Content and licensingview original scan buy a printed copy

Geology of Yell and some neighbouring islands in Shetland Memoir for 1:50 000 geological sheet 130 Yell (Scotland)

By Derek Flinn

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

British Geological Survey

Geology of Yell and some neighbouring islands in Shetland: Memoir for 1:50 000 geological sheet 130 Yell (Scotland)

Derek Flinn

• Contributors
• Age determinations D Roddam, J A Miller
• Stream sediment geochemistry D M A Flight, J A Plant
• Author
• D Flinn, DSc, PhD University of Liverpool
• Contributors
• D Roddam, PhD, MA J A Millar, ScD University of Cambridge
• D M A Flight, MSc, DIC British Geological Survey, Keyworth
• J A Plant, BSc, PhD, FIMM, CEng, FGS British Geological Survey, Keyworth

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.

© NERC copyright 1994 First published 1994. ISBN 0 11 884502 0 Printed in the UK for HMSO Dd 292045 C8 5/94 1679 13110

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

Books

• Memoirs
• The geology of Western Shetland 1976
• British Regional Geology Orkney and Shetland 1976

Maps

• 1:625 000
• United Kingdom (North Sheet) Solid Geology 1979
• Quaternary Geology 1977 Bouguer anomaly 1981
• Aeromagnetic anomaly 1972
• 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

Author's acknowledgements

The author is very grateful to the Unilever Trust for an Emeritus Fellowship (1986–1988) awarded him prior to the NERC contract, which enabled him to obtain the 120 rock analyses and 100 modes on which are based many of the interpretative results presented in the Memoir. Ile is also grateful to Drs M P Atherton, A S Gamil and D Moffat for permission to quote and make use of various unpublished rock and mineral analyses and to the Department of Geology, University of Manchester and Messrs T Hopkins and D Plant for permission to use, and assistance with using, their Geoscan Microprobe. Dr D I J Mallick and Dr F May of the British Geological Survey, Edinburgh provided very considerable assistance with the writing and presentation of the Memoir.

The section on Radiometric ages was written by D Roddam and J A Miller of the University of Cambridge and the author, and that on Stream sediment geochemistry by D M A Flight and J A Plant of BGS, Keyworth; these contributions are gratefully acknowledged.

Notes

• National Grid—The 1:50 000 Yell Sheet lies across the east–west boundary between National Grid squares HU arid 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 Yell Sheet.
• Specimen numbers—Specimen numbers prefixed with S belong to the collection housed in Murchison House, BGS, Edinburgh. Those prefixed with L belong to the University of Liverpool collection.
• Chemical analyses of rocks are by the Geochemical Laboratory of the University of Liverpool, except where stated otherwise. Microprobe analyses of minerals are by the author, except where stated otherwise.

Preface

This memoir, and the 1:50 000 scale map that it accompanies, are the product of a mapping contract between the Natural Environment Research Council and the University of Liverpool. Such contracts stem from NERC policy of encouraging academics with substantial knowledge about specific areas in the UK to transfer their information 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, are then published by the British Geological Survey. In the case of Yell, the contract built on the 40 years of experience of Shetland geology possessed by Professor Flinn, who had earlier surveyed the area of the solid and drift geological maps of Central Shetland (Sheet 128) published by the Institute of Geological Sciences (now BGS) in 1981 and 1982. This memoir contains the results of a large number of rock analyses obtained by Professor Flinn while holding a Leverhulme Emeritus Fellowship.

Yell is a sparsely populated island in the north of Shetland and is separated geologically from neighbouring Unst and Mainland by two major faults. Yell has been commonly regarded by geologists en route to the more varied geology of Unst as of little particular interest, being merely gneiss covered by peat. This attitude has been largely due to the absence of a good account of the geology, a deficiency that has now been remedied by Professor Flinn. He has shown by painstaking work, particularly along almost continuous coastal sections, that the geology of Yell is both interesting and instructive, and of more than just local interest.

Most of Yell consists of metasedimentary rocks correlated with the Moine of the Northern Highlands of Scotland, and the new work has shown that a pre-Moine crystalline basement, similar to the Lewisian, may be recognised, in spite of the subsequent deformation and metamorphism and the fact that even now the structure is not fully understood. Of particular importance is the recognition of the variety of gneisses produced, together with an new insight into the processes by which they have been formed. In addition, Professor Flinn has recogised a wide range of igneous rocks, some old and involved in the deformation and metamorphism that affected their host rocks, and others emplaced after the deformation which are relatively unaltered. The metamorphic and igneous rocks constitute part of the interior of the old Caledonian mountain chain which developed in late Precambrian and early Palaeozoic times.

The only younger deposits are of Quaternary age and are rather sparse products of the last (Devensian) glaciation, together with the Holocene beach and peat deposits. There are, however, a number of well-displayed glacial features, including those associated with the former ice limit in the north of the island. The morphology of the coastline is also of interest and provides evidence that the Devensian glaciation had a relatively minor effect on a much older landscape which has undergone a prolonged period of cliff erosion.

So far, little in the way of economic mineral deposits has been proved on Yell, although some of the rocks are usable as local road-stone and building aggregate, numerous pegmatites contain considerable resources of feldspar, and areas of blown sand have been exploited for construction purposes, while the peat has been used from time immemorial for fuel. However, it is the relatively unspoiled beauty of the island, and its attraction for the tourist industry, which is the island's greatest resource.

Peter J Cook, DSc Director British Geological Survey Kingsley Dunham Centre Keyworth Nottingham NG12 5GG November 1993

Geology of Yell and neigbouring islands—summary

Yell is an island in the north of Shetland, about 28 km long by 10 km wide. It is bounded by almost continuous cliffs up to 100 m high in which the geology of the island is magnificently displayed. The cliffs are interrupted in several places by sandy beaches up to 300 m long. Inland peat is extensive and exposures are limited.

The island is formed largely of Moine psammites interlayered with a number of lenticular bodies of mica schists and quartzites , together with a number of Lewisian inliers. Early in their metamorphism the Moine rocks were intruded by sills of basic rocks and several granitoid bodies. During the metamorphism the latter became orthogneisses and the metasedimentary rocks were widely but variably gneissified. The Moine rocks are bounded to the east by a remarkable band of microcline megacrysts which can be traced for 80 km along the east cost of Yell and across the Mainland of Shetland. East of this band, and exposed in the cliffs on the east side of Yell, are psammitic rocks and para-gneisses belonging to the Boundary Zone, a tectonic unit which, in Shetland, separates the Moine and Dalradian rocks.

Late in the metamorphism, the area was extensively injected by veins and masses of pegmatite and aplite and, locally in south-east Yell, by sheets of tonalite and globular bodies of ophitic metadolerite. At the end of the metamorphism and prior to 500 Ma, a zone, 250 to 500 m wide ' of hornblende-blastomylonite developed. This contains residual masses of hornblende gneiss, including Lewisian-inlier-like gneisses, and is extensively exposed in the cliffs of north-east Yell. Much later, proba bly in Mesozoic times, the rocks of Yell were intersected and offset by the three splays of the dextral Nesting Fault.

During the Late Devensian glaciation, the ice shed of the Shetland ice cap trended obliquely across the south of Yell, while the margin of the ice lay across the northwest of the island. Rising sea level following the deglaciation has caused the construction of a series of bars and tombolos, and peat now lies below sea level in the sheltered heads of some drowned valleys.

Geological Succession In Yell, Shetland

Chapter 1 Introduction

Area

The area covered by the 1:50 000 Yell, Shetland Sheet is the central part of the area covered by the one-inch-to-one-mile Northern Shetland Sheet (IGS, 1968a). It includes the island of Yell and many nearby small islands and half-tide rocks (baas). To the north, west and south of Yell the islands of Gloup Holm, the Holm of West Sandwick, Uynarey, Bigga, Sligga Skerrv, Samphrey, The Rumble, Orfasay, Green Holm and the Neapaback Skerries are included in the Sheet, and also the island of Hascosay (except. for its east coast) and several skerries and Baas east of Yell (Figure 1). Yell is about 28 km long and 10 km wide. It is a low-lying island rising exceptionally to 200 m above OD, and is ringed by an almost unbroken line of cliffs ranging up to 100 m in height. The island is mostly covered by hill peat and crofting lands. Exposure is nearly continuous in the cliffs but inland it is very variable; neighbouring inland exposures of rocks are often separated by more than a kilometre.

Previous work

Very little work had been done on the geology of Yell prior to the present study. In general, travellers of all types have passed through Yell, on their way to Unst and Fetlar, describing it in passing as a mass of gneiss covered in peat. In fact it is a very interesting island both geographically and geologically, but it requires close inspection in order fully to appreciate this.

Robert Jameson (1798) was the first geologist to visit Yell. He mentioned the gneiss and the peat and also 'several curious veins of granite' on the west side probably the pegmatites of Lumbister.

Hibbert, who surveyed the whole of Shetland in 1817 and 1818, devotes about one and a half pages to the geology of Yell in a 600 page book which includes a geological map of the whole of Shetland (Hibbert, 1822). He mentions the Yell gneisses, and the granite (i.e. pegmatite) injection zone of Lumbister, which he illustrated with several figures. He also refers to the hornblende gneisses of the south end of Yell (the Houlland Lewisian Inlier), the contact between the Graven Granodiorite (called sienite by him) and the gneisses on Bigga, and he mentions the large-biotite pegmatites of West Sandwick.

In 1879 both Heddle and Peach and Iforne published maps of Shetland. Heddle's map is a faithful copy of Hibbert's map on a modern base. Peach and Home also copied Hibbert's map onto a modern base but made a number of amendments, some of which are correct and some incorrect. Heddle's map accompanied a report on his mineralogical searches undertaken over many years in Shetland (Heddle, 1879), while Peach and Horne's map accompanied their survey of the glacial geology of Shetland (Peach and Home, 1879). Both maps show Yell and Hascosay as formed entirely of gneiss, as on Hibbert's map, but Heddle has added, incorrectly, a small limestone in West Sandwick, while Peach and Home have added incorrect glacial flowlines. Heddle characterises the island as being even more uninviting to the mineralogist than to the geologist. He mentions finding hornblende, epidote and garnet. His analysis of the garnet is very similar to garnet analyses obtained for this work.

Peach and Home's (1879) geological map of Shetland indicates ice flow across Shetland from east to west, in confirmation of Croll's (1870) prediction that, during the glaciation of northern Europe, ice from Scandinavia filled the North Sea and overflowed into the Atlantic over Shetland. This is a misconception of the last glaciation that still dominates glacial work in the area, for at that time Shetland had its own ice cap. C W Peach (1867) had correctly described striae indicating ice flow towards the north-east in the north-east corner of Unst, but he was later persuaded by his son to reverse this flow direction to south-west to fit in with Peach and Horne's (1879) glacial map of Shetland. In the Bluemull Sound area they show three striae pointing to the south-west (including one on Yell), although all the striae found in this area during the present survey point to the north-northwest. On the west side of Yell, just north of West Sand-wick, they correctly identified two north-west-pointing striae, but they failed to see that in the south of Yell the striae all point to the east.

The first significant attempt to investigate the geology of Yell after Hibbert was by J K Allen working fbr the Geological Survey in the years 1929, 1930 and 1931 and the account of his work in the Summaries of Progress for those years are the first modern account of the geology of the island. Nevertheless it is quite extraordinary how little his account adds to Hibbert's brief notes. He brings out the importance of garnet in the area, includes some nongneissic schists among the gneisses and names the Arisdale Quartzite. His clean-copy 1:10 560 maps are available in the offices of the British Geological Survey, Edinburgh.

In 1950 a manuscript map of Shetland at one inch to the mile was available in the Geological Survey, Edinburgh. It was probably drawn by G V Wilson in the years 1936, 1937 and 1938 and possibly later (see Summaries of Progress for those years), when he was working on the petrography of the islands, and surely with the co-operation of Allen. The map of Yell contains much detail, none of it meaningful except for the Arisdale Fault, the Arisdale Quartzite, the hornblendic part of the Houlland Inlier and the general trend of the rocks. Structurally, this map improves slightly on the Yell part of the map of schistosity trends throughout Shetland provided by Hibbert (1822) in what is possibly the world's first structural map.

J Phemister used the one-inch manuscript map in compiling the solid edition of the one-inch-to-one-mile North Shetland Sheet (IGS, 1968a). However, his improvements of the geology of Yell appear, from my 1950 handmade copy of the manuscript map, to be limited to the addition of minor intrusions (lamprophyres and pegmatites) , several microcline-impregnated areas, and hornblende-rich rocks in a number of places – most importantly the West Yell Lewisian Inlier and that part of the Hascosay Slide outcropping along the north-east coast of Yell.

Recent work

Work on Shetland as a whole, since the 1960s, has provided ground-magnetic, aeromagnetic, gravity and geochemical maps covering Yell (see Chapters 11 and 12).

Between 1975 and 1983 Flinn surveyed Yell on 1:10 560 scale, mapping on airphotos in inland areas, and subsequently carried out revision surveying up to 1988. A total of 400 days was spent in the field, some 1700 hand specimens were collected and thin-sectioned and more than 500 photographs taken. About 120 rocks were analysed for major and trace elements and point-counted (this work was paid for by the Leverhulme Trust) and about 450 microprobe analyses of minerals were made (Flinn, 1993). Ground magnetic surveying was unnecessary due to the lack of sufficiently pronounced ground-magnetic anomalies. The survey was clean copied onto a 1:10 000 base, and is available in the office of the British Geological Survey, Edinburgh. A 1:50 000 map of the area has been published (BGS, 199?-). A preliminary account of this work has been published (Flinn, 1988) and a more detailed account is available (Flinn, 1989). (Figure 1) provides a key to the 1:10 000 clean-copied sheets.

Outline of the geology

(Figure 2) shows the relationship of the 1:50 000 Yell sheet to the previously published, adjacent 1:63 360 sheets and it shows the continuation into the adjacent sheets of the geological units occurring in Yell.

Yell is composed predominantly of the Yell Sound Division of the East Mainland Succession of Shetland, a series of psammites, quartzites, schists and gneisses. These rocks have a northerly strike and generally dip steeply west, though in places on the east side of Yell the dip becomes much less steep. The schistose rocks and the quartzites are distributed through the succession in the form of large lenticular masses, the boundaries of which present no indications that they were originally of other than conformable sedimentary nature, though competence differences may have concentrated deformation at them. Nothing has so far been found which gives any indication of direction of younging in the division; it is conventionally described as if it youngs to the east. There is a mineral lineation plunging at 10–20° to the north-north-west parallel to the most common direction for fold axes. The folds are rather irregularly distributed and vary considerably in style. They do not lead to repetition in the succession. The succession also contains a number of lenticular masses of coarse hornblende gneiss and quartzofeldspathic gneiss which in many ways are comparable with the Lewisian inliers of the Moine of Scotland. These masses occur as integral components of the succession and provide no structural or other evidence relating to their origin or emplacement. Thin bands of garnet-studded hornblende schist occur widely. Also widely developed are patches and zones of gneissification in the metasedimentary rocks and the whole division appears to be of kyanite grade. The division continues, below the sea, to the west of the Yell Sheet and also to the southwest into the Central Shetland Sheet until truncated by the Walls Boundary Fault, a major transcurrent fault which last moved in Mesozoic times (Flinn, 1992). This division has been correlated with the Glenfinnan/Loch Eil Division of the Scottish Moine (Flinn, 1967; Flinn et al., 1972).

The Yell Sound Division is bounded to the east (not necessarily conformably) by a conspicuous band of microcline-megacryst porphyroblast gneiss, the Valayre Gneiss, which can be traced along the east coast of Yell and from thence to the south-west of Yell, through the Central Shetland Sheet to close to the Walls Boundary Fault, a distance of more than 80 km. In Yell, to the east of the Valayre Gneiss and forming the eastern cliffs are rocks of the Boundary Zone. This is a tectonic unit which in Shetland separates the Moine (Yell Sound Division) from the Dalradian (Scatsta Division of the East Mainland Succession). In Yell the rocks of the Boundary Zone are dominantly psammites which locally are rich in plagioclase microporphyroblasts and have been intruded by globular masses of ophitic metadolerite. Outwith Yell the Boundary Zone varies up to a kilometre or two wide and is composed of zones of kyanite-bearing schist, or micaceous gneiss and hornblende-banded anatectite gneiss. It is bounded to the east by an augen gneiss very different in appearance from the Valayre Gneiss. This association of disparate units, assembled at an early stage in the metamorphism of the area, appears to form a vertical schuppen-type zone separating the Moine and Dalradian of Shetland.

In the north-east of Yell the Boundary Zone is cut by the IIascosay Slide, a blastomylonite zone about 500 m thick with a north-north-west trend and a westerly dip in the north steepening to nearly vertical in its southernmost exposure in Hascosay. The blastomylonites are very fine-grained hornblende schists which locally contain neocrystalline clinopyroxene and garnet and they were formed prior to 500 Ma. They contain large residual masses of coarse hornblendic rocks, including hornblende gneiss of Lewisian-inlier type. The fabric of the slide zone is orthorhombic, includes a mineral lineation plunging gently to the north-north-west, and contains no indication that it is the result of thrusting. To the east of the Hascosay Slide are Dalradian schists and gneisses of the Valla Field and Lamb Hoga Blocks of Unst and Fetlar (Flinn, 1958).

In the north-eastern part of Yell, the rocks of the Yell Sound Division, the Valayre Gneiss, the Boundary Zone and the Hascosay Slide, which throughout most of their outcrop in Shetland are vertical or dip steeply west, have been folded so that they all dip at a low angle to the west. The rocks of Yell have been heavily injected by pegmatites and aplites and later by lamprophyres. The Graven Complex, an intrusive appinitic granodiorite about 400 Ma old, was emplaced after the lamprophyres. It extends into both the North Shetland Sheet and the Central Shetland Sheet, is offset by the Nesting Fault and is truncated to the west by the Walls Boundary Fault.

Within Yell, the Yell Sound Division, the Boundary Zone and the Hascosay Slide are all cut and dextrally offset by splays of the Nesting Fault which combine just south of Yell and continue south through the Central Shetland and the South Shetland Sheets to the sea. The Nesting Fault is probably of the same age as the last movement on the Walls Boundary Fault.

Yell was glaciated by the Shetland ice cap during the last glacial maximum (Late Devensian) when the ice shed lay obliquely across the southern half of the island and the ice front across the northern end of the island. After the first occupation of the island by man, at about 4650 BP, a thick layer of peat was formed over the island.

Chapter 2 Structural and stratigraphical succession

Yell, except for its east coast, is mostly underlain by a succession of psammites and psammitic gneisses of Moine type, the Yell Sound Division, which is very well exposed in the cliffs. It is bounded to the east, close to the east coast of the island, by a narrow band of microclinemegacryst augen gneiss, the Valayre Gneiss. To the east of the Valayre Gneiss, and exposed in the cliffs forming the east coast of Yell, are the metasedimentary schists and gneisses of the Boundary Zone, a compound tectonic unit which throughout Shetland separates the local Moineand Dalradian-type rocks. To the east of the Boundary Zone, and along the west sides of both Unst and Fetlar, is a succession of Dalradian-type schists and gneisses, the Valla Field and Lamb Hoga succession, which has been correlated with the Lower Dalradian Scatsta Division of the East Mainland Succession of Shetland. These are overlain farther to the east, in both Unst and Fetlar, by the Shetland ophiolite. In the north-east of Yell the Boundary Zone is cut by the Hascosay Slide, a zone of blastomylonitisation. The stratigraphic and tectonic division of the rocks of Yell, based on the metasedimentary lithology prior to the gneissification of the rocks, and including Lewisian-inlier units, is shown in (Figure 3).

The succession is described from west to east beause the East Mainland Succession as a whole is considered to young progressively to the east, and because that succession is truncated immediately west of Yell by the Walls Boundary Fault, while it continues far to the cast of Yell. No indication has been found within the Yell Sound Division itself of the direction in which the rocks get younger. Neither has it been possible to show that the contacts between the various lithological components of the division are other than conformable sedimentary contacts, although some differential movement may have occurred during deformation due to differing competences of adjacent units. The question of the continuity of the division arises from the presence within it of exotic masses in the form of Lewisian inliers. The boundaries of the inliers provide no evidence of the mechanism by which they were emplaced.

Yell Sound Division

Most of Yell is underlain by the psammites, schists, quartzites, gneisses and hornblende schists of the Yell Sound Division of the East Mainland Succession, which has been correlated with the Moine of mainland Scotland (Flinn, 1967). The Division was subsequently redefined as being limited to the east by the Valayre Gneiss (Flinn, 1988).

In Yell, the Yell Sound Division is dominantly composed of banded and laminated psammites with interbanded garnet-hornblende schists. Macroscopically garnet-studded hornblende schists are characteristic of the division. At three levels this psammitic lithology is interrupted by large lenticular masses of quartzite, pelite and semipelite. The resultant four units of psammite can only be separated from each other where the nonpsammite 'lenses' intervene and present sharp lithological changes. The fact that the Division is not a simple and continuous sedimentary succession is shown by the presence within it of lensoid masses of coarse hornblende gneiss of Lewisian inlier type accompanied by quartzofeldspathic orthogneiss. Widespread gneissification of the metasedimentary rocks occurs and the succession has been intruded by granitoid rocks, now orthogneisses.

The westernmost unit, the Yell Sound Psammites, is predominantly composed of psammite and includes the West Yell Lewisian inlier (Figure 3). Several bands of quartzite occur on the coast of Yell opposite Bigga. Some of these quartzites show well-formed, very regular and straight but faint centimetre-scale colour banding [HU 466 784]. Similar quartzites occur in the Yell Sound Division in the Central Shetland Sheet [HU 362 715]. In the Copister region several small deformed masses of steatitite with actinolite-biotite rims occur [HU 468 783]. Similar zoned exserpentinitic 'balls' also occur along the strike to the south-west in the Central Shetland Sheet. Also occurring in the Yell Sound Psammites are broad zones of rapidly alternating interbanded hornblende schist and psammite in the Ness of Sound and Copister areas and several sheets of granodioritic orthogneiss to the south of Grave-land ( [HU 445 957] and [HU 445 935]). The Graveland Lens is composed of layers of thin-bedded quartzites and interlayered quartzites and siliceous psammites, together with massive wedges of coarsely micaceous pelitic schist, all tending to wedge out to the south. These rocks are well exposed in the great cliffs on the west side of the Graveland peninsular. The Graveland Lens may be considered to continue to the south for 10 km in the form of a discontinuously exposed pelitic band richer in staurolite than kyanite. However, no staurolite or kyanite was found in the pelite masses in the main body of the lens. The staurolite schist is best seen near Lunga Water [HU 460 898] and along the crest of Evra Houll [HU 457 840].

The Herra Psammites, the next unit to the east, is made up of a thick sequence of psammites and contains a number of small Lewisian inliers best seen in the cliffs near Head of Bratta [HU 475 990], and a poorly exposed quartzite in the region of Kame of Sandwick [HU 480 880].

The Mid Yell Lens contains the massive, coarsely crystalline and apparently unbanded and unlayered Arisdale Quartzite on its west side, and on the east side the coarsely micaceous Kaywick Schist. In between is a lenticular mass of relatively finely crystalline, very micaceous psammitic schist of such an appearance that it can be readily distinguished from the generality of the psammites. This has been labelled semipelite and called the Mid Yell Schist. To the west of Otterswick, in a roadside quarry [HU 516 856], it can be seen to contain a number of fist-sized calc-silicate nodules. A few others occur elsewhere. The Mid Yell Lens seems to pinch out to the north in a very poorly exposed area extending from the head of Basta Voe to the north coast of Yell. However, on the north coast of Yell, to the west of Gloup Voe where on structural trends this stratigraphic horizon is expected to cross the coast, a considerable number of the calc silicate nodules in the psammites are exposed in the cliffs between Tonga [HP 502 052] and The Smeid [HP 496 054].

The Basta Psammites are not easily delimited due to the lack of inland exposures, extensive gneissification and faulting. A small, poorly defined Lewisian inlier occurs in Gloup Voe [HP 505 051]. Several quartzite bands occur but these may be southward extensions of the Cullivoe Lens. The Houlland Lewisian inlier in the south of Yell occurs in the junction between the Mid Yell Lens and the Basta Psammites.

The Cullivoe Lens is composed of very poorly exposed massive quartzite occurring to the south of Kussa Waters, [HP 520 030], and coarse pelitic schists exposed west of Hill of Breakon [HP 514 040] and intermittently between Breakon [HP 528 043] and Ward of Grimsetter [HP 538 007]. Farther south, on the same general strike, are interbanded quartzites and kyanite-rich pelitic schists, best seen at North Sandwick [HU 551 970]. The Cullivoe Lens contains a large body of granodioritic orthogneiss, the Breakon Gneiss, which distorts the structure of the lens and outcrops along the north coast where the rocks of the lens would otherwise be exposed. Some micaceous rocks, outcropping along the east side of the Bluemull Sound Fault and included in this unit, occur to the west of Gossaburgh [HU 513 833] and west of Burravoe [HU 515 795] where they are associated with quartzites [HU 512 788].

Relationships between the rock types east of the Bluemull Sound Fault are made difficult by the offset which has taken place on the fault, the complexity of the geology in the area and the usual paucity of inland exposure. The Otterswick Psammites contain only the Vatsetter and Rana Lewisian Milers [HU 544 893] and [HU 544 883] as separable lithologies, and on the north shore of Burra Voe several serpentinitic 'balls', a metre or two in diameter (e.g. [HU 527 797]).

Valayre Gneiss

On the west side of the Bluemull Sound Fault the Valayre Gneiss, a microcline-megacryst porphyroblast gneiss, can be traced almost continuously and unambiguously from the north coast of Yell to its intersection with the fault seven kilometres to the south. Along this part of its course it is a band rarely more than a metre or so wide. On the east side of the fault it continues its course to the south, but its exact line is less clear because of local faulting, and lack of inland exposure. It is seen again further south, on the Central Shetland Sheet, forming the core of the Lunna Ness peninsular and continuing south until intersected by the Nesting Fault in the Laxo area. On the west side of this fault it is displaced 16 km to the north and can then be traced via intermittent exposures from the Mossbank area westward, close to the southern contact of the Graven Complex, to Brae in the Central Shetland Sheet. From there it can be followed to the south parallel to the Walls Boundary Fault to a final exposure north of Bixter Voe (Figure 2). All the ground between the Valayre Gneiss and the Walls Boundary Fault is underlain by rocks of the Yell Sound Division and of the younger, intrusive Graven Complex.

As described below (Chapter 6) the Valayre Gneiss is a band of megacrysts in a variety of matrices marking the tectonic junction between the Yell Sound Division and the Boundary Zone and, therefore, belongs stratigraphically to neither.

Several bands of similar porphyroblast gneiss occur, which may be parallel to the Valayre Gneiss rather than a part of it. They are seen in several small isolated exposures of microcline-megacryst augen gneiss, generally less well developed than the Valayre Gneiss and with smaller megacrysts at several places in the Basta Psammites between Gossaburgh and Burrafirth, on the south coast of Yell (Figure 13). Similar rocks also occur in the Central Shetland Sheet immediately' west of the Valayre Gneiss on the west side of Lunna Ness, between the West Voe of Lunna and Boatsroom Voe.

Several small isolated occurrences of microclinemegacryst augen gneiss occur elsewhere in Yell.

Boundary Zone

To the east of the Valayre Gneiss are the rocks of the Boundary Zone, a unit which everywhere, in Shetland, lies between the Moine and the Dalradian rocks of the East Mainland Succession. The Boundary Zone (and the Valayre Gneiss) can neither be shown to transgress the rocks of the Yell Sound Division nor to be conformable to them, due to interrupted exposure, the lenticular nature of the units forming the Yell Sound Division, and the presence of the Graven and Brae Complexes intruding the line of contact. Within the Boundary Zone abrupt changes occur between such rock types as psammites, schists, paragneisses, hornblendic and quartzofeldspathic orthogneisses of anatectic and nebulitic types, with no parallel changes in the Dalradian and Moine rocks on either side. The Boundary Zone appears to be a zone of tectonic slices whose maximum dimensions are to be measured in kilometres, and which arc composed of kyanite-bearing metasedimentary rocks and high-grade gneisses of various types (a schuppen zone). The zone is vertical in most places and bounded on both sides by microcline-augen gneisses, but augen gneisses of very different appearance.

In Yell, the rocks of the Boundary Zone are psammites (locally gneissified), together with some quartzite at Otterswick [HU 547 876]. They contain globular masses of metadolerite associated, in some places, with hornblende schist bands, but the hornblende schist bands lack the macroscopic garnets so characteristic and so prominent in the hornblende schists of the Yell Sound Division.

A particularly interesting component of the metasedimentary rocks of the Boundary Zone, occurring to the east of Gossaburgh, is a body of metadiamictite [HU 535 830]. This takes the form of widely scattered basic to ultramafic clasts 'floating' in the local microporphyroblastic paragneiss. Most common are angular fragments of hornblendite several centimetres across. About ten times as large as these are flake-shaped masses of steatite with hiotite rims and traces of actinolite, which have been derived by metasomatism from serpentinite. Also present is an angular fragment of serpentinite a metre or so across and nearby is a several-metre diameter mass of pyroxene banded-gneiss, partially altered to hornblende gneiss.

The Neapaback Skerries, off the south-east corner of Yell [HU 535 787], are made up of rocks belonging to the Boundary Zone, and are nebulitic gneisses very similar to those on the east side of Lunna Ness, about 4 km to the south (Central Shetland Sheet, IGS, 1981).

Hascosay Slide

Along the north-east coast of Yell the Boundary Zone is obliquely cut and truncated by the Hascosay Slide which is a zone of hornblende blastomylonites some hundreds of metres thick containing closely packed relic masses of coarse hornblendic rocks, including gneisses of Lewisianinlier type. The Hascosay Slide, which has been cut and displaced by the Bluemull Sound Fault, passes to the south-east through the centre of the island of Hascosay and is not seen again farther south, even in the Central Shetland Sheet.

Comparison with mainland Scotland

The Yell Sound Division is a monotonous succession of psammites and psammitic gneisses, which can only be subdivided locally without difficulty and has a present apparent thickness of the order of 10 km or more. The psammites are intensely banded and laminated due to varying mica content and show no signs of cross-bedding or of conglomerates. Cale-silicate nodules occur locally and sparsely. Several thick lenticular masses of quartzite and coarse micaceous schist occur in the Yell Sound Division, but thin bands of hornblende schist occur commonly, many being prominently studded with garnets (no such garnets occur in hornblende schists anywhere to the east of the Yell Sound Division). Lewisian-inlierlike masses of coarse hornblendic gneiss occur. The psammites and schists are widely, variably and patchily gneissified.

The rest of the East Mainland Succession lying east of the Yell Sound Division, and mostly outwith the area covered by the Yell Sheet, also has a present thickness of about 10 km but it is composed of thick, continuous, mappable bands of limestone, calc-silicate, quartzite, psammite, pelite and volcanic units. There are no Lewisian inliers and paragneisses are confined to a single well-defined horizon.

Thus the eastern part of the East Mainland Succession has a Dalradian appearance and the western part a Moine appearance. It is these features which gave rise to the correlation of the Yell Sound Division with the Moine of mainland Scotland (Flinn, 1967) and in particular with the Glenfinnan/Loch Eil Division (Flinn et al., 1972; Flinn, 1988).

Chapter 3 Lewisian inliers

Coarse hornblendic gneisses occur in Yell in the Yell Sound Division and in the Hascosay Slide. They are unlike any other rocks in the East Mainland Succession of Shetland. Similar rocks occur west of the Walls Boundary Fault in Northmaven, between the Virdibreck and Wester Keolka shears, and further west in the Western Gneisses. Both occurrences are considered, by analogy with occurrences in the mainland of Scotland, to be Lewisian gneisses (Flinn, 1988). Coarse hornblende gneisses in the Hascosay Slide are described in Chapter 8. The hornblendic gneisses in the Yell Sound Division in Yell occur as masses of various sizes enveloped in coarse quartzofeldspathic gneisses. In the field, both types of rock have an exotic 'high-grade gneissic' appearance not shared by the rest of the rocks in the Yell Sound Division. This appearance is due to an absence of metasedimentary relics, a generally coarser grain-size and other features described below. The joint occurrences of the two gneiss-es are called 'Lewisian inlier', because of the similarity of the hornblende gneisses to the Lewisian hornblende gneisses of north-west Shetland and to the hornblende gneisses forming the Lewisian inliers in the Moine of mainland Scotland.

Lewisian inliers occur in about a dozen places within the Yell Sound Division to the east of the Nesting Fault, but do not occur in the Yell Sound Division west of the Fault, either in the islands in Yell Sound or in the Central Shetland Sheet. The distribution of the inliers is shown in (Figure 3). The Houlland Inlier, which is intersected and offset by the Arisdale Fault, is the largest and the Gloup Inlier the smallest. All inliers occur to the west of the Valayre Gneiss except that at Rana (6 on (Figure 3)). I Iowever, this inlier lies immediately to the north-east of a sharply formed anticline, with Valayre Gneiss on both its limbs. The inlier probably lies in a synform with Valayre Gneiss in its northern limb hidden under the sea. This interpretation would mean that it is on the same side of the Valayre Gneiss as the Vatsetter Inlier to the south, and therefore belongs to the west side of the Valayre Gneiss.

Hornblende gneisses

Coarse hornblendic gneisses form the most prominent component of the inliers. These gneisses vary from hornblende gneisses to hornblende-feldspar gneisses and to epidote-banded hornblende gneisses. Most are hornblende gneisses or interbanded hornblende and epidotehornblende gneisses ((Plate 2)a and b). Some hornblende gneisses have very fine, parallel, feldspathic laminations. The coarse hornblendic gneisses differ from the hornblende schists of the Yell Sound Division in being coarser, lacking garnet (except at some contacts with the enveloping quartzofeldspathic orthogneisses), and in showing signs of agmatisation, a type of migmatisation observed elsewhere in the East Mainland Succession (outside Lewisian inliers) only in the anatectic gneisses on the east side of Lunna Ness. In thin section the grain size of the hornblende is seen in most samples to be 2 mm or more, whereas the hornblende schists have a grain size usually less than 1 mm.

The hornblendic gneisses occur as masses varying from a kilometre or more across (in the Houlland Inlier) to lumps less than a metre across. In all the inliers, smaller masses, lenses, pods and lumps of hornblende gneiss 'float' like 'fish' in the associated quartzofeldspathic gneiss ((Plate 2)c). Where junctions can be seen they are sharp and in places are marked by the development of garnet in the edge of the hornblende gneiss. In the better-exposed junctions it is possible, in places, to observe that banding in the hornblende gneisses is truncated, generally very obliquely, by the foliation in the quartzofeldspathic gneisses. At Houlland ([HU 504 801]; (S94729)) coarse hornblende-banded gneisses are truncated at their contact with the enclosing quartzofeldspathic gneisses by a narrow zone of garnet-hornblende schist. The general lack of continuity between the hornblende 'fish' is such as to provide no grounds for or against interpreting them as former intrusive sheets or dykes. The impression given by all these exposures is that the hornblendic gneisses are xenoliths, large and small, within the quartzofeldspathic gneisses. This impression is reinforced by the fact that banding in the hornblendic gneisses of the Houlland Inlier is nearly horizontal, while the foliation of the enclosing gneisses dips to the west much more steeply.

In thin section the hornblende gneisses are seen to be composed of olive-green and occasionally green-brown hornblendes, generally between 2 and 3 mm in diameter ((S91250); (S91298); (S90787)). Both quartz and plagioclase are usually present, and biotite occasionally so. In most, but not all rocks, some degree of lattice preferred orientation is developed parallel to the banding and lamination present, but grain-shape preferred orientation is generally absent ((S94730); (S90787)). In some places the epidote-banded hornblende gneisses contain clinopyroxene ((S91018); (S90723); (Plate 2)a). Very rarely the gneisses have a calc-silicate-rock appearance, and are composed of clinopyroxene with microcline (S90784) rather than calcic plagioclase.

Also present in the inliers are small pods or 'fish' of relatively fine-grained hornblendite (less than 2 mm grain size) which generally contain garnet and sometimes prominent rutile ((S91738); (S94729); (S90727)). Some show signs of agmatisation (S91620). The hornblende in the fine-grained garnet hornblendite is seen in thin section to have a grain size of between 0.5 and 1.0 mm (garnets are larger). They generally have a good lattice preferred orientation, but a granular, equidimensional texture so that the rocks do not appear schistose in the field, although some have a preferred shape-orientation (S91298). They contain minor amounts of quartz and/or plagioclase, and sphene or rutile. Rutile is much more common in these rocks than it is in the somewhat similar garnet-hornblende schists and hornblende schists occurring outwith the inliers.

In the Houlland Inlier, in the Ness of Galtagarth, a band of dark, relatively fine-grained rock about a metre thick and apparently untectonised, appears to cut the banded coarse tectonised hornblende gneiss [HU 497 797]. In thin section this rock is seen to be a pyroxenite with some hornblende and plagioclase, a grain size of about 0.5 mm, a moderate preferred orientation of the hornblende and a granular texture (S90779). It may be a basic intrusion emplaced in the gneiss before it became part of an inlier.

Quartzofeldspathic gneisses

The other main component of the inliers is coarse garnet-biotite quartzofeldspathic gneiss completely lacking in metasedimentary relics. These gneisses are distinguished from coarse quartzofeldspathic paragneisses occurring outwith the inliers by a number of characteristic features, the most prominent of which is the presence of the small to giant xenolithic masses of the coarse hornblendic gneiss. The proportions of hornblende gneiss and quartzofeldspathic gneiss vary widely from inlier to inlier.

The quartzofeldspathic gneisses vary from inlier to inlier chiefly in the local importance of the following features. They contain a good deal of garnet in grains up to several millimetres in diameter, but characteristically they also contain larger garnets between 2 and 5 cm in diameter. Some of these have well-formed crystal faces. Also very characteristic are 'blebs' of pegmatite several centimetres in diameter (S90726) composed solely of quartz and albite. The term 'bleb' has been used here in order to emphasise the difference in their appearance from the pegmatitic-looking quartzofeldspathic augen and 'sweat-out' veins and streaks so common in the paragneisses. The 'blebs' have a xenolithic or fragmental appearance, and the feldspar grains within them are often sufficiently large and well developed for albite twinning to be discernable in the field. Both the large garnets and the pegmatite 'blebs' are rather sparsely developed so that at best only one of either can be expected to be seen in a square metre of rock surface ((Plate 2)c). Less prominent and more unevenly developed arc microporphyroblasts of plagioclase 2 to 3 mm in diameter (S91016).

The quartzofeldspathic inlier gneisses with these characteristic features are of two types. The most common are sufficiently micaceous for them to have a marked schistose appearance in the field. Less commonly the gneisses have an equigranular, nonschistose, appearance due to the presence of fewer and smaller mica flakes.

Some occurrences of both types of quartzofeldspathic gneiss contain thin strings and bands of hornblende and garnet and, much more commonly, the fine-grained hornblende 'fish' mentioned above. Within the inliers, garnet tends to occur with hornblende only where it is also in immediate contact with, or very close to, quartzofeldspathic gneiss.

In thin section the quartzofeldspathic gneisses are seen to he quartz-garnet-biotite-plagioclase rocks with a good schistosity defined by the biotite (S91539). The grain size is generally between 1 and 2 mm, but garnets and rounded equidimensional porphyroblasts of plagioclase vary up to two to three times this size ((S91016); (S90913)). These plagioclase microporphyroblasts are indistinguishable in appearance from those in the microporphyroblast gneisses of the Boundary Zone (Chapter 6). Hornblende occurs in some gneisses as an accessory mineral, K-feldspar never and muscovite only rarely (S91182), and then in very small proportions. It is difficult to determine exactly the boundary between the Lewisian inlier quartzofeldspathic gneiss and the enclosing gneissose metasedimentary rocks and it is possible that the few samples of inlier rocks found to contain muscovite may, in reality, be tectonically intermixed rocks of metasedimentary origin.

