Content uploaded by Brian John
Author content
All content in this area was uploaded by Brian John on Nov 10, 2015
Content may be subject to copyright.
OCTOBER 2015 No. 137
QUATERNARY EVENTS AT CRAIG RHOSYFELIN,
PEMBROKESHIRE
Brian John, Dyfed Elis-Gruffydd and John Downes
Note: Quaternary Newsletter is a peer-reviewed publication of the Quaternary Research
Association. It is issued three times a year, and goes out to more than 1,000 members who have a
professional interest in glacial geomorphology and other Quaternary disciplines.
Note: This short note is intended for a readership of Quaternary specialists. It will be followed by a
more comprehensive paper relating to the site stratigraphy and archaeological issues, to be published
in the 2015 issue of the journal “Archaeology in Wales”.
Citation: Brian John, Dyfed Elis-Gruffydd and John Downes (2015).! "Quaternary Events at Craig
Rhosyfelin, Pembrokeshire."! Quaternary Newsletter, October 2015 (No 137), pp 16-32.
Page 1
QUATERNARY EVENTS AT CRAIG RHOSYFELIN,
PEMBROKESHIRE
Brian John, Dyfed Elis-Gruffydd and John Downes
Abstract
The Afon Brynberian valley is claimed to contain Britain’s most important Neolithic quarry, used
for the extraction of bluestone orthostats destined for Stonehenge. Archaeologists argue that an
exposed rock face within a meltwater channel at Craig Rhosyfelin is a quarried surface, and that an
eight-tonne block found five metres away was prepared for transport but then abandoned. Site
investigations have revealed scoured surfaces, faceted and abraded erratic boulders, glacial till,
fluvioglacial sands and gravels, and widespread rockfall and solifluction deposits. All the features
associated with the “proto-orthostat” are considered to be natural.! There are currently no visible
prehistoric landforms or sediments that are demonstrably anthropogenic in origin.
Introduction
Craig Rhosyfelin is a rocky spur in the valley of the Afon Brynberian (SN117362) on the
northern flank of Mynydd Preseli in Pembrokeshire, UK (Fig. 1). ! The spur is aligned
approximately NE-SW. It is about 80 m long and 15 m wide, tapering to a narrow tip at its
downslope end. A central ridge of fractured rhyolite bedrock is surmounted by fragile
and precarious outcrops from which many large slabs and sub-angular blocks have fallen. !
Summit slabs are currently being broken up by biological and other processes. Rock
exposures are visible on both flanks, but the SE flank is heavily vegetated.
The site has gained recent notoriety in archaeological circles because an unusual foliated
rhyolite, exposed on a series of sub-planar fracture surfaces and on detached blocks, has
been matched to some “bluestone” fragments in the debitage at Stonehenge (Ixer and
Bevins, 2011, 2013, 2014). !The NW flank of the spur has been postulated as a "Neolithic
bluestone quarry" from which orthostats were carried to Stonehenge around 5,000 years
ago (Parker Pearson, 2012; Ixer, 2012). We have examined the site and the surrounding
landscape and have assessed the evidence for human quarrying activity, and this paper
reports our observations.
One major benefit from a five-year archaeological dig is that it has provided a unique
opportunity both to study an unusual volcanic outcrop and to consider some of the
processes, landforms and sediments typical of the north Pembrokeshire Quaternary. On
the cleared and excavated rock face of the Rhosyfelin crag there is ample evidence that
many different processes have contributed to its current appearance, to the widening of
Page 2
joints and to the accumulation of fallen rock debris. Within the confines of the
archaeological dig (Fig. 2) there are also layered and intercalated sediments up to 3 m
thick, suggestive of a complex history over a great period of time. On the basis of its
geomorphological interest, a notification has been submitted for RIGS designation.
Figure 1. Satellite image of the Afon Brynberian channel, the Craig Rhosyfelin rhyolite spur and
the subsidiary meltwater intakes near the bend in the road. The sampling points used by geologists
Rob Ixer and Richard Bevins in 2010, prior to the archaeology dig, are also shown.
(Acknowledgement: Richard Bevins)
Page 3
Background
During the Quaternary north Pembrokeshire has been affected by both Welsh and Irish
Sea Glacier ice. Mynydd Preseli may at some stages have supported a small cold-based ice
cap (Patton et al., 2013), similar to that proposed for Dartmoor by Evans et al. (2012). It
may also have been affected intermittently by ice flowing W or SW from the Welsh Ice Cap
(Hubbard et al., 2009). !However, most of the glacial field evidence (relating to striae,
erratic transport and sedimentology) in West Wales relates to an ice stream flowing
onshore from the N and NW. !It is widely accepted that the Irish Sea Glacier, Britain’s
largest ice stream, flowed across the county and was thick enough to submerge even the
Figure 2. The upper part of the dig site in 2012. The rock face on the flank of the spur is just off
the left edge of the photo. The litter of broken blocks was completely buried beneath rockfall and
slope deposits prior to the commencement of the dig in 2011. The large elongated block in the centre
of the photo is the “proto-orthostat” which has attracted much attention. In the foreground the
upper surface of the till layer is exposed.
highest parts of the landscape during the Anglian Glaciation (Campbell and Bowen, 1989;
Walker and McCarroll, 2001). !During the LGM, Irish Sea Glacier ice was less extensive,
but it affected much of the northern coastal strip of Pembrokeshire and inundated the
landscape around Rhosyfelin (Hambrey et al., 2001; McCarroll, 2001; Etienne et al., 2005).
