Victorian gifts: new insights into the Stonehenge bluestones

The recent rediscovery of a series of rock samples collected during the Victorian period has allowed new analysis of some of the stones of Stonehenge. Rob Ixer, Richard Bevins, Nick Pearce, and David Dawson explain more.


Today, samples from the Stonehenge bluestones – the smaller uprights that make up the monument’s Inner Circle and Inner Horseshoe, and which were quarried far from Wiltshire in west Wales (see CA 313, 345, and 366) – are almost all held in museum collections. They fall into two major groups: the first comprises the very few that have been taken directly from the orthostats (stone uprights set in the ground) themselves, whether above-ground or buried, and which are therefore proven to have come from bluestones. The other, much larger group contains the many, many thousands of small pieces – the so-called debitage – that litter the Stonehenge Landscape but are not attributed to specific stones. Demonstrating that fragments from this latter group belong to a given orthostat and are not random rocks, or even glacial erratics, has been and continues to be a challenge.

This drawer, containing rock samples collected by the Victorian geologist William Cunnington III, was recently rediscovered in the Wiltshire Museum’s attic. The top row has thin sections of rhyolite, while those on the bottom are andesitic tuff. Two, labelled S70 and S70E, are from orthostat SH32c, one of the Stonehenge bluestones.

Victorian geologists directly sampled the bluestones (most of the pieces they collected are now curated by the Natural History Museum in London), but their samples are few in number, some are missing, and there are doubts over their correct attribution. It was to address this that, in the early 1990s, a group of Open University archaeologists and geochemists directly sampled 15 above-ground orthostats alongside 22 pieces of debitage – but since then, just two more direct orthostat samples, both dolerites, have been added to the list (in 2006). Given this lack of material, any happenstance that increases the number of examples within the tiny but critical directly sampled group is a cause for celebration. Last August was such a happening. After more than 50 years, three thin section samples – taken from one standing orthostat and two buried orthostats – have been rediscovered. Two of them are dolerites, but the one that is not is especially significant, as we will explain.

These two slides, S70 and S70E, come from stump S28 (now known as SH32c or Stone 32c) and represent a rare example of samples taken directly from a Stonehenge bluestone.

How did this happy discovery come about? Every museum has dark and dusty corners, and in the case of the Wiltshire Museum, this was an attic where some uncatalogued geological specimens had been placed in the 1980s following their last examination by the British Geological Survey (BGS) a decade earlier. The museum’s director, David Dawson, takes up the story of what happened next: ‘Last year, to reduce the museum’s carbon footprint, we were installing extra insulation and had to clear the boxes out of the attic. One held a tray of microscope slides labelled ‘Stonehenge’, and we immediately recognised the handwriting of William Cunnington III [geologist and principal founder of Wiltshire Archaeological and Natural History Society]. Having read all the latest research into the source of the bluestones, we knew they would be important – and a quick phone call with Rob Ixer started the ball rolling on his research.’

In this photograph of excavations at Stonehenge in 1954, Stone 32c is the closest of the three stone stumps that can be seen. Its strongly laminated structure is similar to much of the Andesite Group A debitage. Photo: Historic England Archive.

Significant samples

In fact, the attic yielded not one but three trays of microscope slides, containing a total of 34 thin sections from 32 separate rock samples collected by William Cunnington in the period 1876-1881. This is a nationally significant collection numerically, historically, and scientifically, as it represents a moderate-sized, unbiased sampling from the last collection of surface finds from within the Stonehenge Circle. It concentrates on ‘foreign rocks’, meaning that there are no examples of sarsen stone (the material of Stonehenge’s famous trilithons, which was quarried locally – see CA 367), though there is a single sample of Greensand (Cretaceous) packing stone. The samples are historically important because they provide first-hand information attesting to mid-Victorian geological- and Stonehenge-related thinking, and are examples of early thin-section manufacture before mass production. And, scientifically, its re-study allows Victorian and later mid-20th-century rock names and descriptions to be incorporated into our 21st-century nomenclature, eliminating incorrect or obsolete rock names. The plethora of rock names in existence has been abused by some to suggest that the range of bluestone rock types is far wider than it really is, a misdirection made in order to promote the belief that they are random erratics rather than having been quarried and transported by humans.

A fragment of the Altar Stone, collected in 1844 and still bearing its original label. It measures 7cm by 4.5cm in size, and it was the source of thin section Wilts 277 – the first detailed petrographical description of the orthostat.

