The bluestones of Stonehenge

FEATURE Feature The bluestones of Stonehenge How and why the bluestones arrived at Stonehenge, the UK’s most revered ancient monument, has long held people’s imagination. The key to understanding these questions relies heavily on the location of their sources. Following early studies in the late eighteenth and early nineteenth centuries, which proposed various places but in particular south-west England, H.H. Thomas, in 1923, suggested that they came from the Mynydd Preseli, in north Pembrokeshire, Wales. Thomas proposed a number of key locations for the geographical origin of the stones. However, recent investigations have called those locations into question, identifying different sources albeit from the same broad area in north Pembrokeshire. Identification of these proposed new sites has led to archaeological excavations and important new discoveries including new suggested routes for the transport of the bluestones from the Preseli Hills to Stonehenge some 230 km away. Stonehenge, in the English county of Wiltshire, is one of the most visited and most distinctive ancient monuments in the world, and certainly architecturally the most sophisticated ancient stone circle known. This iconic site is protected, together with the stone circle at nearby Avebury, as a World Heritage Site. The significance of these sites to World Heritage was neatly summarized in the UNESCO designation in 2009: ‘Stonehenge is the most architecturally sophisticated prehistoric stone circle in the world, while Avebury is the largest in the world. Together with inter-related monuments and their associated landscapes, they help us to understand Neolithic and Bronze Age ceremonial and mortuary practices. They demonstrate around 2000 years of continuous use and monument building between c. 3700 and 1600 bc’. For at least 700 years people have speculated on Rob Ixer1 & Richard Bevins2 1 Institute of Archaeology, University of London, WC1H 0PY, UK [email protected] 2 Department of Natural Sciences, National Museum of Wales, Cardiff, CF10 3NP, UK Richard.bevins@ museumwales.ac.uk Fig. 1. Oblique aerial view of Stonehenge. (Photograph courtesy of Adam Stanford.) 180 © John Wiley & Sons Ltd, The Geologists’ Association & The Geological Society of London, Geology Today, Vol. 33, No. 5, September–October 2017 FEATURE Fig. 2. Simplified map of north Pembrokeshire showing the location of Mynydd Preseli and the outcrop of the Fishguard Volcanic Group. the origin of the stones of Stonehenge and even more on how they were transported to their site on the distinctive chalk downlands of the English county of Wiltshire. For most of that time the ideas were rather fanciful (some still are) but now serious speculation has centred around two main agencies, namely mankind or ice. Geological studies of the Stonehenge lithologies are less than 150 years old and interest has been sporadic, with decades of inaction between bursts of activity; the last decade is one of those latter times and so the rocks have been subjected to intense study, largely involving detailed petrography and whole-rock and mineral geochemistry, together with microfossil and radiometric dating techniques. This work, jointly led by ourselves but involving other Earth science specialists, including Peter Turner, Nick Pearce, Peter Webb, Jane Evans, Stewart Molyneux and Andy Gize, has produced unexpected results that really can claim to rewrite the history books—even if only slightly. The stones of Stonehenge (Fig. 1) comprise two main classes, the bigger attention-grabbing sarsens that make up the ‘trilithons’ (two large upright stones with a connecting lintel) of the large outer circle and inner horseshoe, and the far smaller bluestones (here, defined as any non-sarsen rock used as an ‘orthostat’ or standing stone) that make up the inner circle and the inner horseshoe, plus one anomalous, very large, calcareous Devonian sandstone, the recumbent Altar Stone. Some of these bluestones are present in the circle only as buried stumps and perhaps up to half of the original stones (bluestones and sarsens) are missing and are now only present as rock debris, most of it small and lightweight, sometimes found on the surface but usually buried, throughout the Stonehenge Landscape. The sarsens (which were the subject of a note in Geology Today, v.33, n.1, p.17) are a silcrete sandstone, which are local to Salisbury Plain and may have only been moved dozens of kilometres, unlike the transposition of hundreds of kilometres