Content and licensingview original scan buy a printed copy

Geology of Unst and Fetlar in Shetland. Memoir of the British Geological Survey, Sheet 131 (Scotland)

Bibliographical reference: Flinn, D. 2014. Geology of Unst and Fetlar in Shetland. Memoir of the British Geological Survey, Sheet 131 (Scotland).

Geology of Unst and Fetlar in Shetland — The district described in this memoir is covered by Sheet 131 (Scotland).

Derek Flinn (1922–2012) To commemorate a lifetime of geological studies on Shetland

British Geological Survey

Geology of Unst and Fetlar in Shetland: Memoir for 1:50 000 geological sheet 131 (Scotland) Unst and Fetlar

D Flinn. Contributor H M Prichard

Author D Flinn, DSc, PhD University of Liverpool

Contributor H M Prichard, PhD, MBA University of Cardiff

(Front cover) Cover photograph Muckle Flugga and Out Stack (the end of the land) (P536015).

(Rear cover)

(Frontispiece) Loch of Watlee, Unst, looking north, [HP 5870 1450] (P536021).

(Dedication) Derek Flinn portrait

This memoir, and the 1:50 000 scale geological map that it describes, are the products of a mapping contract between the Natural Environment Research Council and the University of Liverpool. The interpretations presented are those of the authors.

Keyworth, Nottingham: British Geological Survey 2014. © NERC copyright 2014 First published 2014. ISBN 978 0 85272 755 3

Your use of any information provided by the British Geological Survey (BGS) is at your own risk. Neither BGS nor the Natural Environment Research Council gives any warranty, condition or representation as to the quality, accuracy or completeness of the information or its suitability for any use or purpose. All implied conditions relating to the quality or suitability of the information, and all liabilities arising from the supply of the information (including any liability arising in negligence) are excluded to the fullest extent permitted by law.

Copyright in materials derived from the British Geological Survey’s work is owned by the Natural Environment Research Council (NERC) and/or the authority that commissioned the work. You may not copy or adapt this publication without first obtaining NERC permission.

Contact the BGS Intellectual Property Rights Manager, Keyworth, Nottingham. You may quote extracts of a reasonable length without prior permission, provided a full acknowledgement is given of the source of the extract. The National Grid and other Ordnance Survey data. © Crown Copyright and database rights 2014. Ordnance Survey Licence No. 100021290.

Other publications of the Survey dealing with this district and adjoining district

Books

• Memoirs
• Geology of Yell and some neighbouring islands in Shetland. 1994
• The geology of Western Shetland 1976. British Regional Geology
• Orkney and Shetland 1976

Maps

• 1:625 000
• United Kingdom (North Sheet) Bedrock Geology 2007
• Quaternary Geology 1977
• Gravity Anomaly 2007
• Magnetic Anomaly 2007
• 1:250 000
• Shetland Sheet (60°N 02°W) Solid Geology 1984
• Sea-bed Sediments & Quaternary Geology 1984 Bouguer anomaly 1978
• Aeromagnetic anomaly 1986
• 1:63 360
• Northern Shetland Solid Geology 1968
• Drift Geology 1968 Western Shetland Solid Geology 1971
• Drift Geology 1971 Central Shetland Solid Geology 1982
• Drift Geology 1981 Southern Shetland Solid Geology 1978
• Drift Geology 1978
• 1:50 000
• Yell
• Solid and Drift Geology 1994 Unst and Fetlar
• Solid and Drift Geology 2002

Acknowledgements

S J Blackbird, R Thomas, R N Brotherton, and Drs B Atkins, R M Aziz, D T Moffat, A S Gamil and T Brewer assisted with rock analyses; A McCormack assisted with computing; and D Plant, Geology Department, University of Manchester assisted with all the electron-microprobe analyses presented in this memoir.

Chemical analyses of rocks are by the Geochemical Laboratory of the University of Liverpool, except where stated otherwise and electron-microprobe analyses of minerals are by the author, except where stated otherwise.

Compilation and overall editing of the memoir for the BGS was by D Stephenson and S D G Campbell, with detailed scientific editing by A G Leslie (chapters 3, 4 and 5), M R Gillespie (chapters 6, 7, 8 and 11), J W Merritt (Chapter 9), A G Gunn (Chapter 11) and M T Styles (chapters 2, 5, 6, 7, 11 and 12). Figures were drafted by C R Woodward. Photographs are all by the author except where stated otherwise.

The cost of printing this memoir was generously supported by the Shetland Amenity Trust.

The word ‘district’ used in this memoir refers to the geological 1:50 000 Series Sheet 131 (Scotland) Unst and Fetlar.

Notes

• The sheet lies across the east–west boundary between National Grid squares HU and HP, but since there is no repetition of grid references within the area surveyed, grid letters are not quoted, except where localities lie outwith the area of the sheet.
• Most of the photographs taken by the authors and used in this memoir are registered in the BGS photographic collection and have been assigned a P number (PXXXXXX). The collection can be accessed via GeoScenic, available on the BGS website, www.bgs. ac.uk. In addition there are references to P numbers throughout the text. These can also be viewed in GeoScenic.

Foreword

This memoir, and the 1:50 000-scale geological map that it accompanies, are the products of a contract between the Natural Environment Research Council and the University of Liverpool to produce the map; and a subsequent contract between the British Geological Survey and Professor Derek Flinn to produce the memoir. Such contracts stem from the NERC policy of encouraging academics with substantial knowledge about specific areas in the UK to transfer their infor- mation into the public domain. This is done by funding them to extend mapping from areas in which they have worked to the boundaries of BGS map sheets. The maps, and accompanying descriptive memoirs (supported by open file reports and archives), are then published by the British Geological Survey. In the case of Unst and Fetlar the contracts have built on more than fifty years of experience of Shetland geology possessed by Professor Flinn, who resurveyed the solid and drift geology of the 1:50 000-scale Unst and Fetlar sheet (Sheet 131) and is the main author of this memoir. He had also earlier surveyed the area of Central Shetland (Sheet 128), published by the Institute of Geological Sciences (now BGS) as 1 inch to 1 mile geological maps in 1981 and 1982, and more recently produced the 1:50 000-scale geological map (Sheet 130) and accompanying memoir for the neighbouring island of Yell, published by the BGS in 1994. The 1:50 000-scale geological map of Unst and Fetlar (Sheet 131), to which this memoir relates, was published by the BGS in 2002. Some minor changes in interpretation by the author since that date have been incorporated in this memoir, which includes information gathered and interpretations made up to and including 2007.

Sadly Derek Flinn passed away in June 2012, whilst this memoir was undergoing final preparations for publication. The British Geological Survey is proud to publish it in his memory, reflecting a lifetime of dedicated geological work on Shetland in general and on Unst and Fetlar in particular. The printing of this memoir has been funded with the support of the Shetland Amenity Trust in recognition of Professor Flinn's contribution to Shetland Geology.

The geology of Unst and Fetlar is most notable for the presence of the Shetland Ophiolite-complex, arguably the most complete and convincing example of an ophiolitic succession in the British Isles. The Shetland Ophiolite-complex is interpreted as being preserved in two nappes, the Lower and Upper nappes, which were obducted onto a basement comprising three tectonic units, informally referred to as the Lamb Hoga, Valla Field and Saxa Vord blocks. These basement units are largely composed of Dalradian metasedimentary rocks and are separated from the inferred Moinian metasedimentary rocks of Yell, at the western edge of Sheet 131, by a tectonic feature referred to as the Boundary Zone. The Skaw Granite was intruded into, and thermally metamorphosed, rocks of the Saxa Vord Block and the Shetland Ophiolite-complex at about the time that the Upper Nappe of the ophiolite-complex was emplaced upon the Lower Nappe.

In compiling the present memoir, Professor Flinn was able to synthesise the wide array of his own extensive international publications, which have dealt with virtually every aspect of the geology of Unst and Fetlar. Although he drew on the work of others, where appropriate, the interpretations and overall models which he presented in this memoir are substantially his own, and have already withstood the test of peer review in a host of international publications over many years. His publications are extensively referred to throughout the memoir and remain the primary sources of ideas and interpretations in most cases. The only excep- tion to this is the description and interpretation of the platinum-group element (PGE) mineralisation of Unst, which has been undertaken by Dr Hazel Prichard of Cardiff University/Prifysgol Caerdydd. The Shetland Ophiolite-complex was one of the first ophiolite-complexes in which significant platinum and palladium concentrations were identified, and has been an important stimulus for research into PGE mineralisation in other ophiolite-complexes around the world. In describing the PGE, Dr Prichard has drawn upon her own extensive research, that of co-workers, and on work by the British Geological Survey. The economic potential of the PGE mineralisation has been assessed in detail, but is yet to be realised. However, significant chromite deposits have been worked at several times in the past, and their economic potential has been the focus of re-assessment. Steatite has had an even longer history of exploitation, dating back to Viking times, and has been extracted recently. Other minerals that have also been worked in the past also include serpentinite and kyanite. Limestone has been extracted locally.

The only younger deposits are of Quaternary age and are products of the last (Late Devensian) glaciation, together with Holocene beach and peat deposits, and many small deposits of diatomite that were discovered during the present survey. The peat has been exploited from time immemorial as a source of fuel, although its use has declined recently.

There are many well-displayed glacial features. Glacial drainage patterns are better developed in Unst than elsewhere in Shetland, and there is evidence of an ice-dammed lake (Milldale Lake) being present during deglaciation. The monadnock-like form of the Shetland archipelago is thought to have originated, whether by marine or subaerial erosion, prior to the Mesozoic Era and the cliffs of the present outer coast are the result of cliff retreat at intervals since then.

In addition to the data presented in this memoir, the British Geological Survey holds further data relating to the area, including a large number of photographic images of the area and geochemical data obtained by Professor Flinn and his co-workers, which can be made available to interested parties.

John N Ludden, PhD Executive Director British Geological Survey Environmental Science Centre Keyworth Nottingham NG12 5GG

Tectonostratigraphical succession and Quaternary sequence on Unst and Fetlar, Shetland.

Chapter 1 Introduction

Area

The area covered by the 1:50 000 Unst and Fetlar, Shetland Sheet (BGS, 2002), referred to here as the district, is the eastern part of the area covered by the one-inch-to-one-mile Northern Shetland Sheet (IGS, 1968a). It includes the islands of Unst and Fetlar, several smaller islands, and many nearby stacks, holms, skerries and baas (half-tide rocks). Between Unst and Fetlar are the islands of Linga, Uyea, Haaf Gruney, Sound Gruney, Urie Lingey, Daaey and to the west of Fetlar the island of Hascosay. Several smaller islets and rocks occur close to the north coast of Fetlar. Off the east coast of Unst are the islands of Balta and Huney and off the west coast many much smaller islets. Off the north-east coast of Unst are the Holm of Skaw and associated smaller rocky masses. To the north-west, lie the stacks and skerries associated with Muckle Flugga. These end in Out Stack, the most northerly land in the British Isles (Figure 1). Also included in the district are several promontories on the east coast of Yell, in particular Burra Ness, North Sandwick, Head of Gutcher and part of Crussa Ness, that were earlier described in the Yell Memoir (Flinn, 1994a). Unst is about 20 km long and 10 km wide and rises to a height of 280 m. Much of the island is low lying with a valley extending from the north coast to the south coast rising to a watershed at a height of 40 m above sea level in the middle of the island. The island is ringed by an almost unbroken line of cliffs ranging up to 160 m in height. Fetlar has a maximum dimension of 8 km from north to south and 10 km from east to west. It is surrounded, like Unst, by almost continuous cliffs, which rise to a maximum height of 124 m, but has less high ground, reaching a maximum height of 158 m in the centre.

Unst and Fetlar are both covered by crofting lands, now largely devoted to sheep grazing. Unlike the rest of Shetland, little or none of the original peat cover remains. This is partly due to exploitation of the peat and partly to large areas of both islands being underlain by serpentinite, on which peat never formed. As a result, inland exposure is better in these two islands than in much of the rest of Shetland, and exposure is nearly complete along the coastline.

Previous work

Geology

More accounts of the geology of Unst have been published than for any other part of Shetland; more than 120 between 1798 and 1993. However, much less attention has been paid to Fetlar. The interest in Unst has always been due to the presence of the ophiolitic complex, with its unusual lithology and mineralogy, and its apparent economic prospects.

Robert Jameson (1798) was the first geologist to visit Unst and Fetlar. He provided a clear account in a few words of the sequence of lithologies met during a coastal traverse of the two islands. Hibbert devoted seven and a half months in 1817 and 1818 to a survey of Shetland and presented a detailed account, with a quarter-inch-to-the-mile geological map drawn on an archaic base (Hibbert, 1819, 1822). He divided Unst into the same series of structural units used later by Read (1934a), listing the lithological components in some detail. He recognised the western (Valla Field Block) gneisses, the phyllites of the north (Saxa Vord Block) and the central valley, the gneissose granite (Skaw Granite), and the serpentinite and metagabbro. He correctly correlated the Fetlar rocks with those of Unst and recognised the metamorphic and conglomeratic nature of the Funzie Conglomerate. His was probably the very first detailed account, anywhere, of a complex metamorphic area. Even more innovative was his systematic account of the dip and strike of the rocks, area by area, throughout Shetland, with the presentation of a strike map of Shetland (the first ever structural map). He also presented the first ever description of rock fabrics and lineations, the type area for the latter being The Cliffs, Norwick [HP 655 150]. He also discovered the chromite.

In the following hundred years many geologists visited Unst and Fetlar, but only Peach and Horne (1879), as part of a glaciological survey of Shetland, made any significant changes to the geological map. They transcribed Hibbert’s geological map onto a modern base. In doing so, they made some amendments, in particular converting some of Hibbert’s boundaries to major transcurrent faults (the Nesting Fault and the Walls Boundary Fault). Major faults were not known in Hibbert’s time. However, some of their amendments were incorrect. This transcribed map remained the only available geological map of Shetland until the publication of the quarter-inch-to-one-mile map of the Institute of Geological Sciences (IGS, 1962), which followed the earlier systematic survey of Shetland by the Geological Survey of Great Britain.

Heddle (1878) carried out a mineralogical survey of Shetland based on visits in 1846, 1848, 1856, and 1873. He found, described and analysed numerous minerals, many for the first time in Shetland, and some for the first time ever. Detailed, recognisably ‘modern’, accounts of the geology of Unst and Fetlar started with unpublished work by Anon (1926–1927) and a petrographical study of the serpentinite (Phillips, 1927).

Geophysics

J-B Biot, the French physicist, visited Unst in 1818 to determine the acceleration due to gravity by observing the length of the seconds pendulum (Flinn, 1989). He did so in conjunction with the Trigonometric Survey (later to become the Ordnance Survey of Great Britain), and as part of a series of observations, extending from the Mediterranean, to establish the oblateness of the geospheroid. Kater (1819) visited Unst the year after Biot had done, and repeated and confirmed his measurements. One hundred and forty years later, McQuillin and Brooks (1967) carried out ground magnetic and gravity surveys of Shetland, including Unst and Fetlar, for the Institute of Geological Sciences, and an aeromagnetic map on a scale of 1:250 000 appeared soon after (IGS, 1968b).

Glaciation

The first paper to be concerned with the glaciation of Shetland was by Hibbert (1831) on erratics; he attributed them to the ‘diluvial wave’. C W Peach (1864) interpreted glacial striae at Hagdale in eastern Unst as indicating eastward- flowing ice, but was later persuaded, incorrectly, by his son, B N Peach, to reverse this direction. Under the influence of Croll’s (1870) hypothesis that ice from Scotland and Scandinavia had filled the North Sea and overflowed to the west over Orkney and Shetland, Peach (B N) and Horne (1879) carried out a glacial survey of Shetland. They concluded, incorrectly, that Croll was right, and that the islands had been glaciated by an ice sheet flowing over Shetland from the east. Robertson (1935), however, when surveying central Shetland for the Geological Survey of Great Britain, correctly observed that the distribution of granite erratics indicated ice flow to the east on the east side of Shetland. Recognition of the existence of a Late Devensian ice sheet based on Shetland followed from a series of papers by Flinn (1964, 1967a, 1977, 1983, 1992a and b, 1994b and 1995a).

Recent surveys

In 1929 and 1930, H H Read and J Phemister surveyed Unst and Fetlar respectively as part of the systematic survey of Shetland by the Geological Survey of Great Britain (GSGB) between 1929 and 1933. Ritchie (1930) and Wilson (1931) summarised their results. Read (1934a) published an account of the geology of Unst with a map and later (Read, 1937) a classic account of the polymetamorphic history of the schists and gneisses of the Valla Field and Saxa Vord blocks. Following completion of the GSGB field survey of Shetland, G V Wilson worked on a one-inch-to-the-mile survey of the area until his death in 1941. Further work was curtailed by the Second World War, but a quarter inch-to-the-mile map was eventually published (IGS, 1962), and the one-inch-to-the-mile map of Northern Shetland, which covers the district, was published subsequently (IGS, 1968a). Read’s (1934a) map of Unst was copied exactly for the quarter inch and one inch maps and all the lines are repeated exactly on the current map (BGS, 2002). When the present survey commenced in 1949, no reason was found to vary the lines, or to query Read’s published interpretation of their nature, and the 1:10 000 survey 40 years later confirmed their positions. The present map differs only in presenting a subdivision of Read’s mapped units.

The geology of the district was transformed by the recognition of the ophiolitic nature of the serpentinite complex, published independently in 1973 by Flinn (personal communication in Garson and Plant, 1973) and Williams and Smyth (1973). Since 1980, HM Prichard and associates, especially from the Open University, have published work on the geochemistry, and especially that of the trace elements, of the ophiolite. Since 1970, the British Geological Survey has produced a series of geophysical and geochemical reports of surveys prospecting for economic minerals in Unst.

Starting in 1949, Flinn has worked in the district during several periods, culminating in the 1:10 000-scale survey, between 1983 and 1994, under contract to the Natural Environment Research Council. This survey led to the published 1:50 000-scale map of Unst and Fetlar (BGS, 2002) and the present memoir. The survey was based on 1:10 000-scale aerial photographs and maps, and included ground magnetic surveys of all serpentinite areas and Saxa Vord, and a gravity survey of the land area. A total of 105 weeks was spent in the field, between 1983 and 1994, and subsequently, a further 10 weeks was spent on a gravity survey and checking geological observations. Students from the University of Liverpool provided some assistance in the surveys. The survey was clean copied onto a 1:10 000-scale base, and is available in the office of the British Geological Survey, Edinburgh. (Figure 1) provides a key to the 1:10 000-scale clean-copied sheets and to the 1:10 560-scale clean-copied sheets produced by Read and Phemister in the 1930s. About 4000 hand specimens were collected, mostly from around the coastline where the best exposure occurs, but samples were obtained from most parts, and all were thin sectioned and are held in the office of the British Geological Survey, Edinburgh. Many hundreds of the samples were obtained during previous surveys, especially between 1949 and 1956 and during surveys by PhD students.

Descriptions of the petrography, mineralogy, geochemistry and other features of rock samples appear throughout this memoir. Whole-rock analyses include many by research students and technical staff at the Department of Geology of the University of Liverpool, some contained in the British Geological Survey database, including some in Guppy and Sabine (1956), and others obtained from published papers. Each set of analyses was examined to ensure it was symmetrically distributed, monomodal and thus could be validly averaged. Where sets were compared, they were further checked for no statistical difference. Electron microprobe analyses of minerals from polished thin sections were obtained by Flinn, Aziz and Moffat at the Department of Geology of the University of Manchester, using the Geoscan energy dispersive microprobe. About 2000 colour transparencies taken during the survey are held in the National Archive of Geological Photographs (NAGP) and are identified in the text by their BGS P numbers.

Chapter 2 Outline of the geology of the district

The geology of Unst and Fetlar, together with that of the neighbouring island of Yell (Sheet 130, BGS 1993) to the west, is composed of a series of tectonic units of major, even national, importance, dominated by the Shetland Ophiolite-complex. Yell is mostly formed of Moinean rocks (Flinn, 1994a); the Yell Sound Division. These are typical Glenfinnan/Loch Eil-type rocks, together with hornblende gneiss-type Lewisian inliers. Unst and Fetlar are mostly formed of Dalradian rocks<span data-type="footnote">At the time of publication of 1:50 000 Sheet 131 (BGS, 2002), the lithostratigraphy of the Dalradian rocks of Unst and Fetlar had not been formalised. Consequently the informal nomenclature has been retained in this memoir.</span> and the Shetland Ophiolite-complex.

Basement to the Shetland Ophiolite-Complex

The Moinian rocks are divided from the Dalradian rocks of Shetland by the Boundary Zone, a tectonic diapir or extrusion wedge (Flinn, 1988) (Figure 2). The western edge of the Boundary Zone is marked by a thin band of microcline augen gneiss called the Valayre Gneiss (P547371), which runs down the east side of Yell and continues south through Shetland for 70 km. The eastern edge of the Boundary Zone closely follows the west coasts of Unst and Fetlar, and in Shetland to the south is clearly marked by a narrow band of microcline augen gneiss called the Skella Dale Gneiss (P602813). In Unst and Fetlar, there is no augen gneiss marker band to indicate the eastern boundary of the Boundary Zone and it had to be based on correlation with the clearly marked boundary in the Mainland of Shetland to the south. Thus, the Boundary Zone lies half in Yell and half in Unst.

The Boundary Zone is composed of a series of lenticular masses of a variety of rocks, some of which are otherwise alien to Shetland. In the north-east of Yell are hornblende gneisses of Lewisian inlier type, partially blastomylonitised by a late high strain zone dated at 500 Ma (Roddam et al., 1994a) and named the Hascosay Slide Zone. Other distinguishable lenticular masses in the Boundary Zone in Yell which do not occur elsewhere in Shetland are metavolcanic rocks characterised by spherical dolerite bodies, a mass of psammite containing ultramafic drop stones, and south-east of Yell, a lens of nebulite (Flinn, 1994a).

In Unst to the east, the equivalent of the Hascosay Side Zone is the Westing Ultramafic Zone<span data-type="footnote">As a result of the limitations imposed by separate mapping contracts, the western half of the Boundary Zone in Yell was surveyed before the eastern half in Unst. A high strain zone along the coast of Yell and within the Boundary Zone was misinterpreted as a slide and named the Hascosay Slide Zone. Later, the survey of Unst revealed the error and the high strain zone was renamed the Westing Ultramafic Zone. However, although it was too late to change the name on the 1:50 000 map of Yell (BGS, 1993) and memoir (Flinn, 1994a), the term was considered inappropriate for Unst. The Hascosay Slide Zone and the Westing Ultramafic Zone are equivalent terms and are a component of the Boundary Zone. </span>. Other gneisses are of no known origin or relationship; these together with a metalimestone have been assembled as the Westing Group. The occurrence of the blastomylonite high-strain zone as a 500 Ma event (Flinn, 1994a) within the Boundary Zone, together with a profuse development of late pegmatites unrepresented outside the Boundary Zone, leads to the conclusion that most of the contents of the Boundary Zone are probably at least as old as Neoproterozoic.

The Westing Group in Unst is composed of the following series of tectonostratigraphical units listed in approximate order from west to east. The Orknagable Formation is composed of slightly gneissose psammites and hornblendic lenses. The North Holm Gneiss is an intensely folded muscovite-biotite leucosome gneiss. The Kirkaby Gneiss is a biotitic psammitic gneiss of variable character. The Westing Ultramafic Zone = Hascosay Slide Zone is a highly laminated hornblende schist due to blastomylonitisation and containing lenses of steatitised serpentinite. The Westing Gneiss is dominantly formed of biotitic psammite, but varying to kyanite-staurolite bearing pelite, all moderately gneissose. The Westing Limestone is a zone up to 200 m wide of interbanded marble, calcsilicate, and hornblendic and psammitic rocks. The Valla Field Gneiss is a kyanite-staurolite-garnet psammitic to pelitic gneiss<span data-type="footnote">In draughting the map (BGS, 2002), the eastern boundary of the Boundary Zone was chosen to lie along the western boundary of the Valla Field Gneiss. By the time this memoir was being written, the eastern boundary had been revised to lie along the eastern boundary of the Valla Field Gneiss.</span>.

The Valla Field, Saxa Vord and Lamb Hoga blocks<span data-type="footnote">The term ‘Block’ as used here, and throughout this memoir, follows the original terminology of Read (1934a).</span>

For descriptive purposes, Read (1934a) divided Unst into a series of tectonic units or blocks of different lithology, structure, and metamorphic history. For the purposes of this memoir, the most useful of these are the Valla Field Block and the Saxa Vord Block, together with the Lamb Hoga Block in Fetlar, to the south of the Lamb Hoga Fault (Figure 3).

The Valla Field Block of Unst contains on its west side the Boundary Zone as described above. The block is bounded to the east by the ophiolite-complex. Between the Boundary Zone and the ophiolite-complex, Read (1934a) recognised the Valla Field Group and the Burra Firth Group (now the Burra Firth Formation) dipping, and which have been interpreted to face, east. All three units also occur to the south of the Lamb Hoga Fault in the Lamb Hoga Block. Read’s Valla Field Group has now been divided into two: the Valla Field Gneiss and the Valla Field Schist (Figure 3) (but see footnote). This division was outlined by Read (1934a, p.642) but not carried out on his map. The Valla Field Schist is a coarsely crystalline, Al-rich muscovite-kyanite-staurolite- garnet-rich schist containing chloritoid locally. The Burra Firth Formation is composed dominantly of quartzitic psammites interbanded with pelite and psammitic schists.

These stratigraphical units have been correlated both lithologically and metamorphically with the Scatsta ‘Division’<span data-type="footnote">Now formalised as the Scatsta Group (e.g. Flinn, 2007).</span>, the regionally metamorphosed basal unit of the Dalradian of Mainland Shetland. The Scatsta ‘Division’, like the Valla Field Schist and the Burra Firth Formation, occurs adjacent to the eastern boundary of the Boundary Zone, and in particular where that boundary is marked by the Skella Dale Gneiss. This correlation indicates that the eastern boundary of the Boundary Zone lies between the Valla Field Gneiss and the Valla Field Schist (in the Mainland between the Scatsta Quartzitic Group and the Scatsta Pelitic Group <span data-type="footnote">In draughting the map (BGS, 2002), the eastern boundary of the Boundary Zone was chosen to lie along the western boundary of the Valla Field Gneiss. By the time this memoir was being written, the eastern boundary had been revised to lie along the eastern boundary of the Valla Field Gneiss.(IGS, 1981)</span>). In the Lamb Hoga Block in Fetlar, the Valla Field Schist and the Burra Firth Formation dip to the south-west and are inverted relative to their situation in Unst. Together with the equivalent rocks in the Valla Field Block they form the lower and upper limbs of a large scale recumbent east-facing fold, named the Valla Field Fold (Figure 20)b. The fold is possibly due to impact from the obducting Shetland Ophiolite-complex. The stratigraphically inverted attitude of the lower limb of the fold in Fetlar is a continuation from the south of the inverted recumbent limb of the mega-monoclinal fold involving the whole Dalradian succession in Shetland (Flinn, 2007). The regional metamorphism in Unst and Fetlar was shown by Read (1934a) to have been of medium to high grade and to have been followed by two lower grade local episodes. This regional metamorphism is shown by correlation with the lithology and metamorphism of the Scatsta ‘Division’ to be the first metamorphism of the Dalradian of Shetland. Read’s ‘second metamorphism’ (Chapter 5), involving the crystallisation of chloritoid, was probably associated with the emplacement of the ophiolite-complex and the folding of the Valla Field Fold; his ‘third metamorphism’ certainly was.

The Saxa Vord Block composed of Dalradian rocks, lies to the east of the Valla Field Block, separated from it by the Burra Firth Lineament, replacing a major part of the Dalradian succession. It is possibly a tectonic unit associated with the ophiolite-complex. The Dalradian rocks forming the Saxa Vord Block, the Saxa Vord Group, have been correlated both lithologically and metamorphically with the Clift Hills ‘Division’<span data-type="footnote">Now formalised as the Clift Hills Group (e.g. Flinn, 2007).</span> in the Mainland, the youngest division in the Shetland Dalradian. The Saxa Vord Group is composed of the Queyhouse Flags, the Saxa Vord Pelite and the Hevda Phyllite. The Queyhouse Flags are composed of bedded quartzites alternating with phyllitic semipelites. The Saxa Vord Pelite comprises fine-grained phyllitic chloritoid- bearing Al-rich pelites. The stratigraphically uppermost Hevda Phyllite is composed of garnet-bearing phyllitic pelite to semipelite. After their regional metamorphism, they show evidence of thermal metamorphism at 425 Ma by the Skaw Granite (Flinn and Oglethorpe, 2005) (Figure 2). Their regional metamorphism, like that of the Valla Field Group, was the primary metamorphism of the Dalradian of Shetland.

The Shetland Ophiolite-Complex and Skaw Granite

The Shetland Ophiolite-complex (Flinn, 1999; Flinn and Oglethorpe, 2005) is formed of two nappes of ophiolitic rocks: the Upper Nappe and the Lower Nappe. They lie one above the other, are separated by the Middle Imbricate Zone, and are underlain by the Lower Imbricate Zone, which rests on the Dalradian basement (Figure 4) and (Figure 5). The imbricate zones are composed of tectonic slices of the metamorphic basement, ophiolitic rocks, hornblendic rocks, and volcanic and siliciclastic sedimentary rocks that were eroded, deposited and metamorphosed after the ophiolite obduction. The ophiolite-complex is truncated to the north by the Skaw Granite, which cuts across and thermally metamorphosed the basement and the Lower Imbricate Zone. To the south, the Lamb Hoga Fault truncates the ophiolite-complex and downthrows it against the Lamb Hoga Block. To the east the complex passes under the sea (Figure 6).

The Lower Imbricate Zone occurs structurally beneath the Lower Nappe and is exposed along its western edge in Unst. It contains tectonic slices of the Norwick Hornblendic Schist (see below), which are closely associated with the base of the Lower Nappe. Also present are tectonic slices of the Unst Phyllite Group (see below), and slices of basement rocks from the adjacent Saxa Vord and Valla Field blocks.

The Lower Nappe forms most of the eastern part of Unst, and extends southwards to underlie most of Fetlar (Figure 6). The nappe is formed of an ophiolite sequence comprising three steeply inclined, north-east- trending layers, each of which is several kilometres thick. The formerly deepest layer, which now crops out along the north-west part of the exposed ophiolite sequence, consists of serpentinised harzburgite. This Metaharzburgite Layer represents the top of the mantle. It is rhythmically banded along its upper edge and includes bodies of serpentinised dunite (metadunite), some of which contain ‘podiform’ chromite bodies. East of the Metaharzburgite Layer, the next layer in the ophiolite succession is serpentinised dunite (the Metadunite Layer), representing the basal part of the crust. The Metadunite Layer is several kilometres thick and contains ‘podiform’ chromite bodies and metre- to kilometre-scale xenoliths of interbanded clinopyroxenite and wehrlite. The Metadunite Layer is succeeded to the east by a layer several kilometres thick of altered gabbro (the Metagabbro Layer). The lower and upper parts of this layer are lithologically distinct. The Lower Metagabbro contains xenoliths of interbanded clinopyroxenite and wehrlite similar to those in the Metadunite Layer. The Upper Metagabbro is cut by swarms of subparallel metamorphosed basic dykes (the Quasi-sheeted Dyke-swarm), which trend approximately north-east in Unst and west in Fetlar.

The ophiolite succession in the Lower Nappe differs from the universally accepted lithostratigraphical model (e.g. Gass, 1980). The ultramafic layers are interpreted to have been pervasively serpentinised before obduction. The Metadunite Layer is unusually thick and, together with the Metagabbro Layer, lacks a cumulate layer. Intrusions of the Quasi-sheeted Dyke-swarm are more nearly parallel to the base of the Metagabbro Layer than normal to it. Furthermore, the nappe lies on its side with steeply inclined layers.

The Middle Imbricate Zone overlies the Lower Nappe and rests on the eroded surfaces of the steeply inclined layers of ultrabasic and basic rock. It is exposed in the Norwick [HP 652 145] and Muness [HP 630 010] regions (in the north-east and south-east of Unst, respectively), and in west and east Fetlar. It is formed of tectonic slices of the ophiolite, metasedimentary rocks containing ophiolite debris, metavolcanic rocks, metavolcaniclastic rocks and Dalradian basement (the Strand Gneiss). Other than the basement slices, these rocks are unknown outside the ophiolite-complex. The most voluminous constituent of the Middle Imbricate Zone is the Unst Phyllite Group. This includes chloritoid-bearing phyllites (the Leagarth Pelite), chloritoid-free pelitic phyllites (the Norwick Phyllite), the Norwick Graphitic Schist, phyllitic metavolcaniclastic rocks (the Gruting Greenschist), and semipelitic phyllites (the Muness Phyllite). The Muness Phyllite contains metaconglomeratic bands that are characterised by ophiolite-derived pebbles of either ‘plagiogranite’ or metagabbro. The Funzie Conglomerate, which has an outcrop of about 5 km2 in eastern Fetlar, is a deformed conglomerate with clasts dominantly of quartzite. Tectonic slices of hornblende schists (Norwick Hornblendic Schist) are closely associated with the base of the overlying Upper Nappe.

The Upper Nappe is composed of metaharzburgite (the Metaharzburgite Layer), which is generally indistinguishable from that in the Lower Nappe. It occurs throughout the outcrop of the ophiolite- complex as klippen overlying rocks of the Middle Imbricate Zone and Lower Nappe. Five klippen occur in Unst, of which the largest forms the Hill of Clibberswick [HP 658 130]. The Vord Hill Klippe [HU 622 935] in the centre of Fetlar is the largest in the ophiolite-complex. Half a dozen much smaller klippen occur elsewhere in Fetlar. The island of Sound Gruney [HU 580 962] and the serpentinite block at Muness [HP 635 010] are composed of wehrlite and clinopyroxenite and are probably tectonic slices of the Lower Nappe contained in the Middle Imbricate Zone.

The Skaw Granite is a K-feldspar-phyric biotite monzogranite that has been deformed into L-tectonite augen gneiss. Microcline phenocrysts have a strong linear preferred orientation parallel to a linear matrix fabric plunging to the north-east. The fabric indicates either forcible emplacement when partly crystallised or, more likely, deformation after emplacement but before final cooling. The granite contains cognate xenoliths of microdiorite, and other xenoliths of mica schist, quartzite and basic volcanic greenstones. A swarm of lamprophyre dykes cuts the intrusion. At its western contact, the Skaw Granite intrudes and thermally metamorphoses rocks of the Saxa Vord Group and the Lower Imbricate Zone (Figure 2). Fibrolite and chloritoid are developed profusely in the contact zone within the Saxa Vord Pelite.

Emplacement

The Shetland Ophiolite-complex was emplaced in at least two phases. Strongly serpentinised, infertile mantle and overlying serpentinised and otherwise hydrated oceanic crust, is interpreted to have been obducted westwards onto the metamorphic Dalradian basement. Lizardite is the principal serpentine-group mineral to have formed in the hydrated mantle rock. However, in a zone several hundred metres thick immediately above the obduction thrust the lizardite has recrystallised to antigorite, probably during obduction. At the same time, basic magma was injected into the thrust where it recrystallised as hornblende schist. This rock now crops out in the Lower Imbricate Zone and Middle Imbricate Zone as tectonic slices of Norwick Hornblendic Schist. Following obduction, the upper surface of the Lower Nappe was eroded and sediments, some of which contain ophiolite debris (Unst Phyllite Group), were deposited upon it. ⁴⁰Ar-³⁹Ar ages for hornblende crystals from the Norwick Hornblendic Schist suggest that obduction took place at about 500 Ma (Flinn and Oglethorpe, 2005).

Later, during the Scandian Event (approximately 425 Ma), part of the obducted slab was thrust from the east over the Lower Nappe and its overlying sediments to form the Upper Nappe and the Middle Imbricate Zone. The Upper Nappe was transported on the old obduction thrust. It carried beneath it slices sheared off the Lower Nappe, the Middle Imbricate Zone and slices of the Norwick Hornblendic Schist. Emplacement of the Upper Nappe was accompanied by open folding of the nappe pile on north-north-east axes, and by greenschist- facies metamorphism, especially of the Unst Phyllite Group. The folding was accompanied by extension (due to constriction) of the Middle Imbricate Zone rocks in the fold core parallel to the fold axis, and parallel to elongation of conglomerate pebbles, and development of L-tectonite fabrics. The entire nappe pile was driven farther onto the basement during emplacement of the Upper Nappe, causing the Lower Imbricate Zone (composed of tectonic slices of rocks from the Middle Imbricate Zone and from the metamorphic Dalradian basement) to form beneath the Lower Nappe.

The Skaw Granite was intruded across the Lower Imbricate Zone at about this time. It caused contact metamorphism of the Lower Imbricate Zone and diverted the thrust at the base of the zone (the Hevda Thrust) farther west.

The form of the Lower Nappe, with its steeply inclined layers indicating it is a slice cut steeply through the ocean floor from above, means that it is unlikely that it ever extended much farther west than at present.

Origin

It is suggested that the Shetland Ophiolite originated as the oceanic floor of an intracontinental extensional basin in Laurentia, close to the western side of Iapetus. Remnants of such a basin and its volcanic cover occur in the south of Shetland in Dunrossness (Flinn and Oglethorpe, 2005). The remains of the basin can be traced northwards from Dunrossness, under the sea for 120 km on the east side of Shetland, by positive gravity and aeromagnetic anomalies. This remnant is what remains after final closure of the gap between Baltica and Laurentia during the Silurian Period.

The earlier westward obduction of the Shetland Ophiolite at about 500 Ma was possibly coeval with the eastward obduction of ophiolites in western Norway that lie east of Shetland (Flinn and Oglethorpe, 2005). A second phase of obduction of Norwegian ophiolites during the Scandian Event, and the final closure of Iapetus, probably coincided with emplacement of the Upper Nappe of the Shetland Ophiolite-complex (Flinn and Oglethorpe, 2005).

Chapter 3 Metasedimentary and metavolcanic rocks of the Valla Field, Lamb Hoga and Saxa Vord blocks

Introduction

The Valla Field Block comprises an eastward-dipping, tectonostratigraphical continuous sequence of schists, gneisses and quartzites, part of the basement of metamorphosed Dalradian rocks underlying the Shetland Ophiolite-complex. The Valla Field Block forms a cuesta along the west side of Unst with the ophiolite-complex occupying the eastern part of the island. The Lamb Hoga Block comprises similar, but westerly-dipping, basement lithologies in the western part of Fetlar. The Lamb Hoga Fault separates the Valla Field and Lamb Hoga blocks, juxtaposing rocks of the Shetland Ophiolite-complex and the inverted sequence of basement rocks of the Lamb Hoga Block (see diagrammatic section on (Figure 7)

Read (1934a) originally divided the rocks of the Valla Field Block of Unst into three, easily distinguishable, stratigraphical units which he named, in ascending order, as the Westing, Valla Field and Burra Firth groups. The IGS (1968a) one-inch-to-the-mile map of the region showed the Lamb Hoga Block on Fetlar as formed from similar rocks to those of the Burra Firth and Valla Field groups of Read (1934a). Aziz (1984) further subdivided the succession in the Valla Field Block and showed that Read’s Burra Firth and Valla Field groups did indeed occur in the Lamb Hoga Block, albeit upside-down and dipping to the west. He also argued that the Westing Group was tectonically distinct from the other two (Aziz, 1984). The present survey (BGS, 2002) has shown that the Westing Group belongs to the tectonic Boundary Zone of Yell (Flinn, 1988) and continues to the south-west of the Lamb Hoga Fault, to the islands between Lamb Hoga and Yell, and into Yell. The Boundary Zone separates the Moinian rocks of Yell from the Dalradian rocks of Unst (Flinn, 1994a). In this memoir, the succession in the Valla Field and Lamb Hoga blocks is divided into the Valla Field Group and the tectonostratigraphical Boundary Zone, including the Westing Group. The published map presents the Westing Group as a tectonostratigraphical series, divided into six mappable formations: the Orknagable Formation; the North Holm Gneiss; the Kirkaby Gneiss; the Westing Ultramafic Zone; the Westing Gneiss; and the Westing Limestone (but see further below). The Westing Ultramafic Zone is a high strain zone, which was recognised in Yell as the Hascosay Slide Zone (Flinn, 1994a). The tectonostratigraphical affinity of the Westing Group is uncertain (see below). The Valla Field Group is considered to be Dalradian and is represented on the published map as comprising three formations; the Valla Field Gneiss, the Valla Field Schist and the Burra Firth Formation. This sequence (Figure 7) closely corresponds to that of Read (1934a). However, a recent reassessment of the relationships between the Westing and Valla Field groups now places the Valla Field Gneiss within the Westing Group and the Boundary Zone (Flinn, 2007; see footnote 3 in Chapter 2).

It has also become apparent during this survey that the Valla Field Block is part of the upper limb of a major recumbent fold, closing to the east and opposed to the ophiolite-complex. The Lamb Hoga Block south of the Lamb Hoga Fault is part of the lower limb of that fold, which before faulting lay several kilometres deeper than the Valla Field Block (Figure 7).

Nomenclature of the gneissose rocks

The gneisses of Valla Field were distinguished and described by Read (1934a, 1937) as types he had previously distinguished in Cromar (Read, 1927) and central Sutherland (Read, 1931). Flinn (1967b) recognised and described these gneisses on the Shetland Mainland and subsequently on Yell (Flinn 1994a), but he replaced Read’s nomenclature with non-genetic names and proposed different origins. In describing these gneissose rocks, Flinn highlighted the lack of a universally accepted definition of the term gneiss. He therefore used gneiss as a field term, without any consideration of definitions, to label rocks, and to distinguish and/or relate bodies of rock, in which feldspar played a dominant role and where the rock had a tectonite fabric. A similar approach has been used here in describing gneisses in the district. These include three main types of gneiss: plagioclase microporphyroblast gneiss, leucosome gneiss, and homogenous gneiss (diatexite) together with their partial development (semigneisses). In addition, there are augen gneisses, and agmatite gneisses, which occur locally.

All three of the main types of gneiss occur abundantly in the district. However, the style of occurrence in Yell (Flinn, 1994a, e.g. fig. 15), where the gneisses are interpreted as having been derived from psammites and schists by recrystallisation, differs markedly from that in this district. Here, the Westing Group is formed of units distinguishable not by simple changes of lithology but by changes in the relative proportions of a set of different lithologies, i.e. pelite, psammite and quartzite etc., occurring in rapidly alternating layers. The gneisses and semigneisses occur as bands alternating with their ungneissified protoliths and with unrelated rocks. (Figure 8) represents their occurrence.

Plagioclase microporphyroblast gneiss

The plagioclase microporphyroblast gneisses are dark coloured psammitic granofelsic rocks rich in rounded equi- dimensional porphyroblasts of oligoclase generally about 2 mm in diameter, but varying from 1–4 mm in diameter (Flinn, 1994a, p.50–52; (Plate 1a); (P534200), (P534858)). These porphyroblasts occur in a uniform granofelsic biotite-quartz-plagioclase matrix, similar to that of the protolith and with a similar grain size of about 0.3 mm. The porphyroblasts comprise up to 15 per cent of the rock by volume and may locally be in contact with each other (Flinn, 1994a, plate 5d). Modal analyses of 33 gneisses and semigneisses have revealed no significant correlation between total feldspar volume and porphyroblast volume (Figure 9). Available chemical analyses (of six specimens) show little variation in composition despite different porphyroblast contents. There is no reason to suppose the gneissification of these rocks involves new growth of feldspar, as opposed to enhanced growth. Read recognised and named these gneisses ‘felspar-blebbed gneisses’.

The plagioclase microporphyroblast gneisses are characteristic of, but often patchily developed in, the Westing and Kirkaby gneisses (Flinn, 1994a), and elsewhere in the Boundary Zone as far south as Lunna Ness [HU 523 743]. They often overlap the adjacent leucosome gneisses of the Valla Field Gneiss locally by up to 100–200 m. These overlapping rocks are quartz-plagioclase-muscovite > biotite micaceous granofels. They are speckled white with plagioclase grains, about 2 mm in size, in a groundmass 1 mm in size. Although the plagioclase grains are less prominent and less equidimensional, and they are lighter in colour, they resemble poorly developed plagioclase microporphyroblast gneisses. They occur on the edges of microporphyroblast gneiss areas as an extension of that gneiss into areas of muscovite > biotite granofels, and leucosome gneisses (Figure 8). The plagioclase microporphyroblast gneisses have been interpreted by Flinn (1995b) as having developed by the enhanced growth of plagioclase microporphyroblasts in muscovite-poor psammitic rocks. The plagioclase porphyroblast gneiss is associated within the Boundary Zone in Lunna Ness with areas of nebulite (P547276).

Leucosome gneiss

The leucosome gneiss (Flinn, 1994a, p.55; (Plate 1b) is quartzofeldspathic, banded, leucosome-streaked and micaceous (migmatitic), and contains augen. The leucosome gneisses have been interpreted by Flinn (1995b) to have developed completely, or partially, from recrystallization of pelites and mica-rich psammites by the growth of quartzofeldspathic leucosomes. Leucosome gneisses are most strongly and uniformly developed in the Westing Group and especially in the Valla Field Gneiss. They are also locally and patchily and weakly developed in the Valla Field Schist; as these schists are quartz-rich and plagioclase-poor, many leucosomes are pure quartz. By contrast, the only gneissic unit in the Burra Firth Formation, which otherwise comprises mainly quartzites and impure quartzites with relatively little psammite or schist, is the Herma Ness Gneiss.

Homogenous gneiss

Homogenous gneiss (diatexite) is coarse and granoblastic (Flinn, 1994a, p.44 and p.55, and 1995b; (Plate 1c); (P541056), (P541134)) and occurs only very sparsely, for example in the Valla Field Gneiss where the micaceous psammitic rocks have been interpreted by Flinn (1995b) to recrystallise to granoblastic gneisses.

Semigneisses

Semigneisses (cf. Flinn, 1994a, p.44; (P541351), (P541860)) are partially developed gneisses that contain features of both the metasedimentary protolith and the recrystallised gneisses, and which vary from granular, almost granitoidal and almost homogenous rocks, to variously quartzofeldspathic-banded, streaked and augened rocks.

Augen gneiss

Augen gneisses (Figure 8); (P533950), (P540888)) are considered here as those gneisses with feldspar grains one centimetre or more in maximum dimension. In the district, such gneisses occur mainly within the Boundary Zone (Chapter 2) and are mostly microcline augen gneisses, though microcline is otherwise absent from the Boundary Zone (other than in pegmatites and aplites). Occurrences include those in the Westing Gneiss and the Herma Ness Ultramafic Zone.

Valla Field and Lamb Hoga Blocks

Westing Group

Orknagable Formation

The rocks forming the headland area at Orknagable [HP 574 135] and the two islands, Round Holm [HP 563 060] and Lang Holm [HP 563 065] are assigned to the Orknagable Formation. The formation comprises psammite and amphibolite, with minor bands of impure quartzite and pelite. The amphibolite occurs as large lenticular flat-lying masses interlayered with the homogeneous medium-grained dark-coloured (biotite-rich) psammite; the psammite is typically non-gneissose, but some amphibolites are agmatitic (Plate 1d). There is some minor gneissose texture locally developed in the psammites at Orknagable. Some amphibolites are more globular than lenticular, while on Round Holm and Lang Holm, the amphibolites form flat-lying lenses several metres thick, with massive cores and schistose garnetiferous envelopes. The psammite layers enclosing the amphibolites are generally flat lying and distorted into whale-backed open folds by the amphibolite bodies (P534057). The Orknagable Formation can be distinguished from the Kirkaby Gneiss since the latter contains a higher proportion of amphibolite.

The amphibolites typically comprise olive-green hornblende, reddish-brown biotite, and plagioclase with grain sizes varying from 0.5 to 1.5 mm. The psammites are mostly plagioclase rich, biotite > muscovite-garnet- quartz rocks of similar grain size and colour to the amphibolites. The psammites are cut by aplopegmatite veins, which are up to a metre or two thick.

North Holm Gneiss

The North Holm Gneiss is limited to the small islands of the North Holms [HP 569 115] and the headland to the north at [HP 573 124] where the formation occurs in contact with the Westing Ultramafic Zone, and extends south into the North Holms [HP 569 115]. It is very different lithologically and structurally from any other rocks in the Westing Group. It is a moderately strongly gneissose, granoblastic, muscovite-rich, biotite garnet leucocratic psammite (P533534). Several bands of amphibolite, one metre thick, occur on North Holms. The formation is intensely folded on axes plunging 10–20º eastwards; fold amplitudes vary from several centimetres up to cliff- sized (here c.40 m high). All the folds are overturned to the south; the smallest folds appear disharmonic. This folding, its easterly trend, the richness in muscovite and quartz and the uniform gneissose texture render the formation different from all other formations of the Westing Group. The rocks are also unusually intensely injected by subconcordant granite to pegmatitic granite veins, up to a metre thick, and composed of quartz and oligoclase, with minor biotite and accessory garnet (P533532). The veins show signs of high strain.

Kirkaby Gneiss

In Unst, the rocks west of the Westing Ultramafic Zone occur in headland areas, so their relationships to one another are largely obscured by the sea. The Kirkaby Gneiss is exposed only in a small area between Houllnan Ness [HP 565 055] and Spoo Ness [HP 565 072]. It is dominantly composed of garnetiferous, biotite > muscovite psammite, which is locally gneissose. In places the psammitic gneiss grades into micaceous psammite and even into pelitic varieties. A large mass of pelitic gneiss occurs on the south side of Wick of Collaster [HP 573 073], while plagioclase microporphyroblastic gneiss forms the tip of Houllnan Ness [HP 565 054] (P534058).

Westing Ultramafic Zone/Hascosay Slide Zone

The present survey of Unst has shown that the Westing Ultramafic Zone and the rocks on either side of it (the Westing Group) are sufficiently similar to the Hascosay Slide Zone and its adjacent rocks (including the Boundary Zone to the west of the Hascosay Slide Zone (Flinn, 1994a)) for them to be equated. The name Westing Ultramafic Zone is applied to the hornblende- rich blastomylonitic high strain zone recognised on Unst. The name Hascosay Slide Zone was given to the hornblende-rich blastomylonitic high strain zone on Hascosay and continuing north on Yell to Migga Ness [HP 540 052]. No evidence was later found in either Yell or Unst for offset or translation across the structure to justify the name slide. The Westing Ultramafic Zone and the Hascosay Slide Zone therefore comprise a single ultramafic high strain zone in the core of the Boundary Zone (BGS, 2002; Chapter 2).

The correlation of the Hascosay Slide Zone with the Westing Ultramafic Zone leads to the association of the Westing Group of Unst/Fetlar with the Boundary Zone of Yell (Flinn, 1994a). The Westing Group has the field appearance of a conformable stratigraphical succession and has been represented as such on the published map and in the lithostratigraphical descriptions which follow. In reality this succession may comprise a set of unrelated lenticular units with wholly tectonic boundaries. The Boundary Zone was defined (Flinn, 1994a) as a tectonic unit forming the eastern limit of the Moinian Yell Sound ‘Division’, and separates Dalradian and Moine rocks in Unst, Fetlar and Yell. The terms Westing Group and Boundary Zone are respectively the lithostratigraphical and tectonic names for the same group of rocks.

The Boundary Zone is believed to have formed from rocks that have risen from depth through tectonic processes. There is no observed constraint upon any differential displacement across or along the high strain zone in the fabric of these rocks, and the Westing and Kirkaby gneisses on either side of the Westing Ultramafic Zone appear similar. The Boundary Zone has the overall appearance of a tectonic diapir or extrusion wedge of a series of lenticular masses of unrelated rocks, pinched and compressed between the Moine and Dalradian successions.

Westing Ultramafic Zone

The Westing Ultramafic Zone can be traced from its truncation by the Lund Fault at Lunda Wick [HP 568 042] northwards to beyond the Wick of Collaster [HP 575 078], where it presumably continues offshore for some 4 km. The zone reappears to the north of the Woodwick Fault [HP 575 118] and extends northwards along the western flank of Sneuga, disappearing offshore again at [HP 576 135], east of Gable. South of the Lund Fault, the Westing Ultramafic Zone is displaced beyond the Bluemull Sound Fault and the western edge of the district, from where it crosses the Bluemull Sound Fault again to the east and joins up with the Hascosay Slide Zone on Hascosay and its southern continuation, as described earlier (Flinn, 1994a).

The Westing Ultramafic Zone is an amphibolite- dominated zone, though hornblendic psammite, psammite and siliceous psammite occurs locally. The amphibolites are dominantly fine grained. They are intensely laminated, striped and banded on scales of less than 1 mm to several centimetres and have a parallel schistosity (Plate 1e); (P534940), (P533960). The laminations are often distorted by semi-parallel shears into an anastomosing, lenticular pattern. The laminations enclose unlaminated coarser hornblendite fish and lenses up to several metres across, some of which show signs of agmatitic texture. The hornblende forming the fine-grained laminated rocks is green, but the hornblendes forming some coarser grained unlaminated inclusions have brown cores. Garnet and biotite occur as minor constituents. The more psammitic rocks have decreasingly prominent lamination. Some hornblendic rocks, and in the Boundary Zone in general, show evidence of agmatisation (Plate 1d). It preceded the injection of the area by pegmatites and aplites and probably preceded the tectonic formation of the Boundary Zone. The agmatisation is not strongly developed and is confined to coarser-grained more deformed amphibolites.

Augen gneiss bands occur at Lunda Wick [HP 567 045] and Collaster [HP 5735 0720] (Figure 8), (Plate 1f) and contain plagioclase augen 2–3 cm in size composed partly of megacrysts and partly of aggregates of triple- junction equidimensional plagioclase grains 0.5 mm in size. The augen occur in a quartz-plagioclase-biotite matrix of 0.3–0.4 mm grains.

A heavily and coarsely sheared layer of steatite varying up to a metre thick can be traced discontinuously from Lunda Wick in the south to Wick of Collaster in the north of the Westing Ultramafic Zone. A similar lithology is a prominent feature of the Tonga and Herma Ness zones farther north. Serpentinite ‘zoned balls’ (Read, 1934b; Curtis and Brown, 1969, 1971; Moffat, 1987) up to 7 m in diameter and fragments of the same are associated with the steatite layer. Where tectonically undisturbed they are enclosed in a series of monomineralic, metasomatic zones resulting from reaction between the ultramafic core and the enclosing quartzofeldspathic gneisses. The zones, in order from the centre outwards, are: antigorite, steatite (talc and carbonate), lustrous green actinolite, chlorite and brown phlogopitic biotite. In places the phlogopite has been transformed to Na vermiculite (‘bronzy-biotite’ according to Read, 1934b; Curtis et al., 1969). Deformation has often partially destroyed the zonal structure by intermixing fragments of actinolitite and biotitite with the steatite and shearing them into adjacent rocks.

The best development of the zoned balls and the steatite zone is on the west side of Lunda Wick [HP 566 045], where the largest zoned ball is 7 m in diameter (P534175). North of Wick of Collaster [HP 575 077] antigorite ‘balls’ up to 4 m diameter and 10 cm lenses of actinolite and bronzy biotite are scattered in the laminated rocks forming the cliff face. The steatite layer is not present there. On the south side of the Wick [HP 574 072] the steatite layer is streaked out in glacitectonic thrust at the back of the beach; actinolitite nodules ((P534036), (P533309)) and bronzy-biotite streaks occur in adjacent rocks. Two abandoned Norse steatite quarries occur on the hillside opposite Brough Holm [HP 570 060]. Steatite occurs discontinuously on the north coast, and again along the south coast of Houllnan Ness [HP 570 055]. There the sheared steatite contains many actinolite and biotite lenses, zoned balls up to 2 m in size and a 0.5 m lens of radiating actinolite coated in biotite.

The presence of similar-sized spherical intrusions of ophitic metadolerite west of the ultramafic zone in Hascosay [HU 555 915] (Flinn, 1994a) suggests that the zoned balls are metamorphosed intrusive globular bodies of peridotite. Analyses of the antigorite core of the zoned balls (Curtis and Brown, 1969; Moffat, 1987) give an ultrabasic bulk composition but contain more Al₂O₃ and CaO than analyses of peridotite from the Shetland Ophiolite-complex, possibly due to metasomatism. However, steatitisation of antigorite in the ophiolite-complex is nowhere accompanied by the development of actinolite or biotite, though chlorite does occur. The volume of steatite in the sheared steatite layer north of the Lamb Hoga Fault appears to be greater than could be harvested tectonically from the few zoned balls present. It is probable that north of the fault a peridotite sheet was intruded as well as the balls. The absence of actinolite-biotite zoning in the steatite layer as compared to the zoned balls is unexplained.

Hascosay Slide Zone

The Hascosay Slide Zone comprises a high-strain hornblende-rich zone of ultramafic rocks along the Yell side of the boundary between the neighbouring Yell sheet and the district (BGS, 1993). The ‘slide zone’ occurs mostly within Yell and was described in some detail in the Yell Memoir (Flinn, 1994a). No blastomylonitic high strain ultramafic rocks occur south of Hascosay. The slide zone is cut and offset by the Bluemull Sound Fault; restoration of the offset (3 km dextral) on that fault shows that all of the Hascosay Slide Zone lies south of the Lamb Hoga Fault (Figure 3).

The Hascosay Slide Zone contains a series of closely packed, structurally resistant/competent, lenticular masses of hornblende gneiss measured in tens of metres. These lenticular masses include hornblendite, hornblendic orthogneiss and some quartzofeldspathic gneiss, contained within a network of very fine-grained and intensely laminated blastomylonite derived through shearing of the resistant masses by differential flow. The occurrences of hornblende orthogneiss are broadly similar to the occurrences of hornblende gneiss forming the Lewisian inliers in the Moine to the west in Yell (Flinn, 1994a, plates 2 and 9).

The finely laminated matrix occurs in three varieties. One type has the field appearance of a tectonised aplite, being a faintly but intensely laminated aplite- blastomylonite (Flinn, 1994a, plate 9b). This occurs only on the south coast of Hascosay, and is uncommon in Unst. Much more common and widespread are laminated and banded hornblende-bearing blastomylonites, considered equivalent to that seen in the Westing Ultramafic Zone though finer grained than those in Unst and more finely and intensely laminated. Grain size ranges between 0.1 and 0.2 mm. The laminations are formed by thin rectilinear strings of coloured minerals, usually biotite or hornblende, and by rows of elongate quartz grains arranged to form continuous parallel-sided quartz lenses 0.1 to 0.2 mm thick and extending 20 or 30 mm. The third type of laminated rock is much less prominent than the first two. It occurs as fine-grained, faintly laminated mica- poor psammitic blastomylonites composed dominantly of quartz and plagioclase, and with some biotite and/or hornblende. The hornblende has probably been added through tectonic mixing with adjacent hornblendic rocks in the high strain zone. Large masses of this third matrix type occur especially on Burra Ness [HU 555 956] where they probably represent blastomylonitic quartzofeldspathic rocks derived from the Yell Sound ‘Division’ to the west.

The resistant masses within the slide zone range in size from tens of centimetres to several tens of metres. The most prominent are coarse hornblendite, hornblende gneiss and hornblende-quartzofeldspathic banded gneiss and rocks of hornblende-gabbroic-appearance. They are spindle-shaped bodies contained within the laminar fabric of the laminated hornblendic rocks. In the field these masses vary from black and inscrutable to coarse black and white speckled rocks. However, in thin section original large grains of hornblende several millimetres across are seen to have been annealled to aggregates of individual grains 0.1 to 0.2 mm across. The large feldspar grains are similarly recrystallised to slightly larger grained aggregates. In the hornblende-feldspar gneiss original pyroxene cores to the hornblende grains have, in nearly all cases, recrystallised to micro-granular aggregates of amphibole.

Zoned balls occur whole and fragmented, as in the Westing Ultramafic Zone on Unst, but there is no sign here of a steatite layer. Lenses and masses of lustrous- green actinolitite varying from fist-sized to 3 m across are also present, associated with talcose lenses containing brown biotite and paler actinolite. Zoned balls occur on the tip of Crussa Ness [HP 553 023], and Burra Ness [HU 550 949]. An analysis of the latter is presented in Flinn (1994a, table 21, analysis 1).

An Ar-Ar step heating age obtained for blastomylonite hornblende from the Hascosay Slide Zone is 496 ± 6 Ma and for associated blastomylonite biotite is 436 ± 7 Ma (Roddam et al., 1994a, p90). The former is considered to be the age of tectonic emplacement of the Boundary Zone and is not significantly different from the age of obduction of the Shetland Ophiolite-complex (498 ± 2 Ma, Flinn et al., 1991). It follows that the protolith of the blastomylonite, the Lewisian inlier-type hornblende gneisses, is Proterozoic or Archaean in age.

Tonga Granite

The Tonga Granite is intruded into the Tonga section of the Westing Ultramafic Zone and is described here as a component of that zone. The granite is exposed in the cliffs of the western tip of Tonga [HP 581 148] and can be followed south for 1 km, through some sea stacks, to intersect the coast at the cliffs at Greff [HP 581 138], where it occupies a 200-m length of cliff face. From Greff, the granite passes inland to the south and is not further exposed. Small exposures of similar rock crop out on the coast at Herma Ness [HP 600 178]. The granite comprises a 200 m-thick layer dominantly composed of pink augen orthogneiss with lenticular to subrectangular, closely spaced, parallel-oriented, Carlsbad-twinned, microcline augen, 10 to 20 mm long (P533296). K-feldspar exceeds plagioclase in volume, both biotite and muscovite are present, and garnet is a common accessory phase. The granite carries a layer-parallel gneissic foliation and occurs as a series of anastomosing parallel sheets, generally more than 10 m thick and conformable with the layering in the adjacent metasedimentary country rock. Locally, a granitic texture is retained and the rock classifies as monzogranite on the basis of its modal composition (Figure 36). K-feldspar exceeds plagioclase in volume, both biotite and muscovite are present, and garnet is a common accessory phase. Whole-rock analyses (Table 1) indicate that SiO₂ is in the range 73.15 to 74.65 and (Na₂O + K₂O) is in the range 7.24 to 8.09. Samples are moderately peraluminous (Figure 37) and have high-K calc-alkaline to weakly shoshonitic affinity.