Page 4
For most of the Quaternary, glacier ice was not present in Pembrokeshire, and there were
frequent climatic oscillations involving interglacial or interstadial conditions and long
episodes during which periglacial conditions dominated. !The Quaternary stratigraphy of
Pembrokeshire, as revealed in BGS mapping, is characterised by periglacial slope
accumulations close to the surface with many expanses of glacial, fluvioglacial and
glaciolacustrine deposits preserved in favourable locations (BGS, 2010). !Coastal sediment
sequences tell a consistent story, at least to the north of Milford Haven (John, 1970a).
However, it is rare to find inland exposures that can be tied in with confidence. Rhosyfelin
provides such an opportunity.
The Rhosyfelin spur
The crags visible along the valley of the Afon Brynberian near Rhosyfelin are erosional
remnants of Fishguard Volcanic Group outcrops (Thorpe et al., 1991; BGS, 2010; Downes,
2011). The spur at Craig Rhosyfelin is the most substantial of these features, with rock
exposures on both flanks. The rhyolitic lavas are splintery blue rocks which weather to a
light grey colour. They are of Mid Ordovician (Llanvirn) age. Some substantial quartz
veins are also visible. !Here the lavas dip steeply north-westwards at about 75º, and
examinations of the exposures reveal deep almost vertical intersecting fractures with
numerous horizontal cross fractures which give rise to sharp-edged clasts ranging in size
from flakes to blocks and slabs more than 3 m long. !The largest visible block, measuring
3.8 m x 1.3 m x 0.6 m and weighing about 8 tonnes, is located about 5 m from the rock face,
where prior to excavation it was embedded in slope deposits. It has a flat upper face, and
is best described as an elongated slab. It has been referred to by archaeologists as a "proto-
orthostat" (Fig. 2). Abundant rock debris, accumulated at the foot of the steep rock face,
has been uncovered during a recent archaeological dig. !The focus of associated geological
research (Ixer and Bevins, 2014) has been an unusual flinty foliated rhyolite with a planar
or lensoidal "Jovian" fabric. The rock face at Rhosyfelin is referred to as a "foliation plane"
but it is actually made up of facets of several distinct fracture planes, some of which
project <150 cm beyond others. !The impressive rock face is the feature that has led the
archaeologists to assume, after clearing away trees, shrubs and many tonnes of debris, that
it is a worked quarry. However, it can be argued that the rock face as it presently appears
is simply an “archaeological artifice” unlike anything that might have existed in the past.
Ixer and Bevins (2014) claim that they have fixed the provenance of certain rhyolite
fragments at Stonehenge to within a few metres close to the tip of the Rhosyfelin spur. On
that basis Parker Pearson (2012) and Ixer (2012) claim to know where at least one
Stonehenge bluestone has come from.
Meltwater channels
Craig Rhosyfelin is situated close to a bend in the Brynberian river gorge. !The main valley
runs N towards a confluence with the Afon Nyfer, which has had a complex history of
meltwater erosion and glacial diversions (Gregory and Bowen, 1966). !The Brynberian
Page 5
gorge currently contains a “misfit” stream and it is suggested that it has been cut by glacial
meltwater through several phases. As seen in Figure 1, there is a small subsidiary dry
channel with two intakes on a col to the W and N of the crag. These features are
reminiscent of many humped and arcuate channels within the Gwaun-Jordanston
meltwater channel system, suggestive of subglacial meltwater flow under great
hydrostatic pressure within subglacial conduits (John, 1970a). Within the channels, as on
many of the tors in the area, there are signs of ice smoothing and block removal; a litter of
rhyolite blocks (some in excess of 1 m long) can be seen in the sediments on the valley
floor. Streamlined and moulded forms are common, especially near the tip of the spur and
on the col (Fig. 3). There are also traces of crescentic gouges on some of the solid rock
outcrops on channel walls and on loosened blocks which are incorporated into sediments.
Close to the enigmatic 8-tonne elongated slab the archaeologists have exposed sections of
what appear to be the channel’s rock floor, partly masked with compressed broken rock
debris up to 30 cm thick. This is the lowest visible layer (numbered 1 in Fig. 4) in the
sediment sequence.
