Some of the ideas discussed in this blog are published in my book called "The Bluestone Enigma" -- available by post and through good bookshops everywhere. Bad bookshops might not have it....
To order, click

Tuesday, 30 August 2011

Irish Sea Glacier Lobe v Celtic Sea Piedmont Glacier

Have a look at these two maps.  They show two fundamentally different scenarios for the Devensian maximum glaciation -- one showing Irish Sea ice flowing in a long narrow lobe from NNE towards SSW -- starting in St George's Channel and extending out into the Atlantic approaches.  That's the scenario accepted by many senior UK geomorphologists today.  As I have argued before, a lobe of this shape is inherently implausible because ice does not tend to flow in long thin lobes when there are no topographic constraints -- it tends to spread out into wide fan-shaped lobes instead.  Also, I cannot see how an ice margin very close to present-day sea-level in the Scilly Islands (if the C14 dates are to be believed) can have also been very close to the same altitude in west Pembrokeshire, more than 200 km away.  that's not just a flattish gradient -- it's a zero gradient.  And however much we might argue about basal shear stresses and saturated sediments on the glacier bed, the basic rule is that a glacier will only maintain forward momentum if it has a continuous surface gradient which accords with the direction of ice movement.

So how do we get an ice edge close to present-day sea-level all the way from Pembrokeshire to the Scilly Isles?  The only way, as far as I can see, is to have the ice gradient sloping down from NW towards SE -- which means powerful ice coming off the SE coast of Ireland, thick enough and active enough to "block off" the ice coming down from the N or NE from the Irish Sea and across Cardigan Bay.  There will also have been contact with Welsh ice, coming down from the uplands of mid and south Wales.  This will give us a very confusing picture in Carmarthen Bay, the middle reaches of the Bristol Channel, and down along the coasts of Devon and Cornwall.  The islands like Caldey and Lundy may be crucial in sorting out exactly what happened.......  and we also need to examine the uplands and coasts of SE Ireland to determine whether the ice was really thick enough to sustain movement right across these Celtic Sea approaches and all the way across to the SW coasts of Wales and England.

The outer limit of Devensian ice shown on the bottom map, from Glamorgan to Devon and Cornwall, may be anything up to 50 km adrift in places.   The field evidence needs to be reexamined in order to sort out its precise position.

Saturday, 27 August 2011

The Celtic Sea Piedmont Glacier

Following my recent posts, I can see no great reason to maintain the pretence that the southern part of Pembrokeshire was totally ice-free during the Devensian.  The evidence from Caldey and elsewhere indicates that the same stratigraphic sequence of Late Pleistocene deposits occurs in South Pembrokeshire as in the north of the county -- ie with glacial and fluvio-glacial deposits overlying head and overlain by sandy brickearth or loess material which forms the basis of the modern soil.  Beneath this whole sequence is the raised beach, cemented solidly where it lies on Carboniferous limestone.

So I'm modifying the map which I pasted on this blog a few months back as below.  My old map is above, and the modified one is below.
If thin ice covered South Pembrokeshire and Caldey Island, it quite possibly coalesced with the ice coming down to Carmarthen Bay from the Towy and Taf valleys.  The signs are that it had a very flat gradient, and did very little, in this area at least, in the way of landscape modification.  I think the ice edge may well have reached the coasts of Devon and Cornwall in places -- maybe not everywhere.

It follows that in eastern Pembrokeshire and western Carmarthenshire there would have been a complex zone of coalescence between Irish ice, Irish Sea ice, and Welsh ice.  Where, when and how?  That is something to be worked on.....

As Henry Patton has shown with his Devensian ice cap modelling, the outermost ice limits in different zones may not have been reached contemporaneously -- the line (if we need one at all)  may be a composite feature, with pulses of forward movement here and there, and out of phase.

And I still think it quite possible that just as a small Preseli Ice Cap existed intermittently in the Devensian,  there may have been thin short-lived ice caps on the uplands of SW England, on Dartmoor, Exmoor, Bodmin Moor and the Mendips, for example.

So the ice edge may have been very fuzzy indeed, with no clear demarcation between forward-moving and far-travelled glacier ice from the west, locally-generated ice on high ground, and areas perennially and intermittently covered by extensive snowfields during the coldest phases.

The Glaciation of South Pembrokeshire

I have been digging around in some of the old records of the Geological Survey -- from about 1905-1938 -- and have found that the old geologists  responsible for the fieldwork in west Wales (Strahan, Cantrill, Dixon, Thomas and Jones) knew all about the glacial deposits scattered across South Pembrokeshire.  In their publications they describe many locations where thin till and glacial erratics are to be seen -- including Bullum's Bay on Caldey Island.  So I'm not the first to describe this till, by any means.  In 1905 EL Dixon wrote of Bullum's Bay:  "..... the glacial deposit appears to overlie the raised beach, although the exposure is obscure, and the evidence of superposition is not so conclusive as in Gower." (Summary of Progress for 1905, Mem Geol Surv, p 70).  He and his colleagues described glacial deposits at Landshipping, in the inner reaches of Milford Haven; till about 7 ft thick at St Florence; glacial sands and gravels at Bubbleton; gravel and sand at Norchard; sandy loam with erratics at Lamphey; and glacial deposits in many pipes and solution hollows at Catshole Quarry, Pembroke, Sandtop Bay on Caldey, and in other places where Carboniferous Limestone is found.  Generally the till in such places is coloured red or pink -- with an obvious association with ORS rocks, and striated sandstone pebbles are also recorded in the deposits, as are pebbles and larger stones of greenstone and felsite.

The Geological Survey maps for the Tenby districts show patches of glacial deposits, as do the maps for the Pembroke and Carmarthen areas.  Some of them are shown on this map, from one of the papers by Prof David Q Bowen:
You can see the deposits on the map in the area  (bottom left corner) that was supposedly ice-free during the Devensian.  Somehow or other, in all the research work that was going on in the 1960's and 1970's, we all slipped into accepting that south Pembrokeshire simply had a few patches of very old till left here and there, and that these were by definition "Older Drift" in comparison with the "Younger Drift" deposits found in North Pembrokeshire.  Nobody worked in South Pembrokeshire, maybe because it was deemed to be not very interesting -- and we all took our eyes off the ball.  Before we realised what was happening, this idea of "new" Devensian deposits in N Pembs and older deposits in S Pembs turned from a working hypothesis into a ruling hypothesis.  I criticise archaeologists often enough for allowing their thinking to be determined by a foolish adherence to a ruling hypothesis -- but exactly the same thing happens in other sciences too, including geomorphology.  With respect to the "old" glacial deposits of South Pembs, none of us took the trouble to re-examine the evidence on the ground as originally recorded by Dixon, Cantrill et al -- and none of us gave their descriptions the respect that they deserved. Mea culpa!

