cemeteries are known from aerial photography alone. These sites have rarely been included in the broader debates on the character and form of post-Roman mortuary practices. This article concentrates on the monumental cemeteries of northern Scotland: square and round barrows and cairns. Aerial photographic and limited excavation evidence is also reviewed for a key region in northern Pictland.
This study focuses on the regions of Aberdeenshire, Moray and Inverness-shire (Fig 1). These modern local government areas 6 loosely cover the heartlands of the regions of Fortriu 6 Inverness-shire is the south-eastern region of the Highland Council local government area. and Cé of northern Pictland. 7 Recent scholarship has suggested northern Pictland was a key region in the development of society in northern Britain in the post-Roman period. Alex Woolf has identified Fortriu, the Pictish kingdom most commonly cited in contemporary sources, as located in the Moray Firth region rather than in central Scotland as was assumed from the 19th century onwards. 8 Historical sources suggest that by the end of the 7th century, Fortriu had established hegemony over most, if not all, of the other Pictish areas and this endured until the 9th century. 9 Archaeological evidence underlines the importance of the region: Class I symbol stones are concentrated here; 10 an early Pictish high-status settlement has recently been discovered at Rhynie, Aberdeenshire; 11 and a series of early ringforts recently dated to the 5th and 6th centuries has been recognised in the same area. 12 This has created renewed interest in areas of northern Pictland that at times have been considered peripheral to the major social and political transformations in northern Britain. 13

PREVIOUS WORK ON PICTISH MONUMENTAL CEMETERIES
In Wainwright's classic 1955 edited volume, The Problem of the Picts, the lack of any conclusively Pictish examples of cemeteries meant that the discussion of burial traditions was limited. 14 Nonetheless, in the same decade, the identification of long cist cemeteries in the Lothians and Fife led to the first characterisation of early medieval burial traditions in eastern Scotland. 15 Recognition of monumental cemeteries of this period followed in the 1960s with the excavations of long cist burials in association with square and round cairns at Lundin Links, Fife. 16 However, the biggest advances in our knowledge of burial traditions in Pictland came with the onset of the aerial survey programme by the Royal Commission for Ancient and Historical Monuments of Scotland (RCAHMS) in 1976, 17 alongside more geographically focused survey programmes by local government Sites and Monuments Record (SMR) teams. 18 Aerial reconnaissance during the hot summers of the mid-1970s revealed hundreds of previously unknown sites as cropmarks, especially in the rich arable zones along the coastlines of eastern and northern Scotland. These surveys identified an entirely new burial type -the square-ditched barrow with central grave. 19 Upstanding monuments were 7 Fortriu is a difficult term to define and appears to have referred to both a territory and to the Pictish overkingship. As a territory it almost certainly included the area of Moray, but may have extended as far north as the Black Isle and other areas of Ross and Inverness-shire: Woolf 2006, 192. It may have encompassed the later medieval bishoprics and earldoms of Moray and Ross: Evans 2014, 68. Cé is more problematic -it only survives in one place name, Bennachie in the Garioch, and is likely to have included significant parts of modern Aberdeenshire : Dobbs 1949;RCAHMS 2007, 116. 8 Woolf 2006;2007. 9 Evans 2014, 58. 10 As noted by Henderson 1958, but largely ignored since; see distribution map in RCAHMS 2008, 11;Woolf 2006. 11 Noble et al 2013. 12 Cook 2011. 13 Fraser 2009, 109;Henderson 1958, 55;RCAHMS 2007, 115-16. 14 Wainwright 1955b, 94-6. 15 Henshall 1956. 16 See Greig et al 2000. Reviews of the antiquarian literature also identified other examples, such as Ackergill, Caithness, where excavation in the 1920s has identified a complex of square and circular cairns: Close-Brooks 1984;Ritchie 2011. 17 Maxwell 1978. 18 For example, the work of Ian Ralston, Ian Shepherd and Moira Greig as part of the Aberdeen Archaeological Surveys and Aberdeenshire Council Archaeology Service aerial reconnaissance programme. 19 RCAHMS 1978, 9-10. also recorded and small-scale excavation provided further characterisation of a monumental Pictish cemetery tradition.
In the 1980s Pictish Studies: Settlement, Burial and Art in Dark Age Northern Britain gathered together excavations and surveys of many important sites for the first time. 20 Synthetic reviews appeared in the same decade, summarising a growing body of evidence, with the long cist, the ditched barrow and the platform cairn identified as the three main categories of early medieval burial evidence in Scotland. 21 The square and round cairns and barrows were frequently identified as a distinctive Pictish burial tradition; however, similar cemetery and burial types were also identified in the Northern and Western Isles, speculatively in Dumfries and Galloway, and also in Wales and western and eastern England. 22 Thus, the monumental cemeteries located in Pictland were identifiable as part of broader British and Irish traditions.
Renewed interest in the monumental cemeteries of Pictland has been stimulated by the publication of past excavations and new research-led and development-led excavations. 23 Together the evidence reveals that Pictish cemeteries vary in construction, size and form, but certain shared traits appear across all cemeteries. 24 Where upstanding examples survive, they are generally low and flat topped, surrounded by either a stone kerb or a ditch and occasionally an outer bank. While round or square in plan, some display a degree of variation including ovoid, rectangular and trapezoidal forms. 25 The ditched examples vary, with some bounded by a continuous ditch and others with broken ditches and in some cases, stone boulders or slabs are located at the corners, for example at Garbeg (Inverness-shire), Unst (Shetland), and Ackergill (Caithness). 26 Particular materials seem to have been significant and the construction methods suggest time and care went into monument creation. 27 Low-lying agricultural land was favoured for cemeteries and burials, with many on low terraces. 28 The majority of barrows cover single graves, but cairns have also been shown to cover groups of individuals (as at Lundin Links, Fife and Ackergill). At Lundin Links skeletal analysis suggests that clusters of interments in particular monuments may be family groups. 29 The orientation of graves is generally E/W with some variability within cemeteries and between sites. 30 Conclusive dating of the cemeteries has been problematic, as lack of excavation and poor bone survival has limited the available material. Only two larger monumental cemeteries have been excavated to any extent, Redcastle, Angus, and Lundin Links neither in the present study area. Radiocarbon dating from both suggests burial activity took place in the 5th to 7th centuries ad (see below). 31

METHODOLOGIES AND CONTEXT
Aberdeenshire, Moray and Inverness-shire have, to date, seen limited close attention. Few of the sites have been excavated and most are documented only through aerial 20 Friell andWatson 1984. 21 Ashmore 1980;Close-Brooks 1984. 22 Ashmore 1980Close-Brooks 1984;Cowley 1996;Longley 2009, 113-15;O'Brien 1999;2009, 148. 23 Eg Dunbar 2012. 24 Winlow 2011. 25 Ashmore 1980. 26 Wedderburn and Grime 1984;Bigelow 1984;Ritchie 2011. 27 Alexander 2005Bigelow 1984, 115;Ritchie 2011, 136;Winlow 2011, 357. 28 Winlow 2011, 346. 29 Greig et al 2000, 603;Williams 2007. 30 Winlow 2011, 343. 31 Maldonado 2013 photographs. The current study was carried out as part of the University of Aberdeen's Northern Picts project, directed by the second author. The Northern Picts project was established in 2012 to investigate the archaeology of Aberdeenshire to Easter Ross, covering the probable extent of the Pictish provinces/kingdoms known as Fortriu and Cé. Major fieldwork, as part of this project, has investigated high-status settlement and ritual sites, the location of a major silver hoard and other sites of early medieval date. 32 A Geographic Information System (GIS) has also been constructed that documents all known sites of early medieval date within the study area, and has contributed to the results presented here. This study also draws on the completed MSc thesis and current PhD research at the University of Aberdeen by the first author. The latter reviewed evidence for monumental cemeteries using the online Canmore database and archive resources of Historic Environment Scotland (HES), Highland Council's Historic Environment Record (HER) and Aberdeenshire and Moray Council's online Sites and Monuments Records (SMR). The categorisation of each site was confirmed through desk-based analysis, and the transcription of individual sites was conducted where possible. 33 Numbers of barrows and cairns were noted and a basic categorisation of cemetery size was conducted. 34 Aerial reconnaissance has played a large role in the success of this study, but of course the aerial record can only ever be a partial record of the original extent and distribution of the archaeology for a particular region. 35