Figure 3. Moulded rock surface in the col between the main meltwater channel and the subsidiary
channel - at the SW extremity of the rhyolite ridge.
Page 6
Glacial sediments
During the 2013 and 2014 archaeological excavations, an extensive layer of coarse-grained
diamicton was exposed (numbered 2a in Fig. 4). It is similar in texture to the “local till”
found on the S side of the St David’s Peninsula (John, 1970b; Campbell and Bowen, 1989).
It appears to rest partly on the channel’s bedrock floor, and partly on rockfall material
Figure 4. Sketch columns (not to scale) of the sediment sequence at Rhosyfelin, at three sites c. 6 m
from the rock face. Both vertical and horizontal relationships are generalised. Locality A is close to
the “proto-orthostat”. Locality B is some 10 m downslope. Locality C is on the valley floor, near the
tip of the spur. Rockfall horizons are discontinuous, suggestive of intermittent rockfall events.
Columns A and C are c. 3 m deep, and column B is c. 3 m deep. Column C is in reality c. 1 m
lower than shown. The Fe-stained “surface” has a continuous slight downslope gradient.
beneath the rock face. Sandy and gravelly materials predominate in the matrix, but in
some locations towards the upper end of the dig site there are higher proportions of silt
and clay. There are incorporated clasts of many shapes, sizes and lithologies (Fig. 5).
Many of the stones are faceted, and some are grooved and striated. They are also “fresh”
in appearance, unlike the clasts in the “rotten” till and fluvioglacial gravels at Llangolman,
Page 7
Figure 5. Devensian till exposed on the floor of the 2014 archaeological dig at Rhosyfelin. At this
point it has a gravelly matrix and contains clasts of many different lithologies including dolerite
boulders.
on the south side of Mynydd Preseli. Most of the local rhyolite stones and boulders are
bluish in colour, and they are angular or sub-angular in shape indicating that they have
not travelled far. !Some quartz pebbles are rough and angular, having been derived from
immediately adjacent bedrock outcrops, possibly as rockfall following extensive frost
weathering. There are also abundant dolerite and volcanic ash boulders and pebbles --
some of them well-rounded. The majority of erratics have come from the Fishguard
Volcanic Series, from outcrops to the N and W. !On the basis of the foregoing, the deposit
at the base of the crag is interpreted as a local till. Where its surface is exposed there is
heavy staining by iron and manganese oxides, a crumbly structure and traces of a crust or
pan. The foxy red colour of the visible exposures is reminiscent of that seen on the surface
of the Irish Sea till at Abermawr, where oxidation has occurred beneath a cap of permeable
fluvioglacial sands and gravels and is consistent with exposure to postdepositional
weathering observed elsewhere in sediments formed during the Last Glaciation in Britain
(Eyles & Sladen, 1981)).
Page 8
Fluvioglacial sediments
Near the end of the spur the frequency of rounded cobbles and boulders increases, and the
local till grades laterally into a poorly-sorted meltwater deposit of gravels with
incorporated erratics which include tuffs, agglomerates, gabbros and sandstones (2b in
Fig. 4). !The matrix is variable. In places it incorporates much silt and clay, and elsewhere
it comprises sands and gravels. !There are no exposures showing that it is underlain by till.
Detached rock slabs in contact with this deposit are heavily abraded, and it is suggested
that at least some of this abrasion is attributable to high-velocity and turbulent meltwater
flow. These gravels are stratigraphically comparable to the fluvioglacial materials
described by Bowen (1982) and Hambrey et al. (2001) in the Cardigan - Monington area,
but here they are much coarser, including some boulders over 1 m in diameter. It is likely
that they were laid down very close to a wasting ice edge, with true glaciogenic and
fluvioglacial deposits laid down contemporaneously. The archaeologists working on this
site have referred to rounded "hammer stones" which in their view have been used in the
quarrying process. However, those we have examined are typical of the rounded and sub-
rounded clasts commonly found in glacial and fluvioglacial sediments. A search has not
revealed a single “hammer stone” with the percussion marks that might be expected on a
well-used Neolithic stone-working tool. Parker Pearson (in a 2014 public lecture) has also
referred to the discovery of a river bank and semi-circular “revetment” made of large
boulders in the fluvioglacial materials near the spur tip, linked by a routeway to the
supposed quarrying area. He has suggested that this is where quarried bluestone
orthostats were loaded onto rafts or sledges. However, these speculations are not
supported by any stratigraphic differences between the supposed dry land and riverine
environments. Recent excavations suggest the presence of a clay-rich horizon beneath
some of the fluvioglacial material, but exposures near the water table are poor and are
subject to flooding. !
