“1 mile SE of Callington on the A388,
from the S side of which a lane leads to Westcott. Site lies on hillside to N. First recognised as a Neolithic henge in 1951, the somewhat ploughed down bank and internal ditch cover an area 90m in diameter. An indistinct entrance occurs on the south side, but there appears to be no corresponding causeway over the ditch. There is no visible evidence to suggest that there were any internal stone or timber settings. Like Cornwall’s other two surviving henges...this site was positioned close to a major prehistoric trackway; trade may have played a part in the function of the henge. Balstone Down, close to the north, was the source of greenstone for the manufacture of Neolithic axes. The bank of the henge reaches a height of 1.0m; the maximum depth of the ditch is 1.2m.
GEOCHEMICAL FINGERPRINTING OF WEST CORNISH GREENSTONES AS
AN AID TO PROVENANCING NEOLITHIC AXES
M. MARKHAM AND P. A. FLOYD
M. Markham and P.A. Floyd. 1998. Geochemical fingerprinting of west Cornish greenstones as an aid to
provenancing Neolithic axes. Geoscience in southwest England, 9, 218223.
Of the large number of Neolithic stone axes made of greenstone, some 392 (referred to as Group 1 axes) are believed to have been
manufactured in west Cornwall around the Mount's Bay area. To aid the location of the greenstone that provided the materials for the axes,
geochemical fingerprinting of the axes and greenstone outcrops was undertaken in the Mount's Bay area to both discriminate the greenstone
localities and provide a basis for matching Group 1 axes. Nondestructive analysis of the axes was determined by a portable XRF
Spectrometer unit that gave comparable results for selected greenstone samples to standard laboratorybased XRF techniques. Geochemical
fingerprinting of the greenstone localities by portable XRF spectrometer provided a degree of discrimination between them, although
preliminary data on the axes suggests that there is not a very strong correlation between axe composition and possible greenstone sites.
Further work is required on other greenstone localities and axes as current data does not conclusively point to an origin in this area of west
Cornwall.
M. Markham, Department of Earth Sciences, Open University, Walton Hall, Milton Keynes, MK7 6AA
P. A. Floyd, Department of Earth Sciences, Keele University, Staffordshire, ST5 5BG
INTRODUCTION
The Devonian and Lower Carboniferous sequences of southwest
England are characterised by episodes of volcanic activity with the
production of largely submarine pillow lavas and highlevel intrusive
sills and sheets (Floyd, 1995). During the Variscan orogeny the rocks
were subjected to lowgrade alteration and the volcanics developed
secondary assemblages indicative of the pumpellyite and greenschist
facies of regional metamorphism (Floyd, 1995). Subsequently, on the
intrusion of the granite plutons, contact metamorphic effects were
locally superimposed with the production of new assemblages and
textures. Overall, the effects of alteration on the predominantly basic
volcanic rocks was two fold: (a) variably developed secondary
assemblages superimposed on a still recognisable primary basic
assemblage (pyroxeneplagioclaseilmenite±olivine), and (b) the
textural effects of deformation with the progressive development of a
foliation and eventually a marked schistosity.
From the industrial viewpoint the mineralogical alteration effects
on the basic volcanics has strengthened them and many of the massive
(nonfoliated) intrusive basic rocks are used for roadstone (Edmonds
et al., 1969). Stone Age man also appreciated the durable qualities of
the altered basic rocks and fashioned many hand axes from what is
generally referred to as "greenstone", that is, lowgrade
metamorphosed dolerites and basalts.
STONE AGE AXES
Stone axes are grouped according to petrological, mineralogical
and textural criteria. Over 40 different axe groups are recognised from
a total of 7625 axes found in Britain (of which only 3546 have been
grouped) (Clough and Cummings, 1988). Just over 1000 axes are
made of greenstone, of which 392 members (referred to a "Group 1")
have been identified as being manufactured in west Cornwall and
possibly originating from the Mount's Bay area (Figure 1), largely on
the basis of the site of origin and the local presence of greenstone
(Stone, 1951).