THE PICTISH CEMETERIES OF ABERDEENSHIRE, MORAY AND HIGHLAND
Of the 57 cemeteries recorded in the HES Canmore database, Aberdeenshire and Moray SMRs and Highland HER for the study area, 27 were considered 'probable' or 'confirmed' (Fig 2; Tab 1). It is these cemeteries that form the basis of detailed comparative study below. Eight of the remaining sites were deemed 'unlikely' and the rest, a total of 22, were recorded as 'possible'. 'Possible' cemeteries are classed here as having only one potential barrow with unclear definition. ABERDEENSHIRE There are fewer sites and a smaller numbers of barrows in Aberdeenshire than in the other two local council areas, but this may be due to the generally poorer definition of cropmarks in Aberdeenshire (Tab 1). 36 Here, sites range from a single barrow at Boynds, to medium-sized cemeteries such as Hills of Boyndie. The latter sits on a high plateau in the parish of Banff (Fig 3). Two square barrows and at least four circular barrows are visible 32 Eg Noble et al 2013;Noble et al 2016. 33 Images were rectified and georeferenced using VectorMap Local, an Ordnance Survey basemap. VectorMap local grid squares for sites in Inverness-shire, Moray and Aberdeenshire were downloaded via Edina Digimap Ordnance Survey Service <http://digimap.edina.ac.uk> [accessed during the months of May to August 2014]. Detailed recordings of the cropmarks identified from aerial photographs were rectified using Aerial 5.10 and transcribed using ArcGIS 10.2. Final images were produced using Adobe Illustrator. 34 Cemetery size was identified as follows: major 11+ barrows; medium 7-10; small 2-6; and single barrows. These categories broadly follow that of Winlow 2011 andHenshall 1956. 35 It is important to note that the distribution of such sites is heavily dependent on patterns of modern land-use, rainfall and the character of soils and drift geology. Thus, across Scotland, cropmark sites are concentrated in areas set to arable, on well-drained soils where crop stress is more likely, and in areas such as the eastern coast where average rainfall tends to be lower. This broad pattern can be illustrated at a very local level, for example, at Mains of Garten (Fig 7), where the visibility of cropmarks depends on local variations in soil depth. 36 Fraser and Halliday 2011, 312. as cropmarks, closely clustered together. None of the cropmark sites in Aberdeenshire have been excavated. One square cairn and two possible cairns were identified in the 1970s at Tillytarmont and two square barrows have been excavated at Rhynie (see below).

MORAY
The barrow cemeteries of Moray are relatively small compared to examples in Invernessshire (Tab 1). They include single barrows at Kinloss Airfield and North Alves, and small cemeteries at Lower Auchenreath, Wester Coltfield and Midtown. Greshop Farm, near the River Findhorn, has complex cropmarking with square barrows evident to the south-east of a stretch of the river and settlement remains and/or further barrows to the south-west. Three of the square barrows at Greshop were excavated in advance of flood prevention works, including one barrow that was more than twice the size of the others (Fig 4). 37 No human remains were found due to acidic soils. 38 The largest and most complex cemetery identified in Moray is Pitgaveny, located close to the former Spynie Loch (Fig 5). Here, the remains of at least 10 square barrows and around five circular barrows are arranged in linear alignment. 37 Dunbar 2012. 38 Ibid.  The barrows are very large, up to 20 m across, and at least four of the barrows appear to enclose smaller square-ditched features, possibly barrows. A further double square-ditched enclosure survives at Wester Buthill, along with two smaller barrows.

INVERNESS-SHIRE
The smaller cemeteries known from cropmarks in Inverness-shire include sites such as Kerrowaid with two barrows and Allanfearn with seven (Tab 1). An intriguing example is Kinchyle, where a large double-ditched square enclosure with a central, possible grave, pit, is clearly visible as a cropmark alongside a dense array of features, many of which may be prehistoric settlement features (Fig 6). Slightly larger cemeteries in Inverness-shire are represented by sites such as Mains of Garten, located at a bend in the River Spey, comprising a cemetery of around 20 round and square barrows. Here, the geology and topography may obscure more examples (Fig 7). At Poyntzfield on the Black Isle, a barrow cemetery of perhaps six barrows is evident, including round and square examples, running in a linear arrangement along a low terrace. The most impressive cropmark sites in this area are Croftgowan and Tarradale. The cemetery at Croftgowan comprises a linear setting of around 27 circular and square barrows located on the slope of Tor Alvie (Fig 8). Areas of deeper soil may obscure more examples at Croftgowan. Tarradale House is located on a terrace overlooking the Beauly Firth (Fig 9). Around 18 circular barrows, the largest 10-12 m in diameter, and eight square barrows each measuring around 5-6 m across, are evident, with two larger enclosures and a trackway.
In addition, we are fortunate in Inverness-shire that a small number of upstanding cemeteries have also survived. The largest is Garbeg, where a total of 26 upstanding barrows have been recorded, including 14 square or sub-rectangular and 10 circular (Fig 10). 39 Here the square barrows range from 3.5 to 5.5 m across, and the round mounds are up to 10 m in diameter and nearly all have surrounding ditches. The mounds are of three types: flattopped stony mounds which rise to a height of around 0.5 m, earthen mounds which stand to a height of c 0.2 m, and low stony mounds separated from their ditches by a berm. Four 39 Wedderburn and Grime 1984. Table 1 Continued barrows were excavated in 1979 following the discovery of a fragment of a Pictish symbol stone associated with one of the round barrows, although no relationship was conclusively proven between the stone and the burial monuments. 40 All examples overlay central inhumations, though the human remains were very badly preserved or non-existent in all cases.

No of barrows
The second largest upstanding cemetery in Inverness-shire is at Whitebridge, situated on a ridge at the confluence of the River Fechlin. The cemetery consists of at least seven burial mounds and five small round cairns. The monuments are 4.5-9 m in diameter, and stand up to 1 m in height. One of the mounds has been excavated, but no human remains survived. 41 Two smaller upstanding cemeteries also survive -Brin School and Pityoulish. Brin School sits on the eastern bank of the River Nairn on a low ridge and comprises five barrows. These two cemeteries contain both square and circular mounds up to 10 m in diameter and up to 0.6 m in height. At Pityoulish, monoliths project from three of the barrows. At this cemetery, one barrow was excavated in 1953 and human remains were recovered from a pit adjacent to the standing stone. 42