On the west face of Tonga, individual granitic sheets are typically separated by thin screens of non-granitic rocks (P533306). These are usually less than a metre thick but several swell to 10 m and more. They include layers of typical Westing Ultramafic Zone laminated hornblende schist ((P533348), (P533353)) The most common rock in the screens is steatite, sheared out and locally containing lenticular fragments of actinolitite and biotitite ((P533339), (P533456)). Other rocks in the screens, sometimes accompanied by steatite, are well- crystallised granofels composed of such minerals as green hornblende, biotite, plagioclase, and quartz with sparse muscovite and/or garnet, and metasedimentary psammitic schists (P533454). Some of the hornblende- rich rocks are partly agmatitic.

At Greff, thick layers of granite ((P533453), (P533460)) are separated by sheets of steatite, aplite, pegmatite and laminated hornblende schist (P533456). The Tonga Granite is directly structurally overlain to the east by up to 40 m thickness of coarse-grained orthoquartzite, containing a 20 m-thick central layer of diopside-epidote- quartz calcareous quartzite (P533457). On the west face of Tonga [HP 583 147], a 50 m layer of quartzite is interbanded with 10 to 20 m of calcareous quartzite, where it is difficult to reach and mostly weathered (P533341). The rock is typically composed of quartz and diopside with accessory sphene, biotite and garnet and zoisite.

The Tonga Granite has not been dated directly. The range of lithologies present in association with the augen orthogneiss suggest a link with the Westing Ultramafic Zone/Hascosay Slide Zone.

Herma Ness Ultramafic Zone

The Herma Ness Ultramafic Zone is exposed along the cliffs at Herma Ness [HP 600 175] and lies to the west of the Herma Ness Gneiss. This ultramafic zone shares similarities with the Westing Ultramafic Zone and, although critical relationships are obscured in cliff exposures and beneath the sea, it is regarded as a probable northward continuation of the Westing Ultramafic Zone. The Herma Ness Ultramafic Zone occurs as a 50 m-thick band of black granofels dipping east at 40º, which can be traced along, and about half way up, the cliff face from [HP 603 167] at Sothers Stack to [HP 604 183] at Looss Wick ((P533060), (P533065)). South of Sothers Stack, it passes inland and is not seen again to the south. The black granofels forming much of the zone varies from homogeneous to banded to laminated. It is a plagioclase > quartz granofels with biotite and/or green hornblende. Thin bands of hornblendite up to 10 cm thick, and 20 cm lenses of actinolite occur. In places the granofels contains scattered plagioclase microporphyroblasts, ill-defined zones of plagioclase and/or microcline megacryst augen and a band of Tonga Granite orthogneiss 3 m thick (P533099). The augen gneisses occur in the cliff face at Herma Ness [HP 603 183]–[HP 603 170] (Figure 8) as lithologically distinct bands, or as zones, of augen in otherwise uniform granofels (P533033). They vary from several tens of centimetres to several metres wide (Fernando, 1941, plates 6 and 7; Read, 1942) (P533033). Both plagioclase and microcline occur as 3 cm augen in quartz-plagioclase-biotite 0.3 mm matrices. Also present are lenses of 2 to 3 cm augen gneiss including plagioclase and rapakivi-type augen in a biotite-quartz- feldspar matrix.

At Looss Wick, a thin sliver of Tonga Granite (P533104) occurs to the east of the ultramafic zone, followed eastwards by the Herma Ness Marble ((P533034), (P533052)) then minor occurrences of quartzite and epidotic quartzite and then the main outcrop of the Herma Ness Gneiss of the Burra Firth Formation. The base of the zone extends along the foot of the cliff between it and the sea stacks offshore. The stacks from Flodda Stack [HP 599 175] to the Greing [HP 600 182] are composed of biotite psammite grading to muscovite- biotite pelite containing garnet and staurolite. They are patchily gneissose and are similar to the Kirkaby Gneiss that lies immediately west of the Westing Ultramafic Zone. The contact between the Herma Ness Ultramafic Zone and the rocks to the west is in places an east- dipping thrust, running along the base of the cliff, accompanied by a sheared steatite band which ranges from 30 cm to more than a metre thick and contains lenses of biotitite and actinolitite. Lenses of megacryst augen gneiss and granite, similar to the Tonga Granite, are also associated with this thrust.

An 8 m-thick zone of amphibolite and sheared steatite (P533371) followed to the east by a 4 m thickness of quartzite and then 3 m of amphibolite is exposed at Burgar Stack in the western cliffs of Burra Firth between [HP 6120 1453] and [HP 6115 1442] south of the lighthouse station. Apart from four small lenses of hornblende schist this is the only known occurrence of hornblende in the Burra Firth Formation. This occurrence may somehow be related to the ultramafic zone to the north- west, but there is an almost complete lack of inland exposure in the Herma Ness area. The Herma Ness Marble and the quartzite-calcsilicate rocks overlying the Tonga Granite at Greff, on the west face of Tonga, are probably northward extensions of the Westing Limestone.

Westing gneiss

The Westing Gneiss is best seen in the one and a half kilometres of west-facing cliffs south of Skitsack [HP 575 093]. The dominant rock type is a dark coloured biotite-rich garnetiferous psammitic granofels (P533825); reddish-brown biotite is more abundant than muscovite. It varies locally by increasing muscovite content to kyanite- and/or staurolite-bearing pelitic schists. These rocks are widely and variably, but never very strongly, gneissose. Calcareous layers and pods and lenticular layers of amphibolite are a minor component of the Westing Gneiss (P533824); layers of calc-silicate rocks and marble a metre or two wide occur locally on the west-facing cliffs for a kilometre south of Skitsack. The amphobolite layers contain reddish-brown hornblende and some are weakly agmatitic.

The Westing Gneiss continues north from Skitsack into South Holms [HP 574 100] where pods and layers of amphibolite are well displayed. The northernmost area of the formation, west of Sneuga [HP 575 125], is poorly exposed but lithologically distinct from the other areas farther south. It differs in being a dark, biotite-rich garnetiferous psammitic gneiss and in being consistently richer in mica and more strongly gneissose. Furthermore, it is free of amphibolites.

South of the Skitsack area, the Westing Gneiss passes inland to the Lund Fault [HP 575 036], thence across the Wick area south to Snarra Voe [HP 565 020]. Much of the area south of the Lund Fault is underlain by a uniform development of plagioclase microporphyroblastic gneiss (Plate 1c); (P534200). This development grades off to the north-west and is absent approaching Blue Mull [HP 550 045], being replaced by weakly gneissified psammite (P534141). A well-defined band of muscovite- rich schistose pelite occurs within the formation and is assigned to the Snarravoe Pelite. North of the Lund Fault between [HP 575 040] and [HP 572 043], the pelite contains staurolite and kyanite with some chloritoid. The type locality for the Snarravoe Pelite is south of the Lund Fault in the Snarra Voe area [HP 565 030] and [HP 563 022]; there the pelite contains abundant several centimetre- sized chloritoid porphyroblasts in a muscovite-rich groundmass.

South of the Lamb Hoga Fault, the Westing Gneiss forms the east side of Linga [HU 560 980] and lies to the east of the Hascosay Slide Zone on Burra Ness in Yell [HU 556 955] and on Hascosay [HU 560 925]. The rocks in these areas are generally similar in type to those on Unst, but are of higher grade early metamorphism (Chapter 5). The larger amphibolite bodies comprise reddish-brown hornblende and are more strongly agmatitic than those north of the fault. Units of schistose pelite assigned to the Snarravoe Pelite occur close to the Westing Limestone on Burra Ness [HU 557 954] and Hascosay [HU 563 932] and are rich in chloritoid (P541622), chlorite, shimmer aggregate and muscovite. On Hascosay, a well-defined 20 m-wide unit of coarse (2 mm grain size) microcline-oligoclase-quartz-biotite-muscovite-garnet granofels with some feldspar augen (microcline and plagioclase about 1 cm in size) occurs on the north coast at Virdi Taing [HU 563 932] (Figure 8). A further 200 m-wide occurrence is found on the south coast, near Skulia Geo [HU 563 922]. No other rock like it occurs in the Valla Field and Lamb Hoga blocks, as K-feldspar is otherwise limited to granitoids (including pegmatite and aplite), calc-silicate granofels and megacryst augen gneisses.

Westing limestone

The Westing Limestone typically comprises thinly interbanded calcareous and non-calcareous rocks. It is thus very unlike the marble-dominated limestone units of the Dalradian of the Mainland of Shetland and more like the calc-silicate-rock dominated Vidlin Limestone Formation [HU 470 640] of that same region.

The Westing Limestone is best displayed in the cliffs east of Skitsack [HP 577 093] in a complete and accessible section about 200 m wide. The cliffs display alternating vertical layers up to 20 m wide of calc-silicate granofels, marble, quartzite, amphibolite, and gneissose and non-gneissose psammite. The amphibolite varies from centimetre-thick layers in the calcareous granofels to metre or more thick discrete layers and even thicker lenticular bodies. The calcareous granofels comprises zoisite and/or diopside- bearing rocks with lesser amounts of biotite, quartz, oligoclase, sphene and microcline. The amphibolite contains green hornblende with reddish brown biotite.

Between Skitsack and Wood Wick to the north [HP 581 113], the Westing Limestone can be recognised in the cliffs as discontinuous layers of calc-silicate-granofels. The calc-silicate rocks are spatially associated with lenticular amphibolite layers and ‘fish’ in the schistose and gneissose psammite of the adjacent Valla Field Gneiss. The calc-silicate layers are mostly 1 or 2 m thick but locally reach 4 m and in one place 20 m. The calc-silicate rocks of the Westing Limestone can be followed south to the Lund Fault [HP 577 035] always structurally above the Westing Gneiss along the foot of the Valla Field escarpment in a minor valley with rare exposures. South of the Lund Fault, poor exposures reveal the presence of a patch of the Westing Limestone at Snabrough [HP 570 030], while farther south, at Ness of Wadbister [HP 560 016], the formation is 100 to 150 m wide, composed of the same range of lithologies as at Skitsack, and complexly folded on east–west horizontal axes.

South of the Lamb Hoga Fault, the Westing Limestone is exposed along the east coasts of Linga [HU 560 985], Burra Ness [HU 557 955] and Hascosay [HU 565 925] and on Fetlar at Ness of Brough [HU 576 925] ((P534616), (P534618)), and Corbie Head [HU 583 914]. On Burra Ness and Hascosay, the Westing Limestone is found adjacent to the chloritoid- bearing Snarravoe Pelite of the Westing Gneiss. The occurrences at the first three localities are very similar. Layers of marble, calc-silicate granofels, and gneissose psammite are interbanded with lenticular masses of amphibolite. The marble and calc-silicate granofels layers are often intensively internally laminated and striped with thin hornblende-rich layers. The lamination and the schistosity in the psammitic rocks often cuts across layering and banding in the masses of amphibolite due to differential deformation but in all of these occurrences, the metasedimentary layers are complexly distorted in the proximity of the lenticular masses of amphibolite. On Fetlar, at Ness of Brough and Corbie Head, occurrences of marble and calc-silicate granofels, mineralogically similar to the above, are interlayered with hornblende schist and gneissose psammite.

The calc-silicate granofels in all of the Westing Limestone occurrences contain diopside, and south of the Lamb Hoga Fault contain diopside with microcline. Zoisite and colourless actinolitic amphibole are very common and the latter often replaces diopside. Many lenses of amphibolite in the Westing Limestone contain brown hornblende.

Valla Field gneiss

In the east-dipping succession on Unst, the Valla Field Gneiss lies structurally beneath the Valla Field Schist and forms the west-facing scarp slope of the Valla Field escarpment, from Greff [HP 584 145] in the north, to east of Underhoull [HP 583 040] in the south. The Valla Field Gneiss and the Valla Field Schist were included in Read’s (1934a) Valla Field Group, but although clearly distinguished in his lithological description, were not separated by a line on his map. The boundary between the Valla Field Gneiss and the Valla Field Schist is defined in the field by the change from the coarsely schistose, highly micaceous Valla Field Schist to the less strongly micaceous, more psammitic granofels rocks of the Valla Field Gneiss. The latter are strongly and consistently gneissose and contain lenses of hornblende schist.

To the south of the Woodwick Fault [HP 583 115], the western boundary of the Valla Field Gneiss is clearly marked by the trace of the Westing Limestone at the foot of the escarpment. North of the Woodwick Fault, inland bedrock exposure to the west of the Valla Field Schist is largely absent due to the peat cover across Sneuga. The Westing Limestone is not apparently exposed and the western boundary of the Valla Field Gneiss is thought to pass over Sneuga [HP 573 125] to Greff [HP 583 140] rather than along the valley floor to the east (BGS, 2002). The Valla Field Gneiss in Fetlar lies structurally above the Valla Field Schist in a south-west-dipping succession along the coastline in the Ness of Snabrough [HU 580 925] and Rams Ness [HU 605 875] areas.

The rocks forming the Valla Field Gneiss are comprised dominantly of psammite, but vary locally to pelite and hornblende schist. The rocks are nearly as coarsely crystalline as those of the Valla Field Schist with matrix mica and chlorite between 1 and 2 mm. Near the Westing Limestone, the Valla Field Gneiss contains scattered fish and lenses of amphibolite up to a metre in size. The south-west tip of the Lamb Hoga Block is occupied by a variant of the Valla Field Gneiss which is a well developed leucosome gneiss, very rich in fibrolitic sillimanite (P534911).

Valla Field Group

The Valla Field Group comprises the Valla Field Schist and the Burra Firth Formation and belongs to the Shetland Dalradian.

Valla Field Schist

The Valla Field Schist on Unst is an east-dipping layer of coarse muscovite-rich schistose Al-rich pelite forming the summit ridge of the Valla Field Block from Tonga [HP 585 154], where it runs out to sea to the north, to Trona Dale [HP 583 043], where both the schist formation and the overlying Burra Firth Formation cut out against the Lower Imbricate Zone of the Shetland Ophiolite- complex. Rocks assigned to the Valla Field Schist also occur within the Lower Imbricate Zone around [HP 583 035] and in the immediate footwall to the imbricate zone on Hoga Ness [HP 557 007]. The Valla Field Schist is also recognised in the Lamb Hoga Block on Fetlar where it occurs as a west-dipping layer structurally overlying the Burra Firth Formation. It is exposed in the cliffs at Ness of Snabrough [HU 580 937], Hiplin [HU 617 873], and along much of the west-facing cliffs between [HU 586 913] and [HU 596 888].

The Valla Field Schist comprises dominantly coarse muscovite schistose pelite but is patchily gneissose and varies locally to semipelite and micaceous psammite. The schist is dominated by coarse mica and contains quartz segregations. It is rich in staurolite and/or kyanite with garnet and muscovite > biotite and/or chlorite. Plagioclase is often almost absent. The micaceous lithology is locally interlayered with lenticular units of orthoquartzite and impure quartzite, especially so in Fetlar where the quartz-rich lithologies tend to exceed the pelitic component in volume. These quartzose layers vary from 10 cm to 4 m or more in thickness and the thicker ones appear to preserve bedding structure. The pelitic rocks are sufficiently rich in aluminium to bear chloritoid at the appropriate grade.

The schist is most coarsely crystalline to the north of the Collaster Fault on Unst [HP 584 065], where kyanite, staurolite, and to a lesser extent garnet, are commonly visible as 20 mm grains. Mica flakes are generally about 2 mm long. South of the Collaster Fault, the effects of retrogressive metamorphism have generally reduced the grain size of the rocks by recrystallisation and cataclasis, though 20 to 30 mm sized chloritoid grains have grown locally in them in a still later prograde metamorphism (see further in Chapter 5).

Burra Firth Formation

On Unst, the Burra Firth Formation occurs along the eastern side of the Valla Field Block, from Herma Ness in the north to Loch of Watlee [HP 595 055] in the south, where it cuts obliquely out against the Shetland Ophiolite-complex. The rocks are dominantly composed of quartzite and impure quartzite. They are well exposed along the west shore of Burra Firth. On Fetlar, they are exposed only along the eastern shore of Lamb Hoga. The formation also includes: the Herma Ness Gneiss; the Gillis Field Microporphyroblastic Gneiss; and the Sunkir Marble. Superficial deposits of peat in both areas mean that inland exposures of the formation are confined to a few stream sections.

The quartzite and impure quartzite occur in beds, often flaggy, varying in thickness from 20 cm to 2 m. These lithologies alternate with more micaceous rocks forming interlayers ranging in thickness from 1 cm to several metres. The interlayers vary from psammite to pelite; a few are gneissose. Substantial masses of bedded orthoquartzite, lacking interleaving psammitic and semipelitic rocks, are exposed in the cliffs west of Moo Wick [HU 622 876] in Fetlar, inland in Unst at Bordi Knowe [HP 584 044], and most notably in the Muckle Flugga group of skerries [HP 605 195] including Out Stack [HP 613 203] which lie to the west of the Herma Ness Ultramafic Zone. Calc-silicate granofels and marble occur locally; the 5 m-thick planar laminated Herma Ness Marble occurs at Looss Wick, Herma Ness [HP 603 183] (see above) and a much bigger mass, the Sunkir Marble occurs at the south-east tip of Lamb Hoga at [HU 623 875]. Cliff sections to the east of Looss Wick, and on Fetlar between Lamb Hoga Head [HU 625 877] and [HU 618 890], reveal small lenses and streaks of calc-silicate granofels and marble (within the Gillis Field Microporphyroblast Gneiss) generally no more than 10 cm thick. A larger exposure of calc-silicate rock occurs on the north shore of Papil Water [HU 603 907]. The Herma Ness Marble and calcareous quartzite lies adjacent to the ultramafic zone on the west face of Tonga and so may represent a northward continuation of the Westing Limestone (see above).

The nature of the Burra Firth Formation changes drastically to the north of The Fidd [HP 616 167] on the western shore of Burra Firth. The succession remains dominated by quartzite and impure quartzite, probably to the same extent as to the south, but the rocks suddenly become much more strongly gneissose. In addition, they have been recrystallised to grain sizes of 2 to 3 mm from less than a millimetre. As a result, the quartzite, and more especially the impure quartzite, gain a gneissose appearance while the psammitic rocks become thoroughgoing quartzofeldspathic gneiss. These occurrences are assigned to the Herma Ness Gneiss. Overall however, they appear to be of no higher metamorphic grade than the rocks to the south. Garnet occurs widely but is sparsely distributed, and fibrolite is almost absent. The recrystallisation and gneissose textures are not apparent in the quartzite-rich Muckle Flugga Skerries or Out Stack to the north [HP 605 195].

In addition to the metamorphic changes, these northern occurrences of the Burra Firth Formation take on an east–west oriented L-tectonite fabric and are deformed into a complex array of folds which both pre and postdate the L-tectonite fabric. High cliffs and lack of inland exposure make the rocks difficult to study.

On Fetlar, in the south-east corner of the Lamb Hoga Block [HU 620 884], there is an ill-defined zone of psammitic rock that has a gneissose appearance due to the presence of closely spaced rounded plagioclase grains 2 mm in diameter or, more rarely, microcline grains. This rock has been assigned to the Gillis Field Microporphyroblast Gneiss. These rocks texturally resemble plagioclase microporphyroblast gneiss characteristic of the Kirkaby or Westing gneisses of the Westing Group, but have a muscovite > biotite matrix instead of a biotite > muscovite matrix. If the microporphyroblastic texture is genuinely the result of metamorphic recrystallisation, then these are thus the only occurrence in Shetland of plagioclase microporphyroblast gneiss outwith the Westing Group but the possibility remains that the Gillis Field occurrences are in fact psammite containing relict grains of clastic plagioclase.

Composition of the rocks of the Valla Field Block

A suite of 61 pelitic rocks from Valla Field and Lamb Hoga was collected, analysed and interpreted by Aziz (1984) with the purpose of reassessing Read’s (1934a, 1937) classic work on polymetamorphism in the area. These analyses, with 11 analyses by Key (1972), are summarised in (Figure 10).

Major element data are displayed on the Al₂O₃–FeO–MgO + CaO + Na₂O + K₂O molecular plot devised by Khoo (1974; Flinn et al., 1996c) to discriminate between pelites capable of crystallising chloritoid at appropriate grade (chloritoid-type pelites) and those not so capable (Chapter 5). This plot is used, as previously, because the distinction between these two types of pelite is essential to an understanding of the metamorphism of the Valla Field rocks. (Figure 10) shows that none of the pelites from Valla Field and Lamb Hoga plotting below the chloritoid limit contain chloritoid. However, many which plot above contain no chloritoid. Most of the latter occur in areas of coarse kyanite-staurolite-garnet schist of too high grade for chloritoid. This high grade applies to the specimens belonging to the Valla Field and Westing gneisses plotting above the chloritoid-limit and about half the Valla Field Schist specimens. However, the remaining seven specimens of the Valla Field Schist, and three Snarravoe Pelites specimens from the Westing Gneiss, are pelites of chloritoid-type but lack chloritoid, although they occur closely associated in the field with chloritoid-bearing rocks of similar composition. All chloritoid-bearing rocks plot in the chloritoid field above or on the chloritoid limit.

The pelitic rocks of the Valla Field Schist and Snarravoe Pelite are dominantly composed of chloritoid- type pelite to an extent that allows them to be mapped as chloritoid-type pelite units. Persistent bands of chloritoid-type pelite are a prominent feature of the Shetland Dalradian succession (Flinn, 1967b; Flinn et al., 1996). The origin of pelitic rocks of this composition is not clear (Khoo, 1974).

Relation of the Westing Group to the Valla Field Group

The rocks of the Westing Group are distinctly different from those of the Valla Field Group and lithologically unlike any other part of the Shetland succession. With the exception of the North Holm Gneiss, the rocks of the Westing Group comprise dominantly dark-coloured biotite-rich psammite to gneissose psammite with up to 30 per cent by volume of hornblende schist, amphibolite and hornblende gneiss. The Westing Group includes the ultramafic, hornblende-rich, high-strain Westing Ultramafic Zone/Hascosay Slide Zone and has been extended to include the rocks assigned to the Boundary Zone of Yell (Flinn, 1994a). South of the Lamb Hoga Fault, the Westing Group extends to the east of Linga, Burra and Hascosay as far as the west coast of the Lamb Hoga Block, where it occupies the Ness of Snabrough–Brough Lodge area [HU 575 925] and Rams Ness [HU 605 875] (Figure 7) (see footnote 3 in Chapter 2). The Westing Group thus encompasses the entire contents of the Boundary Zone and represents a tectonic assemblage of poorly defined and unrelated lenses of different gneisses and other lithologies of Proterozoic or Archaean age.

The Valla Field Group entirely lacks hornblendic rocks, contains very little limestone, being dominated by quartzitic psammites with a major band of coarse Al-rich micaceous schist, and very little gneiss. The group has been correlated with the Scatsta ‘Division’ of the East Mainland Succession of Shetland and thus with the lower Dalradian of Scotland (Flinn, 2007). The Scatsta ‘Division’ has a quartzite–psammitic lithology with bands of coarse Al-rich micaceous rocks. The G-base regional geochemical map for boron in Shetland (IGS, 1978a) shows the outcrop of the Valla Field Group in the district to be less than half as rich in boron (at 40 ppm) than the Scatsta ‘Division’ rocks on the Mainland west of the Nesting Fault. Nevertheless, tourmaline is an unusually common and prominent accessory mineral in the Valla Field Group rocks on Unst and the reported concentrations of boron are little less than that associated with lower Dalradian rocks in Scotland (Plant et al., 1984). The contrast with the adjacent Moinian rocks of Yell is stark. There, as in Scotland, boron occurs at background level. On the other hand uranium, potassium and rubidium are at higher levels in Yell than in the outcrop of the Valla Field Group. These differences are also typical of Dalradian and Moine rocks in Scotland (Plant et al., 1984).

Saxa Vord Block

The Saxa Vord Block comprises those rocks in Unst that crop out to the east of Burra Firth and Loch of Cliff (Figure 11) and to the west of the Skaw Granite and the Lower Imbricate Zone of the Shetland Ophiolite- complex. At its south end, where the trace of the Lower Imbricate Zone passes the southern end of the Loch of Cliff, the Saxa Vord Block pinches out into the Lower Imbricate Zone as one of its tectonic slices (Figure 2). The Saxa Vord Block lies structurally beneath the Shetland Ophiolite-complex.

The metasedimentary rocks forming the Saxa Vord Block have been classified as the Saxa Vord Group. That succession comprises the Queyhouse Flags, Saxa Vord Pelite and Hevda Phyllite (BGS, 2002; (Figure 11). The Summit Pelite and the Stabba Pelite occur within lower and upper parts respectively of the Queyhouse Flags. In contrast to the lithologies present in the Valla Field Group, no gneissose lithologies occur in the Saxa Vord Group.

Saxa Vord Group

Queyhouse Flags

The Queyhouse Flags dominantly comprise bedded orthoquartzite and impure quartzite. The Stabba Pelite comprises a unit of schistose pelite in the west of the formation, close to the Burra Firth Lineament [HP 613 135]. The Summit Pelite comprises a wedge or lens-shaped unit of phyllitic pelite centred on the summit of Saxa Vord [HP 630 165].

The Queyhouse Flags quartzites are spectacularly displayed in the cliffs forming the east shore of Burra Firth, extending from the Burra Firth [HP 617 142] beach to Hols Hellier [HP 627 177] in the north (Plate 2a), but they are not everywhere accessible. To the east of the Summit Pelites, the quartzites are exposed on the col at Loomer Shun [HP 633 159], at the foot of the cliff at the west end of Brei Wick [HP 635 175] (P533147), and farther north in the cliffs at The Lug [HP 636 180] (P533133) and immediately west of The Lug. They are also well exposed in the banks of the Burn of Sulerdale [HP 622 135], and along the shores of Loch of Cliff [HP 603 125], including the eastwards branch at Quoys.

The beds of orthoquartzite are generally about 50 cm thick though they vary up to 1 m in thickness. They are often glassy in appearance; in some exposures they appear graded. Individual beds and groups of beds of quartzite alternate with layers of weathered, fissile, fine- grained phyllitic pelite and semipelite, which vary from centimetres to metres in thickness. These micaceous lithologies occur both as blackish (biotitic) and silvery grey (muscovitic) units in all the exposures of the Queyhouse Flags (P533386). They are best seen along the east end of the Burra Firth beach [HP 617 144], where polishing by the wave action shows them to advantage. However, these rocks are too fissile to be easily collected at this locality, due to a strong S-tectonite bedding- parallel schistosity and weathering effects. In the hinge areas of recumbent intrafolial isoclinal folds exposed west of Buddabrake [HP 6174 1419] and [HP 6135 1403], the alternating quartzite and black micaceous layers are seen to have preserved an original grading indicating younging to the north. Some of these micaceous layers are both graded and preserve a very fine lamination of sedimentary origin (Plate 2b) and (Plate 2c). Other exposures to the north and west of Buddabrake ( [HP 6196 1469] and [HP 6175 1425]) preserve highly deformed flame structures (Plate 2d). In fold limbs, the graded laminations have become attenuated with a superimposed strain-slip cleavage such that the grading is no longer detectable. On Burra Firth beach these rocks have been strongly kaolinised in places, e.g. [HP 616 141].

Boulders of quartzite on the beach of Ura Geo [HP 633 178], probably fallen from the quartzite band crossing The Noup from The Lug, contain spindle- shaped elongate pebbles one centimetre long, in an L-tectonite fabric.

The Stabba Pelite (Figure 11) comprises a unit up to 50 m wide, close to the western edge of the Saxa Vord Block. The unit strikes out to sea from a small headland in the middle of the Burra Firth beach [HP 614 141], where weathered fissile staurolite-garnet muscovite schist composed of interbanded pelite and semipelite occurs. Muscovite flakes are about 2 mm long. Staurolite grains are often altered to shimmer aggregate but are observed to truncate the primary schistosity, later microfolding is superimposed. Garnet occurs as small grains, less than 1 mm in diameter, which truncate mica flakes in the matrix. The pelite member passes gradually into quartzite east of the headland in a zone of interbanded lithologies. The pelite member appears again in exposure 500 m to the north-east of Burra Firth beach at [HP 617 145] (P533400) from where it can be followed along the foot of the cliffs for 2.5 km. The pelite appears less weathered in these exposures.

The Stabba Pelite is the source of a linear positive magnetic anomaly of about 1000 nanoteslas amplitude (Chapter 10). The anomaly can be followed inland to the south of the Burra Firth beach for a kilometre, and the subcrop of the pelite member has been traced on that basis to where the magnetic anomaly pattern ends just north of the Quoys branch of Loch of Cliff [HP 609 134]. Several magnetic anomalies of similar amplitude but limited extent occur to the east of the main band in the Sotland area [HP 612 134]. These effects may be due to the occurrence of separate lenses of the magnetic pelite or to infolded or disrupted fragments of the main unit. No magnetic outcrop can be associated with any of these subsidiary anomalies, but stone walls near them are composed of fragments of typical Stabba Pelite schist.

The Summit Pelite (Figure 11) is a mass or lens of pelitic rock varying from schistose pelite in the west of the outcrop to phyllitic pelite in the central and eastern parts. Mica flakes in the schistose pelite are greater than 2 mm in length whereas those in the phyllitic pelite are less than 0.5 mm in length. Where sufficiently clearly exposed, the phyllitic pelite is seen to be thinly laminated, but in most exposures the flaky nature of the lithology hides this detail. Several lenticular units of magnetic bedrock are apparent in the geomagnetic data but are undetected in the surface exposure. The outcrop of the member is about 0.75 km wide at its maximum and extends for some 3 km southwards from the cliffs north and west of Saxa Vord [HP 632 167]. Further exposure occurs in several places on, and to the south of, the summit of Saxa Vord [HP 632 167]. Considerable quantities of the rock have been excavated from the military compound on the summit of Saxa Vord and are available on spoil heaps nearby [HP 628 165].

Along its western boundary, the Summit Pelite passes into the quartzite of the Queyhouse Flags through a zone of alternating quartzite and semipelite which includes a quartz-rich calc-silicate granofels unit less than 1 m wide. This boundary is accessible in places along the cliff top at Hols Hellier [HP 627 177] and [HP 625 165]. The Summit Pelite thins and disappears southwards into the Queyhouse Flags, presumably by lateral facies variation.

Saxa Vord Pelite

The Saxa Vord Pelite has been divided into western and eastern parts by the Hevda Thrust (Figure 11). The Burn of Skaw provides a section with intermittent exposure in the north-western part of the outcrop and in the footwall of the thrust. At the point where the stream crosses the trace of the Hevda Thrust [HP 644 163] the phyllitic Saxa Vord Pelite is reduced to a mass of papery flakes. The pelite is very well exposed at The Noup [HP 634 184] in the far north of the outcrop, but access is very difficult. Except at the extreme ends of the beach, the outcrop of the pelite is obscured or displaced due to slope failure on the cliffs bounding the south side of Brei Wick [HP 637 173]. The pelite is otherwise poorly exposed inland.

The eastern area of the Saxa Vord Pelite is even more poorly exposed than the western. The pelites can be seen along the summit ridge of Ward of Norwick [HP 647 155], but only as drift and as spoil from trenching. However, they are obviously nearly in situ. Pelitic rocks in contact with the Skaw Granite are thoroughly hornfelsed (Chapter 7). Occurrences in the northern cliffs at Hill Ness [HP 652 170], in the Skaw Burn [HP 653 162], midway between these two locations [HP 653 165] and also on the summit of Ward of Norwick [HP 648 154] are thoroughly recrystallised by the thermal effects of the granite.

The aeromagnetic map (IGS, 1968b) shows a large north–south orientated lenticular positive anomaly over the centre of the Saxa Vord Block (Figure 53). More detailed ground magnetic surveying of the area of that anomaly to the west of the Hevda Thrust (Chapter 10; BGS, 2002), reveals a series of narrow elongate anomalies of 1000–2000 nT amplitude. It is a characteristic of these magnetic anomalies that the rocks causing them are even more rarely exposed than the non-magnetic Saxa Vord Pelite. Where sampling was possible, the rock was found to be a coarse muscovite schist, rich in flaky hematite grains orientated parallel to the schistosity.

Apart from the highly fissile phyllonitic pelite exposed along the trace of the Hevda Thrust (P533174) and the narrow zone of recrystallised hornfels against the granite, the Saxa Vord Pelite is typically composed of extremely uniform silvery muscovite-rich chloritoid- staurolite-kyanite bearing phyllitic to schistose pelite, with the pelite becoming more schistose towards the east. Sand-polishing of exposures on the Brei Wick beach [HP 637 173] reveals phyllitic pelite with poorly defined zones containing white spots which are shimmer aggregates after andalusite and some poorly developed silvery (muscovite-rich) and greenish-grey (chlorite-rich) colour banding ((P533137), (P533153)). There are only rare signs of preserved sedimentary structures; e.g. several examples of graded layers occur at Brei Wick beach [HP 6383 1708], 2 cm thick and younging to the east. Beautiful aggregates of blue kyanite rosettes in quartz occur in the outcrop of the Saxa Vord Pelite (Read, 1933).

Hevda Phyllite

The Hevda Phyllite can be distinguished from the Saxa Vord Pelite by the presence of obvious sedimentary layering, by mineral content, and by chemical composition (Flinn et al., 1996).

The Hevda Phyllite occurs as a large infold into the Saxa Vord Pelite from the north (Figure 11). It is exposed along the coast in the northern cliffs from Brei Wick [HP 641 172] to just short of the Skaw Granite contact at Hill Ness [HP 651 171] and along the Skaw Burn ( [HP 647 163] to [HP 653 161]). The boundary with the Saxa Vord Pelite is marked by thin and discontinuous units of quartzite which are associated with kaolinite, e.g. north- east of Ward of Norwick [HP 646 157] and [HP 649 156].

The Hevda Phyllite dominantly comprises massive phyllitic pelite, with chlorite typically present in excess of muscovite. Bands occur that are very rich in garnet, 1 mm in size, and other bands are rich in white and dark spots (shimmer aggregates). There are widely scattered segregations of quartz and pink feldspar (albite). Although the Hevda Phyllite has the general appearance of a phyllitic pelite, it is typically less micaceous and more semipelitic than either of the Saxa Vord or Summit pelites. In cliffs at Hevda [HP 646 174], units and laminations of sedimentary origin are visible, though not prominent, on all scales from millimetres to decameters (P533168). Siliceous laminations generally less than 1 cm thick, and finely internally laminated are a characteristic of this formation (P533139). These siliceous layers often contain a calc-silicate-like core composed of quartz ± garnet ± epidote. Such laminae vary in spacing from centimetres to metres, and in places variation in this spacing provides the basis for a larger scale banding.

The western boundary of the Hevda Phyllite with the Saxa Vord Pelite is well exposed and accessible on the cliffs east of Brei Wick [HP 642 172] in the footwall of the Hevda Thrust. At this location, quartzite layers alternate with phyllitic pelite in widely spaced beds up to 1 m in thickness. Layers of impure quartzite and semipelite are also present, some of which contain prominent quartz grains. Narrow lenses or streaks of quartz-rich calc- silicate granofels occur near Hevda [HP 642 172]. Farther to the east in the Hevda Phyllite, in the footwall of the Hevda Thrust near Hevda [HP 642 173], there are several conformable, sharply defined, lenticular bodies (e.g. 30 cm by 20 m) of fine-grained black rock. Thin sections show them to be composed of chloritoid, muscovite and quartz with uniformly distributed, very finely divided black inclusions. At the water’s edge, very close to the Hevda Thrust at Ritten Hamar [HP 6412 1712], there is a band of semipelitic rock, which is intensely laminated and graded on a 4 mm scale, indicating younging to the east. This observation, and the fact that the Hevda Phyllite does not occur between the Queyhouse Flags and the Saxa Vord Pelite, is considered to show that the Hevda Phyllite stratigraphically overlies the Saxa Vord Pelite.

The eastern boundary of the Hevda Phyllite on the north coast is exposed in the cliffs at Rurhella [HP 652 172], but is difficult of access and partly obscured so that it is unclear whether the narrow band of hornfelsed Saxa Vord Pelite between the Hevda Phyllite and the Skaw Granite reaches the water line, or is cut out by the Skaw Granite contact. The limit of the Hevda Phyllite eastwards is marked by a substantial quartzite layer, or layers, alternating with phyllitic pelite. The boundary quartzites are intermittently exposed as far south as the Burn of Skaw [HP 653 162]; granule grade quartz grains are visible.

Composition of the Saxa Vord Block rocks

Seventy-two analyses of pelitic rocks from the Saxa Vord Block are available (Key, 1972). The compositional fields occupied by these rocks are presented in (Figure 12. It is apparent that the Hevda Phyllite and the Saxa Vord Pelite are significantly different in composition in line with their different mineralogical compositions. The former has higher Mg and the latter higher Al. The Summit Pelite includes both types of pelite. This is shown by the plot in (Figure 12), using a format devised by Khoo (1974) to discriminate between pelites that can develop chloritoid at appropriate temperatures and pressures (chloritoid-type pelites), and those that cannot (normal pelites). The boundary between the two types of pelite is called the chloritoid limit and is marked on (Figure 11) by the line ‘CL’. Only two analysed chloritoid-bearing rocks plot below the chloritoid limit on (Figure 11). Both are Saxa Vord Pelite. All 60 available thin sections of the Saxa Vord Pelite were found to contain chloritoid, whereas none of the 80 thin sections of Hevda Phyllite contain chloritoid, although 4 out of the 15 analysed specimens plot in the chloritoid field.

The tectonic lens of chloritoid-free pelitic phyllite exposed in the cliffs at Norwick [HP 652 150] in the Lower Imbricate Zone is shown by (Figure 12) to have a compositional field very similar to that of the Hevda Phyllite. One out of seven of the analysed specimens plot in the chloritoid field, although it contains no chloritoid. Out of 18 analyses of the Summit Pelite, five contain no chloritoid and all plot in the chloritoid-free field.

On the stream-sediment-based regional geochemical maps (BGS, unpublished), the Saxa Vord Block is an area of relatively high B, Li, Al, Ga and Ti. The relative lack of K compared with Al, as shown by these maps, accounts for the dominance of chlorite over biotite.

Correlation of the Saxa Vord Group

The rocks of the Saxa Vord Group show many similarities with the rocks of the Clift Hills ‘Division’ occurring in the south of Shetland (Flinn, 1967b). The Clift Hills ‘Division’ is the easternmost unit in the East Shetland Succession, and is the infill of an upper Dalradian extensional basin (Flinn and Moffat, 1985; Flinn, 2007).

In the south of Shetland, the Clift Hills Phyllitic Formation, at its lower boundary, comprises bedded quartzites showing evidence of turbidite deposition. The quartzites alternate with layers of fine-grained, thinly banded and graded, phyllitic semipelite, which have the appearance of distal turbidites. These are very like the Queyhouse Flags. The Clift Hills Phyllites continue upwards as fine-grained phyllitic pelite and semipelite passing finally into chloritoid phyllites. This association compares well with the Hevda Phyllite and Saxa Vord Phyllite in Unst. The chloritoid-bearing Clift Hills Phyllites have the same composition, and bear the same minerals, as the Saxa Vord Pelite. Where the former are thermally metamorphosed in the aureole of the Spiggie Granite (Flinn et al., 1996) they develop a very similar sequence of minerals to those produced by the contact metamorphic effects of the Skaw Granite in the Saxa Vord Pelite.

The Saxa Vord Group is therefore considered a continuation of the Clift Hills ‘Division’ and to be upper Dalradian. The high value of boron for the Saxa Vord Block shown by the Regional Geochemical stream sediment maps is characteristic of Dalradian rocks (Plant et al., 1984).

Chapter 4 Structural geology of the basement to the Shetland Ophiolite-complex

Structure of the Valla Field and Lamb Hoga Blocks

(Figure 13) presents stereographical projections of planar and linear structural data for the metasedimentary and metavolcanic rocks making up the Valla Field and Lamb Hoga blocks in the western parts of Unst and Fetlar. These rocks are bound to the east by the Lower Imbricate Zone on Unst, and the Lamb Hoga Fault on Fetlar. The foliation poles were measured unselectively on schistosities and surfaces of compositional layering, as these are invariably parallel.

The data presented on (Figure 13) are subdivided geographically into structurally homogeneous domains. The characteristic structural pattern in each domain reflects the domain location within the overall structural architecture of the district; data for each domain is bound as far as possible by faults. The domains are clearly reflected in the underlying lithology, and in the nature of the structures developed in that lithology, so that domains W1 to W4 cover the Westing Group (and thus the Boundary Zone), while domains VF1 to VF4 cover the varied pelitic to quartzitic lithologies of the Valla Field Group and thus the Dalradian rocks. Domains W1 and W2 cover the cuesta escarpment of the Valla Field Block while domains VF1 to VF3 cover the cuesta dip slope. All of these different aspects of the domains strongly influence both the type of structural feature likely to be present at outcrop, as well as the degree of exposure presented.

The foliation in the Valla Field Group on the east- facing cuesta dip slope of Unst (domains VF1 to VF3) dips consistently to the east or east-north-east, associated with a general pattern of east or east-north-east-plunging down-dip lineations. Foliations in Valla Field Group rocks in Lamb Hoga Block on Fetlar dip generally to the south-west with few lineations which are broadly south- west-plunging and down dip (domain VF4). Foliations in the Westing Group exposed on the cuesta escarpment of the Valla Field Block dip generally east or north-east (domain W1) while foliations in the Westing Group in the southern parts of the Valla Field Block form an array dipping gently west, south or east associated with subhorizontal south-south-west-plunging lineations.

Exposures of Valla Field Group lithologies which occur on the cuesta dip slope in the Valla Field Block are frequently of insufficient depth to reveal folds of much more than a metre or so in amplitude. Few folds are observed in the Valla Field Schist and those are of late development associated with phyllonitisation (see below). Folding in the Burra Firth Formation is observed as intrafolial isoclinal structures but is typically not well exposed. Rectilinear parallel fluting lineations (rodding) are often observed in association with these isoclinal folds and a more widespread occurrence of these folds is often revealed by fluting lineations observed on the surface of quartzite and impure quartzite beds. Somewhat similar flutings occur on the surfaces of conformable and near-conformable aplite/pegmatite sheets injected into the Burra Firth Formation. The folds are often multiple and concertina- like and almost ptygmatic in style. These lineations are considered to arise from the interaction of the schistosity with the aplite/pegmatite sheets as the latter rotate into conformity with the schistosity during flattening.

Fabric lineations (crenulations) mostly arise because of minor folding of the schistosity in fold closures, elongated leucosome segregations and due to preferred parallel orientation (girdle or partial girdle arrangement) of mica in L-S fabrics in the Valla Field Schist, on which strong mineral lineations based on aligned elongate kyanite grains were observed in several places.

The Westing Group (Boundary Zone) (see footnote 3 in Chapter 2) is well exposed on the cuesta escarpment of the Valla Field Block and the coast beyond to the west. Domains W1 and W2 clearly reveal the nature of the folding, which is considered to be of post-Boundary Zone age. North-easterly trending folds with a shallow (down-dip) plunge, and with amplitudes varying up to 80 m or more, are frequently apparent in the cliffs and escarpment south of Woodwick. These structures are usually overturned to the north. Larger folds are often made up of a series of much smaller folds with similar profile. Folds that have more north–south, strike-parallel axes are rarely more than 1 m in wavelength, and are usually overturned to the east. This latter set is typically subhorizontal, south-plunging in domain W2 and north- plunging in domain W1.

Fold axes and other lineations located in or close to the Hascosay Slide Zone (domain W4 on (Figure 13), and based on the blastomylonitisation, trend parallel to the strike of the slide zone and plunge at c. 20° to the north-west (Flinn, 1994a); foliations in or adjacent to the Hascosay Slide Zone dip to the west-south-west or south-west. Flinn (1994a, chapter 8) describes the structure of the Hascosay Slide Zone in some detail. That detail is not reproduced here except to say that preferred lattice and/or grain shape fabrics in the laminated blastomylonitic rocks of the slide zone are regarded as orthorhombic and that these rocks are believed to mark a zone of intense compressional pure shear which has mobilised the rock and forced it to extend by flow towards the north-west and the south-east, parallel to the observed fold axes and lineations. The Westing Ultramafic Zone dominantly comprises intensely and finely laminated hornblendic rocks, equivalent to the blastomylonites of the Hascosay Slide Zone, and only slightly different in that they are more uniformly planar due to the paucity of resistant amphibolite and amphibolite gneiss masses distorting the flow. The foliation in the Westing Ultramafic Zone dips to the east, while the lineations and fold axes plunge to the north-east (domains W1 & W2 on (Figure 13).

The east-dipping nature of the Valla Field succession led Read (1934a, 1937) to position it on the eastern limb of a north–south antiformal structure. However, recognition that the Lamb Hoga Block contains a similar, but inverted, west-dipping succession (Aziz, 1984) necessitated a reassessment of the larger scale structure. Comparing measurements of east-dipping schistosity in domains VF1, VF2, VF3, and W1 on (Figure 13), with that from the west-dipping domains VF4 and W4 has been used to infer the existence of a recumbent fold with a horizontal fold axis whose azimuth is about 160°N — the Valla Field Fold. Domains W2 and W3 are believed to have been tectonically rotated out of this alignment. The broad scale model now presented on the published geological map (BGS 2002; see also Flinn 2001, 2007) has the Valla Field Block as the upper limb of a giant recumbent fold closing to the east with the Lamb Hoga Block as the lower limb of that fold (see (Figure 7). The folds and lineations contained within the large scale recumbent Valla Field Fold cannot be related to it and there is no closure exposed.

Phyllonitisation

The structure of the two blocks is complicated by widespread phyllonitisation. This takes the form of an apparent enhancement of the schistosity of the Valla Field Schist and to a lesser extent of the underlying Valla Field Gneiss. The occurrence starts immediately north of the Woodwick Fault and extends northwards for a kilometre (Figure 14). It limits the southwards extension of the coarsely crystallised Valla Field Gneiss south-east of Tonga [HP 585 130]. The phyllonites are more coarsely fissile than unaffected schists to the north, and have developed a prominent schistosity parallel to the earlier fabric and lithological layering, but cutting earlier folds and pegmatitic lenses. In thin section, the phyllonitic fabric is seen as millimetre-thick, crudely formed bands of microgranulitised rock; in the unaffected schist the minimum grain size is 1 to 2 mm, while in the phyllonitised bands the maximum size is 0.1 to 0.2 mm. The grain boundaries within the bands are crystalline not cataclastic. Phyllonitisation was not accompanied by any mineralogical changes, such as biotite alteration to chlorite. Folding of the phyllonitic fabric is sparse.

To the south of the Woodwick Fault, the schistosity of the Valla Field Schist has been uniformly phyllonitised to an extent easily visible in the field. It leaves these rocks with a degraded appearance. In thin section, the normal coarse texture characteristic of the early metamorphism (Chapter 5) has been cut by millimetre thick zones of fine-grained granulitisation approximately parallel to the earlier schistosity. It is similar in texture to the rocks showing coarsened schistosity, but more intensively developed. Biotite/chlorite proportions are unaffected, but all the minerals present, including chloritoid, suffer cataclasis, kinking and bending. Late crinkle folds result from the phyllonitisation. Easterly trending folds overturned to the north are more common than northerly trending folds overturned to the east. This schistosity is considered a more advanced development of the phyllonitisation described above and to be the consequence of deformation associated with the third metamorphism (Read, 1937; (Figure 14); Chapter 5).

Structure of the Saxa Vord Block

Data from the Saxa Vord Block are presented in (Figure 15). Despite the magnificent cliff sections in the north and west of the Saxa Vord Block, the internal structure of the block is not easily determined. This is due in part to the paucity of inland exposure and to the lithological homogeneity of the phyllosilicate-rich lithologies of the Saxa Vord Pelite and Hevda Phyllite in the east of the block. A contrasting structural style is however apparent in the isoclinal intrafolial folds of bedded quartzite of the Queyhouse Flags in the west. In the east, crenulation microfolding is widespread and commonly leads to the development of secondary spaced cleavages and axial plane cleavages in minor fold closures. L-tectonites result from the development of minor folds with parallel axes, but irregular cross-section.

Structures in the Queyhouse Flags

In general terms, the structure of the Queyhouse Flags is dominated by an S-tectonite structure rather than the L-tectonite structure developed by the rocks to the east. The stereographical projection of (Figure 15)a shows a point concentration of poles to bedding (S0), schistosity (S1) and fold axial planes along with a partial girdle of fold axes, fluting (rodding) and mineral lineations normal to the point concentration.

The foliation of the Queyhouse Flags, as seen in the cliff along the east side of Burra Firth [HP 625 158], dips 40–50° to the east. In the cliff, a number of quartzite beds are seen to be folded into intrafolial isoclinal ‘hairpin- type’ folds (Plate 2a); (P533121), (P533127), (P533128) while others form intrafolial isoclinal folds of several quartzite beds (P533224), (P533233). As displayed on the stereoplot (Figure 15)a the fold axes lie within the foliation as the common axial plane, but scattered on azimuths plunging between north-north-east and south-east. The silty laminated and banded layers enclosed between quartzite layers are crumpled into minor folds (P533994), (P533998), which locally develop a strain-slip spaced cleavage [HP 617 142]. The isoclinal folds are intrafolial with respect to schistosity and bedding. This folding does not repeat the lithology of the Queyhouse Flags at a formation scale, merely individual beds or small groups of beds. Isoclinal intrafolial folds of similar geometry and orientation are exposed in more accessible exposures of this formation along the eastern shores of Loch of Cliff to the south [HP 604 128]. These folds verge both south and north.

Superimposed monoclinal folds along the cliff face of Burra Firth occur, as for example at Hols Hellier [HP 628 177]. These structures verge both north and south on east-plunging axes and step the strata down by as much as half the height of the cliff (100 m) in several places. The Stabba Pelite displays commonly developed dominantly south-east-vergent crenulation folds on north- east and north-north-east axes. This crenulation folding is associated with some development of spaced cleavage.

Structures in the Saxa Vord Pelite and Hevda Phyllite

The primary schistosity observed in the phyllosilicate-rich lithologies of the Saxa Vord Pelite and Hevda Phyllite is invariably parallel to original sedimentary layering (S0) where such layering is visible. Linear structures show a more compact distribution than in the Queyhouse Flags. Thus the stereonet for the eastern half of the Saxa Vord Block presents an L-tectonite appearance, with the lineations and fold axes forming a point concentration and the bedding surfaces, axial planes and schistosities forming a partial girdle (Figure 15)b, the reverse of that for the western half. The mean orientation of this lineation has a plunge of 40° towards 060°N, parallel to the L-tectonite lineation in the Skaw Granite (see relevant section of Chapter 8). It is possible to see large- scale folds with this orientation on the north coast, albeit with difficulty due to the homogeneity of these phyllosilicate-rich lithologies. Very large (cliff-sized) rounded open folds with interlimb angles of c. 90° can be seen in the Saxa Vord Pelite when looking north from the cliff top south of The Noup [HP 634 178] and from the beach in Brei Wick, at Backberg [HP 635 175].

Away from the cliff section, large-scale folding is not generally observable within either the Saxa Vord Pelite or Hevda Phyllite. This is only partly due to lack of exposure but also because the Saxa Vord Pelite has been rendered extremely fissile in the vicinity of the Hevda Thrust. In addition, the pelitic rocks are progressively recrystallised to the east in response to the thermal metamorphic effects of emplacement of the Skaw Granite. One notable exception to this general pattern is the structure delineated by the magnetic anomaly pattern detected in the Saxa Vord Pelite near Ungirsta [HP 624 134] (Figure 11).

The Hevda Thrust

The Hevda Thrust is identified by the occurrence of zones of phyllonitic rock; these are particularly well displayed in the Saxa Vord Pelite in the Burn of Skaw between [HP 643 160] and [HP 645 163]. The thrust is also exposed on the cliff face to the north-east of Ritten Hamar [HP 643 173] at which point the structure has apparently cut upsection so that the intensely fissile phyllonitic rocks at this location are derived from the Hevda Phyllite. Scattered occurrences of easterly- vergent kink folding on an approximately horizontal axis are associated with the phyllonitic rocks. Such kink folding is relatively rare elsewhere in the Saxa Vord Block.

Chapter 5 Metamorphism

Introduction

The Valla Field and Lamb Hoga blocks are formed of three lithologically very different units. These were recognised in a classic work by Read (1934a) as the Westing, Valla Field and Burra Firth groups; a series of successively structurally overlying units. Read did not specify whether they are a stratigraphical or a tectonostratigraphical succession, but made it clear that the three units have been subjected to the same metamorphism. He based this metamorphism on that of the upper part of his ‘Valla Field Group’ (i.e. the Valla Field Schist in the terms used here) and distinguished three successive episodes of metamorphism. The Lamb Hoga Block was not considered by Read, but the same three lithological units and the same metamorphisms are present there also.

Since Read’s time, it has become clear that the Westing Group is a tectonically emplaced assemblage of tectonostratigraphically distinct units; the Boundary Zone. This was not apparent to Read because the boundaries of the Boundary Zone are not obvious faults, mylonite zones or dislocations separating the groups, but are apparently metamorphically and crystographically intact contacts. Since its emplacement, the Boundary Zone has been subjected to Read’s tripartite metamorphism; the emplacement of the Boundary Zone therefore predates his first metamorphism. Prior to that time, the constituent lenses of the Boundary Zone had their own metamorphic histories and possibly a common Boundary Zone metamorphism.

The Saxa Vord Block was mapped and interpreted by Read (1934a). It was subsequently studied and interpreted by Snelling (1958) and also mapped and interpreted by Key (1972). All three came to the conclusion that it had been subjected to three metamorphisms: a kyanite-staurolite-garnet regional metamorphism, a later retrograde kyanite-chloritoid metamorphism and a dislocation chlorite metamorphism. This reflects the influence of Read’s three metamorphisms of the Valla Field Schist. Examination of the evidence presented makes it clear that their first metamorphisms are based on evidence collected from the Summit Pelite (see below) and the second metamorphism was interpreted without appreciating the influence of thermal metamorphism by the Skaw Granite. A complete remapping and examination of the area lead to an alternative interpretation.

Metamorphism in the Valla Field and Lamb Hoga Blocks

Metamorphism of the Valla Field Schist

The schist is a coarsely crystalline pelitic schist of uniformly Al-rich composition which extends from north to south in the centre of the Valla Field Block. It passes from a domain of kyanite-staurolite- garnet metamorphism in the north, designated the first episode or metamorphism by Read (1937), to a domain of chloritoid metamorphism in the south of the block, which he designated the second episode or metamorphism. The Valla Field Schist continues to the south of Unst in the Lamb Hoga Block where it lies within the domain of the second metamorphism.

The ‘first metamorphism’

The ‘first metamorphism’ is best displayed in the region of Libbers Hill [HP 586 137] though it is well developed in Tonga [HP 585 150] to the north and continues for several kilometres to the south before the onset of the second metamorphism (Figure 14). The minerals are very well described by Read (1937) and have been redescribed by Aziz (1984) and for this account, the descriptions are based on more than 1000 thin sections collected from the Valla Field Block and from field observation.

Sillimanite appears to be the oldest mineral but occurs only as fibrolite in other minerals and is almost invariably sericitised or partly sericitised. It occurs in garnet, staurolite, chloritoid, kyanite, tourmaline and late muscovite as single fibres, or very rarely as sericitised groups of needles (faserkieseln).

Kyanite and staurolite usually occur together as major constituents, forming crystals 2 cm in size. From place to place, they range from kyanite only to staurolite only. Together or alone, they locally form three quarters of the whole rock, or more. Lensoidal kyanite segregations, 10 cm thick or more, are common in the Tonga [HP 585 150] to Wood Wick [HP 590 110] areas. Individual grains range in size from millimetres to two centimetres or more, and they vary in shape from irregular shaped grains to crystals. Staurolite commonly contains inclusions of kyanite, garnet and fibrolite, but kyanite less commonly contains inclusions of staurolite, garnet and fibrolite. The schist was drilled in a search for economic quantities of kyanite (Chapter 12).

Garnet occurs throughout the Valla Field Schist. It is very variable in appearance in thin section, due to different inclusion patterns of the minerals kyanite, fibrolite, and less frequently staurolite. Aziz (1984) recognised a series of garnets with different characteristic patterns of inclusions of these minerals. The inclusions, fracturing and replacement by chlorite give the impression of a long and complicated history of garnet development.

Mica flakes are commonly 2 mm long. Biotite is less common than muscovite in the Valla Field Schist, partly due to being accompanied by chlorite and partly due to retrogression to chlorite. As shown by (Figure 16), biotites are divided into those that are reddish brown and those that are brown to faintly greenish brown. The two colours are more closely related to differences in contents of total Fe and Mg than to Ti content.

Read (1934a) attributed chlorite in the Valla Field Schist to retrograde metamorphism, but in the domain of the first metamorphism much chlorite occurs in equilibrium with the biotite. In the Valla Field Schist, there is more than six times as many Al cations as K, leading to a shortage of K to form biotite.

Muscovite occurs in two forms. It occurs as flakes similar in size and form to biotite and chlorite flakes, and interspersed and approximately aligned with them to form the schistosity. The second type of muscovite occurs as large equidimensional plates, or very fat flakes, called late muscovites. They are commonly associated with shimmer aggregate (very fine grained white mica) formed from fibrolite, and more often than not contain relict fibrolite fibres and more rarely staurolite and garnet. The distribution of late muscovite is almost the same as that of fibrolite.

Plagioclase occurs as albite porphyroblast grains several millimetres in size in the highest grade area of the Valla Field Schist between Tonga [HP 585 150] and Wood Wick [HP 590 110]. They are commonly partially sieved by swarms of drop-like opaque grains, and are distinctly different from matrix plagioclase grains. They may form from the breakdown of paragonite in these highly recrystallised rocks.

The ‘second metamorphism’

The domain of the ‘second metamorphism’ is apparently confined to the outcrop of the Valla Field Schist (within the Valla Field Group) both at the south end of the Valla Field Block (Figure 14) and within the Lamb Hoga Block. Its northern limit, south of Clammel Knowes at [HP 583 060], is preceded to the north by some 3 km of phyllonitisation (Chapter 4). Read considered the second metamorphism to be a retrograde metamorphism. It is characterised by the minerals chloritoid and chlorite.

Chloritoid forms strongly twinned grains, up to three centimetres or more in size. They contain inclusions of garnet, staurolite, fibrolite and occasionally kyanite. Although they enclose and grow in contact with staurolite, more often staurolite grains occur free of chloritoid, which shows more affinity for garnet than staurolite, especially by filling networks of cracks in garnet. It is possible it replaces chlorite filling the cracks in garnet. Chloritoid shows no preference for contact with other minerals. It occurs also in tectonic slices of Valla Field Schist within the Lower Imbricate Zone of the ophiolite-complex.

No biotite occurs within the Valla Field Schist in the domain of the ‘second metamorphism’. It has been replaced by chlorite, except as inclusions in garnet. The minerals muscovite, late muscovite, kyanite, staurolite, garnet and fibrolite all occur throughout the domain of the second metamorphism, though in somewhat reduced quantities compared to the Libbers Hill area [HP 586 137] in Unst. Aziz (1984) found a difference in composition between the garnets in Unst and those in Fetlar in the Valla Field Schist.

Read’s three metamorphisms (1934a, 1937)

Read recognised the existence of three metamorphisms in the Valla Field Block. Two have been described above; the third was a ‘dislocation’ metamorphism. This was associated with the cataclastic and mechanical distortion effects coupled with the development of chlorite along the contact with the ophiolite-complex and produced by movement on faults and thrusts.

Read considered that the first episode, or metamorphism, extended over the whole Valla Field Block and was manifested by different minerals in different lithologies (Read, 1937). The second episode occurred later, extended over the southern part of the block, and was represented differently in the different lithologies in the block. The third metamorphism was a still later episode and was distributed along the boundaries of the block. It is now recognised also to occur within the block wherever faults occur. Read recognised the second metamorphism as a retrogressive episode and because of its association with the development of chlorite and the fortuitous presence of areas of third metamorphism, considered it to be both retrogressive and mechanically destructive.

Read’s ‘metamorphic correlation’ (Read, 1937)

Read (1937) extended the first and second meta- morphisms over the very different lithologies of the Burra Firth Formation to the east and the Westing Group (Boundary Zone) rocks to the west, because he found no evidence of faulting or thrusting along the contacts of those units. This survey has confirmed this judgement. In particular, the contact with the Boundary Zone shows no evidence of disruption, confirming that the first metamorphism postdates the emplacement of the Boundary Zone. Read sought to evaluate the effects of the first and second metamorphisms in the Valla Field Schist on the very different lithologies on either side. He called this ‘metamorphic correlation’.

Metamorphism of the Burra Firth Formation

The Valla Field Schist stratigraphically and conformably underlies the Burra Firth Formation, but the metamorphic state of the two units is very different. While the Valla Field Schist is apparently of kyanite or possibly sillimanite grade, the Burra Firth Formation is of low garnet grade judging by infrequent small garnets in the pelitic-psammitic bands alternating with the quartzitic-psammitic bands. No sillimanite, kyanite or staurolite was found anywhere in the Burra Firth Formation. This same metamorphic difference is also shown by the correlated rocks in the Mainland of Shetland to the south. It is suggested that the Al-rich pelite has developed an anomalously high grade (Flinn et al., 1996).

Metamorphism of the Boundary Zone (Westing Group)

Most of the rocks in the Boundary Zone between Unst and Yell are lithologically incapable of developing sillimanite, kyanite or staurolite and mineralogically resemble the Precambrian rocks forming the protolith of the Westing Ultramafic Zone, which are primarily composed of brown hornblende. Their widespread recrystallisation to green hornblende could be due to a Precambrian recrystallisation prior to the Cambrian blastomylonitisation (Roddam et al., 1994a) .The Westing Group units are cut by a suite of numerous pegmatites of a variety of ages. Few pegmatites cut the Valla Field Group. Thus the Westing Group is considered to be composed of rocks which predated, both in origin and metamorphism, the emplacement of the Boundary Zone.