Then we all became obsessed with lines.  If you look at the scores of learned papers written on ice limits in West Wales over the past 20 years, they have all been based on the ruling hypothesis that the Devensian ice limit lay somewhere across Pembrokeshire -- maybe a bit to the north or a bit to the south of the line shown on the map above.

 Three maps -- top, by DQ Bowen; middle, by Jansson and Glasser; bottom, from my blog.  the differences are not all that great.  Note that on my map Caldey Island is shown, slap in the middle of the supposed ice-free area during the Devensian.

There has never been any strong stratigraphic or dating evidence to support the positioning of this line, and I am increasingly convinced that we have all been wrong about this ice limit, and that the glacial deposits of the supposed "ice free" area on the map above are actually the same age as those of the St David's Peninsula and North Pembrokeshire.  They are thin and patchy, and have very little or no surface expression (ie there are no mounds or hillocks or obvious moraines, drumlins or eskers) but I think that might be because the glaciation of this area was short-lived and that the ice was thin.  Maybe it was flowing very slowly over frozen ground, and carried very little debris.

I once wrote a short paper called "Ice Age events in South Pembrokeshire" in which I argued that the animal and human remains in the Caldey and other caves dating from approximately the Devensian maximum showed that all these inhabitants were there while the ice edge lay somewhere to the north and west.  In retrospect, that was not a very sensible argument -- and I think we now have to accept that the ice did cross this tundra area, flowing from west towards east, maybe just for a few centuries, forcing an evacuation of all living things -- which then moved westwards again when the ice began to melt, re-colonising their old hunting grounds (or rather, the hunting grounds of their ancestors.)

Back to the drawing-board, chaps.......

The Glaciation of Caldey

The map above is from the 1950's, and Caldey is mis-spelt.  But the key locations are shown.  The contact zone between the ORS and the Carb Limestone (with gash breccias and a clay band) runs from Sandtop Bay in the west to Drinkim Bay in the east.

Having found that till exposure in Bullum's Bay, I went hunting for more of the same, elsewhere on the island.  Examining the cliff exposures at the west end of the island is extremely difficult,  since there is thick vegetation (mostly gorse and brambles) covering most of the best "exposure" sites, and since there is a herd of semi-wild (ie rather fierce) beef cattle (including cows, very small calves and a bull) roaming about, unconstrained by fences or walls.  So one has to lurk about, trying to avoid these monsters.  The ORS cliffs (south of a line between Sandtop Bay and Drinkim) are generally unpromising, since they are very exposed to wave attack and they are also fringed with thick vegetation.  But from my quick recce I think there may be till exposed in the cliff face of Sandtop Bay on parts of a slope covered with blown sand and small dunes:
I need to go back there to examine the reddish sediment exposures shown high up on the cliff face at the back of the bay, near the twin rock buttresses -- from a distance this looks like till, but this is precisely where the contact zone between ORS and Carb Limestone runs, and the contact is marked by thick gash breccias, with reddish weathering residues in a network of chaotic fissures.  The locals also refer to a band of soft clay, called "ochre" and which has been used in the past for lining ponds etc.  That has nothing to do with the Ice Age, and it may be of Triassic age......

I examined a small stretch of the cliffs between the jetty in Priory Bay and Eel Point, and I think there is thin till on the clifftops and in gullies, overlying a limestone "head" or periglacial deposit.  In places that head looks as if it has neen subject to frost-heaving or cryoturbation -- a pood sign of permafrost prior to the arrival of the glacier ice.

The best exposure of till I have seen on the north coast is immediately adjacent to the jetty.  If you stand on the roof of the whitewashed public shelter, this is what you see just a few yards away:

The till is quite thin, maybe up to 1m thick, but it is full of erratic pebbles of all shapes and sizes -- and it looks just like the till found in Bullum's Bay.

The other geomorphological feature of interest on the island is a little group of small dry channels on the clifftop of Red Berry Bay (on the ORS coast of the south of the island) which look as if they have been cut by meltwater spilling over from the west coast between West Beacon Point and Sandtop Bay;  that suggests to me that the direction of ice movement (and meltwater movement) over the island was from west to east.

What we now need to do is to tie in these deposits to the cave sequence for the island, building on the work of Brother James, Mel Davies and others back in the 1970's.......

Friday, 26 August 2011

Daylight Rock references

I'm increasingly convinced of the great importance of the eastern end of Caldey Island in sorting out the Devensian history of Wales.  Here are some refs I have come across:

Site: Small Ord Point (Daylight Rock Cave), Caldey Island, Dyfed


Bronk Ramsey, C., Pettitt, P.B., Hedges, R.E.M., Hodgins, G.W.L. & Owen, D.C. (2000) Radiocarbon dates from the Oxford AMS system: Archaeometry datelist 30. Archaeometry 42: 459-479.

David, A. (1989) Some aspects of the human presence in West Wales during the Mesolithic. In Bonsall, C. (ed.) The Mesolithic in Europe. Edinburgh, Donald.

Davies, M. (1976) Small Ord Point, Caldy Island. Archaeology in Wales 16: 25.

Gibson, A. (1995) First impressions: a review of Peterborough Ware in Wales. In Kinnes, I. & Varndell, G. (eds) 'Unbaked Urns of Rudely Shape': Essays on British and Irish Pottery. Oxbow Monograph 55. Oxbow Books, Oxford, pp. 23-39.

Hedges, R.E.M. et al. (1994) Radiocarbon dates from the Oxford AMS system: Archaeometry datelist 18. Archaeometry 36: 337-374.

Jacobi, R.M. (1980) The early Holocene settlement of Wales. In Taylor, J.A. (ed) Culture and Environment in Prehistoric Wales: Selected Essays. British Archaeological Reports British Series, 76: 131-206.

Lacaille, A.D. & Grimes, W.F. (1956) The prehistory of Caldey. Archaeologia Cambrensis 104: 85-165.

Lacaille, A.D. & Grimes, W.F. (1961) The prehistory of Caldey, Part 2. Archaeologia Cambrensis 110: 37-39, 60-63.

Valdemar, A.E. (1970) A preliminary report on the archaeological and palaeontological caves and rock shelters of Wales. Transactions of the Cave Research Group of Great Britain 12(2): 113-126.