GENERAL CHARACTERISTICS
The aerial photographic evidence shows that most monumental cemeteries identified in the study area were relatively small in scale: 59% of the probable or confirmed examples contain six or less graves. This corresponds to both Winlow's and Henshall's analyses of cemeteries further south in Pictland, which suggest that many cemeteries were composed of one to six burials. 43 However, the aerial evidence is likely to provide an underestimate of 40 Stevenson 1984;Wedderburn andGrime 1984. 41 Stevenson 1984;Alexander 2000. 42 Rae andRae 1953. 43 Henshall 1956;Winlow 2011, 341. the total number of burials. Many cemeteries undoubtedly still await discovery; the barrows excavated at Rhynie, Aberdeenshire, for example, were not visible on aerial photographs despite the presence of other archaeological cropmarks in the same field (see below). Some cemeteries are also likely to be more extensive than they first appear. For example, at Kinchyle, Inverness-shire, four large square enclosures or barrows and four possible round barrows lie just over 800 m away at Little Kildrummie (Tab 1). These barrows could represent additional elements of a very large cemetery, or a nearby related site. The larger cemeteries in the study area (Garbeg, Pitgaveny, Croftgowan, Mains of Garten, Tarradale) all have upwards of 11 barrows (Fig 11). They display more variety in their size, shape and architectural construction, which could be suggestive of their importance, their longevity, or both. The monumental cemeteries identified in the study area may also include unenclosed graves, exemplified in the south at sites such as Forteviot, Perthshire, and Redcastle, Angus. 44 At Redcastle, for 44 Campbell and Maldonado forthcoming; Alexander 2005. Findhorn. A large square barrow with two adjacent smaller barrows were visible as cropmarks and were confirmed by excavation (Dunbar 2012). Further cropmark evidence suggests a barrow to the east, another to the north, and a cropmark group to the west comprising what appear to be settlement features, but may also include barrows. Illustration by Juliette Mitchell. Base map © Crown Copyright/database right 2016. An Ordnance Survey/ EDINA supplied service.
example, excavation identified at least seven unenclosed graves without a defining cairn or ditch located near to the square and round barrows of the cemetery. 45 There is little evidence of unenclosed graves in the aerial photographic evidence for our study area, but future excavation may well identify them. Long cist cemeteries may also represent an aspect of Pictish burial tradition. However, although in Tayside and Fife around 90 examples are known, 46 in the study area long cist cemeteries are rare -examples are restricted to a handful of sketchy antiquarian accounts and a small number found during development, but all are undated and unpublished. They are largely restricted to the southernmost part of Aberdeenshire and include five long cists found at two different locations in Stonehaven, 47 another two at Johnshaven, Kincardineshire, 48 and a single inhumation at Inverbervie. 49 The lack of long cist cemeteries in the area may represent a real difference in the burial traditions of northern Pictland, or it could reflect a lack of defined detail in the cropmark evidence. Certainly nothing comparable to the long cist cemetery of Hallow Hill, Fife, where at least 150 inhumations were found, has been identified in the study area, though Hallow Hill is also exceptional for Tayside and Fife. In Tayside and Fife over half of the identified long cist cemeteries contain less than five graves. 50 This is comparable to the size of most monumental cemeteries in the study area under discussion here. 45 Ibid, 94. 46 Winlow 2011, 344. 47 RCAHMS 1984, 17, nos 82 and 83. 48 RCHAMS 1982 Ibid, no 134. The remains are held at the University Museums Service, University of Aberdeen. 50 Winlow 2011, 344-6. The general trend, north of the Forth, is for small clusters of graves rather than extensive cemeteries. 51 As well as number of monuments, other characteristics mark some cemeteries out. The majority of barrows and cairns conform to the sizes and types known elsewhere. The barrows are between 4 m and 12 m in diameter, and most lie towards the smaller end of that spectrum. Yet some sites contain barrows that are much larger than the norm, including examples of square and round barrows up to 25 m across. Examples of cemeteries with large barrows include Greshop, Pitgaveny, Wester Buthill (all Moray), Kinchyle, Inverness-shire and Hills of Boyndie, Aberdeenshire. In three of these cases oversized square barrows appear to have been enlargements of initially smaller monuments (Greshop, Pitgaveny, Kinchyle). Tarradale (Inverness-shire) also has larger circular barrows of 20 m diameter, but as yet with no evidence of multiphased development. 51 Eg Maldonado 2013, 9-11 andDunbar andMaldonado 2012. FIG 6 Kinchyle, Inverness-shire. An extensive group of cropmarks was recorded by RCHAMS in the 1970s. This includes a double-ditched square enclosure or barrow with a central feature, possibly a grave cut. There are also a number of circular ditch features measuring 5 m to 9 m in diameter. Some of these may be barrows, but others appear to be later prehistoric houses. The field also has numerous other features including pits and ovoid features.

RELATIONS TO TOPOGRAPHY AND ROUTEWAYS
Aerial photographic analysis of the landscape location of cemeteries in the study area reveals that cemeteries often form linear distributions that follow topographical features. These include areas of higher ground and rivers, similar to patterns observed in Tayside and Fife. 52 It is possible that these linear distributions may reflect routeways through the landscape. At Croftgowan (Inverness-shire) for example, a major cemetery of 27 barrows follows a distinctive linear arrangement. This runs almost parallel to the route north from Stirling to Inverness where the traversable land narrows through the Cairngorm Mountains. Likewise, at Whitebridge, Inverness-shire, a cemetery of at least 12 upstanding monuments is located just north of the road that leads from Fort Augustus to Inverness and adjacent to a number of 18th-century and modern river crossings that include a natural ford. Dalbreck, Aberdeenshire, is also adjacent to a routeway that leads to a ford across the Water of Feugh. The site is also strategically located at the northern end of the route that leads across the Cairn O'Mount, a mountainous pass through the Mounth, an eastward projection of the Cairngorms. 53 At Pitgaveny, Moray, the linear layout again leads to a ford across the River 52 Winlow 2011. 53 Small 1974 argues that this was an important routeway in the early medieval period. At Cairn O'Mount, a cross slab has been found at St Ringan's (NMRS: NO67NE 15). Lossie and there are also archaeological traces of an old road or routeway here. 54 Garbeg, Inverness-shire, may also sit on an old route through the highlands that leads from the mouth of River Enrick at Loch Ness northwards to Beauly and Inverness. 55 Greshop, Moray, is located just off the modern road from Forres to Nairn which crosses the River Findhorn, the same route followed on the earliest detailed maps for the area. 56 Mains of Garten, Invernessshire, may also sit at a river crossing, in this case over the River Spey, near the crossing at 54 NMRS: NJ26NW 51. 55 The modern route to Inverness was dynamited along the western side of Loch Ness. 56   Boat of Garten where a ferry crossing, now replaced by a bridge, was located. On the other side of the river from the Mains of Garten cemetery is a Pictish Class I symbol stone, found at Lynchurn. 57 Other examples of cemeteries show a more clustered layout, but some of 57 RCAHMS 2008, 74. these too may also relate to important land routes. The large barrow cemetery at Tarradale, Inverness-shire, for example, is dissected by a holloway that leads southwards to a landing place on the Beauly Firth (Fig 9). 58

CEMETERIES AND SETTLEMENTS
Likely contemporary sites in the surrounding landscape include settlements and fortified enclosures. Unfortunately, evidence for Pictish settlement is still rare across much of mainland Scotland. In the lowlands there is an absence of settlement remains after the Iron-Age roundhouse tradition ceased, sometime around the 2nd or 3rd centuries ad. 59 In the Northern and Western Isles, we have a richer settlement record that can include a variety of different house 'types', including oval, sub-rectangular, multicellular, semi-subterranean and figureof-eight structures. 60 The few Pictish settlement types known on mainland Scotland include structures with sunken floors at Easter Kinnear in Fife, 61 and byre-houses in the uplands known as Pitcarmick-type buildings. 62 In the study area virtually no unenclosed Pictish settlement evidence is known. The majority of the area is fertile, but intensively cultivated, land. Major agricultural improvements 58 Gregory and Jones 2001, illus 1;Yeoman 1988, 131, no 92. 59 Hunter 2007, 49. 60 Ralston 1997. 61  For these reasons, it is difficult at present to situate the cemeteries within the settlement landscapes of the same period. However, in the upland landscapes towards the Great Glen in Inverness-shire better preservation exists, presenting an opportunity to study landscapes that have escaped the agricultural improvements and modern settlement expansion of the last few centuries. Three upstanding cemeteries are found along the Great Glen. The landscape in the wider environs preserves upstanding archaeological sites including hut circles and platforms, hillforts, and cairns. Fig 12 shows all possible prehistoric to early medieval settlement remains within 5 km of the cemeteries at Garbeg, Whitebridge and Brin School. 67 While none of the settlement remains are dated, and many undoubtedly date to the Bronze or Iron Age, their distribution demonstrates that the cemeteries are located in areas that were densely settled in prehistory. The most intriguing evidence for the potential juxtaposition of cemetery and settlement is at Garbeg, where oval and sub-rectangular houses, types that have parallels dated to the Pictish period in the uplands of Perthshire, are located a short distance away from the cemetery (Fig 13). 68 Barrow cemeteries were, it seems, located in prime agricultural areas, with some degree of importance placed on their accessibility to main routes across the landscape. 69

