Some of the ideas discussed in this blog are published in my new book called "The Stonehenge Bluestones" -- due for publication on June 1st 2018. After that, it will be available by post and through good bookshops everywhere. Bad bookshops might not have it....
To order, click

Saturday, 1 December 2012

Devensian Welsh Ice in North Pembrokeshire

Had a Eureka moment the other day, when I was sitting on top of Carningli, drinking my cocoa and gazing northwards, taking in the fantastic view of Cardigan Bay -- I could see Aberystwyth, Plynlimon, Cader Idris, Snowdonia, Llyn and Bardsey.  Most impressive......

Anyway, I realized that the experts have got it all wrong, and that during the Devensian NE Pembrokeshire was not so much affected by Irish Sea Ice as Welsh Ice.  The most up-to-date publication to portray the position of the Devensian ice edge is that of Etienne et al in 2006.  Their map is shown above.  We are talking about an ice edge position here at around 20,000 years ago.

We will forgive the authors of the article for showing the ice edge as a smooth line -- they were generalising.  But there is no way that the actual ice edge will ever have looked like this. it will have been much more closely adapted to the lie of the land, with fingers into the lowlands and embayments where the higher land acted to divert ice to right or left.  Another summary of this ice edge position is here, showing the local topography rather more clearly:

That's one of my maps, made while I was pondering on the southernmost extent of the Irish Sea Ice and coming to the conclusion that there are so many ice-smoothed slabs and perched blocks on Carningli and Dinas Mountain that the ice must have flowed across the whole of that upland area and reached the spectacular meltwater channel (a very old one) called Cwm Gwaun.  You will also notice that the lime on the map by Etienne et al is about a mile further south than mine.  You will also notice that on both maps there is an assumption that the ice has come in from the N or NNW.

But I have been very worried by the morainic accumulations at Cilgwyn and Waun Mawn, and have been wondering how Irish Sea ice managed to bend around the eastern flanks of Carningli and Carnedd Meibion Owen in order to leave these deposits where they are, when it melted.  Ice can do extraordinary things, but if it can, it prefers to do things the simple way.

So I have got it wrong.  The ice has actually crossed NE Pembs from the N or NNE, meaning that it was propelled by the Welsh Ice Cap rather than by the Irish Sea Ice Sheet.  This is my new scenario:

Suggestions for ice movement directions at the peak of the Devensian Glacial episode in NE Pembrokeshire.  On the map, the blue arrows represent Welsh ice and the orange arrows represent ice from the Irish Sea Glacier.

There is a great deal of evidence in support of this scenario:

1.  Strong evidence of ice streaming on the ice-smoothed slabs on the NE-facing edge of the Carningli upland, overlooking Newport.  This evidence is best explained by ice coming from the NE.
2.  The morainic accumulations at Cilgwyn and Waun Mawn are best explained by ice pressing in across the Brynberian lowlands from the NE.
3.  The evidence of ice smoothing and perched blocks at Carnedd Meibion Owen is best expolained by ice movement from the NE.
4.  The enigmatic accumulations of fluvioglacial sands and gravels in the Monington area (near Moylgrove) are best explained as accumulations on a wasting ice contact edge, between Irish Sea ice to the west and Welsh ice to the east.

Note the "excrescences" or rough areas on this image of the Monington area, just to the NW of the centre of the image and also to the SW of the centre.  These are the accumulations of sands and gravels, exploited today in two gravel pits.  Note that elsewhere the land surface appears smooth. (Thanks to Henry Patton)

5.  Till deposits exposed on the beach at Newport since recent storms have considerable quantities of red, purple and pink sandstone clasts in them -- suggesting strongly a North Wales origin.  There are also strange metamorphic rocks in the Nevern estuary on the foreshore (samples sent to Rob Ixer) which I strongly suspect as having a N Wales origin.
6.  There are faint traces of marginal channels on the north face of Preseli which suggest an ice gradient sloping westwards.  Again this suggests ice coming from the NE.