Rockfall debris
The most noticeable sedimentary features at this site are the scree and accumulated
rockfall debris banked up against the rock face (Fig. 2) (numbered 3 in Fig. 4). Some
blocks are fractured and sharp-edged, and others are smoothed and rounded as a result of
glacial and fluvioglacial erosion. Close to the rock face, there is no readily-observed
junction or “surface” between the till and the rockfall debris, suggesting that on parts of
this site at least, the two deposits are contemporaneous. In places it is possible to observe
the fractures that have occurred on grounded slabs and blocks as a result of percussive
impacts associated with later heavy rockfalls. Many blocks reveal faces that are
particularly heavily weathered and pitted; prior to dislodgment these faces were clearly
those most exposed to the elements. There is a matrix of rubbly debris with organic
inclusions and sharp-edged small rock fragments which are inferred to have moved
downslope and accumulated during a long period of slope evolution punctuated by
rockfalls. The finer sediments have been preferentially removed by the archaeologists in
2011-2014, leaving many of the larger blocks behind (Fig. 2). We are not aware that these
Page 9
sediments have been logged and analysed before being removed to a large adjacent spoil
tip, but the positions of the majority of large clasts have been accurately surveyed and
recorded by the digging team (Parker Pearson, 2012). The 8-tonne “proto-orthostat”
referred to above is in contact with the underlying till at its downslope end, and at its
upslope end the till appears to be about 30 cm beneath its lower surface. It is in a fragile
state: one fracture runs across its upper face, and there are various other fractures which
can be traced on its sides and ends. If any attempt should be made to move it, it would
probably disintegrate. However, it has been left by the archaeologists “perched” on a base
of underlying blocks following the removal of finer sediments and till. This assemblage of
rocks of all shapes and sizes, some of which are broken by percussive impacts, is referred
to by the archaeologists as a set of pillars, pivots, props and “railway lines” put into
position by Neolithic quarrymen in order to ease the movement of the “proto-orthostat”
downslope. However, there is nothing “artificial” about either the position of the large
block or the stones beneath it, and since all were embedded within accumulated slope
deposits and till, they all appear simply to be components of gradually-accumulated
layers of glacial and post-glacial material by natural processes. Near the downslope end
of the “proto-orthostat” there is a broken and elongated block of rhyolite with an upper
surface displaying features described by Parker Pearson as “striations” caused by large
blacks being dragged over it. On careful examination these “striations” are seen to be
outcropping foliations like those seen on the edges of countless other blocks.
Solifluction sediments and colluvium
Interbedded with the rockfall debris and overlying the till and fluvioglacial deposits there
is a layer of slope deposits up to 2.5 m thick (numbered 5 in Fig. 4). The material exposed
adjacent to the large slab is reminiscent of the “upper head” or stony solifluction layer
found above Devensian glacial and fluvioglacial deposits in West Wales coastal sections
(John, 1973). Exposures cut in the sediments in 2012 showed that more than 5 m from the
rock face there is a clear contact between the till and the overlying pseudo-stratified slope
deposits. Within the latter there are at least five different but discontinuous layers, with a
c. 10 cm sandy/silty layer at the base. Above that, there are some layers of fine-grained
sediments and others made up predominantly of elongated stones and flakes less than 10
cm in length. In addition to an abundance of sharp-edged local rhyolite fragments there
are some large slabs and boulders and also, in the lower layers, rounded and sub-rounded
erratics derived from upslope glacial materials. There are many signs of root penetration
through this sequence, and the fine-grained layers contain many streaks of peaty organic
debris. These materials have moved downslope predominantly from the NW, W and SW.
There do not appear to be any ice wedge casts which might suggest the presence of
permafrost at the time of accumulation or at a later date, but in lower horizons there are
some signs of bedding disturbances possibly attributable to frost-heave processes.
Further downslope, where the surface gradient decreases, the stones in layer 5 become less
abundant, and the sediment is predominantly made up of colluvial gravels, sands, silts
and clays. Resting directly on the fluvioglacial deposits on the edge of the Afon
Page 10
Brynberian floodplain, there are at least three bands which are difficult to correlate
precisely with the layers at the upslope end of the dig site. At the base (close to the valley
floor water table) there is a blue-grey layer at least 30 cm thick. !It is clay-rich but
incorporates bands of gravels which appear in places to have been deformed either by
loading or frost-heave processes. Above that is a layer up to 80 cm thick made up of
sands, silts and clays but with some gravel and stone inclusions. It is buff-coloured, and
there is no sharp junction between this and the underlying grey-blue sediment; texturally
they appear to be related.
Figure 6. The sediment sequence in the lower part of the dig site. This shows fluvioglacial gravels
and contained clasts at the base, !overlain by fine-grained sediments which are gleyed and oxidised,
and then by dark-coloured colluvium and soil. Occasional angular clasts of local rhyolite occur
throughout. Trowel for scale.
In this deposit there are occasional fragments of charcoal, suggestive of either natural/
accidental burning of woodland or scrub in the vicinity, or else human occupation.