CEMETERIES AND FORTS
More fortified sites in the study region have been dated to the 1st millennium ad than settlements, though the numbers are still small. As part of the Northern Picts project the data gathered were used in GIS to identify forts, duns or fortified sites lying within 5 km of cemeteries. Seventeen possible sites were identified; however, only two of those forts have confirmed Pictish phases of construction and use. One is the largest fort known in Pictland -Burghead, Moray, which is within 5 km of Wester Buthill, a site with a large square enclosure and three possible barrows. Wester Buthill is located near to one of the modern routes into Burghead from inland areas to the south. Garbeg, Inverness-shire, is located around 4 km north-east of Urquhart Castle. Urquhart Castle, excavated by Leslie Alcock, lies at the head of the Great Glen. 70 Here, a medieval masonry castle overlay a site that Alcock suggested could 63 RCAHMS 2007, 17-24. 64 Cook 2003. 65 Watkins 1981. 66 Cook and Dunbar 2008, 149-59. 67 Around half of the sites that have been surveyed are more ovoid and sub-rectangular than circular, which may suggest an early medieval or later date. 68 NMRS NH53SW 11; Carver et al 2012. The University of Aberdeen Northern Picts project is currently undertaking geophysical survey and evaluative excavation at a number of these house sites. 69 Reynolds (2002, 186) highlights the location of many Anglo-Saxon cemeteries within or near major settlements. The occurrence of 'pit' place names which contain the place-name element pett, 'a piece of land', as found in Pitgaveny, may further corroborate these as settlements, although the age of these names is in some doubt -see Taylor (2011) who argues that the place names relate to the expansion of Gaelic in the 10th century, but see also Evans (2014, 33-7) who argues that these names are older. 70 Alcock and Alcock 1992. have been the 6th-century fort (munitio) of Bridei, son of Mailcon, referenced in Adomnán's Life of Columba. 71 The scale of excavation at Urquhart was very limited, but suggested the presence of a hillfort enclosure on the craggy rock-boss upon which the later medieval motte stands, with possible lower terrace enclosures. 72 Radiocarbon dating suggests activity began at Urquhart in the 5th or 6th centuries, and lasted into the early centuries of the 2nd millennium 71 Alcock 1981, 159-61. 72 Alcock and Alcock 1992, 260. ad. All other forts within the vicinity of cemeteries are undated, but some examples where cemeteries and forts are closely juxtaposed are worthy of further investigation. Brin School, Inverness-shire, for example, is overlooked by the hillfort, Creagan An Tuirc, 'the boar's rock'. 73 The old route adjacent to the cemetery also leads northwards towards Inverness where the spectacular Boar Stone at Knocknagael was located and another possible barrow cemetery. Croftgowan cemetery is located on the south-west slope of Tor Alvie, which has an undated fort on the summit. Other fortified sites may exist in the vicinity of Pictish barrow and cairn cemeteries. As outlined below, the barrows at Rhynie, Aberdeenshire, were located near to a contemporary fortified settlement and at Tarradale, Inverness-shire, excavations in 1991-1993 found a large ditched enclosure, palisade and internal features at a site just to the north-east of the barrow cemetery. 74 Pottery from one of the internal pits has been suggested as early medieval in date, but the enclosure itself remains undated. 75

CEMETERIES AND SYMBOLS
Even before Pictish burial traditions were fully identified, Wainwright postulated a relationship between symbol stones and burial. 76 Since then the relationship of Pictish symbol stones to burial has been widely discussed. 77 Of the 27 confirmed or probable sites in the study area, nine cemeteries have symbol stones or cross slabs within a distance of 5 km. However, few direct relationships can be identified and few are in close association with a cemetery. Fragments of a Class I symbol stone were found in 1974 at Garbeg, Inverness-shire, in association with one of the round cairns (Cairn 1). 78 However, the stone is incomplete and 73 Taylor with Markus 2012, 342, 520. 74 Gregory and Jones 2001, 242, 245, illus 1 and 3. 75 McGill 2001, 255-8. 76 Wainwright 1955b, 87-96. 77 Eg Close-Brooks 1980Clarke 2007, 27-31;Ritchie 2011, 133-4;Foster 2014, 64-5. 78 Wedderburn and Grime 1984, 151-2. no direct relationship with the cairn could be conclusively demonstrated. At Tillytarmont symbol stones have also been found in the close vicinity of potential burial monuments. 79 In 1975 Tony Woodham excavated a small square cairn (4.25 m by 4.25 m and 0.5 m high) made up of large waterworn boulders with evidence for a large central quartz monolith and identified two stone spreads that may have been other cairns. These were found in the same general area as five Pictish Class I symbol stones located on the haughland at Tillytarmont. 80 However, no burial was found at the cairn and no direct association can again be demonstrated. Relationships at other sites are suggestive. At Mains of Garten, for example, a symbol stone was ploughed up in a field near to the cemetery, but in this case the stone was located on the opposite side of the river.

Barrows at Rhynie
The most recent discovery of Pictish burial monuments in the study area is at Rhynie, Aberdeenshire, and this exemplifies a broad juxtaposition between barrows and symbol stone monuments. In 2013 excavations near the village of Rhynie uncovered two square barrows with central graves. Excavations in 2011-2012, less than 0.5 km to the south of the barrow locations, uncovered a fortified settlement of the 5th-6th centuries ad, defined by a plank-and post-built box rampart, inner and outer ditches and evidence for enclosed rectangular buildings. 81 Late-Roman amphorae (B ware), along with fragments of imported glass and high-status metalwork including evidence for production, suggest a high-status site. The archaeological evidence can be set alongside the place name, which derives from early Celtic (Pictish) rīg, 'king', with the overall name likely to mean 'place of or associated with a great king'. 82 Eight Pictish Class I symbol stones are known from Rhynie, and burials and human remains have been recorded in close association with some of these stones since the 19th century. Three of the symbol stones come from the vicinity of the fortified settlement, two from the modern church and three others were found towards the southern end of the village. In 1836 two of the stones from the village (Nos 2 and 3) were removed during the construction of a turnpike road. A 'quantity' of human bones was found near the stones at this time. James Logan also states that Rhynie No 3, which depicts a warrior with a spear, was found in association with a cairn. 83 Antiquarian reports and local newspapers also record the discovery of cists in the same general area. Isobel Henderson, 84 for example, records three parallel cists being found during the construction of Ashvale cottage in the village, and E/W orientated cists are reported being found near the warrior figure, Rhynie No 3. The two square barrows excavated in 2013 were found a short distance to the south of where the two symbol stones and reports of human remains and cist burials were identified. They were found in association with two larger square enclosures that had been located on aerial photographs (Fig 14). These two large square enclosures have short segments of ditch that project in front of an apparent entrance on the N side of each enclosure. The larger square enclosure measures around 20 m across and the smaller 16 m. The dating evidence for these larger enclosures is problematic, but dating of an upper fill of the ditch of the largest 79 See Gondek 2010 for a full discussion of the site and stones. 80 Woodham 1975, 6. 81 83 Logan 1829, 56. 84 Henderson 1907, 163. suggests it was still visible in the 7th century ad and a pit dating to the 5th-6th centuries ad was found inside. Each of the smaller square barrows measured 4 to 4.5 m across. In the centre of one, a stone-lined long cist containing the remains of an adult female was found. The second barrow contained traces of a wooden coffin in the central grave cut, but no surviving human remains. The central burials were aligned ENE/WSW and NE/SW respectively. Radiocarbon dates place these burials between ad 400-570 (see below), contemporary with the high-status settlement discovered in 2011.