Several types of rocks, not mentioned above, have also been found within the inliers. Laminated or banded rocks with the field appearance of quartzite and impure-quartzite occur in the West Yell Inlier, the Houlland Inlier and the easternmost of the inliers at Head of Bratta ( [HU 476 990]—one of the West Lumbister Group of Inliers). Thin section examination shows, however, that these may be early highly tectonised aplites (S90777), or quartzofeldspathic gneisses extremely rich in quartz (S91756). The West Yell Inlier contains some gossanousweathered bands of hornblende-pyrrhotite-pyrite a few tens of centimetres thick (S90910). In the Houlland Inlier, on the north shore of the Ness of Copister, a small mass of actinolite with a biotite selvage occurs, as does another in the West Yell Inlier (S90905). Both have the characteristic appearance of 'zoned balls' of metasomatised serpentinite. Clinopyroxene, pale green in colour and with a salite-like appearance, was found in the Houlland and West Yell Inliers; in both it occurs in epidotic bands in the hornblende gneiss ((S90723); (S91018)). Two 30 cm-thick hands of rock composed of quartz, garnet, clinozoisite and augite (S91019) occur in the quartzofeldspatic gneiss in West Yell [HU 448 838]. A most unusual rock occurs on the shore at Houlland [HU 502 798] in the form of a clinopyroxene-microcline granular rock with accessory sphene and zoisite (S90784).

Pegmatites and aplites occur in all the inliers. Some are tectonised or are heavily contorted or truncated at gneiss boundaries within the inliers and so seem to be of pre-inlier age ((S94682); (S94678)). All of these are white in colour. They are most clearly seen in the Houlland quarry. Others containing large pink single-crystal micro-clines intrude all other rocks except lamprophyres and belong to the late microcline pegmatite suite which occurs profusely throughout the island. These last are particularly common in the West Lumbister and West Yell inliers. The Houlland Inlier is heavily injected by the pegmatitic-aplites (S94681) so common in south-east Yell. Like the pre-inlier-age pegmatites, they are white and some are difficult to distinguish from the older pegmatites. In general late pegmatites are larger than the early pegmatites and often form continuous wandering sheets.

The inlicrs are cut by the lamprophyres which occur throughout Yell and which intrude all other rocks (see Chapter 9).

Field occurrence

The Gloup Inlier ( [HP 506 051]; 1 in (Figure 3)) is dominantly quartzofeldspathic and is exposed in a very accessible coastal section where it is seen to grade imperceptibly into the adjacent Yell Sound Division psammites, making its boundaries very difficult to determine precisely. Hornblende gneiss (S92119) is present as a minor constituent.

The Setter Inlier ( [HU 495 915]; 2 in (Figure 3)) is exposed in a flat, ideally weathered, inland exposure showing blocks of coarse hornblende gneiss (S91298) and fine-grained garnet hornblendite 'fish' in typical pegmatiteblebbed quartzofeldspathic inlier gneiss.

The presence of the Arisdale Inlier ( [HU 484 844]; 3 in (Figure 3)) is revealed by a single tiny exposure of hornblende gneiss in a burn and a string of loose blocks of different types of inlier gneiss distributed along the burn for a kilometre or so.

The East Lumbister Inlier is exposed on the shore at the east end of the Loch of Lumbister ( [HU 493 970]; 4 in (Figure 3)) as a few exposures of coarse biotitic quartzofeldspathic gneiss ((S91644); (S91647)) rendered rather unusual by contorted pegmatitic streaking and xenoliths of coarse hornblende gabbro (S91648). Garnetiferous hornblendite 'fish' are also present.

The Vatsetter Inlier ( [HU 544 884]; 5 in (Figure 3)) is obscured and complicated by faulting, folding and lack of inland exposure in critical areas and, as mentioned above, may once have been connected to the Rana Inlier ([HU 543 893]; 6 in (Figure 3)) to the north before the connection was eroded. The inlier at Rana is a horizontally foliated and easily accessible block of coarse hornblende gneiss lying on a thin envelope of pegmatite-blebbed granular quartzofeldspathic gneiss (S91250). The hornblende gneiss contains some epidotic banding and is cut by ductile shears and tectonised aplites. To the south, across the anticline of Valayre Gneiss, the Vatsetter Inlier is composed of pegmatite-blebbed quartzofeldspathic inlier-type gneiss containing blocks of coarse hornblende gneiss and smaller 'fish' of garnet-hornblendite, but the cliff section is obscured by fallen blocks, a faulted-in mass of non inlier and a large intrusive sheet of Vatsetter-type tonalite (see Chapter 9). The southern boundary is completely gradational from quartzofeldspathic orthogneiss to paragneiss. To the north of the Vatsetter–Rana Inlier, on the east side of the Ness of Vatsetter [HU 540 897] and, less well developed, in the south-west tip of Hascosay [HU 544 910], characteristic inlier-type pegmatite blebs occur in psammites which lack all other inlier-type features.

The West Lumbister Group of Inliers (7 in (Figure 3)), extending from south of the Burn of Lumbister [HU 475 965] to the Head of Bratta [HU 475 989], are well exposed only in the Head of Bratta cliff section, although the Head of Bratta West Inlier is very heavily injected by late microcline pegmatites. The Head of Bratta section exposes the contacts of both the Head of Bratta East and West Inliers with the enclosing rocks. The boundaries can be located, at best, to within a metre or so by the termination of the characteristic features listed above and are accompanied by a local enhancement of the schistosity. Zoisite and rutile occur unusually frequently in the Head of Bratta East Inlier which is also unusual in containing bands of quartzitic appearance (see above). A hornblendic mass which is possibly a small inlier occurs in the cliffs at Stuis of Graveland [HU 463 963].

The remaining inliers are more rewarding to visit. The West Yell Inlier ([HU 450 830]; 8 in (Figure 3)) is well exposed in a very accessible coastal section running parallel to its longest axis. Apart from the several occurrences of unusual rocks mentioned above, this inlier differs from the others in that coarse hornblende gneiss occurs interbanded with quartzofeldspathic gneiss. Careful study of the boundaries reveals that the enclosing quartzofeldspathic gneiss commonly obliquely truncates the banding in the hornblende gneiss. The quartzofeldspathic gneiss is poor in pegmatitic blebs and relatively rich in microporphyroblasts of plagioclase (S91016). Pyroxene-epidote bands occur in both the hornblende gneiss (S91018) and the quartzofeldspathic gneiss (S91019). Early tectonised aplites occur (S94688) and the inlier is cut by numerous microcline pegmatites no different to those occurring throughout Yell. However, as with all the inliers, it can be difficult to distinguish between pegmatites and aplites that belong to the inlier and those emplaced later throughout both the inliers and the Yell Sound Division.

The Houlland Inlier (on the south coast of Yell; 9 in (Figure 3)) is by far the largest, extending as it does from The Rumble, to the south in Yell Sound, for several kilometres into Yell to the north. Much of it is hidden beneath the sea or the peat but good accessible coastal sections are available in Ness of Copister [HU 490 785], Ness of Galtagarth [HU 495 797] and the Houlland quarry [HU 504 802] and provide unexcelled views of the coarse hornblendic gneiss and the early pegmatites. The inlier contains a very large mass of epidote-banded hornblende gneiss underlying these areas. The layering in the mass apparently dips gently to the north, but is affected by widespread chaotic rolling of the handing. The relation of the mass to the underlying quartzofeldspathic inlier-gneiss is well exposed on Ness of Copister [HU 492 782]. Pyroxene-epidote banding in the hornblende gneiss is well exposed nearby at [HU 487 783]. The coastal section at Houlland [HU 503 797] contains good exposures of hornblende-garnet striping in the pegmatite-blebbed quartzofeldspathic gneiss. The coast on the south side of Ness of Galtagarth shows an exceedingly complex mixture of most of the rocks composing the Lewisian inliers, including a poor exposure of what may be a late, but metamorphosed, basic dyke cutting the gneisses ( [HU 496 896]; see above).

Interpretation

No sharply defined tectonic or lithological features mark the limits of the inliers so that boundaries can be located, at best, no more precisely than to the nearest metre. The inliers thus have the appearance of being integral parts of the Yell Sound Division. Nevertheless, they have an exotic appearance due to such features as the lack of any sedimentary relics of undisputable validity within them, the higher grade (coarser and more garnetiferous) than the surrounding rocks, the fact that they are entirely composed of rocks not occurring elsewhere in the succession and the presence in the quartzofeldspathic gneiss of the albite-quartz pegmatitic 'blebs' and giant garnets. All these features represent a metamorphism which took place prior to the incorporation of the inliers into the Yell Sound Division. They are therefore, by analogy with occurrences in the mainland of Scotland, termed Lewisian inliers and interpreted as fragments of an older metamorphic basement incorporated into the succession forming the Yell Sound Division before or early in the metamorphism of that Division. If the Lewisian inliers were part of the basement on which the metasedimentary rocks of the Yell Sound Division were deposited then all structural features associated with that incorporation have been destroyed. If the inliers are the cores of folds, then all evidence of this folding has been destroyed, but in such a way that, as shown below, the sedimentary banding in the enclosing rocks has not been affected.

Chapter 4 Metasedimentary rocks

The metasedimentary rocks underlying most of Yell belong to the Yell Sound Division of the East Mainland Succession of Shetland. Those along the east coast of Yell, to the east of the Valayre Gneiss, belong to the Boundary Zone, a lithologically complex tectonic zone which separates the Yell Sound Division from the Scatsta Division of the East Mainland Succession. The Yell Sound Division has been intruded by innumerable sheets and bodies of basic rock now in the form of garnet-hornblende schists and hornblende schists. These, the Lewisian inliers and the metasedimentary rocks together, are sufficiently similar to the rocks of the Glenfinnan and Loch Eil divisions of the Scottish Moine to suggest a correlation. The hornblende schists and other basic intrusions, together with a number of early granitic intrusions in the Yell Sound Division, are described in later chapters.

Nomenclature

Most of the rocks in Yell are granoblastic in nature like those in the Moine of mainland Scotland and break into angular fragments, not into the flaky fragments characteristic of schists. For this reason, such rocks in mainland Scotland have in the past been called 'Moine granulites'. The term granulite is now universally accepted as referring to certain high grade metamorphic rocks and so should not be used for granoblastic Moine rocks. Since such rocks differ significantly in appearance from schists and arc volumetrically the most important constituent of the Shetland Moine and probably of the mainland–Scottish Moine, and since, as shown in Chapter 6, their response to gneissification is very different to that of schists, they need to be distinguished from schists. In this memoir they will be referred to as granofels (Plate 3).

The term will be used to indicate the structural (fabric) state of a rock, not the composition. Granofels show a continuous gradation into schists through such a series as psammite–semipelite–pelite. The difference between granofels and schist is that the granofels on being broken, naturally or when collected as hand specimens, breaks into angular fragments tending to be bounded by surfaces unrelated to the schistosity, whereas schist fragments tend to be dominated by surfaces parallel to the schistosity. The distinction is complicated by the fact that granofels pass gradationally into schists either by increasing mica content, increasing grain size of the mica or by increasing preferred orientation of the mica flakes or by combinations of these and also by interlamination of different members of the schist-granofels series.

The essential difference between granofels and schists lies in the nature of their mica fabrics. In the granofels the mica flakes occur separated from one another in a granular matrix of quartz and feldspar, or, if touching, having litle or no preferred orientation, whereas in schists the mica flakes are parallel in at least one dimension and run into one another to form continuous rows or layers of flakes. The two types of rock behave differently when deformed and, as described in Chapter 6, when gneissified.

It should be noted that in the transition between granofels and schist, the former can have a higher mica content than the latter. In extreme cases pelites can be granofels and impure quartzites containing very large, parallel-oriented mica flakes can he schists. It is for this reason that such compositional terms as psammite, semipelite, and pelite are insufficient for many purposes.

Yell Sound Division

The metasedimentary rocks of the Yell Sound Division range from quartzite to pelite, but most are of intermediate composition, i.e. granofels of psammitic to semipelitic composition varying to schists of semipelitic to pelitic composition. The quartzites and pelites occur closely associated in large lenticular bodies separated by considerable thicknesses of psammites.

Quartzites

The distribution of quartzites is shown in (Figure 4). There are two large bodies of massive quartzites associated with the 'lenses'. The Arisdale Quartzite in the Mid Yell Lens, and the quartzite in the Cullivoe Lens are very poorly exposed massive, coarsely crystalline, pure quartzites with no apparent banding or lamination. The quartzites in the Graveland Lens or groups of lenses are very well exposed in the cliffs and can be seen to be intensely interlaminated and interbanded pure massive quartzite, impure quartzite and very siliceous psammite. Apart from two poorly exposed, much smaller bodies of quartzite at Cullivoe [HP 535 030] and West Sandwick [HU 478 880], and more notably a section of bedded quartzite cut by the Bluemull Sound Fault at Burravoe [HU 512 789] and [HU 516 802], the quartzites are isolated bands in psammite or groups of bands alternating with pelite, most being a metre or less wide (Figure 4). Quartzites varying from several centimetres to several metres in thickness occur interbanded with pelites at North Sandwick [HU 550 969], less spectacularly on either shore in the entrance to Mid Yell Voe [HU 530 910], and at West Yell in staurolite schist [HU 465 855]. The other occurrences indicated on (Figure 4) are isolated quartzite hands in psammite. Most bands more than about 30cm thick have a bedded appearance.

On Bigga [HU 447 795] and at Copister [HU 466 784] on the south-west tip of Yell there are bands of pure massive quartzite up to several metres thick with very pale colour banding a centimetre or two thick, which, because of the perfectly parallel and rectilinear nature of the bands (where unfolded), give the rocks a very characteristic field appearance. One sample (S90693) was found, on thin section examination, to be an impure quartzite with very minor amounts of biotite and hornblende. Very similar rocks occur more frequently and in bigger masses in the Yell Sound Division of the Central Shetland Sheet to the north and south of Grona Taing [HU362 715]. Thin sections showed them to be pure quartzites. These striped quartzites may be derived from banded radiolarian cherts.

Calc-silicate rocks

Apart from some slices of marble in the Nesting Fault in Samphrey [HU 464 765] no calcareous rocks have been found in the Yell Sound Division in Shetland. The marble slices in the Nesting Fault have possibly been derived from the limestones which occur in the Boundary Zone and exposed locally in the area covered by the Central Shetland Sheet.

In Yell a number of carbonate-free talc-silicate rocks have been found (Figure 4). These are either nodules, mostly fist sized, or are bands varying from several centimetres to several metres in thickness, generally highly quartzitic and often banded on a 1 or 2 cm scale.

The nodules mostly occur in the Mid Yell Schist and are very sparsely distributed. They are best viewed in a roadside quarry at Otterswick [HU 517 856] and, less conveniently, in a burn west-south-west of Vatsetter [HU 527 893]. Such nodules occur most profusely in Yell on the north coast to the west of Gloup Voe [HP 498 054], an area out-with the Mid Yell Schist but on the same stratigraphic horizon. Thin sections of the nodules contain quartz, garnet, and zoisite (S91244) together with a selection of such minerals as pale green hornblende (S94699), plagioclase (andesine-labradorite, (S94724), and clinopyroxene (S92056). The accessory minerals present are sphene and rutile. In thin section (S94700) the garnets contain well-developed geometrical patterns of inclusions indicating rhombdodecahedral growth of the garnet. Quartz is the dominant mineral in the nodules. An isolated 50 cm quartz-free nodule, composed entirely of zoisite and garnet with accessory sphene, occurs in north-east Yell ( [HP 529 055]; (S92172)).

Bands of quartz-rich rock with a talc-silicate appearance in the field also occur in thicknesses varying from a couple of centimetres to several metres. The thicker layers are generally composition-banded on a several centimetre scale. Several such layers, up to 1 m thick, occur associated with the talc-silicate nodules on the west side of the entrance to Gloup Voe. The other exposures are widely scattered. Isolated 2 cm-thick bands occur on the west coast [HP 474 022], and inland in Cullivoe [HP 529 031]. A few bands several centimetres thick, together with lenses of the same rock, occur in Wick of Whallerie [HP 506 052] and the most substantial occurrence of all, a band 3 m thick, occurs in a quarry near Basta Voe [HU 510 968]. An occurrence in a burn to the south-east of the quarry [HU 515 958] is probably a continuation of the latter. All these occurrences are quartz-rich and contain the same minerals found in the nodules and listed above.

Mica schist

Lenticular masses and bands of coarse mica schist are prominent components of the Cullivoe, Mid Yell and Graveland lenses (Figure 5). Several substantial masses of coarse mica schist occur well exposed in the cliffs of Gravel and [HU 450 960] and closely associated with the quartzites of the Graveland Lens. The Kaywick Schist is a coarse mica schist which forms a lenticular band along the east side of the Mid Yell Lens. It is well exposed in the cliffs forming the entrance to Mid Yell Voe [HU 520 910]. It is truncated by the Bluemull Sound Fault, but appears again to the south forming the northern boundary of the Houlland Lewisian Inlier [HU 500 810]. The Camb Schist, a smaller, poorly exposed lens of coarse mica schist, occurs on the west side of the Mid Yell Lens to the north of the Arisdale Quartzite [HU 505 922]. Poorly exposed coarse mica schists form part of the Cullivoe Lens in north-east Yell [HP 530 040] and west of the Breakon Gneiss [HP 513 032]. These two occurrences have been partially gneissified by the development of much quartzofeldspathic streaking. A large lenticular mass of very similar coarsely quartzofeldspathic-streaked mica schist, which occurs to the west of Burravoe [HU 515 790], is probably on the same stratigraphic horizon.

No sedimentary structures or banding, other than the common occurrence of quartzite bands and the less common occurrence of psammite bands, were found within the mica schists. The coarseness of crystallisation and the highly schistose and flaky nature of these rocks prevents the recognition of such features. Only the minerals garnet, muscovite, biotite, quartz and plagioclase were found in these major lenses or bands of mica schist, except at the south tip of the Kaywick Schist [HU 503 815], where both kyanite and staurolite were found, and in the northern extension of the Camb Schist [HU 500 928] where staurolite occurs. Garnet is ubiquitous, and often occurs in crystals up to a centimetre or more in diameter. The micas rarely form more than 50 per cent by volume of these rocks and often much less. Thirty-three out of 57 thin sections of mica schist contained less biotite than muscovite (by contrast biotite is generally the dominant mica in psammites). Plagioclase occurs in equal proportions, or less, to quartz in mica schists, and quartz plus plagioclase occurs in greater volume than the micas in all but 6 out of the 57 thin sections, i.e. the mica schists are mostly semipelites rather than pelites.

Apart from these large masses, coarse mica schists are of limited occurrence in Yell. Bands a metre or two wide, and occasionally several metres wide occur in several places. Recognition is made difficult because mica schist, in places, grades into micaceous psammite and micaceous semipelites and gneissification of these three types of micaceous metasedimentary rock leads to the production of indistinguishable quartzofeldspathic-streaked gneisses. The locations of a number of minor occurrences of coarse mica schist are shown in (Figure 5).

Three minor occurrences of mica schist are of special interest. A band of staurolite-mica schist, often with kyanite, is exposed intermittently for 10 km south of the Gray-eland Lens although no staurolite or kyanite was found within the main part of the lens. Exposures occur at Lunga Water [HU 460 898] and at Hill of Clothan [HU 463 813] but most notably along the ridge from Evra Houll [HU 457 835] to Hill of Noub [HU 465 855] where the schist is associated with a good deal of quartzite banding. Besides staurolite, these exposures contains kyanite, garnet, muscovite, biotite, quartz and plagioclase (S91023); (S91027). Another mica schist band can be traced from south of the Cullivoe Lens to the coast at North Sandwick [HU 551 970], where it is intensely interbanded with quartzite and largely composed of coarsely crystalline blue kyanite rods with staurolite, garnet, biotite, muscovite, plagioclase and quartz (S91682); (S91696). Inland to the north are exposures of kyanite schist and quartz-kyanite segregation lenses 10 cm or more across. The third occurrence is a lens of coarse mica schist, about 10 m across and several times as long, complexly folded on steep axes in the cliffs at Gerherda [HP 475 009] on the north-west coast (Figure 5). This is remarkable for containing kyanite pseudo-morphs of chiastolite as rods up to 2 cm long and square cross-section of several millimetres side, with clearly preserved chiastolite-type inclusion patterns (S91817) clearly visible in the field ((Plate 3)e). It will remain a remarkable exposure only as long as it is left unhammered. Other minerals in this schist are garnet, biotite, muscovite, plagioclase and quartz. No other relics of chiastolite were found in Yell, but some shimmer aggregates which could be replacements of andalusite were found in micaceous psammites at the north-east entrance to Gloup Voe ( [HP 505 055], see below). It is concluded that the chiastolite grew in the sediments as a result of thermal metamorphism before or at the beginning of the regional metamorphism. If this is so, the most likely cause would be an intrusion of granitoid rock immediately seaward of the schist and now under the sea. Such intrusive rocks occur in north-west Yell in the form of the Graveland orthogneisses and in north Yell in the form of the Breakon Gneiss (Chapter 6). However, no relics of thermal metamorphism have been observed adjacent to these gneisses.

The Mid Yell Schist is a large lenticular band of highly schistose semipelite forming the core of the Mid Yell Lens, with the Kaywick Schist to the east and the Arisdale Quartzite and Camb Schist to the west (Figure 5). The rocks forming the Mid Yell Schist appear in the field to be much finer grained than the coarse mica schists described above, yet in thin section the individual mica flakes are seen to be as large, that is 1 or 2 mm long. Throughout their outcrop the schists are laminated, ribbed and banded by psammitic granofels, generally varying up to a centimetre or two in thickness and less commonly up to 10 or 20 cm. The ribs and bands are of sedimentary origin, are parallel, and never exhibit graded bedding. There are no signs of cross-bedding or of gritty or conglomeratic rocks, although many ideally weathered surfaces are available on which such features, if present, would be clearly visible. The S_B crenulation (spaced) cleavage (Chapter 7) is so strongly developed throughout the Mid Yell Schist that in many places the rocks are too flaky for a good hand specimen to be collected without great difficulty. The psammite ribbing is most easily seen where it has been folded and is cut by the SB cleavage. Elsewhere the cleavage is parallel to and overprints the ribbing. The Mid Yell Schist in many places is prominently studded with garnets varying up to a centimetre in diameter. At locality [HU 518 918] there is a 20 cm band of schist containing garnets up to 2 cm in diameter. The Mid Yell Schist contains talc-silicate nodules locally, as described above.

The widespread development of a spaced cleavage, SB, give the rocks of the Mid Yell Schist a more highly schistose appearance than rocks of similar mineralogical content elsewhere in Yell. Mineralogically, they are very similar to the coarse mica schists, the mica flakes being the same size (about 2 mm) and, like those in the coarse mica schists, tend to occur packed into ragged anastomosing lamellae several flakes thick. In the coarse mica schists there is more mica and the quartz and feldspar grains are coarser. As in the coarse mica schists there is more muscovite than biotite in most specimens of the Mid Yell Schist and plagioclase tends to occur in equal or lesser amounts than quartz.

Granoblastic psammites (granofels)

Most of the metasedimentary rocks in the Yell Sound Division are granofels. At first sight they appear to be a monotonous uniform succession made interesting only by the presence of zones and patches of gneissification and partial gneissification. Closer inspection of suitably weathered exposures reveals compositional banding on all scales from millimetres to hundreds of metres. It is seen in individual exposures as parallel-sided and continuous rectilinear bands several centimetres thick, as well as continuous parallel laminations a millimetre or two thick. (Plate 3). The larger-scale handing is apparent. in cliff sections.

Compositional banding and lamination is best developed in the more micaceous granofels. Quartzofeldspathic-rich granofels (e.g. psammites) generally exhibit a bedded appearance in the field due to parallel partings with a spacing of about 20 cm and no apparent compositional variation. These partings are parallel to the schistosity which, in turn, is parallel to the compositional banding and lamination. The bedded appearance is therefore considered to be a relic of original sedimentary bedding. No signs of grading or cross-bedding have been found and any pinch-and-swell structures or other interruptions to the regularity of the banding and bedding appear to be tectonic in origin. No gritty or conglomeratic rocks occur.

The average diameter of quartz and feldspar grains obtained from measurements of several hundred thin sections is 0.3 mm. The mica flakes tend to increase in size with increasing mica content. The mica content of the granofels varies between 10 per cent and 40 per cent. It is sudden changes in the mica content which gives rise to the banding and lamination. Most psammites contain both biotite and muscovite, but a significant proportion of them (up to a quarter of all thin sections) are completely lacking in muscovite; in the schists muscovite is dominant. Biotite is associated with plagioclase and muscovite with quartz. Equal numbers of sections were found to contain plagioclase in excess of quartz, equal in amount to quartz and less in amount. Microcline occurs locally as an interstitial mineral and is considered to be a late addition (Chapter 6). These modal relationships can be discerned in (Figure 6)a–c, but were established by qualitative modal analysis of hundreds of thin sections. Garnet occurs throughout the psammites and is usually large enough to be visible in the field. Kyanite and staurolite are widely, though very sparsely, distributed but not visible in the field in granofels. Sillimanite has not been found except in the thermal aureoles of the Graven Complex and the Gloup Holm Diorite.

Origin

The psammites appear originally to have been sandstones of medium sand-grain size and are mostly of greywacke type. Nearly all of the metasedimentary rocks fall in the greywacke field, five plot in the Ethic arenite field and only one in the arkose field ((Figure 6)e, after Pettijohn et al., 1973).

Paragneisses have been included in this plot together with the metasedimentary rocks, as it is shown in Chapter 6, and is obvious from (Figure 6)f, that no significant change in composition took place during gneissification. The ubiquitous parallel banding and lamination in the paragneisses is of sedimentary origin although the thickness of the bands and laminae has been reduced by pure shear deformation (producing orthorhombic fabrics, Chapter 7). No grading, cross-bedding or other such sedimentary structures were found. Deformation can be seen to destroy grading in similarly banded rocks in Shetland (Saxa Vord in Unst) except in fold closures, but the preservation of bedding-parallel, very fine laminae makes it less likely that cross-bedding has been destroyed in this way. The considerable thicknesses of seemingly constant sedimentary facies suggests that deposition occurred in a deep basin out of range of shallow-water processes. Further evidence in support of this interpretation is provided by the lack of cross-bedding and of original pinching and swelling of the layers, features associated with shallow-water deposition. The fine grain size (c. 0.3 mm) of the granofels reflects a similar original grain size, because the uniform and ubiquitous preservation of the sedimentary banding and lamination in the metasedimentary rocks precludes much grain growth having occurred. Grain growth during gneissification leads to the destruction of the sedimentary banding, lamination and bedding (Chapter 6). The original relatively fine grain size suggests that the sediments were deposited by distal, sand-rich turbidity currents (Flinn, 1988). The conditions under which the lenticular masses of quartzite and mica schist were deposited are more difficult to determine. Both are coarse rocks in which it is difficult to see original sedimentary features. The fact that they lie within a series of more or less uniform psammites probably indicates that they differ from the latter only in the nature of the sediment supplied to the basin and not in the depth of the basin. The association of quartzite and mica schist is too close to be a random product of the deposition, but cannot be due to a simple sorting of a psammitic sediment into two components as the two together contain less feldspar than the psammites; these lithologies may be due to changing sediment input to the basin.

Boundary Zone

Psammitic granofels

The rocks of the Boundary Zone are intermittently exposed on the coast of north-east Yell, where they occupy a narrow zone up to a few hundred metres wide between the Valayre Gneiss and the Hascosay Slide. A traverse through a part of the Boundary Zone is provided by the coastal section on the north coast [HP 537 053]; there are further partial sections to the north and south of Ness of Cullivoe, and on Hascosay. In these areas the rocks are seen to be handed and laminated psammitic granofels and uniform (unlaminated and unbanded) bands of micaceous granofels. Similar rocks occur to the west in the Yell Sound Division, but there the micaceous granofels tend to be laminated and banded. These rocks become increasingly finely laminated, schistose and lineated as they approach the Hascosay Slide and become conformably injected by thin (several centimetre) bands of aplite. All are blastomylonitised in the Hascosay Slide zone.

A much broader tract of the Boundary Zone occurs in south-east Yell, where very good sections are provided by the cliffs of Ness of Quheyin [HU 540 850] and the cliffs from Ness of Gossaburgh [HU 530 830] to Heoga Ness [HU 520 790]. In these cliffs the metasedimentary rocks are of the same types that occur farther north, but here the uniform micaceous granofels underlies large tracts, and may form a continuous band from Ness of Quheyin [HU 540 853] to Bay of Winnifirt [HU 534 827] and Burravoe [HU 520 793]. The laminated and banded quartzofeldspathic granofels occur on either side in the Broch of Aywick area [HU 547 877], west of Ness of Quheyin [HU 543 855] and Heoga Ness [HU 530 790].

The micaceous granofels generally contain very rounded grains of plagioclase (and very rarely of microcline and then only in certain areas) 1 to 2 mm across in a matrix of quartz and mica with a grain size between 0.2 and 0.3 mm. In thin section these large grains are seen to be single crystals and to contain biotite inclusions of the same size and colour as those in the matrix. The biotites in the matrix wrap round them very closely (S90966); (S90864) and give them the appearance of porphyroclastic relics from the sedimentary state. However, for reasons listed immediately below they are interpreted as microporphyroblasts. They vary in amount and size from absent or so sparse and so small that they are only apparent in thin section, to so very profuse and coarse that the rocks have the appearance of gneisses. The latter are well displayed in Ness of Quheyin [HU 540 853], a zone to the north and south of Point of Whitehill [HU 536 821], and the Green Holm [HU 516 787]. The gneiss-like masses grade outwards into normal micaceous granofels containing no microporphyroblasts by decreasing content of such grains. The microporphyroblasts never show a banded or sedimentary-type distribution, such as might be expected of porphyroclastic grains and they are always composed of feldspar. The granofels containing them rarely show handing and such handing as is visible is very weakly developed.

A sedimentary origin for the granofels is supported by the occurrence in them of the Gossaburgh diamictite, a cluster of dropstone-like clasts of ultrabasic igneous rocks. The clasts arc described in Chapter 5 and the occurrence of the clasts in Chapter 2. The wide and sparse distribution of the clasts, the great variation of clast size, strongly skewed towards tiny fragments, and the angular shape of all except the talcose metasomatised fragments point to their emplacement by a sedimentary process. Most likely, they are glacially transported drop-stones derived from an exposed ultramafic complex.

The granofels are considered to be metasedimentary rocks which have been locally gneissified by the production in them of the microporphyroblasts. This gneissification is discussed in Chapter 6.

Calc-silicate rocks

The Boundary Zone in Yell contains several occurrences of calc-silicate rock, but no limestone (Figure 4). In the Central Shetland Sheet impure limestones occur in the Boundary Zone between Grobs Ness and East Burra Firth ( [HU370 640], [HU373 623] and [HU367 576] ), and these may be the source of the limestone slices in the Nesting Fault in Samphrey.

In Hascosay several calc-silicate nodules occur in the psammites [HU 555 915]. Farther north, thin centimetre-scale conformable ribs of talc-silicate rock occur at [HP 549 013] and [HP 542 036]. In the Gossaburgh region, at [HU 530 822], a 12 cm xenolith-like fragment of siliceous rock occurs in paragneiss (S90997). The mineral content of all these rocks is very similar to that of the talc-silicate rocks in the Yell Sound Division. Both the nodules and the ribs contain assemblages of quartz, garnet, zoisite, plagioclase (oligoclase-andesine), and pale-green hornblende, with accessory sphene.

Quartzites

Minor bands of quartzite, laminated and banded, occur in north-east Yell (Figure 4), and more substantial bands are present to the north-east of Aywick [HU 547 884], [HU 546 877] and [HU 544 871] ), where quartzite is interbanded and interlaminated with psammite and is closely associated with microcline-megacryst gneiss bands to the east of the Valayre Gneiss.

Mineralogy of the Yell Sound Division and the Boundary Zone

There is no apparent difference between the metamorphic condition of these two bodies of rock so it is convenient to consider them together.

Micas

Biotite occurs in practically all the rocks of metasedimentary origin, unless it has been altered to chlorite by late retrograde metamorphism. The colour varies from reddish brown through brown to yellow-brown. Of some 700 thin sections examined about 300 contained reddish brown biotite, 300 yellow brown and 100 brown. Two sections were found to contain foxy red biotite typical of thermal aureoles and these came from the edge of the Graven Complex; one also contained fibrolite and the other shimmerised fibrolite ((S90683); (S90706) respectively). Such biotites are common to the south-west of Yell in enclaves within the Graven Complex. Apart from the foxy-red biotites no regularity could be discerned in the pattern of distribution of the different coloured biotites, either geographically or with respect to the gneissified or ungneissified state of the rocks. Some biotite analyses are provided in (Table 1) and in (Figure 7)a. They are very similar in composition to biotites from the hornblende schists and the Hascosay Slide, but plot in two separated groups. This difference in composition could riot be related to lithology or geographic position. Analyses of muscovite and paragonite are given in (Table 2). The paragonite was found by chance in a kyanite schist when microprobing muscovites (S91696). It was not possible to see any optical difference between paragonite flakes and muscovite flakes, so paragonite may occur in any of the rocks sectioned or microprobed.

Biotite is the dominant mica in the granofels and muscovite in the schists. With increasing amounts of mica in the rocks muscovite forms an increasing proportion ((Figure 6)d). In 245 thin sections of ungneissified granofels examined, 134 contained more biotite than muscovite, 50 less biotite than muscovite, while 66 of the former were muscovite free and only two of the latter were biotite free. Out of 80 schists 14 had more biotite than muscovite, and 44 had less biotite than muscovite. (Note: Out of 470 thin sections of gneissified and partially gneissified rocks 285 had more biotite than muscovite including 52 with no muscovite, and only 42 had less biotite than muscovite.) Nearly all the several hundred metasedimentary rocks sectioned have a smaller volume of micaceous minerals than of quartz plus feldspar; this is also true of the schists, only a few of which contained more micas than quartz plus feldspar. In all thin sections containing more mica than quartz plus feldspar, muscovite occurred in greater volume than biotite. However, a higher proportion of schists than granofels were observed to have more or less equal amounts of mica relative to quartz and feldspar.

Plagioclase and quartz

Out of 340 thin sections of ungneissified metasedimentary rocks those with more plagioclase (oligoclase to andesine) than quartz, those with equal proportions and those with less occurred in more or less equal numbers. (Note: of 470 gneissified and partially gneissified metasedimentary rocks 230 had more plagioclase than quartz, 200 had equal amounts and only 40 had less plagioclase than quartz.) K-feldspar occurs in metasedimentary rocks only as a late additive. Its occurrence is considered at length in Chapter 6 under the heading 'interstitial microcline'.

Garnet

Garnet is a prominent component of the metasedimentary rocks of Yell and appears in three quarters of the thin-sectioned rocks in amounts up to about 20 per cent. Out of 340 thin sections of metasedimentary rocks 241 contained garnet. (Note: out of 470 thin sections of paragneisses and partially gneissified metasedimentary rocks307 contained garnet.) Garnet is associated with muscovite in preference to biotite and with micas rather than feldspar and quartz. All model thin sections found to contain more than 10 per cent garnet were of rocks containing more muscovite than biotite and more mica than quartz plus feldspar.

Garnet appears to occur more profusely in the north half of the island than in the south half, but the difference is not very pronounced and difficult to estimate due to the uneven distribution of exposures. However, west of the Nesting Fault, and in Samphrey, garnet is less common than in Yell and this decreased garnet abundance continues farther west through the Graven Complex, within which there is very little garnet in the enclaves of psammite and paragneiss, and continues as far south as Brae in the Central Shetland Sheet. South of Brae and in Lunna Ness garnet is as common in the Yell Sound Division as it is in north Yell, so its near absence in and around the Graven Complex may be a thermal aureole effect.

Garnets containing tiny TiO₂ inclusions (identified by microprobe analysis) are a characteristic feature of the Yell Sound Division in Yell. The TiO₂ takes the form of crystals much smaller than the thickness of a thin section, generally about 10 microns across, but of sufficiently high birefringence to show moderate birefringence colours. At magnifications of x 50 they appear as tiny points of coloured light in the garnets when viewed between crossed polars. The relief of these embedded grains is apparently less than that of ruble embedded in garnet (S91830), so they are possibly anatase.

Within Yell such garnets are most common between Graveland and Gloup Voe, but were not found south of Yell in the Central Shetland Sheet or in the Boundary Zone. Of 548 thin sections of metasedimentary rock and paragneiss containing garnet 130 contained TiO₂ inclusions in more than trace amounts, but they were found in about 250 altogether. It should be noted that similar inclusions were found in garnets in several thin sections of quartzofeldspathic gneiss from Lewisian inliers (S91644).

The TiO₂ inclusions are not apparent in all garnets within a thin section. Within individual garnet grains they vary in amount from a single crystal to concentrations of dozens. Most commonly they form small clusters (S94718). Commonly, by their presence, they define an outer zone of the garnet, leaving an inclusion-free inner zone ((S91633); (S91776)). Many of these outer zones are incomplete, and look as if they have been partially removed ((S91749); (S91237)). In other garnets these inclusions are arranged in a linear manner across the whole garnet, but the orientation varies between garnets and is oblique to the present schistosity ((S91775); (S92075)). However, in one section (S94731) the inclusions are parallel, both in and between garnets, but oblique to a late schistosity (S_B, see Chapter 7). In (S91300) they define an S-shaped pattern of snowball-garnet type and in (S94713) a U-shaped pattern.

Zoning defined by other types of inclusion also occurs in garnets. Several garnets with cores rich in tiny inclusions of quartz and indeterminate material, surrounded by clear inclusion-free rims were found (S91038). One garnet (S91177) was found to contain fine, dusty inclusions defining a cruciform pattern indicating growth of a rhombdodecahedral form, probably very early in the metamorphism. Similar growth in a calc-silicate rock from nearby was noted above. More common are clear, inclusion-free cores with an outer zone containing inclusions. Most of these contain TiO₂ inclusions but others contain quartz, biotite, aluminosilicates or various indeterminate minerals ((S91043); (S91258)). Microprobe analysis revealed only very weak cryptic zoning in the garnets (see below).

Besides the TiO₂ inclusions and the quartz and biotite inclusions, 22 thin sections (out of 548 containing garnet), including 7 of mica schists, contain garnets with inclusions of kyanite ((S92128); (S94704)), staurolite ((S91830); (S91026)) or both ((S90938); (S91300)). Most of these inclusions occur in garnets with no zoning apparent in thin section and are often associated with TiO₂ ((S91843); (S92116)). In a single occurrence kyanite was found enclosed in an inner TiO₂ zone in a garnet which also had an outer TiO₂ zone.

The vast majority of the garnets, however, are ragged grains with inclusions of quartz and biotite arranged in no recognisable pattern and often taking up so much space that the garnet is almost skeletal. Within each thin section a variety of the different types of garnet may be seen. Even allowing for the different appearance of garnets cut at different distances from their centres, it is clear that few rocks contain only one type of garnet.

The relationship of the garnets to the micas varies from rock to rock. The schistosity defined by the micas opens to enclose many garnet grains (S91789); (S90846) which had, therefore, grown earlier than the schistosity, while other garnet grains have grown across the schistosity ((S91484); (S91181); (S91936)), indicating late growth. Many garnets truncate some of the micas forming the schistosity, and yet the schistosity opens to some extent to enclose the garnet (S91300); (S91319). Such occurrences are considered to result from alternate episodes of garnet growth and schistosity development. This even occurs with garnets whose boundaries are crystal faces resulting from late growth or overgrowth. A variant of this relationship occurs in rocks in which the schistosity opens to enclose a group of garnets contained in a patch of non-schistose matrix (S91817); (S92073). In one such example (S91083) several small garnets of later growth cut the micas forming the schistosity.

The compositions of sonic garnets are given in (Table 3) and in (Figure 7)b. Differences in composition were found between the cores and the edges of some of the garnets analysed and the edges were found, but the differences were not large and showed little similarity from one thin section to another. For example in different samples Mn was found to be richer in the core than the rim, richer in the rim than the core and uniformly distributed. Similar variations were found also for Mg, Fe, and Ca. Nevertheless, in eight out of nine sections use of core analyses with matrix biotites gave slightly higher temperatures with biotite-garnet geothermometers than were given by the use of rim analyses and matrix biotites (see below). As shown by (Figure 7)b the garnets from the metasedimentary rocks are significantly poorer in Ca than the garnets from the hornblende schists.