Group 1 axes, originally catalogued in the 1940s (Keiller, 1941),
are broadly described as uralitized gabbro with original pyroxene and
feldspar, with the characteristic development of a uralitic fringe of
bluegreen amphibole around the primary pyroxene; epidote, sphene
and chlorite are common accessory minerals. Because this assemblage
and texture is a very common feature of many Cornish greenstones,
the actual outcrops, representing the stone age factories used by
Neolithic man, have never been positively identified. However, the
assumed manufacture from this area is still largely based on the
apparent similarity between a petrographic examination of thin
sections of greenstone axes and the outcrops in the Mount's Bay area.
To date no definite match has been achieved, such that archaeologists
have suggested that the outcrops are actually underwater and have yet
to be sampled! (Evans, 1962)
PROJECT OBJECTIVES AND APPROACH
The prime object of this work is to determine from where the
Group 1 greenstone axes originated, using a combination of
petrographic and textural data coupled with new geochemical
discrimination techniques. Petrographic comparisons between axe and
potential outcrop are a standard method to identify sources (Markham,
1997) although in this case the results have proved inconclusive.
Geochemical fingerprinting of the axes, especially trace element
distributions, was considered a more sensitive and selective approach.
Also, there already exists a broadbased set of geochemical data on
Cornish greenstones (both intrusives and lavas) that shows some
significant chemical distinctions between different localities
composed of broadly similar mineralogical assemblages (Floyd,
1995). If both axes and greenstone outcrops can be sufficiently
discriminated geochemically this provides a sound basis for
comparison and the strong possibility of linking axe to source.
However, in order to provenance axes by geochemical means, a
nondestructive analytical method is needed to preserve the axe,
although ideally should be a similar technique to that used to analyse
field samples (to minimise inter technique errors). In this connection
we have used Xray Fluorescence Spectrometry, both a portable
instrument (PXRF) and a laboratorybased instrument (LXRF)(see
below).
ANALYTICAL METHODS AND COMPARISONS
XRF Spectrometry determines the concentration of elements in a
sample by measuring the variable intensity of fluorescent energies
given off by a sample when illuminated by xrays.
218
219
219
West Cornish greesstones as Neolithic axes
Figure 1
Map of west Cornwall showing the occurrence of greenstone outcrops
sampled for this study. (Greenstones solid ornament, Granite ‘+’ornament)
Considerable preparation is required prior to analysis, involving
crushing to a fine powder and subsequent pelletizing, and thus
destroying the original sample. This is standard procedure for
laboratorybased instruments, but is not an option for valuable and
irreplaceable axes. In this context the PXRF has a number of
advantages it is portable (carried by one person), is nondestructive,
the samples need no preparation, and can work off mains or battery.
Provided the sample has a reasonably planer surface greater than 25
mm in diameter it can be geochemically analysed. For PXRF the
radiation energies needed to fluoresce the sample are obtained through
the radioactive decay of isotopes of Fe, Cd and Am. The reflected
energies are detected by a high purity HgI2 detector which requires no
external cooling. Table 1 summarises the main features of both PXRF
and laboratory XRF (LXRF) equipment
Before meaningful geochemical comparisons can be made
between the axes and potential sample localities, it is necessary to
check that the PXRF instrument is capable of producing similar
results to laboratory based XRF instruments (at Open University) and
also existing data from the literature derived by similar means (at
Keele University). A selection of greenstone samples collected from
around Penlee Point (Penlee Lifeboat station and Mousehole
foreshore) and from within and shore side of the Penlee Quarry at
Cam Gwavas were analysed by LXRF and PXRF at the Open
University and compared with existing data by Floyd (1976, 1983,
1984) and Floyd and AlSamman (1980). The results for the three sets
of data (PXRF, LXRF, literature) are shown in Figure 2 and Table 3,
and clearly demonstrate that each data source can separately define
and discriminate the two sampled sites (foreshore versus quarry). We
consider that the overlap of geochemical data derived from the PXRF
is sufficiently good to allow discrimination of the axes and sources,
and also compares well with existing LXRFderived data.
FEATURES OF GREENSTONE SITES
Figure 1 shows the sites subjected to preliminary evaluation as
potential axe sources. Initially the choice of locations was determined
by the proximity to Mount's Bay and environs (the suggested origin of
Group 1 axes), and the amount of existing data available. Table 2
summaries the location and nature of emplacement for the greenstone
bodies; detailed descriptions of some of the following sites may be
found in Floyd et al. (1993).
