The Brimham Grit formation crops out across an estimated 150 sq km of countryside, covered in part by glacial deposits. Then we have the Lower Coal Measures................
I am more and more intrigued by the apparent obsession -- on the part of assorted geologists and archaeologists -- with taking a lump of stone from an archaeological context and claiming to have found the exact location, to within a few square metres, of the spot from which it was taken I choose my words carefully -- because there is an underlying assumption that said lump of stone must have been selected and fetched by human beings at its place of origin and then carried across land or sea to its final resting place. Underpinning this assumption is the belief that naturally transported stones (such as glacial erratics) were somehow deemed inadequate, and that the fetching of the stone from a faraway location was a superior activity within the Neolithic belief system. All very weird when we consider that in glaciated parts of Northern Europe burial chambers and stone settings generally were always made of glacial erratics if they were available and from broken bedrock slabs and pillars if they were close at hand -- as at Callanish. But nobody carried stones across country for great distances, as pointed out many years ago by Geoffrey Kellaway, Stephen Briggs, Olwen Williams-Thorpe and others. There is no sign at all that glacial erratics were considered inferior for megalithic building projects. Indeed, we can make a case that they were more highly valued than "quarried" stones:
https://brian-mountainman.blogspot.com/2026/08/the-special-status-of-unhewn-rocks-and.html
And yet Parker Pearson, Ixer, Bevins, Clarke and many others persist with their illogical conviction that monoliths needed to be FETCHED in order to be deemed useful.
All this comes sharply into focus because of the latest offering from Clarke et al on the origins of the Devils Arrows stones. I was forcefully reminded of how easy it is to sell an idea through the careful choice of the method you choose for the portrayal of your data. As I have mentioned in another post, their basic problem is a lack of data. That may sound strange, given that the article looks hugely impressive, packed with technical complexities, tables and graphic representations. But in attempting to assign an origin to the megaliths in question, they have only ONE serious "candidate" sampling point -- Brimham Rocks. On that basis alone, the article should be laughed out of court. But let's look at their Figure 6, which is a "multidimensional scaling plot" based on 60 comparative data sets.
Deliberate prehistoric sourcing of the Devil's Arrows, Britain's tallest stone row Open Access
Anthony Clarke, Jim Leary, Chris Kirkland, Proc. R. Soc. A (2026) 482 (2345): 20260504 .
https://doi.org/10.1098/rspa.2026.0504https://royalsocietypublishing.org/rspa/article/482/2345/20260504/483254/Deliberate-prehistoric-sourcing-of-the-Devil-s
The Devils Arrows "field" or blob on the map is an artifice -- it appears to be an aggregate obtained by amalgamating the sample results from the 4 monoliths. So it is a phantom. It does not represent any of the real world monoliths. This is completely unacceptable in an article that purports to have discovered that monoliths A, B, C and D have all come from point source E. Statistical manipulation gone mad.
The authors of this paper have not demonstrated that the 4 sampled monoliths that constitute the Devil's Arrows have actually come from the same place. Three of the sampled monoliths are incorporated into Fig 3A, but there is no trace of sample 4. It might be wildly different from samples 1, 2 and 3, or it might look rather similar. But it is incorporated into an amalgamated sample anyway. That is completely unacceptable. The number of grains analysed (total 146) is also unacceptably low for a study of this type.
Another fundamental flaw in this study relates to sampling methods. The samples for the 4 monoliths were taken with sticky tape, which introduces a massive bias that should have been spelt out. At best each sample should be spelt out as being "partially representative" of the real world situation. This might have been acceptable if the Brimham Crags sample had been taken in the same way. But we learn that the the sample from this potential source area was taken from a whole rock sample knocked off a bedrock outcrop (no actual location provided) and weighing c 1 kg. So you are not comparing like with like. No matter what statistical wizardry you employ in your defence, this invalidates the whole study. On this basis alone, the study should have been rejected outright by the publishers.
Figure 3A looks very impressive, but we have no idea how similar, or how different, the 3 chosen samples look when seen alongside samples taken in the same way from comparative or control sites. To make matters worse, later in the paper, these sticky tape sample results are compared with multiple samples analysed by Andrew Morton and others from across the Millstone Grit and Coal Measures outcrops which were NOT sticky tape samples. Again, unacceptable.
In Figure 4C the authors claim that "Zircon concordia ages older than 510Ma showing shared peaks between these datasets and divergence from Plumpton Rocks, supporting a Brimham Rocks source." The visual match is not very close at all -- you do not need a computer to tell you that. And because there is a poor match with a "straw man" sample (Plumpton Rocks), that does nothing whatsoever to strengthen the claim of a Brimham Rocks source. That's a bizarre and utterly biased interpretation of what the diagram shows.