Perhaps even more excitingly, though, the collection contains not just representative rocks picked up from the hundreds lying on the grass, but also a limited number of samples gathered close to standing orthostat Stone 33 and from the stump of Stone 61a during the small excavations that Cunnington undertook within the Stonehenge Circle itself in 1881. In terms of its contents, the recently found assemblage is comparable to all other collections that have been described; its lithologies comprise nine Spotted and Unspotted Dolerites, seven Rhyolite Group C tuffs from Craig Rhos-y-Felin, five calcite-bearing Andesite Group A tuffs, two Lower Palaeozoic Sandstones, six Dacite Group B/Stone 38 tuffs, plus a single sample thought to come from the Altar Stone, a Dacite Group D tuff, and a non-bluestone Greensand sandstone. It includes all the major lithological bluestone groups and some of the minor ones. No previously undescribed rocks were recognised, confirming again that all collections of bluestone debitage from the Stonehenge Landscape, mainly taken from excavations, comprise the same limited range of rock types.

The single Dacite Group D sample came as a surprise, but a welcome one, offering additional proof that Group D is a true bluestone. Formerly, because of the very restricted occurrence of this group in the Stonehenge Landscape (all known fragments were found close to the western end of the Stonehenge Greater Cursus), it was believed not to be associated with Stonehenge. Finding an example within the circle plus an additional stone from Aubrey Hole 15, however, suggests that this lithology should be added to the list of genuine bluestones, and its parent orthostat sought. In other words, a new bluestone group can now be added to the list.

Plots of barium vs strontium and rubidium vs zirconium (determined by portable XRF) for Wilts 277, compared with in situ analyses from the Altar Stone.

Perhaps the most spectacular find, though, was the two thin sections labelled S70 and S70E, which come from Cunnington’s excavation and his direct sampling of buried stump S28 (since renamed Stone 32c), alongside four very similar-looking pieces of debitage. The lithology was immediately recognised, creating an additional bonus above the reward of adding Stone 32c to the list of directly sampled orthostats. We now know that 32c is the, or one of the, sources of the thousands of small pieces of Andesite Group A debitage found throughout the Stonehenge Landscape, which had been previously unassigned. Now, one of the major groups of debitage has found at least one secure parental home. In other words, this collection provides yet more proof that the debitage is directly related to the Stonehenge orthostats rather than being a natural, random rock scatter.

Analysing the Altar Stone

The picture is not completely clear, however. The Wiltshire Museum is not alone in having Victoriana, and Cunnington does not make it clear if his Altar Stone sample (petrographically identical to other recent Altar Stone descriptions in the literature) came from direct sampling or not. Adding to the confusion, Victorian and later 1920s excavators labelled a fair number of excavated micaceous sandstones ‘Altar Stone’, whether they were or not. Indeed, the Hunterian Museum in Glasgow even has two thin sections from a collection of the major Victorian geologist Professor Matthew Heddle labelled ‘Alter Stone’ (sic), but one is a sarsen and the other a volcanic tuff. Therefore, the question whether the large prone orthostat Altar Stone (Stone 80) has been directly sampled seemingly remains unresolved… or has it?

One of the ‘foreign rocks’ excavated by Hawley at Stonehenge in the 1920s: a refitted joint block known as RSN18 ENQ 2305. It includes the triangular British Geological Society sample and three Salisbury Museum pieces; the missing area was the destructively analysed sample OU2. Image: Reproduced with the permission of British Geological Survey © UKRI [2022]. All rights reserved.

The answer came from a specimen of the Altar Stone held in the Salisbury Museum (museum label 2010K 240), which was the source of ‘Wilts 277’, a thin section held in the Somerset Heritage Centre at Taunton. This answer did not come from investigation of the rock itself, however, but its partially decipherable label, which reads:

Portion of the underpart of the Altar Stone at Stonehenge – taken by Mr Brown of Amesbury while excavating in the summer of 1844 to ascertain if any interment there – no traces of such discovered – The search was made at the request of a Swedish gentleman who was deputed by an Antiquarian Society of that (sic) Sweden to obtain the information.

This shows a slice and thin section taken from the joint block; the authors have recently reunited and examined the block and all its offcuts and associated thin sections. Image: Reproduced with the permission of British Geological Survey © UKRI [2022]. All rights reserved.