Aluminosilicates

Aluminosilicates (including staurolite and shimmer aggregates) were found in 78 of 800 thin sections of non-gneissic and gneissic metasedimentary rocks, 12 of them from mica schists. They came from localities widely distributed over the area (Figure 8), and occurred both within and outwith garnets even within the same thin section. Of particular interest, and noted above, are the kyanite aggregates which replace chiastolite at Gerherda. In general, both kyanite and staurolite in the matrix occur in two ways which, taken together with the occurrence as inclusions in the garnet, indicate several generations of growth. Both minerals occur as grains contained within and aligned with the schistosity, and as grains growing across arid truncating the micas forming the schistosity. In many of these latter occurrences this growth takes place preferentially on either side of garnets where the mica schistosity is strongest and is tangential to the garnet ((S91277); (S90846)). Two staurolite analyses are included in (Table 2).

No sillimanite has been found, except within the aureoles of the Gloup Holm Diorite and the Graven Complex. The sillimanite on Gloup Holm was fibrolite but is now shimmerised. Associated with the Graven Complex on Bigga both fibrolite and fibrolite clusters (faserkiesel) occur, and farther west in the Graven Complex, sillimanite crystals occur in metasedimentary xenoliths.

Shimmer aggregates occur and nearly all are associated with kyanite, either replacing the kyanite or, in larger patches, possibly after andalusite, now being replaced by kyanite aggregates.

Sequence of mineral growth

Thin section examination shows that the various types of minerals grew at different times both in different parts of Yell and in the same localities, but there is no evidence that enables them to be ordered in a sequence of separable metamorphic episodes. In Chapter 7 it will be shown that folds of different type occur in much the same way. Metamorphism and accompanying deformation was probably more of a continuous process than a sequence of separate events.

Geothermometry and geobarometry

Temperature determinations based on four different published biotite-garnet geothermometers and using nine specimens of rocks from the Yell Sound Division gave an average temperature of about 650°C (Table 4). The temperatures obtained are all shown in (Table 4) without weeding or readjustment, thereby giving rise to a range of values that is not unusual for such a wide spread of specimens and geothermometers. If the several impossibly high values (around 800°C) and unlikely low values (around 500°C) are ignored, the remainder show a not unreasonably wide scatter (for such studies) at about 650°C. The geographic distribution of the temperature estimates (both reasonable and unreasonable) is shown in (Figure 9), in which it can be seen that there is no evidence for a temperature gradient over the area. High values tend to occur close to low values indicating that the scatter of the temperatures in (Table 4) is due to experimental error and geothermometer error. It should be noted that little cryptic zoning was found in the garnets, and that the differences between temperatures obtained from the use of rim analyses as opposed to core analyses were insignificant compared with the temperature differences shown in (Table 4). It is noticeable that the Hodges and Spear (1982) geothermometer comes closest to the averages of all four.

The Henley (1970) muscovite-paragonite geothermometer used with the muscovite and paragonite compositions quoted in (Table 2) clearly indicates a temperature greater than 600°C.

The garnet-plagioclase-kyanite-quartz geobarometer (Dachs, 1990) gives a pressure of about 7 kb for a temperature of 650°C (Table 5). This pressure and temperature falls in the same P-T field as that proposed for Moine rocks in Scotland (Barr et al., 1986, fig. 5).

These temperatures and pressures are shown plotted on a map of Yell (Figure 9) together with temperatures and pressures obtained from hornblendic rocks, which are reported in Chapters 5 and 8.

Geochemistry

Thirty-two non-gneissified metasedimentary rocks (lacking microcline) were selected for chemical and modal (point counting) analysis. They were chosen from more than 300 thin sections as the best representatives, both in field occurrence and thin section, of a variety of commonly occurring types of granofels and schist which occur widely in the area. Microcline-containing rocks were excluded as the microcline is a late addition (Chapter 6). The group of analyses as a whole is, therefore, not to be taken as representing the lithology of the area quantitatively, but as representing the range of rock types composing the lithology. The methods of analysis employed, the analyses and the modes are listed in Flinn (1993).

A plot of the modal quartz + feldspar content of the 32 rocks against their SiO₂ content shows a significant correlation, but the distribution is such that for a given SiO₂ per cent the modal values are scattered over 10 or 15 per cent on either side of the expected modal percent. Four coarse mica schist samples were among those analysed, three from the Kaywick Schist and one from a coarse mica schist band in the Mid Yell Schist. Despite their similar field appearance, i.e. very micaceous and flaky, these contained between 35 and 70 per cent quartz and feldspar by volume and between 58 and 66 per cent SiO₂ by weight. It is usually possible to recognise in the field a granofels very rich in quartz and feldspar as a psammite, and a rock which is extremely rich in mica as a pelite, but very schistose rocks are sometimes found to have similar modes to granofels and some granofels have modes similar to schistose rocks.

The classification of Moine rocks from the mainland of Scotland in such terms as psammite, pelite and semipelite has been based on their content of SiO₂ (Winchester et al., 1981; Winchester, 1984), even though the SiO₂ content cannot be estimated in the field, and can only be guessed at in terms of quartz content. In the field, the identification of Yell metasedimentary rocks as pelites, semipelites, and psammites on the basis of their quartz and feldspar content is difficult and, at best, very approximate; therefore, the Scottish classification was not used in the selection of specimens for analysis.

However, for comparison with rocks from the mainland of Scotland, the analysed Yell Sound Division metasedimentary rocks have been divided into two groups - those with a content of less than 67 per cent SiO₂ and those with more. The averages for the two groups are presented in (Table 6), and may be compared with average analyses of Glenfinnan Division and of Loch Eil Division rocks grouped in the same way, into 'pelites' and 'psammites' by Winchester et al. (1981, table 1) . Differences are many and varied.

Chapter 5 Early igneous rocks

Early basic igneous rocks of three different types occur in Yell. They are termed 'early' because they were emplaced before the end of the regional metamorphism. A fourth occurrence, that of small veins of metabasaltic rock emplaced in the Hascosay Slide at the very end of its formation and of metamorphism in the area, is described in the section on the Hascosay Slide (Chapter 8). Two of the types of early basic igneous rock occur in the Boundary Zone. These are the clasts of the Gossaburgh metadiamictite and the ophitic globular metadolerites. The former consist of a scattering of clasts of both basic and ultrabasic igneous rocks in psammitic granofels and the latter a series of globular masses of ophitic dolerite in the same psammites. The third type of early basic igneous rocks are the hornblende schist bands in the Yell Sound Division and the Boundary Zone, which occur in a manner very similar to that of the amphibolite/hornblende schist bands characteristic of the Glenfinnan and Loch Eil Divisions of the Moine of mainland Scotland. Granitic and granodioritic orthogneisses, including the Breakon Gneiss, appear to have originally been intruded as granitoid bodies into the rocks of the Yell Sound Division, either before, or at an early stage of their metamorphism, and to have been deformed and metamorphosed to orthogneisses.

Clasts in the Gossaburgh Metadiamictite

On the east side of the Ness of Gossaburgh, between Groti Geo and the Bay of Winnifirt ([HU 535 827]; (Figure 10)) the granoblastic psammitic rocks of the Boundary Zone contain a scattering of clasts of basic and ultrabasic rocks. These fragments vary from angular to rounded and range in size from a centimetre or two across to several metres. They occur widely scattered over an area extending for about 400 m along the coast and for about 50 m inland from the sea. There is no discernable boundary to the deposit, merely a lack of any further fragments. They are rarely less than a metre apart, but most occur in the general area of the largest and most prominent block which is easily found at locality [HU 536 830]. This is an angular block of variably, and partially, amphibolitised banded-peridotite several metres across ((S90982); (S90984); (S94695)). It occurs close to the second largest block, one of serpentinite which is less than a metre across (S90986). The most commonly occurring fragments are hornblende-rich and are rarely more than several centimetres across. Less common are larger fragments of serpentinite ((S91003); (S94697)) and altered serpentinite varying from 10 to 30 cm across. The latter are partly to completely altered, in some cases in a concentric zonal manner, to such minerals as talc (S90981), chlorite, actinolite (S90987), and brown mica. These altered serpentinite fragments tend to weather out very easily, so that many of them have been more or less completely lost, leaving characteristically rounded cavities in the surface of the exposure containing, in some cases, remnants of talc or actinolite. Very rare small fragments of pyroxenite (S90983) also occur.

The serpentinite and altered serpentinite fragments are generally rounded and elongated in the direction of the schistosity of the enclosing rock. The other types of fragment are angular or subangular, and are enclosed within the country rock schistosity, but in some cases their long axes are oblique to it rather than aligned in it, the sense of obliquity varying from specimen to specimen. In several cases fragments appear to have been broken and separated during deformation and metamorphism of the enclosing rock. There are no sedimentary structures or banding in the enclosing rocks and the distribution of the fragments shows no signs of a sedimentary pattern. (Table 7) Analyses of minerals from clasts in the Gossaburgh metadiamictite Thin sections of the hornblende-rich fragments show them to be composed of brown hornblende and plagioclase (S91004) with the brown hornblende changing to green hornblende on the outside of the fragment (S91006), or of green hornblende throughout. Most show no preferred orientation of the hornblende ((S91004); (S91001)), but a minority show hornblende exhibiting a preferred orientation (S94691). Several contain garnet (S94696). The grain size of the hornblende generally varies between 0.2 and 0.4 nun.

The 1 to 2 cm pyroxenite fragment (S90983) is revealed in thin section to be an equigranular mixture of bronzite (En₈₀) and a very pale green magnesio-hornblende. Occasional grains of hornblende contain inclusions of pyroxene, otherwise the two minerals appear to have formed at the same time. Small, interstitial and partially serpentinised olivines also occur (Fo₇₉). The serpentine is yellow-brown (2Fe to 240). A trace of phlogopite ((Figure 7) a) occurs together with dull brown chromiferous spinels (Cr/ (Cr + Al) = 16 per cent, Fe*/ (Fe* + Mg)= 54 per cent). Analyses arc given in (Table 7).

Serpentinite fragments vary from uniformly fine-grained lizardite-like aggregates ((S94694); (S91003)) to lizardite-mesh patterned serpentinites which appear to have pseudomorphosed foam-cell-like equigranular olivine grains (S94697). The analysed specimen ((Table 8), (S94694)) is an antigorite-talc rock.

The large banded-peridotite block ((S90982); (S94695); (S90984)) is mostly composed of olivine (and serpentine-pseudomorphs of olivine) and a colourless magnesiohornblende (Table 7), both minerals being similar in composition to those in the pyroxenite fragment (S90983). They occur intergrown in different proportions to give the banding. A small amount of orthopyroxene occurs together with three types of serpentine. A coarse lizardite pattern and a fine lizardite pattern pseudomorph large relic grains of olivine. A yellowish brown variety of serpentine occurs in smaller quantity and is associated with the orthopyroxene relics.

All the clasts in the Gossaburgh diamictite appear to have been derived from a partially serpentinised ultrabasic complex. Some of the hornblende may be the result of metamorphism after the clasts were enclosed in their present matrix. The alteration of the serpentinite fragments to such minerals as talc and actinolite is the result of such a late metamorphism.

Hornblende schists of the Boundary Zone

Hornblende schists of the Boundary Zone in Yell contain green or olive-green hornblendes, often in very pale shades due probably to late alteration. Some sectioned hornblende schists were found to contain a reddish brown hornblende (S90886). However, most of the reddish brown hornblende in Yell occurs in the cores of the metadolerite 'balls' (Chapter 9), and to a lesser extent in the schistose envelopes of the 'balls', and in the fragments in the metadiamictite. The hornblende schists containing the reddish brown hornblende may be small, unrecognised occurrences of the metadolerites.

Reddish brown hornblendes are an important feature of parts of the Boundary Zone outwith Yell. In Lunna Ness, in the Central Shetland Sheet, the Boundary Zone is largely composed of nebulitic gneisses and quartzofeldspathic orthogneisses rich in hornblendite bands and 'fish' and all the hornblende (other than that which has suffered late alteration) is coloured a strong reddish brown. This colour contrasts strongly with the green or olive-green hornblendes (when unchanged by late diapthore tic alterations) characteristic of the Yell Sound Division and the Scatsta and Whiteness Divisions (Lower and Middle Dalradian), and with the blue-green feathery-like hornblende grains characteristic of the Clift Hills Division (Upper Dalradian).

In the Central Shetland Sheet, reddish brown hornblende is not present in the Boundary Zone south of the Lunna Ness area and east of the Nesting Fault. It occurs on the south edge of the Yell Sheet in the Neapaback Skerries, immediately north of Lunna Ness, in rocks like those in Lunna Ness. In the Boundary Zone to the north of the Neapaback Skerries reddish brown hornblende is confined to the metadolerites and is probably unrelated to that in Lunna Ness. The Lunna Ness reddish brown hornblende zone is considered to end immediately north of the Neapaback Skerries, at an old tectonic junction within the Boundary Zone.

The envelopes of the metadolerite 'balls', the schistose hornblendic bands closely associated with them, and indeed the rims of the reddish brown hornblende grains within the metadolerite 'balls', are generally composed of green-brown hornblende.

No occurrences of garnet-studded hornblende schist were observed in the Boundary Zone in Yell and those hornblendic rocks which were found by thin sectioning to contain garnet probably belong to the metadolerites.

The rocks of the Neapaback Skerries are rich in bands, lenses and 'fish' of garnetiferous reddish brown hornblende schists floating in nehulitic gneisses like those to be found in the Boundary Zone, east of the Valayre Gneiss, in Lunna Ness.

Hornblende schists of the Yell Sound Division

Hornblende schists are a characteristic component of the Yell Sound Division throughout its outcrop in both Yell and the Central Shetland Sheet to the south. They are seen to be somewhat unevenly distributed in the cliffs of Yell (Figure 10); the uneven and sparse nature of the exposure inland prevents an estimate being made of their frequency of occurrence there. Throughout the Yell Sound Division in Shetland these rocks are characteristically studded with garnets varying in diameter from 2 to 20 mm. Hornblende schists in the East Mainland Succession east of the Valayre Gneiss are never garnet studded and only extremely rarely contain garnet.

Hornblende schists are unevenly distributed in the psammites of the Yell Sound Division, arc extremely rare in the mica schists and absent from the quartzites. On the west side of Ness of Sound [HU 447 823] and on the east side of Wick of Copister [HU 476 785] hornblende schist hands are so closely spaced and interbanded with the psammites as to be the dominant rock over thicknesses of several hundred metres. Coastal sections in which hornblende schist bands (average thickness probably less than 1 m - see below) occur at a concentration of up to 1 per 10 m are to be found at the entrance to Gloup Voe [HP 500 050] and either side for a distance of a kilometre, Ness of Houlland [HP 525 055] and for a kilometre to the east, Stuis of Graveland [HU 460 970], the 3 km of coast to the north of West Sand Wick [HU 446 892], the coast for 2 km north of Head of Brough [HU 445 850], the east coast of Bigga and the east coast of Samphrey. Coastal sections where the hornblende schists are absent or almost absent are the 5 km to the south of Head of Bratta [HU 475 990] and the 2 km on either side of the north-west corner of Yell [HP 475 050]. Except for these lengths of coast line the hornblende schist bands occur at concentrations of 1 or 2 per km.

Most of the hornblende schists are parallel-sided bands which vary from a centimetre or two to a few metres in thickness. A substantial proportion, perhaps a third, are flat and lenticular, that is, bands which pinch out along the strike. Well-developed boudinage was not observed. On the west shore of Whalfirth a spherical mass several metres in diameter occurs ([HU 464 937], (S91474)) and farther north, at locality [HU 464 967], several sac-form masses up to 7 m in thickness occur, including one in which can be seen folding of fine feldspathic laminations, and another with a coarse-grained core. Similarly large masses occur on the east coast of Bigga ([HU 450 788] and [HU 447 793] ), Burra Ness [HU 510 790], and Bligg [HP 492 054].

Nearly all these bodies are conformable to the schistosity and banding of the psammites and gneisses. However, two, otherwise apparently typical, garnet-hornblende schist bands cut the banding in the psammites and gneiss near Geo of Markamouth (S91815), [HP 475 008]; (S94717), [HP 475 023] respectively).

Several garnet-hornblende schist bands close to the Graven complex on Bigga have been metasomatised to coarse garnet-biotite schists ([HU 444 797]; (S90679); (S90691). Nearby, several hornblende schist and garnet-hornblende schist bands have developed biotite selvages; for example those exposed Uynarey [HU 443 805], at Wester Wick of Copister [HU 473 786] and on the east end of Samphrey [HU 470 760]. Two other garnet-biotite schist bands were observed elsewhere in Yell (S92160). These garnetbiotite schists of metasomatic origin are easily recognisable in the field because of their very high content of unusually large biotite flakes. Many occur in close proximity to the Graven Complex in several localities on the Central Shetland Sheet.

Petrography and mineralogy

Some 60 per cent of about 400 hornblende schist hands seen in the field contained garnet sufficiently large to be immediately obvious. In several areas hornblende schists are notably deficient in prominent garnets such as those at Stuis of Graveland [HU 460 970], Ness of Sound [HU 447 825], and Burra Ness [HU 513 788]. Fifty-six per cent of 90 thin sections of hornblende schist contained garnet, some too small to be easily visible in the field. Garnet is confined to the edges of some hornblende schist hands.

Nearly all the thin sections of hornblende schist examined had no trace of relic igneous texture, the hornblende, garnet and feldspar being uniformly distributed and hornblende having a good lattice preferred orientation. However, in several thin sections from Whalfirth the feldspar grains showed a poorly defined clumping which could possibly he a relic ophitic texture (S91617); (S91613); (S91464). The indications of an original ophitic texture are so poor that the possibility became clear only after a study of the complete range of altered ophitic metadolerites described below. Such textures were mostly found in fat lenses or masses that contained garnet and showed a preferred orientation of the hornblende grains, features which distinguish them from the ophitic metadolerites of the Boundary Zone.

Most of the hornblende schists examined in thin section exhibited a preferred orientation of the hornblende grains, either of grain shape and lattice ((S91525); (S91779)) or of lattice alone (S94725) but a few showed neither (S91740). Equigranular grain shape makes it difficult to recognise the presence of schistosity and lineation in hand specimen. Thin sections cut normal to the lineation, especially in L-tectonites, show little or no preferred orientation of grain shape or lattice although a preferred orientation of hornblende shape and/or lattice is visible in thin sections cut parallel to the lineation. Therefore, tectonites and nontectonites cannot always be distinguished in the field or even in thin section. Since the great majority of these hornblendic rocks appear to be tectonites they have all been called hornblende schists.

The hornblende-lattice fabrics observed varied from Ltectonites with a parallel alignment of hornblende c-axes, to L-S tectonites with a parallel alignment of c-axes and a partial alignment of b-axes (partial girdle), to Stectonites with an unrestricted alignment of c- and b-axes in the s-plane. The hornblende-grain-shape fabrics varied from elongate parallel to the c-axis to equigranular. Hornblende grains are only elongate in rocks with a preferred lattice orientation. The distribution of the different types of fabric shows no apparent relation to any other features of the geology of Yell.

In the hornblende schists the hornblende grains vary in size from 0.2 to 1.0 mm and are rarely more than twice as long as broad. In thin section the hornblendes are brown ((S91929); (S94689)), (but never reddish brown as in the Boundary Zone) or brownish green (S91491) or green (S91833). Bluish green hornblendes ((S94815), (S91220)) are less common, but pale washed-out colours occur widely. The latter are interpreted as the result of late alteration. The brownish colours are dominant throughout the Yell Sound Division. The bluish green hornblendes occur widely but very sparsely; only in Samphrey did all the sectioned hornblende schists have this bluish green colour.

Garnet is very common in many thin sections. Some are inclusion free, except for such accessories as rutile and opaque minerals, but most are rich in inclusions. These occur in two ways. Most common are garnets containing ragged grains of quartz and, less frequently, feldspar and hornblende (S91883). Such inclusions vary in number from a few to be so closely packed that the garnet is merely a ragged interstitial net (S91697). The inclusions are generally smaller than the grains in the rock matrix. Less common are garnets containing closely packed small, well-formed, elongate hornblendes in a parallel arrangement, together with prominent twinned plagioclases, both of which are unmatched in the matrix of the rock (S92112). Many garnets have an inclusion-free outer rim which is not sharply separated from the inclusion-rich core. One specimen contains atoll garnets (S91221).

The relationship of the garnets to the hornblendes of the rock enclosing them is many-fold. In most hornblende schists the schistosity defined by the parallel alignment of the hornblende lattices and grain shapes open to enclose the garnet in an 'eye' ((S94689); (S91779)). Where the inclusions in such garnets define an internal fabric this is usually rotated and oblique to the outside fabric, the internal fabric in some cases being parallel in all garnets (S92112). In some thin sections the garnets appear to be contemporaneous to a strong hornblende fabric (S94685) and in sample section (S91697) the inclusions in such a garnet exhibit a crudely spiral arrangement. Most garnets which appear to have grown contemporaneously with the hornblende fabric occur in thin sections showing no preferred orientation of hornblende grain shape or lattice (S91075). Where the hornblende forms equigranular grains but the lattices show a preferred orientation the latter is usually enhanced where it is tangential to the garnet boundary ((S92175); (S94725)). Sections in which the garnets exhibit good crystal outlines were from rocks with a poor or absent preferred orientation of hornblendes (S91792), as were garnets which appeared to cut (postdate) the hornblendes ((S94717); (S91895)), including some with well-formed faces (S91835). These late garnets are especially rich in inclusions of hornblende. Garnet analyses are given in (Table 9) and (Figure 7)b.

The wide variety of fabric relationships exhibited by the hornblende schists does not appear to be systematically distributed over Yell. This is considered to be the result of repeated rejuvenation of the fabrics, itself the result of a sequence of episodes of deformation and recrystallisation in which the garnet remained unaffected or grew larger while the hornblendes recrystallised. It was not found possible to distinguish a particular sequence of episodes or to relate any given state to an episode registered in the enclosing rocks. The garnets in the hornblende schists and those in the metasedimentary rocks exhibit very similar relationships to the matrices enclosing them.

Biotite occurs as an accessory mineral in many hornblende schists and occasionally as a major constituent. The grains generally show a better preferred orientation, parallel to that of the hornblende, than the hornblende. As mentioned above, on Bigga several hornblende schists with biotite selvages were observed as well as biotite-garnet schists, formed by the replacement of hornblende by biotite in hornblende-garnet schists ((S90679); (S90691); (S92160)). The biotite from one such rock, occurring out-with Yell, has been dated as 391 + 12 Ma (Chapter 10), close to the age of 400 Ma obtained for the Graven Complex (Miller and Flinn, 1966).

Plagioclase and quartz are essential constituents of the hornblende schists, generally in amounts very much smaller than that of hornblende. Epidote (e.g. (S94813)), zoisite (e.g. (S91508); (S91474); (S91809)), sphene, rutile and opaque minerals occur as accessories.

Analyses of minerals from Yell Sound Division garnet-hornblende schists outwith Yell are presented in (Table 9) and (Figure 7)a. In L40557 the amphibole is greenish brown and has the composition of a tschermakitic hornblende (Table 9), plagioclase has the composition of approximately An_a) and the opaque mineral is ilmenite. The biotite and garnet compositions are shown plotted in (Figure 7). The cores of the garnets were found to contain about 2 per cent more MnO than the rims.

Geothermometry

(Table 10) shows the range of temperatures provided by a series of geothermometers for a hornblende schist belonging to the Yell Sound Division outwith Yell. The higher values are more in agreement with temperatures obtained from the metasedimentary rocks (Figure 9). Temperatures based on analyses of the cores of the garnets are usually 10 or 20°C lower than those based on analyses of the rims.

Geobarometry

The Kohn and Spear (1990) garnet-hornblende geobarometer gives a pressure of about 5kb (Table 11) for a temperature based on the garnet-hornblende geothermometer of about 600°C. Both the temperature and the pressure determinations are less than those obtained from the metasedimentary rocks and fall outside the kyanite field. Therefore, the determinations from the psammitic rocks are preferred.

Geochemistry

No analyses of hornblende schists from the Yell Sheet are available. However, two analyses of hornblende schists from the Yell Sound Division of the Central Shetland Sheet are presented here as they appear to be in no way different in thin section or field occurrence to those examined in Yell (Table 8). The two analyses of typical garnet hornblende schists plot within or close to those fields in discriminant diagrams appropriate to subalkaline tholeiitic basalts with MORB affinities (Figure lla-h), except that they have remarkably low Sr contents. Their position in the plots in (Figure 11) is similar to that of analyses of amphibolites and/or hornblende schists from the Glenfinnan Division of the Scottish Highlands (Winchester, 1984; Lindsay, 1988).

Interpretation

Nowhere were rocks found which, in mineral content or appearance, were intermediate between hornblende schists and the metasedimentary rocks. All boundaries to the hornblende schists are sharp. This feature, together with their composition and the preservation of possible relic igneous textures, points to a basaltic igneous origin.

Origin

The hornblende schists of the Yell Sound Division are considered to he subalkaline tholeiite intrusions of MORB affinity which were emplaced in consolidated sediments either before or early in the metamorphism. It is possible that some were emplaced during later stages of the metamorphism. During the metamorphism they suffered varying conditions of deformation and recrystallisation which led to continuing development of the schistosity while it remained parallel to the edges of the hornblende schist hands and to the layering in the enclosing metasedimentary rocks.

Metadolerites

A series of fat-lenticular to near-spherical metadolerites varying in size from a metre or two in diameter to, more rarely, 10 m in diameter occur along the east coast of Yell in an area extending from the south coast of Hascosay to Burravoe, a distance of some 13 km (Figure 10). The metadolerites occur in the psammitic and granoblastic rocks of the Boundary Zone. The hornblende schists of the Yell Sound Division are generally parallel-sided bands, but grade occasionally into fat-lenticular or sac-form bodies. The metadolerites described in this section grade from fat-lenticular to near spherical bodies, and to emphasise the difference from the hornblende schists will be referred to as globular.

The field occurrence of globular metadolerites is more or less coincident with that of the plagioclase-microporphyroblast development (Chapter 6). Except for some occurrences on the coast west of Burravoe, the metadolerites are confined to the east of the Valayre Gneiss, i.e. to the Boundary Zone. The metadolerites do not seem to continue to the south of Yell, except for a single possible occurrence on the east side of Lunna Ness (Central Shetland Sheet).

The metadolerites may be recognised in the field by their globular shape and in thin section by the relic ophitic texture. However, in many cases metamorphism has destroyed, partially or completely, the ophitic texture by converting the outer edges of the masses to hornblende schist. At Head of Gutcher [HP 551 000] a group of globular masses of basic rock up to several metres in diameter occurs closely associated with the Valayre Gneiss, and large masses of basic rock occur in the Hascosay Slide (e.g. at North Sandwick, [HU 553 970], (S91698)) and both occurrences exhibit the relic ophitic features described below, such as saussuritic feldspar and hornblende-pseudomorphed pyroxenes (S91870).

The metadolerites occur in two main ways. In one type of occurrence they take the form of isolated globular and nearly spherical masses in the psammites and the microporphyroblastic gneisses. The edges of the masses are concordant with the schistosity and the layering which opens up to enclose them. This is especially well seen on the coast in Hascosay [HU 555 915], where the 'balls' of metadolerite throw the flat-lying layered psammites into dome-like rolls, indicating the presence of metadolerite 'balls' even where the 'balls' are not exposed. Similar, but less well displayed, are the occurrences on Heoga Ness [HU 523 790] and north of The Poil [HU 546 876].

In the other type of occurrence the 'balls' and sac-form masses, ranging in size from 1 to 10 m, occur closely packed in fairly well-defined zones some tens of metres thick in association with bands of hornblende schist, hornblende-bearing psammitic granofels and psammite. In places the psammite contains the plagioclase microporphyroblasts characteristic of the adjacent rocks and even quartzofeldspathic streaks and augen. Two such zones are especially well developed and displayed in Revri Geo, north of Burravoe [HU 533 820] and at The Poil [HU 547 873].

All the dolerites have been partially or completely metamorphosed but only the outer edges have been tectonised. The least-altered mass is a 7 m-diameter ball on the south coast of Hascosay ((S94705); (S91353); (S94806)) and an example almost as good occurs in Heoga Ness (S90792). In thin section they are seen to be ophitic dolerites formed of large (several millimetres) augites, well-twinned, unzoned labradorite (An_52-70), and ilmenite together with microgranular aggregates of hypersthene (En₅₇). The ilmenite grains are surrounded by a red-brown ferroan pargasite which also rims the hypersthene aggregates and most of the augite grains. In thin section (S91353) clinopyroxene grains contain exsolution orthopyroxene. In several occurrences both garnet and biotite occur in minor quantities (S90792). Mineral analyses are listed in (Table 12) and (Figure 11)d.

On the south coast of Yell the cores of many of the metadolerite masses are more strongly metamorphosed, but always statically so with no sign of deformation or cataclasis. Such metadolerites ((S94687); (S90740); (S90792)) develop red-brown hornblende (ferroan pargasite (S94705); ferroan pargasitic hornblende (S94687)) around the ilmenite grains and in smaller-sized grains around the augites. The reddish brown hornblende tends to grade to a green hornblende (edenitic hornblende in (S94687)), especially on the grain edges away from the pyroxene core. In most metadolerite masses the metamorphism has progressed to the point where the augite core is no longer preserved and is replaced by rnicrogranular aggregates of pale green to colourless hornblende (siliceous edenite in (S94687)). Such augitc pseudomorphs are very similar to those in gabbro masses in the Hascosay Slide. With increasing metamorphism the labradorite becomes saussauritised and any fresh feldspar which occurs is albitic. In several specimens of altered metadolerite garnet and/or biotite occur (S94690). In small masses and towards the edges of the larger bodies the ophitic texture is so fine grained (or absent) that it is swamped by the re-crystallisation and the metadolerites become red-brown or brown equigranular hornblendites with no preferred orientation of the hornblende and grain sizes as fine as 0.1 mm. Such rocks, except for the hornblende colour, are very similar to the fine-grained hornblendic veins cutting the mylonites in the Hascosay Slide. Only the outermost edges of the metadolerites become hornblende schists, often with garnet and biotite (S91373), and thus become very similar to the common hornblende schist bands. However, they usually contain saussurite patches and traces of red-brown hornblende, like those in the cores of altered metadolerite 'balls'. Some of the saussurite is partly or completely recrystallised to coarse zoisite (S91144).

The metadolerites have been cut by the acid pegmatites (Chapter 9), but, presumably because of their competence, less so than the country rock. On Heoga Ness several have been veined by other rocks. A 10 m ball has been veined by an equigranular (grain size c.0.1 mm) brown-hornblende hornblendite (S90741) very similar to the fine-grained hornblendite veins cutting the Hascosay Slide mylonites. The hornblendite in turn has been cut by a vein of garnet-biotite-plagioclasequartz granular rock (S90742) very little different in appearance to the matrix psammites, except for the lack of a tectonite fabric. A similar psammitic-looking vein, without garnet (S94684), cuts another metadolerite mass nearby. These veins look like mobilised country rock, but they have the same mineralogy (except for garnet) as the tonalites (Chapter 9) which occur in the same area, although they lack the prominent, partly altered, plagioclases characteristic of the tonalites. Tonalite veins have been observed to cut the metadolerites.

Geochemistry

The core of the least-altered metadolerite (S94705) from Hascosay and the core of an altered metadolerite from farther south (S94687) have been analysed (Table 8). When these analyses are plotted on discriminant diagrams, they fall in the fields of subalkaline tholeiitic basalts of MORE affinity (Figure 11). There is a significant difference between the two analyses. The analysis of the more altered of the two rocks conforms more closely to the standard tholeiite composition than the analysis of the less altered of the two. The less-altered metadolerite, (S94705), contains less TiO₂, K₂O, Zr and Rb than the more altered sample, (S94687). Thus the two analyses are separated by different amounts on several plots (Figure 11).

Significant differences are apparent between the two analyses of metadolerites and analyses of two hornblende schists from the Yell Sound Division. Both are subalkaline tholeiites with MORB affinities but the metadolerites fall in different parts of the fields appropriate to such rocks. Comparison of the analyses show the metadolerites to contain about 1 per cent by weight more Na₂O and several per cent less total iron oxide. They also contain noticeably less TiO₂, MnO₂, Zr, Zn, Y and V and noticeably more A1₂O₃, Cr, and Sr. The metadolerites are closer in composition to the late hornblendite veins in the Hascosay Slide described elsewhere (Chapter 8, (Table 21) – analysis 2) than are the hornblende schists.

Interpretation

The globular shapes of the metadolerite bodies, together with the preservation of the undeformed ophitic texture to within a few centimetres of their contacts precludes the possibility of them being tectonically disrupted intrusive sheets. They must have been intruded when the country rock had much the same viscosity as the magma. Such conditions could occur during metamorphism at high temperature or during the consolidation of the sediments. Since the metadolerites occur both in the Boundary Zone and in the Yell Sound Division in the Burravoe area, they must have been intruded after the differing metamorphic states of the various components of the Boundary Zone had been established and after the Boundary Zone and the Yell Sound Division had been brought together. Thus they were intruded at a late stage in the metamorphism and deformation of the host rocks, but before they cooled significantly.

The presence of the ophitic texture indicates that the magma was somewhat hotter than the host rocks and thus cooled moderately rapidly at first, and until amphibolite-facies conditions were reached. Thereafter they suffered amphibolite-facies metamorphism and cooled with the country rocks. The production of basaltic magma from the mantle may have been related to the onset of uplift of the area, leading eventually to the cooling.

It should be noted that within the Yell Sound Division, among the hornblende schists, several sac-formed masses with possible traces of ophitic texture and several crosscutting sheets occur. No evidence was found that proved that all the hornblende schists in the Yell Sound Division are of the same age. Some may be younger and even related to the metadolerites.

No other occurrences of globular metadolerites exist in the metamorphic rocks of Shetland. Large sac-form bodies of basic rock 100 m or more in smallest dimension occur in the Cliftlls Phyllites to the west of Lerwick (Flinn, 1967) and are associated with the metavolcanic Asta Spilites. They appear to have been intruded into the unmetamorphosed sediments.

Orthogneisses

Within the Yell Sound Division are a number of bodies of gneissic rock which in the field appear to be tectonised granitoid intrusives emplaced before, or in the early stages of, regional metamorphism. They differ from the paragneisses in being more uniform and homogeneous, and in having sharp boundaries against adjacent rocks. Weakly tectonised varieties occur, which are transitional to undeformed granitoid intrusives, but not to metasedimentary rocks. The most prominent mineralogical variation between the different occurrences is the microcline content which varies from zero to the microcline being the dominant mineral in the rock. Those rocks in which microcline is an essential and major constituent will be called granite orthogneiss. The other rock occurrences, within which the microcline content varies from absent to about 40 per cent of total feldspar (some of which is of interstitial type), will, for convenience, be called granodiorite orthogneiss. The two types are not always easy to distinguish in the field and both are difficult to distinguish from some of the more homogeneous granular para-gneisses, particularly when seen only in small exposures and or in thin section.

Granodiorite orthogneisses

These include the Breakon Gneiss in the Cullivoe Lens, several smaller sheets (the Graveland Gneiss) in the Graveland Lens and two tectonic slices in the Bluemull Sound Fault (Figure 12). The Breakon Gneiss and the Graveland Gneiss are folded by unusually large-scale complex folds of very irregular and noncylindrical style. No signs of an aureole were detected adjacent to either body. However, on the east side of the entrance to Gloup Voe and on the west coast of Yell, there is relic andalusite ((Plate 3)e) which may have formed during an early, thermal phase of metamorphism (Chapter 4), such as might be expected to result from the intrusion of such bodies into unmetamorphosed rocks.

The Graveland Gneiss ((Plate 4)a) forms the core of The Eigg headland [HU 448 956]. In that headland it takes the form of a vertical sheet about 5 m thick and 50 m high folded over to the west on top and to the east at sea level to form a upright pot-hook-shaped body which is of z-type when viewed from the south. The homogeneity of the gneiss, together with a slight pinkish colour, make it easily discernable from the cliff top to the south. It continues to the south of The Eigg for about 300 m in the cliff (here about 100 m high) as land-slid masses and as exposures. In the fold closures it is an L-tectonite and in the limb an L-S tectonite.

About 2 km to the south, on either side of a small embayment in the cliff top above Whalegeo Stacks [HU 445 935], a 2 to 3 m-thick sheet of very similar rock can be seen. It is not folded, but like the sheet at The Eigg it is conformable with the enclosing rocks which, in both localities, are interbanded thin bedded quartzites and quartz-rich psammites with no evidence of gneissification.

Thin sections show that microcline varies from absent to equal in volume to plagioclase over distances of less than a metre. Quartz varies between 30 and 40 per cent in volume and micas vary between 10 and 20 per cent with biotite usually exceeding muscovite in volume. Apatite is a common accessory. The potash feldspar is called microcline for convenience, even though it only rarely shows good microcline twinning. Mostly it shows a very vague, fuzzy and shadowy extinction pattern of obliquely crossing bands.

The Breakon Gneiss ((Plate 4)b) is a much larger more coarsely and strongly gneissose body. In the field it appears to be very uniform and homogeneous over its whole outcrop. However, on the north coast it can be seen to contain several garnet-hornblende schist bodies up to 7 m across (e.g. [HP 523 048]). The contacts, except in the Brim Ness region, are sharply defined and are particularly well shown on the coast in the Breakon area [HP 523 048], along the east side of Ness of Houlland [HP 525 056] and elsewhere. The contact in the Brim Ness region [HP 517 053] is not so clearly defined because the recrystallised semigneissic psammite country rock is very similar in appearance to the Breakon Gneiss.

The Breakon Gneiss either tails off into a vertical sheet to the north or branches into such a sheet, because its eastern boundary, with a parallel schistosity, can be followed round the coast in the Wick of Breakon forming a sharp fold closure (limbs at 90°) at [HP 525 052]. The gneiss is seen again to the north-east where it forms a vertical sheet up to 100 m wide along the east coast of Ness of Houlland.

The gneissic fabric is strongly developed in most places and varies from S-tectonite to L-tectonite ((Plate 4)d), but in no simply interpretable pattern. In many places the gneiss contains ptygmatically folded, somewhat ghostly (i.e. not sharply defined), pegmatite stringers (e.g. [HP 526 052] and [HP 525 049]).

Thin sections of the Breakon and Graveland gneisses are very similar, the Breakon Gneiss being somewhat richer in microcline and with the microcline generally showing very clear microcline cross-hatched twinning. Garnet and sphene are very rare accessories and in some rocks the biotites contain zircon-cored pleochroic haloes. The texture of these orthogneisses is characteristically different to that of the Yell paragneisses in that they are very even grained (1–1.5 mm grain size), the quartz and both feldspars having, generally, similar complex interlocking boundaries. By contrast, in the paragneisses, plagioclase grains control the grain boundary pattern by forming more-equidimensional, less complexly bounded grains than the quartz and microcline ((S94719); (S91528)). The microcline grains in the orthogneisses are, generally, similar in shape and size to the plagioclase grains, whereas in the paragneisses microcline tends to be interstitial or porphyroblastic. Extreme variation in micro-cline content in the Graveland Gneiss, despite its homogeneous appearance in the field and in handspecimen, where microcline-free and microcline-rich varieties were observed to occur in close association, may be due to development of (metasomatic) interstitial microcline as described above. Some of the microcline shows the characteristic invasive habit.

The two tectonic slices of orthogneiss in or against the Bluemull Sound Fault west of Gossaburgh [HU 517 833] and at Wick of Vatsetter [HU 533 903] (Figure 12) differ from the Breakon and Graveland Gneisses chiefly in that they have been subjected to late cataclasis due to movement on the Bluemull Sound Fault. In the field, shearing and lack of good exposures obscures their nature to some extent. In thin section, if cataclastic and diapthoretic effects are allowed for, they closely resemble the Graveland and Breakon Gneisses ((S90939); (S90950)).