Cudden Point [SW 548 275] This is a relatively coarse, massive,
sheet like intrusive body of greenstone composed of metagabbro and
Figure 2
Envelopes enclose geochemical data from two major greenstone
outcrops around Penlee Point and adjacent foreshore, and the Carn
Gwavas quarry, derived by PXRF (MM(PXRF)), laboratorybased XRF
by Markham (MM(LXRF)) and by Floyd (PAF(LXRF)). Note the
general overlap of PXRF and LXRF data from the two localities.
M. Markham and P. A. Floyd
metadolerite, with strongly foliated margins (Floyd and Lees, 1972).
It shows the typical development of pale green uralitic actinolite
fringing large clinopyroxene prisms, some of which may exhibit
pigeonite lamellae. Petrographically it is characterised by rare crystals
of brown primary amphibole, two generations of secondary amphibole
(pale uralite and bluegreen), and replaced ovoids of olivines
surrounded by pyroxene.
Trenow [SW 529 303] This site covers a series of scattered
outcrops around Trenow Cove including a small quarry at
Perranuthnoe. Periglacial "head" covers much of the area behind the
wavecut platforms and it is not clear if more than one massive
intrusive body exists here. The greenstones here are similar to Cudden
Point, but generally finer grained and composed of metadolerite and
metabasalt. The metamorphic mineralogical growth has developed
further with a higher proportion of secondary minerals, especially
amphibole, albite, epidote, chlorite and sphene. In order to decrease
the potential variability of results all data presented below comes from
the west end of Trenow Cove.
Penlee Point [SW 474 269] This includes much of the
Mousehole foreshore around the Penlee Life Boat station to Carn
Gwavas to the north. It includes the two sills found in the inland
quarries at Penlee Point proper. Although within the Land's End
granite aureole, the greenstones here show similar mineralogical and
textural features to lowgrade regional metabasites elsewhere in west
Cornwall. They are mainly metadolerites with uralitic actinolite,
grading to actinoliteplagioclase hornfelses with a weak foliation.
Evidence of contact metasomatism is shown by the development of
rare blue zoned tourmaline, and the patchy hydrothermal "bleaching"
of some outcrops.
Carn Gwavas [SW 470 280] This is restricted to the huge (now
disused) quarry between the shore and the granite inland and the
immediate foreshore east of the quarry. Although often considered
and quarried as "greenstone", this rock type is not basic, but more
intermediate in composition and thus quite distinct mineralogically
and chemically from typical greenstones. It has an interlocking
granular texture composed almost entirely of albite and ragged
amphibole. It displays variable degrees of hydrothermal alteration
with the development of kaolinite, biotite, chlorite, as well as a
complex of sulphide mineralization veinlets (Floyd, 1965). Axes
derived from this source would be very distinctive mineralogically
and chemically.
Zennor [SW 450 394] This forms a small resistant headland of
greenstone west of St. Ives within the granite contact aureole. It is
mainly a finegrained amphibolerich hornfels of massive aspect.
Gurnard's Head [SW 432 387] This headland is largely composed of
a massive sheet like body that grades upwards into a pillow lava
sequence and represents a highlevel intrusion near the sediment
water interface. Again within the granite aureole, the original
mineralogy and texture have been replaced by a hornfelsic matte of
Portable XRF
actinolite with subsidiary plagioclase and rare biotite replacing
amphibole. Chloritefilled amygdales in the pillowed section are still
recognisable.
Table 2 shows the mineralogical features of the greenstones found
at these sites, as well as summarising some of their geochemical
characteristics.
GEOCHEMICAL DISCRIMINATION OF GREENSTONE SITES
In this section we attempt to discriminate the different selected
sites on the basis of PXRF data alone prior to any comparison with
axe data derived by the same technique. The PXRF has been set up to
measure 13 elements such that a large number of binary plots can be
generated. In this preliminary exercise, however, a number of diagrams plotting absolute abundances measured in ppm have been selected to illustrate the chemical diversity of the greenstone sites and
provide a measure of discrimination (Figure 3 and Table 4). We have
chosen to present the data as averages ± 1sd since this presents a clearer indication of the outcrops and reduces clutter on the plots. 2SD
'space' will be used in the final assessment although it is recognised that a multivariate statistical approach will be needed in the future.