that the bluestones have undergone. Of the standing bluestone orthostats most are dolerites from the Mynydd Preseli, in north Pembrokeshire (Fig. 2), the majority (the spotted dolerites, the iconic so-called ‘preselites’) show diagnostic pale-coloured spots (Fig. 3); they are often misidentified as being feldspar clots but in fact are aggregates of low grade, secondary metamorphic minerals that (probably) have replaced feldspar. A few orthostats are fashioned from unspotted dolerite found at different source locations to the spotted varieties. H.H. Thomas in 1923 was the first to recognize that the spotted dolerites cropped out in a number of places on the tops of the Preseli Hills (Fig. 4) and suggested specific localities including Carn Menyn (also known as Carn Meini) and Cerrigmarchogion. In 1991 a team from the Open University resampled many of the bluestone dolerite orthostats from Stonehenge and also dolerites from the Preseli outcrops and, primarily based on their geochemistry along with © John Wiley & Sons Ltd, The Geologists’ Association & The Geological Society of London, Geology Today, Vol. 33, No. 5, September–October 2017 181 FEATURE Fig. 3. A cut face of spotted dolerite from the Mynydd Preseli. The spots were previously thought to represent altered feldspar phenocrysts but X-ray diffraction shows them to be fine-grained intergrowths of chlorite, epidote and albite. The scale bar is 5 cm. work previously published by Bevins, Lees and Roach, suggested that Carn Menyn on the main Mynydd Preseli ridge was the origin for many of the spotted dolerites, a view that had been popular for much of the twentieth century. Our re-examination of the interpretation presented by the Open University team, together with new analyses, has shown that many of the Stonehenge dolerites are in fact a better match to Carn Goedog on the northern slopes of the Preseli (Fig. 5) and very recent excavations at Carn Goedog have revealed evidence for Neolithic working of the outcrop. There are only four above-ground, non-dolerite igneous orthostats (stones SH38, 40, 46 and 48) and these are volcanic rocks, two being dacites and two are rhyolites. They are all different from each other suggesting four discrete geographical origins and, as yet, no specific location can be suggested for them although as evidenced by recent age dating for stone 48, they are almost certainly part of the Fishguard 182 Fig. 4. Oblique aerial view of the eastern extremity of Mynydd Preseli showing some of the key localities that have featured in the bluestone literature. Photograph courtesy of Sid Howells. Volcanic Group, of Ordovician age, which crops out to the north of the main Preseli Ridge. The Fishguard Volcanic Group (Fig. 2) forms a bimodal basic-silicic sequence and the dolerites are the high level intrusive equivalents of the spectacular pillow lava sequence as exposed, for example, at Strumble Head. The silicic rocks (rhyolites and dacites) were erupted as extrusive domes, rare silicic pillowed flows and as rhyolitic ash flow tuffs. All took place in a submarine environment and reworking of the volcanic material is evident as debris flow deposits which commonly incorporate mudstone intraclasts. It must be of significance that amongst the many thousands of debris/debitage samples that have been investigated no more than 25 pieces of debris have been recognized from all four of the upright dacitic/ rhyolitic stones. One orthostat, SH38, is unusual in carrying graphitizing carbon (Fig. 6). This has been rarely recorded in rocks (it is normally seen as an artificial component in coke) and may represent the thermal maturation (‘cooking’) of mud rip up clasts into the base of a hot subaqueous rhyolitic ash flow. It is hoped that this peculiarity will help in finding the exact provenance of this particular Stonehenge stone. The buried orthostats are all unsampled but descriptions made during excavations and photographs suggest that they are a mixture of tuffs and sandstones and indeed it is just those lithologies that dominate the debris/debitage. The debitage is remarkably uniform throughout the Stonehenge Landscape and alongside some sarsen and dolerite it is dominated by a very distinctive, strongly foliated rhyolitic tuff and by more variable, poorly cleaved argillaceous tuffs, as well as with lesser amounts of an indurated Lower Palaeozoic sandstone which shows a poor fracture cleavage. The last rock type is very different from the calcareous (non-cleaved) Devonian sandstone that