Granite orthogneisses

These are homogeneous microcline-rich gneisses very similar in field appearance to the granodiorite gneisses ((Plate 4)c). The granite-orthogneisses were distinguished from the granodiorite-orthogneisses during field mapping on the basis of differences in textural appearance too subtle to describe. The occurrences of graniteorthogneiss are mostly inland, and poorly exposed so that the boundaries are only inferred to be sharp from the occasional close proximity to metasedimentary rocks and paragneisses of very different aspect. The three best occurrences are at Basta Voe and Gutcher ( [HU 525 941], [HU 518 958], [HU 535 998]). Apart from the greater, more uniformly distributed, and essential microcline content they differ in thin section from the granodiorite-gneisses in that most microcline grains exhibit very good microcline twinning ((S91487); (S94710)).

Geochemistry

Analyses of granodiorite-orthogneiss and the average of two analyses of granite-orthogneiss are presented in (Table 13a) and (Table 13b). The analyses are plotted on several compositional diagrams in (Figure 19), and the modes of the analysed rocks are shown plotted on the Streckeisen triangle in (Figure 19)c. According to the latter the granodiorite-orthogneiss varies in composition from tonalite to monzogranite and the granite-orthogneiss from monzogranite to syenogranite. The two types differ chiefly in content of K₂O, Rb and Sr. According to the criteria of Chappel and White (1974) and others both have most of the characteristics of S-type granitoids particularly with respect to their muscovite, garnet and apatite contents. They have normative quartz contents of about 34. Most plot in or near the S-type field on the Na₂O-K₂O plot ((Figure 19)b) although two plot in the I-type field, and molal A/CNK ranges from 1.2 to 1.7. Too much reliance should not be placed on statistics involving K in view of the possible presence of metasomatic interstitial micro-cline described above. On the Pearce et al. (1984) Rb v. Y+ Nb diagram both orthogneisses plot in the VAG field, close to the WPG boundary ((Figure 19)a).

Chapter 6 Gneisses and gneissification

In the past, geologists travelling across Yell towards the spectacular lithological variety of Unst and Fetlar, have dismissed the island as a mass of gneiss covered by peat. In fact, only about a third of the island is composed of gneiss which can he divided into a number of types based on origin, and each of those types display a number of varieties.

Gneisses forming the Lewisian inliers have already been described in Chapter 3. Hornblendic gneisses somewhat similar to the Lewisian inliers occur as relic masses in the Hascosay Slide and are described in Chapter 8. Anatectic quartzofeldspathic gneisses, similar to those forming the east coast of Lunna Ness in the Central Shetland Sheet to the south, occur in some skerries in the extreme south-eastern corner of the Yell Sheet. The Valayre Gneiss is a narrow zone of microcline megacrysts in a matrix similar in type to adjacent rocks, but which shows evidence of being a metamylonite. It forms the boundary between the Yell Sound Division and the Boundary Zone to the east.

Orthogneisses derived from granitoid bodies intruded into the Yell Sound Division before or early in its metamorphism have already been described in Chapter 5. The remaining gneisses, making up the hulk of the gneisses of Yell, are paragneisses of various types.

Nomenclature

In quantifying the types and varieties of gneiss a difficulty is posed by the lack of a universally accepted definition of the term gneiss. In this work the term 'gneiss' will be used but not defined; the rocks labelled gneiss in this Memoir were so named in the course of field surveying, without any consideration of definitions, in order to distinguish and/or relate various bodies of rock. Rocks were named gneiss in which feldspar plays a dominant role in field appearance and when the rock has a tectonite fabric visible in the field.

Where, as commonly happens, gneisses pass gradationally into metasedimentary rocks with no gneissic characters, difficulty arises in drawing a boundary between gneiss and metasediment. In many places intermediate members of the series, partially formed gneisses, containing the features of both the metasedimentary rocks and of the gneisses, occur in the absence of any 'gneiss' or 'metasedimentary rock'. In this work, for descriptive and survey purposes such rocks have been called semigneisses, and they and the gneisses are both referred to as paragneisses. The lack of a name for such intermediate rocks has, in the past, led to insufficient attention being paid to them, even though they contain more evidence for the origin of the gneiss than the gneiss itself.

Yet another difficulty arises from the names currently available for different types of paragneiss. Some of the gneisses and semigneisses in Yell have the appearance of migmatites according to the definition of such rocks as macroscopically mixed, inhomogeneous or composite rocks (Mehnert, 1968; Brown, 1973; Ashworth, 1985). However, they occur intimately mixed with, and obviously formed at the same time and by the same process as non-composite, relatively homogeneous, gneisses and semigneisses which, by definition, are not migmatites. These homogeneous, nonmigmatitic rocks resemble diatexites (as defined by Mehnert (1968), Brown (1973) and Ashworth (1985)), but are not the result of melting or metamorphic differentiation. Therefore, such terms as migmatite, metatexite, diatexite and other related terms will not be used for the Yell paragneisses, even though gneisses which match pictures and descriptions of such rocks occur in Yell. The terms are no longer sufficiently nongenetic and their use would inevitably lead to misconceptions. In the absence of a nonspecific term covering all the processes by which rocks are turned into gneisses, whether migmatites or not, the word gneissification is used in this Memoir.

Before describing the gneisses of Yell the occurrence of microcline arid its distribution and origin has to be considered.

Rocks containing interstitial microcline

Throughout Yell, some of the metasedimentary rocks (except possibly quartzites which were rarely sectioned), the semigneisses and gneisses derived from these rocks, together with several samples of quartzofeldspathic orthogneisses in the Lewisian inliers, contain potash feldspar with a characteristic interstitial habit. Such rocks, even where half replaced by potash feldspar, are little different in appearance to adjacent rocks containing no potash feldspar and so cannot usually be recognised in the field and could not be mapped. In a number of places quartzofeldspathic psammites were seen to have a very slightly pinkish tinge instead of the normal grey or grey-green colour. Some such rocks were found to contain interstitial microcline. However, feldspar which appeared pink in the field, occurring along joints or in some pegmatites, was often found, on thin sectioning, to be stained plagioclase. On the other hand, pink micro-clines in pegmatite were often observed to have been bleached white by weathering, especially where they had been in contact with peat. Feldspar colour is not a good guide to feldspar composition in Yell.

Potash feldspar is distributed in an apparently random manner both geographically and lithologically (Figure 13), as far as can be determined from the irregular nature of inland exposure. The areas of rocks containing potash feldspar appear to vary in size from small patches, perhaps no more than a few metres across, to patchily developed regions a kilometre or more across.

Potash feldspar was found in 237 out of about 1000 thin sections of rocks which were originally sedimentary. In most the potash feldspar varies in volume between 10 and 40 per cent and most commonly forms between 20 and 30 per cent of rock volume. Only a very few of those containing less than 10 per cent contain more than a trace amount and only slightly more contain more than 40 per cent. The average modal amount of potash feldspar found in the rocks containing potash feldspar was about 20 per cent in each of the three groups: metasedimentary rocks, semigneisses, and gneisses.

The potash feldspar was studied by the method of Becke Line staining, as this method is easier to use and was found to give better and more reliable results than chemical staining. However, as with chemical methods, it is less reliable (more difficult to apply) in mica-rich rocks, which may explain why most interstitial microcline was found in mica-poor rocks.

The potash grains have a characteristic appearance. Their boundaries are highly complex, and may be described as interstitial and invasive, because they are formed of convex-inwards arcs joined by sharp and narrow protuberances into and between the adjacent grains. In rocks containing 10 per cent or less by volume the grains are all tiny (0.1 mm or less maximum dimension) shreds, many in grain boundaries or at triple junctions and some within other grains ((S91014); (S91521)). With increasing volumes of potash feldspar the grains become larger and form aggregates but still show only minor microcline twinning ((S94677); [HU591 981] until the potash feldspar makes up more than 20 per cent of the rock when the twinning becomes common ((S90900): (S92019)). The close association of twinned and untwinned feldspars in varying proportions suggests that both are microcline.

Microcline-twinned microporphyroblasts several milli-metres across form in the rocks richest in microcline ((S94803); (S91780)), but preserve their characteristic invasive outlines against the adjacent quartz and plagioclase grains. There is a tendency in the microcline-rich rocks for the grain size of the microcline to approximate to that of the other minerals in the rock. In gneisses and semigneisses containing coarse quartz-feldspar segregadons within a finer-grained rock the microcline (S91897) is concentrated in the segregations as large grains with the usual complex and invasive profiles. In these segregations, and in general where microcline occurs in contact with plagioclase, its invasive contact with the plagioclase gives it the appearance of replacing the plagioclase, but myrmekite is very rare. Furthermore, plots of plagioclase modal volume against microcline modal volume show no correlation, not even the negative correlation which might he expected from the closed nature of the modal data.

In the Brim Ness area [HP 515 054] interstitial microcline development has lead to the formation of rounded aggregates of microcline in the form of pseudomegacrysts reminiscent of the megacrysts in the Valayre Gneiss.

It should be noted that interstitial microcline was the only type of microcline found in metasedimentary rocks and semigneisses, but that noninterstitial microcline occurs as a major and essential constituent of several types of gneiss described below.

Geochemistry

Thirty of the metasedimentary rocks, semigneisses and gneisses analysed for this Memoir contain interstitial microcline. Plots of the oxide and element values against modal volume of microcline show a significant positive correlation for K₂O, Rh, and Ph while TiO₂, FeO, Zr, Sr and Sc show significant negative correlations ((Table 14), based on (Table 15a) and (Table 15b). The correlation coefficients are not very high indicating that besides the variation due to the presence of the microcline there is substantial variation in the amounts of these components in the rocks and also in the data due to a lack of perfect match between the rock sample analysed and that used for point counting. A comparison of the 30 analysed rocks containing microcline with 65 analysed rocks of similar origin but containing no microcline showed significant increase of K₂O, Rb, and Ba and decrease in Tia₎, FeO, Sr and Sc in the microcline-containing rocks (Table 15a),(Table 15b). Other large differences between means are associated with large differences in variance so that no valid comparison can be made. It is clear from (Figure 14)a that interstitial microcline has grown in rocks of very restricted range of A1₂O₃, indeed only in granofels of psammitic type. The average volume of mica in the microcline-hearing rocks is about 20 per cent, while that in the microcline-free rocks is 30 per cent.

Interpretation

Patches of interstitial microcline occur throughout the Yell Sound Division and the Boundary Zone in Yell. They also occur to the south, in the Central Shetland Sheet, throughout the Yell Sound Division in the gneisses and metasedimentary rocks of the Whiteness and Scatsta Divisions (Flinn, 1954, 1967). The heaviest impregnation of interstitial microcline so far seen in the Yell Sound Division occurs at Calna Taing [HU 368 727], in the Central Shetland Sheet, transforming a psammite band into a rock noticeable even in the field as unusually feldspar rich.

The patchy distribution of the interstitial microcline, apparently unrelated to other differences in the protolith or to any structures or intrusions, is interpreted as the result of metasomatism of the rocks at a late stage in their development. It appears to have taken place as a result of the permeation of the area by potassium-rich fluids flowing through irregularly distributed channels, of a variety of diameters, but, judging by the proximity of rocks free of microcline, relatively sharply bounded. The relatively low volume of mica in rocks containing interstitial microcline may be due to the breakdown of mica to form the microcline.

Gneisses of the Neapaback Skerries

The gneisses of the Neapaback Skerries are the northward continuation of the band of nebulitic gneisses forming the east side of Lunna Ness, immediately east of the Valayre Gneiss in the Central Shetland Sheet, and last exposed in Lunna Holm [HU 527 748] 4 km to the south. Such rocks are not exposed north of the Neapaback Skerries. They belong to one of the tectonic units forming the Boundary Zone.

They are formed of coarse (2 mm grain size) granular gneisses with centimetre-sized micaceous 'fish' grading into granitoid gneisses with streaky to laminated appearance due to concentration of mica and garnet in thin layers. In many places inicroporphyroblasts of plagioclase are common ((S90745); (S90746)). The gneisses are rich in garnetiferous red-brown-hornblende schist bands (S94800), some boudinaged, and some as 'fish', all disposed in flowing folds with closures varying up to a metre or so across and of no regular orientation ((Plate 4)e). The gneisses are cut by many late pegmatite and aplite veins which, like the red-brown hornblende schists, are described elsewhere.

In thin section the gneisses are seen to be composed of quartz and plagioclase, together with garnet, biotitc, and muscovite, and to have a grain size of one to two millimetres. The plagioclase tends to occur as compact rounded equidimensional grains about 2 mm in diameter, with interstitial irregular-shaped grains of the other minerals. Plagioclase grains dominate the texture. An analysis of a sample of the gneiss (S90745) containing about 40 per cent plagioclase and 30 per cent biotite, is presented in (Table 13a).

Paragneisses – microcline megacryst augen gneisses

The eastern boundary of the Yell Sound Division, that is the eastern limit of a dominantly psammitic succession with included garnet-studded hornblende schists and Lewisian in-Hers, is marked by a continuous zone, called the Valayre Gneiss, in which the rocks contain large single-crystal micro-cline augen ((Plate 5)a, b and c). Other less-extensive bands of similar augen gneiss, generally with smaller augen, occur in the Yell Sound Division west of the Valayre Gneiss and also to the east, in the Boundary Zone (Figure 12).

The Valayre Gneiss

The Valayre Gneiss marks a lithological break in the East Mainland Succession, because the rocks of the Boundary Zone to the east are significantly different from those of the Yell Sound Division to the west. East of the Boundary Zone lies a Dalradian succession separated from the Boundary Zone in the area of the Central Shetland Sheet, by another augen-gneiss, of different aspect (the Skella Dale Gneiss). It is likely the Valayre Gneiss marks not only a lithological break but also a tectonic break. The lenticular boundaries of some of the units forming the Yell Sound Division and the Boundary Zone, the interruption of the Valayre Gneiss by faulting, poor exposure, as well as the presence of the Graven Complex makes this difficult to prove.

The Valayre Gneiss can be followed from the southern edge of the Central Shetland Sheet close to the Walls Boundary Fault at [HU 345 525], to the Nesting Fault near the northern edge of that sheet at [HU450 752]. After a 16 km dextral displacement on the Nesting Fault it can be traced along Lunna Ness into the Burravoe area of south Yell [HU 521 791] and along the eastern side of Yell to the north coast at [HP 535 053], a total distance of about 77 km ((Figure 2) and (Figure 12)). It is exposed intermittently along this strike length and varies from a few metres to several hundred metres in thickness. In the Yell Sheet the maximum width exposed is less than 100 m. It seems likely that the band is continuous, apart from interruptions by late faulting.

In the Central Shetland Sheet, the Valayre Gneiss is formed by a single band, except on Lunna Ness, where to the north of West Voe of Lunna [HU480 690] there are several parallel bands. In south-east Yell, however, the Valayre Gneiss splits into several closely spaced bands and parallel hands of similar rock occur in a number of places.

On the west tip of Heoga Ness [HU 521 791] the Valayre Gneiss is represented by half a dozen closely spaced bands of microcline megacrysts, each band a metre or two thick. Up to five bands of megacrysts occur along the north shore of Burra Voe, though some are little more than a single string of megacrysts. Between the exposures on the shore at Burravoe [HU 524 794] and that at Gossaburgh [HU 534 835] the Valayre Gneiss is exposed in several places, occurring apparently as several bands. From Gossaburgh to Hascosay it is seen as a single band in a number of substantial cliff exposures, although faulting interrupts the continuity. From Gutcher [HU 551 982] to the north coast of Yell [HP 535 053] the Gneiss forms a single continuous band exposed in a series of inland and coastal exposures.

The Valayre Gneiss is not a normal lithological unit. It is a hand of megacryst augen in a matrix composed, from place to place, of a variety of different lithologies. These lithologies, apart from the presence of the megacrysts and, in places, of an enhanced schistosity, are those of the local adjacent rocks. Thus in the Central Shetland Sheet, at the southern exposure of the Valayre Gneiss, it has a schist matrix [HU366 545], [HU367 620], while in Lunna Ness the matrix resembles the adjacent anatectic gneiss [HUHU 500 700]. In Yell, in the Burravoe area, the matrix is the local rock, a microporphyroblastic psammite ((Plate 5)c). Farther north, at Aywick and Vatsetter the matrix is psammitic granofels ((Plate 5)b, and at [HU 540 865] and [HU 544 892]), while at the Birrier [HU 546 885] and on Hascosay [HU 546 910] the matrix is recrystallised, partially gneissose quartzofeldspathic granofels. In northeast Yell the most common matrix is granofels containing prominent fat quartzofeldspathic lenses about the same size as the potash megacrysts.

Between Aywick and the north coast of Yell, lenses and bands of garnet-free hornblende schist occur within the Valayre Gneiss. The largest masses occur at Aywick [HU 544 872]; farther north they vary up to 3 m in thickness.

The characteristic feature of the Valayre Gneiss, which makes it immediately recognisable in the field, is the presence of smoothly bounded rectangular megacrysts of microcline varying up to 5 cm in greatest dimension and varying from equidimensional to proportions of less than two to one. Where the gneiss is lineated, the augen are most clearly displayed in rock faces normal to the lineation. In faces parallel to the lineation they tend to have tails composed of matrix-sized grains giving them a fat-lenticular shape which, together with the schistosity, gives the rock a mylonitic-look in hand specimen. No two exposures show the same proportions of larger and smaller megacrysts or density of their packing. The separation of the megacrysts varies from tens of centimetres to several centimetres. Where the boundary of the Gneiss is exposed it is defined only by the sudden cessation of the megacrysts, and it only occasionally coincides with a change in matrix lithology.

The best exposure of the rock is on Hascosay [HU 546 910] where, on an optimally weathered surface normal to the lineation, characteristicly smoothly rounded subrectangular single crystals of slightly vitreous pink microcline, several centimetres long, with 1 to 2 mm-thick white rims, occur among many less-regularly bounded microcline megacrysts. Also present are rounded, white, centimetre-sized augen of plagioclase. All these occur in a matrix of coarsely granular, recrystallised psammite ((Plate 5)b). Due to the orientation of the main exposure surface normal to the lineation the rock looks very like an undeformed porphyritic granite. On surfaces parallel to the lineation all the augen form fat lenses and the matrix has a platy, fine lamination which encloses the augen in perfect eye-like structures. A very different development of the Gneiss is well exposed at the tip of Heoga Ness ([HU 521 792]; (Plate 5)c) where similar microcline augen occur, but in a fine-grained psammite, together with 2 to 3 mm-diameter plagioclase porphyroblasts characteristic of the area (microporphyroblast semigneiss – see below).

Other microcline-megacryst augen gneisses

The other bands of megacryst augen gneiss occurring east and west of the Valayre Gneiss and striking parallel to it are distinguished from the latter by containing smaller augen (in most cases), and by being much less continuous. Characteristic of them are augen about 1 cm across but bands containing augen 2 cm and even 3 cm in diameter do occur, for example at the Horse of Burravoe and at Point of Whitehill [HU 533 813] and [HU 536 823]. The range of densities of packing of the megacrysts, and the range of lithologies forming the matrices are the same as those of the Valayre Gneiss. The bands east of the Valayre Gneiss in the south-east of Yell have a more strongly mylonitic appearance in the field than the Valayre Gneiss itself. This is best seen in the bands west of the Ness of Quheyin [HU 537 853] and at Horse of Burravoe [HU 533 813], and most especially east of Burravoe [HU 528 801]. The bands east of the Valayre Gneiss occur close to metadolerite balls at Point of Whitehill [HU 536 823], east-north-east of Aywick [HU 546 873] and at Head of Gutcher [HP 551 000], Those in the neighbourhood of Ay-wick are interleaved with the metadolerites, but no crosscutting relationships being exhibited.

Petrography

Apart from augen size there is little difference in thin section between the two types of augen gneiss. There is no sign of cataclasis in the rocks, and the fine mylonitelike laminations apparent in some hand specimens and exposures are no longer visible under the microscope due to recrystallisation and new mineral growth (S91186), but the micas show a very strong parallel preferred orientation in sections parallel to the lineation. In addition to the megacrysts, microcline also occurs as smaller porphyroblasts ranging down to several millimetres across. Some of the smaller augen are composed of two or more crystals, or even aggregates of microcline. Single- and multiple-grain plagioclase augen up to several millimetres in diameter occur and fat lenticular streaks of quartz, plagioclase and microcline of the same size as the smaller augen are present. The proportions of the different types of augen differ from locality to locality.

The matrix has a grain size generally between 0.2 mm and 0.3 mm, which is often less than that of the adjacent granofels and gneisses. It varies from being nearly absent ((S90963); (S91427)) to being the major constituent. Microcline occurs interspersed with the quartz, plagioclase and micas forming the matrix of the gneiss bands. Microcline occurs also in quartz-microcline and quartz-plagioclase-microcline aggregates forming lenticular augen with grain sizes of half a millimetre or more. Most prominently, it occurs as the single, or multigrain rounded megacrysts varying up to several centimetres in diameter. The latter are often surrounded by envelopes about one grain thick of plagioclase, myrmekitic-plagioclase and microcline ((Plate 5)a and b, (S90802); (S91801)). Such rims never accompany the plagioclase megacrysts which are not then myrmekitic. In some rocks the megacrysts are turned into lenticular augen by the presence of coarse triangular-shaped aggregates of microcline forming the corners of the 'eyes' (S91801). The microcline megacrysts often contain inclusions of biotite that are the same colour as those in the matrix, but larger in size and with a different orientation ((S92058); (S91855)).

Augen of all types, whether single- or multicrystalline, rounded or lenticular, are enclosed within the schistosity as defined by the mica flakes. The mica flakes are deflected where they contact the augen, are everywhere tangential to the edges of the augen and wrap them very closely ((S90794); (S91196)). This relationship differs from the garnet-schistosity relations described above, where the schistosity often encloses not only the garnets but also lenticular areas of the matrix within which the garnet grew prior to the development of the schistosity, or where the garnets cut the schistosity.

Interpretation

The Valayre Gneiss coincides with a major, but not immediately obvious, lithological and tectonic break. In some places in the field it has a mylonitic appearance which in thin section appears to have been destroyed by a later recrystallisation. The augen have the appearance of porphyroblasts, not porphyroclasts, which grew in a movement horizon operative during the regional metamorphism of the area, when the Moine and Dalradian blocks were brought together to form the East Mainland Succession. There is no sign of similar augen-growth in the other shear and slide zones in Shetland and no independent evidence of displacement due to shearing associated with the other augen-gneiss zones.

Megacryst augen in augen gneisses are often interpreted as relic phenocrysts (Vernon 1990). In its more granitoid-like variants the Valayre Gneiss is, indeed, very similar in exposures to the edge facies of the Skaw Granite of Unst, a forcibly intruded and penetratively deformed porphyritic granite of postorogenic type. However, it is very different in thin section, especially in that the megacryst augen of the Gneiss lack the cataclastic features shown by the porphyritic microclines of the granite. Identical cataclastic textural features to these have been described in detail from elsewhere in the world by Vernon (1990). However, Vernon's criteria for distinguishing porphyritic from porphyroblastic megacrysts are not operative for the Valayre Gneiss which was deformed and metamorphosed both during and after the growth of megacrysts, in contrast to the megacryst porphyroblasts of his account. The Valayre Gneiss is similar in appearance, and in its occurrence as an extremely attenuated and extended band, to the Tannas Gneiss of Norway (personal communication, D G Gee). The Tan-lids Gneiss is considered to be a deformed granite sheared out beneath a nappe (Roshoff, 1978). However, the Valayre Gneiss resembles a deformed granite only locally where the matrix containing the megacrysts happens to be of granitoid appearance; it marks a tectonic junction, but it is a band of megacrysts in a variety of matrixes, it is not a band of uniform gneiss.

Paragneiss – microporphyroblast gneiss

In Unst, Read (1934) distinguished a particular lithology in the Valla Field Block, which he named 'blebby gneisses'. They are granular rocks with a grain size varying up to a millimetre, rich in rounded plagioclase grains varying from 1 mm to 4 mm in diameter, the 'blebs'. They resemble the plagioclase blastite or pearl gneiss of Mehnert, (1968, fig. 22). Similar rocks occur among the Unst and Fetlar metasedimentary rocks and gneisses to the east of Yell, on Linga, and on Hascosay immediately east of the Hascosay Slide. They occur in the Boundary Zone, to the west of the Hascosay Slide, in the Central Shetland Sheet in Lunna Ness, for example at [HU485 688], and in the Yell Sheet in south-east Yell and west Hascosay (Figure 15).

In Yell, they are best developed between Horse of Burravoe [HU 530 810] and Ness of Quheyin [HU 540 850] where they form bodies of rock closely packed with microporphyroblasts and showing signs of mobilisation ((Plate 5)d). The gneiss bodies are delimited by a rapid decrease in the density of packing of the porphyroblasts which, in places, become absent within a few metres. However, elsewhere the porphyroblasts are reduced in numbers to the point where the rocks are best called semigneisses, being formed of metasedimentary rocks containing widely spaced, but otherwise similar, rnicroporphyroblasts. In semigneiss areas, the microporphyroblasts vary in density from almost absent to almost as closely packed as in the gneisses, i.e. from separated by centimetres to separated by millimetres. The density of packing varies over the area in a very patchy manner. A complete gradation exists between microporphyroblast gneisses, microporphyroblast semigneisses and the microporphyroblast-free psammitic granofels within which they occur. The porphyroblasts tend to occur in uniform, homogeneous biotitic psammitic granofels but they also occur in associated laminated and banded granofels. In neither do they show in their distribution any tendency to a layered or banded distribution.

In Ness of Quheyin the microporphyroblast semigneisses are seen to pass rapidly east, with increasing density of porphyroblasts, via a zone 10 to 20 m wide, of schistose S-tectonite microporphyroblast gneiss into an area of L-tectonite microporphyroblast gneiss. Suitably weathered joint faces normal to the lineation show a nebulitic rock with signs of mobilisation and flow revealed by very faintly but sharply defined disoriented and contorted fragments of gneiss which are slightly less microporphyroblastic than the gneiss matrix enclosing them. These fragments are all small, being no more than several tens of centimetres in maximum dimension. The fragments are parallel to and even folded on axes parallel to the lineation in the gneiss, so that on rock faces normal to the lineation the L-tectonite gneiss looks like an undeformed nebulite. On faces parallel to the lineation the nebulitic appearance is difficult to detect. The L-tectonite core of the gneiss body is divided into pods, ten metres or so across, by schistose bands less than a metre thick with the form of ductile shears with the schistosity curving outwards on either side along the shear band.

A somewhat similar occurrence of microporphyroblast gneiss is exposed on Green Holm [HU 516 787] where pods of steeply plunging L-tectonite microporphyroblast gneiss several tens of metres across are enclosed in S-tectonite microporphyroblast gneiss passing out into microporphyroblast semigneiss. The gneiss body south of Gossaburgh [HU 530 820] is well exposed in places, but its relation to surrounding semigneisses cannot he viewed as well as in the other two occurrences.

Apart from the zone of microporphyroblast gneiss and semigneiss between Green Holm and Ness of Quheyin, microporphyroblasts occur sparsely scattered in granofels farther north in the Boundary Zone, generally in unbanded and unlaminated biotitic psammites. They occur in places immediately west of the Hascosay Slide in Hascosay and farther north in Bay of Brough [HP 537 047], but nowhere are these microporphyroblasts so closely spaced as to give the rock a gneissic look.

Thin sections of the microporphyroblast gneiss and semigneiss present the appearance of normal garnet biotite psammites of granofels-type with scattered, very rounded, porphyroblasts of plagioclase. These grains are about lmm or more in diameter, and are twinned single crystals. Microprobe analysis shows them to be oligoclase, the same as the matrix plagioclase. The garnets in the rocks are usually the same shape and size. Biotite makes up 15 to 20 per cent of the rocks by volume, quartz 35 to 50 per cent, plagioclase 25 to 35 per cent and muscovite less than 10 per cent. These minerals form a matrix with a grain size between 0.2 and 0.3 mm. The rocks are schistose and the schistosity appears to be less deflected by the plagioclase porphyroblasts than by the garnets ((S90966); (S90864)).

A perfect gradation exists between granular psammites and similar rocks with scattered plagioclase microporphyroblasts (semigneisses) and rocks so packed with microporphyroblasts that they have the field appearance of gneisses. Thin sections show the matrix to remain unchanged mineralogically and in grain size from the psammites through to the gneisses (S91058). Only the micro-porphyroblasts change; in the semigneisses they are spaced at centimetre distances while in the gneisses they are spaced at millimetre distances. However, in the L-tectonite cores of the gneiss areas the microporphyroblasts vary up to 2 mm in diameter and occasionally 3 mm. Some of the rocks containing the closest packed microporphyroblasts show signs of recrystallisation of the matrix, with the quartz grains doubling in size while retaining their angular shapes (891125). Aggregates of such quartz grains occur and, less commonly, augen-like aggregates of plagioclase grains.

Some microporphyroblast rocks contain microcline, but this occurs only in or close to an area between Horse of Burravoe [HU 530 810] and White Hamar [HU 530 8031], and only in semigneisses. Because of this restricted and patchy occurrence, and because microcline was not found in the gneisses, its presence is considered to be the result of an unrelated process, probably the same metasomatism that gave rise to the interstitial microcline occurrences. The microcline occurs in the matrix as grains of the same size and shape as the quartz and plagioclase of the matrix, and rarely, as microprophroblasts of the same shape and size as the plagioclase microporphyroblasts. This may indicate that the microcline was added to the rock prior to the process causing the development of the plagioclase microporphyroblasts.

It should be noted that, particularly in the case of the semigneisses, the microporphyroblasts look very like relic clastic sedimentary grains. Thin section examination revealed no features which would distinguish the two. They have been interpreted as porphyroblasts for a number of reasons. They are invariably formed of feldspar, on exposed surfaces they never show any other than a completely random distribution pattern and in the rocks in which they are most closely packed the matrix shows signs of recrystallisation and the rock as a whole of becoming mobilised (nebulitic).

Geochemistry

Analyses of three of these rocks are presented in (Table 16). (S90835) was chosen by field and thin section examination as a typical microcline-free semigneiss, while (S90732) and (S91013) (average of the two analyses reported) arc typical microcline-free gneisses and (S90902) is typical microcline-hearing semigneiss. It is apparent that most of the differences between the three are carried by the microcline-bearing semigneiss. Its relatively high contents of potassium, rubidium and aluminium can be attributed to the presence of microcline, although aluminium is not positively correlated with microcline content in the other rocks containing interstitial microcline. The lack of a significant difference in composition between the microporphyroblast semigneisses and gneisses is supported by a study of the modes of 20 of these rocks (microcline-free) chosen as representative in both thin section and field appearance and occurrence. In (Figure 14)b microporphyroblast volume, providing a measure of the degree of gneissification, is seen to be independent of total-plagioclase volume and to remain within the range of feldspar volume observed in the metasedimentary rocks. There is no evidence that any increase in plagioclase content has taken place as a result of the gneissification.

Interpretation

Since increasing microporphyroblast content is not associated with increasing plagioclase content, and the total-plagioclase content of the rocks appears to have remained constant during the growth of the porphyroblasts, and since there is no major difference in composition between the ungneissified rocks and the gneissified rocks the process of gneissification is essentially isochemical. Any addition or subtraction of material which may have taken place during the gneissification is very minor and is incidental to that process. Furthermore, the process by which the L-tectonite, mobilised and nebuliticlooking microporphyroblast gneisses were formed cannot have involved melting, the usually cited cause of such rocks, because the matrix separating the porphyroblasts, the part of the rock which would have to have melted, is still largely unaltered fine-grained metasedimentary granofels, only showing signs of the onset of recrystallisation in the most extreme gneisses. The gneissification cannot have been caused by locally raised temperature, because there is no mineralogical evidence that the microporphyroblast gneisses were ever at a higher temperature than the adjacent ungneissified rocks and because there is no evidence for a local source of heat, i.e. no closely associated igneous intrusions. Furthermore, the gneissi fled areas are too small and scattered to be attributed to geothermal heating or isothermal uplift, the usual explanations for anatectic gneissification. The gneissification process which operated on these rocks caused the porphyroblastic growth of plagioclase, and in areas where it acted most strongly it weakened the grain boundaries of the minerals so that the rock could deform, quite freely, by grain-boundary sliding and caused sonie recrystallisation of the matrix. It is, therefore, suggested that these effects were the result of the local permeation of the area by a fluid at a temperature little different to that of the rocks at the time, and that this allowed the plagioclase in the rock to aggregate as single crystal microporphyroblasts, but had a minimal and incidental effect on the composition of the rocks, though in areas of maximum flow it loosened the grain boundaries so that the rocks deformed more easily than the surrounding rocks.

Other paragneisses

All the types of gneiss mentioned above together occupy a considerably smaller area of Yell than the paragneisses now to be considered, which, for convenience, will be called 'the paragneisses'. They occur throughout the island, in smaller and larger patches and bands, within which the gneisses show a continuous gradation via semigneiss to ungneissified metasedimentary rocks, both granofels and schists, and ranging in composition from psammite to pelite. The paragneisses comprise a whole range of types of gneiss and semigneiss which occur intimately associated with one another in the areas indicated in (Figure 15). The different areas of gneiss and semigneiss vary more in size and in proportion of gneiss to semi-gneiss than they vary in the proportions of the different gneiss-types present. It is significant that these areas or patches of gneiss and semigneiss are unrelated in their distribution to the metasedimentary lithology of the area, except that they do not involve the quartzite bands. Except for the quartzites, the susceptibility of the rocks to gneissification does not depend on original composition, but the type of semigneiss (and to a lesser extent, of gneiss) produced does depend on the nature of the original metasedimentary rock. This is revealed by the way the metasedimentary rocks which surround the gneissified patches, and occur as relics within them, are no different to the metasedimentary rocks which have been partially gneissified to form semigneisses.

The semigneisses vary from granular, almost granitoid and almost homogeneous, rocks to variously quartzofeldspathic-handed, streaked and augened rocks, while the gneisses present a more homogeneous appearance, tending towards a finer to coarser granitoid appearance, depending on whether the protolith was granofels or schist. When Yell was mapped in the 1930s these gneisses and semigneisses were grouped together as gneisses and, according to whether they were streaked and/or banded or not they were called respectively injection-gneisses or permeation-gneisses (Allen, 1932). This followed the work of Read (1931) who described very similar paragneisses in Central Sutherland. He named homogeneous granitoid varieties 'permeation-gneisses' and the quartzofeldspathic banded, streaked and augened rocks 'injection-gneisses'. His illustrations of injection-gneisses are indistinguishable from the paragneisses of Yell (omitting the microporphyroblast gneisses).

Since Allen's (1932) survey of Yell, Read's 'injectiongneisses' have generally come to be called 'migmatites' and attributed to a variety of origins or no particular origin. On the other hand, Read's 'permeation-gneisses' have been lost sight of (homogeneous migmatite being a contradiction in terms), and relatively homogeneous paragneisses figure in works on migmatites only as diatexites resulting from complete melting (Mehnert, 1968; Brown, 1974; Ashworth, 1985).

In Yell the permeation-gneisses (non-migmatites) are intimately associated with the injection-gneisses (migmatites), are obviously the result of the same process of gneissification, grade continuously into them and are at least as important volumetrically. Read's terms permeation gneiss and injection gneiss are genetic and no longer used. Currently, the term 'migmatite' tends to he used as a synomym fbr paragneiss, and is even applied to whole gneiss belts containing homogeneous gneisses as well as migmatites, and so will be avoided here. Instead, to emphasise the close association of the two different types of gneiss they will be referred to by the nongenetic terms homogeneous-gneiss or semigneiss and leucosomegneiss or semigneiss.

In Yell the leucosome-semigneisses grade continuously in appearance from rocks in which the coarse quartzo-feldspathic component (leucosome) forms parallel-sided bands (lit-par-lit injection-gneisses of 1932) via rocks with lenticular streaks, to rocks in which it forms fat eye-like aggregates. Rocks with these patterns have been described as stromatic and opthalmitic respectively (Mehnert, 1968), but these names are of little use in Shetland where the fabric of the rocks varies from S-tectonite to L-tectonite. L-tectonites show different patterns on faces normal to the lineation to those on faces parallel to the lineation (lenticular streaks or bands parallel to the lineation and augen normal to it), and all gradations between the two on faces inclined to the lineation. This effect can be seen in the Ltectonite orthogneiss shown in (Plate 4)d. Similar relationships between very different patterns in the same rock are presented by L–S tectonites and S-tcctonites. In Yell the leucosomes all conform in attitude and symmetry to the tectonite fabric and the relic sedimentary layering in the semigneisses. The more complex patterns figured by Mehnert (1968) are not apparent.

Semigneisses, in field appearance, show a complete gradation between leucosome-semigneiss and homogeneous-semigneiss. The gradation involves both a decrease in numbers and prominence of leucosomes and an increasing granular texture for the matrix of the leucosomes ((Plate 6) and (Plate 7)).

Particular types of semigneiss show a more obvious gradation, both in appearance and in field occurrence, between particular types of metasedimentary rock on the one hand and gneisses on the other. The homogeneous semigneisses (homogeneous granular rocks) grade, on the one hand, into granofels with poorly developed lamination and banding but with a bedded look, and on the other into granitoid homogeneous gneisses. The gneissification of these rocks, thus, involves an homogenisation leading to a loss of the bedded look of the metasedimentary rocks and of any lamination or banding present.

The gneissification of the leucosome-semigneisses is also a process of homogenisation, but one in which original inhomogeneities in the metasedimentary rocks are at first enhanced and only in the later stages of gneissification are they partially or completely suppressed. The semigneisses with long parallel streaks of leucosome (litpar-lit rocks) grade, on the one hand, into psammites with alternating layers or laminations of more and less micaceous rock (schistose and granoblastic layers), and on the other hand, into somewhat inhomogeneous granitoid gneisses, as the differences between the leucosomes and the intervening rock become less pronounced. The semigneisses in which the leucosomes take the form of lenticular streaks and augen grade, on the one hand, into schists (micaceous psammites and mica schists) by decreasing density and size of streaks and augen and, on the other hand, into coarse micaceous granitoid gneisses as the streaks and augen increase in density of packing and the matrix containing them coarsens.

In thin section the most prominent difference between metasediment and gneissified rock is an increase in grain size from the less than 0.5 mm characteristic of the metasedimentary rocks to 1 mm or more characteristic of the gneissified rocks. Also characteristic of the gneissified rocks is the preferential growth of the plagioclase as grains more equidimensional and regularly bounded than the grains of quartz. In the metasedimentary rocks the grain shapes and sises of quartz and plagioclase are the same.

These gneisses and semigneisses can be seen in most parts of Yell, but the relative proportions of the different types, and especially the proportions of gneiss relative to semigneiss, differ from place to place. Most widely distributed are isolated or very widely spaced quartzofeldspathic streaks and augen in the more micaceous metasedimentary rocks (Plate 6). The presence of such preliminary manifestations of gneissification has not been indicated on the map (Figure 15). Also too widely distributed to be noted on the map are small bands or areas, usually no more than a few centimetres wide, in which a development of millimetre-sised quartzofeldspathic augening is visible in mica-rich bands in banded psammites. The same incipient augening occurs very commonly within metasediment relics in semigneiss areas.

The semigneiss and gneiss areas indicated on (Figure 15) vary in size from the equidimensional patch of gneiss surrounded by semigneiss little more than 100 m across at West Sandwick [HU 440 883] to continuous bands dominated by gneiss many kilometres in length (e.g. that just east of the West Sandwick patch) and to large areas dominated by semigneiss like that on the north coast. Semigneisses are widely distributed along the west coast to the north of West Sandwick and within these semigneiss areas are several well-defined patches of gneiss, especially well developed south of Point of Bugarth [HU 442 925] and at Burgi Geos [HP 477 035]. The most instructive occurrence for study of the results of gneissification is the one at West Sandwick [HU 440 883]. Especially homogeneous developments of gneiss derived from granofels can be seen nearby at West Sandwick [HU 445 890] and in the cliffs south of the Point of Bugarth [HU 442 930]. Inland, to the south of Muckle Vandra Water [HU 492 885] coarsely micaceous gneisses occur, which are highly gneissified versions of quartzofeldspathic-streaked micaceous semigneisses occurring in many areas and especially well developed at Burravoe [HU 514 79] and Brim Ness [HP 517 053].