DATING
The dates from Rhynie are the only scientific dates available for Pictish burial monuments in the study area. 85 Two radiocarbon dates for the individual in the cist burial indicate the square barrow dates to between cal ad 400-570 (at 95% probability) (SUERC-52935 1559 ± 30 BP, cal ad 420-570; MAMS-21252 1602 ± 24 BP, cal ad 400-540). 86 This corresponds well with the dating for the cairns and mounds of the two largest Pictish cemeteries excavated -Redcastle, Angus and Lundin Links, Fife (Fig 15; Tab 2). Square and round cairns may have been constructed earlier in the 1st millennium ad, 87 but the floruit of this tradition can be placed in the 5th-6th centuries ad and the tradition appears to have largely ceased by the 7th century. 88

DISCUSSION
The 5th-6th centuries ad, when the monumental cemeteries of Pictland flourished, are increasingly seen as a critical period in the formation of the early kingdoms of northern Britain and north-western Europe more generally. In north-eastern Scotland, at this time, fortified sites re-emerged after a hiatus in the later Iron Age. 89 Class I symbol stones appeared and flourished; perhaps associated with new forms of identity and place-making in the post-Roman era. 90 While there are examples of burial monuments and even small cemeteries in the earlier centuries of the 1st millennium ad, 91 the establishment of monumental cemeteries marks an important transition in the visibility of the dead in the archaeological record. 92 Across northern and eastern Scotland, from Shetland to the Firth of Forth, very similar burial monuments were constructed suggesting strong links between the dispersed communities of Pictland. 93 In Scotland generally, burial becomes much more visible in the 85 A date was obtained from human remains recovered from one of the cairns at Garbeg, Inverness-shire, but returned a very late date (11th-12th century ad) and recent resampling suggests contamination (Kate Britton pers comm.). 86 Radiocarbon dates presented in the text and in Table 2 are calibrated using the IntCal13 calibration curve (Reimer et al 2013) and the computer program OxCal v4.2 (Bronk Ramsey 2009), and are given at 95% probability range. 87 Eg Murray and Ralston 1997;Neighbour et al 2000. 88 With the exception of Forteviot where recent dating of grave fill and ditch fill suggests the construction of mounds into the 8th or 9th century ad, but note that skeletal remains were not recovered and the dates were on charcoal in the grave fill: Campbell and Maldonado forthcoming. 89  5th and 6th centuries. 94 The emergence of monumental cemeteries from the 5th century, and the dwindling occurrence of such monuments by the 7th century, mirrors patterns of change evident elsewhere in Britain and Ireland in this period. A shift towards church burial from the 8th century ad is also evident. What marks the Pictish monumental cemeteries out from other traditions in Scotland is the focus on the construction of elaborate earthen mounds and stone cairns to cover the dead. 94 Maldonado 2013, 1. The square and round barrows of Pictland can be paralleled with similar traditions of monumental graves across Britain and Ireland. 95 In Ireland, enclosed cemeteries were established for the first time in the 5th-6th centuries ad, marking a change from the sparser burial evidence of the Iron Age. 96 Some monumental or enclosed cemeteries were also created, including the construction of 'settlement-cemeteries' defined by ditches, ring ditches and cairns. 97 However, while monumental cemeteries and mounds occurred in Ireland, the focus on individuals, which is common to Pictish barrows, remains rare. In Ireland, the enclosing of groups of burials within mounds, cairns or enclosures was more common. Nonetheless, there are examples in Ireland of cemeteries that appear to cluster around founder barrows or graves, and a very small number of these graves were furnished with gravegoods. 98 In Wales 95 Longley 2009, 112-15;Maldonado 2013, 17. 96 O'Brien 2009, 136-8. 97 O'Sullivan et al 2014. 98 Bhreathnach 2014, 125-6. the most typical burials from the 5th century onwards were E/W orientated and unenclosed, but enclosed graves are also known, including square enclosures that were probably barrows. 99 The mound and cairn burials of Pictland, as identified in the study region, are overwhelmingly extended inhumations, unfurnished and generally orientated E/W. This arrangement had become widespread across the western Roman provinces by the mid-1st millennium ad and Pictish traditions seem to fit this more general shift in European mortuary practices. 100 The lack of gravegoods has been seen as a distinguishing factor in the burial traditions of northern Britain, but in Ireland and western Britain contemporary burials were also generally unfurnished. 101 The Pictish monumental cemeteries emerged prior to the elaborate princely burials of the late 6th and 7th centuries in Anglo-Saxon England, instead overlapping with the practice of a more modest barrow building tradition. 102 Early Anglo-Saxon barrows are not generally associated with rich assemblages of gravegoods; it is the investment in construction, time and material that demonstrates wealth or status. 103 Some contemporary Anglo-Saxon cremation cemeteries, in some instances, contained thousands of burials. Pictish cemeteries are more comparable with the Anglo-Saxon inhumation tradition, which involved the burial of smaller numbers of individuals with monumental markers such as barrows or ring ditches. 104 In Pictland, the small number of mounds or cairns found at most sites suggests that this was not a common rite. They were instead acts of selective remembrance, perhaps commemorating only certain individuals. The presence of barrow monuments at sites such as Rhynie (and at the later royal centre at Forteviot in southern Pictland) implies that at least some of these monuments were part of high-status cemeteries, the mounds and cairns built for the few rather than the many. 105 The chronological spread of mounds and cairns within individual cemeteries, with perhaps only a half dozen or so constructed over two or three centuries, also suggests a restricted, probably elite, practice 106 -episodes of construction that would have been memorable, creating powerful statements within the landscape. 107 In Anglo-Saxon England, the building of mounds was in some cases at least, an elite practice. Martin Carver, among others, has connected the practice of mound building with the emergence of powerful hereditary aristocracies. 108 Another connected practice is the reuse of antecedent prehistoric barrows as locations for burials. 109 Both traditions are argued to have signalled a new elite presence in Anglo-Saxon society, with the use and manipulation of both landscape and burial rites employed to make visible statements of real or perceived ancestry and underline claims to power in the present. 110 In Ireland, like Scotland, the lack of gravegoods has meant that questions of status or elite practice have not been addressed to the same level. Nonetheless, occasional gravegoods 99 Longley 2009, 113. 100 Eg Halsall 2012Hines andBayliss 2013, 553. 101 Carver 2005;Dickinson 2011, 230;O'Brien 2009, 145. 102 Carver 2005Dickinson 2011, 230;Geake 1992;Welch 2011, 269. 103 Scull 2009 Dickinson 2011, 229;O'Brien 2009, 137-48. And similar to the generally small size of cemeteries in Scotland in the early medieval period: Maldonado 2013, 10. 105 Campbell and Maldonado forthcoming. 106 Alexander 2005, 110. 107 Cf Mizoguchi 1993. 108 Carver 2002, 136. 109 The phenomenon of monument reuse and its connection to elite strategies has been studied in some depth in Anglo-Saxon contexts, cf Semple 2003;Williams 1997;2006. 110 See for example Lucy 1992;Halsall 2010;Semple 2003;Williams 1996;  Gleeson has gone further, arguing that the multiple burials found in a small number of Irish barrows (as occasionally found within Pictish mounds and cairns) may be the burials of kindred heads of ruling lineages. 113 High-status burials in Ireland also appear to have been inserted into prehistoric barrows, perhaps as a means of asserting the authority of particular lineages through reference to the ancestors or creating contrived lineages of the dead. 114 In Wales, references in the early praise poem Englynion y Beddau suggest that burial in mounds was an elite practice. 