This does not alter the fact that there might have been a later slight advance of Irish Sea Ice across the Pembrokeshire coast in this area  since there is heavy clay till at Gwbert, Parrog and other places which appears to be a very typical Irish Sea till.  There are also striae on the coastal rocks at Gwbert and Parrog which show that ice has come across the coast from the N and NNW.

This is just a start.  I have to give this much more thought.......

The relevance of all this as far as archaeology is concerned?  Quite considerable, since both Rhosyfelin and Carn Meini  (and Preseli generally) are in the frame here.  If we can tidy up on the details of chronology and the precise sequence of geomorphological episodes in this region, that will help us greatly in determining what happened at these sites which certain people are pleased to call "quarries."  By the way, Glacial Lake Brynberian is still in the frame here, as seen in the map at the top of the page.  (It's the lake shown as overflowing via the Rhosddu Channel at 220m asl.  I'm not convinced about that overflow, and that needs to be looked at..........)


olwyn said...

Thanks for this interesting post Brian - does the presence of rhyolitic field stones around Llech-y-dribedd, and at Bayvil, similar to the rocks at Craig-rhos-y-felin fit with your new thinking ? There are also morrainic accumulations near Hermon, outside of Crymych, including blue and white rhyolites, and large wedge shaped erratics (both dolerite and rhyolite) with striations. I continue to scratch my head...

BRIAN JOHN said...

Thanks Olwyn -- interesting! Can you give me some grid refs and field notes on all of this? It will be interesting to see what the Trefael Stone turns out to be made of. The trouble is that NE Pembs generally hasn't had any serious Quaternary field work in recent decades, apart from the work connected with the Teifi Lake story. The problem with glacial deposits that have come from the NE (with Welsh Ice) is that the grits, sandstones and shales of mid Wales are rather nondescript, and I certainly would not feel confident on any provenancing from that great tract of sedimentary rocks between the Teifi and Plynlimon. the North Wales rocks may well hold the key -- but again I'm not familiar enough with them. Maybe Richard and the other guys from Cardiff can help?

Yes, I'm aware that there are rhyolites and even spotted dolerites in the area north of the Nevern valley -- OT Jones drew attention to them many years ago. I am increasingly coming to the view that there have been many different ice movements across this area, in various phases of both the Anglian and Devensian Glaciations. Local ice from Preseli, Welsh ice from the NE and Irish Sea Ice from the NW all waxing and waning at different times. It's a really confusing scenario.....

chris johnson said...

I am surprised you think the exit for the putative lake is to the West.

The North would seem more likely, perhaps even Hermon. Take a look at the streams feeding the Teifi. Even today Prescelli drains in that direction and with surprising force - Cenarth is a good place for lunch!

Olwyn said...

OK - further info - accepting that I'm not a skilled geologist, so my rhyolite might include ashes and tuffs but I can distinguish between these laminated bedded fine grained volcanics and the more massive and crystalline dolerite - at the end of the lane to Llech-y-Dribedd dolmen near Moylegrove I found a pile of rocks raked off the ploughed field, have some samples, they are blue grey, laminated, finegrained, with little cream inclusions ?accretionary lapilli? the grid ref for these is SN10124320. Bayvil - collected samples from the spoil heap of George Nash's recent excavation,(trefael) and peered closely at the chipped side of the cup marked rock - again, blue, fine grained, laminated, much like Craig-rhos-y-felin and nearby outcrops. The Trefael stone is worn and smoothed though. grid ref SN10254025. The dump near Crymych is in the vicinity of Craig-y-Fran SN18553125. This site is marked on the OS map as an area of rocky scrub, it is full of the most amazing rocks. Some years ago there was a bit of archaeological interest when some 'white rhyolite' neolithic stone axe working debitage and polishing stones were found in the fields downslope of the scrub. There is a paper on this somewhere but I don't have the reference immediately to hand (if you're interested though, I could get a copy for you). There does not seem to be an outcrop - the raw material is derived from the glacial deposits. A similar axe working area was also identified near Glandy cross, also using similar material, also in the topsoil. This is thought to be one of the sources of the 'group VIII' axes.

I know what you mean about the sandstones from up in Ceredigion ! Again, I have a few samples,and am fairly convinced that what I found at Llech-y-dribedd is not the same.