The nebulite lens south-east of Yell, extending into Lunna Ness [HP 520 174] and the Kirkaby and Orknagable formations, is equally unlikely to show visible signs of superimposed metamorphic effects of the ‘first metamorphism’. The Snarravoe Pelite of the Westing Gneiss is the most likely unit in the Westing Group to have been metamorphosed by the ‘first and second metamorphisms’. It is an Al-rich pelite and in Unst, Burra Ness [HU 558 955] and Hascosay [HU 563 931] has developed large plates of chloritoid. In Unst north of the Lund Fault [HP 573 036], it includes kyanite and staurolite besides chloritoid, but to the south in Burra Ness and Hascosay, only chloritoid, chlorite, shimmer aggregate and muscovite occur. No sillimanite occurs. Paragonite is indistinguishable from muscovite in thin section. Key (1972) X-rayed 5 samples of Snarravoe Pelite from the type area [HP 563 022] and found four containing paragonite.

The Westing Limestone is another unit of the Westing Group which has probably been influenced by the ‘first and second metamorphisms’. Read (1937) provided a very long list of the minerals that he found in the Westing Limestone, of which diopside is the indicator of highest grade. The calcsilicate granofels in all Westing Limestone occurrences contain diopside, and south of the Lamb Hoga Fault contain diopside with microcline. Zoisite and colourless actinolitic amphibole are very common in the domain of the ‘second metamorphism’ and the latter often replaces diopside. Many amphibolitic lenses in the Westing Limestone contain brown hornblende; an indicator of very high grade in the district.

The Westing Gneiss is a dark coloured (because biotite-rich) garnetiferous red-brown biotite > muscovite psammitic granofels. It varies locally with increasing muscovite content to kyanite- and/or staurolite-bearing pelitic schist. These rocks are widely and variably, but never very strongly, gneissified. Calcareous layers in the Westing Gneiss occur together with pods and lenticular layers of amphibolite, some of which contain reddish-brown hornblende, and are weakly agmatised (Plate 1d). The kyanite and staurolite may be indicative of the effect of the ‘first metamorphism’, but the gneissification and agmatisation are considered to be older than the emplacement of the Boundary Zone.

The Valla Field Gneiss (see footnote 3 in Chapter 2) is less strongly micaceous, and more psammitic and granofelsic, than the Valla Field Schist and strongly and consistently gneissified. The rocks vary to pelitic in places, with micas and chlorite between 1 and 2 mm long, and staurolites and kyanite. Garnet is variable in size but rarely reaches a centimetre in diameter. Near the Westing Limestone, the Valla Field Gneiss contains scattered fish and lenses of amphibolite up to a metre thick, but locally reaching 4 m and in one place 20 m thick. In the Lamb Hoga Block, the Valla Field Gneiss overlies the Valla Field Schist and dips to the south-west into the sea along the coastline in the Ness of Snabrough [HU 580 925] and Rams Ness [HU 605 875] areas. The Valla Field Gneiss at Rams Ness is remarkable among the Westing Group rocks for its prolific development of fibrolite. The Valla Field Gneiss matches the apparent metamorphic grade of the Valla Field Schist, but the inclusion of amphibolite lenses and the local development of plagioclase microporphyroblast gneiss makes it very likely that its primary metamorphism occurred prior to the emplacement of the Boundary Zone. (This precludes the Valla Field Gneiss being in the Valla Field Group, and instead contrary to the published map (BGS, 2002) warrants its inclusion within the Boundary Zone, — see footnote 3 in Chapter 2).

Discussion

The Valla Field Group has been correlated with the Scatsta ‘Division’ in the Mainland of Shetland (IGS, 1981) both in respect of the lithology and the metamorphism. The minerals in the Al-rich correlate of the Valla Field Schist are the same, except for the fibrolite inclusions, as are the minerals in the correlate of the Burra Firth Formation. The metamorphism of the Scatsta ‘Division’ is considered to be the first suffered by the Dalradian of Shetland and to have been the regional metamorphism. Therefore, the ‘first metamorphism’ of Read (1934a, 1937) is also the first metamorphism of the Dalradian of Shetland.

Read considered his ‘second metamorphism’ to be retrograde. However, he visited Fetlar on only one day (and that was contrary to his instructions). In Fetlar, the ‘second metamorphism’ does not conform to his ideal model. It is suggested here that the ‘second metamorphism’ is retrograde in the sense that it is of lower grade than the ‘first’, but that it is prograde because the rocks had cooled subsequently to the ‘first metamorphism’ and for the ‘second metamorphism’ were later raised in temperature, but only to that appropriate to crystallisation of chloritoid. This could allow the preservation of fibrolite, kyanite and staurolite as apparently stable minerals. The rise in temperature may have accompanied the emplacement of the ophiolite-complex and resultant depression of the crust. The Lamb Hoga Fault has a downthrow to the north of at least three kilometres so that the lower limb of the east-facing recumbent Valla Field Fold is expected to continue beneath Unst in the metamorphic state it has in the Lamb Hoga Block. Some 25 kilometres south of the Lamb Hoga Block, in the Scatsta ‘Division’ in the island of Whalsay, the correlate of the Valla Field Schist contains large plates of chloritoid as in the Lamb Hoga Block. If the ophiolite-complex is the cause of Read’s ‘second metamorphism’, then prior to its erosion it continued at least as far south as Whalsay.

Metamorphism in the Saxa Vord Block

The Saxa Vord Block is formed of three lithologically contrasted units: the Queyhouse Flags, the Saxa Vord Pelite and the Hevda Phyllite (Figure 11).

Queyhouse Flags

The Queyhouse Flags are metamorphically of low grade judging by the silty to schistose pelitic and semipelitic bands alternating with the quartzite beds (Plate 2b), (Plate 2c), (Plate 2d). These include both black (biotitic) and silvery (muscovitic) bands with no apparent garnet. The Queyhouse Flags contain what appear to be metamorphically higher grade enclaves (Figure 17). The Stabba Pelite (Figure 11) is a schist composed of 2 mm muscovite flakes and containing staurolite, often schimmerised and distorted by later microfolding, together with garnet less than 1 mm in diameter, both of which cut the schistosity. The Summit Pelite is a schistose to phyllitic pelite composed of c.0.5 mm muscovites, formed from both Al-rich and normal pelite (Figure 12). It contains chloritoid, garnet, kyanite and staurolite as well as completely schimmerised rectangular andalusite. Read (1934a, 1937) accepted the aggregates as having been formed after andalusite, whereas Snelling (1958) claimed to have found relict andalusite in one aggregate, and Key (1972) attributed the aggregates to staurolite). The chloritoid forms rectangular grains up to 0.5 mm aligned with the muscovite fabric. They are commonly included in garnet grains, which vary up to 1 mm in diameter. Kyanite and staurolite enclose garnet and are included by garnet. Kyanite in crystal form occurs in shimmerised andalusite. The Summit Pelite is considered to contain elements of three metamorphisms. It is a low grade (greenschist facies) chloritoid-muscovite-chlorite regional metamorphic phyllite, which previously suffered a thermal metamorphism producing the centimetre- sized andalusites. After the regional metamorphism, kyanite, staurolite and garnet grew without significant recrystallisation of the micas. The only intrusive rock capable of causing the crystallisation of andalusite within the Saxa Vord Block is exposed as a schistose two-feldspar-muscovite-biotite granite on the shore of Loch of Cliff [HP 604 133].

Saxa Vord Pelite

The Saxa Vord Pelite occurs several hundred metres east of the Summit Pelite (Figure 17). Its outcrop is divided into two by the Hevda Thrust, which produces a hiatus in the metamorphic grade. The western part is a regional metamorphic chloritoid phyllite closely resembling the phyllitic state of the Summit Pelite (Flinn et al., 1996). Rectangular chloritoid, up to 0.5 mm, occurs aligned with a fine-grained schistose muscovite-chlorite matrix. As in the Summit Pelite, centimetre-sized square to rectangular shimmer aggregates occur within the schistosity, but without any relics of andalusite in any of the 300 thin sections available. However, some contain crystal-shaped kyanite grains of later growth. Chloritoid, kyanite, staurolite and garnet grains all occur and are of obvious later growth, unlike the same minerals in the Summit Pelite. Chloritoid forms ragged grains up to 1 mm in size, cutting the schistosity and commonly overgrowing the regional metamorphic grains. Staurolite and kyanite occur as grains cutting the schistosity, the former being frequently partly, or completely, shimmerised in a coarser manner than the andalusite shimmer aggregates. Garnet was found in only 9 of the 60 thin sections available. They vary up to 2 mm in diameter and mostly contain rectilinear dusty inclusion patterns. The post-regional metamorphic minerals increase in frequency and development towards the Hevda Thrust, as expected of a thermal aureole caused by the Skaw Granite.

To the east of the Hevda Thrust, the phyllitic regional metamorphic fabric has been recrystallised to a schist and increases in grain size towards the granite contact. There is no sign of staurolite or its shimmer aggregates. Chloritoid has grown to up to a centimetre in size and exhibits polysynthetic twinning. Kyanite is more profuse than to the west and increases in size towards the granite contact, but disappears a hundred or so metres before it to be replaced by shimmerised fibrolite. Along the granite contact the rock is formed of masses of shimmerised faserkiesel fibrolite, overgrown by one to two centimetre-sized plates of chloritoid, late muscovite and chlorite. The latter three minerals must have crystallised as the temperature in the aureole fell.

Hevda Phyllite

The Hevda Phyllite is dominantly composed of massive chlorite > muscovite regional metamorphic phyllites. In the field, the only distinguishable minerals, other than chlorite and muscovite, are garnet, and white and dark spots (shimmer aggregates). Garnet occurs throughout the outcrop in 36 out of 50 thin sections, generally as grains no more than 1 mm in diameter. Several are zoned with a sharp separation of dusty core and clear rim. Some have rectilinear inclusion patterns, parallel to the schistosity. In the Hevda Ness area [HP 644 174] the dark spots in the rock (see above) contain very small garnets, often with moderately good crystal forms. Some are the only garnets in the rock. Staurolite was found in only one rock.

Although the Skaw Granite does not cut the Hevda Phyllite, it does cut the Lower Imbricate Zone of the ophiolite-complex where it is formed by a tectonic lens of Hevda Phyllite-like rock [HP 653 150]. In contrast to the Saxa Vord Pelite, the thermal effects of the granite on this rock are small. In a 200 m by 10 m zone along the granite contact, sparsely distributed fibrolite fibres and groups of fibres, generally shimmerised, were found in a dozen or so places. Several of the available thin sections contain late muscovite. Garnet occurs in 11 out of 40 thin sections, mostly as grains up to 0.3 mm across, often with crystal forms and dusty cores, and cutting the schistosity. In the south part of the lens, garnet (less than 0.1 mm diameter) was found in only 2 out of 60 thin sections.

Discussion

Andalusite apparently crystallised prior to the regional metamorphism of the Saxa Vord Pelite because each of the shimmer aggregates lie enclosed within the phyllite schistosity and deform it. There is no occurrence of andalusite associated with the Skaw Granite thermal metamorphism. The occurrence of centimetre-sized crystals, situated as remotely from the granite contact as 2 km and beyond the other recognisable thermal minerals, precludes the precursor of the shimmer aggregates as being due to thermal metamorphism by the Skaw Granite. The absence of andalusite from the aureole of the Skaw Granite indicates that the intrusion took place at a depth greater than that equivalent to the pressure of the aluminosilicate triple point, i.e. approximately 4 kbar, and equivalent to a depth greater than 12 km.

The extent of thermal metamorphism due to the Skaw Granite in the Saxa Vord Phyllite is less than 2 km, and at that distance has only a minor effect. Therefore, it seems unlikely that the post-regional metamorphism of the Summit Pelite is attributable to the Skaw Granite.

In both the Saxa Vord and Valla Field blocks, a gross difference has been shown to occur between Al-rich chloritoid-bearing pelitic rocks and Mg-rich pelitic rocks. The former crystallise to coarser-grained rocks and develop minerals of apparently higher grade than the Mg-rich pelites. Such an effect was noted above in the case of the Valla Field Schist and the Burra Firth Formation. It was also noted by Flinn et al. (1996) in the correlated rocks in the Mainland of Shetland to the south (Chapter 3). There, regional metamorphic chloritoid phyllites, closely resembling the chloritoid phyllites in the Saxa Vord Pelite, have been thermally metamorphosed by the Spiggie Granite [HU376 182] to coarsely crystalline rocks mineralogically the same as those in Saxa Vord Block. The lithological and metamorphic correlates of the Saxa Vord Pelite and the Queyhouse Flags in the Clift Hills ‘Division’ experienced the first regional metamorphism of the Shetland Dalradian. The rocks of the Saxa Vord Block are considered to be have been metamorphosed by the same regional metamorphism.

Chapter 6 The Shetland Ophiolite-complex

Introduction

The Shetland Ophiolite-complex consists of two nappes of ophiolite (the Lower Nappe and Upper Nappe; Flinn, 1958), beneath each of which is a zone of imbricated tectonic slices (the Lower Imbricate Zone and Middle Imbricate Zone, respectively) of metasedimentary rock, ophiolite, acid and basic metavolcanic rock and hornblende schist. The ophiolite succession is lying on its side in the Lower Nappe (Figure 18), exposing a layered sequence, originally of harzburgite, dunite and gabbro. Each layer is several kilometres thick, near vertical, and striking approximately north-east. The sequence is arranged with the outcrop of the formerly deepest layer in the west and that of the shallowest layer in the east (Figure 4) and (Figure 6). Nearly all of the ophiolite has suffered intense alteration; the ultrabasic units are thoroughly serpentinised, and the basic units are amphibolitised and saussuritised. To reflect this widespread change, the term ‘meta’ is prefixed to the name of each lithology (e.g. metaharzburgite, metadunite).

From west to east the ophiolite layers are: Metaharzburgite Layer; Metadunite Layer; and Metagabbro Layer. Bodies up to kilometre scale of generally unaltered clinopyroxenite-wehrlite and of metadunite crop out throughout the Metadunite Layer and Metagabbro Layer. The Metagabbro Layer has petrologically distinctive upper and lower parts; these are referred to as the Upper Metagabbro and Lower Metagabbro. Numerous sheets of mainly basic rock intrude the Upper Metagabbro. These are generally not ‘sheeted’ (i.e. they are not contiguous, but typically separated by country rock), but they have similarities to a sheeted dyke-swarm and are therefore referred to here as the Quasi-sheeted Dyke-swarm. The Metaharzburgite Layer represents infertile mantle. The boundary between it and the Metadunite Layer is interpreted as the petrological Moho. The boundary between the Metadunite Layer and the Metagabbro Layer is interpreted as the geophysical or seismic Moho.

The Upper Nappe is largely eroded, and consists now of a number of relatively small klippen that overlie the Lower Nappe and Middle Imbricate Zone. The klippen consist almost entirely of metaharzburgite, which is indistinguishable in exposures from that forming the Metaharzburgite Layer of the Lower Nappe. The petrological Moho is not exposed in the Upper Nappe.

The Lower Imbricate Zone in Unst crops out in a long, thin, continuous strip adjacent to the western edge of the ophiolite, separating the ophiolite from underlying metamorphic units in the basement. Units within it consist mainly of tectonic slices of the Saxa Vord and Valla Field groups, and of hornblende schist (Norwick Hornblendic Schist). The Middle Imbricate Zone is largely eroded. However, large isolated remnants crop out in several localities; the largest are in north-east and south-east Unst, and in central and east Fetlar. The Middle Imbricate Zone consists mainly of tectonic slices derived from several units of metasedimentary rock, which are assigned to the Unst Phyllite Group.

The ophiolite-complex overlies Dalradian basement (the Valla Field, Lamb Hoga and Saxa Vord blocks) to the west and north. To the south it is truncated by the Lamb Hoga Fault, across which it is juxtaposed with the Lamb Hoga Block. To the east, where it passes under the sea, it is possible to follow the boundaries of serpentinised units with the aid of aeromagnetic data (Chapter 10).

The Ophiolite Nappes

Distribution and structure

Lower Nappe

The Lower Nappe (Figure 6) and (Figure 18) is exposed widely in the eastern part of Unst, being overlain only by metasedimentary rock of the Middle Imbricate Zone in the Norwick [HP 645 146] and Muness [HP 615 017] areas and by several relatively small klippen of the Upper Nappe. To the west, it is bounded by its inferred obduction thrust, across which it is in contact with the Lower Imbricate Zone. Its outcrop continues south of Unst, through the western half of Uyea [HU 600 990] and the small islands of Urie Lingey [HU 595 955] and Daaey [HU 603 950], and into north-west Fetlar. It is well exposed along the north-facing shore of Fetlar, but between there and the Lamb Hoga Fault to the south it is much obscured by units of the Middle Imbricate Zone and klippen of the Upper Nappe. The Lamb Hoga Fault truncates it to the south. A small area of the Lower Nappe is exposed in Hesta Ness [HU 664 925] in north-east Fetlar.

The thrust that underlies the Lower Nappe crops out along the east side of the central valley in Unst, between Belmont [HP 565 010] and Ungirsta [HP 623 130]. It underlies a west-facing escarpment rising above the outcrop of the Lower Imbricate Zone, and is accompanied to the west by a narrow, negative ground-magnetic anomaly (Chapter 10). North of Ungirsta, the trace of the thrust passes north-east to Nor Wick [HP 655 145], separating the Lower Imbricate Zone from the Middle Imbricate Zone. North-east of Ungirsta, the Lower Nappe is concealed beneath units of metasedimentary rock in the Middle Imbricate Zone; as a consequence, the trace of its boundary runs east from Ungirsta, and it loses the escarpment and negative ground-magnetic anomaly.

Upper Nappe

The Upper Nappe (Figure 6) and (Figure 18) crops out only as relatively small klippen, with segments of the inferred obduction thrust as their basal thrusts. The klippen rest directly on the Lower Nappe and rocks of the Middle Imbricate Zone. The largest klippe, at Vord Hill [HU 622 935] in central Fetlar, has a synformal base resting on rocks of the Middle Imbricate Zone. A large aeromagnetic anomaly links the Vord Hill Klippe to the lithologically similar island of Haaf Gruney [HU 635 983], which lies some 3.5 km north of Fetlar. A short distance farther north is the klippe-like mass of serpentinised rock at Mu Ness [HP 637 010]. However, this is serpentinised clinopyroxenite-wehrlite and is therefore likely to be a tectonic slice in the Middle Imbricate Zone. West of the Vord Hill Klippe the basal thrust flattens and forms the lower boundary to thin klippen capping several hill tops: at Hamara Field [HU 600 930], Northdale [HU 607 920], Southdale [HU 600 914] and Oddsta [HU 587 937]. A positive aeromagnetic anomaly over Sound Gruney [HU 580 962] is probably a result of the basal thrust dipping beneath the island, making it either a klippe of the Upper Nappe or, more likely, a tectonic slice of ophiolite (clinopyroxenite-wehrlite) immediately underlying the Upper Nappe (like that at Mu Ness). On the east side of Fetlar, the thrust below the Vord Hill Klippe reaches the surface a short distance (up to 250 m) west of the Aithbank Fault. Ground east of the Aithbank Fault is downthrown, and the Upper Nappe crops out again on its east side.

The Hill of Clibberswick [HP 658 130] in north-east Unst is a klippe of the Upper Nappe resting on a steeply east-dipping thrust marked by a strong negative ground- magnetic anomaly (Chapter 10) and a poorly defined, west-facing escarpment. Aeromagnetic evidence (Chapter 10) suggests this klippe extends offshore for about 1 km to the north, 1.5 km to the east, and more than 4 km to the south-south-east. A small, positive aeromagnetic anomaly over a submerged high point on the sea floor at [HP 648 043] may mark the position of another small klippe of the Upper Nappe. This feature and the klippen of Hill of Clibberswick, Vord Hill, Haaf Gruney, and Muness form a broadly north-north-east- trending zone, which may be the result of a synformal downfold of the base of the Upper Nappe along a north- north-east-trending axis.

The Muness Klippe is separated from underlying metagabbro in the Lower Nappe by a thrust plane that is exposed clearly and accessed easily in many places. The thrust plane is flat-lying east of Springfield at [HP 638 009]. However, near Taing of Noustigarth [HP 637 005] it forms a vertical, isoclinal synform in the cliffs, with the closure only just above sea level. The synform has a northerly axial trend and amplitude of about 10 metres. Shearing parallel to the thrust is displayed very well in both the serpentinite and the underlying metagabbro of the Lower Nappe. The shear zone includes a thin lens of blackish phyllite [HP 6388 0068], which resembles phyllite in units assigned to the Unst Phyllite Group.

The Gallow Hill Klippe [HP 570 010], the Watlee Klippe [HP 600 055], the Caldback Klippe [HP 609 068], and possibly the Sound Gruney Klippe [HU 580 960], lie along the extreme western edge of, and directly on, the Lower Nappe. They may be remnants of a north-north-east- trending downfold of the base of the Upper Nappe, which is broadly parallel to a similar feature on the east side of Unst that was postulated above.

Metaharzburgite Layer

A section several kilometres thick through the Metaharzburgite Layer crops out at the northern end of the Lower Nappe in Unst, but this thins significantly farther south and the layer is excised entirely by the basal thrust. North of Balta Sound the boundary between the Metaharzburgite and Metadunite layers strikes approximately east–west and dips to the north at 70° or more. The boundary can be followed easily from the coast on the north side of Wick of Hagdale [HP 645 106] through inland exposures to the Gossa Dale Fault [HP 626 104]. The boundary between the Metaharzburgite and Metadunite layers is not detectable west and south of the Gossa Dale Fault. Close to the western edge of the Lower Nappe, extensive antigorite serpentinisation, locally poor exposure, and klippen of the Upper Nappe make it very difficult to distinguish between metaharzburgite and metadunite (especially clinopyroxene-bearing metadunite). It is unclear whether observed boundaries represent the petrological Moho or discrete masses of metaharzburgite in the Metadunite Layer.

The Metaharzburgite Layer consists of harzburgite that has been strongly, pervasively serpentinised; samples rarely contain more than 10 per cent by volume of unaltered primary minerals. Of 180 thin sections of metaharzburgite, 80 contained no primary minerals other than chromite. Alteration products and occasional unaltered relics confirm that the metaharzburgite was composed dominantly of olivine before it was serpentinised. Twenty-three of 180 thin sections of metaharzburgite contain clinopyroxene, all of it very fresh; two samples contain sufficient clinopyroxene to be classified as metalherzolite. In the region of The Punds [HP 6446 1108], cleavage planes of clinopyroxene can be seen glittering in the sun. Crystals of orthopyroxene, like olivine, have been at least partly altered; only 7 of 180 thin sections contain reasonably fresh orthopyroxene relics. Crystals of olivine and pyroxene are coarse- grained (c. 5 mm). Crystals of chromite are largely fresh, but some have black altered rims with or without a fringe of chlorite (see below).

Averaged values for available whole-rock analyses of metaharzburgite are presented in (Table 2). The averaged composition of metaharzburgite samples recalculated to a dry base is approximately equivalent to a volume mode of 81 per cent olivine, 17 per cent orthopyroxene and 2 per cent clinopyroxene (calculated on the basis of SiO₂ content, (Table 2)). These figures concur with those for the primary mineral mode derived by point counting (Moffat, 1987): olivine 74–92 per cent, ortho- pyroxene 7–26 per cent, clinopyroxene 0–3 per cent, and chromite 0.25–1.5 per cent. Generally low values for orthopyroxene may reflect loss of silica during serpentinisation. The mean Mg# (MgO/[MgO + FeO]) in 61 whole-rock samples of metaharzburgite is 82 (see note on (Table 3)).

Rhythmic modal banding (Plate 3a) is developed close to the upper boundary of the layer in ground south of Haroldswick [HP 637 123]. The bands are sharply bounded, rectilinear and dip steeply northwards; the near east–west strike is parallel to the petrological Moho. They are formed by alternations in the relative proportions of olivine and pyroxene, and are ungraded. A typical pair of olivine- rich and pyroxene-rich bands is about 30 cm thick, and can rarely be followed for more than a few metres. The bands rarely show signs of disruption or folding, and are untectonised. Banding is developed best in a zone several hundred metres wide on the north side of the boundary between the Metaharzburgite and Metadunite layers, between the coast north of Wick of Hagdale [HP 644 106] and Nikka Vord [HP 628 105]. Farther north, well-developed localised patches of banded rock become increasingly rare.

Several areas of very poorly defined rhythmic banding are exposed in the Upper Nappe. In the Hill of Clibberswick Klippe [HP 658 130] the bands are orientated c. 045º/90º, while north-east of Vord Hill the orientation is c. 135º/70º E.

Bodies of metadunite are distributed sporadically within the Metaharzburgite Layer, and become more common towards the boundary with the Metadunite Layer. The mean Mg# (MgO/[MgO + FeO]) in 21 whole-rock samples of the metadunite bodies is 85 (see note on (Table 3)). A few of the larger bodies contain ‘podiform’<span data-type="footnote">Podiform bodies of chromite are said to occur in the mantle layers of most ophiolites. In the Shetland Ophiolite such chromite bodies are far from podiform in shape and vary from pancakes to vein stockworks. For this reason, the term ‘podiform’ will be referred to in inverted commas.</span> masses of coarsely-crystalline chromitite. Most metadunite bodies in the northern part of the outcrop are sub-equidimensional, 1 metre or so in diameter, with no discernible preferred orientation. Larger bodies, up to 100 metres or more across, commonly have a very irregular shape. West of the Gossa Dale Fault the metaharzburgite lacks rhythmic banding, which is a feature east of the fault, and metadunite bodies within it are much less common and less ordered; these observations suggest displacement on the fault.

Sheet-like bodies of metadunite increase in proportion relative to more equidimensional bodies towards the Metadunite Layer. Within several hundred metres of the petrological Moho, in the rhythmically banded zone (P533580), metadunite sheets up to 20 cm thick and several metres long are locally as voluminous as the host metaharzburgite. Sheets of metadunite in the Metaharzburgite Layer adjacent to the petrological Moho are subparallel to rhythmic banding and the Moho (Plate 3a). A high proportion cut their host at a low angle (10º or less), while a smaller proportion cut it at higher angles, for example north-east of Muckle Heog ( [HP 6363 1134], [HP 6373 1124]), or are parallel ((P533680), (P533717), (P533788)). The small bodies of metadunite, especially those close to the petrological Moho, commonly have thin, wedge-like and sheet-like protuberances that cut the enclosing metaharzburgite in an intrusive manner.

Widely and sparsely scattered bodies of metadunite also occur in metaharzburgite of the Upper Nappe. In the Hill of Clibberswick Klippe they range from 1 to several tens of metres in diameter and have a roughly equidimensional outcrop. The largest mass, 100 m in diameter, crops out at Athans [HP 660 115] and encloses several apparently detached masses of metaharzburgite. The Vord Hill Klippe contains widely scattered bodies of metadunite, including several large, irregular masses south-west of Vord Hill [HU 613 933] and near Swart Houll [HU 643 916]. Parallel sheets of metadunite a few metres wide and up to 100 metres or more long crop out in a partly defined arc to the south, west and north- west of Loch of Winyadepla [HU 640 930]. They have an average trend of about 135º and a steep dip to the north-east. Several contain small chromite prospects. Any occurrences of metadunite in other klippen of the Upper Nappe are obscured by antigorite serpentinisation.

Many thin veins of metapyroxenite cut the Metaharzburgite Layer. They are generally no more than several centimetres wide, and fill planar joint-like fractures (P533719). They have been called dykes (Prichard, 1985), but are referred to here as veins to avoid confusion with dykes that cut the Upper Metagabbro. The pyroxene crystals are coarse (1 cm or more in diameter) and always pseudomorphed by an intergrowth of antigorite and talc, so it is unclear if the original crystals were orthopyroxene or clinopyroxene. Orthopyroxene in the host rock is usually thoroughly altered while clinopyroxene is usually fresh, so they are likely to have been orthopyroxene. The veins also lack calcic secondary minerals. Many veins strike parallel to the rhythmic banding where it is present, but they only rarely have the same dip. They cut the banding, and are in turn cut by bodies of metadunite in the Metaharzburgite Layer. They have no preferred orientation in areas lacking compositional banding. The veins are non-planar in several localities, including west of Girr Wick [HP 6600 1367] and at Haroldswick [HP 6376 1220]. An irregular sheet about 10 cm thick of fresh, olivine-free orthopyroxenite crops out south of The Punds [HP 6446 1108].

West of the Gossa Dale Fault, near to the inferred obduction thrust in the extreme north-west part of the Lower Nappe near Queyhouse [HP 610 120], the Metaharzburgite Layer locally exhibits a weakly developed schistose fabric that is parallel to the rhythmic banding developed east of the fault. The age and origin of this fabric are unclear. In the Lower Nappe, in both the Metaharzburgite Layer north of Balta Sound and (with decreasing frequency) in the Metadunite Layer to the south, widely distributed patches, generally no more than 1–2 m across, have a north-easterly trending structure, similar to spaced cleavage. Individual planes of cleavage are separated by several millimetres. Around 80 such occurrences were recorded in the Hill of Clibberswick Klippe (Upper Nappe), including good examples at [HP 6609 1414], [HP 6556 1349], [HP 6396 1233] and [HP 6275 1136]. Thin sections reveal the ‘cleavages’ are sub-millimetre wide tension cracks filled by chrysotile- like serpentine with cross-fibre texture.

Interpretation

The Metaharzburgite Layer in the district has been cited as a tectonite (Gass et al., 1982; Prichard, 1985; Moffat, 1987). Two vertical foliations, intersecting at high angles and made recognisable by flattened crystal shapes arising from simple shearing during mantle flow, were reported by Bartholomew (1983, 1993) to be present throughout metaharzburgite in both nappes of the Shetland Ophiolite-complex. No such features were found during the present survey.

In published models for the Shetland Ophiolite- complex (Gass et al., 1982; Lord et al., 1994), bodies of dunite within the harzburgite mantle layer formed by olivine crystallising from basic or picritic magma produced by adiabatic melting deep in the mantle. None of the innumerable bodies of metadunite in Shetland contain evidence for cumulate crystallisation. Their homogeneity and uniformity is disturbed only rarely, by distorted vein-like bodies and segregated bands of coarse chromitite. An alternative mechanism for the origin of dunite bodies in the mantle (Kelemen et al., 1995) proposes that basic magma produced by adiabatic melting deep in the mantle would be undersaturated in pyroxene and would rise through the mantle by intergranular flow, dissolving pyroxene crystals in mantle rock and crystallising olivine. The magma would focus along conduits made more porous than their surroundings by removal of pyroxene; the conduits would thus become columns of dunite. Enhanced porosity in the dunite (magma:dunite ratio of 1:10 according to Kelemen et al., 1995) could so lower the cohesion of olivine crystals that the rising magma would cause the column of olivine to fluidise. The mobilised olivine crystal mush could then flow up the conduits, producing masses of dunite with intrusive margins in the enclosing harzburgite. The crystal mush would then pass through the Moho to form an overlying intrusive layer of dunite (the Metadunite Layer, see below). When flow ceased, the conduits could fragment to form discrete bodies of dunite in the mantle, like those recorded in the Metaharzburgite Layer.

Metadunite Layer

Metadunite occurs in two settings: as discrete bodies within the Metaharzburgite Layer, as described above; and forming the Metadunite Layer which crops out to the south and east of the Metaharzburgite Layer in Unst. The outcrop of the Metadunite Layer reaches the sea between Wick of Hagdale [HP 645 106] and Swarta Skerry [HP 650 080] in the north, and extends down the centre of Unst to Uyeasound [HP 690 010] in the south. In Fetlar, the Metadunite Layer is exposed very poorly and is covered extensively by klippen of the Upper Nappe and by rocks of the Middle Imbricate Zone. Several occurrences of metadunite form mappable units within the outcrop of the Lower Metagabbro along the north-west coast of Fetlar (e.g. at [HU 597 938], [HU 598 938]). Determination of crystal size and shape is made difficult by pervasive serpentinisation, which leaves only scattered remnants of olivine. The metadunite is coarse-grained (c. 5 mm on average), and its crystal boundaries have triple-junction geometry. Other than local concentrations of chromite, and very rarely clinopyroxene, the metadunite is lithologically uniform and texturally isotropic. There is no preferred crystal orientation or tectonite fabric, other than locally flattened chromite crystals.

Disseminated crystals of chromite, mostly less than 0.5 mm in diameter and commonly less than 0.1 mm, occur in accessory proportions. The crystals are usually translucent reddish brown in thin section, though some are black (opaque) along some or all edges or cracks. They are generally distributed uniformly, as in the metaharzburgite, but locally they occur in loose clusters (schlieren) typically several centimetres thick and around ten times as long, in which individual chromite crystals are separated by several millimetres. Such schlieren are well exposed on the west coast of Wick of Hagdale [HP 643 105]. They are more common inland, to the north of Balta Sound [HP 625 105] and in the area between Helliers Water [HP 610 047] and ground east of Loch of Watlee [HP 600 050] (P533771). A slight flattening of crystals in the plane of the schlieren is discernible locally, e.g. at Keen of Hamar [HP 6474 0962], and very rarely the schlieren display a rectilinear, layer-like distribution (P533757).

Where they are more common, a local alignment of chromite schlieren gives the otherwise isotropic Metadunite Layer a weakly foliated appearance. In the chromite-quarried area north of Baltasound, they are aligned approximately parallel to the boundary between the Metaharzburgite and Metadunite layers. This has led to them being termed ‘layers’ or ‘bands’ and attributed to cumulate crystallisation, but their appearance suggests they are more likely to be a result of deformation during bulk flow of the Metadunite Layer. Veins of chromitite are extremely rare because they were targeted by the miners.

Clinopyroxene crystals are a very minor component of the Metadunite Layer, and are absent from metadunite bodies in the Metaharzburgite Layer. They occur very locally and rarely as sub-millimetre size, fresh, interstitial crystals forming small schlieren, in exposures of the lower part of the Metadunite Layer between Balta Sound [HP 635 085] and the boundary between the Metaharzburgite and Metadunite layers. Clinopyroxene- bearing metadunite is more common in the upper parts of the Metadunite Layer, especially next to bodies of clinopyroxenite-wehrlite. Schlieren of clinopyroxene crystals are much less common than schlieren of chromite, and are easily mistaken for them.

Averaged values for major element oxides and trace elements in samples of metadunite from both its settings (i.e. in the Metaharzburgite Layer and in the Metadunite Layer) are not significantly different (Table 4) and (Table 5). The mean Mg# (MgO/[MgO + FeO]) in 45 whole-rock samples of the Metadunite Layer is 83 (compared with 85 for metadunite bodies in the Metaharzburgite Layer; (Table 3)). Other than water content, the composition of olivine is very similar to those of serpentine and the metadunite, so it can be concluded that serpentinisation has had little effect on bulk composition. There are no significant differences in the trace element contents of metaharzburgite and metadunite.

No statistically robust evidence for cryptic variation is discernible in the whole-rock and mineral compositions of metadunite samples. Accessory chromite in 35 samples collected from a 600 metre long traverse across the Metadunite Layer (traverse 1 of Gass et al., 1982) yielded invariable Cr-number (‘Cr#’; Cr₂O₃/[Cr₂O₃ + Al₂O₃]) and Mg-number (‘Mg#’; MgO/[MgO + TFeO]) values. In an extensive study of platinum-group elements and associated trace elements in borehole cores and traverse samples, Lord and Prichard (1998) found much variation on parts per million and parts per billion scales, and attempted to connect local relatively high concentrations laterally over several kilometres to form a ‘stratigraphy’.

The boundary between the Metadunite Layer and the Metagabbro Layer is not exposed. An abrupt, non-gradational boundary is indicated by patches of closely-spaced exposures on either side, and by a rapid decrease in the ground-magnetic field of up to several thousand nanoteslas within several metres of the boundary. A ground-magnetic survey has confirmed the non-rectilinear nature of the boundary indicated by the distribution of exposures. These observations, and the absence of sheared rock in exposures adjacent to the boundary, indicate that it is not faulted.

Interpretation

The origin of the Metadunite Layer, like that of metadunite bodies in the Metaharzburgite Layer, has been ascribed (Gass et al., 1982; Lord et al., 1994) to cumulate crystallisation of olivine from a basaltic or picritic magma that rose through the metaharzburgite mantle and accumulated above it, forming a magma chamber. However, such a model is inconsistent with the considerable thickness (several kilometres) of the Metadunite Layer, the absence within it of any signs of layering attributable to cumulate crystallisation, its uniformity, and its resemblance to a flow-deformed intrusive body. The Metadunite Layer is best explained, therefore, as ‘re-emplaced Alpine peridotite’; an intrusive body that rose through the mantle, as proposed by Thayer (1969), carrying with it ‘podiform’ bodies and leaving behind in the mantle (i.e. in the Metaharzburgite Layer) small and large bodies of dunite, some of which also contain similar chromite deposits.

Serpentinisation

The Metaharzburgite and Metadunite layers have been thoroughly intensely serpentinised (Rushton, 1973). ‘Loss on ignition’ values for 40 samples of metaharzburgite and metadunite average 14 wt per cent (range 10–15 wt per cent), whereas serpentine contains 13 wt per cent H₂O. 239 samples of metaharzburgite and metadunite have an average density of 2.64 g/cm³, which is close to that of the serpentine group minerals (2.55–2.6).

Serpentinisation occurred in two main phases. The first was pervasive (i.e. not controlled by fractures), and transformed crystals of olivine into mesh-type lizardite that commonly encloses olivine relics. Orthopyroxene was transformed into aggregates of serpentine, talc and amphibole, while clinopyroxene and chromite remained largely unaltered. A second, later, phase of pervasive serpentinisation transformed all primary and secondary minerals, except chromite, into antigorite (and locally talc and magnesite) in a zone up to several hundred metres wide immediately above the base of both nappes (i.e. above the obduction thrust). The antigorite serpentinisation extends along some internal shears within the nappes, none of which are sufficiently large to disrupt the ophiolite layers significantly. The stability ranges of serpentine minerals (O’Hanley, 1996) suggest the lizardite serpentinisation occurred below 300ºC in prehnite-pumpellyite facies, while the later antigorite serpentinisation occurred at a higher temperature, but below 500ºC.

The zone of antigorite serpentinisation above the base of the Upper Nappe has suffered strong secondary shearing. This is displayed particularly well on the west side of the Vord Hill Klippe, north of Stackaberg [HU 611 936], where intensely sheared antigoritised metaharzburgite encloses unsheared lenses of the same rock. On Haaf Gruney [HU 635 983] (P603057) and on the coast south of Ruvrapund [HU 645 921], antigoritised metaharzburgite in the lowest few metres of the klippe has been sheared into north-north-east-elongate phacoids up to 1 m across, above which it is extensively shattered (P534503). Where the base of the Upper Nappe crops out along the western edge of the Hill of Clibberswick Klippe, there is a zone up to around 200 m wide in which all evidence of a protolith has been obliterated by alteration; in the north (e.g. at The Taing [HP 653 145]) the zone consists of sheared antigoritite and in the south it is steatite-mylonite, the latter displayed clearly on the beach of Cross Geo [HP 652 121] ((P533604), (P533627)). The Clibberswick and Vord Hill klippen are sufficiently thick to be composed largely of ochrous- weathering lizardite serpentinite, which overlies their white-weathering antigoritised bases. The Gallow Hill, Watlee and Caldback klippen are so thin that they lie entirely within the zone of white-weathering antigoritised metaharzburgite that overlies their basal (obduction) thrust. Shearing and shattering is not a feature of the antigorite serpentinised base of the Lower Nappe.

The two types of serpentinisation are distinguished easily in the field, because weathered surfaces of lizardite serpentinite are a bright ochrous colour (to a depth of about 1 cm), and altered crystals of orthopyroxene weather white or pale pink, while weathered surfaces of antigorite serpentinite are white (P534229). Textural and lithological relationships are therefore revealed in great detail on the extensive, weathered, glacially smoothed exposures. Where it is lizardite-serpentinised, metadunite in the Metadunite Layer has weathered to an ochrous colour that is significantly brighter and stronger than that displayed by metaharzburgite (P533605). It also has a smoother surface due to the absence of pyroxene relics, and it contains a greater density of accessory chromite crystals. The two variants are therefore distinguished easily in the field where they display ochrous weathering; it is possible to distinguish clots of metadunite a few centimetres across enclosed in metaharzburgite. However, where they are antigorite serpentinised (and especially where the metadunite is clinopyroxene bearing) both variants weather white, and distinction in the field becomes difficult or impossible.

Lizardite, with characteristic mesh texture, is ubiquitous in ochrous-weathering rock, with or without relict olivine and younger antigorite. Two hundred and fourteen thin sections of ochrous-weathering metadunite and metaharzburgite all contain mesh-textured serpentine; of these, 137 have olivine relics, 47 contain patches of antigorite, and 39 (18 per cent) contain fine-grained magnetite in more than trace amounts. Antigorite (which has slightly higher birefringence and a generally more flake-like morphology than lizardite) is ubiquitous in the white-weathering serpentinite. One hundred and five thin sections of white-weathering metadunite and metaharzburgite lack olivine and mesh- textured serpentine, but 56 of them (53 per cent) contain fine-grained magnetite in more than trace amounts. In rock containing lizardite and antigorite, the antigorite encloses at least as much fine-grained magnetite as lizardite. It is concluded, therefore, that the ochrous weathering colour derives in part from the presence of relict olivine (Figure 19).

Lizardite and antigorite can commonly be distinguished in thin section by their diagnostic textures (especially mesh-textured lizardite and flake-like antigorite; O’Hanley, 1996) and optic orientation. Chrysotile forms serrated, vein-like patterns, and other less distinctive patterns, in association with lizardite. The two minerals can also be distinguished by XRD.

Orthopyroxene breaks down more easily than olivine during serpentinisation. Only 7 of 180 thin sections of metaharzburgite contained relict orthopyroxene crystals, all of them partly altered. The lizardite serpentinisation event caused orthopyroxene crystals to be replaced by fine-grained aggregates of talc, serpentine and amphibole (in order of decreasing abundance). A smaller proportion of crystals have been pseudomorphed by opaque, pinkish light-brown, close-packed fibres (uralite); electron microprobe analysis of these revealed actinolite and some diopside compositions. Orthopyroxene crystals occur in bodies of clinopyroxenite-wehrlite in the Metadunite Layer, but all have been uralitised. Fine-grained magnetite concentrated along original mineral cleavage traces reveals the location of antigorite-replaced orthopyroxene crystals in antigorite-serpentinised rock. On weathered surfaces, altered orthopyroxene forms white or pinkish- white spots with a slightly greater relief than serpentine- replaced olivine.

Graphitic serpentinite occurs in metaharzburgite of the Lower Nappe near Dalepark [HP 623 117] and Hagdale ( [HP 644 106], [HP 644 101]), always in a fault or shatter zone marking the centre of a white-weathering (antigorite) zone. It forms shattered, angular fragments associated sometimes with steatite or talc, which are a more common feature of these zones. Charcoal-black hand specimens are distinguished readily from greenish- black, slightly lustrous, graphite-free serpentinite. The concentration of graphite ranges from barely noticeable to so concentrated that it renders thin sections opaque. In all cases the particles are cryptocrystalline. Powder-XRD examination of graphite-rich samples revealed antigorite and a very small graphite peak. Powder-XRD analysis of around 50 samples of serpentinite revealed brucite in 12 lizardite-serpentinised samples, at or just above the lower limit for powder-XRD identification. Brucite was not identified in thin section.

Olivine expands in volume by up to 60 per cent during serpentinisation (Shervais et al., 2005). To accommodate this change, a body of dunite must also increase in volume, or it must lose volume by metasomatic transfer of material to adjacent rocks (the actual value depends on the serpentinisation reaction). The density of dunite is reduced by serpentinisation to 2.65 g/cm³ (O’Hanley, 1992). The average density of 239 samples of metaharzburgite and metadunite from the Shetland ophiolite is 2.64 g/cm³ (Flinn, 2000), suggesting that serpentinisation is virtually complete. However, there are no signs in Shetland of the structures ascribed by O’Hanley (1992) to solid state expansion, nor is metasomatism apparent in rock adjacent to serpentinised material. Furthermore, the field and geophysical evidence cited above suggests that the boundary between the Metadunite and Metagabbro layers is intrusive. It is therefore concluded that the Metadunite Layer must have been in a mobile and hydrated state when its boundary with the Metagabbro Layer formed.

The Upper Nappe contains several occurrences of rodingite (a massive, dense, buff to pink rock formed by metasomatic alteration, typically rich in grossular garnet and calcic pyroxene, and enveloped in serpentinite) (Bates and Jackson, 1987). They were discovered and described by Phemister (1964) and later by Moffat (1987), and have not been re-examined during the present survey. The best exposures occur in the sheared zone at the base of the Upper Nappe between Colbinstoft [HU 612 937], to the north of Stackaberg, and Tressa Ness [HU 620 950] in Fetlar (P534490). There is a separate occurrence [HU 600 935] to the north of Hamara Field in Fetlar. A similar occurrence on Sound Gruney [HU 580 960] includes a rodingitised sheet of lamprophyre. A rodingitised lamprophyre intruding the Gallows Hill Klippe at [HP 573 005] is the only recorded occurrence of rodingite in Unst.

Chromite in the ultrabasic lithologies

The term chromite is used here for crystals of spinel containing more than trace amounts of chromium that occur in the Metaharzburgite and Metadunite layers and in the clinopyroxenite-wehrlite bodies. The vast majority are iron-bearing magnesiochromite, which is translucent reddish brown in thin sections of normal thickness. The occurrence and composition of accessory chromite in the Shetland Ophiolite-complex have been described and discussed in detail by Brzozowsky (1977) and Gass et al. (1982) on the bases of 411 microprobe analyses. More recently, Flinn for the purposes of this survey, made 93 analyses of accessory chromite.

Two main types of chromite are distinguished. ‘Accessory’ chromite occurs as crystals of <1 mm (generally <0.5 mm) disseminated throughout the units of metaharzburgite, metadunite and clinopyroxenite- wehrlite. ‘Ore-type’ chromite occurs as crystals of 1 to 10 mm and in concentrations ranging from centimetre- size streaks to ‘podiform’ deposits 100 m across. Accessory chromite is richer in iron than ore-type chromite, and is accompanied by very fine-grained magnetite in enclosing serpentinised olivine. Ore-type chromite crystals are accompanied by magnetite crystals of similar size; these range in modal volume from almost zero to 100 per cent, and are detectable in the field only with the aid of a magnet. Ore-type chromite grains contain inclusions of magnetite varying in volume from absent to sufficient to be picked up by a magnet.

Crystals of accessory chromite generally have angular outlines, giving them a fragmented appearance. Larger crystals are usually heavily fractured and veined by serpentine. A small proportion has embayed outlines enclosing now-serpentinised minerals. The number of crystals in each thin section varies according to lithology. Metaharzburgite of the Upper Nappe contains twice as many crystals as metaharzburgite of the Lower Nappe, while the Metadunite Layer contains twice as many again. Accessory chromite crystals appear in the field to be distributed uniformly, but locally they occur in very loose elongate clusters (schlieren) (P533784). Such schlieren, extending from 10 cm to a metre, are well exposed on the coast at Wick of Hagdale [HP 643 105], inland to the north of Balta Sound [HP 625 105], and in the area [HP 600 050] between Helliers Water to the east and Loch of Watlee to the west. Within these areas the schlieren have a roughly parallel alignment. Very rare, parallel, clinopyroxene-rich schlieren also occur in these areas. The schlieren are very rarely rectilinear (P533793).

The chromite ore deposits are now entirely destroyed by mining, leaving only the spoil heaps as evidence of their previous existence. The spoil heaps provide evidence of two types of ore, named high grade and low grade respectively (Plate 3d). The high grade ore takes the form of veins of chromitite, which are now exceedingly rarely exposed due to having been the targets for prospecting at intervals during the last 200 years. Scattered fragments are preserved in spoil heaps. The vein shown in (Plate 3d) is the only one preserved in situ that is more than several centimetres thick. The low grade ore is preserved only in the spoil heaps, where it is seen to vary in content of chromite from very little to 90 per cent, and to be limited by apparently irregular boundaries and no regularity of grain distribution. Sandison pumped out many water-filled pits and mined the Mid Brake deposit. He reported (Sandison, 1936) that the two largest deposits, Hagdale and Mid Brake, have a near vertical swollen vein of chromitite along the north side of the pit. Parallel to it and separated from it by a barren zone, a metre or two thick, is a band of low grade ore a metre or two wide. In the Hagdale deposit, the vein swelled to 3.6 m wide, but in the Mid Brake deposit, it was no more than 2 m wide. In both cases, the vein was no more than 100 m in length. The other worked deposits were much smaller and less well known. It appears from the spoil heaps that the proportion of high grade to low grade ore varied considerably. In the spoil heaps, clots of high grade ore less than 25 cm in size, floating in unsheared serpentinite, are a not uncommon feature (P533710), (P533792), (P533646). They are best explained as being derived from high grade veins disrupted during mobilisation of the metadunite. Similar sized clots of low grade ore presumably have the same origin (P533643), (P533699), (P533663). Nowhere did the veins of chromitite, or the grains forming the low grade ore, exhibit any order or regularity, or display any rectilinearity.

The chromite deposits are long worked out, the pits are now water-filled, and spoil heaps are dispersed. Material in the spoil heaps reveals that the deposits suffered two main phases of deformation: the deposits were deformed during an older phase. Distorted and broken fragments of vein-type and segregated ore now ‘float’ in apparently undeformed and unsheared serpentinite. Later, the ore and the serpentinite were faulted and sheared during a younger phase The quarry shapes and relict exposures of the original in-situ ore bands reveal a crude alignment of the banding with nearby accessory chromite schlieren.

Composition

Accessory (disseminated) and ore-type (massive) chromite are compositionally distinct (Gass et al., 1982, fig. 25) (Figure 20); the accessory chromite is richer in iron than the ore-type chromite. Ore-type chromite is iron-rich magnesiochromite, while accessory chromite straddles the boundary between magnesiochromite (MgCr₂O₄) and chromite (Fe₂+Cr₂O₄).

Opaque crystals and crystal rims of chromite, from bodies of clinopyroxenite-wehrlite in the Metagabbro Layer, have higher Cr# and lower Mg# than translucent crystals of accessory chromite (Figure 20). The composition (in terms of Mg# and Cr#) of translucent crystals of accessory chromite in the Metaharzburgite Layer, in metadunite bodies of the Metaharzburgite Layer, and in metadunite of the Metadunite Layer, does not in general vary significantly (Figure 20).

Gass et al. (1982) attributed lower Mg# values in accessory chromite to late alteration. However, in almost all analyses of altered chromite rims there is an accompanying marked increase in Cr#, due to loss of Al. There is also very little overlap between the compositional fields for accessory- and ore-type chromite. These facts, together with the different habit and occurrence of the two types of chromite, suggest they crystallised at different times. The ‘podiform’ chromite may have crystallised from a more primitive (Mg-rich) magma as it rose through mantle rock containing accessory chromite.

Interpretation

There are no published models for texturally complex chromite deposits like those in Shetland. Zhou et al. (2001) have shown that progressive silicification of the magma rising through ‘conduits’ (as in the model of Keleman et al., 1995; see above) in the mantle would lead to chromite crystallisation. While such crystallisation might lead to the formation of a pod, it does not offer a mechanism for forming the Shetland deposits. Even more difficult to explain is the occurrence of identical chromite deposits in the Metadunite Layer, where simple stratiform cumulate crystallisation might be expected. Gass et al. (1982, figs. 19 and 20) showed that chromite compositions in the Shetland Ophiolite match those in other ophiolites, but differ markedly from those of stratiform complexes. It is proposed that the Shetland ‘podiform’ deposits formed in the mantle, and were transported with the intrusive bodies that rose into the crust (cf. Thayer, 1969) to form the Metadunite Layer (see above). The crude foliation discernible locally in the Metadunite Layer as a weak alignment of chromite layering is interpreted as being due to flow deformation formed during emplacement.

Metagabbro Layer

The Metagabbro Layer crops out south and east of the Metadunite Layer on Unst, and continues southwards onto the north-west coast of Fetlar through Uyea [HU 600 995], Daaey [HU 603 948], and Urie Lingey [HU 597 957]. It is concealed locally in Muness [HP 620 010] and eastern Uyea [HU 606 990] by Unst Phyllite Group rocks of the Middle Imbricate Zone, in the Mu Ness [HP 638 010] and Sound Gruney [HU 580 962] areas by tectonic slices of ophiolite in the Middle Imbricate Zone, and by klippen of the Upper Nappe (Figure 18).

The boundary between the Metagabbro Layer and the Metadunite Layer is unexposed, but is considered to be intrusive and unfaulted (see above). In several places, the Metagabbro Layer forms marked protrusions downwards into the Metadunite Layer, most notably at Hill of Colvadale [HP 625 065] and on the north side of Helliers Water [HP 607 055]. Three bodies of metagabbro lie completely within the Metadunite Layer, north-west of Virda Field [HP 620 070], and south-west of Uyeasound ( [HP 585 005] and [HP 585 003]). The boundary is interrupted in several places by large bodies of clinopyroxenite- wehrlite, notably north of Sobul [HP 605 045] and at Uyeasound [HP 590 013]. The boundary between the Metadunite and Metagabbro layers marks the greatest lithological change in the ophiolite succession: north of the boundary there is no feldspar and south of it there is no orthopyroxene. The only olivine south of the boundary occurs within bodies of clinopyroxenite- wehrlite. The gabbro–peridotite boundary is widely considered to be the geophysical or seismic Moho.

The Metagabbro Layer has been divided into two parts on lithological grounds. The Lower Metagabbro is medium-grained (1 to 2 mm), green and white speckled rock that is generally uniform, isotropic, and even more thoroughly and uniformly hydrated than rock in the Metaharzburgite and Metadunite layers. It is clear from field observations that this early hydration is distributed uniformly, like that in the serpentinised ultrabasic layers, and is not concentrated in and around fractures as is generally held necessary for hydration by seawater (e.g. Schroeder et al., 2002). Throughout the ophiolite succession accessory sulphide minerals have a magmatic sulphur isotope signature (Maynard et al., 1997). Thus, two lines of evidence suggest that the hydration was not caused by seawater.

The general uniformity of the Lower Metagabbro is interrupted locally by widely scattered patches of rhythmic compositional banding, in which individual bands up to several centimetres thick (Plate 4a) occur in groups from less than one to several tens of metres across. A wide range of orientations has been recorded, but in general the banding dips gently towards the east, south-east and south (Figure 21)b. The largest areas of continuous banding occur on Balta [HP 660 085] and at Sheetaberg [HP 648 076]. A patch of well-developed banding crops out east of Urie, on Fetlar [HU 5980 9390]. A single example of isoclinally folded banding was recorded on Balta (P533933); otherwise, the bands are ungraded and undeformed.

Large areas of the Lower Metagabbro have been brecciated and healed, and the fragments (each no more than several centimetres across) have all been rotated slightly (Plate 4b). The healed rock is as resistant to hammering as unbrecciated rock, and the fractures are not discernible in thin section. The brecciation occurred before common and widespread late shattering, shearing and greenschist-facies metamorphism. It is particularly well developed on the east coast of Balta [HP 661 075] and on the south-east coast of Mu Ness [HP 639 005], and occurs commonly along the east coast of Unst between Huney [HP 640 062] and Qui Ness [HP 621 032] (Figure 21)a. Inland, brecciated metagabbro is well developed between Uyeasound [HP 595 015] and Sobul [HP 605 040]. Rhythmic banding is difficult to detect in areas of healed breccia.

The Lower Metagabbro consists of plagioclase replaced by nearly opaque saussurite, and clinopyroxene replaced by single-crystal plates of pale green to colourless hornblende that is close to actinolite in composition. The latter are locally partly replaced by acicular actinolite formed during late greenschist-facies metamorphism. The rock weathers to a speckled pattern of pale green and white. Original crystal boundaries are preserved, revealing medium-grained (1 to 2 mm), equigranular crystals. Rare, interstitial crystals of primary brown pargasitic hornblende are preserved. Only 4 of 50 thin sections contained fresh relict clinopyroxene, and none of the sections contained olivine (or any indication that olivine was ever present), opaque minerals, apatite, titanite or carbonate minerals, and only one contained epidote. The thin sections, like the exposures, vary little in appearance.

The boundary between the Lower Metagabbro and the Upper Metagabbro is not defined sharply but is accompanied by a number of clinopyroxenite-wehrlite bodies that are indistinguishable in the field from those occurring above and below the boundary between the Metadunite and Metagabbro layers. In the Mu Ness peninsula, the metagabbro outcrop is divided into two parts by a north-west-striking, steeply dipping zone about 10 m wide of gneissose or flaser-gabbro that crosses the peninsula from [HP 638 008] to [HP 634 017]. South of this zone exposed rock is typical of the Lower Metagabbro. North of the zone, exposed rock is typical of the Upper Metagabbro, and contains numerous north-east-trending dykes (of the Quasi-sheeted Dyke-swarm). Metagabbro exposed along the north-west coast of Fetlar is intruded by similar dykes with a broadly east–west trend.

Compared to the Lower Metagabbro, the Upper Metagabbro is lithologically and texturally complex. The unit ranges from rock that is indistinguishable from that forming the Lower Metagabbro to darker, finer grained mafic rock forming irregularly shaped bodies from a few metres to hundreds of metres across. Lithological boundaries within the Upper Metagabbro are very difficult to discern due to gradation between rock types, extensive phyllonitisation accompanying late greenschist-facies metamorphism, and the presence of many dykes (forming the Quasi-sheeted Dyke-swarm, see below). Hornblende-plagioclase pegmatite, and more rarely plagiogranite (see below), form sheets and veins generally no more than 1 m thick. The north tip and north-east side of Ham Ness [HP 635 020] are formed of an intrusion of epidote-hornblende metagabbro. The Vere [HP 646 033], a half-tide rock in the sea 1.5 km north-north- east of Ham Ness, is formed of a similar lithology. This intrusion cuts the Quasi-sheeted Dyke-swarm, but is itself cut by several similar sheets of similar orientation, a vein of plagiogranite and by several sheets of lamprophyre.

The finer grained, darker mafic rock that is characteristic of the Upper Metagabbro has generally suffered low-grade metamorphism that has largely destroyed pre-metamorphic texture. The rock consists now of an isotropic aggregate of green, feathery hornblende crystals, 0.3 to 0.4 mm long, with (probably secondary) albitic plagioclase, and titanite ± ilmenite. Carbonate minerals and epidote occur in some rocks, but are generally in, or associated with, veins. Several of the darker bodies of fine-grained rock have suffered less from late alteration, and consist of brown hornblende and plagioclase. One such sample contains brown crystals of tschermakite-pargasite amphibole that are interstitial to saussuritised plagioclase, plates of blue to green pleochroic hornblende, and ilmenite. It has a crystal fabric similar to, but finer grained than, rock samples from the Lower Metagabbro. More common are rocks of transitional type containing brown hornblende with albite and accessory actinolite, epidote and titanite.

Pegmatitic metagabbro composed of coarse brown hornblende and hydrated plagioclase forms minor irregular bodies throughout the Metagabbro Layer, but especially in the Upper Metagabbro. They cut rhythmic banding and are cut by dykes of the Quasi-sheeted Dyke-swarm. Pegmatitic metagabbro fills a network of joint-like fractures in the Lower Metagabbro north of Muckle Head [HP 660 085] on Balta. In Fetlar, large masses of pegmatitic metagabbro crop out at Vallahamars [HU 615 904] and in a tectonic slice of ophiolite in the Middle Imbricate Zone, south-west of Houbie [HU 619 902]. Lenses, sheets and masses of plagiogranite (described below) crop out sparsely in the Metagabbro Layer, especially in the Upper Metagabbro.

Geochemistry

Twenty analyses of samples from the Lower Metagabbro and 17 from the Upper Metagabbro are available (Table 6) and Table 7); Thomas, 1980; Gass et al., 1982; Gunn et al., 1985; Flinn, unpublished; Spray, 1988; Spray and Dunning, 1991). All of the rocks are highly altered; geochemical interpretations should therefore be treated with caution. The Lower Metagabbro and Upper Metagabbro appear distinct, although they are not necessarily unrelated, in terms of their major element compositions (Figure 22)d and e. On the TAS diagram (Figure 22)d Upper Metagabbro samples have generally higher (Na₂O + K₂O) at a given SiO₂ concentration; Lower Metagabbro samples cluster mainly in the basalt field, while Upper Metagabbro samples straddle the foidite, picrobasalt and basalt fields. The mean Mg# (MgO/[MgO + TFeO]) in 26 whole rock samples of the Lower Metagabbro is 64, and in 15 samples of the Upper Metagabbro it is 35 (Table 3).

The analyses scatter across tectonomagmatic and petrological discrimination diagrams (Figure 22), but most samples from the Upper Metagabbro have mid-ocean-ridge-basalt (MORB)-like characteristics, whereas those from the Lower Metagabbro do not (Figure 22)a, b. The Ti, Zr, and Y contents of Lower Metagabbro samples are so low that they generally plot outside the tectonic setting fields on discrimination diagrams (Figure 22)a and b. High field-strength element (HFS)-depleted, isotropic gabbro such as that forming the Lower Metagabbro is generally interpreted to be cumulate from which the intercumulus liquid has been expelled, taking with it the missing HFS elements.

Interpretation

The general absence of cumulate-type layering within or below the Lower Metagabbro, and the presence instead of several kilometres of unlayered, non-cumulate metadunite, provides no evidence for a magma chamber in the crustal layer of the Lower Nappe. The absence of evidence for a magma chamber overlying the Metaharzburgite Layer does not support a model, as advocated by Gass et al. (1982) and Prichard (1985), for the Upper Metagabbro being the roof region of a high- level magma chamber at a constructive plate margin. The composition of the Upper Metagabbro indicates an origin by melting of fertile MORB-like mantle.

Flinn (1996b) has argued that batches of magma that formed the Upper Metagabbro, together with late felsic differentiates such as hornblende pegmatite, plagiogranite and swarms of basic dykes, cannot have been intruded from below, where there is in any case no trace of them; they must have been intruded from above.

Bodies of clinopyroxenite-wehrlite in the Metadunite and Metagabbro layers

Bodies of interbanded clinopyroxenite, wehrlite, and metadunite are scattered widely through the Metadunite and Metagabbro layers, especially close to the upper or eastern borders of these layers (Figure 23). Particularly high concentrations occur in Swinna Ness [HP 652 092], near Skeo Taing [HP 644 083], and in the area between Helliers Water and Loch of Watlee [HP 607 048]. They range in maximum dimension from centimetres to hundreds of metres, and in shape from lenticular to irregular-equidimensional. Elongate bodies tend to have a north-north-east strike. Large bodies of irregular shape occur at Sobul [HP 602 038] and Sandwick [HP 611 021].