Devensian Till at Bullum's Bay, Caldey Island

The fresh (Devensian?) till at Bullum's Bay, near the easternmost tip of Caldey Island.  It is preserved in a limestone gully, just above HWM.

On a trip to Caldey island, to give a couple of lectures to the Brothers, I managed to do a bit of exploring, and -- to my great surprise -- found a deposit of till very close to the eastern tip of the island, in Bullum's Bay.  This was at grid ref 147966.  The location is  a gully or small chasm in the Carboniferous Limestone, where the strata are standing almost vertically.  The gully is only about 4m wide and a similar depth, but somehow or other till has been deposited here and has survived.  The exposure is not far above HWM -- immediately below the exposure, which can be easily examined, there is a pile of flotsam and jetsam flung up in winter storms.  Quite possibly, during extreme gales coming from the SE, storm waves actually affect the exposure itself...... and wash stones out of it, which then collect on the beach below.
 Close-up of the till exposure, showing the reddish matrix of sandy and gravelly material and the frequent stones, many of which are rounded or sub-rounded.  See the daisy for scale!
 Some of the erratic stones and pebbles taken from the till face.  There is a considerable variety of stone types.

The till has a reddish colour, which means that it is composed mostly of ORS materials although it is on white or grey Carboniferous Limestone.  It has a sandy and gravelly matrix, and is packed with stones of all shapes and sizes.  The most frequent stone types are ORS sandstones -- some more red than others --and limestone, but there are also shales, conglomerates and (I am pretty sure) volcanic rocks which are blue and green in colour and which are strongly reminiscent of the rocks found on Skomer Island and in the Marloes Sands area.  Have they come from that far away, or are there other potential source areas closer at hand?  I didn't see any striations on the stones (the biggest are c 12" across) but I didn't have time to sample the till properly, or to clean and examine the contained stones as closely as I would have liked.  But I did notice that some of the stones are very rotten, and are coated with manganese.  One stone which I took from the face of the exposure was very heavy and shiny -- and reminiscent of the "ironstone" nodules found in the Coal Meaures and Millstone Grit in Pembrokeshire.

There is no doubt at all in my mind that this is till -- and that it is fresh.  In this location, ancient tills (maybe dating from the Anglian Glaciation) would certainly be solidly cemented or concreted by now as a result of highly-charged groundwater percolation over 450,000 years or so.  I know of one such till, not far away on the headland at Lydstep.  The till also reminds me of the till at West Dale Bay and the till found on the south coast of the St David's Peninsula, where the Irish Sea Glacier had picked up coarse-grained "land" sediments rather than sea-floor sediments composed of silts and clays with contained sea shells.  The till isn't cryoturbated or otherwise redistributed by slope processes or frost cracking -- it seems to me to be in situ and fresh.

As far as I could make out, the overall sequence of deposits here is as follows:

6.  Modern soil
5.  Sandy loess or brickearth (c 1m thick)
4.  Reddish till with ORS and other erratics (c 3m thick)
3.  Blocky broken limestone head (up to 2m thick)
2.  Concreted raised beach and included angular limestone fragments (up to 50 cms thick)
1.  eroded raised beach platform up to 10m above MSL.

There is nowhere where you can see this full sequence, but bits of it (including other patches of till) can also be seen in the cliffs to the south of the main exposure.

What was the Devensian Irish Sea Glacier doing here?   I had thought it skidded to a halt quite a few miles to the west, off the mouth of Milford Haven.  Time for another revision of my map.........  it is now incontrovertible that the ice flowed across Caldey Island.  This till could not have been deposited by a glacier coming in from the east.  But the ice might have flowed from the SW towards the NE or from W towards E.  There is ORS exposed in Sandtop Bay, on the west side of the island -- maybe that is where the ice came in from the west and where it picked up its load of ORS debris before carrying it across the island and dumping it in Bullums Bay.

By the way, Daylight Rock, where the bones of many Ice Age beasts were found, is on Small Ord Point, just a couple of hundred metres away.  Did those beasts roam this territory before, or after, the ice flowed across the island?  That's another nice problem that we need to sort out.............

Thursday, 25 August 2011

The Monsters from Daylight Rock (2)

 Possible post-glacial coastline positions around Caldey  -- as at the dates indicated.  Click to enlarge.

Given that Ice Age creatures were inhabiting caves on Caldey and on the adjacent hillsides near Tenby between 30,000 and 18,000 years ago, when sea level was far below its present position and when the whole district was dry land, we need to ask where the ice edge might have been at the time.  Cave lions, woolly rhinoceros, cave hyaena, cave bear, horse, woolly mammoth, hippopotamus, brown bear, giant Irish deer, bison, wolf and reindeer were all tundra animals, living in a landscape of cold-tolerant plants including dwarf birch and dwarf willow, grasses, mosses, bilberry etc -- very much like the tundra of today, as we see it in East Greenland and the Arctic fringes.  There might have been some areas of scrub birch and even conifers, but with the ground surface underlain by permafrost most of the countryside would have been open, with great expanses of bogs and standing water and with perennial snowpatches surviving throughout the year in shady locations.

As the Devensian cooling occurred snowpatches would have coalesced, lakes and open water areas would have contracted, and copses of trees might well have died as summers shortened and winters lengthened.  The animals might well have hung on, just as bears, wolves, reindeer and musk-oxen survive very close to the edges of active glaciers today.  Were they hunted by human inhabitants?  Very probably, since the dates of the Eel Point and Paviland human remains place them solidly into this period of Ice Age habitation.

So where was the ice edge at the time of the Devensian maximum?  Until yesterday, I thought it was located somewhere out to the west, near the mouth of Milford Haven.  Now I'm not so sure, following my discovery of Devensian till right at the eastern tip of Caldey Island........

The Monsters from Daylight Rock (1)

This is a photo of Daylight Rock, one of the key locations in unravelling the story of the Ice Age in Wales.  It's near the eastern end of Caldey Island, off the south Pembrokeshire coast.  This is where a piece of bone from a cave lion (pictured below) was found long ago, in a deep crevice (there is no actual cave here, unlike Nanna's Cave, Potter's Cave, Eel Point and Ogof-yr-Ychen).  There is quite an assemblage of bones from these Caldey sites:  woolly rhinoceros, cave hyaena, cave bear, horse, lion or tiger, mammoth, hippopotamus, brown bear, giant Irish deer, bison, wolf and reindeer.  This is a typical Ice Age cold-climate menagerie from the late Pleistocene -- and the radiocarbon dates confirm that some of the animals were alive on the island )which would not then have been an island) around 20,000 - 25,000 years ago.  One of the rhinoceros bones found by Brother James in Ogof-yr-Ychen was dated at 22,340 yrs BP, and other dates from South Wales caves seem to cluster around 22,000 - 28,000 yrs BP, with some outside this range also.  There is one older date for a reindeer bone which is around 30,000 yrs BP.