Microcline occurs in the paragneisses in the form of 'interstitial microcline', as described above. Microcline also occurs as an essential mineral in several bands of paragneiss described below under the heading 'microcline-paragneisses'.

Geochemistry

Samples of metasedimentary rock (26), serniparagneiss (14) and paragneiss (25), all lacking microcline were analysed. The samples were all chosen by field and thin section examination as representative of commonly occurring types of these classes, but not in representative proportions. Since there is no mineralogical or chemical index available as a measure of gneissification, the three classes based on field appearance had to be used. The average analysis for each class is given in (Table 17).

The mineral content and the oxide and trace element contents obtained from each individual analysis were plotted against the appropriate rock class. Spear-man's rank correlation coefficient proved significant at the 95 per cent level in only six of the 32 elements and oxides examined. In all of these the correlation coefficient was only about 0.2, which indicates a very poor correlation. Furthermore in the cases of quartz, SiO₂, MgO, and Co the values observed for semigneisses were higher than those for either the metasedimentary rocks or the gneisses. The correlation for SiO₂ is significant only because of the very restricted variation of SiO₂ content and the difference is very small. The changes of composition compared to the changes in appearance caused by the gneissification are either riot significant or are so small as to indicate that they are a incidental biproduct of the process. It is concluded that the paragneisses of Yell are not the result of a metasomatic process, but of a gneissification process which may have caused some incidental and not very significant changes in composition.

(Figure 14)c–f support this conclusion by showing the rocks plotted according to the three classes, metasedimentary rock, semigneiss and gneiss and according to their mineralogical contents (point-counted modes). It is apparent that there is no systematic difference between the modes of the metasedimentary rocks on the one hand and the gneissified and partially gneissified rocks on the other. The gneissification has not altered the type of minerals present or their proportions on a scale commensurate with the alteration in appearance in the field and in thin section. (Figure 14)c–f include plots of semigneisses and gneisses which contain postgneissificadon interstitial microcline, and it is apparent that these are richer in total feldspar than the microcline-free rocks.

Microcline-paragneisses

Several bands composed entirey of microcline-rich gneiss occur, as shown on (Figure 15). Such gneisses are very similar in appearance to the granitic-orthogneisses and are difficult to distinguish from them in inland areas where exposures are not large enough to reveal boundaries or the variability of the gneiss ((Plate 7)e). The microclineparagneisses are less homogeneous than the orthogneisses and have gradational, instead of sharp, boundaries. Nevertheless, the gradation is rapid, so that they form more clearly defined masses than the other paragneisses; no independent or widespread areas of semigneiss having been found. In the microcline-paragneisses the microcline forms grains of the same size as the plagioclase grains, giving the rock an altogether different look to paragneisses lacking microcline or containing only interstitial microcline.

The average of four analyses of microcline-paragneisses is given in (Table 17). The average differs significantly in respect of several major and trace elements from the paragneisses, all in a manner that places the microclineparagneisses between the paragneisses and the graniteorthogneisses (Table 15) in composition. The modes of the microcline-paragneisses are shown in (Figure 14).

Interpretation

When Yell was first mapped in the 1930s the quartzofeldspathic banded, streaked and augened gneisses were thought to have been created by the injection of granitic magma into the metasedimentary rocks (injection-gneisses) and the homogeneous granitoid varieties of semigneiss and gneiss (permeation-gneisses) were thought to have been the result of recrystallisation by the juices' from the magma (Read, 1931). The rocks were, thus, considered to be migmatites (mixed rocks). Nowadays, such rocks are still labelled migmatites, but are generally considered to be the result of partial to complete melting by regional heating or isothermal uplift (Mehnert, 1968; Brown, 1974; Ashworth, 1985), or are attributed to metamorphic differentiation or segregation which is merely a description of the rocks in the absence of a proposed cause and mechanism and which cannot be applied to the homogeneous-gneisses.

The field appearance and occurrence of the rocks in Yell can be produced by neither origin. The whole area, gneissified and nongneissified, has been heavily invaded by a variety of pegmatites and aplites, but only after the gneissification and in a style very different to the leucosomes. There is no evidence of the leucosomes cross-cutting and veining of the rocks in the way that the pegmatites do.

The geographical distribution of the patches of paragneiss is such that they cannot be the result of anatexis. Anatexis is usually attributed to geothermal heating. Had the gneissification in Yell been caused by geothermal heating the heating would have been more or less uniform over the area leading to a more uniform distribution of the gneisses. The same result would be expected from anatexis due to isothermal uplift. Neither are the gneiss areas associated with intrusive rocks which could have raised the temperature of the rocks locally. There is no evidence in the distribution of minerals or of temperature determinations obtained from geothermometers of local variations in temperature which would be expected if the gneisses were due to anatexis or even partial anatexis. Nor is the localisation of the gneissification due to a patchy distribution of susceptible metasedimentary rocks. The Yell paragneisses are not the result of anatexis.

The essential difference between the metasedimentary rocks and gneissified metasedimentary rocks, both semigneisses and gneisses, is that the gneissified rocks have been recrystallised. The recrystallisation has resulted in grain growth and has taken place, as shown above, without any essential addition or subtraction of material.

The granofels are uniformly granular rocks with a grain size of less than half a millimetre. The quartz, plagioclase and biotite grains are evenly dispersed within bands or laminae, obeying the rule that in well-crystallised rocks grains tend to avoid contact with grains of the same type (Flinn, 1969). As seen in thin section, grain boundaries tend to be straight and to form triple point contacts. In a rock which has crystallised in this manner, the grain boundary system is very stable. The energy stored in distorted lattices and in irregular grain boundaries has been released and used during the crystallisation from the sedimentary state and the energy available from further decrease in grain boundary area is insufficient to drive the processes needed for grain growth, i.e. to cause small and otherwise unfavoured grains to break down and diffuse through the grain-boundary system to grains of similar type and lower free energy, there to add to their size.

The grain-boundary system has become locked. In monomineralic rocks such as orthoquartzites and marbles and rocks very rich in micas, grain growth accompanies rising grade, but not in the multimineralic rocks, in which grains of the same composition tend to be separate. In rocks containing flakes of mica separated from one another the grain boundary system is further locked due to the relatively high surface energies of the mica flakes. This results in the lower energy quartz and feldspar grain boundaries being strongly attached to the ends of the mica flakes so that grain growth has to take place uniformly through all the grains at the same time or these attachments have to be broken, which would require more energy (Voll, 1962).

It is apparent that rising temperature alone is insufficient to unlock the grain-boundary system of multimineralic rocks until the rock starts to melt. The granofels of Yell, and even the less micaceous schists, have reached kyanite grade with their sedimentary lamination and banding sharply preserved, yet the first effect of recrystallisation during gneissification is to destroy this sharpness. It is, therefore, concluded that following the original recrystallisation of the protolith sediments to meta-sedimentary rocks no grain growth has taken place.

The gneissification process which operated in Yell caused an increase of grain size to double or more with no other essential accompanying change; no essential mineralogical changes took place and no significant compositional changes occurred, except possibly in the microcline-paragneisses. In the field the recrystallisation of the granoblastic rocks is seen in the coarsening of the rock, the loss of bedding-type structures and the loss of lamination and banding as the boundaries become increasingly diffuse. The end product is uniform, granular, granitoidlooking gneiss with a grain size of 1 to 2 mm i.e. homogeneous-gneiss. In thin section these rocks have a foam-cell, triple-junction-type grain-boundary system, coarser than before but with the feldspars playing a more dominant role than the quartz grains. Whereas before gneissification they formed grains of the same shape and size as the quartz grains, afterwards they have more compact shapes to which the quartz grains have adapted. The mica flakes have usually lost some degree of preferred orientation.

The response of the schists to gneissification by recrystallisation is different. The schists, including rocks containing no more mica than some granoblastic micaceouspsammites, differ from the granofels in containing parallel, touching, mica flakes. This arrangement has facilitated their growth into matted streaks and laminae of mica flakes. Under these circumstances quartz–quartz, quartz–feldspar (oligoclase), and feldspar–feldspar boundaries are attached to mica basal planes and not to the ends of the mica flakes, as in the granofels. These boundaries meet the mica basal planes at 90° due to the high energy of the basal planes (Voll, 1961, 1962). The boundaries are, thus, not anchored as they are when attached to mica-flake ends but can slide freely along the mica flakes as the quartz and feldspar grains grow. This greater freedom to grow during gneissification recrystallisation leads to the growth of coarse quartzofeldspathic bands, streaks and augen in mica rich rocks i.e. the production of leucosome-semigneisses; the particular leucosome patterns developed depending on the mica fabric. This easy growth in the quartzofeldspathic laminae between the mica laminae leads to some swelling (segregation) as a biproduct of the recrystallisation. With continued recrystallisation grain growth also takes place in the layers between the leucosomes; the leucosome-semigneiss becomes more uniform and gneissic.

The process of gneissification described above is easily observed and clearly displayed in the field and in thin section. The cause of the unlocking of the grain-boundary system in some areas allowing grain growth to proceed, while in others areas identical rocks retain their original grain size, is not so obvious. The restriction of recrystallisation (gneissification) to apparently randomly distributed areas of irregular size unrelated to the lithology, is most simply explained in terms of these areas being permeated by a fluid which wets the minerals grains, i.e. has a lower surface tension against them than they have against each other. Under these conditions the dihedral angle of the fluid at triple junctions would be far less than the 120° characteristic of triple junctions between three minerals of the same type, so that the fluid would enter the grain-boundary system and flow through it (Voll, 1961). This would both unlock the system and facilitate diffusion between separated grains of similar composition, thus facilitating grain growth. In view of the observed lack of change of composition of the recrystallised rocks such a fluid would most likely be water.

A name is necessary to distinguish leucosome-gneisses and homogeneous-gneisses resulting from recrystallisation from rocks of a similar appearance resulting from the various types of migmatisation which have been proposed. To avoid confusion with the products of these other gneissification processes, these rocks will be called recrystallisation-paragneisses.

The cause giving rise to the recrystallisation process, that is the permeation of the rocks by a fluid, is the same as that proposed for the formation of the microporphyroblast gneisses and the interstitial microcline. The same conclusion was reached in each of the three cases, because in each the changes had taken place in a patchy manner in a terrain in which there is no structure, intrusion or variation in the nature of the protolith which can explain the localisation of the changes and in which there is no evidence of temperature or pressure differences. The different responses to the permeation are probably due to the permeation taking place under different conditions. The interstitial-microcline rocks are due to permeation by potassium-rich fluids at a late stage in the metamorphic history, probably when the rocks were a lot cooler than during the recrystallisation gneissification. The microcline-paragneisses may have been formed in the same way, but earlier and under higher-grade conditions. The conditions obtaining during the formation of the microporphyroblast gneisses are not so easily suggested. However, it should be noted that widely scattered and less prominent plagioclase microporphyroblasts occur in recrystallised granofels, in quartzofeldspathic Lewisian inlier gneisses and in the nebulite gneiss-es of the Boundary Zone. They are not as well formed as in the microporphyroblast gneisses, but they occur as a natural development of the preferential growth of plagioclase during the recrystallisation, as noted above.

Similar paragneisses of both homogeneous and leucosome-type were found during the survey of the Central Shetland Sheet arid areas farther south. In the Yell Sound Division in that sheet area patches of homogeneous- and leucosome-gneiss and semigneiss occur in psammitic granofels host rocks as described above for Yell. Such patches occur in Lunna Ness, in enclaves within the Graven Complex and farther south on the west side of the sheet.

In the Dalradian part of the East Mainland Succession, the Colla Firth Permeation Belt, a band of similar para-gneiss and para semigneiss, can be traced from the Spiggie area in the south of Shetland into the Central Shetland Sheet west of Scalloway and can be followed to the north and east across the Nesting Fault to the Out Skerries (Flinn, 1954, 1967; Flinn in IGS, 1981). This Belt differs from the paragneisses of Yell in being more closely confined to a single stratigraphic band, and in being more continuous, though it is broken in at least one place and is accompanied by several isolated elongate patches. It contains the same types of gneiss and semigneiss as occur in Yell but homogeneous-gneiss is dominant and leucosome-gneisses are a minor component, generally limited to the edges. It is for this reason that it was named 'the Colla Firth Permeation belt' (Flinn, 1954). Unlike the gneisses of Yell, the Colla Firth Belt is in rocks of biotite grade, but is itself of higher grade. Calc-silicate and marble bands within the belt, but not those outside it, are rich in diopside and microcline, while micaschist relics contain sillimanite. As in Yell, apart from calc-silicate rocks, microcline occurs in the Belt and outside it only as interstitial microcline indistinguishable in thin section and in field occurrence from the interstitial micro-cline of Yell. Modes of the gneisses and semigneisses plot in the same quartz-feldspar-mica field as those of the metasedimentary rocks, as do those of the Yell rocks. The Colla Firth Belt and the adjacent metasedimentary rocks have been heavily injected by aplogranites which were tectonised before metamorphism ended and by pegmatites, most of which, like those in Yell, were not. The aplogranites have been dated at 531 Ma (Flinn and Pringle, 1976).

On the basis of the same arguments used above for the gneisses in Yell, the gneisses of the Colla Firth Permeation Belt were attributed to recrystallisation by a permeating hot watery fluid (Flinn, 1954, 1967).

Chapter 7 Structures

This chapter is concerned with schistosities, folds, and faults, while a major shear zone, the Hascosay Slide, is described in the next chapter. Two schistosities are recognised in the rocks of Yell. They have been labelled S_A and S_B to avoid confusion with schistosities labelled S₁, S₂ etc. in the mainland of Scotland, which may or may not be their equivalent. SA is parallel to S_O, the sedimentary layering, and has been rejuvenated a number of times. S_B is a late spaced-schistosity or cleavage based on crenulation folding of S_A in mica-rich rocks. Folds occur on a wide range of wave-lengths, with a variety of profiles and provide evidence of having formed at different times. In most fold closures S_A is parallel to S_O, but in others S_A, as seen in planes normal to the fold axis, changes to a more or less complete girdle about the fold axis (L-tectonite to L-S tectonite), and in yet other fold closures an S_B develops as an axial plane schistosity in the closure region.

Folds vary in wave length from millimetres—crenulation folds—to kilometres—large folds. Most folds have a wave length of a metre or so, and formed during the regional metamorphism. Late folds occur which fold the lineation and other folds. Intrafolial folds occur which originated before or at an early stage in the regional metamorphism and which continued to develop through the later stages of the metamorphism by becoming increasingly compressed. The folds vary so much in size and profile that this classification can only be applied loosely. As usual in the metamorphic rocks of Shetland, it is not possible to set up a rigid sequential classification of folds in the form F₁, F₂ etc. because the behaviour of S_A in the fold closures depends as much on the lithology of the folded rock as on the time the folding took place. Furthermore, there is no reason to suppose that, because the crenulation folding giving rise to S_B looks the same in different parts of Yell, it has formed at the same time all over Yell.

The banding, foliation and schistosity in Yell is mostly steeply inclined, with a northerly strike, but approaching the Hascosay Slide in north-east Yell, the dip flattens to about 30°W and locally becomes horizontal. Large-scale folding is almost restricted to cliff-scale noncylindrical contortions associated with the Graveland and Breakon Orthogneisses. Isoclinal folding of medium scale and on nearly horizontal axes occurs along the north coast. These are upright in the west and become overturned to the east as the Hascosay Slide is approached.

Numerous small faults of unknown offset, some of possibly zero net offset, can be seen in the cliffs. Three faults with 5 km dextral offsets intersect the area and are probably of Jurassic agc.

Schistosity

Throughout Yell a penetrative schistosity formed by the parallel orientation of mica flakes occurs parallel to the compositional layering of the rocks and generally dips steeply west. The layering is interpreted as being of sedimentary origin. The origin of the schistosity is not so simply determined. The fabric defined by the mica flakes is in most places an S-fabric or at least an S > L fabric, but locally it changes to an L-fabric or an L > S fabric. The linear component of the fabric, the mineral lineation, is developed unevenly throughout the area and in places seems to depend in part on the nature of the lithology. It generally has a shallow plunge to the north-north-west (Figure 16). The L-dominated fabrics occur in highly folded areas and in fold closures.

Some folds can he seen to fold earlier folds and the S_A schistosity shows a variety of relationships to the fold closures. Furthermore, garnets in many rocks have been rotated so that strings of inclusions marking an early schistosity have been rotated out of alignment with the schistosity of the enclosing rock. Consequently, although SA is the earliest detectable schistosity in the area, it was reactivated at a number of different stages during the metamorphism of the rocks, and it was in existence both before and after the formation of the garnets and of some of the folds mentioned above, and it still retains its parallelism to the sedimentary layering in the rocks.

The S_A schistosity shows no signs of having ever been an axial plane cleavage. Folds formed before or early in the metamorphism continued to develop during the later stages of the metamorphism by flattening or rotation of their limbs towards parallelism with their axial planes and both rotating towards parallelism with the local trend of S_o. Such folds thus become intrafolial folds with limbs and axial planes parallel to the adjacent foliation. In the more micaceous rocks (schists) S_A can be traced round the flattened-fold closure, but in less micaceous rocks (granofels) the S_A schistosity has recrystallised in the closure region from an S-fabric to an L-fabric, the mica flakes taking up a decussate arrangement when viewed in the plane normal to the fold axis. However, most fold closures are of mica-poor bands in which it is not possible to be sure of the nature or orientation of the schistosity. In a few of the flattened or intrafolial-type folds the mica flakes in the closure region have been reoriented to an L-S- or even S-type fabric and so form an axial planar schistosity in the closure region. This is interpreted as having developed by continued flattening of the linear-rearranged micas in the fold closure region of intrafolial folds. It does not appear to be a relic of early-formed axial planar cleavage.

The S_A schistosity which throughout its period of development has remained parallel to the sedimentary layering, probably originated as a mimetic schistosity based on the original parallel or semiparallel compaction-enhanced arrangement of the micas in the sediments. This parallel relationship occurs throughout the East Mainland Succession of Shetland, except locally in the Clift Hills Division (Flinn, 1967).

In some coarsely micaceous schistose rocks there occurs a late, spaced cleavage or schistosity, S_B ((Plate 8)a). It cuts obliquely through S_A. This spaced schistosity results from the fusing of alternate limbs of systematically spaced and very regular crenulation-microfolds of S_A. It is a typical crenulation (strain-slip) cleavage. With progressive development S_A is suppressed and S_B enhanced to become a penetrative or nearly penetrative schistosity. It only occurs in rocks which are sufficiently mica-rich and with a sufficiently well-formed S-tectonite schistosity for the microfolding to take place. It is common throughout the outcrop of the Mid Yell Schist but is less clearly recognisable in the adjacent Kaywick Schist. In the very coarse mica-rich schists of the latter it is possible that S_B, the new schistosity, has taken over completely from S_A. The coarseness of the Kaywick Schist has led to the destruction of all evidence of sedimentary banding. Scattered occurrences of S_B occur elsewhere in the Yell Sound Division, especially in the closure regions of folds in the more micaceous rocks. S_B can only develop after S_A has developed, but S_B may develop as an axial planar schistosity in folds of micaceous bands at the same time that S_A is being folded smoothly round the closures of less micaceous folds, or is being converted to L or L > S fabrics in the closures of even less micaceous rocks. Therefore, all occurrences of S_B were not necessarily formed at the same time.

In the great majority of occurrences S_B is formed some 20° anticlockwise (in plan and looking down) from S_A; both dipping steeply, while in a few occurrences the rotation is about 20° clockwise. In the former case the crenulation folds viewed down the plunge are z-folds and in the latter s-folds. Both s- and z-folds of this type occur in schists between two splays of the Nesting Fault and it should be noted that vertical-axis z-type kink folding dominates over s-type in the fine-grained phyllites along the Nesting Fault in the southern part of Shetland (Flinn, 1977b). The main development of S_B probably accompanied the formation of the Nesting Fault, since z-type folding matches the dextral displacement on that fault.

Folds

Very few folds, other than steep-axis crenulation folds and intrafolial folds, are visible in inland exposures in Yell, because most folds plunge at shallow angles. Most folds are seen in cliff sections where they are widely, but very patchily, distributed. For descriptive purposes they may be divided into five classes. Large folds occur which are individually as large or larger than the cliffs within which they are seen, the largest being bigger than the highest cliffs in Yell at about 100 m. Smaller than these but larger than crenulation folds are medium folds which most commonly have wave lengths less than a metre or two, but grade into the large folds. The most widespread folds are small intrafolial folds with amplitudes of only a few tens of centimetres and which tend to plunge steeply down dip. The axial planes of these folds (other than the crenulation folds) are parallel to the general attitude of the adjacent layering and schistosity. As well as these four types of fold , there are also late folds, a very variable group grading into kink folds and folds formed by shattering.

Large folds

These are especially well developed along the west coast for 5 km southwards from the north-west extremity of Yell and for several kilometres along the coast in the region of the Graveland orthogneisses (Figure 16). They show no regularity or symmetry in profile and are strongly noncylindrical. The complex curvature of the axial planes seems to be original and not the result of later folding. Although they are noncylindrical their axes, in general, do appear to approximate to a north-north-west plunge of about 10 to 20°. The large folds in the Grave-land area are closely associated with the Graveland Gneiss bodies. A large fold at Breakon with a north-westerly plunge folds the northern part of the Breakon Gneiss. The gneisses are early intrusions and their presence may have controlled the folding in the adjacent areas. At the south entrance to Basta\Toe is a large reclined double fold of s-type with a north-north-west plunging axis. The fold closure can be seen on the coast at locality [HU 537 937]. However, this may be a late fold resulting from stresses set up between the Bluemull Sound Fault and the Arisdale Fault. Another large-scale monoclinal fold with a near horizontal axis occurs in northeast Yell and results in a decreasing dip of axial planes and schistosity across the Yell Sound Division–Boundary Slide contact and into the Hascosav Slide.

Medium folds

These grade in size into either intrafolial and minor folds or into large folds. Some occurrences are ideally shaped isoclinal folds of s- or z-type but others are of irregular profile like the large folds. Most plunge between 0° and 30° to the north-north-west. However, on the west coast, 1 or 2 km to the north of Lumbister [HU 477 998], and at Gerherda [HP 475 009], there are small areas of steep north-north-west-plunging folds. In these areas fold axes and lineations can be seen to change continuously from shallow-angle plunges to near down-dip plunges within distances of a few metres.

In L-tectonite areas the medium folds are cylindrical and, in conformity with L-tectonite symmetry, have no regularity of orientation of axial planes as seen in the plane normal to the fold axes. In L > S tectonite areas such folds are often isoclinal with axial planes tending to parallel the foliation, but the folds arc not as flattened as intrafolial ones. The mica fabric in the closures of medium folds varies according to the nature of the lithology being folded. Some closures of very micaceous rocks show an axial-planar spaced schistosity, but others show a simple folded schistosity, as do contrast-layered lithologies. In folded schistose psammites the S-tectonite schistosity in the limbs can, in many places, be seen in the plane normal to the lineation in the closure region to change to an L-tectonite fabric by reorientation of the micas to a decussate instead of a parallel arrangement, as described above.

The different fabrics in the different fold closures are the different responses of different lithologies to the folding; they are not the result of different deformations at different times. Nevertheless, folding had undoubtedly occurred over a long period prior to the pegmatite injections. Small flattened intrafolial folds are probably the oldest folds and can occasionally be seen to have been folded by the later larger ones. Their flattening resulted from continued deformation during which the larger folds were formed.

Intrafolial folds

These are generally small isoclinal fold closures strongly flattened in the plane of the foliation. Occasionally two fold closures can be found linked in the form of s- or z-folds. While there is some tendency for groups of s-folds or groups of z-folds to occur, the two types probably occur in equal numbers over the whole island. Such folds are common for several kilometres along the west coast north of Ulsta and are widely though unevenly distributed along the whole west coast. They are less common in the rest of the island.

The fold axes of intrafolial folds lie within the plane of foliation of the rocks and are most commonly parallel or inclined at up to 40° to the dip direction of the foliation, showing considerable divergence between adjacent folds even within small exposures. Thus they are, in general, steeply plunging folds. This variation in plunge within the plane of the foliation is especially well developed in areas of strong S-tectonite development, as for example in the coast south of West Sandwick.

In L-tectonite areas these intrafolial folds are much less common and their axes tend to plunge at shallow angles to the north-north-west, approximately parallel to the general direction of the lineation. They are not obvious structures because, in conformity with the L-tectonite pattern of the rocks, they are not flattened and so difficult to distinguish from other larger folds. Such minor folds show a particularly clear relationship of s- and z-folding to the major (large) fold of the Breakon Gneiss. The short limb of this major monoclinal z-fold is the site of a series of s minor-folds while z minor-folds are confined to the longer limbs of the major fold.

Late folds

Late folds occur throughout the area. Some are kink folds and some are cataclastic folds. Others are recognised as late folds because they fold the lineation and other folds and because they have divergent orientations. South of Gerherda on the west coast [HP 477 000] there are several patches of kink-folding with short limbs of the order of 20 cm in length. Small faults have developed along several of the kink planes.

Much distortion of the foliation takes place in shatter zones associated with faults and in shatter zones with no detectable displacement. Where this occurs on a large scale, as in the south-east of Bigga, it is impossible to determine whether it is due to shattering of a fold or folding during shattering.

Field occurrence

Most folds in Yell have near horizontal axes and in inland areas are difficult to detect except where they are on a smaller scale than the exposure. The best traverse for the study of folding in Yell is provided by the cliffs along the north coast of the island. Folding is less common on the west coast and even less common in the other coasts so that the occurrence of folding in the north cliffs is not typical of Yell as a whole.

Folding in the cliffs of the north coast of Yell varies in abundance and style. There is a lack of systematic textbook-type pattern to the folds. The folds have very shallow plunging axes with a north to north-north-west trend. There is a persistent tendency for them to be overturned towards the east, but sense of overturn is often difficult or impossible to determine. A sufficiently large section of fold profile may not be available due to the fold being larger than the cliff or the cliff face may have been rendered inscrutable by weathering or by massive and/or intense pegmatite injection. In some cases adjacent closures (synformal and antiformal) have been separated by deformation so that long and short limbs cannot be distinguished. The folds may not show a sense of overturning or the folding is so complex that the rocks have been structurally homogenised (L-tectonites). These difficulties apply in varying degree to all the coastlines.

Along the north coast the folds vary in wavelength from centimetres to 100 m or more. Except for the large Breakon fold, medium- and small-sized folds are more common in the east arid larger-sized folds in the west. In profile, folds vary from ideal similar to isoclinal, and while some have single, simple closures others are more complex and have multiple closures in the closure region. Many of the folds are flattened in the plane of the generally west-dipping foliation. The axial planes of such folds, therefore, tend to dip steeply west in the west and less steeply west towards the east end of the north coast.

Some fold closures show the S_A schistosity to be folded parallel to S_O but in the more tightly folded closures the S-tectonite-type S_A schistosity has changed to an L-S- or even L-type schistosity as described above. However, in many folds no schistosity can be detected due to the paucity of mica or due to poor weathering. Extremely rarely a spaced-cleavage axial-plane schistosity (S_B) occurs (e.g. [HP 483 053] ) and even more rarely a penetrative axial plane schistosity (e.g. [HP 513 056]). Folds are cut by aplites, pegmatites and lamprophyres.

On the evidence cited above most of the folds appear to have been formed after the emplacement of the Breakon and Graveland granodiorites, after the initiation of the S_A schistosity, after the emplacement of some early pegmatites and after the initiation of the recrystallisation gneissification and at different times. However, the folding (other than the late folding) occurred when the rocks were still at metamorphic temperatures which allowed continued development of the SA schistosity during folding. The lamprophyres, tonalites, and most of the pegmatites and aplites were emplaced after the main folding but before the late folding.

The continuity of the lineation, fold axes and schistosities across Yell and into the Hascosay Slide (Chapter 8), and the ductile nature of both folding and deformation in the slide zone, suggests that the 'sliding' took place towards the end of the period of folding and regional metamorphism. In this case the persistence of overturning towards the east across the north of Yell to the Hascosay Slide zone was interpreted by Flinn (1988) as evidence that Yell was thrust upwards towards the east over the Hascosay Slide during sliding (or Unst was thrust down to the west). However, since then a more detailed study of the Slide, reported in Chapter 8, has failed to find any evidence of sliding (monoclinic movement). The Slide appears to be the result of extreme compression which led to some flattening of the rocks to the west, thus rotating the folds and their axial planes into a westerly dip and producing the appearance of an easterly overturn. This effect can only have been enhanced by the large-scale folding on near horizontal axis of the rocks of north-east Yell.

In Unarey and Bigga, to the west of the Nesting Fault and close to the Graven Complex, the schistosity and sedimentary layering tends to dip to the east and the folds plunge to the east, unlike anywhere east of the Nesting Fault in Yell. Many of the folds show well-developed overturning to the south. The folds are unevenly distributed and very irregular in form and include many near-kink folds and folds formed during fracturing. Most have wave lengths of no more than several metres, but some are much larger. Where very schistose rocks are folded intense crenulation folding occurs in the fold closures, but very rarely gives rise to spaced (S_B) cleavage. In some places the rocks have been crumpled into L-tectonite zones of irregularly formed subparallel folds and micro-folds showing no vestige of symmetry or regularity as seen in the plane normal to the fold axes. These folds are mostly oblique to the fabric lineation which in the rest of Yell tends to be parallel to axes of folds other than those of intrafolial and late type. The folds of Unarey and Bigga are late folds which have probably been localised and controlled by the adjacent Graven Complex.

Faults

Faults are a common feature of the Yell coastline. Only three are of any great significance: the Nesting Fault and its two splays, the Bluemull Sound Fault and the Arisdale Fault ((Figure 2) and (Figure 16)). The two splays combine with the Nesting Fault just south of the Yell Sheet where it shows a dextral offset of 16 km. The Bluemull Sound Fault and the Arisdale Fault each appear to offset the geological pattern of Yell about 5 km dextrally, leaving a dextral offset of about 6 km for the Nesting Fault in the area of the Yell Sheet. The Nesting Fault and its splays form a short cut across a major bend in the Walls Boundary Fault and thus may be of the same age, that is Jurassic (Flinn, 1977, 1991).

The Nesting Fault is seen in the cliffs of Samphrey and also farther north in Ness of Sound [HU 450 826]. In both localities it is accompanied by very heavy and extensive shattering of the adjacent rocks. In Samphrey the Fault is steep and multiple and includes several lenses of transported rocks, including marble [HU 464 765]. The granodioritic and psammitic rocks on either side of the Fault in Samphrey [HU 465 760], are intensely shattered, being crushed down to grain-sized fragments leading to a loss of identity in places, especially close to the Fault (Flinn, 1977b). At Ness of Sound the rocks are coarser-grained gneisses and pegmatites and are sheared and shattered into a coarser breccia. Along the shore of Yell at Taing of Setter and to the south [HU 454 817] anastomosing thin shears are to be seen in the cliff. They are roughly parallel to the Nesting Fault, which lies just offshore, and are associated with patches of cataclasis, shattering and networks of small fractures containing a white-weathering, very fine-grained, infill ((Plate 8)b). These fractures with infill occur also on West Sandwick Holm [HU 434 893] and farther north on Sweinna Stack [HU 436 917]. They are also common in the Orka Voe and Calback Ness areas of the Central Shetland Sheet and are a common feature of the cataclastic fault rocks along the Walls Boundary Fault (Flinn, 1977h). They appear to be of hydrothermal origin and are formed of laumontite-leonhardite.

The Arisdale Fault is nowhere exposed, but exposures close to it, for example in the Arisdale Burn [HU 483 823] to [HU 479 837], are highly shattered. However, its entire course in Yell is marked by a prominent valley drowned by the sea at either end.

The Bluemull Sound Fault is exposed on the south coast of Yell and is a simple, steep fracture enclosing a band of powdered rock, up to 1 m wide, and a zone of less intense cataclasis extending tens of metres to the west. At Otterswick [HU 523 848] and at Vatsetter [HU 533 900] the Fault is a multiple fracture dipping about 45⁰ to the west and enclosing cataclastic slices of local rocks up to 10 m thick. West of Burra Ness [HU 548 954] and [HU 550 958] the Fault is obscured or only partly exposed but at North Sandwick [HU 552 970] it forms a broad zone of powdered rock 100 m wide containing large masses of highly cataelastic rock. The powdered rock, like much of the shattered rock, is probably deeply weathered cataclasite. In the next exposure to the north of North Sandwick [HU 551 982], it is a single narrow vertical fracture lying in a zone of strong shattering several tens of metres wide.

The Gamla Fault is an observable fracture in the cliffs, separating local rocks of very different type in the northern part of its course. The displacement cannot be determined.

The Breakon Fault has a small dextral offset, as have several small faults to the south of Cullivoe. South-east-trending faults at Aywick and Otterswick are not exposed but their presence is indicated by sinistral offset of litho-logical boundaries crossing these bays.

Prominent fractures are abundant common in the cliffs; possibly as common as every 10 m or so in places. Some may be joints, but most show some sign of movement having taken place, either by the interruption and occasionally the offset of lithological boundaries, or by the presence of cataclasis and/or shattering varying from zones of fractured rock 10 m or more in width to a fracture-infill of a centimetre or more of powdered rock. Such powdered rock is almost always too coarse and too incoherent to be classed as gouge. Where lithological boundaries can be traced across such fractures the apparent offset is small and could be due to either normal, reverse or transcurrent offset. These faults have probably been activated at different times under different stresses. In such cases although the final net of^-1kt may be small the total gross offset during the working of the fault may have been large, thus explaining the large amount of shattering associated with the small net displacement (Flinn, 1977b).

Chapter 8 Hascosay Slide

The Hascosay Slide is a zone measuring up to a kilometre in width, and displaying the results of intense deformation at high temperature. It is called slide because, to some extent, it resembles the 'slides' in the Moine of mainland Scotland in being a zone of intense deformation and recrystallisation separating lithologically different areas. It is a zone of closely packed masses of coarse-grained, dominantly hornblendic rocks, cut by, and contained within, fine-grained and dominantly hornblendic blastomylonites. However, it presents no evidence of a monoclinic sense of shearing. Only in Hascosay is the whole width of the zone completely exposed.

The Hascosay Slide can be traced along the north-east coast of Yell and passes through the centre of Hascosay ((Figure 3) and (Figure 16)). In the north, in Migga Ness [HP 540 050], the western edge of the zone is exposed. In Papil Ness [HP 540 040], Crussa Ness [HP 547 030] and Cullivoe Ness [HP 552 024] both the western margin and the central part of the zone are exposed. In North Sandwick [HU 550 970] the central part of the zone and in Burra Ness [HU 555 955] the eastern margin are exposed. In all these areas the exposure is exceedingly good in the cliffs. In Hascosay the zone is poorly exposed on the north coast east of Taingar [HU 550 934], extremely well exposed on the south coast east of Ramna Geo [HU 556 915], and not at all inland. The zone is intersected and offset dextrally some 5 km by the Bluemull Sound Fault.

The slide zone has two main components, a blastomylonite schist and contained within it residual masses of partially to completely sheared and recrystallised gneiss varying in size from tens of centimetres to several tens of metres. The foliation in the blastomylonite generally dips to the west at 30–45° and trends NNW-SSE, parallel to the zone outcrop.

The rocks to the east of the Slide belong to the Dalradian part of the East Mainland Succession (the Valla Field and Lamb Hoga Blocks of Unst and Fetlar – Flinn, 1958), while those to the west belong to the Boundary Zone. There is no apparent lithological discordance at either junction, but both junctions are marked by a continuous but rapid increase in the intensity of development of the schistosity.

Blastomylonite

There are three types of blastomylonite. One type—the aplite-blastomylonites—has the field appearance of a tectonised aplite, being a white aplite-like rock which is faintly but intensely laminated ((Plate 9)b). In thin section the laminations are seen to be due either to rectilinear strings of tiny biotite flakes or, more commonly, to laminae one grain thick of long quartz grains (cf. streaked-out quartz grains in mylonites). The rocks have the appearance of very fine-grained schists (0.1 to 0.2 mm) of very uniform grain size with no signs of strain or cataclasis. The aplite-blastomylonites occur interbanded with banded-blastomylonites (see below) on the margins of the slide zone, where they grade by decreasing blastomylonitisation into crudely schistose thin-banded aplites occurring conformably in the adjacent metasedimentary rocks. The aplite-blastomylonites differ from nearby younger aplites in containing biotite instead of muscovite, somewhat more epidote and generally much less microcline.

The banded-blastomylonites are as fine grained as the aplite-blastomylonites, and contain laminae as fine and as continuous, parallel and rectilinear, but the laminae and the thicker bands which accompany them are formed of hornblende ± biotite and, locally, clinopyroxene and garnet ((Plate 9)c). In hand specimen they often appear to be a highly sheared rocks, having alternating dark and light laminae up to a millimetre or two thick, and black hands a centimetre or two thick. Banding is absent in some places and much thicker in others. In thin section, however, such rocks are seen to be a very fine-grained highly schistose rocks lacking any sign of cataclasis, and with a grain size generally nearer to 0.1 than 0.2 mm. The laminae are more obvious in thin section than in hand specimen, being mostly 0.1 to 0.2 mm thick. They 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 laminae 0.1 to 0.2 mm thick and extending across the whole width of the thin section.

The banded-blastomylonites have a strongly developed orthorhombic fabric. Thin sections cut parallel to the lineation and normal to the foliation show sharply defined rectilinear laminations, while in sections cut normal to this and to the lineation the laminations are much less sharply defined. The hornblende and biotite flakes have a strongly preferred lattice and shape orientation parallel to the laminations in thin sections cut parallel to the lineation ((S92149); (S92154)) and a more decussate arrangement as seen in the sections cut normal to the lineation ((S92149); (S92154)). At the southern end of Cullivoe Ness [HP 552 022] the lamination is well developed, but the coloured minerals show no preferred orientation of lattice or shape (S91920). In some nearby areas hornblende has a lattice preferred orientation but no shape orientation and the biotite has a moderate lattice and shape preferred orientation (S91918).

The laminated- and banded-blastomylonites, unlike the aplite-blastomylonites, occur throughout the entire width of the slide zone passing through and closely wrapping round the residual masses of gneiss ((Plate 9)f). Their origin is revealed in the envelope zones of the individual hornblende-feldspar gneiss masses where they show all gradations from fine, continuous, parallel lamination parallel to and wrapping the sides of the gneiss masses to anastomosing and streaked-out networks of thicker coarser-grained laminae, where they are being drawn out from the ends of the gneiss masses ((Plate 9)e). These drawn-out lamellae can be traced back into the residual masses, into weakly deformed and even completely undeformed hornblendite or hornblende gneiss. This process is especially well displayed at Migga Ness [HP 539 053].

The third type of blastomylonites, the psammitic-blastomylonites, are much less obvious. They occur as fine-grained, faintly laminated mica-poor varieties of psammite occurring in the margins of the slide zone inter-banded with the two varieties described above; at Migga Ness they appear to be mylonitised country rock drawn into the slide zone. In the field they differ from the aplite-blastomylonites in having a pale brown colour like the psammites, instead of the white sugary appearance of the aplite-mylonites. They are composed dominantly of quartz and plagioclase, with some biotite and/or hornblende. The hornblende has probably been added mechanically from adjacent hornblendic rocks in the Slide. Larger masses of psammitic-blastomylonite occur within the slide zone, especially in Burra Ness [HU 555 956], where they appear to be blastomylonitised quartzofeldspathic psammites and gneisses similar to those to the west in the Yell Sound Division. Where such rocks occur within the Slide zone it is difficult to determine the limits, as there is no lithological break. The boundary to the slide zone at such places is marked only by a continuous and gradual decrease outwards in deformation.

The foliation varies from S-tectonitc to L-tectonite. The blastomylonites with the finest, most continuous and parallel laminations are mostly S-tectonites even though in hand specimen a faintly developed mineral lineation is often visible on the S-surfaces. The best L-tectonites occur at either end of the gneiss masses, in the envelopes, where they are being drawn out into blastomylonite ((Plate 9)e). Gneiss masses of suitable composition, such as coarse hornblende-plagioclase rocks, can often be seen to have been deformed to L-tectonites. In general, on the scale of the whole slide zone, the foliation of the blastomylonites appears to he parallel to the zone in both strike and dip. The mineral lineation plunges on average about 20° to the north-west.