Passing upwards in this layer iron-staining becomes more and more prominent, until the
Page 11
sediment has a distinct foxy-red colour similar to that on the till surface elsewhere on the
site. It is noteworthy that the iron-enriched band transgresses the junctions between
stratigraphic layers, suggesting that it is a pedogenic feature related more to water table
oscillations than to age (Iron "pans" are common in podzol soil horizons across north
Pembrokeshire). !Finally there is a grey-brown surface layer made of accumulated fine-
grained slope materials, passing upwards into modern soil. !There is a high content of
organic matter, and more fragments of charcoal. !This layer is up to 80 cm thick on the
lower part of the site (Fig 6).
Discussion
Craig Rhosyfelin is related to many other tors made of Fishguard Volcanic Series rocks.
The greatest concentration is to be seen on the upland ridge of Mynydd Preseli at locations
including Carn Meini, Carn Breseb, Carn Alw, Carn Goedog and Cerrig Marchogion. On
the lower northern slopes of Mynydd Preseli there are other tors at Carnedd Meibion
Owen, in Coed Tycanol and at Felin y Gigfran. !Some of these incorporate ice-moulded
surfaces and perched erratic blocks, and some are so heavily denuded that they have
almost disappeared (McCarroll, 2001). !It is likely that Craig Rhosyfelin was once higher
and more prominent than it is today: its upper surface is broken and irregular, with many
loose blocks close to collapse and many others already incorporated into the rockfall
debris on its flanks. Close to its tip there is much joint-widening and evidence of block
detachment and collapse guided by fracture planes, strongly suggesting that it has also
been reduced in length. There are no slickensides on the densely-jointed exposed rock
faces, and this suggests that there has been little or no fault movement. It is therefore
possible that the extreme fracturing is simply a result of ancient cooling and contraction.
However, in the light of the scale of destruction of other Pembrokeshire tors, it is also
possible that the fragility of Craig Rhosyfelin is partly a Quaternary legacy. !Joint creation
may have followed compression beneath thick and mobile glacier ice, during unloading
and pressure release. This process is thought to have led to slope collapse and landsliding
within the South Wales Coalfield (Woodland and Evans, 1964) and in West Wales at
localities including Traeth Cell-Howell and Druidston (John, 2008). !Glacial entrainment
processes and large-scale block removal from "vulnerable" crags such as this are thought to
have contributed to the long-term modification of the local landscape (John and Jackson,
2009), and it is probable that the most dramatic changes occurred during the Anglian
glacial episode during and after a deep inundation beneath ice which was thick enough to
maintain forward momentum at least as far as the coastlands of Somerset (Kellaway, 1971;
Campbell and Bowen, 1989; Thorpe et al., 1991; Hubbard et al., 2009; Gibbard and Clark,
2011) . !
On the matter of bluestone provenancing, and the claimed matching of rhyolite fragments
at Stonehenge to a bedrock exposure within a “few square metres” near the tip of the spur,
a degree of scepticism is in order. While not denying the care and the professional skill of
geologists Richard Bevins and Rob Ixer, it appears to us that their sampling point density
is inadequate for the degree of precision claimed; and a close reading of their published
Page 12
data suggests that some of the Stonehenge foliated rhyolite fragments might have come
from other outcrops in the Pont Saeson area. Furthermore, they do not appear to have
considered the possibility that the fragments at Stonehenge have come from parts of the
crag which have been entirely removed by glacial erosion.
It is suggested that the Afon Brynberian channel is very ancient, having been cut during
the glacial phase responsible for the Gwaun-Jordanston subglacial meltwater channel
system which has its main intake some 6 km to the west (John, 1970a). !Again, the Anglian
glacial episode is the most likely candidate, since some channel mouths at the coast
contain sediments that pre-date Late Devensian tills and meltwater deposits, and since the
scale and the complexity of the humped channels are difficult to explain by reference to
marginal and sub-marginal meltwater flow close to the Late Devensian ice limit. !The main
channel at Rhosyfelin and the subsidiary channel on the NW flank of the rhyolite ridge
may have been modified during subsequent glacial episodes, but currently there is no firm
evidence for dating any of the phases, apart from clear signs that the main channel has
been deepened by at least 10 m below the intake point of the smaller channel. The rock
face exposed during the archaeological dig is essentially a channel wall which carries
evidence of high-velocity turbulent meltwater flow (Fig. 3). !It may be a relict Anglian
feature, but it is likely that it was freshened during the Late Devensian ice wastage phase
and that it is closely related to the bulk of the sediments at the site. This hypothesis is
reinforced by the presence of many large sub-rounded and sub-angular locally-derived
rhyolite blocks in the sediments.!