As seen from the plots in Figure 3, a number of chemical features
characterise the sites which can be summarised below:
(a) In all plots, the Cam Gwavas body is geochemically
distinctive relative to the rest, being typified by high Zr and Y,
coupled with low Ti, Fe, and Sr. Also, as mentioned above, it is also
mineralogically and texturally distinctive. These features provide a
good discrimination for this type of "greenstone", whereas some of
the
other
greenstone locations have more overlapping
characteristics.
(b) Cudden Point also has a distinctive chemistry relative to the
other greenstones, with very low incompatible element contents,
especially Ti, Zr, Y and Nb; the Zr/Y ratio is also low at between 26.
(c) The Trenow data shows the largest standard deviation of data
often overlapping other greenstone sites, with the exception of
Cudden Point and Cam Gwavas. There is a degree of chemical
overlap between Penlee Point, although this locality is within the
granite aureole and would be expected to show some mineralogical
and textural differences.
(d) There are small chemical differences between the actinolite
bearing greenstones (hornfelses) within the granite aureole at
Gurnards Head and Zennor. In general, Gurnards Head has
systematically higher Fe, Y and Ti than Zennor.
As seen (Figure 3) discrimination is by no means clear cut for all
greenstone compositions. The log plot of Ti versus Zr provides one of
the better discriminations with different locations showing partial
separation on a curved trend, starting with (i) Cudden Point, (ii)
Zennor, (iii) Gurnards Head + Penlee, (iv) Trenow and finally (v)
Carn Gwavas. The log plot of Y versus Zr, on the other hand, has an
almost linear trend with the same progression.
Precision (relative usually 10% or better
λd XRF
error)
Major elements to 0.5%
Minor elements to 1 5 %
(of value measured)
Operation One person portable, hand held or stand
mounted sensor, battery or mains powered
Laboratory based, high power xray tube and
cooling required
Detection limit Major elements to 100ppm
Range of Elements
Heavier than potassium
Heavier than boron
Table 1. Comparison of portable and laboratory based XRF equipment (after Markham, 1997: Potts, 1995)
220
West Cornish greesstones as Neolithic axes
Figure 3
Geochemical data derived by PXRF data (averages ± 1 sd) from selected greenstone localities in west Cornwall. Note that some diagrams provide a degree of discrimination between selected greenstone localities.
PRELIMINARY AXE DATA
At the moment only a preliminary assessment and comparison between the greenstone sites and the axes can be made. This is a consequence of discovering that the geochemistry of the Group 1 axes
was more heterogeneous than expected with some axes having
compositions that lay well outside the normal range. Histogram plots
of axe elements showed a nonstandard distribution and upon
investigation some of the previously classified axes may have to be
regrouped. Further work on existing petrographic thin sections and
accession records is being carried out and this will need to be
concluded before axes are eliminated from Group I. Hence this feature
suggests that all the socalled Group 1 axes may not belong to this
group or indeed be derived from Cornish greenstones at all. Thus,
some 50 Group 1 axes have been measured by PXRF and an average
together with a characteristic range calculated. The range identified in
Table 4 has been taken from the aforementioned frequency histograms
of the axe data and does not represent a standard deviation for the full
data set, that is, it excludes those axes with apparently nonGroup 1
chemistry.
The axe average and range is plotted in Figure 4 (with data in
Table 4) and compared with the geochemical fingerprints for the
greenstone sites. The Ti versus Zr plot chosen to illustrate the
comparison shows the axe range is comparable with both Penlee and
Gurnards Head, with Carn Gwavas, Cudden and Zennor outside the
axe range. Whilst encouraging, these are early results and more work
will be required to build up a conclusive picture.
CONCLUSIONS
One of the most important conclusions to draw from this type of
geochemical fingerprinting of archaeological artefacts is that a PXRF
spectrometer can provide sensitive and accurate results without the
destruction of the specimen. We have demonstrated that comparable
data can be obtained from a PXRF instrument relative to a standard
laboratorybased XRF spectrometer and without the problem of sample
221
223 222
M. Markham and P. A. Floyd
Outcrop Geological location Type of Intrusion Mineralogy Texture Geochemistry
Cudden Point Outside L.E. aureole, Major sill Primary: olivine clinopyroxene ophitic meta gabbroic/ high Cr & Ni
but affected by plagioclase opaques, rare primary meta doleritic foliated low incompatible (Y,Zr,Ti, etc,.)