The bodies consist largely of fresh pyroxene or amphibole. By contrast, the host rock is generally strongly altered (saussuritised metagabbro and serpentinised metadunite). Hence, those bodies larger than a few metres across form prominent exposures that are identified readily even where covered by vegetation. The bodies are strongly magnetic, and those that are too small to form prominent exposures may be located and mapped by magnetometer where they lie within the outcrop of the relatively non-magnetic Metagabbro Layer.

The bodies are not easy to study in the field. In inland areas, the mound-like outcrops and the low ground between them are generally covered by vegetation. Where they are exposed, they lack the smoothly weathered surfaces that reveal so clearly the petrography and mineralogy of the enclosing metagabbro and metadunite. Their relationship with the host rocks is exposed more clearly in the coast sections at Skeo Taing [HP 645 085] and Swinna Ness [HP 650 090]. One lenticular body, 10–20 m long, is exposed in full on the beach at Skeo Taing [HP 647 083]. Inland, between Helliers Water and Loch of Watlee [HP 600 050], and south of Hagdale [HP 640 100], several small bodies (up to several metres across) are exposed clearly ((P534091), (P534242)). They are surrounded by an ‘aureole’ of interstitial clinopyroxene in the enclosing metadunite. Several bodies hosted in metadunite are accompanied by swarms of smaller (up to 10 cm diameter), angular bodies in the adjacent rock.

Within the bodies, lithology ranges from clinopyroxenite through olivine-clinopyroxenite to wehrlite. Intimate ‘mixing’ of, and gradation between, the lithologies renders them difficult to distinguish in the field, and they have been mapped using the general term ‘clinopyroxenite-wehrlite’. Some bodies are rhythmically modally banded, with bands typically several centimetres thick. The bands are difficult to detect due to a lack of significant colour contrast between the lithologies, and they lack the abrupt, planar interfaces typical of rhythmic banding in the Metagabbro and Metaharzburgite layers. The best examples are in the Skeo Taing area [HP 645 085], but even here the banding is poorly to vaguely developed and extremely local (Plate 4c).

The rock is dominantly granular, being composed of sub-equidimensional, unzoned, medium-grained (1–2 mm) crystals. Crystals of clinopyroxene tend to partly enclose crystals of olivine, but they rarely enclose them fully; a lone occurrence of coarse, poikilitic clinopyroxene enclosing olivine was recorded in Huney [HP 650 065]. Orthopyroxene crystals are enclosed in clinopyroxene more commonly than in olivine crystals. Chromite occurs as inclusions in all these minerals.

The bodies of clinopyroxenite-wehrlite appear in the field to be unrelated to their host rocks; they are either xenoliths (Plate 4d) or lenticular intrusions. They show none of the continuity, and internal or external regularity, of cumulate banding expected in the lower crust layers of ophiolite. The orientation of rhythmic banding in the metagabbro host rock close to the larger irregularly shaped bodies varies even more widely than usual, e.g. south of Sobul [HP 603 036]. Some bodies in Balta are well-formed lenses, one of which cuts rhythmic layering in the host metagabbro. Both the lenses and rhythmic layering are cut by pegmatitic metagabbro.

Bodies of clinopyroxenite-wehrlite hosted in the Metadunite Layer consist of clinopyroxene, serpentinised olivine, and a small proportion of uralitised orthopyroxene. The metadunite host contains no orthopyroxene, and clinopyroxene occurs only very locally as disseminated accessory crystals, mostly near to the bodies of clinopyroxenite-wehrlite. Bodies hosted in the Metagabbro Layer consist of fresh clinopyroxene (except where local, late amphibolitisation has occurred) and serpentinised olivine; plagioclase and orthopyroxene are absent. Very fine-grained (0.01 mm), almost isotropic, pale green chlorite (Figure 19) forms uniform, monomineralic patches apparently filling former voids between crystals of pyroxene and serpentinised olivine. The patches are ‘invaded’ by needles of actinolite formed during late greenschist-facies metamorphism. The host metagabbro is dominated by saussuritised plagioclase and amphibolitised clinopyroxene, and lacks olivine. Clinopyroxene in metagabbro contains more total Fe than that in the clinopyroxenite-wehrlite bodies. However, both the host metagabbro and enclosed bodies of clinopyroxenite-wehrlite contain accessory, primary, interstitial, pale-brown hornblende.

The hydration event that caused extensive development of lizardite in the Metaharzburgite and Metadunite layers had the same effect in bodies of clinopyroxenite-wehrlite, including those enclosed in the Metagabbro Layer. The result was perva- sive serpentinisation of olivine, and uralitisation of orthopyroxene; clinopyroxene and brown horn- blende were left unaltered. Enclosing metagabbro was also pervasively and uniformly hydrated, resulting in saussuritised plagioclase and clinopyroxene transformed to plates of colourless hornblende. The (altered) clinopyroxene in metagabbro is richer in Fe and Al than fresh clinopyroxene in the bodies of clinopyroxenite-wehrlite. The latter is similar in composition to unaltered clinopyroxene in metaharz- burgite of the Metaharzburgite Layer.

Bodies of clinopyroxenite-wehrlite hosted in the Metagabbro Layer suffered a later greenschist-facies metamorphism locally. In the northern part of the outcrop of the Metagabbro Layer this took the form mainly of minor replacement of clinopyroxene by actinolite. However, farther south, near Sandwick [HP 613 024], Uyeasound [HP 590 010] and in Uyea [HP 595 001], clinopyroxene in the clinopyroxenite- wehrlite bodies has been replaced completely by actinolite, and secondary carbonate has replaced and veined the rock locally. Colourless hornblende that has replaced clinopyroxene in the enclosing metagabbro has itself been replaced by actinolite.

Geochemistry

Whole-rock analyses of samples from bodies of clinopyroxenite-wehrlite are available from various sources (Table 8) and (Table 9). Highly variable mineral modes make it difficult to evaluate differences in the whole-rock composition of clinopyroxenite- wehrlite bodies hosted in different layers of the ophiolite. However, the bulk compositions of clinopyroxenite-wehrlite bodies hosted by the Metadunite Layer, the Lower Metagabbro and the Upper Metagabbro are distinct. The mean Mg# in clinopyroxenite-wehrlite hosted in the Metadunite Layer is 85 (compared with 83 for the Metadunite Layer; (Table 3)), whereas that in clinopyroxenite- wehrlite hosted in the Lower Metagabbro is 76 (compared with 64 for the Lower Metagabbro; (Table 3)), and that in clinopyroxenite-wehrlite hosted in the Upper Metagabbro is 55 (compared with 35 for the Upper Metagabbro; (Table 3)). It is noticeable that the differences between the mean Mg# of the clinopyroxenite-wehrlite bodies and their hosts become greater in shallower layers of the ophiolite.

Interpretation

Bodies of clinopyroxenite-wehrlite are an enigmatic feature of the Shetland ophiolite, and their origin has not been specifically addressed in publications. The Metadunite Layer, with its xenolith-like bodies of clinopyroxenite-wehrlite, has been presented hitherto as an ultramafic cumulate in which dunite grades upwards into wehrlite and then pyroxenite. This description has been accompanied by idealised sections of ophiolite pseudostratigraphy showing distinct, successive layers of dunite, wehrlite and clinopyroxenite (Gass et al., 1982; Prichard, 1985; Lord et al., 1994). The absence in the field of persistent layers of wehrlite and clinopyroxenite has been explained as a result of tectonic disruption by north-east-trending sinistral faults (Lord et al., 1994). However, these faults are not apparent on the ground, or in aerial photographs, and the boundary between the Metagabbro and Metadunite layers appears unfaulted (see above).

It is proposed that rocks making up the bodies of clinopyroxenite-wehrlite formed by cumulate crystallisation within what are now their host rocks, and that the discrete bodies are autoxenoliths formed during mobilisation and flow of the host rocks into their present positions as layers in the ophiolite. This occurred prior to lizardite serpentinisation.

Quasi-sheeted Dyke-swarm

The Upper Metagabbro is cut by many subparallel mafic sheets (Figure 24), particularly along the east coast of Unst from Pund Stacks [HP 621 033] and Qui Ness (Plate 4e), to as far north as Huney [HP 650 065], and in the Ham Ness area [HP 635 015]. A smaller grouping of similar sheets crops out at Urie in Fetlar [HU 595 942], and a few are exposed in the south of Balta [HP 658 070] and in the Holm of Heogland [HU 575 990]. Their character and position in the ophiolite suggest they are part of a sheeted dyke- swarm. They are referred to here as the Quasi-sheeted Dyke-swarm because (unlike many occurrences of sheeted dykes) the dykes occupy at most 50 per cent by volume of the outcrop, being separated from each other by screens of metagabbro country rock. Furthermore, most of the ‘dykes’ are more nearly parallel to the base of the Metagabbro Layer than normal to it, though they are oblique to rhythmic banding in the host rock. Dykes are displayed particularly well on the beach at Brei Wick at [HP 638 017], and at Pund Stack [HP 622 034]. Gass et al. (1982) presented a reconstruction of the occurrence at Mu Ness [HP 638 014].

The dykes range from off-white to nearly black, but the vast majority are pale greenish grey. They are very similar in colour and weathered texture to fine- grained rock forming the Upper Metagabbro, which makes them difficult to distinguish from their host. Dykes are identified more easily in broad coast exposures than in smaller inland exposures, because of their continuity. However, they are a common feature between the east coast of Unst and the zone of clinopyroxenite-wehrlite bodies that marks the boundary between the Upper Metagabbro and Lower Metagabbro. Only three examples were found of dykes that cut the clinopyroxenite-wehrlite bodies; at Mailand [HP 613 022], Vord Hill (on Unst) [HP 617 034] and Brough Taing [HP 627 049]. A very few dykes were found west of the gap in the screens, at Framgord [HP 610 030]. Many dykes are no more than several centimetres wide, most are 30–60 cm wide, and a few are 1 m or more wide. Even the most closely spaced dykes are separated by screens of country rock, which are about the same width as the dykes themselves. The dykes are essentially lenticular bodies, few being traceable for more than several metres. A small proportion has chilled margins, and some are emplaced within other dykes.

On the east coast of Unst the dykes have a north- east strike, in Huney it is nearly north–south, while on the north coast of Fetlar [HU 597 938] and the south coast of Balta [HP 659 069] it is broadly east–west (Figure 24). Within these areas the strike varies as much as 30º from the average, with the result that some dykes intersect; younger dykes tend to be have strike orientations that are rotated clockwise relative to older dykes. Many younger dykes have a penetrative vertical schistosity, which crosses them at an angle approximately 30º greater than their strike.

Petrography

One hundred and ten samples of the Quasi-sheeted Dyke-swarm were collected to represent the full range of field characteristics. Almost all are composed largely of amphibole, most of it secondary. A decussate arrangement of laths of former plagioclase gives around half of the dykes a subophitic texture. The laths are very fine- to fine-grained (0.1 to 0.5 mm) and altered to saussurite or saussurite + albitic plagioclase. Amphibole crystals are the same size as feldspar crystals and form randomly orientated aggregates of fibrous to spiky, colourless to pale green crystals that locally enclose relict brown hornblende. Two dykes in Fetlar with chilled margins are indistinguishable petrographically, being composed of fresh brown hornblende in ophitic intergrowth with saussuritised plagioclase. Such feldspar as they do contain is interstitial to amphibole. One such example contains brown pargasitic hornblende with younger actinolite, albite and titanite. Plagioclase- and hornblende-phyric chilled margins occur rarely. Epidote occurs in some dykes, as does ilmenite rimmed by titanite and, less commonly, chlorite. Carbonate minerals are virtually absent. Dacite dykes consist of a matrix of very fine-grained (0.1 mm) laths of sodic plagioclase cut by an irregular network of fractures filled by fine-grained epidote.

Geochemistry

Forty-six whole-rock geochemical analyses of the Quasi- sheeted Dyke-swarm are available (Moffat, 1987; Prichard and Lord, 1988; Spray, 1988; Flinn, unpublished). The proportion of dykes recorded as dacitic in the field and in thin section is greater than that represented in the analysed samples. All of the analysed samples are highly altered. In 10 analyses presented by Spray (1988) there are both quartz- and hypersthene-normative variants, Y/Nb>2, and K, Ti and Zr contents are low: these are features indicative of low-K tholeiite. Spray (1988) also recorded several high-Mg dykes (MgO c.18 wt per cent) of boninitic affinity, and considered them to be ‘typical products of early marginal basin development’. Gass et al. (1982) observed similar compositions in 23 analyses of the dykes. Their trace-element tectonomagmatic discrimination diagrams showed most to be consistent with ‘island arc’ compositions. Re-using 6 of the 23 analyses and 2 new ones, Moffat (1987) concluded that the dykes have a subalkaline tholeiitic composition, and formed by continuous fractionation from a high-Mg or boninitic parent magma.

MgO in the dykes is in the range 1.27 to 18.87 wt per cent (Figure 25)a. Most have MgO in the range 2.0 to 10.0 wt per cent and classify as basaltic trachyandesite, basaltic andesite and basalt (Figure 25)b; they are here called the basaltic group. Dykes with MgO in the range 10.0 to 19.0 wt per cent form a separate group (Figure 25)a, here called the ‘boninitic group’ (cf. Spray, 1988), though they plot outside the boninite field of Crawford et al. (1989, fig. 1.3). Three samples classify as dacite (Figure 25)b and are referred to as the dacite group. In the classification of Le Maitre (2002) a minority of the dyke samples classifies as ‘high- Mg’ volcanic rocks; a few are boninite (SiO₂>52 per cent, MgO>8 per cent, TiO₂<0.5 per cent), and several are picrite (SiO₂ 30–52 per cent, MgO>12 per cent, [Na₂O + K₂O]<3 per cent).

On selected diagrams that discriminate tectonic setting using trace element concentrations, the samples range from volcanic arc basalt/island arc-like compositions, to compositions (mainly those with higher Mg) that lie outwith defined fields (Figure 25)c and d.

Interpretation

Intrusions of the Quasi-sheeted Dyke-swarm cut the Upper Metagabbro and were therefore emplaced into already hydrated rock. Like their host, they retain evidence that they crystallised directly as hydrated hornblende-plagioclase rock before being metamorphosed to greenschist facies. Many are too primitive to have been derived from the Lower Metagabbro, even if it was once part of a magma chamber. They do not generally cut the clinopyroxenite-wehrlite screens marking the boundary between the Lower Metagabbro and Upper Metagabbro, and do not cut the Lower Metagabbro. Their absence from all parts of the ophiolite sequence below the Upper Metagabbro is characteristic of sheeted dykes in the roof of a magma chamber, but there is no magma chamber in the ophiolite forming the Lower Nappe. The dyke compositions indicate they are derived from the mantle by melting, but there is no trace of them in the Metadunite Layer and Lower Metagabbro. It is proposed that the dykes cannot have been intruded into their present position from a deeper level in the ophiolite, and must have been intruded from above, or laterally.

‘Plagiogranite’

‘Plagiogranite’ (a synonym for leucotonalite and ‘oceanic trondhjemite’ (in the sense of Coleman and Donato, 1979)) is a characteristic and nearly universal component of ophiolites. Two variants of ‘plagiogranite’, quartz-albite leucotonalite and quartz-albite-phyric felsite, crop out in the Shetland Ophiolite, almost exclusively in the Upper Metagabbro (Figure 36) and as erosional detritus in the Middle Imbricate Zone. The field appearance of both variants is very similar to that of other light-coloured rocks in the Upper Metagabbro, including intrusions of the Quasi-sheeted Dyke-swarm and feldspathic veins. Around 20 intrusions have been recorded (Figure 26), most of which are lenses or sheets less than 1 m thick. At South Ship Geo, a group of ‘plagiogranite’ veins, about 10 cm thick, fill vertical brecciation fractures, cutting the quasi-sheeted dykes and trending about 120º. This occurrence has been the subject of radiometric dating by Spray and Dunning (1991; see Appendix) (Plate 4f). Net-vein complexes are also developed in several places. A relatively large mass of quartz-albite-phyric felsite occurs in a tectonic lens of the Lower Metagabbro beneath the Vord Hill Klippe, at Staves Geo [HU 653 887]. A large (c. 500 m long), ill-defined mass composed of both ‘plagiogranite’ variants occurs at Hesta Ness [HU 665 925] in north-east Fetlar, in tectonic contact with the body of serpentinite and overlain unconformably by the Funzie Conglomerate.

The leucotonalite variant is composed almost entirely of quartz and unzoned sodic plagioclase (Figure 36). Mafic minerals are interstitial and occur in little more than accessory proportions. Most of the mafic component has been replaced by aggregates of epidote, chlorite and opaque minerals; hornblende occurs in a few specimens and is the likely precursor. All the intrusive bodies of ‘plagiogranite’ have been deformed cataclastically. The boundaries of albite crystals have recrystallised to a fine granular mosaic, and quartz crystals have recrystallised to granular aggregates. Samples of the felsite variant have uniform, isotropic, microcrystalline matrices usually containing phenocrysts of quartz and/or albite, some of which show signs of resorption. Most are cut by thin shear zones and comminuted veins of quartz.

Interpretation

‘Plagiogranite’ samples (Table 10) contain less Ca, Mg, and Fe than samples of dacite (which have similar SiO₂ contents) from the Quasi-sheeted Dyke-swarm. They are similar in composition to ‘plagiogranite’ in other ophiolites, in particular in their very low K content Values of 0.04–0.31 wt per cent K₂O (Gamil, 1991) place them in ‘oceanic trondjhemite’ class of Coleman and Donato (1979).

‘Plagiogranite’ and the Norwick Hornblendic Schist were the last components to be emplaced in the Shetland Ophiolite before and during the obduction. The ‘plagiogranite’ and its variants cut intrusions of the Quasi-sheeted Dyke-swarm and, like them, do not occur below the Upper Metagabbro. Possible processes for the formation of ‘plagiogranite’ have been proposed (Lippard et al., 1986; Pedersen and Malpas, 1984; Spray and Dunning, 1991). These require either a magma chamber or melting within the ophiolite nappe, neither of which, it is argued, is possible within the Shetland Ophiolite. However, the formation of the Norwick Hornblendic Schist by adiabatic partial melting in the obduction thrust in the mantle (Flinn, 1993 and see below) at about the same time that the ‘plagiogranite’ was emplaced , had its parallel in another environment elsewhere in Shetland. In the course of rifting of the Laurentian Plate at the end of deposition of the Shetland Dalradian (see below), complete and partial melting of the mantle occurred, producing an outpouring of komatiitic and basic lavas. These rocks are intruded by veins of Na-rich 'keratophyre' of the composition of ‘oceanic trondjhemite’, that is of ‘plagiogranite’ (Moffat, 1987). It is possible that ‘plagiogranite’ is a product of adiabatic melting of the mantle. Spray and Dunning (1991) obtained a U/Pb zircon age of 492 ± 3 Ma for a ‘plagiogranite’ intrusion at South Ship Geo [HP 623 043] ((P536027), (P536030)).

The Imbricate Zones

The Lower Imbricate Zone (LIZ) and Middle Imbricate Zone (MIZ), which formed beneath the Lower and Upper nappes respectively, consist of tectonic slices derived from various sources (Figure 27). No mélange- type rocks occur. Some slices derive from the ophiolite nappes, some from the Valla Field and Saxa Vord blocks (the basement on which the nappes rest), and some from units of metasedimentary rock that were deposited on the Lower Nappe after it was obducted and before it was over-ridden by the Upper Nappe. The latter, which include units of metavolcaniclastic and metasedimentary rock containing eroded ophiolite debris, are united within the Unst Phyllite Group. Also present are tectonic slices of hornblende schist, of a type associated commonly with ophiolite-complexes; these comprise the Norwick Hornblendic Schist. Other than the basement rocks, none of these lithologies occur elsewhere in Shetland, or are related to lithologies anywhere in Shetland outside the ophiolite-complex.

Unst Phyllite Group

The Unst Phyllite Group encompasses six tectonically defined and lithologically distinct units: the Muness Phyllite, Norwick Graphitic Schist, Norwick Phyllite, Leagarth Pelite, Gruting Greenschist and Funzie Conglomerate. All six occur in the MIZ, while slices of the Muness Phyllite and Norwick Graphitic Schist also occur in the LIZ (Figure 27). The imbricate nature of the zones, and generally poor exposure, make it impossible to arrange the units in stratigraphical order. The Leagarth Pelite and Funzie Conglomerate occur only in Fetlar, and the Norwick Phyllite crops out only in Unst.

Muness Phyllite

The Muness Phyllite consists of low-grade, dominantly silty, semipelitic and often finely laminated, metasedimentary rock (Plate 5a) and (Plate 5b). The rock contains clastic grains of quartz and albite, is rich in quartz segregations, and is strongly deformed. The unit crops out in the type area of Muness within the MIZ on Unst, and extends south from there into Fetlar, west of the Vord Hill Klippe. It also crops out within the LIZ on Unst.

The rock is generally grey to black and laminated locally in combinations of grey, black, cream, green and reddish brown (Plate 5a). The laminae are displayed clearly in the field only on sand-polished surfaces at the back of beaches ((P536052), (P536059)). On a larger scale, the rock is banded into coarser grained siliceous layers and finer grained micaceous layers. The coarser bands are generally 20–30 cm thick (P534818), but considerably thicker examples occur. Centimetre-scale graded bands recorded in several places (e.g. the north [HU 6044 9968] and east [HU 6101 9844] coasts of Uyea, Quida [HP 6247 0216] and Littlegarth [HU 6164 9998] in south-east Unst, and west of Vord Hill on Fetlar [HU 6084 9366]) and are the right way up.

Thin sections of finer grained phyllite reveal alternating laminae of muscovite and/or chlorite and fine-grained quartz. The nature and condition of the laminae varies considerably, reflecting a range of sedimentary processes and the effects of deformation. Micaceous minerals form a schistosity parallel to the lamination. However, later crenulation folding has distorted this into a spaced cleavage that cuts the lamination ((P534317), (P534378)). Clastic grains of quartz and/or albite are common. Carbonate impregnation is distributed widely and very irregularly. Reddish-brown laminae result from weathering of carbonate-impregnated, quartz- rich laminae ((P534316), (P533627)). Many samples of the Muness Phyllite contain minor proportions of small, ragged crystals of biotite that grew after the schistosity developed. Several thin sections contain crystals of garnet that have grown across the schistosity and are not affected by later deformation, e.g. at Yogli Geo [HU 649 916].

The unit contains variable proportions of ‘plagiogranite’ debris, in the form of quartz-albite-phyric felsite, leucotonalite and granophyric leucotonalite. These lithologies occur as pebbles in conglomerate ((P534342), (P534367)) and as fragmental debris, from granule grade to coarse breccia, in phyllitic rock. In thin section, the felsite occurs as patches of very fine-grained quartz (and albite?), commonly with phenocrysts of quartz and/or albite. In hand specimen, such rocks tend to be more compact and felsic than ‘normal’ semipelitic phyllite. Several highly deformed, lenticular masses, about 1 m wide, of quartz-albite-phyric felsite crop out on the coast north of Muness, north-east of Castleton ( [HP 6307 0168], [HP 6322 0148], [HP 6308 0168]). Another crops out to the south-east of Millburn [HP 608 003], on the south coast of Unst. Clastic grains of quartz and less common albite that are distributed unevenly throughout the Muness Phyllite are probably erosional debris derived from intrusions of quartz-albite-phyric felsite.

Two types of conglomerate form bands, up to 1 m or so thick, in the Muness Phyllite. One conglomerate type is characterised by pebbles of ‘plagiogranite’, including tectonically elongated pebbles of quartz-albite-phyric felsite and subspherical pebbles of leucotonalite and granophyric leucotonalite; it also contains pebbles of quartzite and of other fine-grained siliceous lithologies of unknown protolith. The pebbles are supported in a matrix of semipelitic Muness Phyllite. This type was recorded at Ramnageo Ness [HU 6243 9974], on the east side of Uyea ( [HU 613 993], [HU 612 989], [HU 620 985], [HU 605 982], [HU 606 983], [HU 610 980]), and on Wedder Holm [HU 615 975]. The other type of conglomerate contains close-packed, elongate pebbles of metagabbro, with features (in hand specimen) typical of both the Lower Metagabbro and Upper Metagabbro. This type also contains pebbles derived from ophitic and non-ophitic variants of the Quasi-sheeted Dyke-swarm, and pebbles of hornblende diorite, ‘plagiogranite’ including quartz-albite-phyric felsite, granophyric and non-granophyric leucotonalite, and metajasper (pink quartzite; recorded at Cliva [HU 599 998] on Uyea and on Stongir Holm [HU 5970 9466]) (P534471). A single pebble of two-feldspar-quartz-garnet- tourmaline aplite contains the only recorded occurrence of K-feldspar in the entire ophiolite-complex (other than in the Funzie Conglomerate). Phemister (1929–63) reported a pebble of jade-type nephrite. No pebbles of serpentine-bearing rock have been recorded. The matrix consists of short needles of actinolitic amphibole in chlorite. The conglomerate type containing metagabbro clasts crops out on the islands of Urie Lingey [HU 595 955] and Stongir Holm [HU 597 947], on nearby skerries and baas (submerged reefs), and on the north shore of Fetlar between Wester Gors Geo [HU 604 935] and Easter Gors Geo [HU 610 937] where several bands of conglomerate containing ‘plagiogranite’ pebbles also crop out. One hundred and fifty pebbles extracted from conglomerate on the north shore of Fetlar have an average spherical equivalent diameter of about 5 cm. Pebbles in the conglomerate exposed at Ramnageo [HU 6243 9974] in southern Unst are about the same size (based on measurements made from joint surfaces).

Occurrences of microconglomerate (with clasts of small pebble-to-granule grade) contain aligned, flake-like clasts of phyllite. Some of these are probably tectonically disrupted laminated phyllite, but others are well formed pebbles probably derived by erosion of laminated phyllite. Two small lenses of fine-crystalline marble occur in the Muness Phyllite on Uyea, south of The Hall [HU 605 983]. Quartz veins, and less commonly quartz- calcite segregation veins, are common and even profuse in places (e.g. on Wedder Holm [HU 614 974]). Lamination and schistosity in the Muness Phyllite are deformed strongly against the veins, so they clearly developed before (or at an early stage of) metamorphism. Late, well- formed, planar, quartz-filled tension veins are normal to the lineation locally, and cut all other structures.

The sediment must have eroded, at least in part, from upper levels of the Lower Nappe. However, the absence of serpentinite debris and the common presence of metagabbro and ‘plagiogranite’ debris suggest that the debris was sourced from east of the present eastern limit of the nappe. K₂O concentration generally exceeds that of Na₂O in samples representing units of the Unst Phyllite Group (Table 11), and muscovite is a major component of Unst Phyllite Group rocks. As the ophiolite is almost devoid of potassium, the Muness Phyllite must contain detritus derived from K-bearing basement lithologies. Heavy mineral populations in <63 μm fractions extracted from samples of sandstone, semipelite and laminated phyllite from the south coast of Muness are dominated by chromite, while magnetite is absent. Around 1 per cent of the population extracted from each lithology consists of columnar crystals of brownish tourmaline (schorl). Tourmaline in Dalradian rocks to the west of the Shetland Ophiolite-complex is coarse-grained and allotriomorphic, whereas tourmaline mineralisation in the metagabbro is similar to that in the heavy mineral fractions.

Analyses of typical samples of Unst Phyllite Group units are presented in (Table 11); sample 7 (albite schist) comes from a small group of close-spaced exposures north of Brough Lodge in north-west Fetlar [HU 589 931] that appear to be Muness Phyllite, though they occur in a very poorly exposed area in which other exposures are of Gruting Greenschist. A thin section of the sample reveals chlorite-muscovite phyllite, rich in porphyroblastic crystals of albite (albite schist), with many large plates of muscovite that overgrow the schistosity.

Traces of gold have been reported (Buchanan and Dunton, 1996) in a prominent strata-parallel, partly oxidised, pyrite-mineralised zone several metres wide at Ramnageo Ness [HU 627 996] (P536033). Limonite-stained zones occur in coastal exposures of the Muness Phyllite unit in other places, but none is as well developed as this one. Gold concentrations of up to 1.2 ppm are associated with arsenic in channel samples. Channel samples from the occurrence also contain up to 4000 ppm of Cr, more than ten times the levels found more generally in the phyllite (Table 11). The Cr values are associated with ppb-level concentrations of Ni, Pt and Pa. However, X-ray backscatter images of polished sections showed no gold or arsenic (Chapter 12).

Despite an absence of serpentinite debris, the presence of elevated concentrations of Cr and associated elements appears to indicate that ultramafic parts of the ophiolite were exposed to erosion in the catchment area supplying the basin in which the Muness Phyllite accumulated. Micron-thick flakes of molybdenite occur in one sample, perhaps indicating that molybdenite-bearing veins in the Lower Nappe (recorded by Neary and Prichard, 1985) formed prior to deposition of the Muness Phyllite.

Norwick Graphitic Schist

In Unst, the Norwick Graphitic Schist rests on the Lower Nappe in the type area at Norwick [HP 630 130] (Read, 1934a), and farther north underlies the Hill of Clibberswick Klippe [HP 648 136] of the Upper Nappe. Both occurrences are in the Middle Imbricate Zone. There are further occurrences in Unst in the Lower Imbricate Zone, west of Baltasound ( [HP 605 085], [HP 607 075]) and at Belmont [HP 565 005]. The most graphite-rich occurrences are in the Middle Imbricate Zone on Fetlar, at Tresta [HU 612 903], Yogli Geo [HU 648 916], and to the east of Honga Ness [HU 653 913]. The presence of weakly graphitic zones in the Muness Phyllite suggests the Norwick Graphitic Schist and Muness Phyllite may be parts of the same depositional unit (P536059).

The Norwick Graphitic Schist consists of laminated siliceous phyllites, which commonly contains sufficient graphite to soil the fingers. They grade into black phyllite (e.g. at Belmont [HP 565 005]), which in thin section appears to be graphite-free. The composition ranges from quartzitic to pelitic, but is dominantly semipelitic. Exposures in the Norwick area contain fine-grained (c. 0.1 mm) crystals of garnet.

The Muness Phyllite and Norwick Graphitic Schist units were deposited below wave depth, in fluctuating redox conditions. While graphite indicates the presence of living organisms (algae?), preserved laminae suggest an absence of bioturbation and probable deposition in fresh water. A systematic search by the collector RJA Eckford in 1954 yielded no fossils in the graphitic schist of the Litla Sand area, immediately east of Hoga Ness (Phemister, 1929–63).

Norwick Phyllite

Ground underlying a series of poor, widely separated exposures on the hill immediately west of Norwick [HP 643 140] is assigned to the Norwick Phyllite. The fine-grained, silvery-looking pelitic phyllite composed of muscovite, chlorite, quartz, garnet (0.1 mm) and sparse crystals of tourmaline crops out in the Middle Imbricate Zone north and west of the two occurrences of Norwick Graphitic Schist. The rock is similar petrographically to the Hevda Phyllite of the Saxa Vord Group, which crops out nearby at The Cliffs [HP 654 150] in the Lower Imbricate Zone; this part of the Norwick Phyllite may therefore be a tectonic slice of the Hevda Phyllite.

Ephemeral exposures on the beach at Norwick, between Upper Nappe serpentinite exposed at The Taing [HP 653 146] and Lower Imbricate Zone rocks exposed in The Cliffs, are not like the Hevda Phyllite. Immediately west of The Taing [HP 652 145], below the thrust underlying the Hill of Clibberswick Klippe, a small exposure of phyllite hosts a boudinaged sheet of lamprophyre (Read, 1934c) and veins of plagioclase pegmatite. Read (in IGS, 1968a) identified an occurrence of phyllite conglomerate in these rocks at The Cliffs [HP 6527 1487] (P533513), which is now at low tide level and under a partial covering of boulders derived from the Skaw Granite. These rocks are difficult to correlate with those of any other unit.

Leagarth Pelite

Fine-grained, pelitic, chloritoid-chlorite-muscovite phyllite assigned to the Leagarth Pelite crops out only on Fetlar: at Ness of Gruting [HU 650 915], to the north-west, and west of Leagarth [HU 630 900], and as a large enclave in the Lamb Hoga Fault south of Sugil [HU 588 924]. The boundaries of these units against rock other than serpentinite are not exposed. The Ness of Gruting occurrence may be conformably interbanded with the Muness Phyllite, but the other occurrences are tectonic slices. By contrast with the Muness Phyllite, the very fine-grained, muscovite-dominated rock lacks banding or lamination and contains laths of primary chloritoid, 1–2 mm long and orientated parallel to the schistosity. Samples of Leagarth Pelite have the composition of chloritoid-type pelite (Table 11) and are similar petrographically to regional-type chloritoid phyllite in the Saxa Vord Group and to chloritoid phyllites in the Clift Hills ‘Division’ in the south Mainland of Shetland.

Gruting Greenschist

The Gruting Greenschist (Figure 28) crops out within the Middle Imbricate Zone in both Unst and Fetlar. In Fetlar, it is displayed magnificently in cliffs at the north end of Ness of Gruting [HU 653 920], in cliffs on either side of the serpentinite lens at Vallahamars [HU 616 903], and along much of the east shore of Wick of Gruting in the area of Strand [HU 660 914]. There are small coastal exposures at Smiddy Geo [HU 6090 9045] and Yogli Geo [HU 648 917] in Fetlar, and on the beach at Sand Wick in Unst [HP 631 022]. The unit underlies a large area of Mires of Oddsetter in west Fetlar [HU 590 930], but is exposed very poorly there. In the Norwick area of Unst, an outcrop of Gruting Greenschist is known almost entirely from ephemeral pits and trenches in a former military village. Contacts with other units are obscured by late deformation. A well-exposed contact between Gruting Greenschist and Muness Phyllite on the beach south of Sand Wick [HP 623 022] is difficult to interpret due to deformation. Contacts between the different outcrops of Gruting Greenschist and other units in the Middle Imbricate Zone are obscured, sheared or phyllonitised, and it is therefore not possible to determine whether they were deposited in their present position on the Lower Nappe or were transported there tectonically during emplacement of the Upper Nappe. ‘plagiogranite’ (see above), which occurs as erosional debris in the Muness Phyllite, forms several minor intrusive veins in the Gruting Greenschist; the Gruting Greenschist may therefore be older than the Muness Phyllite.

The Gruting Greenschist consists of dark, fine-grained phyllite that varies little in appearance in hand specimen. Amphibole-bearing varieties are harder than amphibole-free varieties; the former can be difficult to distinguish in the field from Norwick Hornblendic Schist type 2 (described below), though the two are distinguished readily in thin section. A crude, lenticular, metre-scale layering discernible in cliff sections is due to slight differences in colour and texture on weathered surfaces. Sparsely distributed bands of microconglomerate contain parallel-orientated ellipsoidal clasts up to 1 cm long.

About 200 samples of the Gruting Greenschist were collected for sectioning from the best exposures in Fetlar. The rock is characterised by an exceedingly small grain size. Most samples are formed entirely of a very fine-grained to fine-grained matrix (generally about 0.01 mm; rarely up to 0.1 mm) composed dominantly of feldspar, chlorite and epidote; also present are titanite (with and without cores of ilmenite), biotite (commonly as late porphyroblasts), amphibole (generally colourless fine rods and needles), and carbonate minerals (mostly as cross-cutting veins). Metamorphic recrystallisation has coarsened the matrix locally. In many samples the matrix contains fine- and medium-grained fragments of broken plagioclase crystals. Chlorite and epidote commonly form modally distinct laminae (visible only in thin section) that are sometimes complexly and irregularly folded. Most of these minerals occur as grains only a little larger than the (very fine-grained) feldspar. Where a schistosity is discernible it is defined by chlorite flakes.

Bands of microconglomerate consist of fragments of highly feldspathic, very fine-grained greenschist embedded in a matrix of less feldspathic greenschist (P534822), (P534812). In several samples the matrix is composed of randomly orientated laths of plagioclase c. 0.2 mm long. Some samples are free, or nearly free, of feldspar. They are instead composed of equigranular, very fine-grained and deformed aggregates of epidote, titanite, amphibole and carbonate minerals. Clasts of texturally distinct varieties of greenschist preserved within single thin sections of microconglomerate indicate that the textural variants developed before deposition and metamorphism.

Fine-grained (about 0.1 mm) epidote-feldspar-chlorite greenschist occurs in contact with partly steatitised serpentinite at Gallafield [HU 614 902], near Strand [HU 658 913], and at Aith [HU 6302 8912]. Greenschist near the contact contains magnificent rich green hornblende ‘garbenschiefer’ (radiating, elongate crystals of porphyroblastic hornblende), forming near hemi-circles. Such occurrences probably developed in association with steatitic metasomatism of the adjacent serpentinite. They are not a thermal aureole to the body of serpentinised ultramafic rock, despite giving the impression in the field of increased crystallinity at the contact.

Thirty-three analyses of Gruting Greenschist samples are available (Table 12) and (Table 13). Plotted on a TAS diagram (Figure 29), the data show considerable scatter, very widely ranging compositions (SiO₂ range 37.97 to 65.95 wt per cent; Na₂O + K₂O range ~0.5 to 10.0 wt per cent), and no preferred concentrations. There is no textural or mineralogical evidence for intermixing with non-volcanic sediments (e.g. siliciclastic debris). The very wide range of compositions shown in (Figure 29) is divided into alkaline and subalkaline groupings on the basis of MgO content, supported by such discrimination plots as Ti–Zr, Cr–Y and N-MORB normalised trace element patterns (spider diagrams).

Funzie Conglomerate

The Funzie Conglomerate is a very large occurrence of a deformed conglomerate in eastern Fetlar (Figure 30) and has a similar metamorphic and tectonic state to units in the Unst Phyllite Group of the Middle Imbricate Zone. It is therefore interpreted to be part of this sequence. Like other components of the Middle Imbricate Zone, the Funzie Conglomerate was deposited before the Upper Nappe was emplaced. The unit is polymict but dominated by pebbles of quartzite (Flinn, 1956). The conglomerate, which extends across about 5 km² and exposes around 1 km thickness normal to the schistosity, is displayed magnificently in cliffs bounding the east end of Fetlar. Hibbert (1819, 1822) and Anon (1926–27) provided early maps and descriptions, and Phemister (in Wilson, 1931) produced a lithological and petrographical account. Flinn (1956) and Hafez (1972) have described the structure of the unit.

The entire mass of exposed conglomerate has a remarkably uniform appearance in the field, other than a systematic variation in clast shape (described below). Rare clast-free, or almost free, bands of psammite (P535979) define a lithological layering, which consistently dips moderately to steeply west. These occurrences are typically lenticular or poorly defined, up to 30 cm thick, and are composed of rock that is indistinguishable from that forming the matrix of adjacent conglomerate. They are distinguished only by the lack of clasts. Several such occurrences extend north–south over a distance of approximately 1 km to the west of Funzie Bay [HU 662 898] and a number of other occurrences are noted (Flinn, 1956). This lithological layering is transected obliquely by the superimposed linear tectonic fabric trending north-north-east–south- south-west, for which the conglomerate is renowned.

The steep to vertical western boundary of the Funzie Conglomerate with the Lower Nappe of the ophiolite- complex is well exposed in the western cliffs of Strandburgh Ness [HU 669 929] and is partially constrained inland where the ultramafic rocks of the Lower and Upper nappes are clearly distinguished by magnetic data. The boundary extends north-north-east–south-south- west from Strandburgh Ness and is thus oblique to any observed lithological layering; layers of phyllite occur between the conglomerate and ultramafic rocks west of Loch of Funzie (Flinn, 1956) but no depositional contact can be observed. The conglomerate is not clearly unconformable on the rocks of the Lower Nappe. The northern sector of the contact seems therefore most likely to be fault defined; there are subparallel faults in rock immediately west of the boundary. South of Loch of Funzie, the conglomerate is overthrust by rocks in the Upper Nappe. This contact is very well displayed in cliffs adjacent to Staves Geo [HU 654 887].

The conglomerate matrix is phyllitic semipelite, rich in sand- and granule-grade grains of quartz, and of uniform appearance. Muscovite is the dominant phyllosilicate mineral. The clasts are matrix-supported, but differential deformation of clasts and matrix has brought the clasts into contact locally. Clasts of quartzite form about 75 per cent of the rock according to estimates by Read and Phemister (Phemister, 1930, 1931) [see references in Flinn, 1956), while clasts of leucotonalitic rock (Table 14) form about 3 per cent. The latter are Na₂O rich (5.82 to 7.51 wt per cent) and have K₂O in the range 0.83 to 2.62 wt per cent (Gamil, 1991); in terms of composition they are therefore typical of continental-type trondhjemite, and are not a part of the ophiolite ‘plagiogranite’ suite (see above). Clasts of marble are extremely rare. Clasts of phyllitic rock are also present but are easily overlooked; they have a similar mineralogical composition and texture to the matrix of the conglomerate and to phyllite of the Unst Phyllite Group. Like the matrix, these are deformed more strongly than the quartzite clasts. The predominance of quartzite as a clast lithology, with subordinate proportions of leucotonalite and marble, suggest the Funzie Conglomerate is sourced, at least in part, from the basement.

The quartzite clasts range from a few cm³ to ‘6 x 3 x 1 ft³’ (Phemister, 1929–63). Five hundred and thirty-three clasts extracted from conglomerate at 33 sites distributed throughout the unit average 14.2 cm spherical equivalent diameter (SED). Within individual sampling sites the averaged SED ranges from 6.0 to 26.3 cm with statistically significant differences.

Tectonic structures in the Unst Phyllite Group

The Muness Phyllite comprises an L-tectonite or L-dominant tectonite. The sedimentary lamination in the Muness Phyllite along the western part of the south coast of Muness dips steadily east at a low angle, and is everywhere parallel to the schistosity ((P536046), (P536059)). In more micaceous varieties, the lamination is cut by an impersistent, and more steeply east-dipping, spaced-cleavage developed on small-scale crenulations (strain-slip cleavage) ((P534609), (P536057)). In some places there are two such cleavages, of different ages. The continuity of sedimentary lamination and bedding is also interrupted in places by minor shears and groups of L-tectonite folds when viewed in coastal sections normal to the tectonic lineation. The folds range from centimetre to 10 m scale; the folding appears chaotic (L-tectonite style) when viewed parallel to fold axes (P534334), being overturned to the east or west, or upright. In less micaceous rock, phyllosilicate grains form an L-tectonite-type girdle fabric, giving the rock a lineation parallel to fold axes and the long axes of deformed clasts in the conglomerate layers. The conglomerate at Ramnageo [HU 6243 9974] is an ideal (Y=Z) L-tectonite conglomerate (Plate 5c) and (Plate 5d). The Muness Phyllite has been deformed by a constrictional deformation that caused it to fold and extend parallel to the north-north-east-trending fold axes, contracting equally in all directions normal to those axes.

The Muness Phyllite suffered a later deformation resulting in sets of closely spaced kink folds with vertical axial planes orientated normal to the L-tectonite lineation ((P533419), (P534314)) (Flinn, 1952). These are displayed well on the north and south coasts of Muness [HP 620 010], where they show opposite senses of rotation as viewed from the east; clockwise in the north and anticlockwise in the south.

Graphitic phyllite in the Norwick Graphitic Schist has been folded, crenulated and disrupted to such a degree that the structure visible in sections cut normal to the lineation and fold axes is indecipherable (L-tectonite). Laths of primary chloritoid in the Leagarth Pelite are aligned with a fabric formed by muscovite flakes and distort the later developed spaced cleavage.

The Gruting Greenschist has a weakly discernible, orthorhombic L–S-tectonite fabric. Parallel-orientated ellipsoidal clasts up to 1 cm long in sparsely distributed bands of microconglomerate clearly conform to the L–S-tectonite fabric. The lineation is subhorizontal, north-north-east-trending, and parallel to the lineation, fold axes, cleavage intersections and pebble elongation in the Muness Phyllite.

Measurements of X, Y and Z dimensions of quartzite pebbles demonstrate a systematic variation in strain state across the outcrop of the Funzie Conglomerate (Flinn, 1956). A total of 533 pebbles were measured at 35 sampling stations and averaged. They were found to delineate a number of zones in which the strain state, and thus the deformation process, are judged to have been similar (Figure 30)a and b. The clasts become increasingly elongated in X through zones VI to III (Figure 30)b, from a least deformed state around Head of Hosta [HU 677 918] towards a maximum elongation (X/Y) of c. 2.5 at The Snap [HU 658 878]. The average sphericity of the clasts in their least deformed state is 0.54 (where sphericity equals diameter, d/X; Flinn, 1956). The majority of the data tracks along the plane strain axis, plotting just into the field of constrictional strain (Figure 30)b. At The Snap [HU 658 878], the clasts occur within L–S-tectonites as triaxial ellipsoids (X>Y>Z) of almost ideal shape in which Y = (XYZ)1/3 (Plate 5e) and (Plate 5f) (P535908), (P535985).

From the Snap [HU 657 878] to Staves Geo [HU 654 887] elongation in X decreases slightly, but is still positive, while elongation in Y increases rapidly until Y becomes equal to X and clasts of quartzite are flattened into almost perfect oblate spheroids (X=Y>Z) at Staves Geo at the western boundary of the conglomerate. These distinctive oblate clast shapes (S-tectonite) occur in the immediate footwall of the thrust carrying the Upper Nappe of the ophiolite-complex over the conglomerate (P534902). The systematic variation of the shapes of the pebbles from south to north across the conglomerate is interpreted in terms of a single continuously varying progressive deformation.

The longest axes of the clasts are approximately parallel, so that the conglomerate has a marked lineation along a general north-north-east trend. In the northern part of the area, where measured strains indicate the weakest deformation, the lineation is horizontal, or plunges a few degrees towards the north-north-east. Farther to the south-south-west, the lineation progressively steepens and plunges up to 20–30º to the southwest.

The attitude of the clasts is everywhere reflected in the attitude of the micaceous minerals in the matrix. Thus, near Staves Geo a preferred parallel orientation of phyllosilicates (mainly chlorite and sericite with subsidiary muscovite and biotite) defines an S-surface and the oblate clasts lie in the plane of the schistosity, which dips towards the south-west beneath the Upper Nappe. In the northern area, the same phyllosilicate minerals define a lineation and the schistosity becomes progressively more difficult to identify. The clasts are orientated with their X-axes parallel to a matrix extension lineation defined by a girdle (in L-tectonite areas) to partial-girdle (in L–S-tectonite areas) preferred orientation of mica flakes in the matrix. The Y-axes of the clasts exhibit a girdle to partial girdle orientation similar to that of the matrix mica at each locality. The geometry described so far is interpreted to be of orthorhombic symmetry (with higher symmetry at Staves Geo), indicating a coaxial deformation of both clasts and matrix. During deformation, the matrix was able to flow past the clasts, causing them to rotate; this produced a high degree of preferred orientation of clast axes and fine striations on the X-axis extremities of the clasts.

Quartz crystals within the deformed clasts are generally c. 0.3 mm and vary up to 1 mm. The shapes of quartz crystals within individual clasts reflect the shape of that clast i.e. they have similar k-values, where k= (X/Y)/(Y/Z). Clasts with L–S-tectonite shapes have quartz crystals of L–S-tectonite shape while clasts with S-tectonite shapes have quartz crystals of S-tectonite shape. However, the crystals have a very much higher sphericity than the clasts, indicating that clast deformation involved grain-boundary sliding within the confines of the deforming clast (analogous to the movement of individual grains in an elastic bag of sand).

Analysis of the quartz mineral optics shows that the quartz crystals exhibit a weak preferred c-axis orientation in which the symmetry of the c-axis fabric conforms to the symmetry of the shape of the host clast (Flinn, 1956; Hafez, 1972). Flattened oblate clasts (S-tectonites) at Staves Geo have a point concentration parallel to Z (i.e. normal to the S-plane). Quartz crystals in clasts associated with L–S fabrics demonstrate partial girdles of c-axes in the plane normal to X (the extension lineation).

In all parts of the unit many clasts are cut by sets of fractures ((P535969), (P534918), (P535977)). These ‘fracture cleavages’ are parallel to the average direction Y at any one locality and dip south-west at about 45º to the average YZ plane. It is usually possible to find at least an occasional clast at each locality with it fracture cleavage planes dipping north-east and such geometries are dominant in several scattered localities. The cleavage does not always pass right through the clast, but movement on each fracture cleavage causes a slight offset of the surface of the clast which is invariably down to the south-west by a few millimetres on each fracture. This weak monoclinic (simple shear) deformation postdates the main orthorhombic (pure shear) deformation. The observation that movements on these fracture cleavages are orientated parallel to the average Y-strain axis at a particular locality, rather than an individual clast Y-axis, suggests that these more brittle failures were developed strictly later than the more ductile clast elongation but could have occurred in response to a continuous deformation rather than to discrete and episodic imposed stress. Both the orientation and sense of movement on these fractures are as would be expected of failure under pure shear.

A still later orthorhombic constriction deformation created a set of tension veins cutting clasts normal to their X (elongation) axes. These veins are tens of centimetres long, 2–3 mm wide, and filled with pink adularia ((P535972), (P535975), 535986). They cut the clasts, fracture cleavage in the clasts, the conglomerate matrix, and sheets of lamprophyre that intrude the conglomerate.

The orthorhombic (and locally higher) symmetry of the conglomerate clasts and matrix are interpreted here as the result of three-dimensional pure (co-axial) shear. It is further argued that clasts with flattened oblate spheroid shapes at the contact with the overlying Upper Nappe were also the result of co-axial shear and that the conglomerate has not been deformed by simple shear, even at the thrust contact, during nappe emplacement. Deformation is best envisaged as the result of a constrictional deformation (X/Y > Y/Z) in which the conglomerate was squeezed by more competent rocks enclosing it, and forced to flow in a ductile manner by grain-boundary sliding. The only monoclinic element in such a deformation is caused by gradients in the flow magnitude.

Basement rock in the imbricate zones

Tectonic slices of rock from both the Saxa Vord and Valla Field blocks, which form the metasedimentary basement below the ophiolite-complex, occur in both the MIZ and LIZ.

Saxa Vord Group

The LIZ includes slices of rock derived from the Saxa Vord Group. The northernmost lens in the LIZ on Unst consists of pelitic phyllite, which has been assigned to the Hevda Phyllite. The rock is well exposed in The Cliffs at Nor Wick [HP 653 150]. The lens is truncated to the north by the Skaw Granite, also at The Cliffs [HP 655 150], along a sharply defined, welded contact dipping about 45°. It is cut out to the south within the LIZ by a tectonic lens of Norwick Hornblendic Schist type 1.

The lens is composed of pelitic phyllite that is indistinguishable lithologically from other exposures of the Hevda Phyllite (Chapter 3) and from the Norwick Phyllite (see above) of the MIZ. Phyllite exposed in The Cliffs is fine-grained, unlaminated, chlorite-quartz-muscovite pelite. At the contact with the Skaw Granite there is a weak development of contact metamorphism (Chapter 7). Similar metamorphic effects occur around a small mass of appinitic hornblende diorite with attached lamprophyre sheets at The Cliffs at [HP 654 150]. The phyllite has a similar chemical composition to samples elsewhere from the Hevda Phyllite (Chapter 3). Another lens of the Hevda Phyllite occurs in the LIZ north-east of Loch of Watlee [HP 598 057].

Poorly exposed tectonic slices of the Saxa Vord Pelite crop out intermittently within the LIZ at Feall [HP 630 133], Cliff [HP 603 110], Caldback [HP 603 073], the west side of Hill of Voesgarth [HP 607 077], Bordi Knowe [HP 589 044], Gardin [HP 573 023], and as far south as Belmont [HP 560 007]. These include slices of regionally metamorphosed, fine- grained chloritoid phyllite with shimmer aggregates after andalusite, and slices of thermally metamorphosed coarser variants containing chloritoid, kyanite and staurolite.

Tectonic slices of the Queyhouse Flags crop out within the LIZ on the north-west of Hill of Voesgarth [HP 605 081] and at Gudon [HP 598 060]. Substantial amounts of kaolin are present at the latter locality (Chapter 12).

Valla Field Group

Slices of rock derived from the Valla Field Group occur in both the LIZ and MIZ. A few very small exposures in ground between the Burra Firth Lineament (which is assumed to be the western edge of the LIZ) and the imbricate slices immediately west of the Lower Nappe between Loch of Watlee [HP 594 057] and Loch of Cliff [HP 599 103] consist of rock that has similarities to the adjacent Burra Firth Formation; a large lens of the Burra Firth Formation is assumed to underlie the area. Lenses of kyanite-staurolite-chloritoid-garnet schist, similar to chloritoid-bearing pelitic components of the Valla Field Schist (and to the Snarravoe Pelite of the Westing Gneiss [Westing Group]), crop out in the LIZ of Unst to the south of Loch of Watlee ( [HP 580 033], [HP 577 030], [HP 573 023] and [HP 560 010]). Lenses of the Loch of Cliff Limestone are exposed in the LIZ at Cliff [HP 602 111], south of Baliasta [HP 602 088], and Loch of Watlee [HP 596 058]. The rock is coarsely crystalline, grey marble containing isolated accessory grains of quartz, plagioclase, muscovite and biotite. Several thin stripes of biotitic semipelite occur within it at Cliff [HP 602 113]. The rock resembles closely the Weisdale Limestone Member and Laxfirth Limestone of the Whiteness Division in Mainland Shetland (which is located stratigraphically between the Valla Field and Saxa Vord groups) (Flinn et al, 1972; Mykura, 1976). It does not resemble the Herma Ness Marble [HP 603 183] or Sunkir Marble [HU 623 875] of the Valla Field Group, or the Westing Limestone of the Westing Group.

The MIZ of Fetlar contains several tectonic lenses of muscovite-chlorite (after biotite)-quartz-plagioclase gneiss, with segregations of quartz and plagioclase (leucosome gneiss). The rock is similar lithologically to the Valla Field Gneiss (Westing Group). However, it is of a type that is common throughout Shetland so the name Strand Gneiss is used here to distinguish this unit. The largest mass, which crops out in cliffs at Everland [HU 656 913], is about 100 m across and consists of highly fractured and crushed rock. Lenses of similar rock, up to 1 m across, are associated with Norwick Hornblendic Schist type 2 on Houbie beach [HU 626 905], in a mylonitic zone at Aith Ness [HU 633 895] ((P534893)), and at Virva [HU 646 920].

Augen Gneiss of uncertain affinity

Several metres of augen gneiss interbanded with phyllonite are exposed against or within the basal thrust of the Lower Imbricate Zone at Hoga Ness, south-west Unst [HP 557 006] (Figure 8) (P534274). This occurrence has some field resemblance to the Skelladale Augen Gneiss, the eastern boundary of the Boundary Zone. It contains 0.5 to 1 cm aligned microcline augen with small plagioclase inclusions. They occur in a phyllonitised muscovite>biotite matrix (0.1–0.2 mm).

Tectonic structures in slices of basement rock within the imbricate Zones

Folds with half wavelengths of up to several metres occur in phyllite in the northernmost lens of Saxa Vord Group rock at The Cliffs [653150] within the LIZ on Unst (Hevda Phyllite). The phyllites are very intensely crenulation-folded, with a spaced cleavage approximately parallel to the axial planes of larger folds. The rock contains two lineations; one due to the crenulation axes, the other to intersection of the spaced cleavage with compositional banding. The intersection lineation generally plunges about 30° more steeply than the gently north-plunging crenulation lineation. Both lineations trend about north-east, parallel to the coast. The phyllite is an L-tectonite, due to the folding and to mica girdles in mica-poor semipelite. The phyllite has been kink- folded on axes plunging at a shallow angle to the north-west locally for several hundred metres south-west of the contact with the Skaw intrusion. These locally form spectacular conjugate sets of S and Z folds near Swartling [HP 654 152] (P533422).

Ophiolite rock in the imbricate zones

The LIZ contains only one lens of ultramafic rock derived from the ophiolite; a small, steatitised body no more than a few metres across which is located in the basal thrust of the zone at Norwick [HP 638 145] in Unst. The ultramafic klippe-like mass at Muness [HP 635 010], and the one forming Sound Gruney [HU 580 960], are formed of clinopyroxenite, wehrlite and metadunite derived from the crustal part of the ophiolite. As they underlie the mantle-derived Upper Nappe, they must be tectonic slices within the MIZ.

The MIZ in Fetlar contains many tectonic slices derived from the ophiolite. These range from metre to kilometre scale, and include both serpentinite and metagabbro. It is rarely possible to identify the protoliths of serpentinite slices, as they are usually metamorphosed to antigorite. The largest slice, at Houbie [HU 625 910], is in contact locally with the Upper Nappe but elsewhere is separated from it by a zone of intense shattering, as occurs east of Skutes Water [HU 625 920] and, between Houbie and Aith Ness, by a complex association of MIZ lenses. It is formed mainly of antigoritised metaharzburgite but also contains bodies of metagabbro, north and south of Houbie respectively ( [HU 622 910], [HU 620 903]), clinopyroxenite-wehrlite, at Houbie itself [HU 622 908], metadunite, south-east of Houbie [HU 624 905], and bodies of Norwick Hornblendic Schist type 3 (see below), also south-east of Houbie [HU 622 905]. Small, apparently intrusive bodies of metagabbro, pegmatitic metagabbro and quartz-albite-phyric felsite occur within it at Blo Geo, south-west of Houbie [HU 6195 9025]. Coastal exposures of this lens, south-east of Houbie [HU 625 904], are very shattered and sheared, and steatitised locally.

The nearby ophiolite slice at Vallahamars, south-east of Tresta [HU 615 903], is also formed of locally steatitised antigorite serpentinite. It too contains bodies of clinopyroxenite-wehrlite and metagabbro, and includes a 100 m-wide mass of pegmatitic gabbro. The northern boundary is a well-defined thrust dipping south at 45°. The upper surface of the lens is covered by a lens of the Gruting Greenschist, but is well exposed in the cliff face at Billaclett, south-east of Tresta [HU 617 902]. Here, it forms a smooth, almost polished, near-planar surface dipping south at 30º, with scattered patches of slickensides trending 030–045º. Below this, a set of anastomosing shears is exposed which separates thin imbricated slices of serpentinite. The shears have been impregnated with carbonate weathered to reddish brown; thin sections reveal that the carbonate has cemented finely shattered serpentinite. The slice of Gruting Greenschist overlying the serpentinite is folded isoclinally, with axial planes parallel to the surface of the slice.

Poorly exposed lenses of commonly steatitised serpentinite, some no more than a few metres across, crop out immediately west of the Funzie Conglomerate at The Heog [HU 657 905], north of Everland [HU 663 918], in Ness of Gruting [HU 653 918], at Aith Ness [HU 630 900], north-west of Stackaberg [HU 606 936], and near Tresta, west of the ophiolite slice at Vallahamars ( [HU 613 905], [HU 6135 9027] and [HU 6151 9028]). Several lenses of metagabbro crop out immediately beneath the Upper Nappe in the Stackaberg area [HU 614 928]; one of these, at Staves Geo [HU 654 887], contains an intrusive mass of quartz-albite-phyric felsite.

Norwick Hornblendic Schist

Read (1934a) used the name Norwick Hornblendic Schist for a number of tectonic lenses of hornblende schist in the thrust zone beneath the Lower Nappe in Unst (Figure 31). The name was used later in Fetlar for units of similar rock, which form lenses in the thrust beneath the Upper Nappe (Flinn, 1958). Along boundaries with other units, the rock is locally sheared and metamorphosed to phyllonitic greenschist. This is displayed particularly well at Norwick [HP 652 147] in Unst, and at Aith Ness [HU 633 895] and Swart Houll [HU 646 917] in Fetlar. Williams and Smyth (1973) interpreted the Norwick Hornblendic Schist in Unst, and similar developments of hornblende schist in other ophiolite successions, to be a dynamothermal aureole created during obduction of hot mantle over sea floor mafic rock. Prichard et al. (1981) and Spray (1988) described the Norwick Hornblendic Schist in some detail, and interpreted its origin in the same way. However, recent detailed survey work (Flinn et al., 1991; Flinn, 1993) has resulted in the recognition of three subdivisions of the Norwick Hornblendic Schist, and has revealed intrusive relationships between the rocks and overlying serpentinite of the nappes.

Norwick Hornblendic Schist type 1

Type 1 crops out at Read’s (1934a) type location for ‘Norwick Hornblendic Schist’ in the LIZ of northern Unst, near The Taing [HP 652 147]. It consists of poorly exposed, deeply weathered, plagioclase-phyric metabasalt. Two small tectonic lenses of the same lithology crop out farther south in the LIZ, at Hill of Voesgarth [HP 608 082] and Queyhouse [HP 612 124].

The margins of the largest lens, at Norwick [HP 642 145], have been sheared and metamorphosed to create strongly schistose and well crystallised chlorite-epidote- plagioclase phyllite. This retrograde facies is developed particularly well where the lens crosses the coast at The Cliffs [HP 652 148]. Read (1934a) correlated the retrograde rocks with retrograde greenschist rocks (derived from Norwick Hornblendic Schist type 2) from elsewhere in the LIZ of Unst. Such rock is even better displayed in Fetlar, at Aith Ness [HU 632 895] and extensively to the east of Virva [HU 647 917]. Read (1934a) assumed that the protolith of the retrograde rock at Norwick was hornblende schist.

Norwick Hornblendic Schist type 2

Type 2 is the most widely distributed of the Norwick Hornblendic Schist variants. The largest lens, 7 km long and up to 400 m wide, is in the LIZ in the south of Unst, for example north of Loch of Stourhoull [HP 580 030], while several smaller lenses are distributed widely in the MIZ on Fetlar ((P534690), (P534769)). All occurrences lie just below the inferred obduction thrust, though they are generally separated from it by zones of cataclasis and, in places, by tectonic lenses of other post-obduction units. In hand specimen, the rock is black, feldspar is barely visible, and a weak lineation is discernible locally. A schistosity is well developed, but the rock does not break easily along it. In units of the MIZ, the schistosity occurs at various angles to the obduction thrust due to tectonic rotation.