The Rhosyfelin till, for reasons enunciated above, is interpreted as a product of a complex
ice wastage environment. It is more likely to be a melt-out till than a lodgement till; but
we have not been able to examine its internal structure or to ascertain whether it is in situ.
In view of its position beneath a layer of slope deposits capped by modern soil, it is
assumed to be the equivalent of the Late Devensian Irish Sea till exposed on the North
Pembrokeshire coast (John, 1970a; John and Elis-Gruffydd, 1970). However, that till is
composed partly of recycled sea floor sediments, with a large clay component (Rijsdijk and
McCarroll, 2001). !The inland till generally contains less clay in its matrix. As indicated
above, abundant studies suggest that the Irish Sea Glacier flowed across the site of Craig
Rhosyfelin and terminated close to the crest of the Preseli upland ridge (Campbell and
Bowen, 1989; BGS, 2010). !
Rockfall debris is incorporated into the till and also overlies it, replicating the situation in
abundant Pembrokeshire coastal exposures (John, 1973). At some of those exposures, ice
wedge casts and frost-heave features indicate that permafrost conditions persisted for at
least part of the interval between the LGM and the Younger Dryas. Slope deposits from
this time contain angular clasts and blocks dislodged by freeze-thaw processes on exposed
cliffs and other rock surfaces. It is therefore suggested that at least some of the stratified
slope deposits exposed beneath modern soil at Rhosyfelin have accumulated under
periglacial conditions following the disappearance of glacier ice and at a time when
vegetation was sparse. !Further, there may well have been accelerated rockfalls here
during the Younger Dryas (John, 1970a; Campbell and Bowen, 1989), at which time the
Page 13
climate was severe enough for glacier growth in the uplands and for pingos and other
periglacial features to have been created in West Wales (Walker et al., 2001). !It is likely that
rockfalls occurred intermittently during the Palaeolithic, Mesolithic, Neolithic and later
periods. Because the 8-tonne “proto-orthostat” !was not deeply buried, there is a
possibility that it was emplaced within the past few thousand years. Nowadays biological
processes associated with root expansion and plant debris accumulation in fracture cracks
are responsible for triggering ongoing rockfalls. The crag is still being reduced in size.
The materials which overlie both the glacial and fluvioglacial materials, and which pass
laterally into the rockfall accumulations, are not easy to interpret, but they appear to
represent maybe 18,000 years of late-glacial and post-glacial climatic oscillations. !More
work is needed on these deposits, including the radiocarbon dating of scattered charcoal
fragments and perhaps luminescence dating of the colluvium. It is suggested that some of
the sediments at the base of the sequence might have accumulated in a temporary slack-
water or lagoonal situation following ice retreat. !There are no thick laminated silts and
clays such as might be expected from a long-lasting pro-glacial lake (Hambrey et al., 2001;
Etienne et al., 2006); !but further exposures and systematic sampling at Rhosyfelin and on
Brynberian Moor will assist in the interpretive process. Recent studies by one of us (BSJ)
on Brynberian Moor have revealed the presence of thin slack-water sediments in some
stream cuttings, but elsewhere bedrock exposures are not capped by laminated silts and
clays, and lake deposits do not overlie or underlie till deposits where they are exposed in
stream cuttings.
Regarding the colour variations in the finer-grained sediments, it is suggested that these
are the result of gleying processes in a wet environment (cf. soil series Hafren, 0654a). !The
staining might be the result of a long period of surface weathering following ice retreat;
but as indicated above, it is more likely to have been associated with pedogenic processes
operating beneath an accumulating later layer of solifluction debris and colluvium.
It is evident from the slight terracing on the floor of the Afon Brynberian valley, and from
the coarse gravels with erratics seen in the river banks, that there has been much Holocene
reworking of older sediments as the river has changed its position on the flood plain.
Human Activity Traces?
As indicated above, there is a conviction among archaeologists involved in the recent dig
that this is a Neolithic quarry site (Parker Pearson, 2012, 2013; Ixer, 2012), and that the
rockfall debris has accumulated on a "quarry floor" as a result of human intervention.
There is nothing that might qualify as a quarry floor or working surface, and there are no
major unconformities or stratigraphic disruptions in the sediment sequence. All the
features referred to by the archaeologists as proto-orthostats, hammer stones, wedges,
sliding striations, pillars, props, fulcrums and railway tracks, are more convincingly
interpreted as entirely natural features associated with multiple late-glacial and Holocene
rock face collapses.
Page 14
It is also argued by Parker Pearson that the Rhosyfelin rhyolite was revered or special
enough to have justified a vast expenditure of time and effort on the part of Neolithic
tribesmen, without metal tools, with a view to extracting bluestones and transporting
them from this site to Stonehenge. However, Rhosyfelin rhyolite has not been used in any
of the Pembrokeshire standing stone settings or burial chambers, so it was clearly not
deemed to hold any mystical or magical properties or to have value for constructional
purposes. Furthermore, as indicated above, the “proto-orthostat” which is the focus of so
much attention is so fragile and seriously fractured (like other large stones at the site) that
its chances of surviving even a short haulage expedition by land or sea would have been
minimal.