Then we have the human remains -- the bones of the "Red Lady of Paviland" now dated as c 26,000 yrs BP  and a human humerus from Eel Point dated to 24,470 yrs BP (Schulting et al, 2005).

While there has been much discussion of the accuracy of the radiocarbon dates for all these Ice Age samples, including errors introduced post-collection storage and down to groundwater and other types of ground contamination, and down to the nature of the diets consumed by the animals (including humans) that have been dated, not a lot of dabate has centred on the position of the Devensian ice front in the period we are talking about.  The peak of the Devensian glaciation (ie the time of most extensive ice) is assumed to have been somewhere around 23,000 - 20,000 BP in this area -- maybe the timing was different elsewhere.  That becomes very intriguing -- there is no obvious "break" in the sequence of C14 dates between 18,000 and 30,000 yrs BP.  Does that mean that the Devensian ice never reached Caldey Island, Paviland Cave or any of the other Pleistocene bone caves found in the Carboniferous limestone areas of South Wales?

To be continued.....

"................. the majority of the abundant bone from the (Eel Point) site was sold by the quarrymen for manure, or simply dumped into the sea (Smith, 1860). Species noted as present by Smith, Laws and W. Boyd Dawkins include woolly rhinoceros, cave hyaena, cave bear, horse, lion or tiger,
mammoth, hippopotamus, brown bear, giant Irish deer, bison, wolf and possibly reindeer. Investigations were renewed in 1950 by van Nedervelde and W. F. Grimes, but by then it seemed that quarrying had largely destroyed the cave (Lacaille and Grimes, 1955). However, persistent investiga-
tions by van Nedervelde led to the discovery in 1986 of sealed fissures containing more Pleistocene fauna (Davies, 1989; Davies and Smith, 1989).  Pleistocene faunal remains and stone tools are well known from most of the Caldey Island sites, although the association between the two is far
from certain. A deep, infilled swallet at Ogof-yr-Ychen was found to contain the bones of rhinoceros and hyaena together with many red deer antlers, as well as two struck adinole flakes and one chert flake (David, 1990: 22). Adinole is a raw material usually associated with Upper Palaeolithic tool assemblages in Wales. A radio-carbon determination of 22,340 G 620 BP (Birm- 340) was obtained on one of the rhinoceros bones (van Nedervelde et al., 1973)."

Source:  A Mid-Upper Palaeolithic human humerus from Eel Point, South Wales, UK
Rick J Schulting, Erik Trinkaus, Tom Higham, Robert Hedges, Michael Richards, Bernice Cardy
in Jnl of Human Evolution 2005

Sunday, 21 August 2011

On intellectual convolutions

While enjoying a brisk walk on Carningli this morning, I pondered a bit on the extraordinary intellectual convolutions that have gone on with respect to the HH Thomas theory of bluestone transport.  When you look at HHT's 1921 lecture and at his paper published in 1923, you see that all he actually did, scientifically, was to demonstrate that there was a very close match between certain rock-types in the Stonehenge bluestone assemblage and certain rocks sampled in the eastern part of the Preseli Hills in Pembrokeshire.  No more and no less.  He did not demonstrate, through a convincing scientific argument, that the stones were carried from Preseli to Stonehenge by human beings.  He SPECULATED about that, with extraordinary conviction, but that is a different matter....

HHT argued, very naively, that the Irish Sea Ice Sheet never crossed the Bristol Channel, although others had already shown, pretty convincingly, that it had.  He argued that the maximum ice front was located just to the south of Pembrokeshire.  He then concluded, with considerable panache, if not foolhardiness, that his argument "permanently disposes of the idea of glacial transport for the foreign stones of Stonehenge."  Permanently??!!  Strange that a professional geologist, whose core belief relates to impermanence, should have used such a word.

That was when the intellectual convolutions really got going.  First, there was the debate about the Altar Stone.  HHT thought it might have come from the Senni Beds of South Wales, and thought that that indicated a likely overland route for the transport of all of the stones from Preseli to Stonehenge.  Later on, Richard Atkinson (the real high priest of the human transport myth) and many others argued that the Altar Stone had come from the Cosheston Beds around Milford Haven -- and that is when the sea transport story kicked in and became a mainstream part of the myth.  In recent years, when Ixer and Bevins showed that the Altar Stone had not come from the Cosheston area, some archaeologists returned to the land transport idea, and it was even mooted that there was a separate transport route for the Altar Stone, carried (supposedly) by Neolithic tribesmen more or less along the route of the A40 road via Brecon and Abergavenny.  More and more convolutions, and less and less common sense.  Then Ixer and Bevins suggested that there were many bluestone sources around Carn Meini, and a long way from it -- and that some of the stones and fragments in the "debitage" had come from sources still unidentified. "Ah yes," said the archaeologists, "that shows that the builders of Stonehenge collected sacred stones from many different sacred sites -- not just one."  And in a final bit of mental gymnastics, when Ixer, Bevins and Pearce showed that some fragments had come from Pont Saeson and other localities north of the Preseli Hills, those same learned gentlemen said "Ah yes, that proves that they collected stones up from the north side of Preseli, and carried them via Newport Bay and round the Pembrokeshire coast before crossing the Bristol Channel."

All of these convolutions were performed by the archaeologists quite shamelessly, without a single scrap of evidence which might contradict the glacial transport theory.  Are they all stupid?  I sometimes think that they must be, since they seem to have no clue about how hypotheses are formulated and tested to destruction.  HHT's human transport hypothesis had been tested to destruction even before it was formulated, since what he did was to replace a hypothesis (the glacial transport theory) which had SOME evidence in support of it with another hypothesis (of human transport) which had NO evidence to support it.  It should have been rejected on that basis alone in 1923, but those who accepted it with acclamation must have been a pretty feeble bunch of wimps, since they seemed to have been quite incapable of subjecting it to intellectual scrutiny.

Since 1923, the glacial transport theory has seen evidence in support of it mounting up inexorably -- and yet the archaeologists still reject it on the very dodgy grounds cited by HHT when he gave his famous lecture.  And over that period of ninety years or so, not a single scrap of evidence has been adduced to show that a single bluestone (let alone 82 of them) was transported by Neolithic tribesmen from Preseli to Stonehenge, either by land or by sea.