Hornblende is bluish green and exhibits a strong preferred orientation of both the crystal lattice and crystal shape within the S-dominant tectonite, which has little evidence of a lineation. The schistosity is emphasised locally by close-spaced, feldspar-rich laminae. Variable proportions of epidote and chlorite are associated with retrograde alteration. Titanite is a common and often profuse accessory phase. Crystals of hornblende are generally fine grained (0.1–0.2 mm) but in some rocks can be coarse or very coarse grained (10–20 mm); the latter tend to be greenish brown or brown. Widely but sparsely disseminated crystals of garnet occur at Belmont [HP 560 005]. The schist passes into lower-grade rock always as a result of superimposed retrograde metamorphism, but nowhere is it seen to pass into higher grade rocks.

Norwick Hornblendic Schist type 3

Type 3 forms intrusive bodies, no more than 10 m across, in the base of the serpentinite Upper Nappe close to the obduction thrust [HU 640 943], and within units of Norwick Hornblendic Schist type 2. The rock is black and locally difficult to distinguish in the field from type 2. The bodies have a high-grade, granofelsic core of coarse-grained brown hornblende, diopside, garnet and feldspar, enclosed in a ‘chilled zone’ of low-grade minerals. The brown (pargasitic) hornblende has altered to greenish-brown at crystal edges and displays a weak preferred orientation of lattices. Hornblende crystal boundaries form a triple-junction network indicative of strong recrystallisation under isostatic conditions. The boundaries of clinopyroxene crystals cut hornblende crystals in convex outward- facing arcs, showing that pyroxene grew later than hornblende and replaced it under static conditions. Relict fragments of hornblende remain within some crystals of clinopyroxene. Crystals of garnet (c. 2 mm) occur within, and also contain inclusions of crystals of hornblende and clinopyroxene. Hornblende, clinopyroxene and garnet contain numerous inclusions of fine-grained (c. 0.1 mm) titanite. Ilmenite, apatite, rutile and small amounts of quartz occur with minor amounts of altered (albitic) feldspar.

All samples show at least some sign of retrograde alteration, including veins of chlorite, carbonate minerals and prehnite. Albite and epidote have replaced plagioclase, actinolite has replaced clinopyroxene in places, and brown hornblende is variably strained and altered to green hornblende along its boundaries. Crystals of garnet have been shattered intensely and cut by very thin veins of chlorite.

These high-grade cores grade into chilled margins, 1 m or more wide, developed against adjacent units. They pass outwards from an increasingly chaotic mixture of fine-grained amphibole, chlorite, epidote, albite, ilmenite and titanite to a zone of close-packed anthophyllite needles in a matrix of chlorite. This passes into an outer zone of fine-grained (<0.1 mm) aggregates of chlorite with scattered opaque crystals and very rare crystals of clinozoisite 0.2–0.3 mm across.

The intrusive nature of units of Norwick Hornblendic Schist type 3 is shown most clearly in the base of the serpentinite cliff opposite The Clett [HU 640 943], on the north coast of Fetlar (Figure 31) (P534596). Just above the basal thrust of the Upper Nappe, there are two parallel sheets, about 1 m thick and several metres apart, dipping about 20º due south. Both have the chilled margins that are characteristic of type 3 bodies. A more accessible example occurs on the hillside south of Virva, at [HU 6446 9157] and [HU 6442 9185], where a 1–2 m-thick sheet with chilled margins cuts and is welded to the serpentinite base of the Upper Nappe. Intrusive lenses, several metres thick with high-grade cores, occur at Honga Ness [HU 653 913], Houbie [HU 620 905], and Velzie [HU 603 910]. A similar small mass crops out within the lens of Norwick Hornblendic Schist type 2 at Belmont [HP 564 006], in Unst.

Spray (1988) and others have described in detail an outcrop of hornblende schist on the coast at Virva [HU 6445 9208] as grading from high-grade granofels to low-grade greenschist with increasing distance from serpentinite of the Vord Hill Klippe (Upper Nappe). This survey has shown that high-grade Norwick Hornblendic Schist type 3 extends along the coast for some 40 m from the thrust. The chilled margin against serpentinite to the west has been partly destroyed by brecciation, and the contact against Norwick Hornblendic Schist type 2 to the east has been partly obscured by a shear. Type 2 schist immediately east of the outcrop of type 3 rock (granofels) contains several widely scattered, lenticular streaks of clinopyroxenite and albite pegmatite, while some tens of metres farther east it contains two thin lenses of garnet-biotite psammitic gneiss (Strand Gneiss). Several hundred metres east of the boundary with the Upper Nappe, the hornblende schist becomes almost unrecognisable due to crushing and late retrograde metamorphism to greenschist facies.

Geochemistry

Twenty-two analyses of samples of Norwick Hornblendic Schist type 2 and type 3 are available (Spray, 1988; Flinn et al., 1991). The two types are indistinguishable in terms of composition; they classify as basalt and picrobasalt on the TAS diagram (Figure 32)d. The data points cluster in one relatively compact field on plots designed to discriminate between tectonic settings (Figure 32)a and b, and probably have a MORB-type protolith. Most are olivine normative and Mg# is in the range 44–41.

Both types 2 and 3 are generally enriched very weakly relative to N-MORB, with minor selective enhancement of the LIL elements (Figure 32)c. The profiles resemble those presented by Pearce (1983, fig. 3) as representing evolved tholeiitic MORB derived by crystal fractionation from a mantle giving rise to N-MORB, or by partial melting of the same. Norwick Hornblendic Schist type 2 and type 3 may be separate batches of tholeiitic MORB.

A crystal of brown hornblende from an outcrop of Norwick Hornblendic Schist type 3 at Virva [HU 6445 9208] yielded a 40Ar-39Ar step heating age of 498 ± 2 Ma (Flinn et al., 1991). Hornblende from the same and similar exposures yielded four K-Ar ages ranging from 479 ± 6 to 465 ± 6 Ma (Spray, 1988). The difference in ages is considered to be due to Ar loss resulting from cataclasis and retrograde alteration common in these rocks.

Interpretation

Flinn (1993) suggested that Norwick Hornblendic Schist type 3 is mid-ocean-ridge-type basaltic rock formed by partial melting of mantle due to adiabatic decompression during obduction-driven decoupling along the obduction thrust. The melt intruded up the obduction thrust into already hydrated (serpentinised) metaharzburgite at a maximum temperature of 500°C (based on the stability range of antigorite: see above). Water released from the adjacent serpentinised nappe and from Norwick Hornblendic Schist type 2 may have caused the chlorite-rich ‘chilled margins’ around the type 3 bodies, and associated hydrous amphibole and epidote, to form. The ⁴⁰Ar-³⁹Ar age of 498 ± 2 is considered to be approximately the age of intrusion and therefore the age of obduction. Units of the Norwick Hornblendic Schist type 2 protolith were emplaced in the same way at an earlier stage, when the nappe was being obducted, during which they developed a tectonite fabric. It should be noted that hornblende schists of aureole and Norwick-type do not occur in Norway. Instead, the age of ‘plagiogranite’ is used to date ophiolites. The Shetland Ophiolite has been dated in this way at 492 ± 3 Ma by Spray and Dunning (1991).

Structure of the Lower Imbricate Zone

The LIZ can be followed through the central valley of Unst from the south coast at Belmont [HP 563 006] to the north-east coast at Nor Wick [HP 650 145]. About 1 km south of Belmont, the LIZ is truncated by the Lamb Hoga Fault, while in the far north of Unst it is truncated by the Skaw intrusion. From Belmont to Ungirsta [HP 623 130] it is bounded to the east by a prominent escarpment marking the western edge of the Lower Nappe. From Ungirsta to the sea at Nor Wick it is bounded to the east by phyllite of the MIZ; this boundary lacks a topographical marker.

The LIZ is exposed so poorly inland that the map has been drawn as anastomosing lines separating distinguishable and unrelated lithologies into lenticular areas. This produced some 30 imbricate lenses consisting of 10 different lithologies. Good exposure in the coastal section at Belmont reveals slices of rock within the LIZ that are too small to be delineated on the published 1:50 000-scale map (BGS, 2002), so the true number of tectonic lenses in the LIZ at outcrop must be larger.

The different units are arranged with some regularity. Lenses of Norwick Hornblendic Schist from beneath the Lower Nappe generally crop out closest to it. Slices of the Unst Phyllite Group crop out next to the west. Lenses of the basement (Saxa Vord Pelite, Queyhouse Flags, Valla Field Schist, Burra Firth Formation and Loch of Cliff Limestone) crop out still farther to the west.

Poor exposure in the central valley of Unst means there is little structural evidence from which to make a kinematic interpretation of the LIZ (Figure 33). At Belmont, the boundary between serpentinite and the LIZ was until recently exposed in the sea cliffs, dipping 70º to the west (P534277). The exposure is now buried beneath the ferry terminal. The boundary at that point is not the obduction thrust, but a fault through the Gallow Hill Klippe overlying the obduction thrust. All of the lenses of Norwick Hornblendic Schist in the LIZ are in faulted contact with the nappe along at least part of their length. All have a schistosity that is parallel to the basal obduction thrust of the nappe. In the bed of the stream entering Loch of Watlee at [HP 593 048], there is continuous exposure through the boundary of the nappe with a body of Norwick Hornblendic Schist. The schistosity in the hornblende schist is vertical and apparently parallel to the boundary with serpentinite, but both lithologies have suffered cataclasis (P534071). The few exposed boundaries between imbricated lenses within the LIZ are schistose (blastomylonite?) (e.g. at [HP 597 055] to the east of Loch of Watlee, and [HP 5572 0105] west of Loch of Belmont), which generally dip steeply to the east. Lineations within slices of metasedimentary rock in the LIZ are a result of crenulation folding; they plunge 10–20º north in the north of Unst and 10–30º south in the Belmont area. No structures indicative of the direction of transport of the overlying nappe were observed in exposures of the LIZ during the present survey.

The base of the Lower Nappe, where it is in contact with the LIZ, consists of massive, strongly jointed antigorite serpentinite with patchy, often partial, alteration to carbonate-impregnated serpentinite and steatite. A schistosity, with a variety of orientations, is developed patchily. The general impression given by the LIZ and its boundaries with the nappe to the east is of steep-to-east dipping shears, matching the geophysical interpretation (Chapter 10).

Structure of the Middle Imbricate Zone

The MIZ lies between the Upper and Lower nappes. In Unst, it lies on the upper surface of the Lower Nappe in the Norwick and Muness areas. This surface is, or is dominantly, an erosion surface, though it is questionable whether it was anywhere a deposition surface for rocks of the MIZ that now lie on it. In the Norwick area of northern Unst, the MIZ has been overthrust from the east by the Hill of Clibberswick Klippe of the Upper Nappe. The LIZ truncates it to the west, and it overlies the Lower Nappe to the south. Units of Norwick Graphitic Schist, Gruting Greenschist and Norwick Phyllite, which make up the MIZ here, are known chiefly from temporary exposures in ditches and pits.

Between Uyeasound and Muness in south-east Unst (Figure 6), the Lower Nappe is overlain by the Muness Phyllite, which is exposed continuously in the coast sections north and south of Muness. To the east, the Muness Phyllite is faulted against a block of the Lower Nappe and a tectonic slice of clinopyroxenite-wehrlite (in the MIZ), which overlies it. In the west, the Muness Phyllite rests on metagabbro of the Lower Nappe at Clivocast [HP 602 005], where both units have been sheared to form L-tectonite phyllite. The original boundary could be a thrust or an unconformity sheared by differential deformation. Farther south, in Uyea, the boundary between the base of the Muness Phyllite and metagabbro of the Lower Nappe consists of deformed conglomerate, as at Scarfa Taing [HU 605 997], with clasts elongated in the usual north-north-east direction. The same association occurs on the south tip of Urie Lingey [HU 655 953]. At Winna Ness [HU 6067 9817] on the south coast of Uyea, the boundary between the Muness Phyllite and the Lower Nappe is steep and associated with steep shears, but metagabbro and phyllitic rock at the contact are only weakly sheared.

At Bakka Skeo [HU 627 996], on the south-east tip of Ness of Ramnageo, a strongly sheared ‘boss’ of rock was identified by Read (1934a) as metagabbro (of the Lower Nappe) underlying the Muness Phyllite. The boundary with the phyllite, exposed at Skuda Geo [HU 6264 9952] and Bakka Skeo [HU 6275 9954], displays structural continuity and little lithological contrast. Narrow shear zones within the boss cut it into spindle- shaped phacoids of metagabbro-like rock, several metres across and elongated parallel to the local north-north- east lineations. Schistosity and lamination in the phyllite decrease gradually into the boss as the rock increasingly presents the pale green and white, spotted appearance typical of metagabbro in the Lower Nappe. Thin sections of rock forming the boss reveal it to consist of crushed but fresh, unclouded albite set in a matrix of chlorite and carbonate minerals.

In Fetlar, the MIZ crops out below the Upper Nappe in the centre of the island, and on the Lower Nappe to the west of this and at the eastern extremity of the island (Figure 34). The MIZ, represented by the Muness Phyllite, can be traced south from the southern end of Unst, through Uyea, Wedder Holm [HU 615 975] and Urie Lingey [HU 597 957], and into Fetlar. Much of Fetlar west of the Vord Hill Klippe is underlain by poorly exposed units in the MIZ that are partly overlain by klippen of the Upper Nappe and locally interrupted by inliers of the Lower Nappe. The latter are not always distinguished easily from klippen of the Upper Nappe. An interesting exposure occurs on Stackaberg [HU 613 931], where erosion has cut through the Vord Hill Klippe and tectonic slices of metagabbro immediately beneath it to form a ‘window’ into the underlying Muness Phyllite (Phemister in Wilson, 1931, fig. 2).

The complex interdigitation of lithologies forming the MIZ in Fetlar is exposed fully in coastal sections east and west of the Vord Hill Klippe, but is difficult to interpret in detail (Figure 34). Boundaries between the tectonic lenses have been deformed into zones of schistose or cataclastic rock. Similar shear zones occur within the tectonic lenses. Metamorphism that accompanied the deformation has rendered different lithologies essentially indistinguishable, especially in the field. Metagabbro and lithologies forming the units of the Norwick Hornblendic Schist and Gruting Greenschist have all become chlorite-epidote phyllite of similar appearance. Exposures of the Muness Phyllite and Gruting Greenschist are indistinguishable in terms of structure and colour, though they can be distinguished in thin section. These effects are displayed especially well in the coastal sections between Houbie [HU 622 905] and Aith Ness [HU 633 894], and east of Honga Ness [HU 657 913].

It is possible to discern some order in the arrangement of tectonic lenses in the MIZ. Slices of Norwick Hornblendic Schist, metagabbro or serpentinite immediately underlie the Upper Nappe and overlie either the Lower Nappe or members of the Unst Phyllite Group, especially the Muness Phyllite and Gruting Greenschist.

The MIZ is formed of internally coherent lenses of rock that are often strongly and intensely folded on north-north-east to north-east axes. They never exhibit mélange-type structures. An extension lineation developed parallel to fold axes and the pebble elongation is persistent throughout the MIZ. The lineation is due to the girdle and partial girdle preferred orientation of phyllosilicate minerals, and to crenulation folding. The folding, as seen in the plane normal to the lineation/fold axes, varies from chaotic crumpling on all scales to areas of only crenulations with strain slip cleavage, in some places of several generations and locally overturned to the east or the west, or upright.

The imbricate structure of the MIZ is interpreted to be a result of emplacement of the Upper Nappe over the Lower Nappe. Constriction between the two nappes resulted in folding and extension parallel to the fold axes.

Chapter 7 Late igneous rocks

Skaw Granite

The Skaw Granite consists of deformed microcline-phyric monzogranite that was emplaced in, and has thermally metamorphosed, rock of the Saxa Vord Block and the Lower Imbricate Zone of the ophiolite-complex.

The outcrop

The intrusion is exposed almost continuously along nearly 10 km of cliff in the Skaw peninsula of north- east Unst (Figure 35); by contrast, exposure inland is very poor. The west margin of the intrusion is a simple, intrusive contact with metasedimentary rock, which it has thermally metamorphosed. In the north, the contact between the intrusion and metamorphic rock of the Saxa Vord Block is exposed in cliffs near Virdik [HP 653 171], and in the bed of the Burn of Skaw [HP 655 163]. To the south, the contact with metamorphic rock of the LIZ is exposed at The Cliffs [HP 655 152] (P553412). The exposed rock is part of a larger body that extends offshore to the north and east, but geophysical maps do not indicate its full extent. A borehole drilled (BGS, 1984) about 4 km east of the easternmost land on the Skaw peninsula (Lamba Ness) terminated in granitic rock not dissimilar to the Skaw monzogranite, but apparently lacking microcline phenocrysts.

Several facies of granite can be distinguished in the field. A facies characterised by subequant phenocrysts of microcline crops out in a zone less than 0.5 km wide adjacent to the western margin of the intrusion (Plate 6a). This grades eastwards, towards the interior of the intrusion, into a facies characterised by tabular phenocrysts of microcline (Plate 6b) and (Plate 6c). The interior facies can be divided in the field on the basis of biotite content into relatively ‘light’ and ‘dark’ variants (P533271). However, this distinction is not obvious in thin section or bulk composition data. The intrusion hosts cognate xenoliths and xenoliths of metasedimentary and metavolcanic rock, all of which are distributed in a patchy manner. All the exposed granite has been deformed, to varying degrees, into L-tectonite, as revealed by phenocryst orientations and a non-penetrative foliation. Exposures of fresh granite are very rare, and in most places along cliffs of the Skaw peninsula the rock shows weak granular disaggregation. Spheroidal weathering is well developed in several places (P533282).

Mineralogy

Microcline phenocrysts form about 20 per cent of the rock. In a zone about 400 m wide adjacent to the exposed contact, the phenocrysts range from broadly subequant shapes up to 2 cm across to squat tabular shapes up to 5 cm long (Plate 6c). Farther east, though subequant shapes occur locally, the phenocrysts become larger and markedly tabular over a short distance, with maximum dimensions up to 10 cm and axial ratios of 3:1 (and rarely 4:1). Significant variation in phenocryst shape occurs over distances as short as a few metres. The phenocrysts are fresh but fractured, with common patches of perthite. They generally present well- developed crystal shapes, are usually Carlsbad twinned, show a variable and patchy development of cross-hatched (microcline) twinning, and enclose small (<1 mm) crystals of plagioclase and less commonly quartz. Phenocrysts of microcline similar to those described above also occur in cognate xenoliths of dioritic rock (Plate 6d). On Mainland Shetland, the Aith–Spiggie appinitic granodiorite [HU 350 510] also has a contact zone containing subequant microcline phenocrysts that become tabular farther from the contact.

Irregularly shaped crystals of sodic plagioclase are up to 5 mm across and form about 30 per cent of the rock. They are partly altered to a variably dense network of sericite, which in some thin sections is accompanied by fine-grained crystals of epidote. The alteration is rarely so dense and complete as to destroy all the lamellar twinning.

Dullish brown flakes of biotite are commonly up to 1–2 mm across, but are much smaller where they have been recrystallised.

Crystals of quartz, up to 2 mm across, have been strongly strained; they either show strain extinction (especially near the contact) or have suffered recovery recrystallisation, and now consist of aggregates of crystals or sub-crystals, up to 0.2 mm across.

Accessory crystals of metamict allanite, up to 0.3 mm long, occur with titanite and zircon. Secondary white mica, epidote, calcite, quartz and, rarely, chlorite, occur.

Geochemistry

Gamil (1991) presented a detailed account of the geochemistry of the granitic intrusions of Shetland, including the Skaw Granite. Fourteen whole-rock geochemical analyses of the Skaw Granite are available (Table 15). The intrusion classifies as monzogranite (grading locally into granodiorite) in terms of its modal composition (Figure 36). However, SiO₂ is in the range 64.10 to 70.78 wt per cent so samples straddle the dacite and trachydacite fields on a plot of total alkalis versus silica (‘TAS’). The intrusion is moderately peraluminous (Figure 37) and subalkaline (Na₂O + K₂O from 6.14 to 7.94 wt per cent), with high-K calc-alkaline to weakly shoshonitic affinity (Figure 38). It is an I-type granite.

Xenoliths

Cognate xenoliths

Cognate xenoliths of brownish microdioritic-rock are scattered unevenly through the Skaw intrusion, with a tendency to occur in swarms. The highest concentration occurs between Gorsun Geo [HP 658 169] and Forn Geo [HP 663 170], where there is in places a roughly equal proportion of xenoliths and granite. Cognate xenoliths are common in swarms and as isolated features between Forn Geo and Wick of Skaw [HP 660 166], and on Holm of Skaw [HP 668 171]. Swarms occur immediately south of Wick of Skaw, on the north shore of The Garths [HP 665 158], and on Lamba Ness [HP 675 158]. They are, by contrast, extremely rare on the north coast, between the contact [HP 652 170] and Gorsun Geo [HP 658 169] (P533273), and along the south coast from Lamba Ness [HP 675 155] to the contact at The Cliffs [HP 655 151]. The xenoliths are mostly of rounded appearance, rarely more than twice as long as they are broad, and generally from several centimetres to more than 1 m across. A giant cognate xenolith, about 7 m across, and a sheet- shaped mass of undetermined length that is nearly 1 m thick, occur in the area of closely-packed xenoliths around Gorsun Geo (P550138). A swarm of very thin xenoliths, each up to 1 m long, crops out on the north tip of Virdik [HP 655 175]. South of Wick of Skaw most xenoliths are aligned, with their long axes parallel to the tectonic lineation in the intrusion. North of Wick of Skaw some xenoliths are similarly aligned, but most are not.

Thin sections of 12 cognate xenoliths from different parts of the intrusion vary little in appearance (Plate 6d). They consist of biotite, quartz, plagioclase and epidote, a little chlorite, and rare microcline. These minerals form a matrix of uneven grain size (no more than 0.5 mm) within which crystal boundaries have a ragged appearance. Larger quartz crystals in many xenoliths have suffered recovery recrystallisation, yielding aggregates of parallel, elongate sub-crystals that give the rock a mylonitic appearance. Such crystals form part of a crude foliation in the xenolith matrix that, like that in the granite, flows around phenocrysts. Most thin sections exhibit these signs of deformation, though evidence of deformation is not usually visible in hand specimen.

Many xenoliths contain widely scattered phenocrysts of microcline, similar in appearance to those in the enclosing granite though usually somewhat smaller and less well formed. Their edges tend to overgrow adjacent minerals, like those in the granite.

Metasedimentary rock

Xenoliths of metasedimentary rock, in the form of massive quartzite, laminated impure quartzite, very fine-grained phyllite, and mica schist, are widely, but unevenly, scattered throughout the Skaw intrusion.

All four lithologies are interbanded in a giant xenolith (effectively a screen) forming the cliff at the back of the beach at Wick of Skaw [HP 660 166] (which is often hidden by blown sand). The xenolith extends northwards, for about 1700 m, to form the east side of Forn Geo [HP 663 169] and the west half of Inner Flaess [HP 667 173]. It was once a single body, but has been dismembered by small veins, up to several centimetres wide, of non-porphyritic and porphyritic granite.

A mass of laminated quartzite and fine-grained phyllite, several tens of metres across in its largest dimension, occurs between Gorsun Geo and Forn Geo [HP 663 170]. A mass of mica schist interbanded with both types of quartzite occurs at Bluejibs [HP 664 167]. Most other xenoliths of metasedimentary rock consist of quartzite, or quartzite interbanded with laminated quartzite and/or phyllite. They occur chiefly in the general area of the screen of metasedimentary rock at Forn Geo. Most are 1 m or so across, but they range down to several centimetres.

All these xenoliths have the appearance of low grade, regionally metamorphosed metasedimentary rock rather than hornfelsed rock. The Queyhouse Flags is the closest lithological match of in-situ rocks occurring locally. Many of the xenoliths are prominently lineated as a result of folding, which created mullions and crenulations. Lineations and fold axes plunge between 0º and 30º towards the north-east (between 020º and 065º), parallel to the lineation in the granite (Figure 39). Distortion of the xenoliths is displayed particularly well on Inner Flaess [HP 667 173]. The L-tectonite character of the Forn Geo screen appears to be a result of pre-intrusion folding. However, the exposure is in steep cliffs and the wavelength of the folds is equal to, or greater than, the cliff height so details of the structure are difficult to establish.

Xenoliths of quartzite consist of quartz and some epidote. The character of the quartz is variable, ranging from fine-grained and granular (due to recovery crystallisation) in some samples, to coarse and barely strained in others. Other xenoliths of metasedimentary rock collected from interior parts of the intrusion include coarse muscovite-biotite schist and finer grained muscovite-biotite phyllite; these occur with and without a schistosity, and they are commonly interbanded with layers of quartzose or semipelite composition. As in the quartzite, crystals of quartz in these rocks range from recovery recrystallised to virtually unstrained. Several examples of more pelitic rock contain moderately well formed shimmer aggregates, none of which contain relics of their protolith. Other than these, the xenoliths of metasedimentary rock contain no evidence of thermal metamorphism, in contrast to chloritoid-type pelite of the Saxa Vord Block (Chapter 5) where it crops out in the contact zone of the Skaw intrusion.

Some thin sections of granite from within a few metres of the contact contain small patches rich in shimmerised fibrolite, and sometimes accompanied by large plates of muscovite and coarse chlorite. These small patches are very similar petrographically to the adjacent contact- metamorphosed Saxa Vord Pelite, except for the absence of chloritoid, and are probably small xenoliths of it that have been partly assimilated.

Metavolcanic rock

Xenoliths of basic metavolcanic rock (greenschist) occur within 100 m of the intrusion margin. They are best seen in the cliff face between Virdik [HP 654 176] and the contact on the north coast of Unst [HP 652 171]. In cliff-top exposures of the contact at Rurhella [HP 651 170], the rock is badly weathered and exposure is patchy and difficult to access, but in vertical cliff exposures where the contact reaches the sea, several xenoliths of metavolcanic rock form thin, very elongate sheets aligned roughly parallel to the contact (P533250). Farther north, towards Virdik, sheet-like xenoliths exposed in the cliff are accessible, and the rock is fresh. They include a swarm of close-packed, rather irregularly shaped but broadly equidimensional, blocks north of The Haa [HP 658 168]. The beach at the back of the geo at Rurhella is formed of well-rounded, polished boulders up to 2 m in diameter, which include deformed agglomeratic metavolcanic rock that can be matched to xenoliths in nearby cliffs to the east. Boulders of deformed pillow lava at Rurhella ( [HP 652 170] (P533163) appear to have a similar metamorphic history, but cannot be matched to xenoliths in the intrusion.

There is a small exposure of metavolcanic rock in the Burn of Skaw [HP 652 161], several tens of metres east of the granite contact. Similar xenoliths of metavolcanic rock crop out at The Cliffs [HP 655 152], close to the southern contact of the intrusion, where they range up to 7 or 8 m across and in places lie only several centimetres above the contact. They include agglomeratic types and are intruded by veins of microcline-phyric granite. The granite and xenoliths are very strongly weathered.

Thin sections of metavolcanic xenoliths show them to be free of late strain effects, unlike the other types of xenolith and the granite. Most contain crystals of blue- green hornblende with a weak to moderate preferred orientation. The hornblende is commonly associated with biotite, which shows a good preferred orientation. Epidote, quartz, titanite and plagioclase occur in very variable proportions in different samples. The agglomeratic rock has a matrix of hornblende crystals with brownish cores and blue-green rims. ‘Pebbles’ within the agglomerate are composed of equigranular feldspar crystals up to 0.4 mm in diameter, with hornblende, biotite and epidote (P533277). Samples of pillow lava consist of crystals of hornblende, biotite, epidote, quartz, feldspar and an opaque phase, all of which are distributed unevenly and of highly variable grain size.

The basic metavolcaniclastic xenoliths with pillow lava occurring in the edge of the granite in close proximity to the Saxa Vord Block is very suggestive of the granite having been in contact with the Dunrossness Spilite Formation exposed in the south of Shetland at the eastern limit of the Clift Hills Group, Shetland Dalradian. The Saxa Vord Block has been correlated with the Clift Hills Group. The Dunrossness Spilite Formation is partly composed of basic volcaniclastic rocks with pillow lavas.

Granite fabric

Throughout the exposed outcrop of the Skaw intrusion, the granite displays an L- to L–S- tectonite fabric, though in places the structure is developed so weakly as to be difficult to detect. The fabric is defined most clearly by a preferred orientation of tabular phenocrysts of microcline (Plate 6c), in which the long axes define a lineation while the intermediate axes define a partial girdle, or girdle, normal to the lineation (Figure 39). Thus, the plane of the foliation is poorly defined. Much of the scattering of the schistosity measurements probably derives from inaccurate measurements on faces normal to the lineation where the degree of preferred orientation is at best moderate.

Samples collected from within a few hundred metres of the contact show only strain extinction in quartz crystals. Farther from the contact, cataclasis accompanied by microgranulation strings develops, and rocks with a mylonite-like character are locally developed. Deformation increases locally to the extent that feldspar porphyroclasts are enclosed in a fine-grained intensely foliated matrix. Such zones appear mylonitic and are developed especially well at contacts between the light and dark facies of granite (e.g. at [HP 673 156] near Lamba Ness).

The granite matrix is foliated rather weakly. A matrix foliation parallel to the lineation is apparent on faces parallel to the phenocryst lineation (Plate 6c). The matrix foliation is barely apparent on faces normal to the lineation (Plate 6b), and is not apparent on faces normal to the lineation in L-tectonite. Matrix foliation is often detectable on faces normal to the lineation in L–S-tectonite, but its orientation is variable and ranges from the same as that of the plunge of the lineation to vertical. The matrix foliation in the granite is best studied in large rounded boulders at the west end of the beach at Norwick [HP 652 148].

A plot of the poles to planar elements, determined from observations of phenocrysts and matrix, reveals a poor girdle (Figure 39). The wide scatter of foliation poles in the girdle owes something to the difficulty of making accurate measurements, but is due also to a non-rectilinear structure of the granite. This is displayed clearly in places, for example on Holm of Skaw.

The least strained rocks are protocataclasites, showing disruption of the boundaries of larger crystals, especially microcline phenocrysts. Crystal boundary disruptions range from simple fractures along the boundaries (often enhanced by iron staining) to narrow shear zones marked by microcrystalline aggregates of quartz and, or, feldspar. Tiny flakes of biotite show some preferred orientation parallel to the shear zones. In such rocks, the larger quartz crystals have become microcrystalline (<0.2 mm) aggregates due to recovery recrystallisation, or they show sutured sub-crystal boundaries due to strain-induced boundary migration.

A higher degree of strain has produced rock in which the crystal boundary shears are linked between crystals, creating a set of subparallel, anastomosing cracks, or more commonly microshears, running through the rock. They enclose lenticular domains containing recovery- recrystallised quartz, bent flakes of biotite, and kinked polysynthetic-twinned plagioclase.

At still higher degrees of strain, the shear zones become more prominent, linking large flakes of biotite and causing them to break down into very small flakes orientated parallel to the shears. Microcrystalline quartz becomes elongated parallel to the mesoscopic shear zones, forming quartz-rich pinch-and-swell layers that alternate with biotite-rich shears. By this stage kinking of twinned plagioclase crystals is common and microcline crystals show a strong tendency for their longest axes to be aligned with the foliation defined by the shears. With stronger deformation, the microcline phenocrysts (especially those orientated parallel to the shears) show increasing numbers of tension cracks. These tend to be normal to the lineation and are filled by aggregates of microcrystalline quartz and, or, feldspar, and, or, calcite, or larger unstrained crystals of quartz, suggesting the rock has elongated parallel to the lineation. Such cracks have not formed in plagioclase crystals. Instead, plagioclase crystals, and much less commonly the microcline phenocrysts, show signs of breaking into disoriented fragments in the most strongly deformed samples.

The most highly strained samples were collected from penetratively sheared zones no more than several metres wide, e.g. at Bluejibs [HP 664 167] and near Lamba Ness [HP 673 156]. In such rocks, ovoid feldspar porphyroclasts (plagioclase generally smaller than microcline) are set in a very fine-grained matrix of quartz, feldspar, biotite and white mica. The feldspar clasts lie with their long axes in these shears.

Recovery recrystallisation of quartz has produced aggregates of equigranular crystals in nearly all samples, but several samples contain highly elongate, parallel sub-crystals. Thin sections showed these to be aggregates of rod-shaped sub-crystals that are orientated parallel to the lineation.

Quartz veins

Planar veins from 2 to 15 cm wide of pure grayish white quartz occur prominently throughout the intrusion. They cut veins of aplite, and a band of mylonitic rock at Bluejibs [HP 664 167]. They typically occur singly, but in several places (e.g. Holm of Skaw [HP 667 169] and west of Bluejibs [HP 662 166]) they form sets (i.e. a number of closely spaced, parallel features). Plotted on an equal-area net, the poles of 50 such veins and sets of veins scatter widely, forming a very dispersed complete girdle about a north-west–south-east horizontal axis (Figure 39).

Interpretation

The Skaw intrusion was emplaced at a temperature sufficiently high for the adjacent Saxa Vord Pelite to be contact metamorphosed from chloritoid phyllite to sillimanite-bearing hornfels (Chapter 5; Flinn et al., 1996). However, the intrusion temperature was insufficient to impose a significant contact metamorphic effect on xenoliths of metasedimentary rock.

Granite deformation is interpreted to have occurred in a tectonically constricting environment, which also involved the Saxa Vord Pelite and Hevda Phyllite. The intrusion was extruded obliquely upwards to the south- west at about 20º above horizontal. This seems to have occurred during the final ‘freezing’ of the granite, while it was still hot enough to undergo extensive recovery recrystallisation. During deformation, phenocrysts were rotated to bring their long axes into alignment with the matrix lineation, and stretched so that they extended by tension fracturing. The deformation involved orthorhombic pure shear (L to L>S); there is no evidence in the field or in thin section of a monoclinic component, other than that due to deformation gradients. The granite responded to the deformation partly by cataclasis and partly by recrystallisation.

The absence of andalusite in the contact aureole part of the Saxa Vord Pelite (Chapter 5) suggests intrusion at a minimum of 4 kbar (a depth of 15 km or more). This may have led to a protracted period of cooling.

Age of the granite

A range of radiometric ages has been obtained for the Skaw Granite (Appendix I). The most recent are five ⁴⁰Ar-³⁹Ar step-heating ages, two from K-feldspar phenocrysts in the intrusion and three from muscovite plates grown in the Saxa Vord Pelite in the aureole (Flinn and Oglethorpe, 2005). Attempts to date other minerals (hornblende and biotite) hosted in xenoliths have failed. The ages obtained for K-feldspar (407 and 406 Ma), are considerably younger than those obtained from aureole muscovite (427, 425 and 419 Ma). The muscovite ages are considered more closely to reflect the age of intrusion (Flinn and Oglethorpe, 2005), because K-feldspar gives less reliable ages than muscovite and has a substantially lower closure temperature (c. 130ºC and c. 350ºC, respectively): one muscovite sample, lacking evidence of physical strain and yielding a near- perfect plateau, was selected as representing the age of the intrusion at 425.6 ± 2.6 Ma. The granite intrusion cuts the Lower Imbricate Zone and has produced fibrolite in country rock immediately adjacent to the contact. The intrusion is cut by a swarm of undeformed sheets of spessartite (see below); elsewhere in Shetland, spessartite intrusions have been dated at about 400 Ma (Flinn, 1994a). These facts suggest that emplacement and deformation of the Skaw Granite was simultaneous with Upper Nappe emplacement, during the Scandian Event (c. 425 Ma).

Petester Granite

The Petester Granite is an apparently continuous sheet of granite, around 10 m wide, that crops out parallel to, and several hundred metres west of, the west shore of Loch of Cliff [HP 597 120]. It is exposed intermittently over a distance of about 3 km in several streams flowing into the loch. The southernmost exposure is in a quarry at Roselea (597 107), and the northernmost is on the shore of Loch of Cliff [HP 6025 1310]. The sheet is approximately conformable with layering in the Burra Firth Formation. The rock, analyses of which are presented in (Table 16), has a monzogranite composition. Phenocrysts of microcline are 1–2 cm across, many with poorly developed, thin, white, rapakivi-like rims composed of aggregates of fine-grained feldspar and myrmekite. Primary muscovite forms flakes up to 0.5 cm across, and a smaller proportion of smaller flakes of biotite are pleochroic from brown to opaque. Some boulders of a similar rock on the adjacent loch shore show signs of a tectonite fabric.

Primary muscovite and a peraluminous (Fig. 37), corundum normative composition suggest that the Petester sheet is ‘S-type’, i.e. it has a sediment protolith.

Pegmatitic and aplitic rock

Veins and masses of granitic pegmatitic and aplitic rock crop out in many parts of the basement on Unst and Fetlar, though they are generally observed in coastal sections and are rarely seen inland. They occur widely in the Valla Field and Lamb Hoga blocks, but apart from several intrusions at Ness of Queyhouse [HP 606 130], they are absent from the Saxa Vord Block.

Two types of pegmatitic rock are recognised. White pegmatite forms veins varying from several centimetres to 10 m or more wide, and is often tectonised. Red pegmatite is distinguished by large, irregular patches of bright red, coarsely crystalline K-feldspar, is rarely tectonised, and forms veins and masses that commonly cut white pegmatite (P533311). Red pegmatite is more common in Yell (Flinn, 1994a).

Three types of aplitic rock are recognised. White ‘aplite’ forms medium- to fine-grained veins from several centimetres up to 10 m wide. Uniform, moderately coarse, schistose aplogranite forms intrusive masses 10–100 m across. Aplopegmatite, an intimate mixture of aplitic and pegmatitic rock, is only rarely tectonised and occurs much more commonly in Yell (Flinn, 1994a).

A body of aplogranite crops out on the west side of Burra Firth at the Lighthouse Station [HP 613 148], on a stretch of coast in which aplite veins, generally less than 1 m thick, are relatively common. In Fetlar, two bodies of aplogranite crop out east of Brough Lodge, at [HU 588 922] and [HU 583 932]. Veins of aplite are rare in the Valla Field Block, except in the extreme south, near Belmont [HP 565 010]. Veins of aplite less than 1 m wide are relatively common in the south-west corner of Unst (near Belmont, along the coast near Belmont, and especially in the large outcrop of the Westing Limestone at Snabrough [HP 570 025]), and farther south in outcrops of the Westing Limestone on Linga [HU 560 985] and Hascosay [HU 565 925], and at Brough Lodge in Fetlar [HU 588 921]. In central Hascosay, generally mylonitised aplite (P541624) is associated closely with the ultramafic Hascosay Slide Zone (Flinn, 1994a). Occurrences of aplite in the Lamb Hoga Block consist mainly of cliff-sized conformable and cross-cutting schistose veins that are rarely less than 5 m thick. Aplopegmatite is confined to the Westing area of Unst [HP 575 058].

Intrusions of pegmatite, less than 1 m thick, range from common to absent in the cliffs between Tonga [HP 580 150] and Lund [HP 560 040]. North of Tonga, the high, vertical cliffs are riddled with giant lenses of pegmatite up to tens of metres thick (P533200). Red pegmatite is concentrated in the south-east corner of the Lamb Hoga Block, but rare, isolated veins are scattered widely in the Valla Field and Lamb Hoga blocks.

Exposures of cross-cutting relationships between the various types of pegmatite are rare, but they suggest that red pegmatite is younger than aplopegmatite, which is in turn younger than aplite and white pegmatite. However, several centimetre-wide veins or stringers of untectonised pegmatite and aplite cut thicker veins of tectonised pegmatite and aplite. Sheets of lamprophyre cut all types of aplitic and pegmatitic intrusion.

Intrusions of aplite consist of oligoclase and quartz, with or without microcline, muscovite and biotite; garnet, epidote and titanite are rare accessories. Thin sections were prepared from 36 samples of aplite. Most have a uniform grain size of about 1 mm, but coarser examples, including samples of aplogranite, have grain sizes ranging up to 3 mm. Only 12 sections contained microcline, all of them from samples collected south of the Lamb Hoga Fault. Microcline occurs mostly as discrete crystals. However, in one sample it is interstitial and in another it has replaced most of the plagioclase and forms a poikilitic mass enclosing plagioclase relics and quartz. Most thin sections contain muscovite and (usually dark brown) biotite.

Intrusions of pegmatite and aplopegmatite are too coarse-grained for easy thin section examination. The large red patches in the red pegmatite consist of microcline, but the pinkish feldspar in the white pegmatite is plagioclase.

Interpretation

The distribution of pegmatitic and aplitic rock is difficult to determine accurately because exposure is so variable. Available data suggest the intrusions were emplaced in swarms. Emplacement postdated the high-grade constructive metamorphism in the area and much of the late deformation, but predated emplacement of sheets of lamprophyric rock.

Basic to intermediate intrusions

Numerous intrusive sheets of basic to intermediate composition cut all other rocks in the district, including those of the ophiolite-complex (though one sheet south- east of The Nev [HP 560 045] contains a small net-vein complex of aplite). Most are microdiorite, and a few exhibit textures typical of lamprophyric rock. They are texturally distinct from intrusions of the Quasi-sheeted Dyke-swarm, and some have been tectonised. For convenience, all of these intrusions are grouped on the published map, under the general heading ‘lamprophyre’. Exposed intrusions are confined almost exclusively to coastal sections (Figure 40). Only a few were found inland: a dyke in the Loch of Cliff Limestone on the south-east side of Loch of Cliff [HP 602 112]; a rodingitised (Chapter 6) sheet in the Gallow Hill Klippe at Belmont [HP 567 004]; and a body of appinitic diorite in Westing [HP 576 054]. No sheets were recorded from inland exposures, even in areas where the cliffs are riddled with them and inland exposures are profuse. The same is largely true in other parts of Shetland that lie east of the Walls Boundary Fault.

More than 300 of these minor intrusions have been recorded (though some may have been recorded more than once in areas where they are common). They range in thickness from 1 cm to 10 m, and exposed lengths range from several tens of centimetres to at least 300 m (e.g. at Westing [HP 567 065], and Pund Stack [HP 624 036]). Individual dykes in the swarm that cuts the Skaw intrusion (Skaw swarm) may extend for 2 km.

No consistent orientations of the sheets have been identified, although many, including the Skaw swarm, are sufficiently steep to be termed dykes. A swarm of well-formed sills crops out on the east coast of Lamb Hoga [HU 613 893]. Single sheets containing multiple intrusions are rare, but they occur at Tonga [HP 579 148] and on Hascosay [HU 564 928]. Autobrecciated sheets are more common, but are all of the malchite type (see below). In these, a late phase of injected magma has brecciated the original sheet, angular fragments of which are now set in the later intrusion (P533953). The fragments generally consist of rock that is slightly coarser and slightly darker than the enclosing rock. Good examples occur south of Fill Geo [HP 578 083] and south of Punds Geo [HP 568 067].

Thin sections were prepared from around 140 samples of these basic to intermediate intrusions, selected to represent the entire range of variation rather than the relative proportions in which the variants occur. Four main types can be distinguished with the aid of thin sections: lamprophyre (spessartite and kersantite); aphyric hornblende-biotite microdiorite (referred to hereafter as ‘malchite’ (Rock, 1991), though such rocks should be weakly plagioclase-phyric); spherulitic rock; and appinitic microdiorite. A subset of the malchite type contains crystals of green hornblende with opaque black cores. Another subset of the malchite type is autobrecciated; these are recognisable only where the sheets crop out on beaches and have been ‘polished’ by sand. Of the samples for which thin sections were prepared, 21 are spessartite, 2 are kersantite; 58 are malchite (24 of which belong to one of the malchite subsets); 18 are appinitic diorite; 23 are spherulitic rock (10 of which contain carbonate minerals); and 20 are too altered to be classified. Sixteen samples of the basic to intermediate intrusions from the district were selected for whole-rock geochemical analysis; a subset of these is presented in (Table 17). Four analyses of similar rock from Yell are also available (Flinn, 1994a).

Lithological variants

Spessartite and kersantite

Sheets of spessartite crop out along the east coast of Unst in the Shetland Ophiolite-complex, and along the west coast of Unst and Fetlar in the basement to the ophiolite-complex. Undeformed sheets of spessartite cut the Skaw Granite, the ophiolite-complex and the thrust below the Upper Nappe, and elsewhere in Shetland. They have been dated at about 400 Ma (Flinn, 1994a). The sheets have a typical lamprophyric texture, and contain brown hornblende (commonly as needles 0.2–0.3 mm long) set in a matrix of fresh, finer-grained andesine plagioclase. Fresh spessartite grades into altered variants in which hornblende is partly altered, or replaced by biotite and, or, chlorite with epidote. Even in the most altered variants, pseudomorphed mafic minerals preserve the characteristic lamprophyre texture, and plagioclase remains fresh. Kersantite contains large (0.5 mm) flakes of biotite. A variety containing plates of biotite up to 10 cm also occurs (P541497) Altered samples can be difficult to distinguish from altered spessartite.

Malchite

Samples of malchite (hornblende-biotite microdiorite) are similar in some respects to samples of altered lamprophyre, but lack mafic phenocrysts or aggregates of crystals that have pseudomorphed mafic phenocrysts. The matrix consists of fresh crystals of andesine of about the same size as the mafic minerals (0.2–0.3 mm). Crystals typically have irregular, ‘ragged’ boundaries. Sheets of malchite were recorded on the west coast of Unst, in the Nor Wick area [HP 653 150], and along the south coast of the Skaw intrusion [HP 663 153]. Along a section of the west coast of Unst, stretching for several kilometres south of Tonga [HP 580 150], all the recorded occurrences of lamprophyric rock are malchite with a distinctive petrographical characteristic: hornblende occurs as poorly formed, rod-shaped, rich green crystals with relict brown cores containing irregular patches of very fine-grained, black particles. Sheets with this characteristic occur nowhere else in the area. The swarm cutting the Skaw intrusion trends subparallel to the south coast of the Skaw peninsula. Around a dozen sheets of malchite, up to 1 m thick, can be traced between the opposing cliffs of geos and headlands along this stretch of coast. All examples of autobrecciated lamprophyric rock are malchite.

Appinitic microdiorite

Appinitic microdiorite forms relatively coarse (0.5–2.0 mm) rock in which fresh andesine forms larger crystals than hornblende and biotite. Hornblende is brown and usually forms subequant crystals. Bodies of appinitic rock, up to 10 m across, crop out at The Cliffs [HP 654 150], Westing [HP 577 057] and Belmont [HP 565 006]. Alteration has affected only the mafic minerals (as in other types of lamprophyric rock), and has primarily involved replacement of brown hornblende by biotite or green hornblende. The rock is very similar to hornblende-biotite diorite recorded at Gloup Holm on Yell (Flinn, 1994a). The occurrence at The Cliffs, like that at Gloup Holm, is surrounded by a narrow fibrolite- bearing aureole and offshoots of lamprophyric sheets extending into adjacent rock.

Spherulitic rock

Spherulitic rock is typically slightly coarser than malchite and comprises chaotic mixtures of ragged crystals of plagioclase, chlorite, biotite and epidote. The rock is characterised by various types of ‘spots’, the most striking of which are spherulites, up to 5 mm in diameter, formed of radiating and branching, florescent, and subparallel growths of albite microcrystallites. Others are formed of radiating single crystals of albite. The character of the spherulites varies from sheet to sheet; they can be scattered widely, or close-packed with mafic minerals concentrated in the interstices. Some sheets contain crystals of carbonate mineral up to 0.5 mm across. In several sheets at Aith [HU 628 900] in Fetlar, both the matrix and spots are aphanitic; some of the spots show signs of recrystallising to feldspar spherulites. One occurrence of fresh brown-hornblende spessartite cutting the ophiolite north of Geo of Heuken [HP 636 057] contains well-developed albite spherulites.

Within the district, sheets of spherulitic rock are confined to the ophiolite-complex, and are concentrated in the Middle Imbricate Zone. However, eight sheets of spherulitic rock have been recorded in the metamorphic basement on Yell (Flinn, 1994a). The greatest concentration of sheets of spherulitic rock occurs along the north shore of the Muness peninsula [HP 625 022].

The sheets have fine-grained edges and spherulitic cores, and are from 0.01–0.5 m wide. They tend to strike parallel to sedimentary layering or, less commonly, to spaced cleavage in the phyllite. Others cut the extension lineation obliquely. Some sheets have been folded and deformed with the rocks they cut. In some sheets, the spherulites are elongate parallel to the fold axes and the extension lineation in the enclosing phyllite. A sheet of spherulitic rock at Cross Geo [HP 6522 1212] cuts the steatite and mylonite thrust zone at the base of the Hill of Clibberswick Klippe, and has been dismembered. It has also developed a chlorite envelope through reaction with the steatite.

A sheet cutting the Gallow Hill Klippe on the south-west side of Gallow Hill [HP 568 003], and another in the cliffs of Sound Gruney [HU 579 961], have been rodingitised. The thrust zone at the base of the Hill of Clibberswick Klippe, east of Clibberswick [HP 652 122], and the similar base of the Gallow Hill Klippe near Head of Mula [HU 567 996], contain detached sections of sheets. In both cases, the margins of the sheets have been chloritised to a width of several centimetres. A sheet immediately below the Hill of Clibberswick Klippe at The Taing, Norwick [HP 653 146], has been boudinaged on down-dip axes (Read, 1934c). Sheets that cut the Funzie Conglomerate [HU 653 887] are cut by adularia-filled tension fractures that also cut pebbles and matrix in the conglomerate.

Interpretation

The basic and intermediate sheets of the district were probably emplaced at the same time as those found throughout units of metamorphic rock elsewhere in Shetland east of the Walls Boundary Fault. (Flinn, 1994a, and references therein). On the Mainland of Shetland, similar intrusions, dominantly of spessartite and kersantite, were emplaced just before the appinitic Graven intrusion, at about 400 Ma (Flinn, 1994a). Sheets of spherulitic rock, unlike the other variants, are commonly folded with the Muness Phyllite, and the spherulites are elongated parallel to fold axes; they must, therefore, have been intruded before emplacement of the Upper Nappe.

Microtonalite

Several sheets of microtonalite, less than 1 m thick, cut amphibolite of the Westing Ultramafic Zone near Stacks of Poindie [HP 577 137]. They are similar petrographically to rare sheets of tonalite in Yell (Flinn, 1994a) and, like them, are exceedingly difficult to distinguish in the field from the rock they intrude; they can be distinguished from granofelsic metamorphic rock only by their continuity and contrasting internal fracture patterns.

Chapter 8 The Shetland Ophiolite-complex: tectonics and origin

A model for the Shetland Ophiolite-complex

The Shetland Ophiolite-complex is presented in Chapters 6 and 10 as consisting of two ophiolite nappes, one above the other (Figure 41), each of which is underlain by zones of imbricated slices of ophiolite, metasedimentary rock, metavolcanic rock, hornblende schist and basement schist and gneiss. The ophiolite- complex contains no mélange or olistostrome.

The ophiolite nappe (including the mantle component) was uniformly, pervasively and almost completely hydrated at the time it was obducted. Ultramafic rock had been lizardite serpentinised and basic rock had been amphibolitised (to actinolitic- hornblende) and saussuritised. Lizardite serpentinised ultramafic rock adjacent to the obduction thrust was statically recrystallised to antigorite serpentinite during obduction. Hydration of oceanic or marginal basin floors is generally held to be due to sea water incursion through fractures, and to extend to relatively shallow depth in proximity to the fractures (Pickup et al., 1996; Muller et al., 1997; O’Reilly et al., 1996). However, thorough hydration of the Shetland ophiolite, and the lack of evidence of an association between hydration and fractures, suggest that a different source of fluid is required. Water escaping upwards in a subducting slab is one possibility. It has been suggested (Flinn, 2001) that, before obduction, the Shetland ophiolite was part of the floor of a marginal basin or ocean (see below) that was subducting to the east. Water expelled from deeper parts of the subducting slab rose through the overlying mantle and shallower parts of the subducting slab, causing pervasive hydration. It also facilitated partial melting along the interface between the subducting slab and overlying mantle (cf. Guillot et al., 2000).

A model of subduction preceding obduction is supported by an interpretation (Chapter 6) that the Quasi-sheeted Dyke-swarm and bodies of metamicrogabbro and ‘plagiogranite’ were intruded into the ophiolite from above the subducting slab.

The eastwards-subducting slab, or a part of it, was obducted westwards onto the Laurentian continent and its cover during an early phase of Iapetus Ocean closure at about 500 Ma. Lizardite serpentinisation at < 300º C prior to obduction left the Shetland slab too cold to generate dynamothermal metamorphism in underlying rock. MORB-type basaltic magma was generated by adiabatic partial melting of the mantle, during uncoupling in deeper parts of the obduction thrust. The magma was intruded up the thrust, during which it incorporated fluid released from the hydrated slab. The magma lubricated the thrust, and the resulting hydrated and tectonised basaltic rock was the protolith of the Norwick Hornblendic Schist. High- temperature brown hornblende-pyroxene-garnet rock (Norwick Hornblendic Schist type 3) forms the cores of discrete, lenticular magmatic bodies emplaced at the end of obduction. Selvages of chilled chloritic rock are developed against the serpentinite into which these bodies were intruded (Chapter 6).

During obduction, the lizardite in serpentinite at the base of the slab recrystallised to antigorite. Antigorite serpentinisation occurs at a higher temperature than lizardite serpentinisation. The additional heat may have been supplied by magma that crystallised to form the protolith of the Norwick Hornblendic Schist, and possibly also by underlying metamorphic rock.

The Lower Nappe is not an obducted, multilayered length of sea floor lying on a layer-parallel obduction thrust, as is usual or usually assumed for ophiolite. Instead, it is a slice cut steeply through the layered sequence, which is lying on its side on the thrust. The Lower Nappe presents, at the present surface, a continuous, uninterrupted traverse across steeply dipping ophiolite layers. Geophysical evidence suggests the layers are truncated at a depth of 1–2 km (Chapter 10).

A nappe of this type could be created by an obduction thrust rising through the mantle from the east and cutting through the subducting sea-floor slab from above. A model of the ophiolite as an obduction thrust- parallel length of sea floor and underlying mantle would require complex post-obduction tectonics, for which there is no evidence.

Following obduction, the Lower Nappe and exposed basement were eroded, and the resulting detritus was probably deposited in a freshwater basin lying, at least partly, on the eroded surface of the Lower Nappe. The deposits consisted mainly of laminated sediment, some of which was graphitic, with beds of coarse sandstone and conglomerate. Many of the coarser beds are rich in ‘plagiogranite’ debris, in the form of clastic grains of quartz- and albite-phyric felsite, and pebbles of leucotonalite. The pebbles in other beds of conglomerate consist mainly of metagabbro. The Gruting Greenschist may have been deposited at this time.

Following a period of erosion and deposition, the mantle tail of the obducted slab was thrust westwards over the Lower Nappe and its overlying sedimentary rock to become the Upper Nappe. It transported beneath it tectonic slices cut from the Lower Nappe and from the Unst Phyllite Group (including the Muness Phyllite and Gruting Greenschist). The similar lithological associations of the obduction thrusts underlying both nappes, each of which has an overlying antigorite serpentinised layer and underlying tectonic lenses of hornblende schist, indicate that this was not a second obduction, but a telescoping of the original obducted slab. The antigorite zone at the base of the Upper Nappe, unlike that at the base of the Lower Nappe, has been crudely and often phacoidally sheared during reactivation of the obduction thrust as the Upper Nappe was emplaced.

Emplacement of the Upper Nappe over the Lower Nappe caused the Middle Imbricate Zone to form between them. The Upper Nappe once covered the Lower Nappe, but is now represented by a number of klippen. These crop out in two subparallel chains along the west (Sound Gruney [HU 580 962], Gallow Hill [HP 575 008], Watlee [HP 603 057] and Hill of Caldback [HP 608 068]) and east (Vord Hill–Haaf Gruney [HU 622 936]–[HU 635 983], Muness [HP 636 009], and Hill of Clibberswick [HP 660 130]) sides of the Lower Nappe. This spatial disposition suggests the Upper Nappe had an antiformal structure, as opposed to the synformal nature of the base of the Lower Nappe (Figure 41). The Middle Imbricate Zone has been metamorphosed and deformed between the Lower Nappe and Upper Nappe. Metasedimentary rock has an L-dominant north-north-east-trending fabric resulting from parallel orientation of fold axes, crenulation axes, pebble elongation, girdle and partial girdle axes of phyllosilicate minerals, and intersections of penetrative and spaced schistosities. The resulting rock is L-tectonite or L–S-tectonite. There is no direct evidence for the direction of obduction or of thrusting. However, rocks of the Middle Imbricate Zone were folded and forced to elongate in a north-north-east direction, parallel to the large-scale fold axis, by a constriction deformation. It is reasonable to assume that pressure was from the east-south-east, and thrusting and obduction were towards the west-north-west.

The nappe pile as a whole is inferred to have been driven farther west onto the basement at the time of Upper Nappe emplacement. It overrode slices of metasedimentary rock from the Middle Imbricate Zone, and sheared off and overrode slices from the underlying basement to form the Lower Imbricate Zone. The western edge of the Gallows Hill Klippe was also sheared off parallel to the obduction thrust along the western edge of the Lower Nappe (Figure 41).

The Skaw Granite has a ⁴⁰Ar-³⁹Ar age of 425.6 ± 2.6 Ma, and is interpreted as having been emplaced during Upper Nappe emplacement. It thermally metamorphosed both the Saxa Vord Block and the north end of the Lower Imbricate Zone, and blocked further movement on thrusts bounding the Lower Imbricate Zone. Continued westward movement of the nappe pile resulted in formation of the Hevda Thrust, which bypassed the Skaw intrusion (Figure 27) and offset the thermal aureole. The Burra Firth Lineament may have played a similar role. Simultaneous emplacement of the Skaw intrusion and the Upper Nappe caused the granite and rocks of the enclosing Saxa Vord Block to be deformed.

The Shetland Ophiolite-complex evolved in four stages. It was first subducted eastward as oceanic or marginal basin crust. It was then obducted westwards at about 500 Ma onto the Laurentian plate margin, forming what is now the Lower Nappe. Following erosion of the nappe and deposition on it of sediment, a mantle part of the ophiolite was thrust westwards onto the Lower Nappe, forming what is now the Upper Nappe. This is interpreted to have been contemporaneous with intrusion of the Skaw Granite at c. 425 Ma, and is likely, therefore, to have been a consequence of Scandian deformation. Flinn et al. (1991) described similarities between the evolution of the Shetland Ophiolite- complex and contemporaneous ophiolites in Norway.

It has been suggested that the Shetland Ophiolite, like other Caledonian ophiolites, originated in a marginal basin (Moffat, 1987; Spray, 1988). Such a basin would lie offshore, within Iapetus, east of the Laurentian plate. The Shetland Ophiolite currently lies within the Laurentian plate, separated from its eastern edge by the Clift Hills ‘Division’ that also forms the Saxa Vord Group in Unst. The Clift Hills ‘Division’ is the infill of a late Dalradian extensional basin (Flinn, 1967b; Flinn and Moffat, 1985; Flinn, 2001), which is exposed onshore in the south of Shetland (Figure 42). At its base on the Mainland of Shetland is a sequence up to 0.5 km thick composed dominantly of volcaniclastic greenschist (the Asta Spilitic Formation). This is overlain by up to 2 km of phyllitic and turbiditic rocks, culminating in a layer 100–200 m thick of chloritoid pelite. The phyllitic rock includes thin-banded, graded rock of distal turbidite character, greywacke clastic rock, and a sequence of bedded turbiditic quartzite. On top of the sedimentary succession is a layer, up to 2 km thick, of volcanic rock, the Dunrossness Spilitic Formation. This commences with 0.5 km of clast-supported serpentinite breccia with traces of spinifex-like texture (Plate 7). The breccia has welded into a solid mass, and much of it has been steatitised. The serpentinite layer contains rounded lenticular and globular bodies of metagabbro, up to 100 m across. The ultrabasic layer is interbanded with, and overlain by, metabasaltic lava and basic volcaniclastic rock. The interbanding shows that the ultrabasic layer is an integral part of the volcanic succession. The final unit in the succession is a layer of basic meta-igneous rock dominated by pillow lava. It has an outcrop width of 1 km and is truncated to the east by the unconformity underlying the Old Red Sandstone Supergroup.

Major- and trace-element analyses are available for 12 samples of serpentinite and 17 samples of basic rock (Moffat, 1987). The serpentinite averages 47 wt per cent SiO₂ and 42 wt per cent MgO (dry). Despite the spinifex-like texture, the serpentinite composition is closer to ophiolitic mantle peridotite than to komatiitic peridotite (Moffat, 1987). The basic rocks are tholeiitic and have N-MORB-normalised spidergrams and other compositional features that are similar to features attributed by Pearce (1983) to partial melting of incompatible-element enriched subcontinental mantle.

Eruption of mantle-derived magma on top of sediment infill in the Clift Hills extensional basin is probably a result of continued extension leading to rupture of the underlying crust. Fracturing instead of gradual thinning by necking could cause adiabatic decompression within the subcontinental mantle, sufficient to result in melting and eruption of ultrabasic magma. Later, as the lithosphere and asthenosphere adjusted to the pressure change, partial melting would lead to eruption of basic lava and pillow lava.

In south Shetland, the exposed area of volcanic rock capping the extensional basin is relatively small (Figure 42) due to the unconformably overlying rocks. However, the outcrop is associated with spatially coincident positive Bouguer gravity (Tully and Donato, 1985) and aeromagnetic (IGS, 1972) anomalies. These extend offshore as major coincident lenticular anomalies that stretch towards north-north-east for 120 km. This suggests that the volcanic infill of the basin continues beneath the sea and the younger rocks to a latitude as far north as that of northern Unst (Figure 42). It is proposed that the Clift Hills extensional basin grew into an intracontinental marginal basin, and that the geophysical anomalies delimit what is left of it after subduction, later obduction of part of its floor as the Shetland Ophiolite, and final closure during Scandian deformation.