The “proto-orthostat” appears to have fallen into its present position at some stage after
the deposition of the Rhosyfelin till. The archaeologists appear to have identified the red-
stained till surface as the putative quarry floor. Thus, if the archaeological hypothesis is
correct, all of the sediments above the glacial and fluvioglacial layer at Rhosyfelin must
have accumulated during the last 5,000 years. In other words, in this small valley with
steep bounding slopes, there is no sedimentary record of the period between ice wastage
and the end of the Mesolithic, spanning a period of c. 15,000 years. It is more likely that
most of the sediments above the glacial and fluvioglacial layer at this site have
accumulated gradually over a time-span of about 20,000 years, with the identified upper
layers (Fig. 4) representing a sequence of climatic oscillations yet to be quantified by
radiocarbon and other dating techniques.
While there appears to be no landform, rock mechanics or sedimentary evidence that this
was a Neolithic quarry site devoted to the extraction of bluestone orthostats destined for
use at Stonehenge, we would accept the possibility that there may have been temporary
Mesolithic, Neolithic or later camp sites here over a very long period of time, as in many
other sheltered and wooded locations in north Pembrokeshire. Parker Pearson has
reported (in public lectures) that a hearth and other occupation traces have been found
near the tip of the spur, and it is anticipated that this will be confirmed by radiocarbon
dating and artifact finds. Sites such as this may have been used in the context of a
hunting, fishing and gathering economy involving seasonal migrations (Bell and Walker,
1992). It is also possible that rhyolitic raw materials from this site may have been used in
the manufacture of blades and other cutting implements which changed hands during
tribal trading activities (Pitts, 2013).
We trust that our initial observations at Rhosyfelin will stimulate further detailed research
by those who have access to modern analytical and dating methods.
Acknowledgements
We thank Dr Rob Ixer, Dr Richard Bevins and Professor Mike Parker Pearson and his team
for drawing this site to our attention and for making it possible to examine it on many
Page 15
occasions. We also thank Prof Danny McCarroll, Dr John Hiemstra, Dr Simon Carr, Dr
Martin Bates and Dr Rick Shakesby for their on-site observations, and the geologist
members of the Association of Welsh RIGS (AWRG) for their assistance in creating a
Regionally Important Geodiversity Site record while the essential features could still be
seen. We thank Prof. David Evans and another referee for many helpful comments on an
earlier draft of this paper.
References
Bell, M. & Walker, M.J. (1992). Late Quaternary Environmental Change, Longman, 273 pp.
Bowen, D.Q. (1982). Pleistocene deposits and fluvioglacial landforms on North Preseli. Ch
16 in Bassett, M.G. (ed) Geological Excursions in Dyfed, South-West Wales, GA / National
Museum of Wales, 289-295.
Bowen D.Q., McCabe, A.M., Rose, J., & Sutherland,D.G.!(1986). !Correlation of Quaternary
Glaciations in England, Ireland, Scotland & Wales. Quaternary Science Reviews. 199-340.
British Geological Survey (2010). !1:50,000 Series, England and Wales ( Sheet 210: Fishguard. !
Bedrock and superficial deposits).
Campbell, S. & Bowen, D.Q. (1989). Quaternary of Wales, Geological Conservation Review
Series, No. 2, Nature Conservancy Council, Peterborough.
Downes, J. (2011). Folds, Faults and Fossils: exploring geology in Pembrokeshire. Llygad
Gwalch, 264 pp.
Etienne, JL, Hambrey, M.J., Glasser, N.F. & Jansson, K.N. (2005). !West Wales, !Ch 7 in
Lewis, C.A. and Richards, A.E. (eds) The Glaciations of Wales and Adjacent Areas, Logaston
Press, pp 85-100.
Evans, D.J.A., Clark, C.D. & Mitchell, W.A. (2005). The last British Ice Sheet: A review of
the evidence utilised in the compilation of the Glacial Map of Britain. Earth Science Reviews,
70 (3-4). pp. 253-312.
Eyles, N. and Sladen, J.A. (1981). !Stratigraphy and geotechnical properties of weathered
lodgement till in Northumberland, England. !Quarterly Journal of Engineering Geology and
Hydrogeology, 14, pp 129-141.
Gibbard, P. & Clark, C.D. (2011). Pleistocene Glaciation Limits in Great Britain.
Developments in QuaternaryScience Vol.15, 75-93.
Gregory, K.J. & Bowen, D.Q. (1966). !Fluvioglacial deposits between Newport, Pembs, and
Cardigan. !In Price, RJ (Ed) Deglaciation: Occ. Pub BGRG No 2, pp 25-28.
