By Dr. Sandy Gerrard.

Following the discovery of the Bancbryn stone row in 2012 the Welsh archaeological establishment set about characterising it and after much deliberation concluded that it was not a prehistoric stone row for six main reasons:

  1. Rows are less than 200m long (Consultant reporting to Dyfed Archaeological Trust 14th February 2012).
  2. The overall alignment of the Mynydd y Betws alignment is sinuous in a form which is not typical for prehistoric ceremonial/ritual stone alignments which are….. predominantly straight’ (Cotswold Archaeology Report).
  3. The variable size and shape of the stones (Cotswold Archaeology Report).
  4. The stone density varies along the row, from stones every metre or so through to 10-15m gaps. (Cadw 24/01/2012).
  5. “Inconsistencies in the physical appearance of the stone alignment when compared with currently accepted Welsh prehistoric examples (Cadw response 15/08/13).
  6. The stones were not in sockets (Cotswold Archaeology Report).

Bancbryn stone row

Following the completion of a long term project looking at all the extant stone rows in Great Britain, it is now possible to access all the single rows using the same Welsh style criteria used at Bancbryn to find out how their unorthodox approach to interpretation affects our understanding of British stone rows.

A total of 174 accepted single rows most of which are scheduled as ancient monuments were put through the Welsh style interpretative mill and unsurprisingly only 32 were found to meet their strict criteria. The remaining 142 failed to clear at least one hurdle and many were rejected for several reasons. It is worth having a quick look at some of the fallers.

  • Rows are not longer than 200m in length

The Upper Erme Valley stone row at 3,386m is many times longer than the upper limit suggested.

  • Rows are predominantly straight

The poor old Upper Erme valley stone row also fails in this regard but is in good company.

The stone row at Hingston Hill is too long and far from straight.

Whilst Drizzlecombe 3 wiggles between a large pillar and a cairn.

  • Stone rows apparently are composed of similarly sized stones according to the Cotswold Archaeology assessment at Bancbryn.

It would appear that Drizzlecombe 3 has failed again.

A mixture of different sized stones does not appear to have influenced Cadw’s decision to schedule the stone row at Carreg Wen Fawr Y Rugos.

Whilst at Afon Hyddgen the archaeological evidence would seem to suggest that the Welsh approach to interpretation may need to be reviewed. Single rows composed of different sized stones are fairly commonplace with 49% of single rows sharing this characteristic and yet this was given as a reason for doubting Bancbryn’s authenticity.

  • Irregular stone spacing

Variable gaps between the stones were seen by Cadw as a reason for doubting Bancbryn. In the real world, fieldwork has shown that half of the single rows are far from regular in their layout and stone spacing varies considerably. Clearly, this could be a consequence of later damage and therefore is not really something that should have been considered in the first place.

At Butterdon Hill many of the stones are close together but in places, there are substantial gaps. At Bancbryn, Cadw saw this as a reason to doubt the row.

  • Inconsistencies in regional form

Cadw seems to be under the impression that because the row looks rather different from others in the region that it cannot, therefore, be a row. In reality, single rows vary considerably in form and whilst broad regional similarities have been identified there are many exceptions to this rule.

The stone row at Joan Ford’s Newtake on Dartmoor looks like those seen in western Scotland.

As does the Five Kings row in Northumberland.

And Harolds Stones in South Wales would be much more at home on the Isle of Mull.

Regional oddities are relatively common and the evidence would indicate that it would be unwise to dismiss rows that do not comply with a certain expectation. The single stone rows vary considerably in character, anomalies occur and this should have been acknowledged.

  • A lack of stone sockets

This is a difficult one to assess in the field without excavation.  However, at Butterdon Hill and Colvannick Tor excavations have demonstrated that even large stones sometimes had no sockets and there is strong evidence to suggest that the large slabs at Cut Hill were erected without sockets or packing. The lack of sockets, therefore, should not be deployed as a barrier to the acceptance of the Bancbryn stone row.

Welsh approved Rows

32 single stone rows managed to make it through the “Bancbryn Style” assessment. It is perhaps worth having a quick look at these as they obviously tell us what was expected at Bancbryn. These rows indirectly give us an insight into what the Welsh establishment thinks a prehistoric single stone row should look like. The first surprise is that there only three in Wales: Saith Maen (NW), Harolds Stones, and Bryntwppa.

Map showing the distribution of single rows that passed the Bancbryn style assessment process.

Saith Maen, NW in the Brecon Beacons fulfills the “Bancbryn Style” criteria.

Another “Bancbryn Style” pass is Bryntwppa, but here the stones appear to be in a long mound and may, therefore, be something entirely different.

It is perhaps significant that Saith Maen (NW) and Harolds Stones figured prominently in the formal assessment paperwork. All the other Welsh rows (many of which are scheduled) failed the special Bancbryn Style assessment test. They must have been sitting a different test as they seem to have passed despite failing one or more modules. The South West of England also does rather badly with only three fulfilling all the criteria thrown at Bancbryn. However, if you travel further north into Scotland, examples of perfectly acceptable rows are found in relative abundance.  When the criteria were being drawn up for the Bancbryn assessment there was a complete misunderstanding of the character of the resource being examined. Inevitably, an objective assessment taking full cognisance of all the variables simply did not happen and instead a wholly inadequate, contradictory and plainly inaccurate conclusion was reached. The failure to assess the site properly is beyond any doubt and means the conclusions are worthless. Imagine dismissing a circle of stones as prehistoric because the stones were different sizes and not arranged in a precise circle – sounds far-fetched, but this is essentially what happened at Bancbryn.

Ballymeanoch near Kilmartin meets the criteria and is clearly the sort of row that ticked all the boxes.

Callanish, West is another one that fulfills the criteria. Perhaps this is the sort of thing they had in mind when starting the assessment process.


The criteria used to interpret the stone row at Bancbryn illustrates a total misunderstanding of the archaeology being examined. When the same criteria are applied to the single stone rows in Great Britain only 32 (18%) make the grade. Unless one is willing to accept that there are only 32 prehistoric single rows surviving in Great Britain then the assessment criteria used at Bancbryn must be flawed. When properly assessed by considering all the evidence from the site together with that from all the other rows it is clear that Bancbryn shares characteristics with many other accepted rows and its identification as a plausible prehistoric stone row is both logical and currently beyond doubt.