Why is it that Darvill, Wainwright, Parker Pearson, and many others who should know better, continue with these increasingly convoluted thought processes, based upon an unsupportable hypothesis, and coming up with ever more wacky explanatory stories, when Geikie, Judd, Jehu and others gave them a perfectly adequate glacial theory more than a century ago?  Answers on a post-card please.....

On ice streams

This amazing image is doing the rounds just now -- it's on the BBC web site today.  It looks as if it is a product of one of the computer simulations of ice movement, but it isn't -- this one is based on real surface velocity data, assembled by many different satellites and processed into a complete "image" for Antarctica.

Note that over the thickest parts of the East Antarctic Ice Sheet there is very little ice movement -- ice velocities are generally less than 5m per year.  But out towards the coasts, the ice evacuation speeds are greater -- between 10m and 100m per year.  Then we come to the ice streams, which show up spectacularly as dark blue and purple, sometimes flowing at velocities far in excess of 1 km per year and carrying vast volumes of glacier ice towards the Ronne, Ross and Amery Ice Shelves.

If you had looked down on the Scandinavian / British Ice Sheet around 450,000 years ago, and done similar measurements, you would have seen something similar with respect to the Irish Sea Glacier.......

Friday, 19 August 2011

Stonehenge -- the empty quarter

 Above:  the recent EH plan of Stonehenge, showing the "empty quarter" (the south-west quadrant) very clearly.   Below:  the resistivity survey, showing relatively undisturbed ground in the "empty quarter" and suggesting that there are no buried stones or stumps there.

 About a year ago I made the post which is reproduced below:

"The resistivity survey image (from the chapter by David and Payne, Proc British Academy, 92, 73-113: Science and Stonehenge) shows a large number of "anomalies". The stones are shown in black. The white areas are mostly areas of disturbed ground coinciding with areas of past exploration and excavation. The dark grey areas may represent areas where there are high densities of intersecting pits or sockets, ie areas where stones have been moved about many times. The indistinct lighter grey mottled areas are difficult to interpret -- but the X and Y holes do show up as indistinct blobs. Note that they are not arranged on concentric circles, and that the spacing of these pits is imperfect and even erratic.  Apart from the white blobs marked A, B and C, there are no signs of "missing" stones buried in the turf in places where we might expect them, and in many places where we might expect sarsen and bluestone sockets there are not even dark grey shadows.  The conclusion from this work has to be that the 67 missing stones are not hiding anywhere on the site --- they are indeed missing -- and as I have already suggested, there is no reason to believe that they ever were put into the positions where the archaeologists would like them to have been.........   So there we are then. Gaps galore. Stonehenge never was finished."

Well, I was hoping that somebody would come up with some evidence to show that the Empty Quarter was indeed built on when the monument was being created, and that Anthony Johnson's "immaculate conception" as to what Stonehenge was like in its prime, has some foundation in fact.  Nothing has been brought to my attention, and I'm increasingly convinced that no stones were ever erected in this area.
The immaculate conception -- pure fantasy?

Bluestones summer lecture

My summer lecture last night went off very smoothly, with a good audience of around 60 people from far and wide.  I tried to summarise the recent published research in the fields of glaciology and geology and -- inevitably -- had a go at some of the archaeological fantasies which are currently in fashion.  I expected some growling and snarling from the archaeologists and "ancient wisdom" people who were in the audience, but in the event it was all very civilised, and nobody came up with anything remotely difficult to cope with.    Had a really good chat with many members of the audience afterwards, over a cup of coffee and a slice of cake.  Maybe the times -- and the Stonehenge story -- really are a'changin' .........


The more I think about it, the more appealing the idea of Ikea-henge becomes!  I really like the thought of these guys (giants maybe?  It doesn't matter) being told by some chieftain or other to get on and build this strange thing out in the middle of nowhere, on some windy hillside on Salisbury Plain. He gives them this flat-pack of bits and pieces, and lays out the instructions for them as carefully as he can.  Then he goes off to sort out some rebellious tribe somewhere, and leaves them to it.

The instructions are clear enough for even an idiot to understand,  so they check that the components are OK, and off they go, making sure that they don't do anything stupid, and consulting Merlin now and then for technical support.  At first all goes smoothly enough.  Many decades pass while the work is going on.  But then things start going wrong.  Some of the builders die off, and some of the others get fed up and wander off to other projects.  Those who are left get the bits and pieces mixed up, break one or two valuable stones that should have stayed intact, and find that a lot of bits are missing.  They get increasingly mystified and irritated, and at last decide that the whole thing was a lousy idea in the first place.  Anyway, nobody can actually remember what this bloody structure was for, and why somebody, long ago,  should have been so obsessed with it.  So at last they say "To hell with this for a lark!" and wander off, leaving behind them a partly completed shambles, with some of the bits in the right places, a lot of other bits just lying around, and many bits totally missing.  Sounds familiar?

Well, its nice that that version of events is pretty close to what I have been saying for years, and what the 2010 EH publication by Field and Pearson seems to be saying too.

And it's good to see some modern thinking coming into the frame, as far as EH is concerned.  Just right for the Age of Ikea.

By the way, click on the images to enlarge them -- and don't forget to read the small print.  Brilliant!  No idea who did this, but whoever it was, he (she?) had a great eye for detail......

Thursday, 18 August 2011

More digging in the hills

 A little bird tells me that Profs W and D have been in the Preseli Hills again this summer, presumably digging as part of their SPACES project.  One site being excavated was the old round barrow which had human remains in it, at Croesmihangel not far from Crymych.  Below is a write-up of a PLANED bus trip which paid a visit to the site.  I gather that Dyfed Archaeology was also involved.  By all accounts the two profs were also doing some work in their supposed Carn Meini "quarry" and that a TV crew was in attendance.  Just can't wait for the next blockbuster......

Croesmihangel Dig

With hopes of avoiding the forecasted rain, a bus full took to the Preselis to visit Professor Geoffrey Wainwright and his team who were digging at Croesmihangel, at the foot of Foel Drygarn.

The damaged round barrow on the eastern end of the Preseli ridge was excavated in the 1950s where they found cremated human bone so this dig is examining the remaining mound and ditch.  They had discovered a burial site when we arrived but had not opened it yet.  This dig is part of three being undertaken by the SPACES project.  They have also looked at a burial chamber at Carn Menyn and are now moving on to a pit circle at Withybush.