115 While similar forms of burial architecture are found throughout Pictland, generally there are no gravegoods distinguishing individual burials. Some monuments were made more prominent through the scale of the mounds or cairns. At Tarradale, Inverness-shire, for example, a series of round barrows were built on a much grander scale than the others, and the cemetery also contains an even larger circular enclosure, 36 m across its widest point -whether this too was a barrow, albeit greatly enlarged, remains to be seen (Fig 9). Likewise, at Hills of Boyndie, Aberdeenshire, a cluster of barrows was centred around a large square barrow, 14 m across. The other very revealing phenomenon, laid bare by the plough-truncated form of most examples, is the occurrence of more than one set of enclosing ditches, suggesting barrow elaboration or perhaps enlargement. At Greshop, Moray, a square barrow 10 m across was enclosed by a set of additional ditches, creating a monument 28 m in length/width. The more elaborate monument at Greshop was the largest of three excavated at the site. The others measured 7 m and 8 m across and the inclusion of additional ditches greatly increased the size differential. Thus, from the outset or through time, one of the monuments at Greshop was made to stand out through a greater investment in labour and was more monumental in form. Once complete, it was around four times the size of the other barrows in the cemetery and would have required approximately 16 times the volume of soil in its creation. 116 At Kinchyle, Inverness-shire (Fig 6), a square barrow appears to have been enlarged and elaborated, the monument consists of average sized smaller ditches/ barrows, around 8-10 m across, encased within an additional ditch extending to around 20 m in length/width. The role of the larger square enclosures at Rhynie is uncertain, but given that other cemeteries in northern Pictland have greatly enlarged monuments, it is possible that these too were large burial monuments. 117 If this is the case, then these are four times the size of the smaller barrows identified at Rhynie. 118 Bhreathnach (2011, 53) state that fert and ferta refer to grave mounds, but the term does not necessarily imply a mound: Gleeson 2014, 672-3. 112 Bhreathnach 2014, 125-6. 113 Gleeson 2014, 162. 114 O'Brien andBhreathnach 2011, 55;O'Brien 2009, 149. See also, in reference to Anglo-Saxon practices : Lucy 1992;Halsall 2010;Semple 2003;and Williams 1996;. 115 Longley 2009 Of course this is a plough-truncated monument, but examples such as Hallhole, Perthshire, an upstanding square barrow that survives up to 15 m across, demonstrate that large barrows of this form existed. 117 At Forteviot, the larger square enclosure in the cemetery has been interpreted as a shrine: see Campbell and Maldonado forthcoming. 118 The lack of contemporary internal features in the Rhynie large enclosures could be explained by the presence of an internal barrow. monumental barrows on this scale or enlarging barrows, cairns or enclosures, are little recognised traditions in Pictland. A handful of larger barrows are known in southern Pictland, in Tayside and Fife, such as Hallhole, Perthshire, but there has been little discussion of their significance. 119 The monumental and/or enlarged barrows are evenly spread across the study area, suggesting they may have been regionally significant monuments and/or cemeteries (Fig 16).
The cemetery with the greatest monumental investment in the study area is at Pitgaveny, Moray (Fig 5). Pitgaveny is a farm adjacent to Spynie Palace, the principal residence of the bishops of Moray from at least the 13th century, and both sites were located close to the shores of the extensive former sea loch of Spynie. The cemetery at Pitgaveny consists of at least six very large square barrows (probably more) and three round barrows or ring ditches. At least four barrows appear to have been elaborated -in one case ditches less than 8 m across were contained within much larger square barrows/enclosures some 20 m across. The Pitgaveny cemetery also shows clear evidence for the arrangement of the barrow cemetery in rows, and the joining of barrows through the sharing of barrow ditches, to create linear distributions of interlinked barrows. The barrows were arranged in two main rows, aligned ESE/WNW. The southern row had at least seven barrows, but nine may have been joined together in total.
The conjoining and elaboration of certain barrows may be related phenomena. Both developments suggest the importance of particular members of society, and imply that the creation of lineages of the dead (whether real or fictive) may have comprised an important element in the establishment and maintenance of cemeteries. 120 One of the major transformations of the early medieval period was the instigation of a hereditary aristocracy and the emergence of individuals with sufficient power and authority to call themselves kings. The growing power of these elite rulers seems to be reflected and materialised in the development of fortified sites in Pictland. 121 The occurrence of elaborate fortified enclosures from at least the 5th century onwards implies an increase in social differentiation. At the same time monumental cemeteries proliferated and their architecture suggests that these monuments may also have been implicated in the establishment and maintenance of hierarchy. The transition towards hereditary aristocracy relied on the creation and maintenance of lineage and the linking of leadership to a powerful past where ancestry was critical. 122 In early Irish literature burial places were seen as nodal points through which the Otherworld and the world of the dead could be accessed and cemeteries were one means by which lineages were forged. 123 Pictish barrows and cairns may have acted in similar ways, through these places claims of lineage and kinship may have been materialised and genealogies created that 119 Winlow 2011, 337, fig 10.4. Other possible examples are Wester Denhead, Perthshire, where there is a double-ditched square enclosure at least 20 m across visible in aerial photographs, and Kettlebridge, Fife, where a square barrow around 25 m across is located next to two very large round barrows. The large barrow at Hallhole was enclosed by two ditches, each with external banks, which suggests that in the cropmark record more than one ditch could represent multivallation as well as barrows being enlarged. However, examples such as Pitgaveney, where barrows were enclosed by square ditched enclosures, but not always on the exact same alignment, also suggest that some monuments probably were enlarged through time. In either case, whether multivallation or enlargement, the presence of additional architectural elements suggests that some barrows were marked out as being different to others through greater investment. 120 Williams 2007. 121 Halsall 2012Noble et al 2013. 122 Gleeson 2012, 9. 123 Ibid, 23-4. were instrumental in establishing hierarchical social relations. The establishment of monumental cemeteries in Pictland is commensurate with the first mentions of kings in northern Britain, while the decline of the tradition in the 7th-century coincides with the references to an over-king of the Picts. 124 Thus the creation of the first large formal cemeteries since the Bronze Age in north-eastern Scotland may go hand in hand with the establishment of regional hegemonies across Pictland. Their demise or the cessation of such rites was perhaps prompted by increasingly centralised forms of authority. 125 Claims over land, lineage and rulership of a people are recognised as increasing concerns in early medieval life. The establishment of formal cemeteries would have been an obvious way to legitimise claims to territory, through reference to ancestry, and create connections between particular lineages and locations. 126 The creation of cemeteries also resulted in the identities of particular lineages being much more obviously fixed in relation to particular locations in the landscape. 127 The elaboration and/or enlargement of particular barrows in the study region also suggests that particular ancestors were emphasised or their status even contested. The linking of monuments to one another through the sharing of ditches or the 124 Evans 2008, 9. 125 The 7th century is when Fraser (2011, 27) argues Pictish ethnogenesis took place, and when regional hegemonies and the over-kingship of Fortriu emerged (See also Evans 2008, 7-9). See Noble et al (2013) for a similar discussion on forts, with fewer and larger sites evident through time in northern Pictland. Gleeson (2015, 46) also discusses similar changes in Ireland with reference to 'cemetery-settlements'. 126 Barrett 1994, 61-3;Binford 1971;Härke 2001, 11, 19;Lucy 1992;Saxe 1970;Semple 2013;Williams 1997. 127 Semple 2013  conjoining of barrows may have situated the dead within particular relational networks of ancestry, and may have signalled evolving alliances and powerful emerging lineages. 