In detail, the foliation defined by the lamination and the parallel schistosity in the blastomylonites has a complex orientation arising from two causes. The foliation wraps the included gneiss masses closely whether they have a well-formed eye-like lenticular shape or are more rounded or more angular. Locally the foliation is drawn into boudinage-type invaginations in the masses or passes through the masses as anastomosing, ductile shears. The other cause of complexity in the orientation of the foliation arises from folding of the blastomylonites. Extremely flattened intrafolial folds occur in the psammitic-blastomylonites, especially in Burra Ness. Less-extreme folds are more common in the banded-blastomylonites; they tend to have their axes parallel to the mineral lineation in the slide zone ((Plate 9)d). They often occur as groups of either s-folds or of z-folds. In some places such folds have no regularity of form when seen in cross-section and conform with the symmetry of the L-tectonite rocks containing them. These groups of folds are irregularly scattered in the slide zone and appear to result from the blastomylonite being squeezed between the irrregularly shaped resistant masses of gneiss.

Less easily understood are folds of the mineral lineation itself on exposed planar s-surfaces in the form of tight isoclinal closures with wave lengths and amplitudes of a metre or so [HP 548 027]. Where the exposure allows parallel s-surfaces to be examined, especially ones a centimetre or two apart, it can be seen that the folding pattern on them is completely different, or even absent.

Some fold closures, especially double closures (s or z) are preserved within lenses contained within the foliation. Such folds, having formed early in the deformation, probably by uneven flow of the deforming blastomylonite, have resisted further deformation and have been preserved as a relic lenses contained within the foliation of the blastomylonites as they continued to deform.

Minerals

Quartz, plagioclase, biotite and hornblende are the most common minerals in the blastomylonites. In the five microprobed samples of aplite-blastomylonite and blastomylonite the plagioclase varies from An₃₁ to An₂₀. The biotite analyses are shown in (Table 18) and plotted in terms of Al–Mg–Fe on (Figure 7)a; all are several per cent lower in Al than the biotites in the psammites to the west. Biotites occur as the only coloured mineral in the apliteblastomylonites and very often as an accessory mineral to hornblende in the handed-blastomylonites. Muscovite occurs in 10 out of 80 sectioned blastomylonites and interstitial microcline occurs in 12, but neither of these minerals, nor biotite, are essential constituents of aplite-blastomylonite. Muscovite and microcline occur in the blastomylonites of Burra Ness and more so in those of Hascosay. Only one section of blastomylonite from farther north was found to contain microcline.

Hornblende is an essential constituent of the bandedblastomylonites and is generally green in colour but occasionally green-brown. The analysed hornblendes, when recalculated to a total of 13 cations plus Ca + Na + K, were found to vary between ferroan pargasitic hornblende, ferroan pargasite and tschermakite, without relation to appearance or occurrence. All three occur in a single specimen ((S91920); (Table 18)).

The extremely fine grain size of the blastomylonites makes the recognition of garnet and pyroxene very difficult in the field. Eight of the 80 thin sections were found to contain garnet and four clinopyroxene and these all came from the south part of Cullivoe Ness [HP 552 022] where the blastomylonites exhibit no lattice preferred orientation of the minerals. The analysed garnets (Table 18) are shown plotted on a Ca–Mg–Fe triangle ((Figure 7)b). They tend to contain less MnO than the garnets in the metasedimentary rocks to the west, i.e. 0.5 to 2.0 per cent.

Geothermometry

The microprobe analyses of these minerals have been used with several geothermometers to estimate the temperature at which the minerals have formed (Table 19). For the blastomylonites, the very fine grain size of the minerals, the intimate mixing of the minerals, their perfectly crystalline and noncataclastic state, and their blastomylonite origin indicate metamorphism under ideal conditions for chemical equilibrium to be established between the minerals. However, the temperatures obtained from geothermometry studies vary widely; most lie between 600°C and 900°C, with a maximum of 925°C. The temperatures obtained by garnet-biotite geothermometers are in general about 100°C higher than those obtained by means of the garnet-hornblende and garnet-pyroxene geothermometers. In these and other geothermometry calculations (Chapters 4 and 5) the Perchuck geothermometers were found to give temperatures less scattered for calculations both within and between the individual thin sections. However, the usual scatter of estimated temperatures found when a variety of geothermometers are used on the same analyses is magnified in this case by real differences in the rocks.

The temperature estimates from the two residual gneiss masses (garnet-pyroxene-hornblende-biotite gneiss (S94706) and (S91371) evidently reflect the metamorphic state of the rocks before they became part of the Slide. Blastomylonite sample (S91918) gave relatively high temperatures, and is from the part of the Slide in Cullivoe Ness [HP 550 025] containing pyroxene, and is expected, on thin-section evidence, to have been formed at higher temperature than elsewhere. If these facts are taken into account and the garnet-biotite temperatures are rejected as too high for the rocks concerned (in conformity with common practice in geothermometry studies), the Hascosay Slide seems to have formed at between 600 and 700°C. This is a temperature not very different from that found for the metamorphism of the Yell Sound Division.

Geobarometry

The Kohn and Spear (1990) geobarometer was used with these analyses (Table 20) and gave pressure estimates between 7 and 11 kb. The different values seem to bear no direct relation to the appearance or occurrence of the rocks tested and in general are considerably greater than the values found for the metamorphism of the Yell Sound Division. However, they do bracket the pressure (10 kb) obtained with the same geobarometer for the hornblende schists in the thrust below the ophiolite to the east (Flinn et al., 1991).

Geochemistry

Two of the aplite-blastomylonites have been analysed (Table 21) for comparison with associated unblastomylonitised aplites described in Chapter 9. The aplite-blastomylonites have less SiO₂ and more TiO₂ and MgO than the aplites. The trace elements show greater differences, especially in respect of Va, Ni, Zr, Zn, Ce, Nd, and Ba, all of which are higher in the blastomylonite-aplites. With the exception of Ba, these differences could be attributed to addition of material to the aplite-blastomylonite from the blastomylonitic hornblendic rocks, but the aplite-blastomylonites and the aplites are probably unrelated.

Residual masses within the slide

The relic resistant masses within the slide zone, ranging in size from tens of centimetres to several tens of metres, are mostly coarse hornblendite and hornblende gneiss, possibly because anything less resistant, such as psammite or psammitic gneiss, has generally been deformed into blastomylonite. In the field many of these masses are black and featureless. They can be classified as hornblendite gneiss, hornblende > feldspar gneiss, hornblende-feldspar gneiss, hornblende-banded gneiss, hornblendite and even hornblende gabbro. These resistant masses grade into hornblendic blastomylonite and banded blastomylonite over short (centimetre-scale) distances both within themselves in ductile shears, and at their junction with the envelope-blastomylonite containing them.

All the residual masses have been completely recrystallised. Hornblende gneisses which appear in the field to he coarse and unaltered are seen in thin section to have recrystallised so that hornblendes originally several millimetres across are now aggregates of hornblendes 0.1 to 0.2 mm in size. The feldspars are similarly recrystallised to aggregates of slightly larger grains. In the hornblende-feldspar gneisses the pyroxene cores to the hornblende grains have, in nearly all cases, recrystallised to microgranular aggregates of amphibole. The grains within undefortned aggregates show no apparent preferred orientation. Preferred orientation develops as deformation increases.

The hornblende-feldspar gneisses are often sufficiently coarse to provide a black and white pattern visible in the field and can be seen in places to have been penetratively deformed to L-tectonites. The L-tectonite fabric changes progressively to an S-tectonite fabric as ductile shears are approached. In thin section the hornblendes and feldspars can he seen changing progressively from an equidimensional pattern to, first, a streaked-out pattern, then to a streaked-out net of hornblende containing lenses of feldspar formed from the original feldspar grains. This pattern progresses via anastomosing coarse laminae of hornblende to the parallel rectilinear laminae of the typical banded-blastomylonites. In many of these rocks what appear in hand specimen to be black hornblende grains are clinopyroxenes enclosed in a rim of hornblende.

The coarse hornblende-feldspar rocks are probably metagabbros, but whether they were intruded into the slide zone, like the globular metadolerites described below (Chapter 9), or whether they were emplaced tectonically, cannot be determined. Such rocks occur along the exposed length of the zone ((S91941); (S94712); (S91586)).

In the field the different gneiss masses are very difficult to delimit and to distinguish due to the blackness of the hornblendic rocks and the ubiquitous and very variable blastomylonitisation which not only contains the masses but also passes through them as ductile shears. Banded hornblendic gneisses formed of alternating layers, 10 to 20 cm thick, of hornblendite and of quartzofeldspathic rock are prominent in the North Sandwick exposure ((Plate 9)a). They are similar in appearance to the Lewisian banded-hornblendic gneisses occurring in places in the mainland of Scotland and the Outer Hebrides, but not to the Lewisian inliers of the Yell Sound Division. Most common within the slide zone and most inscrutable are the coarse hornblendite gneisses, very like those in the Lewisian inliers to the west ((Plate 9)e). In the Burra Ness area, and rarely elsewhere, are masses of psammite and psammitic gneiss similar to common rocks in the Yell Sound Division. These are much less resistant to deformation than the hornblende-containing gneisses and deform more uniformly, so that they are difficult to distinguish from the country rocks adjacent to the slide zone.

Only a few occurrences of gneiss were found to contain garnet. At Migga Ness is a garnet-hornblende-pyroxene-plagioclase gneiss (S94728). Others occur on the south coast of Hascosay ((S91370); (S91371); (S94706)) and are microgranular rocks composed of clinopyroxene (safite - (Figure 17)), garnet, biotite, antiperthitic andesine (c. An₃₆) and ilmenite, and with grain sizes of 0.1 to 0.3 mm. The fine grain size is due to the recrvstallisation of a much coarser texture so that original large pyroxene and feldspar single-crystal grains are now aggregates, probably due to recovery re-crystallisation after being strained, though no signs of deformation or cataclasis now remain (S91371). Thin section (S94706) is from a rock which was originally a coarse-grained gneiss but which has now been streaked out into a 1 mm laminated rock with large garnet grains contained in thin laminae of pyroxene and garnet separated by thin laminae of feldspar. With the exception of the large garnets the rock is now a recrystallised microgranular aggregate showing no signs of strain or cataclasis.

The temperatures obtained by geothermometry of the residual masses of gneiss show a wide scatter, but tend to be greater than those for the blastomylonites, probably reflecting the metamorphic state of the protolith ((Table 19) and comments above).

Only two out of more than 100 thin sections of gneiss masses contained microcline. (S91497) is a thin section of a coarse (2.0 mm grain size) microcline-plagioclase-hornblende diorite and (S91720) a microcline>plagioclase hornblende gneiss.

Ultramafic rocks also form resistant masses within the slide zone. They are of two types. The most easily recognisable are lenses and masses varying from fist size to 3 m across of dark lustrous-green actinolite. They are associated with talcose lenses and bands containing brown biotite and paler actinolite. Also present are masses of serpentine and carbonate in association with some of the above minerals. They are all parts of disrupted zoned balls, i.e. masses of serpentinite which by metasomatism have developed concentric zones of such minerals as talc, chlorite, phlogopite, actinolite and anthophyllite. Similar, but better-preserved, bodies occur on Unst, where they have been described by Read (1934) and Curtis and Brown (1969). Zoned balls occur in Migga Ness [HP 539 053], Brough [HP 540 045], Papil Ness [HP 545 040], Crussa Ness [HP 547 028] and Burra Ness [HU 549 949]. The coarse antigorite-carbonate-opaque mineral core (S94722) of an actinolite-talc lens has been analysed ((Table 21), analysis 1).

More difficult to evaluate in the field are the less-altered ultramafic masses occurring at the southern tip of Ness of Cullivoe [HP 553 022]. These are very rounded, up to 7 m in diameter, with thin skins of talc and/or serpentine and are extremely difficult to sample with a hammer. Thin sections cut from the core samples contain clinopyroxene and orthopyroxene in fine-grained aggregates as well as such minerals as actinolite, antigorite, and biotite ((S91922); (S91923); (S91926)).

White quartz-feldspar pegmatites and aplites occur in some of the resistant gneiss masses and show deformation and crude blastomylonitisation in conformity with that in their host rocks. Lenses of schistose quartz-feldspar pegmatite occur between the individual bands of the banded-blastomylonites, especially in the western edge of the Slide in the neighbourhood of aplite-blastomylonite. They are probably residual masses of pre-slide zone pegmatite occurring in the Boundary Zone and the Yell Sound Division to the west.

Late episodes in the development of the slide

Minor but tectonically important components of the slide zone are small vein-like occurrences of hornblendic rock with a grain size of 0.1 to 0.2 mm which crosscut, or are contained as lenses within, the blastomylonite foliation. They occur along the whole exposed length of the slide zone and are usually less than 10 cm thick, although larger masses do occur. They appear to be intrusions of basaltic composition which were emplaced late in the formation of the slide zone. They were emplaced in two phases. Those emplaced in the first phase are strongly schistose amphibolites (S94727) with a fabric similar and parallel to that of the adjacent blastomylonites. In places such veins cut obliquely through the blastomylonite foliation and in others they are contained within it as lenses. The veins emplaced in the later phase are more obviously cross-cutting and are very fine-grained amphibolites with at best only a weak fabric parallel to their edges (S91715). More than 27 occurrences were observed and five of these, some schistose and some not, contained garnet. One garnetiferous rock ((S94709); (Table 21)) was analysed and was found to be of subalkaline tholeiitic composition with MORB affinities (Figure 11).

All the rocks in the Slide are cut by veins and sheets of quartz-plagioclase-microcline aplite and pegmatite. These are distinguished from older aplite-blastomylonites, and from the relatively crudely tectonised aplites and pegmatites contained within the residual masses, by their lack of tectonite fabric and by their cross-cutting relations to the blastomylonites. Most are in the form of veins of the order of 10 cm thick. The pegmatites cut the aplites in several places and thus appear to be slightly younger. The pegmatites mostly contain large reddish single-crystal microcline patches and are indistinguishable from the microcline pegmatites occurring throughout Yell to the west. They typically form continuous wandering veins cutting through all the other rocks. The late aplites are much less continuous and tend to be concentrated at the western margin of the slide zone near to the aplite-blastomylonites.

The late aplites and pegmatites and all the other rocks of the slide zone are cut by lamprophyre sheets similar to those found throughout Yell (Chapter 9). These tend to cut the slide zone approximately normal to its 'length and they are untectonised. Spherulitic and autobrecciated types occur.

The youngest rocks in the slide zone are rare, very thin stringers of microcline pegmatite (Chapter 9), usually 2 or 3 cm thick, which in several places cut across the lamprophyres. Such late pegmatite veins can be found, very rarely, cutting the lamprophyres elsewhere in Yell and in the area of the Central Shetland Sheet.

The slide zone has suffered the same amount of late folding, faulting and shattering that is found throughout Yell.

Field occurrence of the Hascosay Slide

The best area to study the relationships in the western margin of the Slide is the cliff top at Migga Ness [HP 539 053], where large ball-shaped masses of residual hornblendic gneisses of various types can be seen both enclosed within and being streaked out into banded-blastomylonite. Intimately interbanded aplite-blastomylonite, psammite-blastomylonite and handed-blastomylonite also occur, enclosing lenses of schistose-pegmatite and isolated fold-closures of blastomylonite less strongly developed than that enclosing them. The eastern margin is best exposed in the cliff top at south-east of Hascosay [HU 560 915] where relations between aplite-blastomylonite and younger cross-cutting aplites are displayed and a large relic mass of garnet-pyroxene gneiss occurs (S91350). The eastern margin is more accessible but less completely exposed in Burra Ness [HU 555 956]. Here balls of ultramafic rocks several metres in diameter occur within the blastomylonites. A sand-polished exposure in the nearby beach at locality [HU 552 956] displays the relationships of many of the components of the Slide in a remarkably clear manner. The coastal section at North Sandwick [HU 553 976] reveals clearly the relationship of the residual masses of Lewisian-type hornblende-banded gneiss with the enclosing blastomylonites. Spectacular folding of the bandedblastomylonites can be seen in the coastal section between Crussa Ness and Cullivoe Ness, especially at locality [HP 548 028]. Biotite-actinolite-talc lenses are best seen on the south side of Papil Ness [HP 545 040], pyroxenite lenses at the south tip of Cullivoe Ness [HP 553 023] and an anthophyllite-actinolite-talc ball (S94814) at Papil Ness [HP 540 045]. Wherever the coastline intersects the slide zone the exposures are continuous, accessible and instructive.

Structure

The Hascosay Slide is a narrow zone of extreme deformation. It is formed of a series of numerous closely packed resistant masses of dominantly hornblendic rock enclosed in a matrix of very fine-grained intensely laminated material derived in part from the resistant masses and in part from intrusive aplites. The fine grain size of the residual rock masses is due to recovery recrystallisation of large grains which were heavily strained but not significantly deformed. The fine grain size in the laminated rocks is due to the breaking down of the mineral grains in the original rocks by recrystallisation during the deformation, while the lamination is due to grain-boundary flow of the fine-grained material during this deformation. The temperature of 600 to 700°C during the deformation allowed the comminuted material to continuously recrystallise as it flowed so that no cataclastic, bent- or strained-crystal textures were developed. The rocks are like normal regional metamorphic schists but of very much finer grain size. No subsequent coarsening recrystallisation has taken place because the multimineralic nature of the rocks inhibited collective recrystallisation. It should be noted that the quartz laminae in the apliteblastomylonites, mentioned above, formed originally by the streaking out of quartz grains so common in mylonites, have recrystallised to a series of long thin unstrained single crystals, but the process has operated only within the quartz lamellae.

Throughout the whole exposed length of the Hascosay Slide the blastomylonites show the same range of grain size, and of fineness and perfection of lamination and banding in both field, hand specimen and thin section. However, in Crussa Ness there is an area within which even the finest and most perfectly laminated blastomylonites shown no preferred orientation of lattice or of grain shape of the biotite, muscovite, hornblende or pyroxene grains forming them. In the rest of the slide zone the minerals exhibit strongly defined S to L-fabrics of lattices and, often also, of grain shapes, exactly conformable to the lamination and banding. Thus the blastomylonites have an orthorhombic symmetry on the scale of handspecimen. Only three out of 80 thin sections showed an element of monoclinic symmetry. In these, a schistosity defined by lattice and grain shape intersects the laminations at an angle of 20° or so. Two of these ((S94809), (S94812)) were blastomylonitised vein-like bodies cutting ductile shears intersecting the lamination of banded-blastomylonites.

Orthorhombic symmetry for the slide zone as a whole is indicated by larger-scale structures. The foliation of the blastomylonites tends to strike parallel to the zone and dips to the west more or less steeply in different parts of the zone, but in detail it follows a highly contorted course as it twists and turns around, through and between the residual masses. In the field, not only can such irregular contortions be seen everywhere, but it can also be seen how the laminated and banded blastomylonite has been extruded from between close-spaced residual masses into less-restricted spaces and there been thrown into flowing folds ('extrusion folds'). The axes of the extrusion folds have a considerable scatter, but tend to plunge 20° to the north-north-west, like the lineations in the blastomylonites. The axial planes of the folds, apart from containing the fold axes, have no preferred orientation. The folds show no regularity in profile. Folds overturned one way can nearly always be matched within a few metres by folds showing the opposite sense of overturn, without any anticlinal structure between.

An interesting demonstration of orthorhombic symmetry in the blastomylonites is provided by an east–weststriking ptygmatically folded pegmatite vein which cuts banded-blastomylonite normal to the banding [HP 544 031]. The ptygmatic-fold axes have an east–west trend, while the axial planes are parallel to the blastomylonite foliation. The vein suffered an apparent shortening normal to the foliation of the order of x10, and this must have taken place during the blastomylonitisation.

The Hascosay Slide is in contact with psammites of the Boundary Zone to the west, and Dalradian schists and gneisses of the Valla Field succession of Unst and Fetlar to the east. The high temperature ductile deformation which occurred in the slide zone has led to a 'smeared' _.junction with the adjacent rocks. In detail, schistosity, lineation and lithology pass gradationally from the rocks on either side into the slide zone. Disruption of the lithology on a larger scale is not apparent, whether it exists or not, because of lack of lithological variation in the Boundary Zone on the western side and lack of sufficient exposure in the eastern side. There is no evidence for, or against, any great displacement across the zone, because it lies within the Boundary Zone which itself appears to be a series of tectonic slices. The eastern edge of the Slide may conform to the eastern edge of the Boundary Zone or may cut it out.

Two pieces of evidence may limit the magnitude of the displacement across the Slide. Firstly, the microporphyroblast gneiss can be followed from Lunna Ness through Yell to Hascosay close to the west edge of the Slide, and then can be traced from the east edge of the Slide in Hascosay to well to north in Unst. Secondly, the Slide is not seen to the south of Hascosay. If it continues southward it must pass between Fetlar and the Out Skerries to the south-east, yet the metasedimentary rocks of Unst and Fetlar, to the east of the Slide, can be correlated in some detail with the East Mainland Succession to the west of the Slide. Finally, while some of the residual masses of rock within the Slide may be matched with rocks outwith the Slide (principally with rock to the west), other masses cannot.

Radiometric age

The unaltered state of the blastomylonites, the parallel relationship between the lattice and grain-shape fabrics of the hornblende and biotite grains and the lamination and banding, and the lack of evidence for later alteration makes it very likely that the Ar-Ar step-heating ages for these minerals are closely related to the time at which the Slide was formed (Chapter 10). Since hornblende has the higher closure temperature its age of 496 ± 6 Ma is taken as the minimum age for this event. The age 436 ± 7 Ma given by the biotite probably reflects the time taken for the Slide to cool from the closure temperature for hornblende to the closure temperature for biotite.

It should be noted that the age found for the blastomylonitisation is not significantly different to that found for the onset of obduction of the ophiolite in Fetlar to the east, also based on hornblende (Flinn et al., 1991).

Interpretation

The occurrence of high-temperature minerals of unusually fine grain size, which are perfectly crystalline, non-cataclastic and unstrained, in very highly deformed rocks with coincident lithological and mineralogical fabrics indicates mylonitisation-like deformation at temperatures so elevated as to allow complete recrystallisation during the mylonitisation. For these reasons the rocks have been called blastomylonites. The blastomylonites have been derived in this manner from a variety of hornblende-rich gneisses of 'Lewisian' aspect, from gneissified and ungneissified quartzofeldspathic granofels belonging to the Boundary Zone and from aplites.

Since the preferred lattice and/or grain-shape fabrics of the blastomylonites are orthorhombic, these fabrics, and therefore these rocks, must be the result of orthorhombic deformation. Since the slide zone, itself, shows orthorhombic symmetry on the large and medium scale and often on the small scale, it too must be the result of orthorhombic deformation. There is thus no fabric or structural evidence of sliding or thrusting in the slide zone. It is a zone of intense pure shear and compression within which rock has been mobilised, squeezed out, and forced to flow parallel to the lineation towards the north-north-west and the south-south-east. The systematic overturning of folds along the north coast of Yell towards the Slide to the east, was earlier taken as evidence of thrusting to the east (Flinn, 1988). In Chapter 7 the folds are explained as having been rotated towards parallelism with the Slide zone by the same compression that created the Slide. This interpretation leaves unexplained the source and emplacement of the residual masses in the zone, but it conforms with the evident lack of a major displacement across the Slide.

Late in the development of the Hascosay Slide it was invaded by thin veins of a rock of tholeiitic subalkaline composition and of MORB affinity, very similar in composition to the globular ophitic metadolerites in the adjacent rocks to the west, but differing from the much older hornblende schists in the Yell Sound Division in being significantly richer Na₂O and FeO + Fe₂O₃. After the formation of the Hascosay Slide, it was invaded by pegmatites and later by lamprophyres similar to those elsewhere in Yell.

The Hascosay Slide is evidently a different type of structure to the slide zones in the Moine of mainland Scotland; perhaps in recognition of this fact it would be better to call it the Hascosay Welt.

Chapter 9 Late intrusive rocks

The late intrusive rocks were emplaced after, or at the end of, the regional metamorphism of the area. They include three different types of minor intrusion and part of a major intrusive mass, the Graven Complex. The minor intrusions are, in order of frequency of occurrence, pegmatites and aplites, lamprophyres and tonalites. All Yell has been heavily injected by pegmatites and, to a lesser extent, by aplites. The pegmatites occur most profusely in the west and north of the island while the aplites are concentrated in the south-east. They have been emplaced as veins, sheets, masses and networks and in places so profusely as to make up 50 per cent of large areas. A number of different varieties of pegmatite and aplite can be distinguished, and it is apparent that they were emplaced in several different episodes. The lamprophyres are spessartites and kersantites. They were emplaced after all but the very youngest phase of pegmatite emplacement and are unevenly distributed throughout the area. The tonalites are a very minor group and were probably emplaced before most, but not all, of the other minor intrusives. The tonalites are concentrated in the south-east of Yell. The south-west corner of the sheet area includes the edge of the Graven Complex, a granodiorite of appinitic affinities, about 10 km in diameter, which was emplaced soon after the lamprophyres at about 400 Ma.

Tonalites

In the field, these fall into two clearly distinguishable types of rock, both of which are concentrated on the south-east coast of Yell in the Boundary Zone (Figure 18). The Burravoe-type arc common at Heoga Ness [HU 525 790], as veins of complex form and as networks in the psammites. They are very difficult to detect because they have the same colour and grain size as the psammites in which they occur and because of their irregular form. It is mainly their lack of tectonite fabric which enables them to be distinguished from the rocks they intrude. Their time of emplacement relative to the lamprophyres cannot be determined directly as no lamprophyres occur in the same area. However, they are cut by both pegmatite-aplite and microcline pegmatites which are themselves cut by the lamprophyres. The tonalite veins are, thus, dated as postmetamorphic and prelamprophyre, but their lack of tectonite fabric indicates that they were emplaced after the earliest pegmatites.

In thin section their fabric is seen to be dominated by rounded, equidimensional and occasionally zoned, plagioclase grains up to a millimetre in diameter, which are often very strongly altered. Quartz, which is less abundant than plagioclase, is interstitial between the plagioclase, and is accompanied by biotite and/or hornblende, together with very variable amounts of epidote, including allanite, and spheric. The Burravoe-type tonalites do not have a tectonite fabric ((S94686); (S90742)).

The other type of tonalite, the Vatsetter-type, occurs mostly between Aywick and Vatsetter. It is mineralogically similar to the Burravoe-type but forms large, north-trending sheets with a weak tectonite fabric parallel to their boundaries. The presence of this fabric probably indicates that they were emplaced earlier than the Burravoe type. These sheets are easier to detect in the field than intrusions of Burravoe-type because of their regularity of form and their size. In thin section the Vatsettertype tonalites are very similar to the Burravoe-type except for the presence of a weak schistosity. All the tonalites are characterised by feldspar-dominated textures.

At Aywick, a 10 m-thick, north-trending sheet of Vatsetter-type tonalite which dips some 60°W divides into a number of smaller sheets when traced to the south [HU 538 869]. It has a coarse core ((S91178) – grain size 1 to 2 mm) with finer-grained edges containing ill-defined darker lumps of tonalite (S94804). Some metres to the east is a 3 m-thick parallel sheet, and a similar distance again to the east occurs yet another mass of the same rock type. They are all weakly tectonised, with a lineation plunging some 16° to the north.

In the cliffs to the north-east of Aywick a sheet of tonalite can be followed for about a kilometre and possibly for three (Figure 18). It is exposed in the back of the Stoal Geo and in the west side of the Snute as a double sheet [HU 547 874]. It cuts the southern limb of the Vatsetter Lewisian Inlier parallel to and just east of the Gamla Fault, where it is about 10 in thick. It is cut and offset to the south by the Gamla Fault in the vicinity of Ern Stack [HU 544 883] and is seen again in the cliffs immediately to the south, just east of the fault, where it is about 7 m thick and is accompanied by another parallel sheet of tonalite nearly 3 m thick. Two sheets occur in the north tip of Vatsetter Ness and may be a northward continuation [HU 537 902]. One, a metre wide is cut by a spherulitic lamprophyre, the other (S91246) has a core of coarser tonalite.

On the east side of Ness of Gossaburgh [HU 537 835], a sheet of rock of tonalitic appearance but granodioritic mineralogy, a metre or so wide, can be followed for some distance to the south. In its northern parts it contains only a little microcline (S90991) but at its southern limit, where it is several metres wide, microcline is a major constituent (S90991). A hundred metres or so to the east, in Saamari Geo [HU 535 825], is a sheet less than a metre wide of coarser (2 mm grain-size) biotite-granodiorite.

There are several intrusive sheets with a field appearance similar to that of the tonalites in North Sandwick, and others farther north. One cuts a residual hornblendic-gneiss mass in the Hascosay Slide [HU 553 970], but contains biotite and no hornblende (S91698). Two small sheets occur west of the Bluemull Sound Fault in the same area [HU 550 969]. One cuts the schistosity and layering of the metasedimentary rocks but is itself folded. The other is cut by a sheet of tonalite-like rock. Several tonalite-like intrusive sheets occur in the north-east of Yell (Figure 18).

The tonalite at Horse of Burravoe has characteristics of both types of tonalite. It forms a continuous sheet, like the Vatsetter types, but strikes east–west and dips north at a shallow angle and, lacking a tectonite fabric, resembles the Burravoe-type tonalites in petrographic appearance.

In the Burravoe area two of the globular metadolerites are cut at their contacts by short veins up to 10 cm thick of rock indistinguishable in the field from Burravoe-type tonalite (Y3.12 [HU 527 788] and [HU 594 684] [HU 526 788]). In thin section they can be seen to be composed of quartz, fresh plagioclase, biotite and garnet and have the appearance of psammitic granofels. It is possible that they are mobilised psammite.

Geochemistry

Analyses of Burravoe and Vatsetter-type tonalites are included in (Table 22). There is little or no significant difference between the two analyses and when plotted on discriminant diagrams show both tonalite types to be talc-alkaline rocks (Figure 19). It is noticeable that the two analyses are, for most elements, very similar to the group of lamprophyre analyses and to the average analysis for the Graven granodiorite; all are presented in the same (Table 22) and (Figure 19). However, the tonalites differ significantly from these other rocks in having much lower Cr and Ni contents and they are relatively low in MgO and Sr.

Pegmatites and aplites

The whole of Yell has been heavily injected by sheets and masses of pegmatite and to a lesser extent by aplite. In places, the results are so spectacular in the field that they attracted the attention of Jameson (1798) and were especially noted by Hibbert (1822) and Heddle (1879).

A number of different types of pegmatite and of aplite can be distinguished in the field, but gradation between types and the bleaching of pink microcline by weathering under peat makes it impossible to allocate every occurrence to a particular type. However, it is possible to distinguish the following types:

Ptygmatic pegmatites

These are ptygmatically folded quartz and feldspar pegmatite veins usually 2 or 2 cm thick. They occur sparsely and patchily throughout the Yell Sound Division in Yell, in the islands of Yell Sound, and on the Mainland, but nowhere else in Shetland. Most occur in the northern half of Yell. Wherever intersecting relationships with other pegmatites are observed, the ptygmatically folded ones are the earlier. The most interesting occurrence is the ptygmatically folded pegmatite vein which cuts the blastomylonites of the Hascosay Slide at Cullivoe (Chapter 8).

Albite pegmatites

The albite pegmatites contain large white feldspars up to about 10 cm in diameter, which show fine lamellar twinning in the hand specimen. Quartz is the only other mineral apparent in the field. They appear to contain no microcline and none has been found in the several thin sections cut. The most spectacular occurrence is the giant sheet forming the Ern Stack in the cliffs of Graveland [HU 454 963] in north-west Yell (Plate lob). Except in their more prominent occurrences they are difficult to distinguish from some of the other types of pegmatite.

Biotite pegmatites

These are not the only pegmatites to contain biotite, but are distinguished by the common occurrence of groups of several biotite plates, each up to 15 cm across and a millimetre or so thick, intersecting in a common line, and splayed over 10 or 20° so that where the line of intersection is normal to the exposure surface they look like the footprints of a large bird (hence the field name of crow's foot pegmatite - (Plate 10)a). They are very easy to recognise in the field and in Yell are especially common along the west coast, to the north of West Sandwick.

Large biotites in pegmatites were noticed by Hibbert (1822), though whether in this type of pegmatite or the next to be described, is not clear. In the Central Shetland Sheet, to the south, they occur in the general area of Orka Voe as xenoliths in the Graven granodiorite. A biotite from one such xenolith was dated (K-Ar) at about 400 Ma but this is probably the date for contact metamorphism by the enclosing granodiorite (Miller and Flinn, 1966). Some biotite pegmatites contain no potash feldspar and grade into the albite pegmatites; others vary towards the muscovite and biotite pegmatites.

Muscovite and biotite pegmatites

This name is given to two-feldspar pegmatites seen to contain good fat books of muscovite, up to 4 cm across, and/or less commonly, biotite books of the same size. They become increasingly difficult to recognise as the numbers and size of the mica books decrease. Good examples are exposed especially profusely along the coast north and south of West Sandwick and to the north of Lumbister.

Microcline pegmatites

This name is confined to those pegmatites distinguished by the presence of large irregular 'blotches' of pink to red microcline, up to 15 cm in diameter, occurring in quartz-plagioclase pegmatites. The individual 'blotches' are generally formed of, or contain, a single large crystal of microcline. They are not the only pegmatites to contain microcline but, especially where the microcline has not been bleached by weathering, they are an easily recognised group. They are the most commonly occurring, and most widely distributed, of the distinguishable pegmatites and are unevenly distributed throughout the island. They are the last of the pegmatites to be emplaced, except for rather rare thin stringers of micro-cline pegmatite which cut the lamprophyres. The stringers, although of rare occurrence, are very widely distributed.

Undifferentiated pegmatites

These are all the pegmatites which for any reason, including poor exposure, cannot be allocated to the classes defined above. They are composed of quartz and feldspar and some contain small amounts of microcline biotite ± muscovite. They form the vast majority of the pegmatites observed and occur unevenly distributed throughout Yell.

There are three types of aplite (Figure 20).

Aplite

Typical aplites occur, composed of quartz, plagioclase, microcline and occasionally muscovite and have the usual white, fine-grained appearance.

Pegmatitic-aplite

These are an intimate mixture of aplite and pegmatite. In appearance, they could have been emplaced as a mixture of two 'magmas' or be due to later, very patchy re-crystallisation of the aplitic component. The pegmatite component forms small patches in the aplite but without the appearance of being xenoliths, as they have very irregular crystal-controlled boundaries which in places appear gradational due to irregular distribution of grain sizes. Within the pegmatitic patches very large white feldspars are often prominent. Pegmatitic-aplites are not to be confused with vein complexes composed of aplite and of pegmatite; these also occur in Yell.

Aplites and pegmatitic-aplites occur preferentially in south-east and south Yell and in Bigga. Stated in this way the northern 'limit' to the frequent occurrence of aplites and pegmatite-aplites appears to have an east–west trend. However, this is an artefact of the three 5 km-offset dextral faults that cut Yell. Restoration of the offset on these faults rotates this limit of aplite and pegmatitic-aplite into a NE–SW-trending boundary to a field of aplite injection which extends southwards from Yell into Lunna Ness in the Central Shetland Sheet.

Blastomylonitic aplites

Blastomylonitic aplites occur in the Hascosay Slide and are described in Chapter 8 (aplite-blastomylonites) as penetratively and uniformly schistose aplite-like rocks, which contain biotite and generally lack microcline. They must be distinguished from the schistose varieties of the aplite and the pegmatitic-aplite described above, which occur very sparsely.

Schistose varieties of pegmatite and aplite occur widely but sparsely distributed throughout Yell, including the Lewisian inliers but more commonly within the Hascosay Slide. The schistosity in such rocks is crudely developed as a network of anastomosing ductile shears enclosing relics of pegmatite or aplite. This fabric is recognisable in the field in the pegmatites and in thin section in the aplites.

(Table 22) shows analyses of a typical aplite showing no signs of deformation and a nearby aplite which exhibits a late nonpenetrative schistosity (not an aplite-blastomylonite). The analyses show noticeable differences in Ca, K and Rb contents, but in the absence of further analyses the status of these differences cannot be determined. The rocks were analysed for comparison with the less siliceous blastomylonite-aplites in the Hascosay Slide (Table 21).

Distribution and occurrence

All the varieties of pegmatite and aplite mentioned above are unevenly distributed, the different types tending to have their own geographic concentrations. In general the aplites are concentrated in the south of Yell and in the Hascosay Slide, and the pegmatites in the west and the north. However, the paucity of inland exposure makes the true boundaries of the concentrations impossible to determine.

Both pegmatites and aplites occur as sheets varying in thickness from centimetres to tens of metres, and also as masses larger than the cliffs in which they are exposed so that their shape cannot be determined. The largest masses of pegmatite occur in the cliffs of the west coast. Similar large masses occur as xenoliths in the Graven granodiorite in the area of the Central Shetland Sheet. Such bodies, up to half a kilometre across, led to the suggestion that a pegmatitic 'granite' formed an early phase of the Graven Complex (Flinn, 1954). This now seems most unlikely, as the pegmatites within the Graven Complex, whether isolated within the granodiorite or forming part of xenoliths, enclaves and screens of metasedimentary rocks and gneisses of the Yell Sound Division, are so similar to those throughout the Yell Sound Division of Yell.

Lamprophyres

About 270 minor late intrusions of lamprophyre type have been recorded in the cliffs of Yell (Figure 21). Very few have been found in inland areas. Their absence from inland exposures and along the inner coast is difficult to explain in areas where they occur profusely in nearby cliffs of the outer and intermediate coasts, but must be due to failure of exposure rather than to their absence.

They mostly occur as sheets inclined to the layering of the host rocks, only about one quarter being conformable to the layering. Particularly good examples of horizontal, conformable sheets, varying up to 2 m thick, occur in the south-west of Hascosay [HU 542 912]. A stereo-plot of poles to all the measured lamprophyres shows a wide scatter of poles across the south-east quadrant and overlapping the adjacent quadrants by 20°, indicating a very general dip to the north-west. The majority are single sheets, although several are seen to pinch in and out in an en-echelon manner across the strike. In several places (e.g. North Sandwick) lamprophyre forms a small net-vein complex (e.g. [HU 553 973]), and a small number of multiple sheets have been found (e.g. [HU 539 893], [HU 550 969]). On Gloup Holm [HU 485 063] a small coarse-grained mass of hornblende diorite (S92183) contact metamorphoses the country rocks and appears to be directly connected to nearby lamprophyre sheets (S92156). Some lamprophyre sheets show chilled margins.

The lamprophyres vary in thickness from 4 cm to 10 m, but most are 30 to 100 cm thick. The largest are the 10 m thick, east-west-striking dykes at Hascosay [HU 547 909], Cullivoe [HP 549 022]), and Breakon [HP 524 029] to [HP 522 029]. The Breakon dyke lies within the Breakon Fault. In Sligga Skerry [HU 434 800] xenoliths of lamprophyre (spessartite, (S90699)) occur in the granodiorite of the Graven Complex and in Bigga [HU 451 785] a xenolith of kersantite occurs in an intrusive sheet of the granodiorite. At several localities, including two occurrences on Papil Ness [HP 546 043], the lamprophyres are cut by veins of micro-dine pegmatite 1 or 2 cm thick. Apart from these pegmatites, the lamprophyres are the youngest hard rocks on Yell and can be observed cutting the other pegmatites in a number of widely scattered places. In places they have been shattered, faulted and/or extensively altered. Very few are unaltered.

From field observations there appear to be two main types of lamprophyre present – spessartites and kersantites – with the former predominant (Figure 21). The unaltered representatives are green hornblende-phyric (S94703) and red-brown hornblende-phyric ((S94805), (S92043)) spessartites and red-brown biotite-phyric kersantites (S94811). However, as a result of alteration and variation most lamprophyres are difficult to classify in the field or even in thin section. Phenocrysts, even when present, are often indeterminable due to alteration. Several occurrences were found of lamprophyre sheets containing both biotite and hornblende phenocrysts.