Brimham Rocks -- spectacular, but that does not mean it was a source for distant monoliths
Back to Figure 6. The extra data sets portrayed on the diagram were of course carefully chosen from the literature. The scale of the graphic representation is very small, with the samples for Brimham Rocks and the Devils Arrows shown close together in the middle. The closest match to the Devils Arrows is not Brimham Rocks but a sample from the Pennine Basin Lower Coal Measures, in an article by Morton et al, 2024. (This is a Langsettian sandstone from the Pennine Basin Lower Coal Measures at Binchester Crags, approximately 70 km to the north.) But there is no proper discussion of this as a possible source for the Devils Arrows, apart from a cursory dismissal on the grounds that the Late Devensian or LGM ice flow directions were not deemed suitable for ice transport. Again this is completely unacceptable, since as the authors acknowledge, erratic monoliths might have been moved multiple times during a number of glacial episodes.
Thorpe RS, Williams-Thorpe O.1991 The myth of long-distance megalith transport. Antiquity 65, pp 64–73. (doi:10.1017/S0003598X00079308)
https://www.cambridge.org/core/journals/antiquity/article/abs/myth-of-longdistance-megalith-transport/F16F6E44F8B8E1040E705748E4DCF9D0
If we eliminate all the "noise" from this graphic representation, and use a large scale instead of a small one, we see that Brimham Rocks is NOT a perfect match, and on this basis alone the authors should have undertaken much more comprehensive sampling across outcrops of both the lower Coal Measures and the Kinderscoutian Brimham Grit to the north and west.
It's all too easy to forget, in this obsessive hunt for precise provenances, that in nature we have continuums and gradations and not neatly packaged boxes of rocks that can be given different labels. Obviously, in our search for order and in the quest for knowledge, we do use labels (sandstone, limestone, dolerite, rhyolite etc) but the parameters that we use are man-made artifices, including all the units of measurement, and there are endless gradations between one rock type and another. Hence the debate between me and Ixer and Bevins about the number of rock types represented in the Stonehenge erratic assemblage.
One of the big lessons I took away from my doctorate research in West Wales in 1962-65 (on the Quaternary sediments and glacial chronology of North Pembrokeshire) is that it is easy to create labels -- Irish Sea till, Lower Head, Upper Head, Raised Beach, Blown Sand etc -- but remarkably difficult in the field to identify these strata because the gradations are so subtle. I spent hundreds of hours in the lab weighing, drying, sampling, sieving, and identifying stone shapes and sizes so that I could convince others that I knew what I was talking about, but I knew that the methods I chose for the representation of my results were those that best suited my purpose. But I had to be prepared to face the challenges of others, not least in a long face-to-face scrutiny session with my examiners.
One of my triangular plots from 1965, showing particle size distributions for key sediment types in the Quaternary succession in North Pembrokeshire. The labels are of course subjective and arbitrary, and there are overlaps and gradations everywhere........
One of my triangular plots from 1965, showing the particle size distributions. No statistical manipulation. Useful, but what the plots show is that nature is very messy......
Map of glacial and related deposits around Rhos-y-Clegyrn. A lot of fieldwork went into this, but the boundaries are invisible in the real world, and are at best educated guesses. You cannot sample every square meter of territory, or improve reliability by statistical manipulation.When I was trying to demonstrate the particle size distributions of samples from different stratigraphic units I used triangular diagrams, and indeed they did show considerable lithologiocal differences between Irish Sea Till, fluvioglacial and other deposits......... But while nature is very organized and obeys the laws of physics, to us it often appears disorganized; and within the Irish Sea till group of exposures there were some samples with an unusually high content of gravels, and within the fluvioglacial group there were some samples with a high clay content. The same variability is found within solid rock formations, and it is thus very frustrating when Clarke et al claim to have represented the Brimham Grit after collecting just one sample. That sample might just be wildly unrepresentative, and a much better match might be found elsewhere.
If somebody was to present me with a 2 kg bag of Irish Sea Till and ask me to tell them where it comes from, my response would be "Man, you must be joking...."
We do not know everything about everywhere. The claims by Ixer and Bevins to have found the precise source for the Stonehenge spotted dolerites at Carn Goedog and and for certain Stonehenge foliated rhyolite fragments at Craig Rhosyfelin are simply absurd, since they have no idea where the same rock types may exist in the unsampled landscape. Besides, they have never produced any perfect matches.
Precision provenancing is a fantasy, sold to the media and a gullible public by academics who should know better.
===============
PS. Of course, the fatal flaw in the MPP narrative and the Ixer / Bevins provenancing work is that it is all based on a single premise -- namely that the glacial transport of the bluestones is unsupported by any field evidence, and that it therefore could not and did not happen. The premise is false. Glacial transport was perfectly possible and indeed probable, given the wide range of supporting evidence presented in the literature.
No comments:
Post a Comment
Please leave your message here