128 Within cemeteries such as Greshop, only a single barrow seems to have been emphasised through the construction of a much larger monument, but at others such as Pitgaveny, four barrows were emphasised or enlarged and the majority of the barrows in the cemetery were changed into large monuments up to 20 m across. This suggests that while the burial places of some individuals were elaborated, and perhaps increasingly venerated through time, at Pitgaveny the whole cemetery population was marked as important through investment in the construction of the earthen mounds and ditches that enclosed the burials. While particular individuals buried in certain barrows may have been important in life, what perhaps mattered more were the ways in which the living community manipulated the status of the dead and the architecture of the cemetery for their own needs. 129 The aggrandisement of particular barrows and cemeteries was an act that would have been socially and politically charged. The elaboration and/or enlargement of existing barrows may have happened during the creation of other monuments or as part of other important social events, whereas the creation of new mounds could have been occasions when social relations were established, reworked and maintained. In this respect, social structure was not simply reflected in architecture of this kind; it was actively forged and manipulated through the creation of cemeteries. 130 Each mound or cairn constructed altered the form of the landscape through a process of accretion, each adding to an evolving narrative that embodied the community of both the living and dead. 131 The monumental cemeteries of Pictland may allow us to glimpse how new forms of social order were established in the 1st millennium ad. The mounds, ditches and cairns, for example, may have been constructed through bonds of clientship: by taking part in the creation of architecture like this, people were actively creating the material frameworks that underpinned an emerging social order and hierarchical society. The materials used -the mounds of earth and stone -may have been designed to add a literal and metaphorical permanency to the social relations being expressed, 132 and through their solidity acted as a powerful material mnemonic that helped create and maintain a new social order. 133 How these cemeteries were situated within the early medieval landscape is also important. While some of these barrows and cairns could reach proportions of up to 25 m in diameter or more, they were not generally located in highly conspicuous locations that would have been visible for kilometres around. The analysis of the barrows in the study area suggests they were located on more locally visible terraces and knolls. In addition a number were situated in 128 Williams 2007 129 Barrett 1994, 51. For a contemporary example see Ó Corráin (1998) who discusses the use of unimportant or invented ancestors by lineages who acquired power centuries later. We thank Patrick Gleeson for highlighting this example. 130 Similar arguments regarding the use of burial architecture to underline particular familial relations and the establishment of lineage have been made for other Pictish and early medieval cemeteries more generally. At Lundin Links, Fife, Howard Williams has suggested that the acts of containing multiple dead within single monuments and the linking of particular monuments through new acts of building created genealogies through architecture: Williams 2007. Maldonado (2013 has also suggested that cemeteries in general continually reforged relationships between the living and the dead, perhaps even creating a form of distributed personhood where the dead were considered an active part of living society. The linking of barrow and cairn monuments in Pictland may have been one other way in which individuals -or dividuals (Fowler 2004;Strathern 1988) -were understood through a relational conception of self, in which one's kinship relations were central to how people were understood in a society where bonds of clientship, lineage and status were increasingly fixed in space and time. 131 Barrett 1994, 113, 123;Bradley 2007, 165. 132 Parker Pearson andRamilisonina 1998, 309-11;Williams 2002, 68. 133 Jones 2011, 12;Thomas 1999, 60. areas tied into the geographies of routine movement: on routeways, at fording points and on general lines of movement through the landscape. Some cemeteries were also located at key transitional points in the landscape. The cairns at Tillytarmont, Aberdeenshire, for example, are located at the confluence of two rivers (Fig 17) and a boundary location between two old counties, the division between three later medieval parishes and the site of a number of fords and crossing places. The place name itself derives from tulach -a term often associated with assembly. 134 Tillytarmont is the findspot of five Class I symbol stones since the 19th-century suggesting the cemetery was a component in wider landscape of Pictish power. In Anglo-Saxon England some field cemeteries seem later to have become important assembly places. 135 Gleeson has also highlighted the roles that so-called 'cemetery-settlements' may have played in assembly practices in Ireland. 136 The location of Tillytarmont on an important natural and cultural boundary and its place-name evidence may similarly suggest the cemetery was located in an area that functioned as a place of assembly. This particular cemetery also shares characteristics with the burial places in Ireland known as ferta, which acted as places of legal assembly and as boundary markers and protectors of the land and territory of particular kin groups. 137 The second element of the Tillytarmont place name, derives from an tearmainn 'sanctuary', which implies that in a later period there was also an important church nearby. 138 Tillytarmont may represent an important site of assembly and burial that originated in a pagan context, but continued as an important place in a Christian milieu. 139 The landscape setting of these monumental cemeteries also reveals something of the wider geographies of northern Pictland. With the exception of Rhynie, there are no known close juxtapositions between confirmed Pictish fortified sites and cemeteries in the study area. The cemeteries thus may offer important clues to additional important nodes in the Pictish landscape and may signal elements of an emerging multifocal landholding system that formed the basis of power for high-status Pictish communities. 140 What is also notable is the lack of a clear relationship between monumental cemeteries and Pictish stones. Symbol stones have been found in close proximity to monumental cemetery sites in only a handful of cases. 141 Indeed, GIS analysis casts further doubt on the connections between burial sites and symbol stones: as part of the Northern Picts project the locations of 337 Pictish Class I and II carvedstone monuments were compared to those of 233 long cists or long cist cemeteries and 110 square barrow or cairn locations (as recorded in the Canmore database), across Perthshire and Fife to Caithness. Only 14 of 337 Pictish carved-stone monuments were found within 134 O'Grady 2014, 114-19; Taylor with Markus, 2012, 519-20. The name specifically means 'hillock' or 'mound' and in Ireland can simply mean ridge or hill (the latter not appropriate to Tillytarmont), but there are two examples of later court sites being held at mound sites with this place-name element in Scotland: O'Grady 2014, 114. The place-name element is also found in later medieval central names such as Tulliallan, Fife and Murthly, mòr-thulack, 'big mound', the centre of a secular barony from at least the 14th century. It is also a place-name element of Kintillo, Perthshire, one of Scotland's chief centres of legal assembly: Taylor with Markus, 2012, 519. Tulach names are concentrated in north-east Scotland in the areas that marked the core of Pictland: see O'Grady 2014; and Taylor with Markus, 2012, 519-20. See also Gondek 2010 for discussion of the place-name evidence. 135 Brookes and Reynolds 2011, 235-40; see also Sanmark 2013 andWilliams 2002;. Williams argues that some of the large-scale cremation cemeteries of Anglo-Saxon England were a form of central place. 136 Fitzpatrick 2004;Gleeson 2015, 45. 137 O'Brien and Bhreathnach 2011, 54-5. See also Charles-Edwards 1993 Simon Taylor (pers comm) has pointed out that the nearest church is Ruthven which is over 2 km east and would indicate a sanctuary of a considerable scale and importance. See Gondek (2010) for a discussion of the place-name evidence and wider early Christian activity in the environs of Tillytarmont. 139 Brookes and Reynolds 2011, 88. 140 Ross 2006. 141 Rae and Rae (1953) suggest the undecorated monoliths at Pityoulish could have been painted.