Chapter 9 Cenozoic

Landscape evolution

There is no known direct record of geological events in the district that occurred between Mesozoic faulting, involving the Walls Boundary Fault and the associated Bluemull Sound Fault with its 3 km dextral offset, and the last glaciation. The present landscape developed during this time when the islands were subjected to prolonged periods of subaerial and marine erosion, but some major elements of the landscape originated even earlier. For example, the Mid Devonian landscape unconformity to the south lies very close to the present surface. The crystalline rocks forming the Shetland islands are roughly at the same level relative to the earth’s surface as they were in the Mid Devonian, when they were buried beneath Old Red Sandstone strata. It is evident that the islands were an isolated protuberant mass rising above the surrounding area at the beginning of the Mesozoic. The unconformities at the base of the Mesozoic formations extend locally beneath the sea to within a few kilometres of the present coasts of Shetland, so that the Shetland archipelago is an erosional remnant retaining many characteristics of a monadnock today (Figure 43) (Flinn, 1973).

The indented form of the coastline of Shetland suggests that marine erosion by cliff retreat is a dominant process shaping the landscape. This is misleading, however, because even the hard rock outer coasts are relatively close to Mesozoic unconformities (Figure 43) indicating that there has been relatively little overall retreat considering the time that has elapsed. Furthermore, there are no extensive subaerial or submarine wave-cut platforms, or convex-upwards offshore profiles indicating prolonged cliffline retreat (Flinn, 1964, 1973). Such retreat would have been interrupted during periods of lower sea, and during high sea level stands when the cliffs were drowned and buried beneath younger sediments. Overall, the monadnock- like form of the archipelago originated as an erosional remnant, whether by subaerial and possibly marine erosion, prior to the Mesozoic and the cliffs at present forming the outer coast are the result of cliff retreat at intervals since then (Flinn, 1977).

Subaerial processes have been even less effective than marine erosion in reshaping the inherited monadnock landscape of Shetland. The slow net erosion rate may have arisen in part from periods of burial and protection beneath younger sediments now completely removed. However, as time has passed the decreasing area of the monadnock has reduced the efficacy of stream erosion. At the present time, no locality in Shetland is more than 5 km from the sea. Few streams are very much longer, especially on Unst and Fetlar. Under the present conditions of rainfall these short streams are mostly incapable of eroding the bedrock. Even in periods of heavy rain and cloudbursts the rain water tends to flow straight down the hillsides into the sea, minimising stream erosion. In Unst, only those streams draining the steep eastern slopes of Valla Field and the Burn of Skaw show signs of having eroded their beds. Throughout Shetland most streams are misfits and barely or not at all incised. Many have beds of peat or till they are unable to erode. Commonly the peat overgrows streams leaving them to flow for considerable distances in tunnels. Stream sediments arise not from bedrock erosion but from winnowing fine material out of till and regolith. Residual fragments are left to form an armoured bed. Glacially eroded features are also of minor importance (see below).

On Unst and Fetlar, as elsewhere in Shetland, it is clear that the topography and the drainage pattern are largely controlled by the lithology of underlying units. However, to the south in Shetland there are several transgressive east–west valleys (‘gaps’) cut across bedrock-controlled, north–south ridges, such as the Quarff gap [HU 420 350], the Mid Yell gap [HU 490 910], and the Voe gap [HU 430 630]. These appear to have originated as stream-eroded valleys predating the formation of the monadnock (Flinn, 1977). The inland topography of Shetland, and especially of Unst and Fetlar, is largely the result of weathering rather than stream erosion or glaciation.

Evolution of the coastline

The Shetland archipelago has been created by rising sea level during the Holocene (Firth and Smith, 1993). The deeply indented coastline is the result of partial drowning of the monadnock. The archipelago is bordered by a discontinuous outer coast of plunging cliffs forming the steep outer boundary of the monadnock (Figure 44) (Plate 8a). The individual islands and numerous inlets result from the inundation of valleys within the monadnock, forming an inner coast (Plate 8b). There are no raised beaches anywhere in Shetland, only a coastline rich in features characteristic of recent drowning (Flinn, 1974, 1977, 1994a). Rising sea levels have continued drowning of the interior landscape of the monadnock with minimal coastal erosion and have only trimmed up the cliffs of the outer coast.

The outer coast

Much of the south, east and north coasts of Fetlar and the east, north and much of the west coast of Unst forms part of the outer coast of Shetland. The outer coast plunges to –82 m (Figure 44), faces the open sea and experiences no protection or relief from the erosive effects of the ocean waves (Plate 8a) ((P534929), (P533321)). The cliffs of the outer coast truncate both valleys and hills. They are partially drowned so that cliff edges pass in a short distance from the summit of a truncated hill to beneath the sea. This is especially well shown by Hill of Clibberswick in Unst [HP 660 130] and Vord Hill in Fetlar [HU 630 940]. Elsewhere in Shetland, cliffs of the outer coast truncate valleys both above and below sea level (Flinn, 1994a).

Along the outer coast beaches occur only where the cliffs are weakened by shattering or deep weathering leading to the erosion of small cliff-foot platforms (e.g. Brei Wick [HP 640 170] and Looss Wick [HP 603 183] in Unst, and Corbie Head [HU 584 914] and Moo Wick [HU 623 877] in Fetlar). Small serpentinite-shingle beaches derived from screes occur at sea level on the Hill of Clibberswick cliff ( [HP 664 124]–[HP 665 133]) in Unst and the East Neap cliff ( [HU 628 945]–[HU 637 944]) in Fetlar (see Travertine section).

The inner coast

The drowned valleys forming the inner coast are locally called voes, firths, wicks and sounds. In Unst, Balta Sound and Uyea Sound (P534931) are good examples of drowned valleys, or voes. Within the voes the sea has little erosive power, being sheltered from the waves of the open ocean. The sea is only able to erode peat, glacial drift and weathered and/or shattered bedrock. Cliffs are only up to a few metres high. The hillsides pass almost continuously beneath the sea (Plate 8b) (P539931). In the inner recesses of some voes even the peat has not been eroded and has become drowned, as south of Belmont [HP 567 006] and under the pebble spit west of Urie Ness [HU 593 943].

Along the shores of the inner coast, beaches form continuous wave-swept, seawards-sloping platforms that narrow to a metre or less in width at high tide. Some shores are ice-scoured bedrock surfaces swept clean of peat and glacial drift by the sea, with glacial striae still preserved. Others are covered by pebbles or by sand. Spits, bars and tombolos are less common in the district than along the inner coast of the rest of Shetland (Flinn, 1977, 1994a), largely because there is relatively less coast of the inner type (Figure 44). These features are characteristic of recently drowned coastlines associated with postglacial sea-level rise. A sand tombolo joins Huney to Unst at very low tides [HP 647 065] (P534009) and a boulder tombolo joins Holm of Heogland to Unst [HU 575 993] (Plate 8c). The spit on the east side of Brough Holm [HP 567 058] owes its existence to a local unusually rich supply of till. Very substantial mid-bay bars of sand have formed across the heads of some voes to form lochs, such as Loch of Cliff, at the head of Burra Firth [HP 610 140] (Plate 8d) (P533360), and Papil Water [HU 605 903] in Fetlar (P534732). Several small lochs are dammed by a combination of sand and boulders at Uyea Sound [HP 594 013] and [HP 598 014], Heogland [HU 576 996] and Urie Ness [HU 596 943] (P536006).

The intermediate coast

In places in Shetland there is a rapid change from outer to inner coast (e.g. at the entrance to Balta Sound). Elsewhere there is an extended intermediate coastline between the inner and the outer coasts, where some of the characteristics of both occur (Flinn, 1994a). The coastline of Bluemull Sound [HP 550 020], Skuda Sound [HP 610 000] and between Wick of Collaster [HP 573 073] and Hoga Ness [HP 557 005] are all good examples. There, cliffs alternate with small, often sandy, beaches. Geos are characteristic features. They are narrow and deep inlets eroded into low cliffs, along faults, shatter zones and prominent joints (Plate 8e). The best example in Unst is probably Longa Geo at Westing [HP 575 088] (P533828) and Longariva Geo at Funzie Ness [HU 656 884] in Fetlar.

Beaches

Sandy beaches

Sandy beaches (Table 18) up to 600 m in length (Sandwick [HP 610 020]) occur in the district. Nickson (1980) conducted an exhaustive and meticulous search and found at least some sand on 60 beaches. Small patches of coarse, shelly sand were observed in cracks in the rocks in the intertidal zone at the foot of cliffs, and on shingle and boulder beaches, especially at mid-tide level. Double (1939) collected and described samples of sand from 12 beaches in Unst. Mather and Smith (1974) included nine from Unst and two from Fetlar in an account of Shetland sandy beaches. During the survey for this memoir 19 sandy beaches were recognised on Unst and its islands, including six associated with blown sand areas. Eight sandy beaches on Fetlar include two with blown sand (Figure 44). Allen (1983) made a study of the sea-floor superficial deposits around Orkney and Shetland largely based on BGS collections. He found a patchy distribution of shelly carbonate sands and gravels generally less than a metre thick. Along the east side of Unst and Fetlar these were generally 75–100 per cent carbonate while north-west of Unst the carbonate content was found to be 50–75 per cent.

Aerial photographs and detailed charts show that the larger sand beaches are onshore extensions of sand patches on the sea floor, e.g. Sandwick [HP 620 020] and Tresta [HU 606 903]. At very low tides some shingle beaches to pass down into sand beaches e.g. Muness [HP 634 015] and Uyeasound [HP 595 014].

The carbonate content of sand varies considerably from beach to beach, varying between 0 per cent and 90 per cent . The beaches richest in carbonate occur on the east coast of Unst (>50 per cent carbonate). They are closest to the outer coast and the sand-covered sea floor (Allen, 1983). Unusually low carbonate contents (<25 per cent) were found for beaches close to the outer coast at Skaw [HP 660 165] and Inner Skaw [HP 663 157]. This is probably due to copious production of sand from the Skaw Granite, which is strongly weathered and verges on friable in places.

The heavy-mineral content of the beaches varies widely due to variable sorting and concentration by wave action. The Burra Firth beach [HP 610 140] provides a particularly good example of this. The western half of the beach is rich in muscovite. The eastern half of the beach is rich in garnet and magnetite leading to black and red streaks in the sand due to sorting by wave action and fresh water flowing from springs at the back of the beach.

Double (1939) found a close association between types of heavy minerals present in the sands and local bedrock mineralogy. He attributed this to local erosion of the bedrock. However, the till is also closely associated mineralogically with the local bedrock and is more easily eroded.

Beg (1990) reported the heavy mineral fraction in seafloor sediments east of the district to be ‘unusually high’ in kyanite (>5 per cent), sillimanite (<1 per cent), staurolite (>4 per cent) and garnet (>20 per cent) and attributed it to erosion of till. He found an isolated group of samples from the floor of Wick of Gruting [HU 650 920] to be rich in heavy minerals (>5 per cent), especially magnetite and chromite. The beach sands and the phyllites forming the cliffs and shores of the Wick of Gruting are rich in magnetite.

It seems likely that most of the non-carbonate content of the sand on Shetland beaches has been derived locally from till, and drift. A proportion of the carbonate content has also been derived locally from animals in the intertidal zone. Beaches close to the outer coast have generally derived most of their sand including the carbonate content from sea floor sands. Sands dredged from offshore of carbonate-rich beaches (in the south of Shetland, unpublished results, D. Flinn) are generally richer in carbonate than the beaches and become increasingly so with increasing depth/distance from the beach.

Shingle and boulder beaches

Shingle beaches composed of well-rounded pebbles several centimetres in diameter occur in many places. Beaches entirely composed of serpentinite pebbles are common along the serpentinite coast (Plate 8f). At Burra Firth [HP 615 140], Westing [HP 569 057], Uyeasound [HP 595 013] and Muness [HP 635 014] pebble beaches form barriers cutting off marshy areas or even fresh-water lakes from the sea. Boulder beaches, including storm beaches are common along the outer coast, but rare at the foot of cliffs due to their plunging nature. Many boulder beaches in Unst and Fetlar are raised storm beaches, and include some so overgrown by grass that they may be considered to have been ‘abandoned’. Abandoned storm beaches occur on extreme outlying points of the outer coast where the cliff top is 10–20 m above sea level: Ham Ness [HP 636 020], Balta [HP 658 090], Holm of Skaw [HP 668 172], South Holms [HP 573 103], Spoo Ness [HP 568 071] and Rams Ness [HU 608 872].

Pleistocene

Late Devensian glaciation

Most now agree that Shetland was covered by an independent ice cap during the last glaciation (Flinn, 1977; Sutherland and Gordon, 1993). It extended about 65 km to the south-east into the Fladengrund area of the North Sea. There the position of a former ice margin facing south is marked by a series of giant tunnel-valleys in the sea floor excavated by melt water escaping from beneath the ice (Flinn, 1967a, 1978). These valleys have not been filled with younger sediments, unlike those elsewhere in the North Sea cut during earlier glaciations. Cameron et al. (1987), Johnson et al. (1993), Peacock and Long (1994) have reported an eastern limit to the Shetland ice cap about 70 km east of Shetland. In the north of Yell and Unst the presence of subaerial glacial, and subglacial, drainage channels in areas otherwise lacking evidence of glacial erosion suggests a northern margin thereabouts (Flinn, 1983, 1994a, 1994b). This is supported by the submarine work of Long and Skinner (1985) and by Clark et al. (2004) who reported offshore moraine which contacts the north coast of Unst. There is evidence of an ice margin to the west of Shetland, on the edge of the continental shelf (Stoker and Holmes, 1991).

From Shetland the ice flowed to the south over Fair Isle (Flinn, 1978) to meet the ice over Orkney.

Within Shetland an ice shed extended from Dunrossness in the extreme south to southern Yell and from there north-eastward across the sea to Uyeasound and up Unst on the east side of the central valley (Figure 45). The ice flowed to the west and to the east on either side (Flinn, 1977).

It has long been apparent that the Shetland ice cap is of Late Devensian age on account of the unweathered state of the till and of glaciated surfaces protected from current weathering. Also, the pattern of dates, found since Hoppe (1965), of basal deposits of peat of about 12 000 to 10 000 BP, match those of Late Devensian glaciated areas outwith Shetland. The oldest dated peat deposit resting on till in Shetland has a corrected age of 12 950 BP (Birnie et al., 1993). Seafloor studies mentioned above, support this conclusion, which is now generally acknowledged (Sutherland and Gordon, 1993). Older glaciations recorded by occurrences of peat below till occur elsewhere in Shetland (Mykura, 1976). No supporting evidence has been found in the district. However, on the House Wick beach [HU 550 920] of Hascosay, it is sometimes possible to find pebbles of flint, tönsbergite and rhomb porphyry. These rocks from London and Tönsberg came to Shetland as ballast in the ship Krageröe that was wrecked on Hascosay in 1803 (Flinn, 1977). By coincidence, a large block of tönsbergite, originating from south of Oslo, was found in a roadside till quarry in the south of Shetland [HU 403 159] (Finlay, 1926; Flinn, 1992b; (P573856)), suggesting an earlier glaciation of Shetland by ice from Norway. Similarly, in the same area of Shetland, rocks occurring along the east coast have been carried westward over the Clift Hills (Flinn, 1977).

Ice flow indicators

The flow pattern of the Shetland ice cap in the area of Unst and Fetlar is revealed by glacial striae, parallel ‘glacial lineaments’ (visible on aerial photographs) that may be incipient glacial flutes and gouges and, less precisely, by the distribution of erratics, stream sediments and glacial drift (Figure 45).

Some 60 sets of glacial striae were found in Unst, but only 12 in Fetlar (P534202). About two thirds give the sense of direction of flow, as well as the azimuth of flow, and point westwards on the west side of Unst and eastwards on the east side of Unst and throughout Fetlar.

Many well-formed roches moutonnées occur in south- west Unst, with several farther north on the top of the Valla Field ridge to the south of Woodwick [HP 582 115]. Roches moutonnées are especially well developed between Lunda Wick [HP 570 040] and Snarra Voe [HP 560 020]. The best-developed roche moutonnée in Shetland occurs at Snarra Voe [HP 5635 0250] (P534199). Glacial lineaments are prominent on 1:25 000-scale aerial photographs of Unst, but not on the ground. They are well developed between Belmont [HP 565 010] and Wick of Collaster [HP 573 075] and north and south of Wood Wick [HP 580 115] (Figure 45).

Many glacial erratics are distinguishable on Unst and Fetlar, including serpentinite, metagabbro, pegmatitic metagabbro, clinopyroxenite-wehrlite, Norwick Graphitic Schist and chloritoid-bearing schists. Erratics occur in Unst as far north as Skaw in the north-east and Wood Wick in the west. Both metagabbro and serpentinite are prominent erratics on the Valla Field Block. On the east side of Unst, due to the trend of the metagabbro- serpentinite boundary parallel to the ice shed direction, transported rocks tend to remain within their outcrop. However, blocks of metagabbro have been carried eastwards onto the Muness Phyllite in the Muness and Uyea areas (Figure 45).

In Fetlar blocks of chloritoid-bearing pelitic schist from the west side of the Lamb Hoga peninsula have been carried eastwards and are now widely but sparsely scattered across Lamb Hoga and the rest of Fetlar (P534886). Metagabbro erratics are to be found within the serpentinite outcrop in the centre of Fetlar and serpentinite blocks on the Funzie Conglomerate farther east. Erratics of many other lithologies occur to the east of their outcrops, in particular graphitic phyllites in central Fetlar, and in Unst on Hill of Clibberswick and the south coast of the Skaw Granite.

On the higher parts of Saxa Vord, erratic-like boulders of vein quartz and of chloritoid-kyanite segregations in vein quartz occur within the outcrop of the Saxa Vord Pelite, their source rock. The lack of evidence of glacial activity in these areas (see below) suggests they are remanié boulders rather than erratics.

Glacial deposits

Small exposures of till, up to 5 m thick, are visible in the cliffs around the islands, showing that till forms small pockets filling hollows in the bedrock surface, rather than a continuous sheet (Figure 46). Evidence of more than 50 such pockets were found in Unst and half as many in Fetlar. Till is seen only in cliffs or the banks of streams. Elsewhere it is covered by peat and/ or vegetation or is indistinguishable from the thin drift or regolith layer. In Unst, the most substantial deposits are exposed in the cliffs west of Wood Wick [HP 575 120], on Houllnan Ness in Westing [HP 565 055] and between Uyeasound [HP 605 004] and Ness of Ramnageo [HU 625 995]. The till is of lodgement type, well compacted and usually matrix-dominated. The clasts are generally no more than a few centimetres in greatest dimension and the matrix contains more sand than clay. This till is very similar in appearance and occurrence to the till found elsewhere in Shetland. However, in south-west Unst, especially between Lunda Wick and Uyea Sound, the till is dominated by boulders, possibly associated with glacial tectonics (see below). At the back of the small sandy beach on the east side of Lunda Wick [HP 573 043], half a metre of bouldery drift or till can be seen overlying not less than several metres of the lodgement till. Much less till occurs in Fetlar.

In areas not covered by peat or vegetation the bedrock is seen to have a thin mantle of ‘glacial drift’ or regolith. It is probably residual from the thin and the very patchy till layer.

No glacial drift (or till, erratics or striae) was found in the Herma Ness and Saxa Vord areas and, as with a similar area in north-west Yell (Flinn, 1994a), these localities are considered to have been unglaciated in the last glaciation. On the north-west coast of Herma Ness, at Taing of Loosswick [HP 603 184] and for about one kilometre to the south and on Tonga [HP 580 150] the bedrock at the cliff edge is covered by a metre or so of ‘dirty’ coarse sand locally rich in clay. This deposit is very similar to several cliff edge exposures of coarse poorly-sorted sand resting on the bedrock in the area of north-west Yell also considered to have been unglaciated in the last maximum (Flinn, 1983). In both areas the sandy deposit may be a dune deposit or be niveoaeolian in origin (P533294).

The deduced pattern of ice flow

In south-west Unst, south and south-east of Wick of Collaster [HP 573 073], the transport of metagabbro fragments onto the serpentinite outcrop west of the metagabbro and of both metagabbro and serpentinite fragments onto the Valla Field rocks as far west as the west coast, confirm that the ice shed lay to the east of the metagabbro–serpentinite contact in the Uyeasound area (Figure 45). Farther north, the presence of serpentinite and the absence of metagabbro as erratics and drift in the Woodwick–Loch of Cliff area show that the ice shed lay entirely within the serpentinite outcrop between Baltasound and Ungirsta [HP 623 120]. Between Woodwick [HP 580 114] and Wick of Collaster [HP 573 073] and to the east, serpentinite fragments occur only very sparsely in the drift and erratics are few. It is possible that the paucity of serpentinite erratics in this area is due to the ice shed lying close to the western edge of the serpentinite outcrop up ice-stream of this area. On the other hand, the Valla Field ridge reaches its greatest height hereabouts and may have dammed the ice flow and reduced its transporting effectiveness.

The combined evidence of the striae, the lineaments and transported rock shows that the ice to the west of the ice shed flowed to the west-north-west. In south-west Unst flow was diverted westwards round the south end of the Valla Field cuesta ridge and after that to the north to join the strong Bluemull Sound ice stream flowing to the north-north-west between the Valla Field ridge and the north-east coast of Yell (Flinn, 1994a and b). On the east side of the ice shed in Unst striae are sparsely and poorly developed, there are no roches moutonnées, and no photolineaments have been detected. Furthermore, lithological boundaries tend to run parallel to the direction of ice flow rather than across it, so that direct evidence of rock transport is limited to easterly pointing striae. Only in south-east Unst and on Uyea does the carriage of metagabbro (and rare serpentinite) erratics onto the Muness Phyllite outcrop east of the metagabbro provide evidence confirming the eastward sense of flow obtained from the striae.

The only evidence of ice flow in the north of Unst was found in the Norwick area. On the south shore of Nor Wick [HP 656 141], striae pointing to the north-east occur associated with till and drift containing graphitic phyllite derived from the Norwick area to the south-west. Erratics of serpentinite and areas of drift rich in graphitic phyllite and serpentinite fragments derived from the south-west occur on the Skaw Granite along the north shore of Nor Wick [HP 670 150]. North of Lamba Ness several pockets of till occur, which do not contain serpentinite, but do contain chloritoid phyllite derived from Saxa Vord to the west (Figure 46). This till is probably the product of a small glacier filling the Skaw Burn valley.

Throughout Fetlar evidence suggests that ice flowed from the west, but the effects on the landscape have been minimal. As on the eastern side of Unst, there is a paucity of striae and till and an absence of photolineaments that can be attributed to glacial processes. On the islands between Unst and Fetlar and in the north of Fetlar the flow was to the northeast. In the south tip of Lamb Hoga [HU 608 872] and Funzie Ness [HU 655 885], in the south of Fetlar, the flow was to the south-east. The topographical masses of Lamb Hoga and Vord Hill seem to have diverted the ice flow to the north and south.

Many of the Regional Geochemical single-element maps of Unst and Fetlar based on stream sediment analysis show a strong glacial influence. The elements Mg, Cr, Ni, Co, Pt, Pa, characteristic of the serpentinite rocks of the ophiolite are strongly represented on the basement to the west of the ophiolite between Tonga and Wood Wick and farther south in the Westing area. In these areas the elements characteristic of the basement rocks (Figure 47), in particular Rb, Li, K, Ga, Al and Si, are partially blotted out, though Ti, Sr and Ba are very little affected. The areas concerned are those of strong westward flow of ice from the Uyeasound area, round the south end of Valla Field, to join the Bluemull Sound ice stream in the Westing area. Also, the area of flow from the ophiolite across the basement between Wood Wick and Tonga is similarly affected. The Saxa Vord and Herma Ness areas considered to have been free of ice have only background levels of the ophiolitic elements in common with the area between Woodwick and Westing, shown to have been less strongly glaciated than the areas to the south and north. The ice flowing to the east has had less influence on the stream sediments due to lesser contrast within the bedrock. However, chromite and garnet in panned heavy mineral concentrates have been displaced eastwards into outcrop areas lacking them in situ (Figure 48). The strength of the glacial influence on the stream sediments arises from the fact that they have been winnowed out of the glacial drift/till rather than eroded directly from the bedrock.

Evidence of glacial tectonics

Evidence of thrusting caused by ice flow can be found in many coastal sites in Unst (Figure 46). In all these places crudely formed horizontal thrusts occur close to the base of the till, either within it or in the underlying bedrock. Where these thrusts lie within the till they form sharply defined ruptures. Where they lie within bedrock or below a slice of bedrock overlying till, they are zones no more than half a metre thick of coarsely ruptured bedrock (Plate 9a). The bedrock below the thrust, and more especially above it, is always loosened along its internal fractures and partially disrupted. Thrust slices of bedrock vary from a metre to several metres thick and in some places can be traced for several tens of metres passing into very bouldery till dominated by blocks of local rocks. The best occurrence is on Brough Holm in Westing [HP 565 058], but very good and more accessible examples are to be seen in Wick of Collaster ( [HP 572 074], [HP 573 073]–[HP 576 073]) (P533951), Clivocast ( [HP 603 004]–[HP 611 003]) and Balta Sound ( [HP 644 091]–[HP 649 091]). In the Westing area thrusting has been facilitated in places by the flat-lying steatite layer in the Westing Ultramafic Zone (Chapter 3) (e.g. Wick of Collaster [HP 753 073]; [HP 560 050]) (P534297). Most exposures of till between Snarra Voe [HP 565 019] and Uyea Sound show signs of thrusting. Glacitectonic thrusting is rare elsewhere in Shetland so the presence of these examples in Unst may be related to the close proximity of the former ice front. Basal ice is likely to be colder near the front than within the ice cap and more likely to be frozen to the bedrock.

Boulder fields and periglacial phenomena

On the east side of Unst, between Hill of Clibberswick and Uyeasound, and in central Fetlar, there are many areas strewn with perched boulders a metre or more in diameter. The boulders have not been classed as erratics because they occur within the outcrop of the rock from which they have been derived and so cannot be shown to have been transported. Examples in Unst occur on the serpentinite on the south end of Hill of Clibberswick, and on the Crussa Field [HP 615 110]–Nikka Vord [HP 625 108] ridge and on the metagabbro outcrop between Ward of Clugan [HP 640 070] and Vord Hill [HP 610 030]. Similar fields occur in Fetlar on the serpentinite on Vord Hill and at Still [HU 044 910] (Figure 46).

On both serpentinite and metagabbro outcrops boulder fields tend to be associated with ice-tumbled protuberant exposures, (Unst [HP 628 056], [HP 604 045]; Fetlar [HU 631 929], [HU 636 921]). In these, joints have been opened and the joint-defined blocks have been rotated from their original positions, without being carried away. Such joint blocks are similar in shape and size to the boulders forming the boulder fields.

On the southwards facing slope of Hill of Clibberswick a somewhat different type of boulder field occurs. Between the 50 and 100 m contours, and running approximately parallel to them, are three parallel low ridges composed of piled-up boulders ((P533595), (P533601)). The boulders vary up to several metres in diameter, and are half buried in the turf so that it is difficult to decide which, if any, are exposures of bedrock. These ridges extend from about [HP 656 127] in the west to about 200 m short of the cliff edge in the east. The ridges are probably boulder moraines. Immediately above them is an irregular row of shallow hollows in the hillside. The hollows are especially well developed along the 100 m contour in the neighbourhood of the planticrub at [HP 660 120]. They are rocky backed and have flat horizontal grassy floors, many of which are the sites of ephemeral springs (P533596). They have the appearance of abandoned small quarries, the entrances to which are about 50 m or so wide. These pseudo- quarries are probably nivation hollows that formed above a glacial trimline surrounding the summit of Hill of Clibberswick, which stood above the surface of the ice cap as a nunatak. The hollows are closely associated with the boulder moraines.

Well-defined fields of frost-shattered bedrock occur in three places in Unst (Figure 46). The Keen of Hamar [HP 645 095] occurrence (better known for its botanical interest) is within the metadunite outcrop. Here, the hillside extending from the summit of the Keen [HP 649 099] to the Hagdale quarry [HP 639 103] is a continuous, loosely packed layer of metadunite fragments one or two centimetres in diameter. Exposures of bedrock within this area are densely cracked isotropically. The cracks are tight, but form a weakness that has allowed frost action to separate the rock into fragments. Neither the cracking of the bedrock nor the regolith sheet of fragments occurs elsewhere. In places the surface is covered by a heavily vegetated soil layer, but where the frost shattered layer is exposed it is apparent that active periglacial processes are producing well-formed stone stripes on the steeper slopes (Carter et al., 1987) (P533745).

Blockfields of much coarser, frost-shattered debris occur on Virda Field [HP 628 068] (Plate 9b) and Sobul [HP 602 036] (P534190) within the metagabbro outcrop and on Hill of Caldback [HP 610 066] within the serpentinite outcrop. These are the result of frost action producing in-situ layers of frost-shattered bedrock at least a metre thick. The layers are composed of angular fragments ranging in size from 10 cm to half a metre or more across. The Virda Field occurrence is limited to the north by the metagabbro contact against the serpentinite. Although this junction occurs on the steepest part of the hillside, relatively few metagabbro fragments have crossed the boundary (P534014). The other boundary of this occurrence and the boundaries of the other two are obscured by growths of sphagnum moss and heather. The lack of dispersal of the frost-shattered fragments on the Keen of Hamar and on Virda Field, Sobul and Hill of Caldback indicates that the shattering took place after the ice sheet had melted.

Glacial erosion

The small volume of till present in the district, and indeed in Shetland as a whole, is consistent with the fact that no part of the islands, and especially Unst, is very far from the former ice shed. This proximity and the relatively low topography of the area limited the amount of glacial erosion that could take place within the islands, and therefore, the amount of till that could be generated there.

The lakes and glacial lineaments provide clear evidence of glacial erosion. However, the lineaments are too insubstantial to be detected on the ground and many of the lochs and lochans are the result of damming by peat and till, rather than of erosion. The topography of Unst and Fetlar, including the sea-cliffs, is probably much as it was before the glaciation (except for glacial drainage channels). The central valley along the Lower Imbricate Zone, the klippen forming the Hill of Clibberswick and Vord Hill, the Saxa Vord, Valla Field and Lamb Hoga blocks are structurally and lithologically controlled. They lie across any conceivable ice flow direction, whether from a local ice cap or, during earlier glaciations from Norway, and like the rest of Shetland topography, they show no signs of being glacially shaped.

Deglaciation

Glacial drainage patterns are better and more widely developed in Unst than in Yell (Flinn, 1995a) and are almost absent from the rest of Shetland. Drainage features were formed by water flowing under the ice, subaerially, or both in succession.

Drainage channels

Subglacial channels are best seen at the south-west corner of Burra Firth [HP 610 150]. It is argued below that they were formed by water draining away under the ice from the Milldale Glacial Lake into Burra Firth (Figure 49). These channels are characteristically U-shaped with steep sides and sharply formed rims. Their courses follow contours, cross them obliquely and run up and down hill indicating that they were eroded by water flowing hydrostatically, as in a siphon (Flinn, 1992a) (Plate 9c). They vary in size from several metres to several tens of metres in width. The best example in Unst is that occupied by the car park at Herma Ness Lighthouse Station [HP 612 149] ((P533364), (P533048)). At Norwick [HP 643 142] a short length of a much smaller subglacial channel, formed by an englacial stream, cuts across the top of a ridge. Other examples of subglacial channels occur on the eastern flank of Libbers Hill [HP 566 136] where water issued from beneath the ice to fill the Milldale Glacial Lake. Most are rather minor features that start abruptly on a hillside and continue for some distance down it as U-shaped gulleys with no other signs of water having flowed in them.

On a larger scale the Norwick Burn drains water from the Northdale area through the watershed at [HP 643 135] (P533477) to the sea at Nor Wick. This probably started as a subglacial drainage channel, which later carried subaerial ice-melt water and now drainage water from the area to the west of Haroldswick. Water from this area also drained westwards through a topographical ridge now occupied by the Quoys branch of Loch of Cliff ((P533471), (P533469)). A more substantial drainage channel cutting through a watershed is that through Dale of Woodwick [HP 587 105]. This was probably formed under the ice, as its floor slopes to the east on the east side of the watershed, and to the west on the west side ((P533543), (P533628)).

Ice-dammed lake

During deglaciation an ice-dammed lake formed in the Milldale area (Flinn, 1992a). In this region, between Tonga [HP 580 150] and Burra Firth [HP 595 143], the land slopes south towards the ice margin. Water issuing from beneath the ice on the east flank of Libbers Hill [HP 596 136] cut two parallel subglacial gulleys there, and deposited to the north a glacial outwash fan of material washed out from beneath the ice. The contents of the fan are now exposed in the banks of the Milldale Burn [HP 598 140] (P533463). The area filled with water up to about 100 m above present sea level and overflowed the cliff edge on the south side of Tonga, eroding the rocky-sided Cat Houll gorge [HP 588 146], which falls 30 to 40 m to the west in 300 m (Plate 9d). Water from the lake also leaked away to the east into Burra Firth under the ice through the network of subglacial valleys incised into the hillside south-west of the Lighthouse Station. Because of erosion lowering the entrance to the Cat Houll gorge, the level in the lake fell to that of the col (95 m above present sea level) west of Mouslee Hill [HP 604 153], thus dividing the lake into two. Leakage under the ice to the east lowered the level in the north end of the lake. Water overflowing the col from the south cut a subaerial glacial drainage channel on the north side of Mouslee Hill [HP 605 157] (P533208). Eventually both parts of the lake leaked away to the east into Burra Firth. No shorelines have been detected in the area of the lake, the whole area being covered by a layer of hill peat several metres thick. The lifetime of the lake was probably very short.

Glacial drainage

Many of the drainage channels (Figure 49) can be arranged in a chronological order of use. While the Milldale Lake was held in place by ice filling Burra Firth and to the east and the south, glacial drainage from the central valley of Unst south of Burra Firth took place to the west via subglacial channels forming the Woodwick gap [HP 590 103] in the centre of Unst and the Gunnister gap [HP 588 039] in the south. In the Woodwick gap overflow started at 81 m above present sea level and cut down successively to 73 and finally 61 m above present sea level in different channels. In the Gunnister gap, overflow started at 62 m above present sea level and later cut down to 41 m above present sea level. This outlet is farther from the ice front than the one at Woodwick or the one at Burra Firth so it probably developed later. Drainage from the Gunnister gap took place via the deeply cut channel leading to Lunda Wick [HP 573 038], now partially filled with younger deposits. Also draining into this channel are several well-formed subglacial channels south of the Gunnister gap at [HP 583 030]. Flow through these gaps ceased when the ice in Burra Firth and to the south melted. Then the present pattern of drainage from the central valley into Burra Firth in the north and Uyea Sound in the south became established.

As long as the central valley was under ice, subglacial water overflowed into it from the east in a series of five subglacial channels along the Quoys–Saxa Vord ridge [HP 600 120] to [HP 620 150] at heights varying from 29 to 58 m above present sea level. However, once Burra Firth was free of ice, drainage water flowed freely westwards from the Haroldswick area, cutting a major subaerial glacial drainage channel through the water shed at the Quoys gap [HP 610 130]. The gap probably commenced as a subglacial channel, but ended as a subaerial channel, which still today drains the area west of Haroldswick, by means of a peat-floored stream too small to have eroded the channel it occupies.

Elsewhere in Unst and Fetlar there exist well-formed grass-covered incised drainage channels on hillsides that otherwise show no evidence of recent water flow. These include two small but lengthy channels (Plate 9e) close to the road east of Loch of Watlee [HP 5985 0516] (see below) on Unst, several in the Muness area [HP 625 020] (P534238) and a prominent group at the north end of Lamb Hoga [HU 595 905]. It is possible that these and several other less noticeable occurrences were formed by meltwater draining out from beneath residual ice masses perched above them on the hills on which they occur.

Glaciofluvial sedimentation

The only known glaciofluvial deposit in the district is a small mound of well-rounded pebbly gravel on the south-west flank of Saxa Vord at Buddabrake [HP 619 144]. This may be the locality referred to by Peach and Horne (1879) (P533388) as containing Skaw Granite-type glacial erratics (i.e. westerly transported drift). No such erratics were found in it during this survey.

Late Glacial events

The limit of the main Late Devensian ice cap appears to have remained 70 km or so to the east of Shetland close to the beginning of the Late Glacial Interstadial (Peacock and Long, 1994). The Loch Lomond Readvance was of very limited extent on Shetland (Sutherland and Gordon, 1993) and no evidence of it, or the preceding Late Glacial Interstadial has been found in the district. Evidence of the Loch Lomond Readvance is preserved in the area of the Central Shetland Sheet (Flinn in IGS, 1982), but much has now been removed as a result of road improvements.

Holocene

The study of cores obtained from peat bogs and lake deposits in several parts of Shetland has provided detailed vegetational histories of the Late Glacial period and Holocene. Elements common to them can be considered to apply to the district. Late Glacial peaty deposits began to form about 12 000 BP in Shetland (Hoppe, 1965, 1974; Birnie, 1993) although they have not so far been found in Unst. After the Loch Lomond Stadial an open herbaceous vegetation was first established. About 9500 BP birch woods appeared, to be followed after a hundred years or so by hazel, oak, juniper and alder. From 8000 to 5000 BP shrubs declined and heather increased. This change has been variously attributed to a wetter climate (Johansen, 1985) or human interference by Mesolithic man (Bennett and Sharp, 1993; Bennett et al., 1992). From 4650 BP Plantago lanceolate pollen becomes a significant constituent of the peat indicating agricultural activity in the area (Johansen, 1976). Woodland clearance extended from about 4000 to about 3000 BP (Bennett et al., 1992). From then on the vegetation was dominated by man’s activities, rather than by the climate.

Detailed studies of the peat in Shetland (outwith Unst) have been carried out by Johansen (1985), Bennett et al. (1992) and Birnie et al. (1993). The peat in the district is nearly all of the blanket type. In Orkney such peat started to accumulate at 3400 BP according to Keatinge and Dickson (1979), but in Shetland started to become important about 1400 years earlier according to Bennett et al. (1992), several hundred years before agricultural communities arrived. Recent work in Unst (Birnie et al., 1993) has shown that peat older than the blanket peat occurs in the bottoms of the glacial drainage channels reaching the sea at Norwick and Lunda Wick. There, peat at 5 m below present sea level is over 7000 years old (Smith, 1993). Two thin persistent layers of silty clay, with ages of about 4000 and 5000 years old, also occur there within the peat. At Norwick the basal layer is very woody. Fragments of wood occur in the hill peat near Burrafirth [HP 600 138] and Muness [HP 630 010].

Peat

The hill peat or blanket bog that covers or once covered most of Shetland up to depths of several metres, covered Valla Field, Saxa Vord and Lamb Hoga. How much of the rest of Unst and Fetlar was covered is not clear. It is widely held that the serpentinite outcrop was never covered, but no explanation for this has been offered, though the chemical properties of serpentinite are usually blamed. The lack of present-day vegetation on the frost shattered regolith of the Keen of Hamar has been attributed to aridity rather than chemical factors (Slingsby, 1981; Carter et al., 1987). It is not clear if the present absence of peat from the metagabbro is due to exploitation or other factors. The absence of peat from the Skaw Granite and the Funzie Conglomerate and the near absence from the Muness Phyllite between Muness and Uyeasound, on the eastern half of Uyea and on Turra Field in Fetlar [HU 615 915] is almost certainly due to exploitation for fuel. The peat cover of the Saxa Vord and Valla Field blocks is currently being actively removed. The Lamb Hoga Block was finally cleared about 1950.

Alluvial diatomite

During his survey of Unst, Read (unpublished field map) noted the occurrence of diatomite in the bottom of a long-since filled-in ditch draining a marshy area at Haroldswick [HP 639 124]. In 1939, an occurrence of diatomite at Small Waters [HP 587 029] in the south of Unst (B2 on (Figure 50) was reported (Sandison, 1948). The Geological Survey of Great Britain published a report with a chemical analysis (Haldane et al., 1940). Since then, diatomite (Tomter, 1958; Johansen, 1975) and diatoms in peat (Hoppe, 1965; Mykura and Phemister, 1976; Birnie et al., 1993), have been noted elsewhere in Shetland. Carter and Bailey-Watts (1981) have provided a detailed list of diatoms currently living in standing water in Shetland, but also outwith Unst and Fetlar.

During the present survey (Flinn, 1996a), the Sandison discovery was found to be only one of many small deposits of diatomite occurring within the peat-free outcrop of the metadunite south of Balta Sound (Figure 50). The deposits are shown by SEM images to be formed of close packed diatom frustules. Most deposits are presently covered by turf and are exposed only where the turf has been eroded, mostly in the banks of streams. They form lenticular layers, varying from a metre to several hundred metres across and from a few centimetres to 0.75 m thick, comprised of a white clay-like substance that is off-white when wet and brilliant white when dry. They overlie the glacial drift.

These diatomite deposits are of particular interest because they all occur on sloping hillsides within the metadunite outcrop, where standing water could not have collected since the melting of the ice cap. To distinguish them from the more common lacustrine diatomite they have been called alluvial diatomite (Flinn, 1996a). Highly diatomaceous muds occur in lochs and lochans in the same area, and elsewhere in Shetland, but always at depths greater then 0.5 m due to the effect of wave disturbance and where the surface of the water is below the base of any nearby alluvial diatomite deposits.

The deposits occur in a variety of situations. Some form lenticular bodies on sloping hillsides below ephemeral springs type A on (Figure 50). A7 is a turf-covered lens, 25 by 20 m in area and 30 cm thick, lying 10 m below a spring on a slope with a gradient of 1:150 (P534224). When the spring dries up, the water course is covered by a white crust, about a millimetre thick, composed of dead diatoms. A1 is a mound of diatomite less than 2 m in diameter and several centimetres thick (P534180), lying on stony glacial drift a few metres below an ephemeral spring and in the vicinity of its soak- away. The turf has been eroded from the area and the diatomite mound is an erosional remnant.

Another type occurs in the beds of the grassed-over glacial drainage channels e.g. C1 in (Figure 50). This channel is U-shaped, about one metre wide and half a metre deep, on a hillside with a gradient of 1:20. It locally widens to 3 m. In one such widening north-east of Heilia Brune [HP 5986 0516] diatomite was found below the turf as a one centimetre-thick layer of white diatomite, with faintly developed brown diatomite laminations parallel to the stream bed, resting directly on the gravelly stream bed. Immediately above it is a layer of peat, 1 to 2 cm thick, which gave a 14C age of 1570 ± 40 BP (Flinn, 1996a, figs 3a and 3b). Above the peat is a layer, about 1.5 cm thick, of pale brown (peaty) laminated diatomite and above that, up to 4 cm of white diatomite, capped by 3 cm of turf (Plate 9e).

Another type of deposit is exposed in the banks of active streams, or artificial drainage ditches B in (Figure 50). They are more substantial than the examples mentioned above and vary in thickness from several centimetres (P533878) to 0.75 m. They occur in very gently sloping areas, with no downstream obstructions to the free flow of water; standing water could not have existed. An example is provided by the example reported by Sandison at Small Waters, B2 [HP 588 029], which is on a col. Another example underlies a grassed-over floodplain-like area several hundred metres wide at B1 ( [HP 593 025], (Figure 50), across the course of the Clay Burn. The diatomite lies on the surface of the till as a layer up to 0.75 m thick, and is generally pure white and homogeneous, though near the base it locally contains brown (peaty) laminations and thin bands parallel to the base. A ¹⁴C age of 2380 ± 40 years BP has been obtained from a peat layer, several centimetres thick, underlying the diatomite at [HP 5931 0250] (Flinn, 1996a, fig. 2 shown upside down) (P534225).

Flatness of the upper surface of the B-type deposits gives them some of the appearance of filled-in lakes. However, the Clay Burn flows in an incised (artificial drainage?) channel through them on a till floor below the base of the diatomite. The burn also flows through lochs lying below the level of the diatomite layer, which have diatomite-free floors. Furthermore, there are no barriers or remains of barriers where the diatomite now occurs, which could have dammed the stream to form ponds.

It is concluded that the diatomite deposits of Unst accumulated from running water rich in diatoms, in places where standing water could not have existed. Being unable to settle in the running water they accumulated where it soaked away into the ground and through earlier formed accumulations. Therefore, these alluvial diatomite deposits are a type of sieve deposit. Their occurrence seems to depend on water draining from metadunite and may also depend on periods of wetter conditions than at present.

Travertine

Two types of postglacial cemented breccias occur in several serpentinite cliff localities in Unst and Fetlar, but are not known elsewhere in Shetland (Flinn and Pentecost, 1995). One type comprises brucite-hydromagnesite-cemented scree on cliff faces. The other comprises aragonite-cemented scree beach deposits at high tide level at the foot of the cliffs. In a number of places the till also has been lithified or cemented.

Cemented scree

Serpentinite sea cliffs sloping at 40o to 50o towards the sea, from heights of between 100 and 200 m, and plunging steeply beneath the sea, occur on Hill of Clibberswick [HP 666 130] in Unst and Vord Hill [HU 623 925] in Fetlar (Figure 51). The cliffs are faced with serpentinite bedrock, scree and grass. A recent scree at Clibberswick (Figure 51), locality X0 [HP 665 131]) is a cone of debris extending from just below the cliff top to sea level and below (P533618). It completely buries the bedrock cliff face. The scree is composed of serpentinite fragments varying in size from fine powder, especially near the top, to blocks 10 m or more in diameter at sea level.

On the Vord Hill cliff in Fetlar, erosional remnants of several older screes occur at several localities at X1 [HU 6289 9448]–[HU 6299 9442] above Il Holm (Figure 51). Where not covered by grass, the interior of the scree is exposed as serpentinite fragments loosely cemented by a white, very fine-grained coating varying up to one centimetre in thickness ((P534593), (P534523)). This coating is composed of one or more crudely formed layers of spherulites (P534518). The basal layer in contact with the serpentinite surface is usually 0.1 mm thick or less, and is composed of contiguous, perfectly formed hemispherulites (centres resting on the serpentinite) of close-spaced, very thin concentric laminations of brucite as confirmed optically by ‘fast-along extinction brushes’. Above this layer, there are spherulites to hemispherulites, up to one millimetre in radius, of aragonite needles intergrown with brucite. Between and on top of these large spherulites, there are much smaller and rather coarsely crystalline spherulites of hydromagnesite, exhibiting ‘radial slow-along extinction’. Aragonite also occurs as small radiating bunches (crude hemispherulites) of pale brown crystals, about 0.2 mm long, within, and on the surface of, the hydromagnesite coating. Calcite occurs as columnar crystals up to 0.5 mm long, sparsely distributed throughout the cementation layer, and replacing the other minerals. The minerals were identified by microscopic and X-ray examination.

Cemented beach deposits

Well-cemented beach deposits occur at the foot of the cliffs at Hill of Clibberswick at four localities at Y1 [HP 6652 1292]–[HP 6651 1233], and at Vord Hill, there are three localities at Y2 [HU 6290 9451]–[HU 6350 9440] (Figure 51). These were first noticed by an Anon (1926–27). Small occurrences of similar breccias occur elsewhere (Figure 51). The beach breccias are formed of unsorted angular fragments of serpentinite ranging in size from millimetres to metres thick. They occur only where the bedrock cliff face forms a near-horizontal ledge at about sea level and accumulates a ‘beach’ of rock fragments. The ledge is natural cliff face, not an erosional shelf (P534590). The cement is a layer of close-packed needle-like crystals about 0.3 mm long and 0.03 mm across, on the serpentinite surfaces. The crystals are pale brownish aragonite with the c-axes orientated approximately normal to the serpentinite surfaces in the form of crystal ‘plush’. Where the serpentinite fragments are sufficiently close, the layers of needle-like crystals meet and securely cement them together. In spaces between those fragments too large to be bridged by the layers of crystals, they enclose empty vug-like spaces.

Origin

The absence of springs on the cliff faces, other than from the screes, shows that the only water available to form cement is rain, often contaminated by salt spray. This has percolated through the scree, dissolving out Mg and Ca from the fine powdery component, and has redeposited it as the cement composed of brucite and hydromagnesite, with some aragonite, on the surfaces of scree fragments, especially where the solution entered empty spaces in the scree, causing it to lose CO₂ and/or water. According to Lippmann (1973, figs 33 and 50), at low CO₂ content, Mg would be deposited as stable brucite, while at higher CO₂ content, brucite and/or hydromagnesite would be deposited as metastable minerals. The high affinity of Mg for water would prevent the deposition of the stable mineral magnesite. Due to the relatively high concentration of Mg, CaCO₃ would be deposited in the metastable form, aragonite. The calcite found in the cementing layers grew later, evidently as a stable replacement mineral, possibly due to the action of low-Mg rainwater percolating through the already cemented scree. The beach deposit cement was formed at sea level within the scree. When the Mg-rich water, which had cemented the upper part of the scree, reached sea level, it enriched the sea water within the scree in Mg. This was especially the case where the sea-water was ponded by flat or concave-upwards bedrock surfaces and was thus protected from free circulation with the open sea. The enhancement of the Mg content of the sea water within the scree, possibly assisted by evaporation in times of drought and calm seas, was evidently sufficiently high to reduce the solubility of the Ca causing it to crystallise as the metastable mineral, aragonite. The precipitation of the stable mineral calcite was inhibited by the high content of Mg (Lippmann, 1973 p.113). The exclusive deposition of carbonate indicates a higher content of CO₂ than in the scree above. This may have been due to waves breaking over the scree at sea level and continually forcing air, containing CO₂, through it.

The beach cementation must have taken place recently. Sea level has been rising in Shetland since the end of the glaciation. It has risen at least 9 m in the last 5500 years (Hoppe, 1965, 1974), and the deposits are at about sea level. The deposit at Hagdale (Y4, (Figure 51) is not associated with a cliff or a scree, but occurs at the outlet of the stream draining the Hagdale chromite quarry (Plate 8f). For more than 200 years, many tens of thousand tons of chromite ore and serpentinite were mined and concentrated by grinding, and the water used passed over this beach. This water acted like the water percolating through the screes at Wick of Hagdale [HP 632 107], leading to cementation of the beach where it entered the sea.

Lithified till

In places, the till layer topping the cliffs has been hardened or lithified so that it projects slightly beyond underlying, hard but shattered, rocks and the overlying, but unlithified, till (Flinn and Pentecost, 1995). In thin section, the lithified till looks no different from the normal till nearby. It does not react with dilute acid. The hardening appears to occur generally where the cliff crosses a minor downfold in the land surface, sometimes near to a stream (P534889), e.g. at Geo of Litlaland [HU 653 887] and Houbie [HU 623 905]. Inland areas of strongly lithified till are exposed locally, especially near to streams. The occurrence near the School at Houbie [HU 628 915] is particularly well developed (P534676).

Chapter 10 Geophysics

Introduction

The published 1:250 000-scale regional Bouguer gravity anomaly map of Shetland (McQuillin and Brooks, 1967) and the 1:625 000-scale aeromagnetic map of north Britain (IGS, 1972) show clearly defined high amplitude gravity and magnetic anomalies associated with the outcrop of the Shetland Ophiolite-complex. A strong positive gravity anomaly overlies the metagabbro outcrop in Unst, while a major positive aeromagnetic anomaly overlies the serpentinite outcrop in Unst and Fetlar. By contrast, the metamorphic schist and gneiss basement rocks immediately west of the ophiolite-complex are characterised by a relatively subdued magnetic anomaly pattern. Onshore, these lithological units are readily accessible and well exposed enabling the geological structure to be studied in detail at outcrop, and also for the magnetic properties and densities of the lithologies to be directly measured. The Shetland Ophiolite-complex was thus deemed a suitable target for geophysical modelling to quantify the form and dimensions of the qualitative model inferred from the geological mapping (Figure 4).

Gravity and magnetic data

For the purpose of geophysical modelling, much more detailed gravity and magnetic maps were needed than those previously available. Gravity data coverage for Unst and Fetlar was therefore improved by the author by carrying out a new gravity survey (1997–1998) and incorporating the results with gravity data surveyed by Taylor (1984). (Figure 52) is a Bouguer gravity anomaly map based on 1195 gravity observations within the land area. These include 235 determinations for Unst and Fetlar taken from the BGS gravity database, used in the compilation of the 1:250 000 Bouguer gravity anomaly map for Shetland (IGS 1978b), and the earlier Bouguer anomaly map of McQuillin and Brooks (1967). Taylor (1984) provided 374 new determinations in Unst, and a further 586 from both islands were obtained for Flinn’s survey, using the same Lacoste and Romberg gravimeter as that used by Taylor. Several traverses within and just outside the area of the map were incorporated from BGS marine gravity survey data to provide additional coverage offshore.

The land-based data were computed using a variable Bouguer reduction density. The mean density of the rocks below each gravity station was determined from density measurements obtained from more than eleven hundred rock samples collected from Unst and Fetlar (Table 19). For the marine gravity data a reduction density of 2.75 Mg/m3 was used. Terrain corrections supplied by BGS were then applied to the land data using height information from a digital terrain model and using a reduction density of 2.75 Mg/m³. Due to the low topography of the islands, the corrections were only significant for several of the highest cliffs.

The aeromagnetic map (Figure 53) is based on the original 1:63 360-scale plotting sheet for Shetland used in the construction of the 1:250 000-scale map (IGS, 1968b, 1980). The airborne survey from which it was derived was flown at a mean terrain clearance of 305 m on east–west flight lines that were 2 km apart, and north–south tie lines that were 5 km apart. The relatively widely spaced flight lines produced artificial anomaly closures between flight lines when contoured by computer, so the final map has been contoured by hand to preserve anomaly values along flight lines and interpolate contour values between flight lines to reflect the geological trend more closely.

Geophysical properties of the rocks

More than 1100 rock samples were collected in the field in order to measure the density and magnetic properties of the main rock types prior to the geophysical modeling. (Table 19) is a summary table showing the mean density values obtained for the main rock types outlined in (Figure 55)." data-name="images/P937279.jpg">(Figure 54) (after Flinn, 2000).

The highest density igneous rock of significant volume is the metagabbro with a fairly uniform density of 3.1 Mg/m3. This is significantly higher than the underlying basement rocks and would account for the regional gravity high over the eastern part of Unst and offshore. By contrast the adjacent serpentinites have a much lower and more variable density of c.2.61–2.66 Mg/m³ and may be low enough to account for a local gravity low over the serpentinite rocks in the central part of Unst. A pronounced elongate north–south gravity low in the central part of Fetlar is also probably due to serpentinite. The other notable low density rock, within the Middle Imbricate Zone, is the Muness Phyllite which has the same density as the serpentinite. Basement rocks have a variable density with the highest values associated with metasedimentary rocks within the Valla Field Block (c. 2.81–2.94 Mg/m³), thus dense basement rocks may account for the regional gravity high in the northern part of this block.

All of the serpentinite rocks within the ophiolite- complex are highly magnetic with higher magnetic susceptility values where these have been antigorite serpentinised. In general the higher the susceptibility the greater the range. Assigning a bulk magnetic susceptibility value to these rocks for modeling purposes is therefore difficult. The metagabbros by contrast are only weakly magnetic. Basement rocks show a wide variety of magnetic susceptibilities with the schists and flags showing the highest values. Some pelitic rocks in the Saxa Vord Block contain distinct magnetic units as marked on the 1:50 000-scale Unst and Fetlar geological map (BGS, 2002).

Geophysical modeling on Unst

Seven traverses have been surveyed across the Shetland Ophiolite-complex on Unst and Fetlar for geophysical modelling purposes (Flinn, 2000). Three of the five surveyed on Unst (Figure 55)." data-name="images/P937279.jpg">(Figure 54) and (Figure 55) are presented here. They were chosen to cross the Lower Nappe of the Shetland Ophiolite-complex normal to its long axis and its layering, and approximately perpendicular to the contours of the regional field (Taylor, 1984). Within these constraints, the profiles were sited to cross the ophiolite body where the geological structure seemed simplest. Gravity stations were sited every few 100 m. The profiles were extrapolated over the sea to the east of Unst in an attempt to find the eastern limit of the ophiolite. The data for this extension had to be based on contours interpolated between the marine and terrestrial data, so for these areas the results are less reliable than for land areas. Magnetic values for the stations were obtained from the hand contoured aeromagnetic map.

Modelling was based on the geological interpretation of the Lower Nappe as a steeply layered thrust mass lying on a basement formed of rocks with an average density of 2.75 Mg/m³. The BGS Gravmag program (Pedley et al., 1993) was used to carry out 2.5 dimensional gravity and magnetic modelling using half-strike lengths deemed suitable for the different rock units. The Gravmag program was used to adjust the geological model so that the calculated gravity and magnetic fields matched the observed ones. This mostly required adjustment of the depth to the base of the ophiolite, the general form remaining largely unchanged.

The results of the modeling are shown in (Figure 55). The observed gravity and magnetic fields are shown as dashed lines with dots representing gravity stations. Solid lines are the calculated fields. The geological components on the models are represented as a series of polygons filled using the same ornament as in (Figure 55)." data-name="images/P937279.jpg">(Figure 54). The numbers displayed within each polygon are the rock physical properties used (derived mainly from (Table 19) and (Table 20) and refer to the density in Mg/m³ with the magnetic susceptibility (beneath) expressed in SI units.

Profile X

Profile X (Figure 55) crosses the Lower Nappe where the geology is at its simplest, and both the gravity and magnetic anomalies are near their maximum amplitudes and present their simplest and most symmetrical forms. The model shows the nappe structure as an ophiolite body comprising serpentinite and metagabbro with a near vertical contact between the two (based on field evidence). Geological indications are that the nappe rests directly on the basement or on imbricate slices of the basement rocks. This model, after adjustment to produce calculated gravity and magnetic anomalies matching those observed, has a simple downfolded base to the nappe. The ‘dip’ in the observed gravity profile at ‘A’ may be due to a continuation of the low density Muness Phyllite basin offshore, resting on the metagabbro (not modelled).

A magnetic susceptibility value of 0.13 for the serpentinite was required to produce a calculated magnetic anomaly that matched the positive magnetic anomaly of c. 2000nT.

Profile Y

Profile Y (Figure 55) also shows serpentinite and metagabbro separated by a steep contact. In this case the serpentinite is differentiated into serpentinised metaharzburgite and metadunite components and a unit of clinopyroxenite-werhlite with a slightly lower magnetic susceptibility value. The base of the ophiolite body again has a simple downfolded base.

This configuration has been achieved by thickening the serpentinite wedge and compensating for the resultant loss of mass by extending the metagabbro below the serpentinite. This solution enables the same magnetic susceptibility value for the serpentinite to be used as that for profile X. This produces a very good match between the calculated and observed magnetic profiles, without further adjustments to the model. The geometry of the model suggests that the basal part of the metagabbro is more like a tectonic slice or nappe rather than a fold limb of the Lower Nappe. The ‘dip’ in the observed gravity profile at ‘A’ is probably caused by the northward continuation of the Muness Phyllite beneath the sea, as in Profile X.

Profile Z

Profile Z (Figure 55) includes not only serpentinite and metagabbro in steep to vertical contact, but also the low density Muness Phyllite overlying the metagabbro. In computing a gravity model for this profile, either the depth to the base of the nappe beneath the phyllite basin, or the depth of the basin, has to be arbitrarily or geologically fixed. In Profile Z, the base of the nappe has been extended as a smooth slope from the west, beneath the phyllite basin. Based on this assumption, for the computed gravity anomaly to match that observed, the thickness of the Muness Phyllite below datum has to be about 300 m. The minor downfold in the west end of the basin, required by the observed gravity anomaly, is reflected in the mapped boundary of the basin on Uyea island [HU 603 997], to the south.

The metagabbro along the east coast of Unst has been intruded repeatedly by basic dykes which form the Quasi-sheeted Dyke-swarm. It is possible that the ophiolite nappe beneath the sea to the east of the coast, is a sheeted dyke-complex. The dykes have a density of

2.91 Mg/m³. The dotted line U shows the form of the base of the nappe if the rocks east of the fault in Profile Z had this density.

Discussion

Profile Z crosses the metagabbro where the amplitude of the gravity anomaly is much reduced compared with those in Profiles X and Y to the north. This is probably due to the absence of the basal metagabbro nappe proposed above. As a result, the mass of the high density metagabbro is insufficient to compensate for the adjacent body of low density serpentinite to the west. Therefore, a more substantial negative gravity anomaly is observed over the serpentinite than in the similar position in Profile Y. In Profiles X and Y the absence of a noticeable negative gravity anomaly over the serpentinite is due to the influence of the much greater mass of the adjacent metagabbro.

The validity of the estimates of the depth to the base of the Lower Nappe suggested by the geophysical modeling depends on the background density chosen for the underlying basement. The same density (2.75 Mg/m3) was used in all of the profiles, so, assuming it does not vary significantly over the area, the depths can be compared relatively. An increase in density for the basement of 0.1 Mg/m3 approximately doubles the depth of a metagabbro body resting on the basement and halves that of a serpentinite body in the models described. The 2.75 Mg/m³ estimate for basement density used in the models must be considered a minimum value. It has been used to conform to BGS work in the area. It is possible that the eastern boundary of the metagabbro lies close inshore to Unst. Here, the model best fitting the geophysical evidence has the metagabbro pinching out at depth in about the same distance offshore as the profiles show it pinching out at the sea floor (Profile X–dotted line V). There is even the possibility that the ophiolite has been overthrust from the east by metamorphic basement.

Geophysical modelling on Fetlar

Geophysical modelling on Fetlar is more problematic due to the wide variety of rocks with different physical properties that make up the Middle Imbricate Zone of the nappe structure. The rocks here form horizontal lenses and layers of variable thickness. In addition the gravity and magnetic anomalies are generally more subdued than those over the ophiolite-complex on Unst making it especially difficult to detect the eastern limit of the ophiolite.

The positive gravity anomaly over Unst extends south towards Fetlar and continues to decrease in amplitude towards the south. This is consistent with speculative modeling on Fetlar, which demonstrates that the ophiolite is much thinner in Fetlar than in Unst. The regional anomaly trend changes direction from north-east over Unst to south-east over the western part of Fetlar. This is thought to be the result of the truncation of the ophiolite by the Lamb Hoga Fault which introduces lower density basement rocks against the western flank of the anomaly, adjacent to the higher density gabbro to the east. The magnetic anomaly map shows a magnetic high over Fetlar coincident with the ophiolite body and the magnetic data suggests the ophiolite body continues offshore to the north as indicated by the magnetic boundaries marked on the 1: 50 000-scale geological map (BGS, 2002).