makes up the flat-lying Altar Stone (once again, like the other four extant bluestones the Altar Stone lithology is almost totally absent from the debitage). A systematic search of outcrops within the Fishguard Volcanic Group suggested a petrographical similarity between the main strongly foliated rhyolitic debitage and outcrops in the Fishguard Volcanic Group around Pont Saeson (see Fig. 7). More detailed sampling of the area showed that a 70 metre long outcrop, Craig Rhos-y-felin (Fig. 2), was the source of this debitage and small scale, systematic sampling along this outcrop demonstrated Stonehenge debitage could be pin-pointed to a large (at the time hidden) north-west facing planar face, a surface that does not look natural (Fig. 8). Subsequent archaeological excavations have shown features consistent with ancient quarrying. More recent detailed mineral geochemistry and high precision radiometric dating have continued to confirm the initial petrographical © John Wiley & Sons Ltd, The Geologists’ Association & The Geological Society of London, Geology Today, Vol. 33, No. 5, September–October 2017 FEATURE provenancing of this material to Craig Rhos-y-felin, whilst also showing that the four rhyolitic and dacitic orthostats do not come from there. Flakes of argillaceous and calcareous tuff are almost equal in amount to the Craig Rhos-y-felin rhyolite debris in the Stonehenge Landscape. Their fine-grained mica-rich nature make them difficult to work on microscopically and they are compositionally highly variable, with many rhyolitic, vesicular lava and micro-tonalite clasts contained within a finegrained quartz, white mica and chlorite-rich matrix determined by X-ray diffraction. Finding a provenance for these tuffs will no doubt be very exacting and they may turn out to be the most difficult of all the rock types to investigate. The final significant lithology in the bluestone assemblage is the Lower Palaeozoic sandstone. Although petrographically this is very different from the Altar Stone sandstone, some of the archaeological literature has conflated and confused them, even suggesting that they come from shore-line outcrops at Mill Bay close to Milford Haven. However, the Lower Palaeozoic sandstone is an indurated rock with a poor fracture cleavage and with mudstone intraclasts and has been palaeontologically dated, using acritarch assemblages, to a likely Late Ordovician age, clearly separating it from the Altar Stone and the Devonian rocks cropping out at Mill Bay. Eliminating any Milford Haven connection with Stonehenge has archaeological significance. The Altar Stone may be the most interesting of all the bluestones; it is certainly the largest and heaviest and the only one lacking a north Pembrokeshire area origin. It is a slightly unusual, probable Devonian carbonate-cemented, fine-grained sandstone, with few if any, ‘unique’ petrographical characteristics. It could come from anywhere within a broad outcrop pattern of the Senni Beds from west Wales to the Fig. 5. Oblique aerial view of Carn Goedog, now thought to be the source of the majority of the analysed spotted dolerite bluestones. (Photograph courtesy Adam Stanford.) Herefordshire hills. Unlike the other lithologies, where routine microscopy or geochemistry are the main provenancing tools, it may be that field studies, and in particular matching the dimensions of the Altar Stone to bedding-jointing characteristics within the many Senni Beds outcrops, will lead to its geographical origin. Other lithologies found within the Stonehenge Landscape are minor to very minor in amount and often found in molehills or rabbit burrows; they reflect the local/regional geology or are post-Victorian imports and include rail ballast, road metal and New Age rock crystals. This lithological litter has also been investigated and where possible identified and provenanced and then eliminated from any ancient Stonehenge association. The task of separating true Stonehenge lithologies from adventitious material is important, as Ice Age proponents have suggested that these rocks (or rather the wide range of rock types) are indicative of glacial dumping, despite glacial deposits elsewhere on Salisbury Plain being remarkable for their absence. Although our studies are far from over, indeed our first pass has only just finished, it is clear that the accepted archaeological explanation (up to 2010) for the origin and the transport route of the bluestones (Fig. 9), namely that they were obtained from outcrops on or close