Godolphin granite amphibole margins
Secondary: uralite & blue green
amphibole, sphene, albite
Trenow Cove As above sheet like intrusion Primary: clinopyroxene plagioclase metadolerite sub ophitic less primitive than Cudden, more
opaques texture of uralitic actinolite fractionated
Secondary: actinolite albite epidote
chlorite sphene
Carn Gwavas Within L.E. aureole pipe like intrusion Primary: None remaining broadly granular, result of alkaline, with high incompatibles
<600 m from granite Secondary: albite biotite chlorite recrystallisation through (esp Ti,Nb,Y & Zr)
outcrop sphene common sulphide mineralisation contact metamorphism
Penlee Within L.E. aureole small sills Primary: clinopyroxene opaques metadolerite sub ophitic to typical of intraplate basalts (OIB),
<800 m from granite Secondary: uralitic actinolite albite rare foliated at edge of intrusion with wide variation due to fractional
outcrop tourmaline crystallisation
Gurnards Within L.E. aurole pillow lavas & high Primary: None remaining (probably metadolerite with hornfelsic tholeiitic affinities with enriched
Head < 1km from granite level intrusions pyroxene plagioclase ilmenite). texture. E MORE features
outcrop Secondary: actinolite plagioclase
opaques biotite, subsequent alteration to
chlorite
Zennor Within L.E. aureole unknown but Primary: None remaining metadolerite no detailed geochem available,
< 1 km from granite probably sill like Secondary: actinolite plagioclase opaques,
outcrop plus chlorite
REFERENCES
West Cornish greesstones as Neolithic axes CLOUGH T.H.McK., CUMMINGS W.A. 1988. Stone Axe Studies, CBA
Research Report 67
EDMONDS, E.A., WRIGHT, J.E. and WILLIAMS, M. 1969. British regional
geology: SouthWest England, 3rd. ed. Institute of Geological Sciences,
HMSO, London.
EVANS, E.D., GRINSELL, L.V., PIGGOT, S. and WALLIS, F.S. 1962.
Fourth report of the subcommittee of the South Western Group of Museums
and Art Galleries on the petrological identification of stone axes. The
Prehistoric Society 10, 209266
FLOYD, P.A. 1965. Argillization of basic homfelses from the Land's End
granite aureole, Cornwall. Clay Minerals, 6, 4558.
Figure4
TiZr plot with average and range of geochemical data for Group 1 axes
superimposed on the fields for greenstones derived from various
localities in west Cornwall. Note that the mean overlaps Penlee and
Gurnards Head, but is separate from other localities.
preparation. In this context the PXRF data was used to generate a
chemical database from a number of greenstone localities in west
Cornwall and also analyse a selection of Neolithic axes (referred to as
Group 1 axes) thought to have been manufactured from greenstone
material in the Mount's Bay area.
A selection of major and trace element binary plots allows a
degree of selective discrimination of the different greenstone
localities. Different plots highlight the specific chemical features of
the greenstones, although there is often considerable overlap between
some localities. More work needs to be carried out to identify the best
set of geochemical discriminants. At the present time the TiZr
diagram provides the best discrimination, but this must be used in
conjunction with other specific characteristics, such as high or low
abundances of Ni, Cr, Fe and Y.
Finally the average and range of Group 1 axes plotted on the TiZr
diagram reveals an encouraging overlap with one Mount's Bay site
(Penlee) and only partial overlap with the other greenstone locality
data. The hypothesis that Mount's Bay was the source of Group I axes
remains unproven, but these results and ongoing work have a good
chance of resolving the 50 year old issue.
ACKNOWLEDGEMENTS
We would like to thank Peter Webb and Olwen WilliamsThorpe
for valuable comments on early drafts of this paper and to The Ian
Gass Fund of the Open University for assistance with the provision of
the PXRF.
FLOYD, P.A. 1976. Geochemical variation in the greenstones of SW England.
Journal of Petrology, 17, 522545.