Three subvarieties of the lamprophyres have been mapped (Figure 21); spherulitic, autobrecciated and xenolithic. Spherulitic (variolitic) lamprophyres (both spessartite and kersantite – (S91614); (S94708)) occur widely, the spherulites occurring in a close-packed manner in the centres of the sheets. Most spherulitic lamprophyres occur on the east coast of Yell and they probably form the edge of a swarm of such rocks, which is much better developed to the east in Unst and Fetlar.

Autobrecciation of lamprophyres is also to be observed along the north-east coast of Yell, but may be confined to the spessartites. In autobrecciated sheets angular fragments, varying up to about five centimetres across, are seen 'floating' in a matrix of very similar appearance, but of slightly finer grain size. In general this pattern is only visible on sand-smoothed intertidal surfaces. Lamprophyres containing xenolithic material are uncommon, except locally. Corroded xenocrysts of quartz are most common. In North Sandwick [HU 550 960] to [HU 550 965] a series of subconformable dyke-like sheets of kersantite (S91589), up to 3 m thick, are particularly rich in xenoliths, especially along their eastern edges. The xenoliths include quartz, red feldspar, country rock and also clinopyroxenite fragments containing biotite similar in appearance to that in the lamprophyre.

The lamprophyres occur very sparsely in the south of Yell and increase in numbers toward the north-east (Figure 21). The kersantites tend to be concentrated into several small areas, while the spherulitic and autobrecciated lamprophyres are almost entirely confined to the north-east coast.

On Gloup Holm there occurs a mass of coarse hornblende-rich diorite several tens of metres across. Thin sheets of it penetrate the country rocks as spessartites. The country rocks at their contact with the diorite appear, in the field, to be somewhat contact metamorphosed. In thin section they are seen to contain considerable amounts of shimmerised fibrolite ((S92184); (S92185)). Thin sections of the diorite (S92183) show well-formed equidimensional red-brown hornblendes, 2 to 3 mm across, in a matrix of equally coarse plagioclase; a texture differing from some of the lamprophyres only in the coarseness and the equidimensional rather than elongate shape of the hornblendes. The offshoots from the diorite are typical red-brown hornblende spessartites (S92186).

In the extreme south-west of Yell, two plagioclasephyric lamprophyre sheets occur. At Copister [HU 472 785] there is a sheet of plagioclase-phyric kersantite (S90711), and the rocks underlying the broch at the Holm of Copister [HU 473 779] are cut by a sheet of plagioclasephyric spessartite. These are the only two plagioclasephyric lamprophyres found in the Yell Sheet, although porphyrite sheets are very common in the Central Shetland Sheet. The plagioclase-phyric lamprophyres of Yell form a link between the lamprophyres and the porphyrites which, like the lamprophyres, were emplaced prior to the Graven Complex.

Minerals

Of the 90 samples of lamprophyre examined it was found that no two thin sections were identical, except where they were possibly from different parts of the same intrusion. Most lamprophyres are partially to heavily altered to epidotic minerals, white micas, chlorites and opaque minerals, all in a matted and inscrutable mass. However, in the better-preserved sheets an extremely wide variety of textures is exhibited. The feldspars (plagioclase) always form a groundmass to the phenocrysts but vary from microgranular to as coarse as 0.5 mm and they are always clouded by very fine-grained minerals of various types including epidotes and amphiboles. In a number of sheets the feldspars are arranged to form poorly developed variolitic groups, the better formed of which can be seen in those sheets showing spherulitic patterns in the field. These variolites are nowhere as well developed as the very beautiful textbook examples to be found in the spherulitic lamprophyres of Fetlar. Microprobe analysis showed the presence of oligoclase (S92043) and albite (S94703) in spessartites, of oligoclase/andesine (S94708) in biotite-hornblende lamprophyre, and of albite (S94715) in kersantite. Due to the widespread alteration shown by the lamprophyres, especially by the matrix, it is not clear whether these are original or alteration compositions. Most lamprophyres contain at least some epidote. An interesting variant of kersantite is a sheet in the entrance of Gloup Voe containing clinopyroxene (S94726).

The phenocrysts in the lamprophyres vary considerably in size up to about 2 mm and in degree of alteration from completely destroyed to apparently perfectly unaltered. The analysed specimens were chosen from samples showing least alteration. Biotites vary in colour from dull brown to a red-brown in thin section. Those analysed ((S94715) and (S94708)) are shown plotted on (Figure 7)a and listed in (Table 23). If 67 per cent mg is taken as the division between phlogopite and biotite then the mica in (S94715), the kersantite, is a phlogopite (mg = 73 per cent), while the mica in (S94708), the hornblende-biotite lamprophyre, is a biotite (mg = 59 per cent). These two micas plot in the same compositional fields as micas from Scottish Caledonian lamprophyres (Rock, 1991, fig. 4.5a, b and f.

Hornblendes, where not altered, are either green or reddish brown in thin section. All gradations from colourless (apparently bleached by alteration) to strong versions of these colours are exhibited. Spessartites occur with only reddish brown hornblende, or only green hornblende or with both together. Analysed phenocrysts of green hornblendes were found to be magnesio-hastingsitic hornblende (S94703) and magnesio-hornblende (S94708) in biotite-hornblende lamprophyres, while reddish brown phenocrysts occur as tschermakitic hornblende (S94703) and magnesio-hastingsite (S92043). The green hornblende in (S94703) has colourless rims of magnesio-hornblende and is accompanied by a mass of tiny tschermakitic hornblende needles in the matrix, while the reddish brown hornblende in (S92043) has green rims of magnesio-hornblende. These identifications (Table 23) are based on a recalculation of the hornblende composition to 13 cations + Ca + Na + K for 23 oxygens and the Leake (1978) classification. These hornblendes plot in the same compositional fields as hornblendes from mainland Scottish Caledonian lamprophyres (Rock, 1991, fig. 4.4 a).

If it is assumed that the strongly coloured, unaltered phenocrysts of hornblende grew in a melt then the aluminium-in-hornblende geobarometer can be used (Johnson and Rutherford 1989). The green magnesio-hornblende in (S94708) gives a pressure of 4.4 kb, the brown magnesio-hastingsite in (S92043) 6.1 kb, while the brown tscherrnakitic-hornblende in (S94703) gives a pressure of 5.6 kb and the green edenitic hornblende in the same rock 5 kb. It is possible that the decreasing aluminium content of the amphiboles as the colour changes from brown to green and from green to paler greens in overgrowths reflects a rise of the magma in the crust as the minerals grow.

Geochemistry

Analyses of one of each of the four types of lamprophyres are presented in (Table 22). All four analyses are of calc-alkaline type according to the discriminant diagrams ((Figure 19)a and d) and plot in the same fields as the tonalites. The fbur analyses of lamprophyres are less similar to each other than the two tonalites are to each other. Some of the differences are systematic and predictable. K₂O and Rb increase from spessartite, through the hornblende-biotite lamprophyre to the kersantite. Most other differences are less systematically arranged. The major element analyses in (Table 22) show noticeable differences from Rock's (1991) 'World Averages' for spessartite and kersantite, all four being higher in SiO₂ and the spessartite lower in K₂O. However, all fall within the ranges given by Rock (1984, fig. 3) with the possible exception of the SiO₂ determination for analysis 5, (Table 22). They all fall in the (K₂O/Al₂O₃ v. Fe₂O₃/SiO₂) and (CaO v. MgO) fields for Caledonian mainland Scottish lamprophyres (Rock, 1991, figs. 5.3 and 5.4) and plot in or very near the same fields as the mainland Scottish late Caledonian dykes (Rock et al., 1988, fig. 3). The trace clement values in (Table 22) fall within the ranges for talc-alkaline lamprophyres quoted by Rock (1991, fig. 5.7), except that the values for Ni, Cr, Nd and Ba in (S94708) (the hornblende-biotite lamprophyre) differ noticeably from the equivalent values from the other three analysed lamprophyres and all four plot as 'outliers' on fig. 5.7 of Rock, 1991.

It is apparent that the lamprophyres of Yell are very similar in composition to the talc-alkaline lamprophyres of Scotland described by Rock (1991, fig. 5.7); in particular the high Sr content of the lamprophyres falls within the range quoted for Scottish Caledonian lamprophyres.

All the lamprophyres in Yell appear to have been intruded at about the same time, just before the emplacement of the Graven Complex c. 400 Ma ago (see below). They do not cut the Graven granodiorite, but are found in it as xenoliths. A lamprophyre on Unst about 2 km east of Yell and of similar appearance to many on Yell has been dated at 407 + 7 Ma by Ar-Ar step heating (Taylor, 1984).

Graven Complex

The granodiorite of the Graven Complex intrudes the Yell Sound Division in the south-west of the area ((Figure 2) and (Figure 21)). It is an appinitic granodiorite complex 10 km or more in diameter emplaced about 400 Ma ago (Miller and Flinn, 1966; Flinn, 1985). It causes a strong negative gravity anomaly and is associated with a sillimanite-bearing aureole. Psammite xenoliths and enclaves within the complex are generally much richer in sillimanite than the psammites of the aureole. In the latter fibrolite needles can be detected for some distance from the granite, but within the complex the xenoliths are often rich in sillimanite, containing both fibrolite, compact streaks of matted fibrolite, clumps of fibrolite (faserkiesel), and coarse crystals of sillimanite. The sillimanite mats and faserkiesel grow from the breakdown of biotite and are associated with K-feldspar (of interstitial type) only locally.

The complex contains many large screens, enclaves and xenoliths of country rock (Yell Sound Division with its pegmatites), often hundreds of metres across, but varying down to several metres, which retain their orientation, but which are distributed in a very patchy manner within the complex. Some of the smaller xenoliths of country rock, usually a metre or less in size, are disoriented. The granodiorite intrudes both the surrounding rocks and the xenoliths within it in a net-vein or sheeted manner, so that it is difficult to define a boundary for the granodiorite on maps of the scale of 1:10 000. Within the area of central Shetland to the south-west of the Yell Sheet it is possible to trace the schistosity and the zones of gneissfication of the Yell Sound Division through the complex by means of the oriented xenoliths. These include all the various types of metasedimentary rock, and paragneiss making up the Yell Sound Division, as well as the Valayre Gneiss, the hornblende schists, garnet-hornblende schists, the pegmatites and the lamprophyres. Lewisian inlier rocks and orthogneisses have not been recognised however. By retaining their orientation and position relative to each other the xenoliths provide a 'ghost stratigraphy'.

The granodiorite intrudes all the rocks of the area where it occurs, including the lamprophyres but is itself intruded by thin microcline-pegmatite stringers similar to those which cut the lamprophyres, and by thin aplite veins, both of which are very sparsely distributed. The complex is truncated to the west by the Walls Boundary Fault and is cut and offset 16 km dextrally by the Nesting Fault in the east.

The granodiorite is characterised by the presence of 1–2 cm sized coarse hornblendite cognate xenoliths. There are usually several of these in every square metre of exposed surface but they vary in concentration from absent to close packed, though the latter occurrence is rare except along the edges of the complex. The granodiorite is a medium-grained pinkish granitoid rock consisting of essential plagioclase and quartz with very variable amounts of biotite, hornblende and microcline, any of which may be absent. Sphene and allanite are common accessories. (Table 24) presents the average of eight modes of granodiorite collected outwith Yell. The grain size varies up to about 3 mm but is mostly between 1 and 2 mm. The rock varies from almost white to a pinkish brown darkened by a variable content of blackish hornblende grains. In places, xenoliths of the darker varieties occur in the lighter varieties, both containing the small hornblendic cognate xenoliths.

Field occurrence

The Graven Complex is exposed in the south-west of the Yell Sheet, on the west of the Nesting Fault in Bigga, Sligga Skerry, Uynarey and on the westernmost tip of Samphrey. On the east side of the Nesting Fault the Complex occurs on Samphrey between the fault and the Arisdale Fault.

On the beach at the southern tip of Uynarey, the Complex is a relatively light-coloured, fine-grained biotite-microcline granodiorite with only occasional cognate xenoliths (S90700). It can be traced north for 100 in or so, becoming increasingly rich in country rock screens until it ends as a few sheets of granodiorite in the country rock.

On the Sligga Skerries and Bigga the granodiorite is unusual in containing 1 to 2 cm porphyritic plagioclases and somewhat smaller bleb-like groups of quartz grains (S90682). This facies is relatively leucocratic and poor in cognate xenoliths. Locally it is cut by some aplite and thin pegmatite veins. Country-rock xenoliths arc widely scattered, but are most common near to the boundaries of the Complex as marked on (Figure 21). Xenoliths of spessartite occur in the Sligga Skerries (S90699). A xenolith of kersantite occurs in a granodiorite sheet cutting the country rocks in south-east Bigga. Beyond the mapped boundaries intrusive sheets of granodiorite occur, but for the most part less frequently. In some cases they are difficult to distinguish from the country rock and thus resemble to some extent Burravoe-type tonalites. However, in the Easter Hcvda Wick area, on the west coast of Bigga [HU 444 796], the country rocks are so heavily injected by granodiorite and related rocks that it is debatable whether or not they should be included within the boundary of the complex.

It is characteristic of the Graven Complex that, especially adjacent to its boundaries, hornblende-rich varieties of the intrusive rock occur. This is particularly well shown in Bigga, in the Easter Hevda Wick area, and at the contact at the south of the island where, however, relationships are obscured by late dislocations. In these areas, there are complex assemblages of country rock, medium and dark-coloured granodiorite with hornblendite cognate-xenoliths, and also dark-coloured granodiorite or diorite closely packed with hornblendite cognate-xenoliths ((S90687); (S94676)). The darker types of granodiorite are usually sharply separated from the lighter-coloured granodiorite and occur both as intrusive veins and as xenolith-like patches in the lighter-coloured granodiorite and in the country rock. Aplite veins several centimetres wide cut these rocks in the Easter Ilevda Wick area and in several other places including the Sligga Skerries. A 2 m-wide sheet of aplite cuts the granodiorite in the south-west corner of Bigga.

In Samphrey, the Nesting Fault can be seen cutting the Complex. On the south coast typical granodiorite is faulted against psammites of the Yell Sound Division [HU 465 760]. The granodiorite has been crushed to a suhgrainsize microbreccia for 100 m west of the Fault (S94802).

Further from the Fault, to the west in Bunglan, many cognate xenoliths and larger xenoliths of darker granodiorite can be seen in lighter-coloured granodiorite. In a possible late vein, the rock is biotite-hornblende-plagioclase granodiorite with the relatively fine grain size of 0.2 to 0.3 mm (S94680). A more usual grain size of about 2 mm is shown by the adjacent biotite-microcline granodiorite (S94679). On the west coast of Samphrey, to the east of the Nesting Fault [HU 464 765], the granodiorite is highly cataclastic and rotten. Its junction with the country rocks, which are also cataclastic and rotten, is an alternation of psammite and granodiorite sheets, the latter decreasing in frequency of occurrence until no more granodiorite appears. The whole exposure is difficult to interpret because of the cataclasis associated with the Fault.

Much more interesting are two breccia pipes of appinitic-type granodiorite found on the north-east coast of Samphrey completely isolated from the main Complex. The contact effects of the larger of the two pipes at locality [HU 470 763] extend over about 100 m along the coast. A smaller breccia pipe is exposed about 100 m along the coast to the north-west.

The north part of the larger pipe-like mass, where exposed on the coast, is composed entirely of a breccia of the country-rock psammite (with interstitial microcline) made up of angular fragments varying from 8 cm to 1 m across, disoriented and packed tightly together with no apparent matrix and with all the fractures healed by crystallisation (S94801). In contact with it to the south the igneous rock is composed of closely packed rounded fragments of variably to coarsely crystalline arid compositionally very variable rock composed of hornblende ± clinopyroxene ± biotite ± microcline (S90752) and more commonly hornblende-biotite rocks, both with some plagioclase (S90766). The fragments are about 4 cm in diameter but decrease to 1 to 2 cm towards the breccia. This highly xenolithic rock changes sharply to a dark variety of the Graven granodiorite containing many hornblende cognate-xenoliths (S90768) and also a xenolith-like patch of close-packed hornblendite cognate-xenoliths (S90767). At the southern contact the psammites are unbrecciated. The smaller pipe is composed of a few metres of breccia west of a small patch of close-packed hornblendite cognate xenoliths, which is in contact with unbrecciated psammite to the south.

On Yell, opposite to Uynarey and Bigga, there are four sheets of felsite which are 2 to 3 m wide and strike approximately north-east (Figure 21). In thin section these are seen to be altered rocks, dominantly composed of randomly oriented feldspar laths about 0.5 x 0.1 mm. In one sheet the feldspars are seen to be albitic (S90845) but in others they appear to be alkali feldspars (S90710). Several contain recognisable biotite flakes (S90843) but interstitial epidote and chlorite are common constituents (S90710). One sheet contains small xenoliths of coarse (grain size 2 to 3mm) granite (S90845) in which the alkali-feldspar has been altered to a poorly developed chessboard albite and the nonfelsic minerals are biotite and epidote-carbonate aggregates.

The felsite sheets are located in the vicinity of the Graven Complex, though, due to the Nesting Fault-splay, they were originally closer to the northern extension of the Complex in Yell Sound than to the main mass in Uynarey and Bigga. They may, therefore, be related to the Graven Complex. The felsite sheet at Copister [HU 472 785] is cut by a sheet of plagioclase-phyric kersantite (S90711) which has been allocated, above, to the lamprophyres. If the felsites are related to the Graven Complex, they are an early phase, since the lamprophyres are cut by the main Complex.

Minerals

Microprobe analyses of the green hornblendes from two of the analysed samples of Graven granodiorite from out-with Yell revealed the following types: magnesio-hornblende and tremolitic hornblende (Table 26). The magnesio-hornblende appears to be a product of early crystallisation and the aluminium-in-hornblende geobarometer (Johnson and Rutherford, 1989) indicates a pressure of crystallisation of 3.2 kb, significantly lower than that found for the earlier emplaced lamprophyres of very similar composition. The plagioclase in these rocks is oligoclase (averaging An₂₃ for cores and An₁₇ for rims) and the biotite compositions are plotted in (Figure 7)a.

Geochemistry

(Table 22) contains the average of eight analyses of Graven granodiorite taken from various places outwith the Yell Sheet. This average composition of the granodiorite is very similar to that of the lamprophyres and tonalites (Table 22) and like them plots in the talc alkaline fields of the A-F-M and FeO0*/MgO-SiO₂ discrimination plots (Figure 19). However, like the kersantite it falls in the 'high K' field of the K₂0-SiO₂ plot. The most notable feature of the granodiorite composition is the high Sr content. (Table 25) contains some analyses for rare earth elements in the granodiorite.

Interpretation

The Graven Complex was emplaced about 400 Ma (K-Ar dating - Miller and Flinn, 1966), it is a post-tectonic intrusion postdating even the lamprophyres and it is truncated to the west by the Walls Boundary Fault. It is associated with hornblendic diorites and more basic rocks, and it has high Na₂O, Sr and Ba content. All these features are also displayed by the Brae Complex and the AithSpiggie Complex which occur to the south, in the Central Shetland and South Shetland Sheets, although the three complexes are very different from each other in their field occurrence and component parts (Flinn, 1985; Gamil, 1991). The lamprophyres and the Graven Complex appear to be closely related because of the similar dates they provide (c. 400 Ma), and their composition and, in particular, the high content of Sr.

The decreasing pressures of crystallisation (aluminium-in-hornblende geobarometer of Johnson and Rutherford, 1989) obtained from red-brown-hornblende lamprophyres, green-hornblende lamprophyres and green hornblende from the Graven granodiorite, respectively, is probably a result of the magma rising as it crystallised.

These three post-tectonic Shetland granitoid complexes have obvious similarities with the Newer Granites of mainland Scotland and in particular with the Argyll Suite of Stephens and Halliday (1984). They contain much hornblende, are associated with appinites, and contain high Na₂O, Sr and Ba contents. The Sr content of some samples of the Shetland rocks is considerably higher than any quoted for the Argyll Suite. However, they were emplaced at about 400 Ma which is younger than the 410–415 Ma intrusion date quoted for the Argyll Suite. The Shetland granitic complexes and the Argyll Suite occur to the east of, and are truncated by, the Great Glen-Walls Boundary Fault.

Chapter 10 Radiometric ages

Contribution by D Roddam, J A Miller and Derek Flinn

Lewisian inliers

A suite of five samples of hornblende gneiss from the Houlland Quarry [HU 506 801] in the Houlland Lewisian inlier have been dated by the K:Ar method (Brook, 1977). The samples were chosen as typical of the hornblende gneisses forming that quarry. They are almost entirely composed of hornblende arid are typical of the hornblende gneisses of all the Lewisian inliers of Yell. The ages obtained have been recalculated with constants appropriate to 1989 and are presented in (Table 27).

A ⁴⁰Ar/³⁶Ar v. %K₂O/³⁶Ar plot of the data was made but no isochron was obtained. It is considered that the rocks have probably suffered varied degrees of overprinting and should be interpreted as giving minimum ages for formation or maximum ages for overprinting. There could of course have been more than one overprinting event.

Potassium metasomatism

Two rock types described above (pp. **), are regarded as having been formed by potassium metasomatism. They are the biotite schists, with and without garnet, derived from garnet-hornblende schists and hornblende schists and the rocks containing interstitial microcline.

In the Central Shetland Sheet, in the vicinity of Brae adjacent to the Graven Complex, many such garnet-biotite schists and biotite schists occur. The biotite from one has been dated by the ⁴⁰Ar/³⁹Ar method at 391 ± 12 Ma (average of two determinations, (Table 27)). This is very close to a K:Ar date of 392 ± 6 Ma (recalculated from 385 ± 6 Ma published in Miller and Flinn, 1966) obtained from an altered diorite containing biotite as the only potassium-containing mineral in the nearby Brae Complex. An early biotite from an unaltered facies of the same diorite (pyroxene-mica diorite) gave an age of 437 Ma (recalculated from 430 Ma quoted by Gill, 1965) and dated by Miller (Table 27). (Table 27) presents the details of all these dates. It is concluded that the Brae Diorite and the metasomatised hornblende schists adjacent to the Graven Complex in both the Yell and the Central Shetland Sheets were metasomatised at the same time, soon after the emplacement of the Graven Complex.

The rocks containing interstitial microcline were metasomatised at a late stage in the metamorphic history of the area, but no evidence was found to link this metasomatic event with the metasomatism of the hornblende schists.

Hascosay Slide

The blastomylonites of the Hascosay Slide, with their strongly tectonised noncataclastic fabrics and general lack of any sign of later cataclasis or alteration, seem to offer suitable material for determining the date of the mylonitisation.

Two samples of blastomylonite were selected for 4oAr/39Ar dating. Hornblende was separated from sample (S92154) (Cambridge and Liverpool number 59967) and biotite from sample (S94728) (Cambridge and Liverpool number 68306). The two mineral separates were irradiated in the Petten Research Reactor, Holland. The standard used to monitor the neutron flux was 133Bi, with a Q value of 0.7504. The samples were analysed on the Cambridge omegatron according to the methods described in Fitch et al. (1969), and the results are presented in (Table 27).

Both samples yield plateau ages over a major portion of the ³⁹Ar release (Figure 22). The hornblende sample (S92154) gives a plateau age of 496 ± 6 Ma for heat steps 7 and 8, representing 80 per cent of total ³⁹Ar release. The biotite sample (S94728) gives a plateau age of 436 ± 7 Ma for steps 6 to 9, representing 62 per cent of total ³⁹Ar release.

As the blastomylonites were formed at temperatures significantly higher than the closure temperatures of hornblende and both rocks were formed at the same time, both samples are likely to record cooling ages postdating the mylonitisation. The large discordance between the two ages suggests that cooling was extremely slow. Assuming typical closure temperatures of 500°C for hornblende and 300°C for biotite, the cooling rate is estimated as approximately 3.3°/Ma. Both dates and cooling rate probably apply for the regional metamorphism throughout Yell.

The results leave the age of the blastomylonitisation undetermined, but not less than 500 Ma, the age that was obtained from hornblende which crystallised during the emplacement of the Shetland Ophiolite Complex (Flinn, et al., 1991).

Chapter 11 Geophysics

Magnetism

That part of the aeromagnetic map of the Shetland area (IGS, 1968b, 1980) covering the area of the Yell Sheet has been redrawn from the original plotting sheets (1:63 360 scale) used in the construction of the aeromagnetic map (Figure 23). Flight lines are shown because it is only where the contours cross the flight lines that the contours are accurately positioned. Ground surveying with a magnetometer on a scale of 1:10 000 in the area of the Central Shetland Sheet revealed a number of places where the contours on the aeromagnetic map had been incorrectly connected between flight lines. However, the magnetic 'topography' of the Yell area is formed of such broad and simple anomalies that there is little chance of such errors in (Figure 23).

A very broad and shallow trough-shaped anomaly extends from north-east Yell to south-west Yell. It connects a large and complex negative anomaly over the Valla Field Block of Unst to the north-east of Yell to a broad weakly negative area associated with the Graven Complex to the south-west of Yell. This negative zone is partly the result of the area being underlain by poorly magnetic rocks, and partly the effect of having highly magnetic rocks to the west (the Lewisian gneiss area of north-west Shetland west of the Walls Boundary Fault) and also to the east (serpentinite of the Shetland Ophiolite in Unst and Fetlar). The negative anomalies in the region of north-east Yell and shown in (Figure 23) are probably, at least in part, due to the influence of a strong but narrow negative ground anomaly along the western edge of the Ophiolite. Since the anomalies overlie the sea, this cannot be checked by ground surveying. Elsewhere in Yell the anomalies are so broad and shallow it is clear that they are not due to bands or areas of magnetic rock reaching the surface, so that there was no purpose in ground magnetic surveying in Yell.

Gravity

The Bouguer anomaly map of Shetland (McQuillin and Brooks, 1967) includes about 230 gravity determinations in Yell. (Figure 24) shows a compilation of these results together with the results of the marine gravity survey of the surrounding sea area (IGS, 1978b). Gravity-determination sites on land are very irregularly distributed, and are shown for the reason given above for including the flight lines on the aeromagnetic map.

For Yell the map shows a rather featureless gently sloping gradient of gravity between the high over Unst to the north-east and the large negative anomaly centred over the Graven Complex to the south-west. A small positive anomaly in the south of Yell at locality [HU 497 813], and the bunching of contours to the south over Ness of Copister, are associated with the outcrop of the hornblende gneiss-es of the Houlland Inlier. The small positive anomaly overlies an outcrop of this gneiss and the closely spaced contours coincide with the Arisdale Fault where it truncates the hornblende gneiss. No other exposed geological feature can be correlated with anomalies on the Bouguer map for Yell.

Chapter 12 Pleistocene and Recent

There is no direct evidence of any geological activity in the general area of the Yell Sheet between the Mesozoic faulting and the last glacial maximum. During the Late Devensian glaciation Shetland was covered by a local ice cap. The ice shed for this ice cap continued from Moss-bank area on the north edge of the Central Shetland Sheet obliquely across the southern half of Yell and extended to the north-north-east into Unst. An ice front lay across the north end of Yell and continued eastwards across the north end of Unst (Flinn, 1983). After the glaciation the sea level rose against the land and continued to rise so that there are no raised beaches in Shetland, only a drowned coastline. The Shetland Islands form a monadnock-like feature rising above the surrounding sea floor so that the drowning has given rise to two distinct types of coastline; an outer coast of cliffs plunging more and less continuously to the flat sea floor at about -120 m, and an inner coast where the sea floods subaerially formed valleys, a few of which may have been glacially modified. Along both types of coastline the post-Pleistocene erosion has been minimal so that the monadnock form of the islands must be of great antiquity, probably dating back to the Mesozoic or earlier (Flinn, 1977a). The coast is rich in beaches and bars of the types characteristic of recently drowned coastlines (Flinn, 1974, 1980). There are many sandy beaches rich in shell material, probably derived from the adjacent intertidal zone. Garnet, derived from the local rocks, is a common consitituent of the beach sands. Man arrived on the islands about 4650 BP (Johansen, 1976). The deposition of the hill peat started about 3400 BP (Keatinge and Dickson, 1979).

Glaciation

During the last glaciation Shetland (including Yell) was covered by a local ice cap. No evidence has been was found in Yell of any earlier glaciation (Flinn, 1964, 1977a). A beach on the south side of Hascosay contains pebbles of flint, tonsbergite and rhomb porphyry. These rocks from London and Tonsberg (south of Oslo) came to Shetland as ballast in the ship Kr^-ago^-6e which was wrecked on Hascosay in 1803. It is an interesting coincidence that a large block of petrographically identical tonsbergite was found in a roadside till-quarry in the south of Shetland [HU 403 159] in the early 1900s (Flinn, 1977a). This erratic is the only evidence, so far found, of Norwegian ice reaching Shetland, probably in an earlier glaciation.

The age of the ice cap is probably Late Devensian. This is indicated by the fact that the oldest dated peat deposit resting on the till from this ice cap gives an age of 12 090 BP (Hoppe, 1974). The till is unweathered, and glaciated surfaces are unweathered, even where they appear never to have been covered with more than the very thinnest mantle of glacial drift. To the south-east of Shetland the position of an old ice margin facing south is marked by a series of giant tunnel-valleys in the sea floor (Flinn, 1967b). The fact that these valleys have not been filled with younger sediments, like those elsewhere in the North Sea cut by earlier glaciations, shows that this was the southern margin of the youngest ice sheet which has existed in the area. The conclusion is inescapable that this is the southern margin of the Shetland ice cap which, therefore, is of Late Devensian age. Recently Cameron et al. (1987) have reported an eastern limit to the Shetland ice cap at about the same distance from Shetland.

The flow pattern of the Shetland ice cap in the Yell area is revealed by glacial striae, lineaments visible on air-photos and, less clearly, by the distribution of erratics (Figure 25).

Glacial striae

About 160 examples of glacial striae were found in Yell, of which about half gave the sense of direction of flow, as well as the azimuth of flow. Poorly formed niches moutonnées were found in several places, including the Ness of Sandwick at locality [HU 441 881] and an area north of Lumbister at locality [HU 482 990]. Glacial lineaments identified on 1:25 000 scale aerial photographs of Yell are best seen in the north-east coastal region, in the area along the coast north of Lumbister, and in the Burravoe area. They are parallel to the local striae, but are on too large a scale and too poorly formed to be detectable features on the ground.

Erratics

Well-defined source rocks in Yell, which provide erratics identifiable beyond their source areas, are the Graven Complex, the Houlland Lewisian inlier, and the Valayre Gneiss. The evidence provided by the erratics serves, at best, to confirm the flow pattern shown in (Figure 25).

All this evidence shows that on the west side of Yell the ice flowed to the west-north-west or north-west, on the south and south-east of Yell it flowed to the east and the south-east, and from central-east Yell it flowed to the east-north-east across Fetlar. In Bluemull Sound area there was a strong stream of ice flowing to the north-northwest. The ice-shed area is defined as that area lacking striae and having striae pointing outwards on either side. It extends from Ulsta in the south of Yell via Mid Yell and Hascosay to Burra Ness and from there north-eastwards towards the Uyeasound area of Unst.

Glacial deposits

Small exposures of till, up to 10 m thick, occur throughout the island in such a way as to indicate that till forms small pockets rather than a widespread deposit (Figure 26). The till is a lodgement till, strongly compacted and matrix dominated. The clasts in the till are usually no more than a few centimetres in greatest dimension and the matrix contains more sand than clay. The till is very similar in appearance and occurrence to the till found elsewhere in Shetland. Over most of the area the exposed bedrock is covered by a thin mantle of 'glacial drift', that is of small stones and 'soil' which is probably residual from a thin and very patchy till layer. Where the ground is now covered by peat or turf it is not possible to determine, without trenching, whether till or glacial drift rests on the bedrock.

In the extreme north of Yell, arid especially in the north-west of the island, there is an absence of striae and till pockets are also absent. Instead some pockets of head and of crudely layered fluvioglacial gravels occur on the cliff edges, and inland several elongate mounds of the same material lie along a stream forming, possibly, in poorly developed esker [HP 504 006]. On the north coast there are several exposures of coarse poorly sorted sand, which could be part of a glacial dune deposit (Figure 26). In the same area there are several small areas thickly strewn with local erratics up to about a metre in diameter. All these features in the extreme north of Yell are considered to have developed along the ice front.

Glacial topography

The small volume of till present in Yell, indeed in Shetland as a whole, is consistent with the fact that no part of the islands, and especially Yell, is very far from the ice shed. This, together with the rather low topography of the area, limited the amount of glacial erosion which could take place, and therefore, the amount of till that could he generated.

The lakes and the glacial lineaments identified on aerial photographs provide large-scale evidence of glacial erosion. However, the lineaments, although of some length, are otherwise very minor features which, while clear enough on the air photos cannot be detected on the ground, and merely testify to the passage of the ice. They are etched onto hillsides and hill tops which show no other sign of being glacially sculptured. No major landforms, other than enclosed hollows now occupied by lakes, occur which require glacial erosion to explain them, except in the area immediately west of Bluemull Sound where the hills are streamlined parallel to the striae and the lineaments.

The general topography of Yell, including the sea cliff's, is probably much as it was before the Pleistocene. The valley along the trace of the Arisdale Fault and the hill ridge along the Arisdale Quartzite are clearly structurally and lithologically controlled and lie across any conceivable ice-flow direction, whether from a local ice cap or, during earlier glaciations, from Norway. The Whale Firth–Mid Yell Gap [HU 495 915] follows a complex course across the lithological banding and is parallel to no known structural feature. It has been associated with the other east–west gaps, the Quarff Gap (Southern Shetland Sheet), and the Yell Sound Gap between Yell and the Mainland. All are probably attributable to very old drainage lines, perhaps relics from a pre-Tertiary land surface (Flinn, 1977a). They show no signs of being glacially formed.

Glacial drainage

In the north of Yell, and also in two places in the middle of Yell at localities [HU 475 966] and [HU 445 903], subglacial drainage channels are prominent features of the topography (Figure 26). The most prominent of these are drainage channels cut through watersheds and major ridges. Gloup Voe [HP 505 035] is a major drainage channel ((Plate 11)e), which has been cut through a 100 m high watershed, and now allows a stream system south of the watershed to drain northwards. Other prominent examples occur at Gutcher at locality [HU 542 993], and Lumbister at locality [HU 475 965] ((Plate 11)b). In most cases, however, the ridges penetrated are only a few metres high, for example at West Sandwick at locality [HU 443 904] and Cullivoe at locality [HP 533 028]. In some places isolated notches characteristic of englacial streams have been cut in the crests of minor bedrock ridges which intersected their courses within the ice, for example locality [HP 530 030].

Very common on the hillsides of north Yell are sharply incised drainage channels which are at full size near the top of the hill, or even on the watershed itself. One particularly fine example of this is the Omand's Dale, south of Gloup Voe, which at its commencement looks like an overflow channel on the top of a major dam (locality [HP 504 016]). Typical channels of this type are about 3 to 5m wide and have the same depth ((Plate 11)a). All of these channels are presently dry or contain misfit streams or local drainage. None of the streams of Shetland are at the present time capable of eroding fresh or unshattered bedrock; many indeed flow over beds of peat or through tunnels of peat where the peat has grown faster than the stream can remove it. Thus, there is no question of any of these drainage channels having been formed under the present climatic conditions.

In north Yell, [HP 517 027] two lochs (Kussa Waters) lie just to the south of the watershed cut by the Gloup Voe drainage channel (Figure 26). They are peat dammed along their south-west banks. A raised shoreline marked by the edge of the blanket peat shows that they were once a single loch overflowing to the north through a notch cut in the watershed. A discontinuous subglacial channel lies between the northern exit and the sea to the north-east, and may indicate an original line of overflow to the north when the lochs were clammed to the south by ice. The lochs now drain to the south through the peat dam, in channels which were probably artificially cut. Water draining out of lochs in Shetland is generally unable to erode the peat flooring the exit channels, but channels are often artifically cut in the peat to drain, or partially drain, lochs in order to increase the area available for grazing. It is noticeable that the line of the peat dam coincides with a line on the Thematic Mapper satellite image and is parallel to the watershed. On the image, the line marks a distinct change in the fine pattern of the topography (Figure 26), the topography being smoother to the south than to the north of the line. Kussa Waters are, therefore, interpreted as an old ice-dammed lake at the ice front, now preserved as a lake by the preferential growth of peat on moraine along the line of the old ice dam.

The association of sub- and englacial channels, fluvioglacial deposits, and the lack of both till and striae in the north-west of Yell, and in the north of Unst (Flinn, 1983), probably means that the margin of the most recent Shetland ice cap lay across the north-west corner of Yell and the north of Unst. Such features are absent from the rest of Shetland.

Loch Lomond Stadial

No evidence of a recurrence of glaciation at this time has been found in Yell, although evidence is preserved to the south in the area of the Central Shetland Sheet (Institute of Geological Sciences, 1982).

Coastline

The coastline of Shetland results from the drowning by the sea of the monadnock-like Shetland archipelago (Flinn, 1977a). The islands are in consequence ringed by an outer coast of cliffs of crystalline and Old Red Sandstone rocks varying in height above sea level from a few metres to more than 400 m (100m in Yell) and plunging to a depth of about 120 m beneath the sea, with little sign of any erosion platforms. The outer coast faces the open sea and experiences no protection from or amelioration of the erosive effects of the ocean swell. The course of the outer coast is little influenced by the topography it truncates. In places the cliff edge passes in a short distance from the summit of a truncated hill to the water's edge, and even beneath the sea, before climbing again to the summit of the next hill. It often truncates stream-occupied valleys high above sea level. The north and north-west coasts of Yell form part of this outer ring of cliffs which form the outer coast of Shetland (Figure 27).

As a result of rising sea level, valleys between the hills have been drowned to form an inner coast which is protected from erosion by the swell from the open ocean. The resultant drowned valleys are locally called voes, firths and wicks. In Yell good examples are provided by Gloup Voe, Basta Voe, Whale Firth, Mid Yell Voe, Hamna Voe and Burra Voe. Within the voes the sea has little erosive power and is, at best, only able to erode the peat and the glacial drift overlying the bedrock, together with weathered and/or shattered bedrock. Thus, within the voes the cliffs are no higher than the thickness of these layers, so that the hillsides pass more or less smoothly beneath the sea interrupted by cliffs no more than a few metres high ((Plate 11)e) and in the inner recesses of the voes by no cliffs at all, not even of peat where it has been drowned instead of eroded. In Yell an example of the latter can be seen at Salt Ness at the head of Hamna Voe [HU 485 805].

Along the shores of the inner coast occur continuous wave-swept seawards-sloping platforms, which narrow to a metre or less in width at high tide. They are riot proper wave-cut erosional platforms but often merely the ice-scoured bedrock surface swept clean of peat and glacial drift; occasionally glacial striae are still preserved below the high tide mark. Alternatively, the sloping platforms may be covered by reworked glacial drift, occasionally by pebble beaches, and rarely by sandy beaches. Along the outer coast beaches are extremely rare and occur only where the cliffs are very heavily shattered, leading to rapid erosion locally.

Between the inner and the outer coasts, the coastline has some of the characteristics of both. It is afforded some protection from the ocean swell by adjacent coastlines, but underwater it does not plunge directly below the level at which the ocean swell can act, so that wave action has a different effect on this coast to that on the other two. This is the intermediate coast (Figure 27).

It is along the cliffs of the intermediate coast that most of the 'geos' that are so characteristic of the Shetland coastline occur. Geos are very narrow and deep inlets eroded into the cliffs along faults, shatter zones and prominent joints. Geos do not occur as frequently in Yell as in some other parts of Shetland, but one of the very best formed geos in Shetland occurs in the entrance to Whale Firth ( [HU 463 956]; (Plate 11)d), eroded along a close-spaced set of prominent joints. The intermediate coast is an alternation of cliffs and beaches. Most beaches occur on the intermediate coast, where valleys reach sea level or the rocks forming the cliffs are sufficiently smashed for an erosional bench to have been formed. Such beaches are composed of rounded shingle or sand, or a mixture of both.