The rest of the label is difficult to decipher but it is dated to 19 March 1845, signed RHB, and refers to a documented excavation. The rock is incorrectly identified as Blue Lias (a Jurassic-age rock that is far too young); but clearly sample ‘Wilts 277’ was directly taken from the Altar Stone, and so it can be added to the first group of bluestone descriptions. In turn, this gives a secure parentage to the recently described and very similar samples of calcareous micaceous sandstone debitage. This assertion is supported by direct comparison of recently obtained portable X-ray fluorescence (pXRF) chemical analyses between the in situ Altar Stone and ‘Wilts 277’; these data show them to be the same.

Cunnington and Kellaway

The Cunnington collection was stored away for around 70 years until Geoffrey Kellaway – a survey geologist and very keen proponent that the all the Stonehenge bluestones were glacial erratics – asked colleagues in the British Geological Survey in 1971 to re-examine and re-identify the thin sections. This they did, revising the Victorian rock names but perhaps wisely saying little about their geological provenance. However, Kellaway also asked them to section and look at a few other ‘foreign rocks’ excavated by Hawley in the 1920s from Stonehenge, including one labelled RSN18 ENQ 2305, a rhyolitic tuff (ignimbrite) that R S Newall (RSN) and Kellaway believed showed glacial striae (scratches or gouges from glacial abrasion). This broken joint block has taken on great significance in the ‘man versus ice’ debate of how the bluestones came to Stonehenge, for in 1991 Kellaway suggested that the rock was a better match to tuffs from North Wales than Pembrokeshire, and so represented tangible proof of glacial dumping on Salisbury Plain.

This image highlights the pale brown wispy stilpnomelane, which is a defining characteristic of rhyolites from Craig Rhos-y-Felin.

For 30 years, this stone and the Boles Barrow dolerite have been and are presently being used as the best evidence for this erratic glacial theory, often stressing a North Wales origin for the tuff. Today, the rhyolite joint block is safe in the collection of the Salisbury Museum, and has been linked to its excavation record by Tim Daw of; a rock slice and the BGS thin section taken from it are also held in storage in the BGS collections at Keyworth. We have recently reunited and examined the joint block and all its offcuts and associated thin sections, and the rhyolitic tuff shows all the key characteristics needed to assign it to Rhyolite Group C from Craig Rhos-y-Felin in north Pembrokeshire, over 170 kilometres south of North Wales – the ‘glacial striae’ are in fact seen to be linear expressions of the internal fabric of the rock, or slickensiding, along one joint plane. Additionally, geochemical data obtained from sample RSN18 ENQ 2305 by the Open University team in the early 1990s showed that it (alongside others of their debitage samples) belonged to Rhyolite Group C – although the group and its origin would only be recognised 20 years later. Indeed, very recent analysis by pXRF on all pieces of the joint block plus two other visually similar rhyolitic tuffs from the same Newell collection (RSN9 ENQ 2295 and RSN10 ENQ 2296) clearly show that these fragments are compositionally Rhyolite Group C, confirming the petrographic identifications.

Plots of zirconium vs barium and rubidium vs niobium (determined by portable XRF) for three Newall samples collected from Stonehenge, along with analyses from Craig Rhos-y-Felin in west Wales, the source of Rhyolite Group C.

Ironically, this Open University research was designed to demonstrate that all the bluestones were erratics, and RSN 18 ENQ 2305 (‘OU2’ in the list of samples that the OU examined) –originally hailed as unique/indisputable proof of glacial transport – is clearly shown to be just one of hundreds of similar rocks fragments from the Craig Rhos-y-Felin quarry site now scattered across the Stonehenge Landscape.

The rediscovery of the Cunnington slides and the Altar Stone sample and its label, plus re-examination of the ‘North Wales’ joint block are important reminders of the value of museum collections in preserving important research materials, especially early collections from settings where sampling is no longer possible on ethical or cultural grounds. It also shows the critical value of cataloguing – if we don’t know what is in museum or other institutional collections, then pivotal samples can’t be available for research. There are Victorian treasures in all their attics, with the promise of future joy as previously unrealised attributes are seen with our new eyes using modern analytical techniques that respect and preserve the material.

Dr Rob Ixer is an independent researcher, Honorary Visiting Senior Research Associate at UCL’s Institute of Archaeology.

Richard Bevins is an Honorary Research Fellow at Amgueddfa Cymru–National Museum Wales, and Honorary Professor in the Department of Geography and Earth Sciences, Aberystwyth University.

Nick Pearce is Professor of Geochemistry in the Department of Geography and Earth Sciences at Aberystwyth University.

David Dawson is an archaeologist and has been Director of the Wiltshire Museum in Devizes since 2008.
All images: Richard Bevins unless otherwise stated.