FLOYD, P.A. 1983. Composition and petrogenesis of the Lizard Complex and
preorogenic basaltic rocks in southwest England. In: The Variscan Fold belt
in the British Isles, Ed: P.L. Hancock. Adam Hilger, Bristol, 130152.
FLOYD, P.A. 1984. Geochemical characteristics and comparison of the basic
rocks of the Lizard Complex and the basaltic lavas within the Hercynian
troughs of SW England. Journal of the Geological Society of London, 141, 61
70.
FLOYD, P.A. 1995. Igneous activity basaltic volcanism in the
Rhenohercynian Zone, N. Europe. In: PrePermian geology of Central and
eastern Europe. Eds: Dallmeyer, RD., Franke, W. and Weber, K. Springer
Verlag, Berlin, 5981.
FLOYD, P.A. and ALSAMMAN, A.H. 1980. Primary and secondary
chemical variation exhibited by some west Cornish volcanic rocks.
Proceedings of the Ussher Society, 5, 6875.
FLOYD, P.A. and LEES, G.J. 1972. Preliminary petrological and geochemical
data on the Cudden Point greenstone. Proceedings of the Ussher Society, 2,
421423.
FLOYD, P.A., EXLEY, C.S. and STYLES, M.T. 1993. Igneous rocks of SW
England. Geological Conservation Review Series no.5, Chapman and Hall,
London, 256pp.
KEILLER, A., PIGGOT, S. and WALLIS, F.S. 1941. First report of the sub
committee of the South Western Group of Museums and Art Galleries on the
petrological identification of stone axes. The Prehistoric Society 2, 12731277
MARKHAM, M, 1997, Geology and Archaeology: a search for the source
rock used by British Neolithic Axe Makers. Open University Geological
Society, 25 Anniversary Edition.
POTTS, P.J., WEBB, P.C. and WILLIAMSTHORPE, O. 1995. Analysis of
silicate rocks using field portable xray flourescence instrumentation
incorporating a mercury (II) iodide detector: A preliminary assessment of
performance. Analyst 120, 1273 1278
STONE, J.F.S and WALLIS, F.S. 1951. Third report of the subcommittee of
the South Western Group of Museums and Art Galleries on t
Higher Kiln Quarry
We have very limited information about the early days of Higher Kiln Quarry but believe that it, and Bakers Pit Quarry at the top of Buckfastleigh Hill, became significant sites for the extraction of limestone early in the 19th century and continued in operation until about the time of the First World War. They were ued principally to feed the limekilns on the Higher Kin Quarry site with crushed limestone that could be turned into lime for use in agriculture.
The operation of the limekilns is described in more detail on
The section above shows the extent of the quarries in Buckfastleigh Hill. The limestone is shown in blue and it sits on top of the old volcanic ash, now tuff (in brown), and shales (green). Only Bullycleaves Quarry survived into the late 20th century but is now closed; limestone from this qaurry was used in the rebuilding of Buckfast Abbey. Caves (in dark grey) exist throughout the limestone – the largest of these is the Reeds-Bakers Pit system with a total passage length of some 5 km. Higher level passages in this system come close to the surface in the area of the churchyard.
The quarries were developed downwards and into the Hill, destroying parts of the cave systems as they went. The tuff (of no interest to the quarrymen) is exposed in the floors of both Bullycleaves and and Higher Kiln quarries and was visible in Bakers Pit quarry before is was used as a landfill site. Some of the caves were known at the time of William Pengelly’s visit in the mid 19th century but there’s little evidence that mostwere explored until the period between the two WorldWars.
Rivers and navigable creeks, p. 36. Tamar, Lynher, p. 38. Tide, or Tidi, p. 40. Seaton, ib.
Loo, or Eaft-Loo, ibid. ProfpoCt of Loo Bridge," ib. Duloo, or Weft Loo river, p. 41. fawy, ib.
Fal, 42, and it’s harbour. Hel, or Heyl river in Kerricr, p. 43. Lo or Low river in Kerrier, p 44.
A1
Heyl in Penwith, ibid. Ganal creek, p. 45. River Alan, al Lamel, ibid. Wade navigable rivers in
may be made notbeneficial, p. 47. Subject: to obftrudtions, p. 49.