Friday, 12 August 2011

Those computer simulations

Blogger did something strange to those simulations mentioned by Henry.  I can't edit the comments, but I hope this will work in a mainstream post!  Click on the following and they should work!

The Preseli ice cap

Henry Patton's simulation of the Devensian ice cap near its maximum extent, c 23,700 years ago.  Note the little ice cap over Preseli, detached from the big ice cap which extended out from the upland ridge of Wales.  this little ice cap was far enough to the west never to have been incorporated into the Welsh ice cap -- at least, not in the Devensian glacial episode.

Henry has two Welsh Ice cap simulations on his web site, here:

They are highly recommended viewing!  One of them shows a single phase of ice expansion -- with a lot of oscillations and "flickering" -- and the other shows TWO phases of expansion, one around 30,000 years ago and the other 7,000 years later.  Which of those is likely to be the more accurate?  As we speak, Henry is probably trying to decide which of the two alternatives is most closely matched by the evidence on the ground -- and especially in the coastal sections of till and other sediments.

What interests me in particular about these simulations is the appearance, over and again, of small ice caps on all of the main mountains and hill masses of Wales, waxing and waning in response to climatic and other factors, and sometimes lasting for no more than a few dacades.

Note that these simulations only give us a part of the picture. There was another factor -- the Irish Sea Glacier which must have interacted in some way along the northern and western edges of the Welsh Ice cap.  They were probably out of phase, because small glaciers respond to climate change at a faster rate than big ones, but this aspect of the research makes it very interesting indeed. 

How long did the Preseli ice cap survive for?  A few centuries?  If we can find the answer to that, maybe we have a clue as to what might have happened during the earlier, bigger glaciations of the UK, just beyond the ice sheet edge.  Here we are talking about Dartmoor, Exmoor, the Mendips, the Cotswolds -- and maybe even Salisbury Plan and the chalk downs?

Thursday, 11 August 2011

On ice edges

Lines on a map.  One of the lines shows the position of the present coastline, which may or may not be correct, depending on the state of the tide.  The other line shows the suggested edge of the Devensian glaciers in the area -- but how sharp was the edge in reality?
Lines on a map, with added colour.  The Somerset Levels are in blue, and the other colours are used to indicate land surfaces between specific upper and lower altitudes.  Might it be that the line marking the transition from yellow to green also shows approximately the extent of the snow-covered or glaciated uplands at the time of the Greatest British Glaciation?

Why is it that we all love drawing lines on maps?  Well, maybe not ALL of us -- maybe geographers are worse than normal people in that they are taught to love lines on maps, and to fill up empty spaces by drawing lines on them.  Much more satisfactory than putting dots on maps -- and very often, if we have lots of dots, we seek some sort of order by joining up the dots with lines, or putting in lines which separate out those areas with high densities of dots from those with lesser densities.  It can all become very mathematical too -- if not absurd -- as I discussed in some posts about ancient wisdom and sacred geometry a few months back.

I got thinking about this the other day when I had the privilege of looking at one of Henry Patton's animated computer models for the growth and waning of the Devensian Welsh ice cap.  These models are very sophisticated as well as being exciting to look at -- if you are into this sort of thing.  I was reminded of some of the earlier computer simulations too, from Alun Hubbard and others, which I have discussed in earlier posts.

On this blog I have to admit to a preoccupation with lines -- many of my posts are concerned with the precise positioning of the Devensian ice limit, or the Anglian ice limit, or the extent of permafrost of various kinds.  Geologists and geomorphologists draw lines around areas of certain types of sediments or landform assemblages, although we all know that in reality what we have are fuzzy zones and transitions, with very few SHARP contacts between one zone and another.  OK -- we are just trying to simplify things and to make them more comprehensible to ourselves and to those who might read our learned papers. We are not being deceitful in doing this, and we are certainly not guilty of scientific fraud, but we are distorting reality.  In some of my field maps of glacial deposits, I know full well that in an undulating landscape of tills and fluvio-glacial deposits it is incredibly difficult to draw lines showing where the transitions are; I might draw a line HERE, and another field worker, in all honesty, might draw it THERE.  Sometimes these transitions are referred to as "feather edges" rather than lines -- and that's a nice descriptive term.

Back to ice edges.  What Henry's animations show, when several thousand years are compressed into a second, is that there is a sort of flickering at the ice edge as it advances and retreats across the landscape.  The ice cap edge, even on the models, is artificially sharp -- and in reality what we have, especially in a waxing ice situation as a glacial episode deepens and as the ice inexorably expands, is a snow-covered landscape with hill summits beyond the glacier edge supporting their own semi-permanent snowfields, ice fields or even small upland ice caps,  all waxing and waning seasonally and over decades and centuries.  The ice cap does not start in one place and gradually spread across the landscape, engulfing everything that lies in its path -- instead, it grows through the coalescence of many small local ice caps on all of the upland areas of Wales.  Gradually, these little ice caps join together until the main ice cap has a critical mass -- and then there is a major expansion involving a sort of positive feedback mechanism.

If you have the right plugins on your computer, and good broadband, have a look at the animations here:

So what are the implications of all this for the Anglian and Devensian glaciations?  Watch this space...

Wednesday, 10 August 2011

The Boles Barrow Bluestone

 The famous bluestone in the garden of Heytesbury House (thanks to Pete G)

On Edward Watson's Hobgoblin blog called "The Ancient Art of Enchanting the Landscape" there is a long analysis of the Boles Barrow bluestone enigma.  The post was put up on 13 April 2011.  It's fair to say that Edward and I do not see eye to eye on many issues relating to Stonehenge, but in the true spirit of enquiry I encourage you to read the article.  Of course, it would be very inconvenient to Edward if the Boles Barrow bluestone really was a piece of spotted dolerite collected from the vicinity and built into the barrow well before the builders of Stonehenge were alive, let alone dreaming up great plans for the iconic monument.  He argues that there were TWO stones (one little, one large) which have, over the years, become confused with one another, and of course he sides with Chris Green et al, who argue that the boulder in Salisbury Museum really came from Stonehenge -- and not from Boles Barrow -- via Cunnington's garden in Haytesbury.

This is a convoluted and interesting story, well told by Edward.  But I am not convinced by his arguments -- based upon the idea that the provenance of the stone is disputed.  He argues that the big bluestone found its way to Heytesbury after being robbed from Stonehenge, more than ten miles to the east;  but there is absolutely no evidence to support that contention, which was examined and rejected by Aubrey Burl.