A prominent gravity and magnetic feature on Fetlar is the north–south-trending gravity low and coincident magnetic high over the Vord Klippe. This is formed of low density magnetic serpentinite which forms the core of a synform. The Vord Klippe has been modeled as a synform extending to a depth of about 1 km.

Ground-magnetic results

Detailed ground-magnetic surveying in the northern part of Unst in the general area of Norwick, including the Saxa Vord Block, and the northern end of the ophiolite-complex was carried out during the geological survey of the area (Flinn, 2000). The boundaries of all serpentinite bodies were surveyed and also various basement rocks within the Saxa Vord Block. Ground- magnetic surveying was undertaken with the aid of a fluxgate Scintrex magnetometer with analogue output. The boundaries of magnetic rocks identified from the ground survey are shown on the 1:50 000-scale geological map (BGS, 2002).

The survey revealed that the serpentinite contacts are accompanied by extremely steep gradients with changes of several thousand nanoteslas (nT) occurring over a distance of a few metres. The western margins of the main serpentinite bodies typically show a narrow negative anomaly with an amplitude of more than one thousand nT. This suggests that the western margins of the serpentinite bodies have a steep easterly dip assuming the magnetisation is induced by the Earth’s present day magnetic field orientation. Over the serpentinite bodies themselves, high-amplitude short- wavelength positive anomalies occur. These demonstrate that the magnetic properties of the serpentinite are highly variable (see (Table 20)) probably due to the irregular distribution of magnetite within the fabric of the serpentinite. A pronounced broad linear positive aeromagnetic anomaly overlying the Saxa Vord Block has been resolved on the ground as a series of continuous narrow north–south positive anomalies (a few metres wide), and more substantial lenticular anomalies associated with magnetic phyllitic pelite units within the Saxa Vord Pelite.

Chapter 11 Platinum-group elements

H M Prichard

Introduction

There are six platinum-group elements (PGE) and in mafic and ultramafic complexes they are commonly concentrated by magmatic processes into specific igneous lithologies. Osmium (Os), iridium (Ir) and ruthenium (Ru) are usually concentrated with chromitites and platinum (Pt), palladium (Pd) and rhodium (Rh) with base metal sulphides. Crystallisation of chromite from a magma is often associated with crystallisation of sulphides and where sulphides occur with chromitites then all six PGE may be present. This is commonly the situation in large layered complexes that are the major hosts for Pt, Pd and Rh such as the Bushveld (Lee, 1996).

Traditionally, ophiolite-complexes have not been considered to have significant Pt, Pd or Rh concentrations but have long been known to contain Os, Ir and Ru, often hosted in discrete platinum-group minerals (PGM), such as laurite [Ru(Os, Ir)S2], in podiform chromitites.

The Shetland Ophiolite-complex was one of the first ophiolite-complexes in which Pt and Pd concentrations greater than 1 ppm were identified. Research on this ophiolite showed that PGE-rich magmas that form ophiolite-complexes concentrate PGE by similar magmatic processes to those operating in layered complexes. Since their discovery in Shetland, many Pt and Pd occurrences have been described from ophiolite-complexes elsewhere in the world.

History of research

PGE were first were described as occurring in Unst by Hitchin (1929), in chromite concentrates from the crushing mill [HP 638 103] near Baltasound. Much more recently, Os, Ir and Ru PGM have been identified in chromitite (Prichard et al., 1981) and further studies have revealed the presence of extremely anomalous PGE- and Au-rich lithologies at Cliff [HP 608 112] (Neary et al., 1984; Prichard et al., 1984, 1986; Gunn et al., 1985). Subsequently Pt and Pd concentrations were found in other ultramafic rocks and in soils (Gunn, 1989; Prichard and Lord, 1993; Lord et al., 1994; Lord and Prichard, 1998).

Distribution of PGE concentrations within the ophiolite

In the Shetland Ophiolite, Os, Ir and Ru are concentrated within chromitites and Pt, Pd and Rh are associated with base metal sulphide-bearing ultramafic lithologies, usually in close proximity to ‘podiform’ chromitite. Thus, the distribution of the PGE broadly corresponds to the distribution of the chromitite occurrences (Figure 56). PGE values occur at a number of ‘stratigraphical’ levels within the ultramafic lithologies of the ophiolite (Figure 57). The most anomalous PGE grades occur at Cliff [HP 608 112], but highly anomalous values also occur at Harold’s Grave [HP 630 113]. At both of these localities, chromitite lenses are enclosed by metadunite situated within mantle metaharzburgite. Higher in the ophiolite sequence, there is a zone of anomalously enriched PGE concentrations in the vicinity of the disused chromite quarries north of Baltasound in crustal metadunite and more minor occurrences are present in chromitites in mid Unst. Stratigraphically above this, there are lower concentrations of Pt and Pd in base metal sulphide-bearing clinopyroxenite-wehrlite bodies in the upper part of the metadunite. Low values of Pt and Pd occur in ultramafic wehrlite lenses in the metagabbro, which mark the highest occurrences of anomalous Pt and Pd concentrations in the ophiolite. The lack of high values of PGE in south Unst and Fetlar corresponds to a lack of podiform chromitite in these areas. The following paragraphs describe the PGE concentrations in more detail, starting at the lowest most enriched Pt and Pd values at Cliff, and working upwards through the ophiolite.

Metadunite lenses in metaharzburgite; Cliff and Harold’s Grave

Metadunite lenses occur within the metaharzburgite, which occurs below the petrological Moho, marked by the junction between the residual mantle metaharzburgite and the overlying lower crustal metadunite. These are variably enriched in chromitite, base metal sulphides and the PGE. Base metal sulphide- poor chromitite in one of these lenses, at Harold’s Grave [HP 630 113], contains high Os, Ir and Ru with Os + Ir + Ru values of 1000 to 4000 ppb. Another chromitite-rich metadunite lens in metaharzburgite, at Cliff [HP 608 112], 300 m from the inferred basal obduction thrust of the ophiolite-complex, hosts the extremely anomalous Pt + Pd values of 60 000 ppb (Prichard et al., 1986). The Cliff locality consists of five small pits formerly containing en échelon chromitite lenses over a distance of 100 m by 20 m (Figure 58). These lenses are surrounded by thin envelopes of metadunite, which are variably sulphide bearing. One sulphide-bearing metadunite from the Cliff locality contains a Pt + Pd value of 4000 ppb. The only other PGE concentration in this area is in a small chromitite closer to the basal obduction thrust. The surrounding metaharzburgites have very low PGE values. Typical values of PGE from the different lithological units are given in (Table 20). All these values are from samples that were in situ. There are much higher values of Pt and Pd recorded in samples from the spoil tips at Cliff, with values of Pt and Pd exceeding 100 ppm and the best intercept recorded by drilling reported 5.82 ppm Pt and 16.99 ppm Pd over 1.83 m representing a true thickness of 0.5 m (Gunn and Styles, 2002).

Metadunite

The metadunite lies above the petrological Moho, within the basal part of the lower crustal ultramafic rocks. Base metal sulphide-bearing metadunite, located in close proximity to the ‘podiform’ chromite, has Pt + Pd values of 1000 to 3000 ppb and Pt/Pd ratios of 1:2. The metadunite hosts the longest (2.5 km) and broadest (0.5 km) zone of PGE enrichment in Unst. This is situated in the area north of Balta Sound, extending from Hagdale Wick [HP 644 105] and The Nev [HP 649 111] in the east, to Nikka Vord [HP 623 113] in the west and includes abundant pods and discontinuous layers of chromitite, many of which have been extracted leaving numerous small, disused quarries. The most anomalous Pt and Pd values in the metadunite occur in the sulphide-bearing metadunites adjacent to chromitite layers, in close proximity to the disused chromite quarries; especially those situated within the metadunite rather than those located at the boundary between the metadunite and metaharzburgite to the north. Maximum Pt and Pd concentrations occur in a cluster of quarries south-east of Nikka Vord (Figure 56).

Collection of basal soil samples at the soil–bedrock interface extended the zone of PGE and base metal sulphide mineralisation from areas around chromite quarries south-east of Nikka Vord eastwards into a broad envelope 2.5 X 0.5 km in size (Lord and Prichard, 1998). In particular, Pt + Pd anomalies up to 540 ppb were located in soils along a north-north-west–south-south-east traverse south of Muckle Heog [HP 631 109], in an area where chromite quarries are rare and outcrop is poor.

Subsequent diamond drilling at 45 degrees to the horizontal in a southerly direction at three sites in dunite above the Moho south of Nikkavord [HP 632 103], [HP 628 103] and [HP 629 102], provided evidence of downdip continuation of PGE mineralisation located in the soil anomalies. High Pt and Pd values, of up to 200 ppb Pt and 300 ppb Pd are also present in other chromitite lenses within metadunite farther south [HP 596 037] in mid Unst

Clinopyroxenite-wehrlite bodies

Clinopyroxenite-wehrlite bodies occur within the metadunite, especially its upper part. They lie in a zone extending westwards from both Swinna Ness [HP 650 090], north of Balta Sound, and Skeo Taing [HP 644 086] (west of Ordale), south of Balta Sound, and then south-south- west in a zone through mid Unst to approximately 1 km north of Uyeasound [HP 590 020]. South of Baltasound airport, a base metal sulphide-bearing pegmatitic clinopyroxenite [HP 624 068] gave Pt and Pd values of 510 and 590 ppb, and at Helliers Water [HP 607 053] a Ru value of 173 ppb is associated with a thin Fe-rich chromitite lens within clinopyroxenite. Pt and Pd concentrations also occur in high-level bodies of clinopyroxenite- wehrlite within the metagabbro, which lies above the metadunite. These occur on the east coast of Unst, south-west of Nuda [HP 617 038], where values reach 48 ppb Pt and 18 ppb Pd, and at Sandwick [HP 613 022], with maximum values of 120 ppb Pt and 190 ppb Pd.

PGE and associated elements

Gunn et al. (1985), Gunn (1989) and Flight et al. (1994) concluded that there are few good pathfinders elements for the PGE in Unst with the Ni/MgO ratio being the best and some PGE are located with As, Sb and Te. Prichard and Lord (1993) noted that Pt and Pd concentrations are especially concentrated with chromitites that are either base metal sulphide- bearing or are enclosed in dunites that are base metal sulphide-bearing and give rise to Ni and Cu anomalies in Pt- and Pd-rich areas. The highest values are in the area north of Baltasound with Ni up to 8500 ppm and Cu up to 1500 ppm (Lord et al. 1994). Gold is also slightly elevated in Pt- and Pd-rich areas, for example values of between 20–65 ppb occur in the Pt- and Pd-rich lithologies north of Baltasound (Lord and Prichard, 1998). High Ni, Cu and Cr values are associated with specific primary magmatic lithologies in the ophiolite where sulphur saturation of the magma occurred, especially in close proximity to the ‘podiform’ chromitites (see Chapter 6). Indeed, at the Cliff locality the only PGE values recorded away from the extremely anomalous lenses of chromitite in metadunite surrounded by metaharzburgite is one very small lens of chromitite near the basal obduction thrust (Figure 58). In contrast to the distribution of the Pt, Pd, Au, Cr and base metals that all show a magmatic distribution, the distribution of As in the ophiolite in Unst follows the alteration zones cross-cutting the igneous stratigraphy with the greatest concentrations of up to one percent occurring along the basal obduction thrust. Arsenic is also concentrated along internal faults associated with ophiolite emplacement.

PGE chondrite-normalised patterns

If the PGE concentrations are normalised to chondrite values then the relative enrichment of the elements can be observed ((Table 20), (Table 20). Chondrite values used to plot the chondrite-normalised patterns are Os 514 ppb, Ir 540 ppb, Ru 690 ppb, Rh 200 ppb, Pt 1025 ppb and Pd 545 ppb after Naldrett and Duke (1980)." data-name="images/P937284.jpg">(Figure 59). Base metal sulphide- poor chromitites in the Shetland Ophiolite-complex are relatively enriched in Os, Ir and Ru, and these give negative slope chondrite-normalised patterns. Although the total concentration of the PGE varies considerably from chromitite to chromitite, the relative proportions of the PGE remains fairly constant, as shown by a comparison of the PGE analyses of sample RL070 from the Os-, Ir- and Ru-enriched Harold’s Grave locality and sample RLM003 from a chromitite from the area north of Baltasound. Pt- and Pd-rich concentrations occur in base metal sulphide-bearing ultramafic lithologies including metadunites often associated with chromitites and clinopyroxenite-wehrlite bodies and these give positive slope chondrite-normalised patterns, as in PGE analyses of samples RLM003 and MR11 from Cliff. Analyses of samples from the higher level clinopyroxenite-wehrlite bodies, within the metagabbros, including NA25 and MR169, give positive slope chondrite-normalised patterns with lower chondrite- normalised values. The extremely enriched chromite- rich metadunite from Cliff with the most elevated chondrite-normalised values has a positive slope pattern with all six PGE elevated but Pt and Pd more so than the other PGE.

Platinum-group minerals

The largest PGM located is from Harold’s Grave. It is a laurite with inclusions of irarsite and native osmium and this has a maximum diameter of approximately 300 microns (Plate 10a) and (Plate 10b). A great variety of PGM have been described from Unst (Prichard et al., 1981; Tarkian and Prichard, 1987; Prichard and Tarkian, 1988; Prichard et al., 1994). The PGM assemblage in the rich PGE lithologies at Cliff [HP 608 112] includes: sperrylite PtAs₂; stibiopalladinite Pd₅Sb₂; hongshiite PtCu; geversite PtSb₂; genkinite (PtPd)₄Sb₃; Pt and Pd tellurides; laurite (Ru[Ir, Os]S₂); irarsite IrAsS; hollingworthite RhAsS; rare potarite (PdHg); AuPd alloys; and Pt and Pd ochres (Table 21). Gunn and Styles (2002) stated that the most common PGM in Unst are arsenides and antimonides with minor tellurides that occur in serpentine, chlorite and carbonate minerals in close association with a range of Ni, Fe and Co sulphides, arsenides and antimonides.

In the area north of Baltasound, around the disused chromite quarries, the PGM are often associated with base metal sulphides which are predominantly Ni- and Cu-rich, including pentlandite, heazlewoodite, millerite, chalcopyrite, chalcocite and bornite (Prichard et al., 1994). These sulphides are visible in hand specimen, forming 1–2 per cent of the rock (Lord et al., 1994).

Primary PGM are rare because the ultrabasic part of the ophiolite is extensively serpentinised but they are sometimes preserved in unaltered primary minerals especially chromite where Os-Ir-bearing laurite (RuS₂) is quite common (Plate 10d). In one case, in a wehrlite from south Unst [HP 598 020], hundreds of tiny (1–5 microns) of PGM occur in stringers that cross cut a pyroxene crystal. Pt- and Pd-base metal sulphides occur within primary clinopyroxene and where alteration has occurred in cleavage planes and on the altered edges of the clinopyroxene the PGM are arsenides antimonides and tellurides (Prichard et al., 1994) (Plate 10e) and (Plate 10f). Thus the primary Ru-, Os-, Ir-, Pt- and Pd-bearing PGM appear all to be sulphides.

PGM enclosed by altered silicates are usually As-, Sb- and Te-bearing. These include irarsite, hollingworthite, sperrylite, and stibiopalladinite. These PGM have been altered during serpentinisation that was accompanied by the introduction of As, Sb and Te. Pd and Pt also occur in solid solution in breithauptite and Pd may also be present in Ni-Cu antimonides. Laurite enclosed in chromite is usually euhedral in shape and contains Ir and Os in solid solution. However, where the laurite occurs on the edge of chromite, in contact with altered silicate, or is crossed by fractures in the chromite that are serpentine filled, then the laurite is often irregular in shape. Serpentinisation has caused recrystallisation of the laurite to form pure RuS₂, expelling the Ir and Os that now form irarsite and native osmium, which often accompany the laurite. Ru also occurs in minor amounts in pentlandite (Tarkian and Prichard, 1987).

Where alteration is intense, PGM lose their As, Sb and Te producing PGE-alloys with base metals and gold including hongshiite (PtCu), Pt-Pd-Au-Cu and Pd-Au. Weathered samples contain PGE-oxides. These have been observed at Cliff where sperrylite is commonly altered along the edges to Pt-oxide. Further analysis of precious metals in the soils covering the ophiolite complex was described by Flight et al. (1994).

Discussion

Initially, on the discovery of the extremely anomalous Pt and Pd values in Unst, the origin of the mafic and ultramafic rocks was questioned, as such high values had not previously been recorded in an ophiolite- complex. However the mafic and ultramafic complex on Unst and Fetlar has many of the characteristics of a dismembered ophiolite, including ‘podiform’ chromitite and dykes at the highest level in the gabbro, which Prichard (1985) interpreted as representing the base of a sheeted dyke-swarm. In major PGE deposits, such as the Bushveld complex in South Africa and the Noril’sk deposit in northern Siberia, Pt and Pd precipitated from the magma with base metal sulphides such as Ni and Cu, and PGE are often associated with chromitites, as in the UG2 in the Bushveld (Lee, 1996). In Unst, the PGE show a similar magmatic association with base metal sulphides and chromitites. Subsequent local remobilisation of the PGE and Ni and Cu has disturbed what may once have been good magmatic correlations between these elements, but in the area north of Baltasound the chromitites are often characterised by the presence of minor (approximately modal 1–2 per cent) concentrations of base metal sulphides, and in these cases there are associated concentrations of Pt and Pd. The PGE are concentrated within magmatic cycles marked at the base by a discontinuous layer of chromitite containing concentrations of Os, Ir and Ru, overlain by metadunite containing Pt, Pd and base metal sulphides and grading up into barren metadunite. The variation in δ34S of -1 to +4 over 1.5 m of mineralised core material from this area was explained by magmatic fractional crystallisation (Maynard et al., 1997). Os-isotope analyses from Cliff, Harold’s Grave and chromite north of Baltasound were compared with those of other ophiolites by Walker et al. (2002). In chromite-rich lithologies Os-, Ir- and Ru-bearing PGM occur within chromite grains, and in the interstitial serpentinised silicate matrix with chlorite they are joined by Pt-, Pd- and Rh-bearing PGM. These textural associations indicate an in-situ magmatic fractionation of PGE, from early Os-, Ir- and Ru-bearing PGM occurring within the centres of relict primary chromite to later Pt-, Pd- and Rh-bearing PGM formed after the crystallisation of chromite (Figure 60). The PGM sulphides in the chromite and clinopyroxene probably most closely reflect the primary igneous mineralogy, with the PGE being strongly associated with sulphur in the magma. Introduction of As, Sb and Te during ophiolite emplacement and regional greenschist-facies metamorphism has altered the PGM in serpentinised lithologies to give As-, Sb- and Te-bearing PGM. Further alteration has produced PGM alloys and oxides.

In the Cliff area, elevated PGE, Ni, Cu and Au only occur in the chromitite-bearing metadunite lenses in a similar association to the other magmatic PGE concentrations in this ophiolite. One base metal sulphide-bearing metadunite from the Cliff locality, which contains a Pt plus Pd value of 4000 ppb, has an average δ34S value of +4.2 that is indistinguishable from sulphur isotope values from other metadunite pods. Therefore, by comparison, the origin of the base metal sulphide in the metadunite at Cliff is also magmatic (Maynard et al., 1997). The extremely PGE-rich lithology at Cliff consists of a disseminated chromite containing 10 to 60 per cent chromite surrounded by serpentinised olivine. Arsenic and antimony are enriched in this lithology which often has a greenish colour caused by the presence of nickel carbonate. The presence of arsenic in the extremely enriched PGE lithology may indicate that reconcentration of igneous PGE occurred during the introduction of arsenic along the nearby basal obduction thrust and along internal faults during, or after, emplacement of the ophiolite. It has been suggested (Lord et al., 1994) that these extremely high PGE values at Cliff were formed by local reconcentration of the lower grade magmatic PGE concentrations in the chromite- bearing base metal sulphide-bearing metadunite lenses, which are similar to those in chromitites in base metal sulphide-bearing metadunites elsewhere in the ophiolite.

The lack of correlation of PGE with magmatic pathfinder elements such as Ni, Cu and Cr is probably due to local remobilisation of these elements at different rates. Veins of native copper are common in drill core from the area north of Baltasound. However, there is an association of these magmatic elements with the PGE, as they are all elevated in primary igneous lithologies at the base of magmatic cycles in the crustal metadunite and in chromite pods enclosed by metadunite in the mantle metaharzburgite. A hydrothermal origin for the PGE concentrations is a model favoured by Gunn et al. (1985) and Gunn (1989) who, for example, suggested that there has been extensive hydrothermal activity in the Cliff area. In layered complexes, such as the Bushveld in South Africa, base metal sulphide-poor chromitites have negative slope chondrite-normalised patterns and base metal sulphide-bearing mafic and ultramafic rocks have positive slope chondrite-normalised patterns (Naldrett et al., 2009). Thus, as in layered complexes, the chondrite- normalised patterns in Unst reflect similar processes of magmatic PGE concentration; negative slope chondrite- normalised patterns occur in base metal sulphide-poor ‘podiform’ chromitites and positive slope chondrite- normalised patterns occur in base metal sulphide- and Pt and Pd-bearing ultramafic lithologies, which are sometimes also chromite rich. These positive and negative slope chondrite-normalised patterns in samples from Unst do have kinks that probably reflect some differential mobility of individual PGE alteration. For example, steep gradient positive slopes between Pt and Pd are typical of unaltered lithologies and these samples have magmatic Pt/Pd ratios of approximately 0.5, whereas low gradient positive slopes between Pt and Pd represent altered lithologies with a Pt/Pd ratio of 1 or more, indicating preferential removal of Pd. At Cliff, the most anomalous concentration in the base metal sulphide-poor, arsenic- rich chromite-bearing metadunite gives a positive slope pattern but with all the PGE elevated above chondrite values reflecting the extreme hydrothermal secondary upgrading of the magmatic concentration.

Variation of PGE concentrations in ophiolite-complexes

Oceanic crust, which is subsequently emplaced onto continental crust to form an ophiolite-complex, varies in composition because it is formed from magma produced by different degrees of mantle melting and this is dependant on the tectonic setting in which it was formed. This variability in ophiolites suggests that they may not all be Pt- and Pd-enriched. Only those ophiolites produced by a high degree of mantle melting, sufficient to extract PGE from the mantle, will have PGE-enriched crustal sequences (Prichard et al., 1996).

However, if there is too much mantle melting the PGE will be diluted in the melt and then PGE are not enriched in the chromite-bearing ultramafic sequence but crystallise with base metal sulphides in the gabbros (Prichard et al., 2008). In the Shetland Ophiolite-complex, the degree of melting was high and just sufficient to extract the PGE from the mantle without dilution by further melting of PGE-poor mantle.

Once formed, a PGE-rich magma is able to crystallise PGE-enriched lithologies resulting in magmatic enrichments in the ultramafic crustal magmatic part of the sequence. This, combined with other evidence, led Prichard and Lord (1988) to speculate on the origin of the Shetland Ophiolite, proposing a suprasubduction zone where a sufficiently high degree of mantle melting formed a PGE-rich boninitic magma.

Summary

The Shetland Ophiolite-complex in Unst contains extremely enriched Pt and Pd values at the Cliff locality and also many occurrences of lower grades of PGE elsewhere in the ultramafic sequence. At the time of the discovery of these unusually anomalous Pt and Pd values at Cliff, in the early 1980s, ophiolite-complexes were not thought to be significantly enriched in Pt and Pd. Now, values of a few ppm Pt and Pd and Pt- and Pd-bearing minerals are commonly recorded in ophiolite-complexes (for example in New Caledonia (Augé et al., 1998); the Zambales Ophiolite-complex in the Philippines (Bacuta et al., 1990); Thetford mines in Quebec (Corrivaux and La Flamme, 1990); the Lewis Hills in Newfoundland (Edwards, 1990); Bulqiza in Albania (Ohnenstetter et al., 1999); Leka in Norway (Pedersen et al., 1993); Troodos in Cyprus (Prichard and Lord, 1990); and Pindos in Greece (Tarkian et al., 1996).

The Shetland Ophiolite-complex, however, remains one of the most PGE-enriched ophiolite-complexes, probably due to the particular degree of partial melting of the mantle that occurred during its formation being optimal for the extraction of the PGE from the mantle (O’Hara et al., 2001a and b). The ophiolite contains concentrations of all 6 PGE produced by magmatic processes, as demonstrated by their close association with specific primary igneous lithologies. Non-base metal sulphide-bearing chromitites, such as at Harold’s Grave, are relatively enriched in Os, Ir and Ru giving negative chondrite-normalised patterns typical of PGE concentrations that are traditionally associated with ophiolite-complexes. Base metal sulphide-bearing chromitites, such as occur in the area north of Baltasound, and other base metal sulphide-bearing ultramafic lithologies, such as clinopyroxenites, have Pt and Pd concentrations equal to, or greater than, Os, Ir and Ru concentrations, giving chondrite-normalised positive slope patterns, traditionally thought to be more characteristic of large continental layered complexes. The collection of Pt and Pd by immiscible base metal-rich nickel- and copper-bearing sulphides in the magma, as documented in layered complexes, appears also to be the process that concentrated them in the oceanic magma that crystallised to form this ophiolite-complex. Thus, the process is similar in both types of complex but in ophiolites, although the grades may be similar to those in layered complexes, the tonnages are small because of the dynamic nature of ocean crust formation and the resulting discontinuous characteristics of ‘podiform’ chromitite.

The extremely anomalous PGE values of 10–100 ppm that occur at Cliff are among the highest in the world, not only from ophiolites, but also layered complexes. These have probably been formed by a two stage process; magmatic PGE concentration in the Cu-, Ni-, Au-, sulphide-bearing metadunites and chromitites enclosed in the mantle metaharzburgite, followed by a very local, late, possibly emplacement related, hydrothermal reconcentration of PGE that also involved the introduction of As and Sb.

The PGE in the Shetland Ophiolite-complex have been extensively studied and their distribution, mineralogy and genesis are recorded in many papers. Although the grades of the PGE are locally very high, the tonnages discovered by the exploration programmes to date are small. The characteristics of this PGE occurrence have provided an excellent opportunity to study the processes that concentrate, reconcentrate and disperse the PGE.

Chapter 12 Economic geology

Minerals

Chromite (see also Chapter 6)

Hibbert (1822) discovered blocks of chromitite (then only available in the USA) in the drift on Unst north of Balta Sound and west of Hagdale in 1817. The following year, Thomas Edmondston found an in-situ vein at Buness [HP 630 091]. In 1823, the great Hagdale deposit [HP 639 103] was found, and at about the same time the Nikka Vord north deposit [HP 621 108] was discovered too. Sale of chromite started in 1824, when some 80 tons of glacial drift and quarried ore were sold (Edmondston, 1839). The industry prospered for two or more decades, but chromite had been discovered in Norway in 1830. When the great Hagdale pit closed down in 1862, most of the high-grade ore was exhausted and the prosperous days were over. The Hagdale deposit provided more than) half the total chromite produced in Unst. The quarry had reached a depth of about 35 m and was based on a vein about 3.5 m wide. After almost ceasing, production in Unst revived about 1870. A new manager, John Walker, pumped out and worked several abandoned pits in the Nikka Vord [HP 628 103] (Plate 11) and Quoys [HP 616 100] areas (Gear, 2005) and set up a beneficiation plant. He sold about 3000 tons of ore before he was declared bankrupt in 1877, because of his copper mine at Sandwick in south Shetland. Chromite mining in Unst ended then, basically because of the exhaustion of high- grade ore. In the period 1820–1877 some 40 000 tons of ore was produced from five major quarries and many smaller ones, of which 32 000 tons came from Hagdale. The ore was used for chemical purposes, in particular the production of pigments (Sandison, 1936, 1948).

The Sandison family bought the mines in about 1900 and small quantities of ore were produced in 1908–9 and 1913–4. From 1916 until 1927 production resumed because of the war and because a use had been found for low-grade ore as a refractory in blast furnaces. Old quarries were drained and worked and spoil heaps at Hagdale were reworked. From 1920 to 1927, a concentration plant was installed to upgrade the ore, but was unable to raise the grade much above 36 per cent Cr2O3. A total of 6000 tons of ore was produced from several quarries and from reworking the spoil heaps. This included 1650 tons from the concentration plant (Sandison, 1936, 1948).

A final period of mining took place between 1937 and 1944, helped again by wartime conditions. It was discovered that chromite ore as low as 20 per cent Cr203 combined with local dunite-serpentinite made excellent refractory bricks for use as open furnace and arc furnace linings. By 1942 they were industry standard (Lynam et al., 1942). During this period, 4000 tons of ore with less than 30 per cent Cr203 were mined and sold. 3000 tons came from the Midgarth quarry [HP 636 095] which had not previously been worked (Sandison, 1948).

In all, about 50 000 tons of ore were produced in Unst, of which 32 000 tons came from Hagdale. The chromite mining ended because of a failure to find new deposits and because of the low grade of Shetland ore.

Gold (see also Chapter 11)

Buchanan and Dunton (1996) and others have reported minute particles of gold in both stream and beach sediments (Figure 61), and gold was also detected, in concentrations below 1 ppm, by analysis of stream sediments and overburden samples. It has also been found in the Muness Phyllite, in amounts up to 1 ppm, by analysis of the stratabound Ramnageo Ness sulphide zone [HU 627 995] ((P536034), (P536032)). However, X-ray backscatter examination of polished sections from this occurrence have showed no trace of gold (personal communication D Plant, University of Manchester). Moreover, the stream-sediment gold on the ophiolite in south Unst is accompanied by ice-transported garnet, and that on the Muness Phyllite by garnet and chromite where these minerals do not occur in the underlying rocks. Therefore, although the gold may have a local source in the bedrock, its ice transport from further afield is also possible.

Kyanite

Segregations of nearly pure kyanite, 10–20 cm long, occur in the Valla Field Schist in the area of Tonga [HP 587 140] and to the south and south-east (P534833). In this area the schist is at its coarsest and is often rich in fresh kyanite grains, up to 2 cm long. About 1970 United Steel commissioned a drilling programme to test the economic prospects for the mineral (unpublished records, BGS, Edinburgh).

Drilling was carried out close to Cat Houll gorge, west of Tonga [HP 587 147], at Ward of Petester [HP 587 119] and at White Stanes, close to Balliasta [HP 593 097]. These sites are all within the outcrop of the Valla Field Schist on the 1:50 000-scale geological map (BGS, 2002). At each site, two holes were drilled 120–240 m apart, normal to the schistosity (i.e. at about 45º towards azimuth 250º) to depths of 60–90 m. Considerable amounts of kyanite were encountered at all three sites, but at Petester and White Stanes the stratigraphically lower of the pairs of cores appear to contain less kyanite than the upper. The drilling results were considered satisfactory but during separation tests too much kyanite was lost due to the kyanite grain surfaces being altered to muscovite (shimmerised).

Diatomite

The occurrence of diatomite in Unst and Fetlar is described in Chapter 9.

Kaolinite

There are several occurrences of kaolinite in the Saxa Vord Block, but the most prominent occurrence is at Moo Wick in the Burra Firth Formation in Lamb Hoga [HU 622 877] (P534846). Hibbert (1822) arranged for Dr Henry (of Henry’s Law fame) to have this ‘porcelain clay’ tested by Josiah Wedgwood. Heddle (1878) also tested a sample and published an analysis. The locality was examined more recently by May and Phemister (1968), who reported that the kaolinite is of a ‘fairly well ordered type’, and concluded that it resulted from the action of meteoric water on the local quartzofeldspathic gneiss.

Four small disused kaolinite quarries occur on the north side of Ward of Norwick, beside peat roads at [HP 6472 1583], [HP 6495 1573], [HP 6462 1589] and [HP 6491 1603].

The occurrences are associated with quartzite. Kaolinised semipelitic rocks occur at the head of Burra Firth, where they are interbanded with the quartzites of the Queyhouse Flags and are exposed in the cliffs at the back of the beach [HP 615 140]. Kaolinite associated with quartzite is also exposed north of the Black Loch [HP 598 065]. This is probably a tectonic slice of Queyhouse Flags within the Lower Imbricate Zone.

Steatite

Steatite, a talc-magnesite rock, is known as ‘kleber’ in Shetland, where it has a long history of exploitation. From Neolithic to Iron age times it was used to a small extent, chiefly for tempering pottery and occasionally carved for vessels. With the arrival of Norse settlers about the beginning of the ninth century the making of pottery was largely abandoned in favour of the use of steatite, as was prevalent in Norway at that time. It was used for vessels of a range of types and sizes, lamps, baking plates, loom weights, fishing-line sinkers, drill whorls, spindle whorls, trinkets and toys and for moulds for metal working. The main sites in Unst and Fetlar for extracting and working steatite are in the cliffs. They are prominent due to the preservation of the working face on the cliff (P534709). They include Clibberswick [HP 652 120], Clammel Knowes [HP 584 065], Gorsendi Geo [HP 565 044], Houllnan Ness [HP 570 060], Hesta Ness [HU 663 927] and west of Houbie [HU 620 905] and [HU 616 903] (Figure 61). Many other sites were probably used on a smaller scale. By late Norse times pottery was again coming into use (Buttler, 1984, 1989).

Modern exploitation of the steatite started in 1914 at a site on the west side of Hesta Ness, Fetlar [HU 661 924] (Figure 61), where several east–west dykes have been steatitised. Production ended in 1923 due to the high cost of transportation. A total of 400 tons was produced between 1920 and 1925 including 50 tons in 1922 (unpublished records BGS, Edinburgh and Wartime Pamphlet No. 8).

In 1937, enquiries were made into the possibility of exploiting Unst talc, but it was not until 1940 that wartime needs enabled quarrying to start at Quoys [HP 613 124] (Figure 61). However, a labour shortage caused quarrying to stop after several months. Quarrying recommenced in 1945 and since then steatite has been exported in bulk to the south to be processed in Bristol and later Stockton-on-Tees. The Quoys quarry was worked continuously until 1989 when the quarry walls were considered too high to be safe (P536031). During this time some 415 000 tons of steatite were shipped south. Production rose from 1000 tons per year in 1946 to 10 000 tons in 1971. From the early seventies to the early eighties production was about 20 000 tons, but then declined considerably with the market. Production from the Cross Geo quarry at Clibberswick [HP 652 122] (Figure 61), next to a Norse site, started in 1984 and continues to the present time (P533617). Production fell again in the early 1980s due to closure of UK fertiliser plants. Since 1993 production has been steady at about 5000 tons a year. Total production to the end of 2000 was 102 805 tons.

Shetland steatite quarries work the lowest grade and are the smallest of any steatite quarries in the world. The steatite (Table 22) has been used mostly as a filler, as a dusting and anti-caking agent in the manufacture of roofing felt, fertilisers and pesticides and for making saggars for holding pottery in the furnaces. (Unpublished records, Alex. Sandison and Sons; personal communication S Owers).

Limestone

Both Unst and Fetlar have substantial quantities of limestone in the form of marble. In Fetlar, the Sunkir Marble [HU 623 874] occurs in the remote extreme south of Lamb Hoga close to the kaolinite deposit at Moo Wick (Figure 61). However, calcareous rocks, including minor amounts of marble, occur at Brough Lodge at the western end of the road [HU 578 925]. Unst contains a considerable amount of limestone in the Loch of Cliff Limestone (P533155) and the Westing Limestone. The latter is distributed widely in the Valla Field Block, much of it near to roads and easily accessible. There is an abandoned lime kiln on the main outcrop of the Loch of Cliff Limestone near Cliff [HP 6026 1121] (Figure 61), and a quarry in the limestone at Alma [HP 6015 0888] (Figure 61). These were both in use in the 1840s (Ingram and Ingram, 1845). In Fetlar, the limestone at Moowick [HU 623 876] has been burnt for lime (Anon in Phemister, 1929–63).

Peat

Peat has been used from time immemorial until recently by the people of Shetland as the main source of fuel. However, the large areas of peat-free serpentinite in Unst and Fetlar has limited the peat resources of these islands. As a result, Fetlar ran out of peat in about 1950 when the last peat bank on Lamb Hoga was exhausted. For some years peat was obtained from the adjacent island of Hascosay. Extensive peat remains on Unst (Figure 61), but over the last two decades, its use has declined.

Serpentinite

Two serpentinite quarries have been worked in Unst. A metadunite serpentinite quarry [HP 638 101] close to the Hagdale chromite quarry was worked during and after the Second World War to supply metadunite for use with low-grade chromite ore for the manufacture of refractory bricks (see above). In 1952, Phemister (1929–1963) noted that it had been worked since 1944, and 55 332 tons were produced between 1956 and 1969, with 2000 tons in 1966 (Andrews, 1966; personal communication S Owers, Alex. Sandison and Sons, unpublished archives, 2001).

The Setter quarry [HP 640 110] in the metaharzburgite was opened during the Second World War to supply aggregate for building the radar installation at Skaw. After the war, it was used for the surfacing of the dirt roads of Shetland and the rebuilding of the radar station on Saxa Vord. The Zetland Islands Council extracted 89 000 tons for this purpose between 1946 and 1975.

In 1966 10 000 tons were produced (Andrews, 1966). In the early seventies between 2500 and 7000 tons were extracted annually (personal communication A Johnson, Shetland Archives, unpublished, 2002). Between 1976 and the present 273 694 tons of aggregate for roads and building purposes have been produced from this quarry (personal communication S Owers, Alex. Sandison and Sons, unpublished archives, 2001).

Much use has been made over the years, especially between 1939 and 1945, of the serpentinite waste heaps beside the major chromite pits, and especially the Hagdale pit, for road construction and concrete.

Exploration

The first professional prospector to visit Unst was Alexander Crighton in 1789. Many have followed. Prospecting for chromite until the middle of the 20th century was a matter of visual inspection of exposures, helped by hand drilling and trenching. Following the original discovery of chromitite in the drift by Samuel Hibbert in 1822, all of the deposits were discovered by local inhabitants. The search was aimed at finding veins of coarse chromitite, generally several centimetres wide. These were followed, especially by hand drilling, in the hope of finding where they swelled to form pancake- like thicker lenses (‘podiform’ deposits) (Sandison, 1936). Such lenses varied in thickness, generally up to a metre or more, although Hagdale reached 4 m. This method of prospecting is revealed by the fact that almost no undisturbed in-situ vein-like occurrences of coarse ore-type chromite (chromitite) exist anymore at surface (Plate 3d), however thin and insubstantial. They have all been excavated. Well over 100 such excavations exist (Figure 61). Each, however small, has beside it two spoil heaps; one is formed of barren serpentinite and the other of serpentinite blocks containing segregations of ore-type chromite grains and/or veins of pure chromitite. Occurrences of accessory chromite, even where forming schlieren, are undisturbed. Samples of them are not included in the spoil heaps beside the trial pits. The search for in-situ ore-type chromitite was extended by trenching to expose bedrock and by intensive hand drilling to depths of up to 25 m. In 1870, about 100 such holes were drilled in Haaf Gruney alone [HU 645 985]. Intensive drilling also took place around Hagdale and in the Nikka Vord area. Only two or three minor economic deposits were found by this method.

Geophysical prospecting

From the 1950s onwards, gravimeters and magnetometers became the standard method of prospecting in Unst, usually combined with diamond drilling. Magnetic surveys (e.g. Lynam et al., 1942; Rivington, 1953), proceeded on the suggestion that magnetite might be associated with the chromite deposits, so enabling the chromite to be detected by magnetometer. However, Sanderson (1936) had already shown by experiment that the serpentinite is itself strongly and extremely variably magnetic due to the presence of magnetite, so that nothing would come of such a method. So, the ground-magnetic maps of the serpentinite produced showed the distribution of magnetite, not of chromite, and the prospecting was unsuccessful.

Similarly, gravity surveys were unlikely to succeed because of the relatively low density of the ore-type chromite: Lynam et al. (1942) reported a ‘bulk density’ of 2.99 g/cm³; Johnson et al. (1980) reported densities for massive ore of 3.73, 3.92 and 3.46 g/cm³; and 16 samples of high-grade ore measured for this survey averaged 3.77 g/cm³ (standard deviation 0.13 g cm3, range 3.56–3.98 g/cm³). Similarly, although the maximum density obtained for chromite concentrate during beneficiation in the 1920s was 4.06 g/cm³, the ‘best practical results’ were only 3.62 g/cm. Sandison, (1948), and Gass et al. (1982) found it impossible in beneficiation tests to separate chromite fragments from a coating of serpentine.

Gravity surveying would, according to Rivington (1953), require a 500 ton deposit to be less than 3 m from the surface to be detected, and Johnson et al. (1980) indicated even shallower depths, although even they assumed a higher density contrast between high- grade and low-grade ore (1.4 g/cm3) than is supported by measured densities. Despite this, Johnson et al. (1980, appendix 1) indicated that taking geological considerations into account, any chromite deposit detectable by gravimeter would also be visible at the surface if not covered by overburden, and ultimately recommended against the use of these methods in Unst. It is even possible that Johnson et al. (1980) over- estimated the overall density of chromite deposits by relying on the shape of the pits and the petrographical description of them as being of podiform shape instead of pancake-like veins.

It is noticeable that all chromite prospecting since 1818 has been confined to agriculturally poor areas (scattald) and has avoided arable land (Sandison, 1936). Geophysical prospecting, although suitable for examining the equally geologically prospective arable land along the shores of Balta Sound, has been applied only to the heavily prospected, mined out, and agriculturally useless area north of Balta Sound.

Geochemical prospecting

Langlands (1974) conducted a geochemical survey, based on soil samples from the whole outcrop of the ophiolite-complex, looking for Ni, Cu, Mo, and Pb. Subsequently, following the discovery of in-situ deposits of platinum group elements (PGE) by Prichard et al. (1981), a geochemical survey by Gunn et al. (1985) was aimed at investigating the deposits and methods of prospecting for them. The survey analysed bedrock, stream sediments and panned soil samples from 8 critical small areas, including known PGE-rich areas, for a wide variety of elements. Although no new deposits were confirmed, untested geochemical anomalies were found in some zones.

Regional stream-sediment-based geochemical surveys started in Shetland about 1970. Samples were collected in a standardised manner on a basis of approximately one per 1 km² for analysis for 19 elements and the results were published in 1978 (IGS, 1978a). In 1990–91 more stream-sediment samples were collected and all were analysed for 38 elements, many by improved methods. Several of these maps have been published in accounts of surveys aimed at platinum group elements and gold (Buchanan and Dunton, 1991, 1992; Flight et al., 1994). Buchanan and Dunton (1991) pointed out the necessity to discriminate between geochemical anomalies due to mineralisation and false anomalies caused by surface chemical processes and human activity. The largest

Pt anomaly noted as a potential target, north-east of Balta Sound, is probably mining related as it is based on the stream draining the Hagdale chromite quarry and ore processing area. Buchanan and Dunton (1991) and Gunn et al. (1985) both mention the necessity to allow for the effects of glaciation on geochemical maps based on the analysis of overburden samples. In Shetland and especially in Unst, this allowance is particularly important, as the stream sediments are dominantly composed of the fines winnowed out the glacial drift overburden (Chapter 9). In Unst, the stream sediments reflect the pattern of ice flow more clearly than they reflect the distribution of the bedrock (Figure 47) and related text, and this should be taken into account in interpreting anomalies such as the Pt, Pd, Ni and Cr anomaly, identified as a potential target (Buchanan and Dunton, 1996; Flight et al., 1994) that occurs on the basement in the south-west corner of Unst. This anomaly lies in the path of an ice-stream that headed for the Bluemull ice-stream and carried much ophiolitic debris.

Information sources

Further geological information held by the British Geological Survey relevant to the district is listed below. It includes published material in the form of maps, memoirs and reports, and unpublished maps and reports. Also included are other sources of data held by BGS in a number of collections, including borehole records, fossils, rock samples, thin sections, hydrogeological data and photographs.

Searches of BGS datasets can be made using GeoIndex; a map-based index of information that BGS has collected or has obtained from other sources. This is available online via the BGS website, and in BGS libraries. This searches indexes to collections and digital databases for specified geographical areas. It is based on a geographical information system linked to a relational database management system. Results of the searches are displayed on maps on the screen. The indexes which are available include:

• Borehole records and site investigation reports
• Geology (BGS rock samples and thin sections, list of active mines and quarries)
• Geochemistry (MRP reports, samples locations and data for soil, overburden, rocks and stream sediment analyses)
• Geophysics (aeromagnetic survey areas and flight lines and gravity data recording stations, location and survey data for local geophysical surveys and magnetic and gravity anomalies)
• BGS products (index of memoirs and outlines of maps at 1:50 000 and 1:10 000 scale and 1:10 560 scale County Series)
• Offshore data (released borehole location and other information including geophysical logs, aeromagnetic and gravity survey records)
• Worldwide information from work commissioned in a range of countries
• Enquiries concerning geological data for the district should be addressed to the National Geoscience Information Centre, BGS, Edinburgh.

Maps

• Geology maps
• 1:625 000
• Bedrock geology UK North, 2007
• United Kingdom (North Sheet) Quaternary geology, 1977
• 1:250 000
• Sheet 59 60ºN–02ºW,
• Solid geology, 1984
• Seabed Sediments, 1998
• Seabed Sediments and Quaternary, 1984
• 1:50 000
• Sheet 131, Unst and Fetlar; Solid and Drift, 2002
• Parts of Sheets 130 and 131, Yell and some adjacent islands; Solid and Drift, 1993
• 1:63 360
• Northern Shetland
• Solid (Special Sheet), 1968
• Solid and Drift (Special Sheet), 1968
• Western Shetland
• Solid (Special Sheet), 1971
• Solid and Drift (Special Sheet), 1971 Central Shetland
• Solid (Special Sheet), 1982
• Solid and Drift (Special Sheet), 1981 Southern Shetland
• Solid (Special Sheet), 1978
• Solid and Drift (Special Sheet), 1978
• 1:10 000 and 1:10 560
• H H Read and J Phemister of the Geological Survey of Great Britain carried out a systematic geological survey of Unst and Fetlar between 1929 and 1930, at a scale of six inches to one mile (1:10 560). Publication was delayed until the quarter-inch- to-one-mile publication in 1962, followed by publication at a scale of one inch to one mile (1:63 360) in 1968.
• Clean copy maps at 1:10 000, prepared between 1983 and 1994 by Prof D Flinn, are included in the 1:50 000 scale sheet 131. These, and earlier clean-copied 1:10 560 sheets produced by H H Read and J Phemister in the 1930s, are avail- able for consultation in the Library, British Geological Survey, Murchison House, Edinburgh, EH9 3LA.
• Geophysical maps and data
• Regional geophysics 1:1 000 000
• 60ºN–66ºN, 6ºW–0ºE
• Gravity Anomaly Map, 2000 Magnetic Anomaly Map, 2005 1:625 000
• Gravity anomaly map of the UK:North, 2007 Magnetic anomaly map of the UK:North,2007
• 1:250 000
• Sheet 60ºN–02ºW Aeromagnetic anomaly, 1980 Bouguer gravity anomaly, 1978
• Geochemical atlas and data
• 1:250 000
• The Geochemical Baseline Survey of the Environment (G-BASE) is based on the collection of stream sediment and stream water samples at an average density of one sample per 1.5 km2. The fine (minus 150 μm) fractions of stream sediment samples are analysed for a wide range of elements, using automated instrumental methods.
• The samples from Shetland were published in 1978. The results (including Ag, As, B, Ba, Be, Bi, CaO, Cd, Co, Cr, Cu, Fe₂O₃, Ga, K₂O, La, Li, MgO, Mn, Mo, Ni, Pb, Rb, Sb, Sn, Sr, TiO₂, U. V, Y, Zn and Zr in stream sediments and pH, conductivity, fluoride, bicarbonate and U for stream waters) are published in atlas form.
• Institute of Geological Sciences. 1978. Regional Geochemical Atlas, Shetland. (London: Institute of Geological Sciences.)
• The geochemical data, with location and site information are available as hard copy for sale in digital form under licensing agreement. The coloured geochemical atlas is also available in digital form (CD-ROM) under licensing agreement. The British Geological Survey offers a client-based service for interactive GIS interrogation of G-BASE data.
• Hydrogeological map
• 1:625 000
• Scotland, 1988
• Groundwater vulnerability map
• 1:625 000
• Scotland, 1995 (out of print)
• Digital geological map data
• In addition to the printed publications noted above, many BGS maps are available in digital form, which allows the geological information to be used in GIS applications. These data must be licensed for use. Details are available from the Intellectual Property Rights Manager at BGS Keyworth. The main datasets are:
• DigMapGB (1:625 000 scale)
• DigMapGB (1:250 000 scale)
• DigMapGB (1:50 000 scale)
• DigMapGB (1:10 000 scale)
• The current availability of these can be checked on the BGS website at: www.bgs.ac.uk/products/digitalmaps/digmapgb.html

Publications

Memoirs, books, reports and papers relevant to the district are arranged by topic. Some publications are out of print but may be consulted at BGS and other libraries.

• British Regional Geology: Orkney and Shetland. 1976.
• Geology of Yell and some neighbouring islands in Shetland, British Geological Survey Memoir, 1994.
• The Geology of Western Shetland, Institute of Geological Sciences,1976.
• Thomas, C H. 1980. Analysis of rock samples from the Shetland Ophiolite belt. Analytical Chemistry Unit, Institute of Geological Sciences, Report No. 113.
• Offshore Regional Reports
• Geology of the Northern North Sea, 1994.
• Tully, M C, And Donato, J A. 1985. 1:1 000 000 northern North Sea Bouguer anomaly gravity map. British Geological Survey Research Report, 16, No. 8 (Southampton: Her Majesty’s Stationery Office)
• Economic Geology and mineralisation
• Andrews, R W. 1966. The talc quarry, Baltasound, Unst, Shetland Islands. Institute of Geological Sciences, Overseas Division. Mineral Intelligence Report, No. 69.
• Gunn, A G, Leake, R C, Styles, M T, And Bateman, J H. 1985. Platinum group mineralisation in the Unst ophiolite, Shetland. British Geological Survey Mineral Reconnaissance Programme Report, No. 73.
• Johnson, C E, Smith, C G, And Fortey, N J. 1980. Geophysical investigation of chromite bearing ultrabasic rocks in the Baltasound, Hagdale area, Unst, Shetland Islands. Institute of Geological Sciences, Mineral Reconnaissance Programme Report, No. 35.
• Environmental baseline geochemistry
• Institute of Geological Sciences. 1978. Regional Geochemical Atlas, Shetland. (London: Institute of Geological Sciences.)
• Hydrogeology/geochemistry
• Robins, N S. 1990. Hydrogeology of Scotland (London: HMSO for the British Geological Survey.)

Documentary collections

Borehole record collections
• BGS holds collections of records of boreholes which can be consulted at BGS Edinburgh, where copies of most records may be purchased.
• Most were drilled in search of metalliferous minerals, with others divided between water supplies and site investigation. Index information, which includes site references, for these boreholes has been digitised. The logs are either hand-written or typed and many of the older records are drillers’ logs.
• Site investigation reports
• This collection consists of site investigation reports carried out to investigate foundation conditions prior to construction. There is a digital index and the reports themselves are held on microfiche.
• Mine plans
• BGS maintains a collection of plans of underground mines for minerals other than coal and oil-shale. The mines known to have been worked or trialled within the district are discussed in Chapter 12 (see Figure 61).

Material collections

References

Alexander, P A. 1975. The sandy beaches of S. Shetland. Unpublished BSc dissertation, University of Liverpool.

Allen, N H. 1983. Recent temperate carbonate deposits on the continental shelf north and west of Shetland: distribution sedimentology and reserves. Unpublished PhD thesis, University of Strathclyde.

Andrews, R W. 1966 The talc quarry, Baltasound, Unst, Shetland Islands. Institute of Geological Sciences, Overseas Division, Mineral Intelligence Report, No..69.

Anon. 1926–27. Fieldwork on the geology of Unst and Fetlar 1926 and 1927 and Fieldwork on the geology of Fetlar 1927. Unpublished Land Survey Archives, LSA/169, British Geological Survey, Edinburgh.

Augé, T, Legendre, O, and Maurizot, P. 1998. The distribution of Pt and Ru-Os-Ir minerals in the New Caledonia ophiolite. 141–154 in International Platinum. Laverov, N P, and Distler, V V (editors). (Athens, Greece: Theophrastus Publications.)

Aziz, R M. 1984. Metamorphic history and geochemistry of the Valla Field Block (Unst) and the Lamb Hoga Block (Fetlar), Shetland. Unpublished PhD thesis, University of Liverpool.

Bacuta, G, Kay, R, Gibbs, A, and Lippin, B. 1990. Platinum-group element abundance and distribution in chromite deposits of the Acoje block, Zambales ophiolite complex, Philippines. Journal of Geophysical Research, Vol..37, 113–145.

Bates, R L, and Jackson, J A (editors). 1987. Glossary of Geology. Third edition. (Alexandria, Virginia, USA: American Geological Institute.) ISBN.0.913312.89.4

Beg, M A. 1990. The distribution and disposal of heavy minerals on the continental shelf around the Shetland Islands. Unpublished PhD thesis, University of Strathclyde.

Bennett, K D, and Sharpe, M J. 1993. Holocene vegetation and environment. 18–22 in The Quaternary of Shetland: Field Guide. Birnie, J F, Gordon, J E, Bennett, K J, and Hall,.A.M (editors). (Cambridge: Quaternary Research Association.)

Bennett, K D, Boreham, S, Sharp, M J, and Switser,.V.R. 1992. Holocene history of environment, vegetation and human settlement in Catta Ness, Lunnasting, Shetland. Journal of Ecology, Vol..80, 241–273.

Birnie, J F. 1993. The late-glacial in Shetland. 15–16 in The Quaternary of Shetland: Field Guide. Birnie, J F, Gordon, J E, Bennett, K J, and Hall, A M (editors). (Cambridge: Quaternary Research Association.)

Birnie, J F, Gordon, J E, Bennett, K J, and Hall, A M (editors).

1993. The Quaternary of Shetland: Field Guide. (Cambridge: Quaternary Research Association.)

British Geological Survey. 1984. Shetland. Sheet 59 50º N–020ºW. Solid Geology. 1:250 000. (Southampton: Ordnance Survey for the British Geological Survey).

British Geological Survey. 1993. Yell and some adjacent islands, Shetland. Scotland Parts of Sheets 130 and 131. Solid and Drift. 1:50 000. (Southampton: Ordnance Survey for the British Geological Survey.)

British Geological Survey. 2002. Unst and Fetlar. Scotland Sheet 131. Solid and Drift. 1:50 000. (Keyworth, Nottingham: British Geological Survey.)

Brzozowsky, T J. 1977. VLF electromagnetic surveys and associated studies in north Dartmoor and Unst. Unpublished PhD thesis, Chelsea College, University of London.

Buchanan, D L, and Dunton, S N. 1991. Preliminary report on the mineral resources of Shetland and the results of a pilot geochemical survey. Shetland Islands Council, Lerwick, Shetland.

Buchanan, D L, and Dunton, S N. 1992. Results of a geochemical survey of Shetland and identification of exploration targets. Shetland Islands Council, Lerwick, Shetland.

Buchanan, D L, and Dunton, S N. 1996. Precious-metal distribution in Shetland: refinement of targets for gold exploration. Shetland Islands Council, Lerwick, Shetland.

Buttler, S. 1984. The steatite industry in Norse Shetland. Unpublished PhD thesis, University of Liverpool.

Buttler, S. 1989. Steatite in Norse Shetland. Hikuin, Vol..15, 193–206.

Cameron, T D J, Stoker, M S, and Long, D. 1987. The history of Quaternary sedimentation in the UK sector of the North Sea Basin. Journal of the Geological Society of London, Vol. 144, 43–58.

Carter, J R, and Bailey Watts, A E. 1981. A taxonomic study of diatoms from standing freshwaters in Shetland. Nova Hedwigia, Vol..33, 513–629.

Carter, J R, Procter, J, and Slingsby, J R. 1987. Soil and vegetation on the Keen of Hamar serpentine, Shetland. Journal of Ecology, Vol..75, 21–41.

Clark, C D, Evans, D J A, and Khatwa, A. 2004. Map and GIS database of land forms and features related to the last British ice sheet. Boreas, Vol. 33, 359–375.

Coleman, R G, and Donato, M M. 1979. Oceanic plagio-granites revisited. 149–168 in Trondhjemites, dacites and related rocks. Barker, F (editor). (Amsterdam, Netherlands: Elsevier.)

Corrivaux, L, and La Flamme, J H G. 1990. Mineralogie des elements du groupe du platine dans les chromitites de l’ophiolite de Thetford Mines, Quebec. Canadian Mineralogist, Vol..28, 579–595.

Crawford, A J, Falloon, T J, and Green, D H. 1989. Classification, petrogenesis and tectonic setting of boninites. 2–44 in Boninites and related rocks. Crawford, A J (editor). (London: Unwin Hyman.)

Croll, J. 1870. The boulder clay of Caithness a product of land ice. Geological Magazine, Vol..7, 271–278.

Curtis, C D, and Brown, P E. 1969. The metasomatic development of zoned ultrabasic bodies in Unst, Shetland. Contributions to Mineralogy and Petrology, Vol. 24, 275–292.

Curtis, C D, and Brown, P E. 1971. Trace element behavior in the zoned metasomatic bodies of Unst, Shetland. Contributions to Mineralogy and Petrology, Vol. 37, 87–93.

Curtis, C D, Brown, P E, and Somogyi, V A. 1969. A naturally occurring sodium vermiculite from Unst, Shetland. Clay Minerals, Vol..8, 15–19.

Double, I S. 1939. Notes on some recent sediments from Unst, Shetland Islands. Proceedings of the Liverpool Geological Society, Vol..17, 321–328.

Edmonston, T. 1839. Unpublished letter to the chromate proprietors of the scattolds of Baliasta and Haroldswick, Shetland. Shetland County Library, Lerwick, Shetland.

Edwards, S J. 1990. Harzburgites and refractory melts in the Lewis Hills massif, Bay of Islands ophiolite complex: the base metals and precious metals story. Canadian Mineralogist, Vol. 28, 537–552.

Fernando, L J D. 1941. Petrology of certain felspathised rocks from Herma Ness, Unst, Shetland Islands. Proceedings of the Geologist’s Association, Vol. 52, 110–130.

Finlay, T M. 1926. A tongsbergite boulder from the boulder-clay of Shetland. Transactions of the Edinburgh Geological Society, Vol..12, 180.

Firth, C R, and Smith, D E. 1993. Holocene sea level changes in Shetland. 17 in The Quaternary of Shetland. Birnie, J J, Bennett, G K, and Hall, A (editors). (Cambridge: Quaternary Research Association.)

Flight, D M A, Hall, G E M, and Simpson, P R. 1994. Regional geochemical mapping of Pt, Pd and Au over an obducted ophiolite complex, Shetland Islands, northern Scotland. Transactions of the Institution of Mining and Metallurgy, Section B: Applied Earth Sciences, Vol. 103, 68–78.

Flinn, D. 1952. A tectonic analysis of the Muness Phyllite Block of Unst and Uyea, Shetland. Geological Magazine, Vol..89, 263–272.

Flinn, D. 1956. On the deformation of the Funzie conglomerate, Fetlar, Shetland. Journal of Geology, Vol..64, 480–505.

Flinn, D. 1958. On the nappe structure of north-east Shetland.

Quarterly Journal of the Geological Society of London, Vol..94, 107–136.

Flinn, D. 1964. Coastal and submarine features around the Shetland Islands. Proceedings of the Geologists’ Association, Vol..75, 321–339.

Flinn, D. 1967a. Ice front in the North Sea. Nature, London, Vol..215, 1151–1154.

Flinn, D. 1967b. The metamorphic rocks of the southern part of the Mainland of Shetland. Geological Journal, Vol..5, 251–290.

Flinn, D. 1973. The topography of the sea floor around Orkney and Shetland in the northern North Sea. Quarterly Journal of the Geological Society of London, Vol..129, 39–59.

Flinn, D. 1974. The coastline of Shetland. 13–23 in The natural environment of Shetland. Goodier, R (editor). (Swindon: Nature Conservancy Council.)

Flinn, D. 1977. The erosion history of Shetland: a review. Proceedings of the Geologists’ Association, Vol..88, 129–146.

Flinn, D. 1978. The most recent glaciation of the Orkney–Shetland channel and adjacent areas. Scottish Journal of Geology, Vol..14, 109–123.

Flinn, D. 1983. Glacial meltwater channels in the northern islands of Shetland. Scottish Journal of Geology, Vol..19, 311–320.

Flinn, D. 1988. The Moine rocks of Shetland. 76–85 in Later Proterozoic stratigraphy of the northern Atlantic regions. Winchester, S A (editor). (Glasgow: Blackie.)

Flinn, D. 1989. Travellers in a bygone Shetland. (Edinburgh: Scottish Academic Press.)

Flinn, D. 1992a. The Milldale glacial lake, Herma Ness, Unst. The Shetland Naturalist, Vol..1, 29–36.

Flinn, D. 1992b. A note on the Dalsetter erratic, Dunrossness, Shetland. The Shetland Naturalist, Vol..1, 49–50.

Flinn, D. 1993. New evidence that the high temperature hornblende schists below the Shetland ophiolite include basic igneous rocks intruded during obduction of the ‘cold’ ophiolite. Scottish Journal of Geology, Vol..29, 159–165.

Flinn, D. 1994a. Geology of Yell and some neighbouring islands in Shetland. Memoir of the British Geological Survey, Sheet 130 (Scotland).

Flinn, D. 1994b. Ice flow in Unst during the last glaciation. The Shetland Naturalist, Vol..1, 73–79.

Flinn, D. 1995a. The melt down of the Shetland ice cap in Unst. The Shetland Naturalist, Vol..1, 109–115.

Flinn, D. 1995b. Formation of gneisses of migmatite and diatexite appearance in Yell, Shetland by solid-state grain growth recrystallisation. Geological Journal, Vol..30, 415–422.

Flinn, D. 1996a. The non-lacustrine diatomite deposits of Unst, Shetland. Diatom Research, Vol..11, 31–34.

Flinn, D. 1996b. The Shetland Ophiolite Complex: field evidence for the intrusive emplacement of the ‘cumulate’ layers. Scottish Journal of Geology, Vol..32, 151–158.