Page 16
Hambrey, M. J., Davies, J. R., Glasser, N. F., Waters, R. A., Dowdeswell, J. A., Wilby, P. R., Wilson,
D. & Etienne, J. L. (2001). Devensian glacigenic sedimentation and landscape evolution in the
Cardigan area of southwest Wales. Journal of Quaternary Science 16, 455–482.
Hubbard A., Bradwell, T., Golledge, N., Hall, A., Patton, H., Sugden, D., Cooper, R., &
Stoker, M. (2009). Dynamic cycles, ice streams and their impact on the extent, chronology
and deglaciation of the British-Irish ice sheet. Quaternary Science Reviews 28, 758-776.
Ixer, R.A. (2012). !!Digging into Stonehenge’s past. !Mineral Planning, issue 143 / October
2012, p 13
Ixer, R.A. and Bevins, R.E. (2011). Craig Rhos-Y-Felin, Pont Saeson is the dominant source
of the Stonehenge rhyolitic ‘debitage’. Archaeology in Wales 50, 21–31.
Ixer, R.A and Bevins, R.E (2013) Chips off the old block: the Stonehenge debitage dilemma.
Archaeology in Wales 52, 11-22.
Ixer, R.A. and Bevins, R.E. (2014). !The Vexed Question of the Stonehenge Stones. !British
Archaeology, Sept-Oct 2014, !50-55
John, B.S. (1970a). Pembrokeshire. In Lewis, C.A. (ed.) The Glaciations of Wales and adjoining
regions. Longman, London. 229-265.
John, B.S. (1970b). The Pleistocene drift succession at Porth Clais, Pembrokeshire. Geol.
Mag. 107, 439-457.
John, B.S. (1973). Vistulian Periglacial Phenomena in South-West Wales, Biuletun
Peryglacjalny !22, 185-212.
John, B.S. (2013) A Long History of Rhosyfelin. ! !Scribd.com !http://www.scribd.com/doc/
150104599/A-Long-History-of-Rhosyfelin
John, B. S. and Elis-Gruffydd, I. D. (1970). Weichselian Stratigraphy and Radiocarbon
Dating in South Wales. Geologie en Mijnbouw, 49 (4): 285-296.
John, B.S. (2008) Pembrokeshire Coast Path. Official National Trail Guide, Aurum / CCW /
Natural England, 168 pp.
John, B.S. and Jackson, L. (2009) Stonehenge’s Mysterious Stones, !Earth Magazine, January
2009, pp 8-15
McCarroll, D. (2001). The glacial geomorphology of West Wales. In Walker, M.J.C. &
McCarroll, D. The Quaternary of West Wales: Field Guide. Quaternary Research Association,
London, 9-16.
Page 17
Parker Pearson, M. !(2012). !Stonehenge. Exploring the Greatest Stone Age Mystery. Simon &
Schuster, London, 406 pp. (Chapter 17)
Parker Pearson, M. (2013). Researching Stonehenge: Theories Past and Present.
Archaeology International, 16, DOI: http://doi.org/10.5334/ai.1601
Patton, H., Hubbard, A., Glasser, N.F., Bradwell, T., & Golledge, N.R., (2013). The last
Welsh Ice Cap: Part 2 – Dynamics of a topographically controlled ice cap. Boreas 42(3),
491-510.
Pitts, M. !in Ixer, R.A. and Bevins, R.E. !(2013). !A re-examination of rhyolitic bluestone
‘debitage’ from the Heelstone and other areas within the Stonehenge Landscape
Wilts Arch & Nat Hist Mag 106 (2013), pp 1-15.
Thorpe, R.S., Williams-Thorpe, O., Jenkins, D.G. and Watson, J.S., with contributions by Ixer, R.A.
and Thomas, R.G. 1991. The geological sources and transport of the bluestones of Stonehenge,
Wiltshire, UK. Proceedings of the Prehistoric Society 57, 103-57.
Walker, M.J.C, Buckley, S.L., and Caseldine, A.E. (2001). Landscape Change and Human
impact in West wales during the Lateglacial and Flandrian. In Walker, M.J.C. & McCarroll,
D. The Quaternary of West Wales: Field Guide. Quaternary Research Association, London,
9-16.
Woodland, A.W. and Evans, W.B. (1964). The geology of the South Wales Coalfield, Part 4.
The country around Pontypridd and Maesteg. HMSO Memoir.
Brian John
Trefelin, Cilgwyn
Newport
Pembrokeshire SA42 0QN
brianjohn4@mac.com
Dyfed Elis-Gruffydd
Ty'r Ardd, Llechryd
Ceredigion SA43 2NR
dyfedelisgruffydd@btinternet.com
John Downes
Mount Pleasant Cross, Cosheston
Pembroke Dock
Pembrokeshire SA72 4TZ
jd1936@btinternet.com
Page 18