I was also working on a dig on a hill top southwest of penparc a couple of years ago (pant-y-butler)and saw a blue finegrained lemon-pip shaped erratic boulder there which seemed to have been incorporated in a bronze age round mound. I wasn't so skilled at distinguishing rhyolite from dolerite then, but it was chipped by the digger and quite a bright blue inside. Can't find the right map just now for grid ref, sorry. But there are 'bluestone' gateposts in that vicinity, too.

It IS all very confusing ! Hope this is of interest anyway :)

Olwyn said...

Hi Brian - further to yesterdays message, the grid ref for the Pant-y-Butler mound is SN21454675. At the time the stone was identified by someone from DAT as 'Fishguard Volcanic series'. I have wondered whether there could be further volcanic outcrops out in Cardigan Bay, which have been scraped by the Irish Sea ice and delivered onto the coastal area in N.Pembs/S.Ceredigion. If I survive long enough, I'd like to take some of my samples round the on-land outcrops and compare them..

Olwyn said...

Hi Brian, did you get the grid refs and stuff ?

All the best,

BRIAN JOHN said...

Sorry all -- I have been away in Brussels for 3 days, making presentations to the Parliamentary Petitions Committee.. back home now, feeling exhausted......

Myris of Alexandria said...

Olwen macroscopical description and id of these rocks is a sure way to madness. Sadly they have to be sectioned.
Trefael Stone (about which I can no longer say much without permission) is not Cryf and I am afraid does not really look much like it macroscopically-it has visible feld.microphenocrysts; at one time I thought it looked more like SH48.
It is not that either or trust me we would all have known about that.
Matching by eye or by PXRF can only go so far and a good thin section, better polished thin section is the only certain way.
The point about Cryf is that it is petrographically NOT NOT NOT like most of the rhyolites/rhyolitic tuffs.
These rocks, as the 1990s OU work and the I and B work this century has shown, are not for amateurs. It has taken us years to recognise the subtle differences and to refine the descriptions and classifications and of course remains on-going.

Constantinos Ragazas said...


The “excrescences” at Monington raise an interesting curiosity. Could similar “excrescences” of sand and gravel formed at drained lake bottoms at places with gullies surrounding such plateau? Where the water flow is much less and so less able to carry these coarse material further downstream?

If you look at your posted photo, the plateau where the Monington “excrescences” are located is surrounded by drain gullies. Thus the raised plateau acts like a “filter” catching the coarser matter in the flow. A little like a more “elevated” discussion!

Just some intellectual musing … you can disregard if you like!


Olwyn said...

Rob - you may well be right but I have been mad for years so possibly wouldn't be able to recognise it..if you have the where with all to do thin sections, I have a variety of samples..

BRIAN JOHN said...

I understand where Rob is coming from -- he is after all a petrologist, and he would say that we must all have thin sections, wouldn't he? OK -- I respect that. But I wouldn't be so dismissive of identifications from hand samples. When I was a research student I knew the rocks of the St David's area so well that I could pretty well pick up any chunk of rock from a beach or from a till exposure and tell you where it had come from. I didn't need thin sections to do that. For identifying major groups of rocks (red Cambrian sandstone, grey Carboniferous Limestone, bright red ORS, and even spotted dolerite, for example) it would be a ludicrous waste of time even to think about thin sections as an aid to identification. You bring in that technique if you want extremely accurate provenancing -- and even then you need a lot of luck as well as judgment!

BRIAN JOHN said...

That info on sites is very interesting, Olwyn, and many thanks for it. I will check out some of those locations......

Anonymous said...

Brian, you write

“The enigmatic accumulations of fluvioglacial sands and gravels in the Monington area “

What makes these “excrescences” enigmatic?


Myris of Alexandria said...