Flinn, D. 1999. The Shetland ophiolite. 29–54 in Caledonian Igneous Rocks of Great Britain. Stephenson, D, Bevins, R E, Millward, D, Highton, A J, Parsons, I, Stone, P, and Wadsworth, W J. Geological Conservation Review Series, No..17. (Peterborough: Joint Nature Conservation Committee.)

Flinn, D. 2000. The architecture of the Shetland Ophiolite.

Scottish Journal of Geology, Vol..36, 123–135.

Flinn, D. 2001. The basic rocks of the Shetland Ophiolite Complex and their bearing on its genesis. Scottish Journal of Geology, Vol..37, 79–96.

Flinn, D. 2007. The Dalradian rocks of Shetland and their implications for the plate tectonics of the northern Iapetus. Scottish Journal of Geology, Vol..43, 125–142.

Flinn, D, and Moffat, D T. 1985. A peridotitic komatiite from the Dalradian of Shetland. Geological Journal, Vol..20, 287–292.

Flinn, D, and Oglethorpe, R J D. 2005. A history of the Shetland Ophiolite. Scottish Journal of Geology, Vol..41, 141–148.

Flinn, D, and Pentecost, A. 1995. Travertine-cemented screes on the serpentinite seacliffs of Unst and Fetlar, Shetland. Mineralogical Magazine, Vol..59, 259–265.

Flinn, D, May, F, Roberts, J L, and Treagus, J E. 1972. A review of the stratigraphic succession of the East Mainland of Shetland. Scottish Journal of Geology, Vol..8, 335–343.

Flinn,.D, Frank, P L, Brook, M, and Pringle,.I.R. 1979. Basement–cover relations in Shetland. 109–115 in The Caledonides of the British Isles.–.reviewed. Harris, A L, Holland, C H, and Leake,.B L (editors). Special Publication of the Geological Society of London, No..8.

Flinn, D, Miller, J A, and Roddam, D. 1991. The age of the Norwick hornblendic schists of Unst and Fetlar and the obduction of the Shetland ophiolite. Scottish Journal of Geology, Vol..27, 11–19.

Flinn, D, Key, R M, and Khoo, T T. 1996. Chloritoid schists of Shetland and their thermal metamorphism. Scottish Journal of Geology, Vol..32, 7–82.

Gamil, A S. 1991. Petrology and geochemistry of Shetland granites. Unpublished PhD thesis, University of Liverpool.

Garnett Frizelle, C J. 1979. The sandy beaches of Yell. Unpublished BSc dissertation, University of Liverpool.

Garson, M S, and Plant, J. 1973. Alpine type ultramafic rocks and episodic mountain building in the Scottish Highlands. Nature, Vol..242, 34–38.

Gass, I G. 1980. The Troodos massif: its role in the unravelling of the ophiolite problem and its significance in the understanding of constructive plate margin processes. 23–35 In Ophiolites. Panayiotou, A (editor). Proceedings of the International Ophiolite Symposium, Cyprus 1979.

Gass, I G, Prichard, H M, Neary, C R, and Bartholemew, I D. 1982. The chromite of the Shetland ophiolite. A reappraisal in the light of new theory and technique. A report for the Commission of European Communities. Department of Earth Sciences, The Open University, Milton Keynes.

Gear, R W. 2005. John Walker’s Shetland. (Lerwick: Shetland Times.)

Guillot, S, Hattori, K H, and De Sigoyer, J. 2000. Mantle wedge serpentinisation and exhumation of eclogites: Insights from eastern Ladakh, north-west Himalaya. Geology, Vol. 28, 199–202.

Gunn, A G. 1989. Drainage and overburden geochemistry in exploration for platinum-group element mineralisation in the Unst ophiolite, Shetland, UK. Journal of Geochemical Exploration, Vol..31, 209–236.

Gunn, A G, and Styles M T. 2002. Platinum-group element occurrences in Britain: magmatic, hydrothermal and super-gene. Transactions of the Institution of Mining and Metallurgy, Section B: Applied Earth Sciences, Vol..111, 3–14.

Gunn, A G, Leake, R C, Styles, M T, and Bateman, J.H. 1985. Platinum group mineralisation in the Unst ophiolite, Shetland.

British Geological Survey Mineral Reconnaissance Programme Report, No..73.

Guppy, E M, and Sabine, P A. 1956. Chemical analyses of igneous rocks, metamorphic rocks and minerals, 1931–54. Memoir of the Geological Survey of Great Britain. (London: Her Majesty’s Stationery Office.)

Hafez, A M A. 1972. A petrographic study of the Funzie Conglomerate, Fetlar, Shetland. Unpublished PhD thesis, University of Liverpool.

Haldane, D, Eyles, V J, and Davidson, C F. 1940. Diatomite. Geological Survey of Great Britain Wartime Pamphlet, No..5.

Heddle, M F. 1878. The county geognosy and mineralogy of Scotland. (Truro: Lake and Lake.)

Hibbert, S. 1819. Sketch of the distribution of rocks in Shetland. Edinburgh Philosophical Journal, Vol..1, 296–314; Vol..2, 67–79, 224–242.

Hibbert, S. 1822. A Description of the Shetland Islands. (Edinburgh: Constable.)

Hibbert, S. 1831. On the direction of the diluvial wave in the Shetland Islands. Edinburgh Journal of Science, Vol..4, 89–91.

Hitchin, C S. 1929. Unst and its chromite deposits. Mining Magazine, Vol..40, 18–24.

Hoppe, G. 1965. Submarine peat in the Shetland Islands. Geografiska Annaler. Vol..47A, 95–203.

Hoppe, G. 1974. The glacial history of the Shetland Islands.

Special Publication of the Institute of British Geographers, No..7, 197–210.

Ingram, J A, and Ingram, J O. 1845. Parish of Unst. 36–53 in The new statistical account of Scotland. (Edinburgh: Blackwood.)

Institute of Geological Sciences. 1962. Shetland. Sheets 1 and 2 (combined). Solid. 1:253 440. (Chessington, Surrey: Ordnance Survey for Institute of Geological Sciences.)

Institute of Geological Sciences. 1968a. Northern Shetland.Scotland. Parts of Sheets 129, 130 and 131. Solid. 1:63 360. (Chessington, Surrey: Ordnance Survey for Institute of Geological Sciences.)

Institute of Geological Sciences. 1968b. Aeromagnetic map of part of Great Britain and Northern Ireland. Sheet 16. (Chessington, Surrey: Ordnance Survey for Institute of Geological Sciences.)

Institute of Geological Sciences. 1972. Aeromagnetic map of part of Great Britain: Sheet 1. (Chessington, Surrey: Ordnance Survey for Institute of Geological Sciences.)

Institute of Geological Sciences. 1978a. Regional geochemical Atlas, Shetland. (London: Institute of Geological Sciences.)

Institute of Geological Sciences. 1978b. Shetland Sheet 60ºN–02ºW. Bouguer gravity anomaly map. Provisional edition. 1:250 000 Series. (London: Institute of Geological Sciences.)

Institute of Geological Sciences. 1980. Shetland, Sheet 60º N–02ºW. Aeromagnetic anomaly map 1:250.000. (London: Institute of Geological Sciences.)

Institute of Geological Sciences. 1981. Central Shetland. Scotland Sheet 128. Solid. 1:63 360. (Chessington, Surrey: Ordnance Survey for Institute of Geological Sciences.)

Institute of Geological Sciences. 1982. Central Shetland. Scotland Sheet 128. Drift. 1:63 360. (Chessington, Surrey: Ordnance Survey for Institute of Geological Sciences.)

Jameson, R. 1798. An outline of the mineralogy of the Shetland Islands and of the Island of Arran. (Edinburgh: Creech, Cadell and Davies.)

Johansen, J. 1975. Pollen diagrams from the Shetland and Faroe Islands. New Phytology, Vol..75, 369–387.

Johansen, J. 1976. The age of the introduction of Plantago lanceolata to the Shetland Islands. Danmarks Geologiske Undersøgelse, 1976, 45–48.

Johansen, J. 1985. Studies in the vegetational history of the Faroe and Shetland Islands. Annales Societates Scientiarum Faroensis, Supplementum XI, Torshavn.

Johnson, C E, Smith, C G, and Fortey, N J. 1980. Geophysical investigation of chromite bearing ultrabasic rocks in the Baltasound Hagdale area, Unst, Shetland Islands. Institute of Geological Sciences Mineral Reconnaissance Programme Report, No..35.

Johnson, H, Richards, P C, Long, D, and Graham, C C. 1993. The geology of the northern North Sea. British Geological Survey Offshore Regional Report, No..1.

Kater, H. 1819. Account of experiments for determining the variation in the length of the pendulum etc. Philosophical Transactions of the Royal Society of London, Vol. 109, 337–508.

Keatinge, T M, and Dickson, J H. 1979. Mid Flandrian changes in vegetation in Mainland, Orkney. New Phytologist, Vol..82, 585–612.

Kelemen, P B, Shimzu, N, and Salters, V J M. 1995. Extraction of mid-ocean-ridge basalt from the upwelling mantle in dunite channels. Nature, Vol..375, 747–753.

Key, R M. 1972. A study of the metamorphic histories of the Saxa Vord, Valla Field and Lamb Hoga Blocks of NE Shetland. Unpublished PhD thesis, University of Liverpool.

Khoo, T T. 1974. The mineralogy, petrology and geochemistry of regional and thermal Dunrossness Phyllites from south Mainland, Shetland. Unpublished PhD thesis, University of Liverpool.

Langlands, J G. 1974. Project 0408, Shetland. Geochemical and geological prospecting for nickel, copper, molybdenum and lead, Unst, Fetlar and North Roe. Mineral Exploration and Investment Grants Act (MEIGA) Report AE 136, Noranda Kerr Ltd.

Lee, C A. 1996. A review of mineralisation in the Bushveld complex and some other layered intrusions. 103–147 in Layered intrusions. Cawthorn, R G (editor). (Amsterdam: Elsevier.)

Le Maitre, R W (editor). 2002. Igneous rocks: a classification and glossary of terms. Recommendations of the International Union of Geological Sciences, Subcommission on the Systematics of Igneous Rocks. Second edition. (Oxford, UK: Blackwell Scientific Publications.)

Lippard, S J, Shelton, A W, and Gass, I G. 1986. The ophiolite of northern Oman. Geological Society of London Memoir, No..11.

Lippmann, F. 1973. Sedimentary carbonate minerals. (Berlin: Springer-Verlag.)

Long, D, and Skinner, A C. 1985. Glacial meltwater channels in the northern isles of Shetland : comment. Scottish Journal of Geology, Vol..21, 222–224.

Lord, R A, and Prichard, H M. 1998. Exploration and origin of stratigraphically controlled platinum-group element mineralisation in crustal-sequence ultramafics, Shetland ophiolite complex. Transactions of the Institution of Mining and Metallurgy, Section B: Applied Earth Science, Vol..106, 179–193.

Lord, R A, Prichard, H M, and Neary, C R. 1994. Magmatic platinum element concentrations and hydrothermal upgrading in Shetland ophiolite complex. Transactions of the Institution of Mining and Metallurgy, Section B: Applied Earth Science, Vol..103, B87–106.

Lynam, T R, Chesters, J H, Howie, T W, and Jay, A.H. 1942. Shetland chromite as a refractory material. Transactions of the British Ceramic Association, Vol. 41, 27–44.

Maniar, P D, and Piccoli, P M. 1989. Tectonic discrimination of granitoids. Bulletin of the Geological Society of America, Vol..101, 635–643.

Mather, A S, and Smith, J S. 1974. Beaches of Shetland. (Countryside Commission for Scotland.)

May, F, and Phemister, J. 1968. Kaolin deposits in the Shetland Islands, UK. XXIII International Geological Congress, Vol..14, 23–29.

Maynard, J, Prichard, H M, Ixer, R A, Lord, R A, Wright, I.P, Pillinger, C T, McConville, P , Boyce, A J, and Fallick,.A.E. 1997. Sulphur isotope study of Ni-Fe-Cu mineralisation in the Shetland ophiolite. Transactions of the Institution of Mining and Metallurgy, Section B: Applied Earth Science, Vol..106, 215–226.

McQuillin, R, and Brooks, M. 1967. Geophysical surveys in the Shetland Islands. Institute of Geological Sciences Geophysical Paper, No.2.

Miller, J A, and Flinn, D. 1966. A survey of the age relations of Shetland rocks. Geological Journal, Vol..5, 95–116.

Moffat, D T. 1987. The serpentinised ultramafics of the Shetland Caledonides. Unpublished PhD thesis, University of Liverpool.

Muller, M R, Robinson, C J, Minshull, T A, White, R S, and Bickle, M J. 1997. Thin crust beneath Ocean Drilling Program borehole 735B at the Southwest Indian Ridge. Earth and Planetary Science Letters, Vol..148, 93–107.

Mykura, W. 1976. British Regional Geology. Orkney and Shetland. (Edinburgh: HMSO for Institute of Geological Sciences.)

Mykura, W, and Phemister, J. 1976. The geology of western Shetland. Memoir of the Geological Survey of Great Britain, Sheet 127 (Scotland).

Naldrett, A, J, and Duke, J M. 1980. Pt metals in magmatic sulphide ores: the occurrence of these metals is discussed in relations to the formation and importance of these ores. Science, Vol..208, 1417–1424.

Naldrett, A, J, Kinnaird, J, Wilson, A, Yudovskaya, M, McQuade, S, Chunnett, G, and Stanley, C. 2009. Chromite composition and PGE content of Bushveld chromitites: Part 1. — the Lower and Middle groups. Transactions of the Institute of Mining and Metallurgy, Section B: Applied Earth Sciences, Vol. 118, 131–161.

Neary, C R, and Prichard, H M. 1985. Molybdenum mineral-isation on Unst, Shetland Islands. Scottish Journal of Geology, Vol..21, 197–204.

Neary, C R, Prichard, H M, and Potts, P J. 1984. Chromite, platinoids, gold and moly in the Shetlands. Mining Magazine, Vol. 151, 559–560.

Nickson, A. 1980. A reconnaissance on the sandy beaches of the Island of Unst, Shetland. Unpublished BSc dissertation, University of Liverpool.

O’Hanley, D S. 1992. Solution to the volume problem in serpentinisation. Geology, Vol..20, 705–708.

O’Hanley, D S. 1996. Serpentinites–recorders of tectonic and petrological history. (New York, USA: Oxford University Press.)

O’Hara, M J, Fry, N, and Prichard, H M. 2001a. Minor phases as carriers of trace elements in non-modal crystal–liquid separation processes I: basic relationships. Journal of Petrology, Vol..42, 1869–1885.

O’Hara, M J, Fry, N, and Prichard, H M. 2001b. Minor phases as carriers of trace elements in non-modal crystal–liquid separation processes II: illustrations and bearing on behavior of REE, U, Th and PGE in igneous processes. Journal of Petrology, Vol..42, 1887–1910.

Ohnenstetter, M, Johan, Z, Cocherie, A, Fouillac, A M, Guerrot, C, Ohnenstetter, D, Chaussidon, M, Rouer, O, Makovicky, E, Makovicky, M, Rose-Hansen, J, Karup-Møller,. S, Vaughan, D, Turner, G, Pattrick, R A D, Gize, A P, Lyon, I, and MacDonald, I. 1999. New exploration methods for platinum and rhodium deposits poor in base-metal sulfides–NEXTPRIM. Transactions of the Institution of Mining and Metallurgy, Section B: Applied Earth Sciences, Vol..108, 119–150.

O’Reilly, B M, Hauser, F, Jacob, A W B, and Shannon, P M. 1996. The lithosphere below the Rockall Trough: wide angle seismic evidence for extensive serpentinisation. Tectono-physics, Vol..255, 1–23.

Peach, B N, and Horne, J. 1879. The glaciation of the Shetland Isles. Transactions of the Institution of Mining and Metallurgy, Vol..35, 778–811.

Peach, C W. 1864. On traces of glacial drift in the Shetland Islands. Geological Magazine, Vol..2, 341–343.

Peacock, J D, and Long, D. 1994. Late Devensian glaciation and deglaciation of Shetland. Quaternary Newsletter, Vol..74, 16–21.

Pearce, J A. 1982. Trace element characteristics of lavas from destructive plate boundaries. 526–845 in Andesites. Thorpe, R S (editor). (Chichester: John Wiley and Sons.)

Pearce, J A. 1983. Role of the subcontinental lithosphere in magma genesis at active continental margins. 230–249 in Continental basalts and mantle xenoliths, Hawkesworth, C J, and Norry, M J (editors). (Nantwich: Shiva Publications.)

Pedersen, R B, and Malpas, J. 1984. The origin of oceanic plagiogranite from the Karmoy ophiolite. Contributions to Mineralogy and Petrology, Vol. 88, 36–52.

Pedersen, R B, Johannesen, G M, and Boyd, R. 1993. Stratiform PGE mineralisation in the ultramafic cumulates of the Leka ophiolite complex, central Norway. Economic Geology, Vol..88, 782–803.

Pedley, R C, Busby, J P, and Dabek, Z K. 1993. Gravmag v 1.5 user manual. British Geological Survey Technical Report, Regional Geophysical Series, WK/93/26/R.

Phemister, J. 1929–1963. Unpublished papers concerning Fetlar. LSA/169. British Geological Survey, Edinburgh.

Phemister, J. 1964. Rodingitic assemblages in Fetlar, Shetland Islands, Scotland. Advancing frontiers in geology and geophysics, 1964, 279–295.

Phillips, F C. 1927. The serpentines and associated rocks and minerals of the Shetland Islands. Quarterly Journal of the Geological Society of London, Vol..83, 622–652.

Pickup, S L B, Whitmarsh, R B , Fowler, C M R, and Reston,. T.J. 1996. Insight into the nature of the ocean-continent transition off West Iberia from a deep multichannel seismic-reflection profile. Geology, Vol..24, 1079–1082.

Plant, J A, Watson, J V, and Green, P M. 1984. Moine–Dalradian relationships and their palaeotectonic significance. Proceedings of the Royal Society of London, A 395, 185–202.

Prichard, H M. 1985. The Shetland Ophiolite. 1173–1184 in The Caledonide Orogen: Scandinavia and related areas, Gee,.D.G, and Sturt, B A (editors). (Chichester: Wiley and Sons Ltd.)

Prichard, H M, and Lord, R A. 1988. The Shetland ophiolite: evidence for a supra-subduction zone origin and implications for platinum-group element mineralisation. 289–302 in Mineral deposits within the European Community, Boissonnas, J, and Omenetto, P, (editors). (Berlin: Springer-Verlag.)

Prichard, H M, and Lord, R A. 1990. Platinum and palladium in the Troodos ophiolite complex, Cyprus. Canadian Mineralogist, Vol..28, 607–617.

Prichard, H M, and Lord, R A. 1993. An overview of the PGE concentrations in the Shetland ophiolite complex. 273–294 in Magmatic processes and plate tectonics. Prichard, H M, Alabaster, T, Harris, N B W, and Neary, C R (editors). Special Publication of the Geological Society of London, No..76.

Prichard, H M, and Tarkian, M. 1988. Pt and Pd minerals from two PGE- rich localities in the Shetland ophiolite complexes. Canadian Mineralogist, Vol..26, Part 4, 979–990.

Prichard, H M, Potts, P J, and Neary, C R. 1981. Platinum group minerals in the Shetland ophiolite complex.

Transactions of the Institution of Mining and Metallurgy, Section B: Applied Earth Science, Vol..90, 186–188.

Prichard, H M, Neary, C R, and Potts, P J. 1984. Platinum and gold in the Shetland ophiolite. Mining Journal, Vol..303, No..7772, 77.

Prichard, H M, Neary, C R, and Potts, P J. 1986. Platinum-group minerals in the Shetland ophiolite complex. 395–414 in Metallogenesis of basic and ultrabasic rocks. Gallagher, M J, Ixer, R A, Neary, C R, and Prichard, H M (editors). Institution of Mining and Metallurgy Symposium Volume, Edinburgh, 1986.

Prichard, H M, Ixer, R A, Lord, R A, Maynard, J, and Williams,.N. 1994. Assemblages of platinum-group minerals and sulphides in silicate lithologies and chromite-rich rocks within the Shetland ophiolite. Canadian Mineralogist, Vol..32, No. 2, 271–294.

Prichard, H M, Lord, R A, and Neary, C R. 1996. A model to explain the occurrence of Pt-and Pd-rich ophiolite complexes. Journal of the Geological Society of London, Vol..153, 323–328.

Prichard, H M, Sa, H, and Fisher, P C. 2001. Platinum-group mineral assemblages and chromite composition in the altered and deformed Bacuri complex, Amapa, Northeastern Brazil. Canadian Mineralogist, Vol..39, 377–396.

Prichard, H, M, Neary, C R, Fisher, P C, and O'Hara, M J. 2008. PGE-rich podiform chromitites in the Al'Ays ophiolite complex, Saudi Arabia: an example of critical mantle melting to extract and concentrate PGE. Economic Geology, Vol. 103, 1507–1529.

Read, H H. 1927. The igneous and metamorphic history of Cromar, Deeside, Aberdeenshire. Transactions of the Royal Society of Edinburgh, Vol..55, 317–353.

Read, H H. 1931. The geology of Central Sutherland. Memoir of the Geological Survey of Great Britain. Sheet 108 and 109 (Scotland).

Read, H H. 1933. On quartz kyanite rocks in Unst, Shetland and their bearing on metamorphic differentiation. Mineralogical Magazine, Vol..23, 317–328.

Read, H H. 1934a. The metamorphic geology of Unst in the Shetland Islands. Quarterly Journal of the Geological Society of London, Vol..90, 637–688.

Read, H H. 1934b. On zoned associations of antigorite, talc, actinolite, chlorite, and biotite in Unst, Shetland Islands. Mineralogical Magazine, Vol..23, 519–540.

Read, H H. 1934c. On segregation of quartz-chlorite-pyrite masses in Shetland igneous rocks during dislocation meta-morphism, with a note on the occurrence of boudinage structure. Proceedings of the Liverpool Geological Society, Vol..16, 128–138.

Read, H H. 1937. Metamorphic correlation in the polymeta-morphic rocks of the Valla Field Block, Unst, Shetland Islands. Transactions of the Royal Society of Edinburgh, Vol..59, 195–221.

Read, H H. 1942. Rocks showing felspar porphyroblasts and nodes from Unst, Shetlands. Proceedings of the Geologists’ Association. Vol..53, 107.

Rickwood, P C. 1989. Boundary lines within petrologic diagrams which use oxides of major and minor elements. Lithos, Vol..22, 247–263.

Ritchie, J E. 1930. 81–84 in Summary of Progress for the year 1929. Geological Survey of Great Britain. (London: His Majesty’s Stationery Office.)

Rivington, J B. 1953. Recent chromite exploration in Shetland. Mining Magazine, Vol..88, 329–337.

Robertson, T. 1935. The glaciation of Aithsting, South Nesting, Whalsay and the Out Skerries. Geological Survey Records (Unpublished).

Rock, N M S. 1991. Lamprophyres. (Glasgow: Blackie.)

Roddam, D, Miller, J A, and Flinn, D. 1994. Radiometric ages. 90–92 in Geology of Yell and some neighbouring islands in Shetland. Memoir of the British Geological Survey, Sheet 130 (Scotland).

Rushton, D R A. 1973. The serpentine minerals of Unst. Unpublished MSc dissertation, University of Manchester.

Sandison, Alex and Sons. Personal communications from Stewart Owers. Unpublished archives, Alex Sandison and Sons.

Sandison, C G D. 1936. The chrome ore of Unst. (Baltasound, Unst, Shetland: Alex. Sandison and Sons Ltd.) (Unpublished)

Sandison, C G D. 1948. Historical account of the mineral workings in Unst, Shetland. Scottish Council (Development and Industry), Paper No..49 (1) (Unpublished).

Schroeder, T, John, B, and Frost, B R. 2002. Geologic implications of seawater circulation through peridotite exposed at slow-spreading mid-ocean ridges. Geology, Vol..30, 367–370.

Shervais, J W, Kolesar, P, and Andreasen, K. 2005. A field and chemical study of serpentinisation — Stonyford, California: chemical flux and mass balance. International Geology Review, Vol..47, 1–23.

Slingsby, D R. 1981. Britain’s most northerly desert. (Lerwick: Shetland Life.)

Smith, D E. 1993. Norwick, Burragarth and Sullum Voe. 51–56 in The Quaternary of Shetland: Field Guide. Birnie, J F, Gordon, J E, Bennett, K J, and Hall, A M (editors). (Cambridge: Quaternary Research Association.)

Snelling, N J. 1958. Further data on the petrology of the Saxa Vord schists on Unst, Shetland Isles. Geological Magazine, Vol..95, 50–56.

Snelling, N J. 1963. in Annual Report of the Overseas Geological Survey for 1962. (London: Her Majesty’s Stationery Office.)

Spray, J G. 1988. Thrust related metamorphism beneath the Shetland Islands oceanic fragment, north east Scotland. Canadian Journal of Earth Sciences, Vol..25, 1760–1776.

Spray, J G, and Dunning, G R. 1991. A U/Pb age for the Shetland Islands oceanic fragment, Scottish Caledonides: evidence from anatectic plagiogranites in ‘layer 3’ shear zones. Geological Magazine, Vol..128, 667–671.

Stoker, M S, and Holmes, R.  1991. Submarine end-moraines as indicators of Pleistocene ice-limits off north-west Britain. Journal of the Geological Society of London, Vol..148, 431–434.

Sutherland, D G, and Gordon, J E. 1993. Shetland Islands: Introduction. 51–54 in The Quaternary of Scotland. Sutherland. D G, and Gordon, J E (editors). Geological Conservation Review Series, No..6. (London: Chapman and Hall.)

Tarkian, M, and Prichard, H M. 1987. Irarsite–hollingworthite solid solution series and other associated Os, Ir and Rh bearing PGM’s from the Shetland ophiolite complex. Mineralium Deposita, Vol..22, 178–184.

Tarkian, M, Economou-Eliopoulos, X, and Sambanis, G. 1996. Platinum-group minerals in chromitites from the Pindos ophiolite complex, Greece. Neues Jahrbuecher fur Mineralogic Monatshefte, Vol..10, 145–160.

Taylor, G K. 1984. Geophysical investigations of a Caledonian Ophiolite Complex, NE Shetland. Unpublished PhD thesis, University of Liverpool.

Thayer, T P. 1969. Gravity differentiation and magmatic re-emplacement of podiform chromite deposits. 132–146 in Magmatic ore deposits, Wilson, H D B (editor). Economic Geology Monograph, No..4.

Thomas, C H. 1980. Analysis of rock samples from the Shetland Ophiolite belt. Analytical Chemistry Unit, Institute of Geological Sciences, Report No..113.

Tomter, A. 1958. Kame bog. Shetland. Department of Agriculture for Scotland Moss Survey, Reports No..25.

Tully, M C, and Donato, J A. 1985. 1:1.000.000 northern North Sea Bouguer anomaly gravity map. British Geological Survey Report, Vol. 16, No..6. (London: Her Majesty’s Stationery Office.)

Walker, R J, Prichard, H M, Ishiwatari, A, and Pimentel, M. 2002. The Osmium isotopic composition of the convecting upper mantle deduced from ophiolite chromites. Geochimica et Cosmochimica Acta, Vol..66, 329–345.

Williams, H, and Smyth, W R. 1973. Metamorphic aureoles beneath ophiolite suites and Alpine peridotites: tectonic implications with west Newfoundland examples. American Journal of Science, Vol..273, 594–621.

Wilson, G V. 1931. In Summary of Progress for the year 1930. Geological Survey of Great Britain. (London: His Majesty’s Stationery Office.)

Zhou, M F, Malpas, J, Robinson, P T, Sun, M, and Li, J W. 2001. Crystallisation of podiform chromitites from silicate magmas and the formation of nodular textures. Resource Geology, Vol..51, 1–6.

Appendix 1 Radiometric age determinations

Published K-Ar age determinations of rocks and minerals from Unst and Fetlar

• From Miller and Flinn (1966) — updated
• 28441 muscovite in Valla Field Schist on Libbers Hill, Unst [HP 586 142] 486 ± 30 Ma
• 28459 biotitic rock from zoned ultramafic ball in Westing high strain zone, Quoy [HP 567 066] 431 ± 14 Ma
• 28442 biotitic rock from zoned ultramafic ball in Westing high strain zone Houllan Ness [HP 571 051] 403 ± 10 Ma
• 14496 muscovite (with 2% chlorite and chloritoid) in Snarravoe Pelite, Snarra Voe [HP 564 021] 417 ± 8 Ma
• 28448 whole rock (muscovite + 70% quartz + chlorite + opaque) Muness Phyllite at Clivocast [HP 610 002] 379 ± 9 Ma, 396 ± 6 Ma
• From Snelling (1963)
• Whole rock Muness Phyllite (muscovite) at Muness [HP 633 014] 432 ± 30 Ma
• Muscovite in Valla Field Schist at Tonga [HP 586 146] 427 ± 30 Ma
• Biotite in Skaw Granite at Lamba Ness [HP 663 157] 362 ± 12 Ma
• From Spray (1988)
• LW1 hornblende in NHS — type 2 at Loch of Watlee [HP 593 048] 465 ± 6 Ma, 465 ± 6 Ma
• VB10 hornblende in NHS — type 3 at Houbie, Fetlar [HU 622 905] 471 ± 3 Ma
• VB17 hornblende in NHS — type 2 at Feal, Fetlar [HU 626 905] 473 ± 6 Ma, 476 ± 6 Ma
• HB2 hornblende in NHS — type 3 at Virva, Fetlar [HU 645 921] 479 ± 6 Ma

Published ⁴⁰Ar-³⁹Ar age determinations

• From Flinn et al. (1991)
• 66123 hornblende from NHS — type 3 Virva, Fetlar [HU 645 921] 498 ± 2 Ma.
• From Flinn (1994a)
• (S92154) hornblende from Hascosay high strain zone Migga Ness, Yell [HP 539 053] 496 ± 6 Ma
• (S94728) biotite from Hascosay high strain zone Migga Ness, Yell [HP 5400 0522] 436 ± 7 Ma
• From Flinn and Oglethorpe (2005)
• 66236 K-feldspar phenocryst from Skaw Granite, north coast [HP 660 170] 407.2 ± 1.4 Ma
• 75438 K-feldspar phenocryst from Skaw Granite, Virdik [HP 653 174] 406.5 ± 0.9 Ma
• 77252 muscovite from thermal contact of Skaw Granite [HP 6528 1656] 425.8 ± 1.7 Ma
• 77362 muscovite from pelitic phyllite near Skaw Granite [HP 6525 1485] 427.0 ± 1.9 Ma
• 73770 muscovite from pelitic phyllite 4 km from Skaw Granite [HP 6205 1291] 419.5 ± 0.7 Ma

Published ²⁰⁶U-²³⁸Pb age

• From Spray and Dunning (1991)
• SH26 zircons from ‘plagiogranite’ South Ship Geo [HP 624 043] 492 ± 3 Ma
• Published radiocarbon dates of peat
• From Flinn (1996a)
• SRR 3089 (East Kilbride) peat below diatomite, Loch of Watlee [HP 5986 0516] 1570 ± 40 BP
• SRR 3090 (East Kilbride) peat below diatomite, Uyea Sound [HP 5931 0259] 2380 ± 40 BP

Figures, plates and tables

Figures

(Figure 1) Physiographical and locality map of the district. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 2) a) Simplified map of the Valla Field and Lamb Hoga blocks: 3 km of dextral offset on the Bluemull Sound Fault has been restored b) Diagrammatic cross-section through the Valla Field and Lamb Hoga blocks–the recumbent anticlinal Valla Field Fold.

(Figure 3) Tectonic blocks forming the Unst and Fetlar district.

(Figure 4) Diagrammatic section through the Shetland Ophiolite-complex, parallel to the restored Lamb Hoga Fault.

(Figure 5) Idealised pseudo-stratigraphical section through the Shetland Ophiolite-complex.

(Figure 6) Simplified map of the Shetland Ophiolite-complex.

(Figure 7) a) Geology of the Valla Field and Lamb Hoga blocks (see footnote 2 in Chapter 2). b) Diagrammatic structural succession in the Valla Field and Lamb Hoga blocks and key to map. c) Diagrammatic section through the antiformal Valla Field Fold.

(Figure 8) Distribution of gneisses in the Valla Field and Lamb Hoga blocks.

(Figure 9) Modal plots of plagioclase microporphyroblasts v. total plagioclase in plagioclase microporphyroblast gneisses.

(Figure 10) [Al₂O₃]–[FeO]–[MgO] + [CaO] + [K₂O] + [Na₂O] plot of pelitic rocks and minerals in rocks in the Valla Field and Lamb Hoga blocks: plot design after Khoo (1974); rock data from Key (1972) and Aziz (1984); and mineral data from Aziz (1984).

(Figure 11) Geology of the Saxa Vord Block.

(Figure 12) [Al₂O₃]–[FeO]–[MgO] + [CaO] + [K₂O] + [Na₂O] plot of pelitic rocks and minerals in the Saxa Vord Block. Plot design after Khoo (1974); rock data from Key (1972). Mineral data from Khoo (1974) and Flinn.

(Figure 13) Structure and stereoplots of the Valla Field and Lamb Hoga blocks; equal area projection, lower hemisphere.

(Figure 14) Phyllonitisation of pelites in the Valla Field Block.

(Figure 15) Stereoplot of structures in the Saxa Vord Block; equal area projection, lower hemisphere. a) Queyhouse Flags–S-tectonite. b) Saxa Vord Pelite and Hevda Phyllite–L-tectonite.

(Figure 16) Mineral data from the Valla Field Schist. a) MgO/(MgO + TFeO)–MnO plot of garnets 204 analyses from 30 thin sections (analyses from Aziz, 1984). b) TFeO–MgO plot of biotites: 109 analyses from 23 thin sections. TFeO total iron as FeO (analyses from Aziz, 1984).

(Figure 17) Metamorphic zones in the pelitic rocks of the Saxa Vord Block. After Flinn et al., (1996).

(Figure 18) Distribution of lithological units in the ophiolite nappes.

(Figure 19) Plot of TFeO–Al₂O₃ in serpentine minerals and chlorite. Symbols as on the 1:50 000 map (BGS, 2002).

(Figure 20) Cr# v. Mg# plot of electron microprobe analyses of accessory (disseminated) chromite, with fields for blackened rims and translucent cores: 166 analyses from 43 localities; Cr# = Cr/(Cr + Al), Mg# = Mg/(Mg + Fe²⁺). Comparative fields for massive (‘podiform’) and disseminated (accessory) chromite are after Gass et al. (1982).

(Figure 21) a) Stereoplot of the banding in the Metagabbro Layer: lower hemisphere. b) Distribution of banding and healed breccias in the Metagabbro Layer.

(Figure 22) Geochemistry of the Metagabbro Layer. Trace element compositional data shown on selected tectonomagmatic discrimination diagrams (after Pearce, 1982). WIP–within plate; IA–island arc; WPB–within plate basalts; VAB–volcanic arc basalt; MORB–mid-ocean- ridge-basalt. Spidergram profiles are averages of all available analyses.

(Figure 23) Distribution of clinopyroxenite-wehrlite bodies and clinopyroxene-bearing metadunite.

(Figure 24) Distribution of quasi-sheeted dykes.

(Figure 25) a–d Geochemistry of quasi-sheeted dykes: caption as for (Figure 22).

(Figure 26) Distribution of ‘plagiogranite’.

(Figure 27) Distribution of the imbricate zones and their constituent rocks.

(Figure 28) Distribution of the Gruting Greenschist.

(Figure 29) Geochemistry of the Gruting Greenschist.

(Figure 30) a) Structural map of the Funzie Conglomerate. b) Tectonite fabric variation at 35 sampling stations (shown in (Figure 30)a) in the Funzie Conglomerate.

(Figure 31) Distribution of the Norwick Hornblendic Schist.

(Figure 32) Geochemistry of the Norwick Hornblendic Schist. Trace element compositional data shown on selected tectonomagmatic discrimination diagrams (after Pearce, 1982). WIP - within plate; IA - island arc; WPB - within plate basalts; VAB - volcanic arc basalt; MORB - mid-ocean-ridge basalt. Spidergram profiles are averages of all available analyses.

(Figure 33) Interpretive exposure map of the Lower Imbricate Zone at the Loch of Watlee.

(Figure 34) Interpretive exposure map of the Middle Imbricate Zone.

(Figure 35) Geology of the Skaw Granite.

(Figure 36) Modal composition of granites in Unst and Fetlar. Fields: I granite; II granodiorite; III tonalite; IV quartz-diorite.

(Figure 37) Al₂O₃/(Na₂O + K₂O) vs. Al₂O₃/(CaO + Na₂O + K₂O) plot for the Skaw, Petester and Tonga granites, and leucotonalite pebbles in the Funzie Conglomerate. Subdivisions after Maniar and Piccoli (1989).

(Figure 38) K₂O vs. SiO₂ plot for the Skaw, Petester and Tonga granites, leucotonalite pebbles in the Funzie Conglomerate and ‘plagiogranite’. Subdivisions after Rickwood (1989).

(Figure 39) Structure of the Skaw Granite: equal area projection, lower hemisphere. a) Quartz veins in the Skaw Granite b) Skaw Granite

(Figure 40) Distribution of lamprophyres in Unst and Fetlar; all except three exposed in the cliffs.

(Figure 41) Diagrammatic sections through the Shetland Ophiolite-complex.

(Figure 42) Geological and geophysical map of the Shetland Intracontinental Basin. NEXTMAP Britain elevation data from Intergraph Technologies.

(Figure 43) The Shetland monadnock. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 44) Coastal features of the district. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 45) Indicators of ice flow in Unst and Fetlar. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 46) Glacial, periglacial and glacitectonic features in Unst and Fetlar. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 47) The impact of ice flow on stream sediment geochemical composition in Unst. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 48) The impact of ice flow on heavy mineral distribution in stream sediments in Unst. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 49) a) Subglacial and glacial channels b) Milldale Glacial Lake. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 50) Distribution of alluvial diatomite in Unst. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 51) Distribution of travertine deposits in Unst and Fetlar. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 52) Bouguer gravity anomaly map of Unst and Fetlar. Base level referred to the National Gravity Reference Net 1973. Based on data from Flinn (unpublished), Taylor (1984) and the BGS database. Dots are gravity stations. Contours at 1 mGal intervals. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 53) Aeromagnetic map of Unst and Fetlar. A recontoured version of the original 1:63 360 plotting sheet for the published 1:250 000 Aeromagnetic Map (IGS, 1968b). Dashed lines are flight lines. Contour interval above zero 100 nT. Contour interval below zero 50 nT. NEXTMAP Britain elevation data from Intermap Technologies.

(Figure 55)." data-name="images/P937279.jpg">(Figure 54) Map of the Shetland Ophiolite showing the lithologies modelled and the locations of the profiles shown in (Figure 55).

(Figure 55) Geophysical modeling along profiles X, Y and Z in Unst. See (Figure 55)." data-name="images/P937279.jpg">(Figure 54) for profile locations and polygon rock units. For A, U, V–see text. Observed gravity and magnetic fields — dashed lines with dots depicting gravity stations. Calculated gravity and magnetic fields — solid lines. Numeric values refer to rock physical property values assigned to each polygon–density in Mg/m³ (upper) and magnetic susceptibility in SI Units (lower).

(Figure 56) Distribution of the main PGE localities in Unst showing an association with chromite, Ni, Cu and As: Chromitite with Pt and Pd Ni, Cu and As

(Figure 57) Schematic section through the stratigraphy of the Shetland Ophiolite-complex on Unst showing the positions of the PGE concentrations. (modified from Prichard et al., 1994).

(Figure 58) Geology of the Cliff area [HP 608 112] showing the distribution of high values of PGE and associated elements (from Prichard and Lord, 1993).

(Table 20). Chondrite values used to plot the chondrite-normalised patterns are Os 514 ppb, Ir 540 ppb, Ru 690 ppb, Rh 200 ppb, Pt 1025 ppb and Pd 545 ppb after Naldrett and Duke (1980)." data-name="images/P937284.jpg">(Figure 59) Chondrite-normalised patterns of PGE for a chromite-rich metadunite from Cliff (MR4), a base metal sulphide-bearing metadunite from Cliff (MR11), a chromitite from Harold's Grave (RL070), a chromitite from Nikka Vord South (RLM003), a chromite-rich metadunite from Jimmies Quarry Hagdale (RLM007), a coarse clinopyroxenite from Ordale (NA25) and a high-level clinopyroxenite from mid-Unst (MR169). The PGE analyses of these samples are given in (Table 20). Chondrite values used to plot the chondrite-normalised patterns are Os 514 ppb, Ir 540 ppb, Ru 690 ppb, Rh 200 ppb, Pt 1025 ppb and Pd 545 ppb after Naldrett and Duke (1980).

(Figure 60) Schematic diagram illustrating the distribution of the PGM in chromite-rich PGE-rich rocks. Euhedral Os-, Ir- and Ru-bearing PGM occur within the unaltered part of the chromite grain. In the ferric chromite rim, associated with Cr-bearing chlorite and in the interstitial silicate host, Os-, Ir and Ru-bearing PGM are joined by Pt, Pd- and Rh-bearing PGM (from Prichard and Tarkian, 1988).

(Figure 61) Location of economic minerals. Source of the data listed in the text, and personal communications concerning gold from D A Rushton and H M Prichard.

Plates

(Front cover) Cover photograph Muckle Flugga and Out Stack (the end of the land) (P536015).

(Rear cover)

(Frontispiece) Loch of Watlee, Unst, looking north, [HP 5870 1450] (P536021).

(Dedication) Derek Flinn portrait

(Plate 1a) Gneisses: Boundary Zone. Plagioclase microporphyroblast gneiss, Westing Gneiss, Boundary Zone. South of Fill Geo, Hagdales Ness, Unst [HP 5768 0841] (P533820). Hammer handle 2.5 cm wide.

(Plate 1b) Gneisses: Boundary Zone. Micaceous leucosome gneiss, Valla Field Gneiss. Hevda Hill, Woodwick, Unst [HP 5805 1089] (P533819).

(Plate 1c) Gneisses: Boundary Zone. Psammitic granofelsic gneiss, Westing Gneiss. Snarra Voe, Belmont, Unst [HP 5626 0246] (P534936).

(Plate 1d) Gneisses: Boundary Zone. Hornblende agmatite hornblende gneiss, Boundary Zone. Loch of Bogliarths, Westing, Unst [HP 5780 0794] (P534948). Coin 2.5 cm diameter.

(Plate 1e) Gneisses: Boundary Zone. Laminated hornblende blastomylonite, ultramafic blastomylonite zone, Boundary Zone. Houllnan Ness, Westing, Unst [HP 5685 0560] (P534065). Coin 2 cm diameter.

(Plate 1f) Gneisses: Boundary Zone. K-feldspar megacryst augen gneiss; ultramafic blastomylonite zone, Boundary Zone. East of Gorsendi Geo, Lund, Unst [HP 5667 0430] (P534147). Hammer handle 2.5 cm wide.

(Plate 2a) Queyhouse Flags. Isoclinally folded quartzite beds. Leera Stack, Burra Firth, Unst [HP 627 175] (P533127).

(Plate 2b) Queyhouse Flags. Laminated and graded silty phyllites in closures of isoclinally folded single quartzite beds. South-east beach, Burra Firth, Unst [HP 6166 1418] ((P535997), (P535996)). Younging to the north (left). Coin 2.5 cm diameter.

(Plate 2c) Queyhouse Flags. Laminated and graded silty phyllites in closures of isoclinally folded single quartzite beds. South-east beach, Burra Firth, Unst [HP 6166 1418] ((P535997), (P535996)). Younging to the north (left). Coin 2.5 cm diameter.

(Plate 2d) Queyhouse Flags. Deformed flame structures in bedded silty phyllites. South-east beach, Burra Firth, Unst [HP 6175 1425] (P533382). Coin 2.5 cm diameter.

(Plate 3a) Chromite, metadunite and metaharzburgite. Banded metaharzburgite cut by metadunite sheet, 40 m north of Moho. Little Heog, Haroldswick, Unst [HP 6380 1118] (P533788).

(Plate 3b) Chromite, metadunite and metaharzburgite. Weathering colours of metadunite (foreground) and metaharzburgite (background). Rurra Geo, Clibberswick, Unst [HP 6600 1165] (P533605).

(Plate 3c) Chromite, metadunite and metaharzburgite. Chromitite; high grade ore, vein type. Chromite pit, Quoys, Unst [HP 6162 1198] (P550135). Bar 5 cm.

(Plate 3d) Chromite, metadunite and metaharzburgite. Vein of chromitite; high grade ore. South side of Wick of Hagdale, Unst [HP 6456 1026] (P550137).

(Plate 3e) Chromite, metadunite and metaharzburgite. Cumulate-type ore; low grade ore. Chromite pit, Cliff, Unst [HP 6080 1105] (P533643). f) Cumulate-type ore; low grade ore. Chromite pit, Mid Brake, Unst [HP 6363 0947] (P550136). Bar 5 cm.

(Plate 4a) Lower Metagabbro, Upper Metagabbro, Metadunite Layer and ‘plagiogranite’. Banded metagabbro; Lower Metagabbro. South of Skeo Taing, Balta Sound, Unst [HP 6480 0750] (P534020). Hammer head 15 cm long.

(Plate 4b) Lower Metagabbro, Upper Metagabbro, Metadunite Layer and ‘plagiogranite’. Brecciated-healed metagabbro; Lower Metagabbro. Sandwick, Unst [HP 6218 0350] (P534935). Hammer head 15 cm long.

(Plate 4c) Lower Metagabbro, Upper Metagabbro, Metadunite Layer and ‘plagiogranite’ Cumulate banded pyroxenite-wehrlite-metadunite, Metadunite Layer. Skeo Taing, Balta Sound, Unst [HP 6437 0837] (P533879). Hammer handle 3 cm wide.

(Plate 4d) Lower Metagabbro, Upper Metagabbro, Metadunite Layer and ‘plagiogranite’ Pyroxenite-wehrlite xenolith in metagabbro; Lower Metagabbro. Hill of Colvadale, Unst [HP 6223 0607] (P534091). Hammer head 14 cm long.

(Plate 4e) Lower Metagabbro, Upper Metagabbro, Metadunite Layer and ‘plagiogranite’.Quasi-sheeted dykes; Upper Metagabbro. Qui Ness, Unst [HP 622 032] (P534959). Hammer head 14 cm long.

(Plate 4f) Lower Metagabbro, Upper Metagabbro, Metadunite Layer and ‘plagiogranite’. ‘plagiogranite’ vein (dated). South Ship Geo, Colvadale, Unst [HP 623 043] (P536030). Hammer head 14 cm long.

(Plate 5a) Muness phyllites and conglomerates. S₀ lamination with S1 cleavage; Muness Phyllite. Pallaberg, Ore Wick, Muness, Unst [HP 6041 0377] (P534320). Hammer head 14 cm long.

(Plate 5b) Muness phyllites and conglomerates. Folds of S₀ overturned to the west; Muness Phyllite. Tooa Stack, Ore Wick, Muness, Unst [HP 6068 0033] (P534323). Hammer head 14 cm long.

(Plate 5c) Muness phyllites and conglomerates. Ramnageo Conglomerate (L-tectonite–X>Y=Z) seen parallel to elongation; Muness Phyllite. Ramnageo Ness, Muness, Unst [HU 624 997] (P534952). Hammer handle 3cm wide.

(Plate 5d) Muness phyllites and conglomerates. Ramnageo Conglomerate (L-tectonite–X>Y=Z) seen normal to elongation; Muness Phyllite. Ramnageo Ness, Muness, Unst [HU 624 997] (P534950). Hammer head 3 cm wide.

(Plate 5e) Muness phyllites and conglomerates. Funzie Conglomerate (XZ plane shown). The Snap, Funzie Ness, Fetlar [HU 658 879] (P535973). Coin 2.5 cm diameter.

(Plate 5f) Muness phyllites and conglomerates. Funzie Conglomerate (XY plane shown). The Snap, Funzie Ness, Fetlar [HU 658 879] (P535967). Coin 2.5 cm diameter.

(Plate 6a) Skaw Granite. Contact zone of granite with subequant phenocrysts, Virdik, Hill Ness, Skaw, Unst [HP 6540 1740] (P533283). Hammer head 14 cm long.

(Plate 6b) Skaw Granite. Tabular phenocrysts of microcline (L-tectonite) seen normal to lineation, Beach at Inner Skaw, Unst [HP 6633 1585] (P550140). Coin 2.5 cm diameter.

(Plate 6c) Skaw Granite. Tabular phenocrysts of microcline (L-tectonite) seen parallel to lineation, Beach at Inner Skaw, Unst [HP 6633 1585] (P550139). Coin 2.5 cm diameter.

(Plate 6d) Skaw Granite. Cognate xenolith with K-feldspar phenocrysts, Beach at Inner Skaw, Unst [HP 6632 1584] (P550138). Coin 2.5 cm diameter.

Plate 7 Spinifex texture in serpentinite breccia; Dunrossness Spilites, Cunningsburgh [HP 4261 2744] (P550134). Coin 2.5 cm diameter.

(Plate 8a) Coastal geomorphology. Outer Coast; cliffs plunge steeply to surrounding seafloor at about 82 m. Hill of Clibberswick, Unst [HP 6500 1400] ( P533485).

(Plate 8b) Coastal geomorphology. Inner Coast; drowned landscape lacking cliffs. Balta Sound, Unst [HP 6200 0700] (P533926).

(Plate 8c) Coastal geomorphology. Boulder tombolo connecting the holm to Unst. Holm of Heogland, Belmont, Unst [HU 5700 9900] (P534397).

(Plate 8d) Coastal geomorphology. Mid-bay bar cutting off the fresh water from the sea, Loch of Cliff, Unst [HP 6270 1700] (P533216).

(Plate 8e) Coastal geomorphology. Typical geo. Rurra Geo, Clibberswick, Unst [HP 6620 1150] (P534939).

(Plate 8f) Coastal geomorphology. Serpentinite-pebble beach cemented by aragonite. Wick of Hagdale, Baltasound, Unst [HP 6436 1047] (P533715). Hammer head 14 cm long.

(Plate 9a) Glacial geomorphology and related features. Ice thrust in bedrock. Belmont, Unst [HP 5587 0046] (P534957). Hammer 35 cm long.

(Plate 9b) Glacial geomorphology and related features. Boulder field of frost-shattered metagabbro. Virda Field, Baltasound, Unst [HP 6200 0600] (P534960).

(Plate 9c) Glacial geomorphology and related features. Subglacial drainage channels. Herma Ness Lighthouse Station, Unst [HP 6120 1490] (P533407).

(Plate 9d) Glacial geomorphology and related features. Overflow channel from Milldale Glacial Lake. Cat Houll, Tonga, Unst [HP 5910 1470] (P533349).

(Plate 9e) Glacial geomorphology and related features. Abandoned glacial drainage channel with small deposit of alluvial diatomite. Loch of Watlee, Unst [HP 5985 0516] (P534944). Site of ¹⁴C date.

(Plate 9f) Glacial geomorphology and related features. Alluvial diatomite in the same channel shown in e. Loch of Watlee, Unst [HP 5985 0516] (P534961).

(Plate 10) Photomicrographs of PGM in reflected light showing the different mineralogical associations. Photomicrographs: H M Prichard. a.Largest laurite crystal (pale grey, Ru) located to date in Unst, within chromite (grey, Cr), crossed by a network of altered silicate-filled veins (dark grey, Si) and containing inclusions (see close up of square in (Plate 10b)). Harold’s Grave [HP 630 113] Bar represents 50 microns. b. Close up of Plate 10a showing a lath-shaped inclusion of a mixture of native osmium (pale grey, Os) and irarsite (grey, Ir) within laurite (dark grey, Ru) with a pure composition of RuS₂ and excluding Ir and Os. c. Potarite (white, Pot) on the edge of a chromite grain (grey, Cr) adjacent to silicate (dark grey, Si) from the extremely enriched PGE chromite-bearing dunite. Cliff [HP 608 112]. Bar represents 10 microns. d. Euhedral laurite crystal (white, Ru) enclosed in a chromite grain (grey, Cr) enclosed in silicate (dark grey, Si). Cliff [HP 608 112]. Bar represents 20 microns. e. PGM located in rows within a clinopyroxene crystal (grey, Cpx) and in altered silicate matrix (dark grey, Si) which infills cleavage and also contains magnetite (Mt), all in a wehrlite. PGM (white) are too small for quantitative analysis but were analysed qualitatively on a LEO 360 SEM at Cardiff University using an Oxford instruments INCA EDX system. These PGM are rarely entirely enclosed in clinopyroxene but in these cases are predominantly sulphides, whereas PGM in contact with the altered silicate are Pt- and Pd-arsenides, antimonides, tellurides and alloys. PGM (a-p) are: (a) Pd-Cu-Sn-Pb-Sb-arsenide, (b) Pd-Ni-Cu-Sn-arsenide, (c) Pd antimonide enclosed in a Ni-arsenide, (d) Pd-arsenide, (e) Pd-Cu-arsenide, (f) Au-Cu-alloy enclosed in a Pd-Pb-alloy adjacent to a Pd-arsenide, (g) Pd-Cu-Pb-Sb- arsenide, (h) Pd-Pb-alloy, (i) Pd-arsenide, (j) a composite grain of Pd-Pt-arsenide and Pd-Cu-Te-arsenide, (k) Pd-Cu- telluride enclosed in a Cu-Fe-sulphide, (l) composite PGM entirely enclosed in clinopyroxene and shown in Plate 10f, (m) Pd-Pb-telluride, (n) Pd-Pt-Te-arsenide adjacent to a grain of magnetite, (o) Pd-Cu-Sb-arsenide and (p) Pd-Cu-Sb- arsenide. South Unst. Bar represents 10 microns. f. Close up of Plate 10e showing a sub-rounded bleb of a composite grain composed of (1) Pd-Cu-Fe-Ni-sulphide, (2) Pd-Au-Cu-Fe-Ni-sulphide, (3) Pd-bearing Fe-Ni-Cu-sulphide and (4) Pd-bearing Cu-Fe-Ni-sulphide, all in enclosed in clinopyroxene. Bar represents 1 micron. g. Pt-Pd-Au-Cu-alloy (white, Pt) adjacent to native Cu (pale grey, Cu) in altered silicate (dark grey, Si) in a system of veins across chromite (grey, Cr) in the extremely enriched PGE chromite-bearing dunite. Cliff [HP 608 112]. Bar represents 1 micron. h. PGM from the extremely PGE-enriched lithology at Cliff [HP 608 112]. The PGM comprise a composite grain of sperrylite (PtAs2) and stibiopalladinite (white, Pd) and a grain of sperrylite (white) in altered silicate (dark grey, Si) near the edge of a chromite grain (grey, Cr), which has a paler grey more Fe-Cr-rich altered rim. The PGM lie in serpentine but the sperrylite grain is altered to a Pt-Ni oxide (pale grey, Ox) with a mesh texture where the sperrylite is cross cut by a vein of chlorite (pale grey, Cl). Bar represents 20 microns.

(Plate 11) Abandoned quarry exploiting a chromitite ‘podiform’ deposit in the Metadunite Layer. Nikka Vord, Baltasound, Unst [HP 6285 1033]. (P534932).

Tables

(Table 1) Analyses of the Tonga Granite

(Table 2) Analyses of metaharzburgite

(Table 3) Mg number (Mg#) in units of the Shetland Ophiolite- complex

(Table 4) Analyses of metadunite in the Metaharzburgite Layer

(Table 5) Analyses of metadunite in the Metadunite Layer

(Table 6) Analyses of the Lower Metagabbro

(Table 7) Analyses of the Upper Metagabbro

(Table 8) Analyses of clinopyroxenite-wehrlite bodies in the Metadunite Layer

(Table 9) Analyses of clinopyroxenite-wehrlite bodies in the Metagabbro Layer

(Table 10) Analyses of ‘plagiogranite’

(Table 11) Analyses of rocks of the Unst Phyllite Group

(Table 12) Analyses of the alkaline Gruting Greenschist

(Table 13) Analyses of the subalkaline Gruting Greenschist

(Table 14) Analyses of granite pebbles within the Funzie Conglomerate

(Table 15) Analyses of the Skaw Granite

(Table 16) Analyses of the Petester Granite

(Table 17) Analyses of lamprophyric rocks

(Table 18) Grain size and composition of beach sands

(Table 19) Rock densities (Mg/m3) for selected rock units used in geophysical modelling

(Table 20) Analyses of PGE and associated elements in represent- ative ophiolite lithologies

(Table 21) Selected typical analyses of PGM

(Table 22) Analyses of steatite

Tables

(Table 1) Analyses of the Tonga Granite

Analyses from Gamil (1991) and unpublished

(Table 2) Analyses of metaharzburghite

• N1–number of 'hydrous' analyses; N2–number of anhydrous analyses Mg# = MgO/(MgO + TFeO); TFeO = total Fe as FeO; TFe₂O₃=total Fe as Fe₂O₃ Analyses from BGS (via Gunn et al., 1985), Moffat (1987), Thomas (1980), Gass et al. (1982) and Flinn, unpublished.

(Table 3) Mg number (Mg#) in units of the Shetland Ophiolite-complex

• Mg number [Mg # = MgO/(MgO + TFeO); TFeO=total Fe as FeO

(Table 4) Analyses of metadunite in the Metaharzburgite Layer

• Mg# = MgO/(MgO + TFeO); TFeO=total Fe as FeO; TFe₂O₃=total Fe as Fe₂O₃
• N–number of analyses; N1–hydrous analyses; N2–anhydrous analyses
• Analyses from BGS.

(Table 5) Analyses of metadunite in the Metadunite Layer

• Mg# = MgO/(MgO + TFeO); TFeO=total Fe as FeO; TFe₂O₃=total Fe as Fe₂O₃
• N–number of analyses; N1–hydrous analyses; N2–anhydrous analyses
• Analyses from BGS (via Gunn et al., 1985), Moffat (1987), Thomas (1980), Gass et al. (1982) and Flinn, unpublished.

(Table 6) Analyses of the Lower Metagabbro.

• Mg# = MgO/(MgO + TFeO); TFeO=total Fe as FeO; TFe₂O₃=total Fe as Fe₂O₃
• N = number of analyses
• Analyses from BGS (via Gunn et al., 1985), Thomas (1980), Gass et al. (1982), Spray and Dunning (1991) and Flinn, unpublished.

(Table 7) Analyses of the Upper Metagabbro

• Mg# = MgO/(MgO + TFeO); TFeO=total Fe as FeO; TFe₂O₃=total Fe as Fe₂O₃ N = number of analyses
• Analyses from BGS (via Gunn et al., 1985), Thomas (1980), Gass et al. (1982), Spray and Dunning (1991) and Flinn, unpublished.

(Table 8) Analyses of clinopyroxenite-wehrlite bodies in the Metadunite Layer

• Mg# = MgO/(MgO + TFeO); TFeO=total Fe as FeO; TFe₂O₃=total Fe as Fe₂O₃
• N = number of analyses - NI - hydrous analyses; N2 - anhydrous analyses
• Analyses from BGS (via Gunn et al., 1985), Moffat (1987), Thomas (1980), Gass et al., (1982) and Flinn, unpublished

(Table 9) Analyses of clinopyroxenite-wehrlite bodies in the Metagabbro Layer

• Mg# = MgO/(MgO + TFeO); TFeO=total Fe as FeO; TFe₂O₃=total Fe as Fe₂O₃
• N = number of analyses – NI – hydrous analyses; N2 – anhydrous analyses
• Analyses from BGS (via Gunn et al., 1985), Moffat (1987), Thomas (1980), Gass et al. (1982) and Flinn, unpublished.

(Table 10) Analyses of ‘plagiogranite’

• N = number of analyses
• Analyses from BGS (via Gunn et al., 1985), Thomas (1980), Gass et al. (1982), Spray and Dunning (1991) and Flinn, unpublished.

(Table 11) Analyses of rocks of the Unst Phyllite Group

• 1 Muness Phyllite — impure quartzite — Belmont
• 2 Muness Phyllite — semipelite — Muness
• 3 Muness Phyllite — laminated quartzite/pelite — Muness
• 4 Norwick Graphitic Schist — north-east Fetlar
• 5 Muness Phyllite — pelite — Ness of Gruting, Fetlar
• 6 Leagarth Pelite — chloritoid pelite — Aith, Fetlar
• 7 Muness Phyllite — albite schist — north-west Fetlar
• Analyses from Flinn unpublished
• TFe₂O₃ = total Fe as Fe₂O₃

(Table 12) Analyses of the alkaline Gruting Greenschist

• Analyses from Flinn, unpublished
• N = number of analyses
• TFe₂O₃ = total Fe as Fe₂O₃

(Table 13) Analyses of the subalkaline Gruting Greenschist

• Analyses from Flinn, unpublished
• N = number of analyses
• TFe₂O₃ = total Fe as Fe₂O₃

(Table 14) Analyses of granite pebbles within the Funzie Conglomerate

(Table 15) Analyses of the Skaw Granite

• N = number of analyses
• Analyses from Gamil (1991), Flinn (unpublished)

(Table 16) Analysis of the Petester Granite

Analysis from Gamil (1991)

(Table 17) Analyses of lamprophyric rocks

• 1 Spherulite
• 2 Carbonate-bearing spherulite
• 3 Kersantite
• 4 Spessartite
• 5 Spessartite
• 6 Plagioclase-phyric spessartite
• 7 Biotite-hornblende microdiorite (malchite)
• 8 Biotite-hornblende diorite (appinitic)
• Mg# = MgO/(MgO + TFeO); TFeO=total Fe as FeO
• Analyses from Flinn (1994a) and unpublished

(Table 18) Grain size and composition of beach sands

• * Data from P A Alexander (1975)
• + Data from Garnet-Frizelle (1979)
• Remaining data from Nickson (1980)
• Data used only once

Unst Groups and Mainland and Yell data are based on one sample per beach, taken from the centre of the beach

(Table 19) Rock densities (Mg/m³) for selected rock units used in geophysical modelling (after Flinn, 2000)

• Metaharzburgite, metadunite, and wehrlite partially to completely lizardite serpentinised;
• * antigorite serpentinised

(Table 20) Analyses of PGE and associated elements in representative ophiolite lithologies

(Table 21) Selected typical analyses of PGM