FIG 17
Location of Tillytarmont with symbol stones and cairn marked. Fords and routeways crossing the haughland, as represented on 19th-and 20th-century maps, are also shown. The location of the two symbol stones is based on good locational information. The other three stones were found on the haughland, but these cannot be accurately located. A geophysical plot carried out by Oskar Sveinbjarnarson for the Northern Picts project is also superimposed on the image to show geophysical anomalies that coincide with the routes across the haughland. 500 m of a long cist or square barrow (less than 5%) and only five of those 14 examples were square barrows or cairns. In 2007 Clarke critically examined the relationship between symbol stones and sites of burial and concluded that the case for a connection was weak. 142 Our GIS locational analysis reinforces this impression. It seems that a direct relationship existed between Pictish cemeteries and symbol stones only in exceptional cases. 143 The analysis of the symbol stones suggests that they had more varied purposes than just gravemarkers. In this respect they can perhaps be directly compared to the ogham stones of Ireland. Most ogham stones in Ireland appear to have been inscribed in the 5th-6th centuries ad and, where decipherable, appear to have been used to express the genealogies of elite members of society. In the past these have been interpreted as burial markers but, like Pictish symbol stones, there are very few examples of stones found in direct association with graves. Some may be memorials, but others had a legal function concerned with boundaries and landholding, acting as 'visible, physical declarations of land possession'. 144 Thus, ogham stones may have acted in similar ways to barrows, proclaiming the rights and genealogies of particular elite lineages. 145 The Pictish Class I symbol stones appear, like the Irish ogham stones, to have been markers of elite identities. 146 They may have complemented the role of cemeteries in marking important landscapes of power and perhaps directly indicated land ownership and rights. There are similarities here in the ways symbol stones and barrows expressed ideas about identity: on occasion symbol stones, like barrows, appear to have symbol sets added, and in a smaller number of cases symbols were overwritten by new carvings. 147 This implies the identities expressed on symbol stones were cumulative, and the overwriting of symbols might demonstrate that these identities were reworked or even directly contested. 148 The 5th-6th centuries ad, when monumental cemeteries were at the height of their use, was a period of change in northern Britain. One of the major changes was the growing influence of Christianity. 149 The establishment of the monumental cemeteries of southern Pictland has often been discussed in relation to changing religious practices. 150 Although a linkage to the creation of particular lineages is not in doubt, these burial monuments also drew on other influences, such as the adoption of E/W burial, Christian belief and local practices and traditions. 151 The end of barrow traditions in the 7th century has been connected in this article to the centralising of power in Pictland, and parallels have been drawn with Ireland. 152 Christianity, however, may have played a role too -in Ireland the church influenced the shift away from ancestral burial grounds as it grew in power in the 7th and 8th centuries. 153 In Anglo-Saxon England the recent Bayesian modelling of radiocarbon dates by John Hines and Alex Bayliss has suggested a strong correlation between the end of elaborate furnished burial and the consolidation of the early Church in Anglo-Saxon England in the third quarter of the 7th century. 154 142 Clarke 2007, 27-31. 143 Ibid, 31. 144 Bhreathnach 2014, 42-4. 145 The inscribed stones of Wales may have also acted in similar ways, Longley 2009, 116. 146 Samson 1992Forsyth 1997. 147 Clarke 2007 Thanks to one of the anonymous reviewers for highlighting this potential parallel. 149 Maldonado 2011a. 150 Eg Ashmore 1980;Greig et al 2000, 611;Henshall 1956;Wainwright 1955b. 151 Maldonado 2011ab, 2013. 152 O'Brien and Longley 2009, 149. 153 Ibid, 149-50. 154 Ibid, 548-54. Hines and Bayliss (2013 suggest the end of furnished burial may have been directly connected to the arrival of Theodore of Tarsus at Canterbury and his reform of the Church in the 670s. In Pictland the dating is less precise, but the influence of the church is evident in the establishment of monasteries from the 7th-century onwards -a time when the tradition of constructing monumental cemeteries was coming to an end. 155 By the 8th and 9th centuries, at the latest, important churches were also found at royal settlements. 156 Thus, in Pictland, the ending of the construction of monumental cemeteries may also have been connected to the changing character of belief. The cessation of the construction of monumental cemeteries may have been connected to the ways in which elites incorporated the Christian faith into ideologies of rulership. In particular, we can perhaps highlight the effects that Christianisation may have had on conceptions of the self and genealogy. In a Christian context genealogies appear to have been altered to focus on family and individual relationships with God. With the consolidation of power in 7th-century Pictland, the local and regional genealogies that barrow cemeteries helped create may have been no longer sufficient to express the authority of elites who began to rule larger territories. It is perhaps in this context that we see disinvestment in the architecture of the dead, and greater interest in other forms of elite monument that more clearly expressed the relationship of elites to their Christian faith. The most obvious example is Class II Pictish sculpture, likely 7th-9th century in date. These monuments are carved with depictions of elites, often mounted on horseback, shown in close juxtaposition with monumental Christian crosses and/or the image of David. 157 CONCLUSIONS Monument building often occurs at horizons of social change and centralisation, with the greatest labour inputs occurring in transitional periods in the establishment of elites. 158 As Barrett notes, 159 elites are not merely created through the completion of monumental projects, but hierarchy emerges through the initiation and realisation of such projects. Emulation, competition and a desire for power within an emerging early medieval society may have been a strong motivating factor in the construction of monumental cemeteries in Pictland. 160 In this article the rich aerial photographic archives available for north-eastern Scotland have been brought together for the first time, and used to review the evidence for burial traditions in an area of Pictland central to the establishment of the Pictish kingdoms of northern Britain. While few sites have been excavated, the transcription of the aerial evidence for northern Britain is an important resource in assessing how burial architecture may have been deeply implicated in the creation of new forms of society in northern Europe in the first millennium ad. In this area of northern Europe the appearance of formal cemeteries and the creation of the Pictish monumental cemeteries can be seen to go hand in hand with the establishment of a more hierarchical form of lineage-based society. 161 The end of this tradition can be linked to changes within the social, political and religious order, with the dead becoming a less prominent source of power as supra-regional polities emerged and the Christian faith achieved firmer ground. 155 Carver 2008, 76;2016, 46-7, tab 3.1. 156 Important churches occurred at Forteviot, Perthshire; Meigle, Angus and Burghead, Moray: see Henderson and Henderson 2004. 157 The adventus scenes of Pictish Class II monuments: Goldberg 2012, 155. The David imagery is an obvious focus for elite genealogies that in a Christian context sought to trace their descent ultimately to Old Testament figures. 158 Whittle 1997, 145;Cherry 1978. 159 Barrett 1994, 29-32. 160 Whittle 1997, 167. 161 See Barrett (1988) for a discussion of changing funerary rituals as a focus of new forms of social identity among both the living and the dead. nelle terre dei Pitti, vennero eretti monumenti funebri a forma quadrata e rotonda, per commemorare una piccola quantità di persone -forse una élite emergente nei secoli postromani. Questo studio presenta i risultati di un progetto che ha riesaminato e consolidato le testimonianze dei cimiteri monumentali dei Pitti settentrionali dall'Aberdeenshire all'Inverness-shire, trascrivendo per la prima volta le testimonianze aeree di molti siti. Si valuta anche la posizione dei cimiteri nel paesaggio, oltre che il loro rapporto con le pietre simboliche dei Pitti, con i siti fortificati e con i paesaggi degli stanziamenti del 1° millennio dell'era cristiana. Si mettono in evidenza due particolari elementi dell'architettura tombale della Pittavia settentrionale: l'ampliamento dei tumuli e il collegamento dei tumuli mediante canali comuni tra tumuli e cairn. Si sostiene qui che entrambe queste pratiche fanno parte della creazione di genealogie dei vivi e dei morti durante l'importante periodo di transizione nell'Europa settentrionale in cui le aristocrazie ereditarie accrebbero la loro importanza.