I'm not sure that we ever will get to the bottom of this one.  It suits me to believe that the Salisbury Museum stone did come from Boles Barrow, and it suits Darvill, Wainwright and many others to think that it was stolen by Cunnington and his friends from Stonehenge.  All good fun.  See my post of 23 March 2011, which contains a detailed discussion with useful info from Tony.

Monday, 8 August 2011

A Mendip ice cap?

If you type in "ice caps" in the search box on this blog, you will find quite a few posts on the subject of small cold-based ice caps that may have existed in Southern England at various phases of the Pleistocene.  Exmoor, Dartmoor and Bodmin Moor might have carried the main ones, but what about Mendip?  On the above map, which I have used more than once, I suggest that there was a small ice cap over Mendip, quite possibly during the Devensian.  And if there was one then, there must also have been similar small ice caps on this upland area (small though it is) during earlier glacial phases.

I have been exchanging Email messages with various people interested in the Mendips, and am grateful to Alex and Henry for pushing this up to the front of the queue again.  In the recent posts about Cheddar Gorge and the other anomalous dry valleys (some of which are humped) we suddenly find an explanation that might make sense.  Shall we therefore speculate that the Mendip plateau supported a little ice cap on several occasions, with dimensions of approx 20 miles x 8 miles?  From above, it might have looked like the Barnes Ice Cap  on Baffin, or the Drangajokull in Iceland.  It was probably thin and cold-based, and pretty well incapable of glacial erosion.  If anything, it probably protected the landscape from periglacial processes for centuries or millennia.......

One interesting question is whether this little ice cap was "absorbed" into the Irish Sea Glacier when it pressed in from the Bristol Channel.  Another interesting question is the nature of the debris that might have rested on the ice cap bed -- might there have been enough movement for the creation of a crude till, or would there simply have been a slight reworking of frost-shattered debris or head?  In some polar limestone areas affected by ice, Prof Brian Bird says that the typical regolith is "rubble" -- mostly angular debris with some finer material, but no striated cobbles and erratics, and no clear structure.

It's when an ice cap like this melts that things become interesting.  Suddenly we have substantial volumes of ice (even in a thin cold-based ice cap) being converted into meltwater which flows radially and catastrophically away from the centre towards the ice cap edge.  The quantities of water involved are orders of magnitude greater than those coming from snowpatches (as in the conventional "periglacial snowmelt" model) -- and gorges and other deep and steep-sided valleys could well be the result.  Maybe these valleys were steepened and deepened a bit more, every time there was a glacial episode and every time the Mendip ice cap was regenerated.  Much of the meltwater flow might have been initially beneath the ice, close to the ice cap margin -- and then later in open channels.

What about the Rickford Valley and some of the other anomalous valleys that don't conform to the radial pattern?  Well, these could indicate that the Anglian ice of the Irish Sea Glacier did impinge upon, and maybe even overwhelm, the Mendip ice cap at the time of greatest ice extent (the GBG or "greatest British glaciation") -- leading to the creation of a small number of subglacial meltwater channels like those seen in North Pembrokeshire and elsewhere.

So there we are then,  that's a hypothesis -- with many questions attached to it.  Does this hypothesis adequately explain the associations of landforms -- and especially the cave sedimentary sequences -- that we know from Mendip?  Over to others for the answers......

Cheddar Gorge and Burrington Combe

Old air photos of Cheddar Gorge (above) and Burrington Combe (below).  A lot of trees have grown up since these were taken......

I have found another old article relevant to the matter of the Mendip valleys -- by Donovan 1969.  Below I have pasted his key comments:

Sunday, 7 August 2011

The strange valleys of the Mendips

 An old map showing some of the key features of the Mendips (Dr Allan Frey, Bristol Univ, 1965)Click to enlarge.  Both the deposits and the valleys can give us guidance on the Pleistocene history of the area.

Since my visit to the Cheddar Gorge the other day I have been doing some digging -- in an attempt to discover where the water might have come from, for the cutting of the gorge.  I've been looking at old papers by Brian Hawkins, David Gilbertson, Geoff Kellaway and many others, and have come to the conclusion that there is something strange going on at Cheddar and elsewhere.  I'm investigating these strange goings-on from a distance (I'm now back in Pembrokeshire) -- but I know that others who know the Mendips well are also investigating, and I look forward to seeing what discoveries they eventually announce.

I have also been looking at the GCR volume (big and expensive!) on the Quaternary of SW England, and have been rather disappointed to find that the geomorphologists who contributed to that learned tome are increasingly preoccupied with the micro-interpretation of sediment sequences, and hardly ever mention landscape and landform assemblages.  That, in my book, is what proper geomorphology is all about....... so who's doing it these days in the Mendip - Bristol - Bath area?   Somebody, I hope......

If we look at the map above, we can see that there are anomalous valleys all over the place -- some shown here as "Trias floored gaps" and others as gorges and deep combes.  Cheddar Gorge is the most spectacular, but there are steep-sided valleys on all flanks of the Mendips, some associated with thick alluvial fans of spreads of debris beyond the points where they leave the hills.  These valleys are not "normal" in the sense that they are either deeply incised, or difficult to explain by reference to very small catchments, or else simply cut across cols with no normal valley profile, or else take tortuous routes that seem to defy the normal logic of rivers!

Another of Allan Frey's maps, showing the route followed by students on a geography field trip in the 1960's.  But notice the gaps, gorges and cols all over the place -- the NW corner of Mendip is very interesting indeed....

On the above map Allan Frey has shown all the key valleys of the NW corner of Mendip, and the tight contour map shows the col breaches in particular, including those at Hele Coombe (south of Winscombe), Webbington (the gap followed by the motorway), and the complex valleys in the Shipham - Burrington - Rickford area.    The aerial photo below shows the western area, seen from above the Cheddar reservoir.

In a small and forgotten paper published in 1971, Hawkins and Kellaway reported on a Bristol /Bath field meeting in 1970 (PGA, 82 (2) pp 267-292) and devoted some space to the "anomalous dry valleys" here and to the north.  They mention Dolebury Channel, Burrington Combe, and the Rickford Valley as being possible glacial meltwater spillways, and they mention a number of other valleys further to the north and closer to the coast -- including the Providence Valley near Bristol, the Flax Bourton gap, and three channels across the western end of the Failand Ridge.  There are other gaps and gorges near Avonmouth.  Three other channels are also shown on the western end of Broadfield Down, and one between Broadfield Down and Dundry Hill.  We are talking therefore about more than 20 anomalous dry valleys within a fairly small geographical area.   See the 3 maps below:

Three maps from Hawkins and Kellaway -- above, map showing postulated ice directions of the Irish Sea Glacier which flowed in from the west; middle, Providence valley; below, Rickford valley.

