Pembrokeshire erratics on Lundy Island?

 Lundy Island traces of glaciation.  Above -- two photos of ice moulded (whaleback) forms in the north of the island.  Below:  Perched blocks / erratic boulders resting on glaciated pavements in the northern part of the island.

In a previous post, I speculated about the possibility that glacial features on Lundy Island might date from the Devensian:
http://brian-mountainman.blogspot.se/2011/01/did-devensian-ice-reach-lundy.html

Well, that has now been confirmed in an article that somehow escaped my attention -- published in 2012.  It's a hugely important paper, from Chris Rolfe and colleagues.  I'll devote more space to it in due course, but the key things are these:

(a)  detailed mapping of the glacial features on the island show that they are varied and relatively fresh, confirming the conclusions of Frank Mitchell many years ago that there was indeed glacial ice at one time even on the highest parts of the island;

(b)  cosmogenic dating shows that the last ice cover here was during the Devensian -- although the suggestion is that the ice was present in the Early to Mid-Devensian, and had nothing to do with the Late Devensian ice advance which brought ice to the northern shores of the Scilly Isles.  This is an anomaly that needs to be sorted out....

(c)  Erratic clasts in the glacial deposits or scattered across some ice-scoured pavements on the island suggest that the ice came from Pembrokeshire.

Now we need some help, folks.  Richard, Rob, Olwen and any other geologist reading this -- can you please give us your considered opinions on the rhyolite erratics?  Below are the geochemical analyses presented in Table 1 (click to enlarge).  There are two rhyolite boulders:

The quartzites might also be interesting.........
Anyway, here is the text description:

 =================================

"Paired 26Al and 10Be exposure ages from Lundy: new evidence for the extent and timing of Devensian glaciation in the southern British Isles"
C.J. Rolfe, P.D. Hughes, C.R. Fenton, C. Schnabel, S. Xu, A.G. Brown
Quaternary Science Reviews 43 (2012) 61e73

Extract:
3.3.2. Erratic gravels in the north island watershed
The gravels that are widely spread over the northern island watershed include numerous clasts that are foreign to the local lithologies. Mitchell (1968) identified 100 rounded clasts. The methods used by Mitchell to identify these clasts is not stated in his paper only that they “were collected and identified as follows”:
“Possibly of island origin” 35 clasts including quartz: 22; miscellaneous igneous: 9; mica schist: 2; haematite: 1; limonite: 1 “Probably not of island origin” 38 clasts including grey quartzite: 25; pink quartzite: 13
“Certainly not of island origin” 27 clasts including coarse sand- stone: 9; feint: 8; chert: 3; micaceous sandstone: 2; sandstone with carbonaceous debris: 1; greywacke: 1; ignota: 3

The physical and geochemical differences between the erratics and the Lundy granite are clear from the XRF analysis of 8 clasts in this study (Table 1). The sample of Lundy granite had the following major element composition: 71% SiO2; 16% Al2O3; 6% K2O, and; 3% Na2O. This is representative of Lundy granite and is broadly consistent with major element data from Lundy granite presented in Stone (1990). Three clasts had very different SiO2 contents of 94e98% (PH 2, 3 & 7). These are interpreted as quartz or quartzite erratics. All of the clasts were fine-grained but varied in colour and appearance (PH 2 and 3: light grey; PH 7: pink with quartz veins). Some of the most extensive quartzite formations in the southern Irish Sea area are present on the coasts of both northwest Wales in Anglesey and the Lleyn Peninsula (Phillips, 1992) whilst quartz pebbles are widespread in the Millstone Grit of the Pembrokeshire coalfield (Archer, 1965). The very high SiO2 content of these clasts is higher than some quartzites in Wales (e.g. Phillips, 1992) and it may be that these clasts are quartz erratics reworked from conglomerates such as Millstone Grit. Pink quartz veins are also present in Morte slate, which is present in the SE corner of the island and some of the quartz-rich samples may be of local origin. Nevertheless, there are no quartz or quartzite outcrops in the area of the gravel spreads on Lundy and these clasts are clearly erratic to where they are found. Two other clasts had SiO2 contents of 82% (PH 5) and 85% (PH 1). This suggests that the clasts are not volcanic since even the most quartz-rich rhyolites rarely exceed 80% SiO2. The clasts are more siliceous with lower aluminium oxide contents than most sandstones and mudstones of the southern Welsh basin, for example (cf. McCann, 1991). The clasts have no comparable equivalent within the Lundy volcanics. As noted in the previous paragraph, the southeast tip of Lundy is formed of Morte Slate and this lithology is likely to be present offshore. However, Upper Devonian slates of this region contain much less SiO2 (c. 55e60%) and much more aluminium oxides (>20%) and potassium oxides (!4%) (Cattell, 1998). Thus, these clasts have not been derived from the Lundy country rocks and can be considered to be erratics.

Two samples had similar physical properties (fine-grained light blue-grey clasts) with similar geochemistry with SiO2 contents of 75.8% and 79.5% and combined Na2O and K2O of 2.6% and 3.3%. These samples are similar to some rhyolites. On Lundy, rhyolite dykes are known to be present e including in the north of the island near the gravel spreads where the clasts were found (Thorpe and Tindall, 1992). However, the local rhyolites have much higher values of combined Na2O and K2O (7.8e10.5%). Furthermore, the clasts have much higher iron oxide content (Fe2O3: 6%) than most of the Lundy rhyolites that have otherwise comparable geochemistries (Fe2O3: 1e2%). The clasts have a closer geochemical composition to rhyolites from the thick rhyolite sequences of the Skomer Volcanic Group off the westernmost coast of Pembrokeshire (e.g. Thorpe et al., 1989) although the precise origin remains unknown.

In addition to the siliceous clasts analysed using XRF, several limestone clasts collected from the gravel spreads are clearly erratics since no carbonate bedrock lithologies are found on or near Lundy. Given that rhyolites, quartz/quartzites, siliceous sedimen- tary and carbonate limestones are common lithologies, a wide range of sources are possible as noted in the previous paragraphs. Nevertheless, the presence of these erratic clasts is consistent with transport and deposition in association with an Irish Sea Ice Sheet. The fact that the clasts are well-sorted into cobbles and gravels and predominantly rounded suggests deposition by water. The presence of these deposits on the watershed is consistent with sedi- ment release at the apex of bedrock obstacles in subglacial channels (cf. Lesemann and Brennand, 2009).

=============================

Now the question is this:  do any of the rhyolite erratics show any strong similarities with the rhyolites from Pembrokeshire, or more specifically from the Fishguard Volcanics, or more specifically from the Pont Saeson / Craig Rhosyfelin area?

And another question:  do any of these rhyolite erratics show similarities with any of the rhyolite orthostats or debitage  found at Stonehenge?