MMMM I am channelling
"I agree with experience much can be done, I can now tell SH48 knock-offs in hand specimen (If I have a bit beside me to compare but I only discovered in July that one of Stone's stones aka the shoebox assemblage was a 48 knock-off and I must have handled that stone 30/40 times). IF you think that macro id is better than 70-80% you are fooling yourself.
My comments really were restricted to f.g rhyolitic rocks. Plus we are dealing with precise provenancing there even with lmst you would need TS petrography to give a location.
In approx. 150 years it has been recognised that for id'ing of rock thin section (I would of course say thin section petrography in both transmitted and reflected light my so-called 'total petrography')is the supreme queen of techniques (a bit like leeks and veg).
To recognise SH rhyolitic or volcanic bluestones in the field ONLY TS petrography (better alongside geochem- we have been here before Olwen- did we not all co-write something along these lines? (Note MR Johnson an example of hidden readers being addressed))will show us the one true path.
Do not fret I noted your little pile of stones and it will not be forgotten). Allow me to get Brian's done first- these days favours in academia are as rare as
SH provenance sites.
I did try to tap the owner of the NW England fracking (I always want to say fragging!!!) company for TS money but no luck mind you he tried to tap BP's Tony Hayward years ago in the same vein (WE are all highly and long time associated) but with the same result.
I shall write to Mr Gates. Him I do not know but what is a few hundred pounds to Micronesia.

Ah the mists are clearing and Sublime Apollo as Ra greets us once again.

Myris of Alexandria said...

Oh Dr Ixer is a petrographer -not bright enough for a petrologist- Dr Bevins is a petrologist.
Secretly Dr Ixer feels petrology to be dull all that modelling.

Myris of Alexandria said...

Wrong Olwyn.
People should not use single names to hide behind it is all very confusing.
I am sure Myris would not call a complete stranger mad, he would wait a few minutes, assessing quite how mad and then modify the noun using an extravagant adjective.
I expect he feels rather crushed and may try to be benign for ten minutes.
Thomas Rhymer.

BRIAN JOHN said...

I was about to point that out. Olwyn and Olwen -- both excellent contributors to our little discussions. At least it's easier with names than with multiple people called "Anonymous" -- when you have no idea who they are and what they are up to....

BRIAN JOHN said...

Kostas -- enigmatic excrescences! they are excrescences because they look like nasty boils or bubbles on the skin that have burst -- without getting too graphic about it. They are enigmatic because I can't quite work out why they are just there rather than anywhere else -- in the middle of a large area which has all been glaciated relatively recently, fairly high up, and not in an area where you might expect some topographic control over the position of an ice margin. They have all the appearances of kames, formed through the accumulation of sands and gravels in temporary lakes bounded at least in part by walls of ice at a time of deglaciation. There are signs that some of them might be eskers -- formed in subglacial tunnels. I am still working on it.....

Constantinos Ragazas said...


Thanks for explaining your “enigmatic excrescences”. You have in the past sensitized me to the perils of ideas based on scant knowledge and information. So consider the following brainstorming as tentative intellectual musings. Based on just what I see in your posted image and my “sensible reasoning” on this. For whatever its worth …

From your posted image of these “excrescences”, I see the following. Correct me if I am wrong.

1) A generally flat area where these “excrescences” are located somewhere in the middle.
2)To the right (of the image) of the “excrescences” there is a flat circular lobe extruding with drain gullies along much of its perimeter.
3) The shape of the “excrescences” area appears to be crescent, with the greatest buildup of sand/gravel along the very well defined and steeper inner side of the crescent.

Question: What geological process can account for these facts on the ground?

My (tentative) answer:

Possibly meltwater streams, with ice embankments, flowing through a narrow opening into a wider open flat area (the “circular lobe” in my description above) with egress drainage (the drain gullies in my description) along the perimeter of this circular lobe.

The rate of water flowing into the “lobe opening” through the narrow passage will be much greater than the rate of flow at and over the perimeter of the “lobe”. In fact, (the “rate in”) = (the perimeter of the lobe) x (the “rate out”). Put in simple clear terms, the water velocity at the “lobe opening” will be far greater than the water velocity at the “lobe” and over its perimeter. Thus, any coarse material (like gravel and sand) will be deposited near the “lobe opening” and not further into the “lobe” where the water velocity will be much less.

This explains the position and crescent shape of the “excrescences”. Want to coauthor a paper with me?!