Speaking for myself, I do not know the geomorphology of this area well enough to know what the meaning of all of these "anomalous valleys" may be -- but given the history of glaciation in the area, we should suppose them to be related to glacial meltwater on some way.  (That is a far more reasonable explanation that the one normally used in the literature -- namely that the valleys are "relics" of some ancient drainage pattern let down upon the present-day landscape.)  Hawkins and Kellaway do get themselves into a spot of lather by assuming that the channels are "spillways" or overflow channels related to the drainage of pro-glacial lakes during the deglaciation of the area.  But that means that they have some of the valleys flowing in a direction which must have been opposite to that of the ice surface gradient -- and that is not very likely at all.  Much more likely that the channels are subglacial channels, with orientations directed more or less by the direction of ice flow.  Some of the channels are indeed humped -- and that is a good diagnostic sign.

So there we are then -- a suggestion that there are many valleys in this area which are of subglacial origin; and another suggestion that the valleys themselves can be used as strong indicators of glacial action across quite a large swathe of country in the Bristol - Bath - Mendip - Somerset area.  Over to those who know the field details far better than I do.......

Saturday, 6 August 2011

Cheddar Gorge and Gordale Scar

Above: Gordale Scar, near Malham in Northern England

Cheddar Gorge and Gordale Scar are very similar limestone gorges, the former supposedly outside the limit of glaciation and the latter well within it.  Although the gorges share many physical similarities, there is abundant evidence of glaciation in the Malham area, and so it is widely accepted that the gorge was cut by torrents of glacial meltwater -- possibly within the Devensian glacial episode. 

On the other hand, Cheddar Gorge is more difficult to explain, especially if you are reluctant to admit to the presence of glacier ice anywhere in the vicinity, at any stage of the Pleistocene.  There is a tendency to fit the gorge into the non-glacial working hypothesis, which means that the gorge has to be explained by reference to permafrost and snow-melt.  But what if (as Megan Siever argues) the gorge is itself evidence of glacial meltwater on a substantial scale flowing down from the Mendip heights, across the southern escarpment and into the lowlands of the Axe valley?  The great quantity of "taele gravels" around the Cheddar Reservoir appears to have come from the gorge -- these gravels are mostly made up of Carboniferous Limestone fragments, and must be the products of mechanical erosion in the gorge.  And since glacial deposits are known from Greylake (not far from Glastonbury) and from the Bath area, it is entirely reasonable to argue that the Mendips were, at one stage (or maybe more!) affected by glacier ice.  Was the ice flowing across the hills, or was it flowing round them, from the west and towards the east, as argued by Hawkins, Kellaway and Gilbertson many years ago?  The answer to this question could be very important, since we need to ask this question:  "Could Cheddar Gorge have been cut by subglacial meltwater?"  My own suspicion is that it was not, but that it was cut by meltwater flowing marginally or from a glacier front positioned over the northern part of the Mendips.

Friday, 5 August 2011

More on Cheddar Gorge

 Above:  Cheddar Gorge from near the viewing platform.  Below: Stalagmites, stalactites and other flow features from Cox's Cavern

On the way home from Sweden, we called in on the family in Somerset and had a day out as tourists in Cheddar Gorge.  It's very many years since I was there last -- and the tourist industry has taken over......

But the visit got me thinking about the gorge and its relationship to the events of the Ice Age.  It is indubitably a very spectacular feature, but everything is relative, and it's really quite small fry when compared with the other great gorges of the world.  So how was it formed?  As indicated in earlier posts, there are essentially 3 theories:

1.  The gorge was formed by the collapse of a series of linked limestone caverns.  This theory is now largely discredited...

2.  The gorge was formed by glacial meltwater, as suggested or implied by Geoffrey Kellaway and as argued by Megan Siever and others over the years.  The establishment doesn't like this idea, because it doesn't fit with where the ice edge is supposed to have been either during the Anglian or Devensian glaciations.

3.  The gorge was formed by snowmelt torrents during periods of periglacial / permafrost conditions.  This is the standard spiel given by the Cheddar tour guides and indeed by most geomorphologists.  Here is the BGS summary:


The gorge

Contrary to popular belief, Cheddar Gorge is not a collapsed cavern, but a fine example of a gorge cut by a surface river, and since left high and dry as drainage went underground. The gorge was formed by meltwater floods during the many cold periglacial periods over the last 1.2 million years. During these arctic episodes, the development of permafrost blocked the caves with ice and frozen mud. Snowmelt floods during the brief summers were forced to flow on the surface,  carving out the gorge in the process.   Each successive periglacial episode caused further erosion. During the warmer interglacial periods, drainage was underground, creating the caves and leaving the gorge dry.


I have to say that I incline towards the glacial meltwater view, and I have argued this elsewhere on this blog.  I think that Anglian ice DID reach the Mendips, even though it may not have flowed right over the hills; and I think that the thick gravels (taele gravels) at the mouth of the gorge and elsewhere around the flanks of the hills may be the "torrent debris" associated with large volumes of glacial meltwater.

I also have problems with the precise mechanisms supposed (by most field workers) to have operated in times of permafrost conditions.  How exactly are the snowmelt floods supposed to have originated?  The assumption is that if the Cheddar limestone mass was frozen solid, then these floods would have flowed at the surface, cutting the Cheddar Gorge by mechanical rather than chemical or solutional processes.  But hang on -- the limestone mass could only be frozen solid if the water table was right up at the surface -- and that is very unlikely indeed, given the fact that the gorge is close to the southern edge of the Mendips.  Wet rock can freeze -- dry rock cannot.  It is much more likely that even in a periglacial (permafrost) environment the water table was well down beneath the surface, and that there would still have been downward percolation of snow meltwater into the chambers and caverns that we see something of in Gough's Cave and Cox's Cave.  I need some caving insights into this problem  -- is there any evidence that the whole cave and tunnel system was actually blocked with ice and debris during the big glacial phases?  Or was water always flowing somewhere, at some level within the limestone mass?

The other problem I have is that the drainage catchment feeding into Cheddar Gorge is not that large.  I do have problems envisaging enough snow, and enough snowmelt, to account for the cutting of a very substantial gorge.......

We actually have very few analogies to work from -- but there are some, from Arctic Canada, where limestone plateaux and permafrost coincide.  Watch this space.