to the main ridge of the Mynydd Preseli then moved southwards to Milford Haven where they were transported by sea to Somerset (possibly losing Fig. 6. Graphitizing carbon droplet in stone 38 from Stonehenge. The image is 200 µm across and was taken under oil immersion, reflected white light and crossed polars. (Photograph by Andy Gize.). Fig. 7. Thin section photomicrograph of the foliated rhyolitic tuff from Craig Rhosy-felin showing the strong planar fabric and the presence of included micro-tonalite lithic fragments. Plane polarized light. © John Wiley & Sons Ltd, The Geologists’ Association & The Geological Society of London, Geology Today, Vol. 33, No. 5, September–October 2017 183 FEATURE Suggestions for further reading Fig. 8. The outcrop of Craig Rhos-y-felin showing the prominent northwest facing exposure. (Photograph by Adam Stanford.) Fig. 9. Map showing the various routes proposed for the transport of bluestones from Mynydd Preseli to Stonehenge, a distance of some 230 km. (Drawn by Irene Deluis.) 184 a stone on Steep Holm) is no longer a tenable option. The majority of the spotted dolerites have now been provenanced to Carn Goedog, the main rhyolitic tuff debitage to Craig Rhos-y-felin, both on the northern slopes of the Mynydd Preseli, and the Lower Palaeozoic sandstone to unknown outcrops north of the Preseli region. In addition neither the Altar Stone nor the Lower Palaeozoic sandstone come from outcrops at Mill Bay (Milford Haven) and the ‘lost bluestones’ of Steep Holm are coarse-grained metabasite glacial erratics with no Stonehenge equivalents and of a kind not seen in north Pembrokeshire. There is no geological evidence for or against a sea route, indeed there is no geological evidence for any transport route (sea and/or land) and unless a dropped orthostat is found it is unlikely there ever can be. Indeed, the question of movement by ice is not ruled out by the available geological evidence although consensus favours human transport. The best that can be hoped for is that a number of undisputed Neolithic quarry sites can be found and recognized and that together they indicate the most likely pathway along which the stones were moved. Bevins, R.E., Ixer, R.A. & Pearce, N.J.G. 2014. Carn Goedog is the likely major source of Stonehenge doleritic bluestones: evidence based on compatible element discrimination and Principal Component Analysis. Journal of Archaeological Science, v.42, pp.605–622. Bevins, R.E., Atkinson, N., Ixer, R.A. & Evans, J.A. 2017. U-Pb zircon age constraints for the Fishguard Volcanic Group and further evidence for the provenance of the Stonehenge bluestones. Journal of the Geological Society of London, v.174, pp.14–17. Ixer, R.A. & Bevins, R.E. 2010. The petrography, affinity and provenance of lithics from the Cursus Field, Stonehenge. Wiltshire Archaeological & Natural History Magazine, v.103, pp.1–15. Ixer, R.A. & Bevins, R.E. 2011. Craig Rhos-y-felin, Pont Saeson is the dominant source of the Stonehenge rhyolitic debitage. Archaeology in Wales, v.50, pp.21–31. Ixer, R.A. & Bevins, R.E. 2013. Chips off the old block: The Stonehenge debitage dilemma. Archaeology in Wales, v.52, pp.11–22. Ixer, R.A. & Turner, P. 2006. A detailed re-examination of the petrography of the Altar Stone and other non-sarsen sandstones from Stonehenge as a guide to their provenance. Wiltshire Archaeological & Natural History Magazine, v.99, pp.1–9. Ixer, R.A., Bevins, R.E. & Gize A.P. 2015. Hard ‘Volcanics with sub-planar texture’ in the Stonehenge Landscape. Wiltshire Archaeological & Natural History Magazine, v.108, pp.1–14. Parker Pearson, M., Bevins, R.E. Ixer, R.A., Pollard, J., Richards, C., Welham, K., Chan, B.,K. Edinborough, K., Hamilton, D., McPhail, R., Schlee, D., Schweenninger, J-L., Simmons, E. & Smith, M. 2015. Craig Rhos-y-felin: a Welsh bluestone megalith quarry for Stonehenge. Antiquity, v.89, pp.1331–1352. Thomas, H.H. 1923. The source of the stones of Stonehenge. The Antiquaries Journal, v.3, pp.239– 260. Thorpe, R.S., Williams-Thorpe, O., Jenkins, D.G. & Watson, J.S. (with contributions by Ixer, R.A. & Thomas, R.G.) 1991. The Geological Sources and Transport of the Bluestones of Stonehenge, Wiltshire, UK. Proceedings of the Prehistoric Society, v.57, pp.103–157. Young, C., Chadburn, A. & Bedu, I. 2009. Stonehenge World Heritage Site Management Plan 2009. English Heritage, London. © John Wiley & Sons Ltd, The Geologists’ Association & The Geological Society of London, Geology Today, Vol. 33, No. 5, September–October 2017