Shetland has been submerging since the last glacial maximum. The tombolos, mid-bay bars ((Plate 11)e), bay-head bars, bay-mouth bars, looped bars and barriers which occur so frequently around the coasts, including those of Yell, (Figure 27) are characteristic of drowning coastlines (Flinn, 1974,1980). Especially worthy of notice in Yell is the double tombolo at Ness of Sound ( [HU 435 825], (Plate 1)), the barrier or bay-mouth bar at Wick of Copister ( [HU 483 785]; (Plate 11)c) and Gutcher [HU 548 994], mid-bay bars in Gloup Voe, Basta Voe and especially Southladie Voe [HU 447 877], and a mid-bay bar that has dammed a voe to form a freshwater lake behind it in the upper reaches of the voe at Vatsetter [HU 535 896]. Further evidence of drowning is provided by the occurrence of submarine peat, not only in the head of voes too sheltered for the sea to erode peat, but even on exposed beaches. In south-west Yell a pebble barrier along the coast is being driven up the hillside in front of the rising sea. Pebbles thrown up by the sea in advance of the barrier armour the peat against erosion. The peat, thus protected, remains preserved beneath the pebble barrier as it is rolled past by the sea and can be found uneroded at low tide where the pebble barrier has moved on ( [HU 467 784]; (Plate 11)f). Such peat has been found elsewhere in Shetland down to a depth of 9 m below sea level (Hoppe, 1965).

Interpretation

The inner coast, the beaches, the geos and some of the cliffs along the inner coast have all been created as a result of rising sea level since the deglaciation. However, the outer coast is a very different feature. The erosion surface at the foot of the cliffs of the outer coast is a concave-upwards surface plunging steeply to the sea floor at about 120 m below sea level, well below the level at which rock can be eroded by the sea at the present time. If the cliffs have retreated significantly since the sea started to rise, the material falling from them must have come to rest far below the present sea surface, where it would have been safe from further erosion. There is no sign of any significant accumulation of such detritus underwater at the foot of the cliffs, even those truncating hills rising several hundred metres above present sea level (Flinn, 1964). Therefore, these cliffs can have retreated very little during the current rise of sea level which has merely served to 'freshen-up' the cliff face. The glacial and subsequent subaerial erosion may have served to remove detritus from a previous cycle of cliff retreat before a rising sea, but several such cycles of rising and falling sea level, with or without glaciation, are not sufficient to account for the creation of the outer coast in its present form. It seems likely that the cliffs of the outer coast of Shetland have been formed by repeated rise and fall of sea level over a very long time predating the Pleistocene. Furthermore, the occurrence of basins of Mesozoic sediments at the foot of these cliffs, for example to the west of Fitful Head in the south of Shetland, may indicate that the monadnock-like form of the archipelago may have originated, whether by marine or subaerial erosion, prior to the Mesozoic and that the cliffs at present forming the outer coast are the result of cliff retreat at intervals since then (Flinn, 1977a).

Beaches

A reconnaissance survey of Shetland's beaches in 1974 revealed the presence in Yell of 21 sandy beaches, some of which contained gravel as well as sand (Flinn, 1974). Also recorded in Yell were nine mid-bay bars, ten barrier bars separating freshwater lochs from the sea, two looped bars, one baymouth bar, one double tombolo and three single tombolos which were mostly composed of gravel. Major areas of blown sand occur at West Sandwick, Breakon and Gossaburgh. Mather and Smith (1974) published an account of Shetland beaches in which they included three sand beaches from Yell and mentioned in passing four others. Their accounts are largely geographical and include little information about the beach sands.

In a careful survey of Yell beaches Garnett-Frizelle (1979) recorded 22 sand and gravel and 13 sand beaches. He also recorded the presence of four sand and gravel tombolos and one sand barrier.

The sandy beaches vary in size from small patches of coarse shell material accumulating in cracks in the rocks in the intertidal zone to similar sized patches of sand in shingle or gravelly beaches and, finally, to sandy beaches hundreds of metres long. Consequently, their numbers vary according to the state of the tide, the time of year, the previous weather conditions, and the observer's standards. The numbers of bars of different type reported vary according to how the observer defines them and in some cases whether the tide is in or out.

Many of the sandy beaches are onshore extensions of sand patches on the sea floor. Others, however, are small, wave-sorted, patches of sand at about half-tide level in much coarser gravel or shingle beaches. There are no cliff-foot beaches on the outer coast of Yell and few sandy, or even shingle, beaches on the inner coast, most such beaches occurring on the intermediate coast.

Garnet-Frizelle (1979) reported an average grain size of 0.37 mm from 56 samples of 'sand' with a standard deviation of 0.66 mm and a range of averages of 0.19 to 0.88 mm. Fifteen of the 36 beeches he sampled are carbonate-free, the remainder contain on average 27.6 per cent by volume of carbonate with a standard deviation of 24.7 per cent and a range of 0 to 90 per cent. The carbonate-rich beaches are almost confined to the north coast of Yell. The heavy-mineral content in 36 samples, one from each beach, ranged from 1.32 to 9.65 per cent volume (and one case of 25 per cent) and averaged 4.2 per cent by volume with a standard deviation of 4 per cent. Garnet, opaque minerals, biotite and hornblende were found in all samples and in many muscovite and chlorite were also present. Garnet is a very noticeable constituent of nearly all of Yell's sandy beaches. Red streaks of garnet form on beaches wherever water flows across them and such streaks may be seen in the dunes where the wind is strong enough to sort the sand when it is dry. Such displays of garnet are also seen on beaches on the west coast of Unst and Fetlar, but otherwise are absent from almost all the other beaches in Shetland. Since the rocks of Yell are far richer in garnet than are the rocks anywhere else in Shetland, it is clear that Yell sands are dominantly of local origin and are probably derived from the till.

The carbonate content of the sands appears to derive from animals living in the intertidal zone and below along the neighbouring coast. Patches of coarse shell sand are often found caught up in cracks in the rocks at the foot of the cliffs, even on the outer coast. These are largely composed of echinoid shell fragments and spines with some other types of shells.

Some large patches of sand on the sea floor extend onto the land to form sandy beaches. For example, the sandy beaches at Breakon in the north of Yell and at West Sandwick in the west appear, from airphoto interpretation and from the notation on charts, to extend down to the surrounding deep sea floor. These underwater sand bodies have been the sources for the sand dunes at those places.

Peat

At some time in the past, peat with a significant thickness, say more than about 10 cm, covered about 97 per cent of Yell, the island having an area of about 208 km². These areas were obtained by planimeter measurement of the 1:50 000 map. At the present time unexploited peat with a thickness of half a metre or more covers about 63 per cent of Yell (Figure 27), the rest having been removed by the inhabitants for use as fuel. The average thickness of the peat probably lies between 1 and 2 m (by inspection of all available sections). The peat is nearly all of the blanket type which, in Orkney and therefore probably in Shetland, started to accumulate at about 3400 BP (Keatinge and Dickson 1979). This is more than a thousand years after man first arrived in Shetland at 4650 ± 80 BP (Johansen, 1976). About 10 cm above the base of the peat, in many places in Yell as in the rest of Shetland, are to be found tree roots and branches up to several centimetres in thickness (Figure 27). In many cases they are of silver birch. Layers of sand and/or gravel, generally up to 10 cm thick, are commonly visible in the peat along the banks of streams high up in the hills. Such occurrences extend down to the shores without change of character, and all appear to be the result of streams overflowing their banks as a result of heavy falls of rain.

Detailed work on the peat in Shetland (outwith Yell) has been carried out recently by Johansen (1985) and used as a basis for a vegetational history of the area, which can be considered to apply to Yell. Late-glacial peat deposited about 12 000 BP does occur in Shetland (Hoppe, 1974) but has not yet been studied, so that the vegetational history established by Johansen starts about 10 000 BP, when the climate ameliorated after the Loch Lomond Stadial and birch appeared in the area. Between then and 8000 BP birch, hazel, poplar and other species formed a shrubby open-vegetation cover. In the wet period from 8000 to 5000 BP shrubs declined and heather increased. From the arrival of man, soon afterwards, the vegetation was dominated by his activities, rather than by the climate.

Chapter 13 Stream sediment geochemistry

In 1978 and 1990–1991 regional geochemical surveys of Shetland were carried out by the Institute of Geological Sciences (1978a) and the British Geological Survey (Buchanan and Dunton, 1992) respectively. The efficacy of this method in a small island like Yell needs to be considered, because many of the streams flow for considerable distances over beds of peat which they are unable to erode so that few of the streams in Shetland are capable of eroding the bedrock other than where it is tectonically shattered or significantly weathered. Any sediment in the stream beds is derived from the thin and patchy mantle of glacial drift covering the island, i.e. from the occasional 'pockets' of till and rare mounds of fluvioglacial material and, only locally directly from the bedrock, where the stream crosses highly shattered rock, for example the Arisdale Burn in its lower reaches where it runs along the Arisdale Fault.

During the last glacial maximum Shetland was covered by its own ice cap, with the ice shed lying obliquely across Yell and extending into the southern part of Unst (Figure 25). Therefore, all the glacial drift occurring in Yell has probably been derived from Yell. However, a stream of ice flowed to the north-north-west along Bluemull Sound and deposited ultrabasic and basic material along the southwest coast of Unst. None was found in Yell, but striae in north-east Yell indicate that the flow extended as far west as the coast of Yell. The conclusion that glacial material in Yell is derived from Yell is supported by the fact that both the rocks of Yell and the beach sands on Yell are rich in garnet, while in most of the rest of Shetland the bed rocks are relatively poor in, or devoid, of garnet and the beach sands are also poor in garnet. Even the beaches seem to derive most of their noncarbonate component from till (Chapter 12). Therefore, the material in the samples taken from streams in Yell for analysis for the geochemical surveys is probably derived from Yell, but not necessarily from the immediate locality of the sample.

Geochemical survey of Yell

D M A Flight and J A Plant

The effect of glacial drift on stream sediment geochemical patterns in Scotland has been studied previously. In the heavily glaciated and drift covered terrain of the East Grampians, for example, it has been shown that there is relatively little glacial smearing of the regional geochemical patterns (BGS, 1992). It is evident from the Yell geochemical data that stream sediment material is locally derived; the closest land masses to Yell are underlain by lithologies giving rise to particularly distinctive geochemical signatures, such as the ultrabasic rocks of Unst, of which no evidence can be seen on Yell. The Cr enrichment located on the north-eastern margin of the island is thought to be related to ultrabasic material in the Hascosay Slide rather than glacial smearing. In addition, known outcrops of Lewisian gneiss on Yell are clearly and accurately delineated by the stream sediment geochemistry.

The Shetland Islands have been the subject of two regional geochemical surveys by BGS, the first in 1970 (IGS, 1978a) and the second in 1991 (Buchanan and Dunton, 1992). The 1991 resurvey benefitted from improved analytical and data-processing techniques; it is this newer dataset which will be used here when considering the geochemistry of Yell. Some 230 stream sediment samples (-150pm fraction) were collected from the island at an average density of approximately 1 per 1 km². These samples have been analysed for 32 elements and the data compiled as coloured digital images. The methods used in sampling, analysis and data processing are described in detail elsewhere (Plant and Moore, 1979; BGS, 1991).

The geochemical signature of the area underlain by Yell Sound Division rocks is similar to that of the Moine sequence in mainland Scotland. The elements RI), U, Be, K, Sr, Li and B have distinctive characteristics over the Moine and have been used to differentiate Moine from Lcwisian and Dalradian sequences (Johnstone et al., 1979; Plant et al., 1984). Particularly similar to the Scottish Moine are Li (mean = 30 ppm) and B (mean = 9 ppm) levels. Both are uniformly low over Yell, with the exception of an isolated coincident anomaly in the north-east of the island, probably derived from a pelitic horizon in the Cullivoe Lens, and a minor Li enrichment associated with Rb and K₂O over a semipelitic facies towards the south-east of Yell. The Dalradian of Scotland and of Shetland are characterised by consistently higher average values of Li and B than the Moine, thought to be indicative of the primary sedimentary characteristics of the two units.

The concentrations of Rb, U, Be, K and Sr over the Yell Sound Division also show similar mean values to those of the Moine. (Figure 28) (modified from Plant et al., 1984) shows the signature of the Yell Sound Division compared with that of the Scottish Moine and Dalradian sequences. Whilst the similarities mentioned above are evident it is also clear that the first row of transition metals, for example Co, Zn, Fe, have considerably higher values in the Yell Sound Division than typically ascribed to the Scottish Moine. However, the same relative relationship of higher metal concentrations in the Dalradian compared with the Moine-like Yell Sound Division is strongly mirrored in Shetland. The variation between the Shetland and the Scottish datasets for these elements may be lithological, or it may be the result of enrichment by surface chemical processes such as co-precipitation which may be more significant on Yell than on the Scottish mainland. In southern Yell minor Cu enrichment is associated with enhanced levels of Ni, V and Fe related to the Houlland Lewisian Inlier.

The detection of particulate gold and pyrite in panned heavy-mineral concentrates about 2 km north-west of Gutcher at locality [HP 5297 0072] and locality [HP 5283 0042] is probably the most important finding of the survey over Yell. The gold observations correlate with a suite of arsenic anomalies – up to 225 ppm As over the northern section of the Hascosay Slide. The Au and As anomalies may be indicative of epithermal or lode-gold mineralisation and merit further investigation.

On Yell the distribution of the other elements studied is fairly uniform and typically at or below background levels for Shetland, the most notable exceptions are the markedly higher levels of Cr, V, Ni, MgO and Cu towards the southern end of the island and the Cr anomaly in the north-east referred to above. These anomalies overlie closely the suggested outcrop of the Houlland Lewisian Inlier and are consistent with the more basic nature of the hornblende gneisses than the surrounding psammites and semipelites of the Yell Sound Division. Coincident Ba and Zn anomalies in north-west Yell may indicate weak mineralisation, while anomalous Ba towards the south-west may suggest the presence of barite veining associated with the Graven Complex

Chapter 14 Economic geology

Minerals

The finding of particulate gold with associated As anomalies, reported by Flight and Plant in Buchanan and Dunton (1992), is described in the Stream Sediment Geochemistry chapter (p.00).

Coarsely crystalline galena occurs in a vein up several centimetres wide cutting a quartz-feldspar pegmatite at Houlland at locality [HU 503 798].

A purple, very fine-grained, micaceous mineral forming a mass several tens of centimetres across in a quartz-feldspar pegmatite to the north of West Sandwick [HU 441 902]; it could possibly be a lithium-containing mica.

Albite and microcline in very large quantities in pegmatites throughout Yell. Both albite-quartz pegmatites and microcline-rich pegmatites occur especially profusely, and often in giant masses, in the cliffs from West Sandwick to the north-west extremity of Yell (Chapter 9).

Quarries

A quarry at Houlland [HU 506 801] in the hornblende gneisses of the Houlland Lewisian Inlier was worked for some time for road stone. It was closed down about the time the car ferries were introduced, partly because of complaints from local inhabitants about dust, and also as an economy allowing greater use to be made of Scord Quarry at Scalloway on the Mainland. Recently, a quarry in shattered microcline orthogneiss at Gutcher [HU 549 996] has been worked for road stone.

Sand

The blown-sand area at West Sandwick was exploited for some 15 years prior to 1991 when the quarry was closed and restored. The sand was used in the building of the Sullom Voc Oil Terminal, and later for building purposes on Mainland. Exploitation ceased when the resource was almost exhausted, the dune barrier at the back of the beach having been protected from extraction by the planning authorities. Exploitation of the area has resulted in the destruction of a well-developed local flora, but a grass cover has been restored over the former quarry site.

The sand in the dunes at Breakon is cleaner than much of that which was exploited at West Sandwick and it is richer in carbonate. It has been licensed for use as an agricultural fertiliser (Mather and Smith, 1974), but due to the policy of the owner, has never been exploited. The area is important for recreation, archaeology and botany so that resistance to its exploitation would be very strong.

The beach at Gossahurgh is composed of a 'silver' sand, but the blown-sand area is small and serves as a recreation area.

Peat

Peat was worked from time immemorial by the local people as a source of fuel. Recently there has been a rapid increase in commercial production of peat for fuel for local use. At Setter (west of Mid Yell) a tractor is used to tow a device beneath the surface of the peat. The device extrudes a 'rod' of peat onto the surface where it is left to dry. The tractor wheels in time kill the heather leaving the surface of the peat layer unprotected, exposed to erosion and eventually impassable by the tractor. In another operation at Gutcher, the heather layer is first stripped off the peat, the peat is then systematically processed so that afterwards the ground can be restored.

Provided use of the peat continues at the present rate, and wasteful methods of exploitation are stopped, this resource is virtually inexhaustible.

References

Most of the references listed below arc held in the Library of the British Geological Survey at Keyworth, Nottingham. Copies of the references can be purchased subject to the current copyright legislation.

ALLEN, J K. 1930,1931,1932. In Summary of progress for 192 1930, 1931. GEOLOGICAL. SURVEY OF GREAT BRITAIN. (London: Her Majesty's Stationery Office.)

ASHWORTH, J R. 1985. Introduction. 1–35 in Migmatites. ASHWORTH, J R (editor). (Glasgow: Blackie.)

ASHWORTH, J R. and EVIRGEN, M M. 1985. Plagioclase relations in pelites, central Menderes Massif, Turkey. I. The peristerite gap with coexisting kyanite. Journal of Metamorphic Geology, Vol. 3, 207–218.

BARR, D, HOLDSWORTH, R E, and ROBERTS, A M. 1986. Caledonian ductile thrusting in a Precambrian metamorphic complex: The Moine of northwestern Scotland. Bulletin of the Geological Society of America, Vol. 97,754–764.

BRITISH GEOLOGICAL SURVEY. 1991. Regional geochemistry of the East Grampians area. (Keyworth, Nottingham: British Geological Survey.)

BRITISH GEOLOGICAL SURVEY. 11993. Yell and some adjacent islands, Shetland. Sheet 130. Solid and drift. 1:50 000. (Southampton: Ordnance Survey for British Geological Survey.)

BROOK, M. 1977. K:Ar age determinations on samples from Shetland. Isotope Geology Unit, IGS, Report No. 77/11.

BROWN, M. 1973. The definition of metatexis, diatexis and migmatite. Proceedings of the Geologists' Association, Vol. 84, 371–382.

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.

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.

CHAPPELL, B W, and WHITE, A J R. 1974. Two contrasting granite types. Pacific Geology, Vol. 8,173–174.

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.

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

DACHS, E. 1990. Geothermobarometry in metasediments of the southern Grossvcncdiger area (Taucrn Window, Austria). Journal of Metamorphic Geology, Vol. 8,217–230.

FERRY, J M, and SPEAR, F S. 1978. Experimental calibration of the partitioning of Fe and Mg between biotite and garnet. Contributions to Mineralogy and Petrology, Vol. 66, 113–117.

FITCH F J, MILLER, J A, and MITCIIELL, J G. 1969. A new approach to radio-isotopic dating in orogenic belts. In Time and place in orogeny. KENT, P F, SATTERTHWAITE, G E, and SPENSER, A M (editors). Special Publication of the Geological Society of London, Vol. 3.

FLINN, D. 1954. On the time relations between regional metamorphism and permeation in Delting, Shetland. Quarterly Journal of the Geological Society of London, Vol. 110, 177–202.

FLINN, D. 1958. The nappe structure of north-cast Shetland. Quarterly Journal of the Geological Society of London, Vol. 114, 120–136.

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

FLINN, D. 1967. The metamorphic rocks of the south Mainland of Shetland. Geological Journal, Vol. 5,251–290.

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

FLINN, D. 1969. Grain contacts in crystalline rocks. Lithos, Vol. 3, 361–370.

FLINN, D. 1974. The coastline of Shetland. 13–23 in The natural environment of Shetland. GOODIER R (editor). (Nature Conservancy Council.)

FLINN, D. 1977a. The erosion history of Shetland: a review. Proceedings of the Geologists' Association, Vol. 88, No. 3,129–149.

FLINN, D. 1977b. Transcurrent faults and associated cataclasis in Shetland. Journal of the Geological Society of London, Vol. 133, 231–248.

FLINN, D. 1980. Geological history. 31–58 in The natural history of Shetland. BERRY, R J, and JouNsrox, J L (editors). (London: Collins.)

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

FLINN, D. 1985. The Caledonides of Shetland. 1158–1171 in The Caledonide Orogen - Scandinavia and related areas. GEE, D G, and Sir RT, B A (editors). (Chichester: J Wiley and Sons.)

FLINN, D. 1988. The Moine rocks of Shetland. 74–85 in Later Proterozoic stratigraphy of the northern Atlantic regions. WINCHESTER, j A (editor). (Glasgow: Blackie.)

FLINN, D. 1989. Geology of Yell and some adjacent islands, North Shetland. British Geological Survey Technical Report, WA/89/75.

FLINN, D. In preparation. Rock analyses from Yell, Shetland. Open file report, BGS, Edinburgh.

FLINN, D. 1992. The history of the Walls Boundary fault. Journal of the Geological Society of London, Vol. 149,721–726.

FLINN, D. MAY, F, ROBERTS, J L, and TREAGLS, 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. MILLER, .1 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. and PRINGLE, I R. 1976. The age of migmatization in the Dalradian of Shetland. Nature, London, Vol. 259,299–300.

GAMIL, A S. 1991. Petrology and geochemistry of Shetland granites. Unpublished PhD thesis, University of I.iverpool.

GARNETT-FRIZELLE, C J. 1979. The sandy beaches of Yell. Unpublished BSc dissertation, University of Liverpool.

GILL, K R. 1965. The petrology of the Brae Complex, Delting, Shetland. Unpublished PhD dissertation, University of Cambridge.

GRAHAM, C M, and POWELL, R. 1984. A garnet-hornblende geothermometer: calibration, testing, and application to the Pelona Schist, Southern California. Journal of Metamorphic Geology, Vol. 2, 13–31.

HEDDLE, M.F. 1879. The county geognosy and mineralogy of Scotland. Mineralogical Magazine, Vol. 2, 131–133.

HENLEY, K J. 1970. Application of the muscovite-paragonite geothermometer to a staurolite-grade schist from Sulitjelma, north Norway. Mineralogical Magazine, Vol. 37, 693–704.

HIBBERT, S. 1822. A description of the Shetland Islands. (Edinburgh: Constable.)

HODGES, K V, and SPEAR, F S. 1982. Geothermosnetry, geobarometry and the Al₂SiO₅ triple point at Mt Moosilauke, New Hampshire. American Mineralogist, Vol. 67, 1118–1134.

HOINKES, G. 1986. Effect of grossular-content in garnet on the partitioning of Fe and Mg between garnet and biotitc. Contributions to Mineralogy and Petrology, Vol. 92, 393–399.

HOPPE, G. 1965. Submarine peat in the Shetland Islands. Geographical Annals, Vol. 47A, 195–203.

HOPPE, G. 1974. The glacial history of the Shetland Islands. Special Publication of the Institute of British Geographers, No. 7, 197–210.

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. (London: HMSO.)

INSTITUTE OF GEOLOGICAL SCIENCES. 1978a. Regional geochemical Atlas - Shetland. (London: Institute of Geological Sciences.)

INSTITUTE OF GEOLOGICAL SCIENCES. 1978b. Shetland Sheet 60°N - 02°W. Bouguer gravity anomaly map. (Provisional edition). 1:250 000 Series. (London: Institute of Geological Sciences.)

INSTITUTE OF GEOLOGICAL SCIENCES. 1980. Shetland, Sheet 60°N - 02°W. Aeromagnetic anomaly snap. 1:250 000 Series. (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. 1976. The age of the introduction of Plantago lanceolate to the Shetland Islands. Danmarks Geologiske UndersOgelse 1976, 45–48.

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

JOHNSTONE, G S, PLANT, J A, and WATSON, J V. 1979. Regional geochemistry of the Northern Highlands of Scotland. 117–128 in The Caledonides of the British Isles reviewed. HARRIS, A L, HOLLAND, C M AND LEAKE, B E (editors). (Geological Society of London.)

JOHNSON, M C, and RUTHERFORD, M J 1989. Experimental calibration of the aluminum-in-hornblende geobarometer with application to Long Valley caldera (California) volcanic rocks. Geology, Vol. 17, 837–841.

KOHN, M J, and SPEAR, F S. 1990. Two new geoharometers for garnet amphibolites, with applications to southeastern Vermont. American Mineralogist, Vol. 63, 1023–1052.

KEATINGE, T H, and DICKSON, J H. 1979. Mid-Flandrian changes in vegetation in Mainland Orkney. New Phytologist, Vol. 82, 585–612.

LEAKE, B E. 1978. Nomenclature of amphiboles. American Mineralogist, Vol. 63, 1023–1052.

LINDSAY, N G. 1988. Contrasts in Caledonian tectonics of the northern and central Highlands of Scotland. Unpublished PhD thesis, University of Liverpool.

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

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

MEHNERT, K R. 1968. Migmatites and the origin of granitic rocks. (Amsterdam: Elsevier.)

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

MOFFAT, D T. 1987. The serpentinized ultramafics of the Shetland Caledonidcs. (2 volumes). Unpublished PhD thesis, University of Liverpool.

PEACH, C W. 1867. On traces of glacial drift in the Shetland Islands. Proceedings of the Royal Physical Society of Edinburgh, Vol. 3, 385–389.

PEACH, B N, and HORNE, J. 1879. The glaciation of the Shetland Isles. Quarterly Journal of the Geological Society of London, Vol. 35, 778–811.

PEARCE, J A. 1980. Geochemical evidence for the genesis and eruption setting of lavas from Tethyian ophiolites. 261–272 in Ophiolites. PANAMOIOU, A (editor). (Cyprus.)

PEARCE, J A. and CANN, J R. 1973. Tectonic settings of basic volcanic rocks using trace element analysis. Earth and Planetary Science Letters, Vol. 19, 290–300.

PEARCE, J A, HARRIS, N B W, and TINDLE, A G. 1984. Trace element discrimination diagrams for the tectonic interpretation of granitic rocks. Journal of Petrology, Vol. 25, 956–983.

PERCHUK, L L, ARONOVICH, L YA, PODLESSKII, K K, LAVRANT'EVA, I V, GERASIMOV, V Yu, FED'KIN, V V, KITSUL, V I, KARSAKOV, L P, and BERDNIKOV, N V. 1985. Precambrian granulites of the Aldan shield, eastern Siberia, USSR. Journal of Metamorphic Geology, Vol. 3, 265–310.

PERCHUK, L L. and LAVRENT'EVA, I V. 1983. Experimental investigation of exchange equilibrium in the system cordierite-garnet-biotite. 199–239 in Kinetics and equilibrium in mineral reactions. SAXENA, S K (editor). (New York: Springer.)

PETTIJOHN, F J, POTTER, P E, and SIEVER, P. 1973. Sand and sandstone. (New York: Springer.)

PLANT, J A, and MOORE, P J. 1979. Regional geochemical mapping and interpretation in Britain. Philosophical Transactions o f the Royal Society of London, Vol. B288, 95–112.

PLANT, J A, WATSON, J V, and GREEN, P M. 1984. MoineDalradian relationships and their palaeotectonic significance. Philosophical Transactions of the Royal Society of London, Vol. A395, 185–202.

POWELL, R. 1985. Regression diagnostics and robust regression in geothermometer/geobarometer calibration: the garnet-clinopyroxene geothermometer revisited. Journal of Metamorphic Geology, Vol. 3, 231–243.

READ, H H. 1931. The geology of central Sutherland. Memoir of the Geological Survey of Great Britain.

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

ROCK, N M S. 1984. Nature and origin of talc-alkaline lamprophyres: minettes, vogesites, kersantites and spessartites. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 74,193–227.

ROCK, N M S. 1991. Lamprophyres. (Glasgow: Blackie.)

ROCK, N M S. GASGARTH, J W, HENNEY, P A, and SHAND, P. 1988. Late Caledonian dyke swarms of northern Britain: Some tectonic and petrogenetic implications of distribution and chemical variations. Canadian Mineralogist, Vol. 26, 3–22.

RÖPSHOFF, K. 1978. Structures of the Tannas Augen Gneiss Nappe and its relation to the under- and overlying units in the central Scandinavian Caledonides. Sveriges Geologiska Undersokning C739, 1–35.

STEPHENS, W F, and HALLMAN, A N. 1984. Geochemical contrasts between late Caledonian granitoid plutons of north, central and southern Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences, Vol. 75, 259–273.

STRECKEISEN, A L. 1967. Classification and nomenclature of igneous rocks. Neues Jahrbuch fur Mineralogie, Vol. 107, 144–240.

TAYLOR, G K. 1984. Geophysical investigations of a Caledonian ophiolite complex, N.E. Shetland. Unpublished PhD thesis, University of Liverpool.

TAYLOR, S R. 1964. Abundance of chemical elements in the continental crust. Geochimica et Cosmochimica Acta, Vol. 33, 275–286.

VERNON, R H. 1990. K-feldspar augen in felsic gneisses and mylonites – deformed phenocrysts or porphyroblasts? Geologiska foreningens i Stockholm forhandlingar, Vol. 112, 157–167.

VOLL, G. 1961. New work on petrofabrics. Liverpool and Manchester Geological journal, Vol. 2, 503–567.

VOLL, G. 1962. Zur Frage des Stofftransports auf den Korngrenzen metamorpher Gesteine. Geologische Rundschan, Vol. 51, 395–405.

WINCHESTER, J A. 1984. The geochemistry of the Strathconon amphibolites, Northern Scotland. Scottish Journal of Geology, Vol. 20, 37–51.

WINCHESTER, J A., LAMBERT, R St. J, and HOLLAND, J G. 1981. Geochemistry of the western part of the Moinian assemblage. Scottish Journal of Geology, Vol. 17, 281–294.

Figures, plates and tables

(Figure 1) Physiographic and locality map of Yell.

(Figure 2) The continuation of geological units in Yell into nearby 1:63 360 sheets.

(Figure 3) Structural-stratigraphical succession of Yell.

(Figure 4) Distribution of talc-silicate rocks and of quartzite.

(Figure 5) Distribution of schist bands.

(Figure 6) Modal compositions of gneissified and ungneissified metasedimentary rocks. a. Modal quartz-feldspar-mica b.Modal quartz-muscovite-biotite c. Modal feldspar-muscovite-biotite d. Modal muscovite v. mica e. Metasediments; Pettijohn et al. (1973) discrimination plot f. Paragneisses and para-semigneisses; Pettijohn et al. (1973) discrimination plot.

(Figure 7) Compositions of biotite and garnet from various Yell rocks—a. Biotite. b. Garnet.

(Figure 8) Distribution of aluminosilicate minerals in metasedimentary rocks and paragneisses.

(Figure 9) Geothermometry and geobarometry of various rocks from Yell. T2 temperature in °C after Hodges and Spear 1982. P1 pressure in kb after Dachs 1990. P2 pressure in kb after Dachs 1990

(Figure 10) Distribution of hornblende schists, ophitic metadolerites and the Gossaburgh metadiamictite.

(Figure 11) Compositions of early basic intrusions. a. Ti-Zr-Y. After Pearce and Cann, 1973 b. Ti-Zr-Sr. After Pearce and Cann, 1973 c. Cr v. Y. After Pearce, 1980 d Pyroxenes from ophitic metadolerites.

(Figure 12) Distribution of megacryst microcline augen gneisses and orthogneisses.

(Figure 13) Distribution of rocks containing interstitial microcline and microcline-bearing gneisses.

(Figure 14) Compositions of paragneisses.a. A1₂O₃ v. K₂O b. Modal microporphyroblast v. total feldspar c. Modal quartz-feldspar-mica: semigneiss d. Modal quartz-feldspar-mica: gneiss e. Modal feldspar-muscovite-biotite: semigneiss f. Modal feldspar-muscovite-biotite: gneiss.

(Figure 15) Distribution of various paragneisses and other gneisses.

(Figure 16) Lineation, cleavage (S_B), folds and faults.

(Figure 17) Compositions of pyroxenes in the Hascosay Slide and the Gossaburgh metadiamictite.

(Figure 18) Distribution of tonalite veins and sheets.

(Figure 19) Compositions of igneous and meta-igneous rocks. a. Rb v. Y+ Nb. After Pearce et al., 1984. syn-COL G—syncollisional granite; WPG—Within-plate granite; VAG Volcanic arc granite; ORG—Ocean ridge granite. b. Na₂O v. K₂O. After Chappell and White 1974. I—Lachlan I-type; S—Lachlan S-type. C. Modal quartz-alkali feldspar-plagioclase. After Streckeisen. 3—Granite; 4—Granodiorite; 5 —Tonalite; 9—Quartz monzodiorite. D. FeO*-K₂O 4-Na₂O-MgO. FeO*—Total iron as FeO.

(Figure 20) Distribution of pegmatites, aplites and pegmatitic-aplites.

(Figure 21) Distribution of lamprophyres and the location of the Graven Complex.

(Figure 22) Step-heating results from samples of blastomylonite.

(Figure 23) Aeromagnetic anomaly map.

(Figure 24) Bouguer gravity anomaly map.

(Figure 25) Glaciation—ice flow directions.

(Figure 26) Location of glacial deposits and fluvioglacial features and deposits.

(Figure 27) Coastal features and distribution of peat.

(Figure 28) Mean element concentrations for the Moine and Dalradian rocks of mainland Scotland compared with the Yell Sound Division. All values have been normalised to average crustal abundance (Taylor, 1964). Modified from Plant et al., 1984.

Plates

(Front cover) Cover photograph: Fugla Geo looking north [HU 476 999]

(Plate 1) Ness of Sound and its double tombolo [HU 452 825].

(Plate 2) Lewisian inlier rocks. A.Epidote- and pyroxene-banded hornblende gneiss, Copister Ness [HU 493 783] Hammer 40 cm long. B. Hornblende gneiss, Houlland Quarry [HU 502 798]. Hammer 40 cm long. C.Quartzofeldspathic gneiss with hornblendite fish, large garnets, and pegmatitic blebs, Ness of Vatsetter [HU 539 901]. Lens cap 7 cm diameter.

(Plate 3) Metasedimentary rocks. A. Granofels, Gerherda [HU 478 995]. Lens cap 6 cm diameter. B. Banded psammite, The Eigg [HU 452 957]. Hammer head 3 cm square. C. Laminated psammite, Bergi Geos [HP 475 030]. Hammer 37 cm long. D. Interbanded quartz-rich psammite and psammite, The Hulk [HU 444 937]. Hammer head 17 cm long. E. Kyanite pseudomorph of chiastolite, Gerherda [HP 475 009]. c.1 cm square.

(Plate 4) Orthogneisses.a. Granodiorite orthogneiss, Graveland Gneiss, The Eigg HU [HU 448 954]. Hammer handle 3.5 cm across at end. B. Granodiorite orthogneiss, Breakon Gneiss, Breakon [HP 524 043]. Pencil 8 mm diameter. C. Granite orthogneiss, Gutcher [HU 538 994]. Hammer head 17 cm long.D. L-tectonite granodiorite orthogneiss, Breakon Gneiss, Inna Ness [HP 528 056]. E. Nebulitic gneiss, Neapaback Skerries [HU 535 786]. Hammer 40 cm long.

(Plate 5) Porphyroblastic gneisses. A. Valayre Gneiss, Burra Ness [HU 507 793]. Lens cap 7 cm diameter. B. Valayre Gneiss with psammitic matrix, Hascosay [HU 546 910]. Hammer 40 cm long. C. Valayre Gneiss with microporphyroblast matrix, Burravoe [HU 521 792]. Hammer 40 cm long. D. Microporphyroblast gneiss, Gossaburgh [HU 531 826]. Hammer handle 3.5 cm across at end.

(Plate 6) Para-semigneisses. a. Recrystallised granofels, Lumbister Burn [HU 472 967]. Coin 26 mm diameter. B. Partially recrystallised interbanded granofels and schist, Vollister [HU 468 947]. Lens cap 7 cm diameter. C. Partially recrystallised schist with granofels, Breakon [HP 521 049]. Lens cap 7 cm diameter. D. Partially recrystallised schist, Saddle of Swarister [HU 531 842]. Compass 6 cm across.

(Plate 7) Paragneisses. A. Recrystallised schist, West Sandwick [HU 438 907]. Hammer head 3 cm at widest. B. Recrystallised semipelite, Gossawater [[HU 479 983]. Lens cap 6 cm diameter. C. Recrystallised granofels—viewed normal to the schistosity, Hons of the Roe [[HU 442 927]. Lens cap 6 cm diameter. D. Recrystallised granofels—viewed normal to the lineation and parallel to the schistosity, Ness of West Sandwick [[HU 444 867]. Lens cap 6 cm diameter. E. Microcline paragneiss—viewed parallel to the lineation, Aywick [[HU 546 885]. Hammer 40 cm long. F. Recrystallised schist, Muckle Vandra Water [[HU 499 883]. Hammer handle 3.5 cm across at end.

(Plate 8) Structures. A. S_B in Mid Yell Schist, Garderhouse, Mid Yell [HU 517 931]. Lens cap 7 cm diameter. B. Laumontite-leonhardite filling shears in fault zone, Ness of West Sandwick [HU 439 872]. Hammer 40 cm long.

(Plate 9) Hascosay Slide rocks. A. Hornblende gneiss, North Sandwick [HU 553 974]. Hammer 40 cm long. B. Blastomylonite, Ness of Cullivoe [HP 548 027]. Lens cap 7 cm diameter. C. Blastomylonite, Ness of Cullivoe [HP 551 022]. Lens cap 7 cm diameter. D. Folded blastomylonite, Ness of Cullivoe [HP 548 028]. Hammer 40 cm long. E. Hornblende gneiss mass being drawn out into blastomylonite, Migga Ness [HP 539 053]. Hammer 40 cm long. F. Hornblende gneiss mass wrapped in interbanded aplitic and hornblendic blastomylonite, Migga Ness [HP 539 053]. Hammer 40 cm long.

(Plate 10) Pegmatites. A. Biotite pegmatite, The Hulk [HU 445 945]. Pencil 8 mm diameter. B. Albite pegmatite, Ern Stack [HU 454 964].

(Plate 11) Geomorphology. A. Glacial drainage channel, Westafirth [HP 503 038]. B. Glacial drainage channel, Lumbister Burn [HU 477 966]. C. Bay barrier, from east, Wick of Copister [HU 483 785]. D. Longa Geo, from east, Whale Firth [HU 463 956]. E. Mid-bay bars in drowned glacial drainage channel, Gloup Voe [HP 505 041]. F. Peat preserved on moderately exposed beach by pebble armouring, Copister [HU 467 784].

Tables

(Table 1) Analyses of biotites from metasedimentary rocks and paragneisses.

(Table 2) Analyses of minerals from metasedimentary schists and paragneisses.

(Table 3) Analyses of garnets from various gneisses.

(Table 4) Garnet-biotite geothermometry of various schists and gneisses.

(Table 5) Garnet-plagioclase-kyanite-quartz geobarometry of Yell Sound Division rocks.

(Table 6) Analyses of metasedimentary rocks.

(Table 7) Analyses of minerals from clasts in the Gossaburgh metadiamictite.

(Table 8) Analyses of early basic and ultrabasic rocks.

(Table 9) Analyses of minerals from garnet-hornblende schists, Yell Sound Division, Central Shetland Sheet.

(Table 10) Geothermometry of hornblende schists.

(Table 11) Geobarometry of hornblende schists.

(Table 12) Analyses of minerals from ophitic metadolerites.

(Table 13a) Analyses of orthogneisses.

(Table 13b) Instrumental neutron activation analyses of Othogneisses.

(Table 14) Variation of rock composition with interstitial microcline content.K₂O

(Table 15a) Analyses of metasedimentary rocks and paragneisses with and without interstitial microcline.

(Table 15b) Modes of metasedimentary rocks and paragneisses with and without interstitial microcline.

(Table 16) Analyses of microporphyroblastic rocks.

(Table 17) Analyses of metasedimentary rocks and paragneisses.

(Table 18) Analyses of minerals from the Hascosay Slide.

(Table 19) Geothermometry of rocks from the Hascosay Slide.

(Table 20) Geobarometry of rocks from the Hascosay Slide.

(Table 21) Analyses of rocks from the Hascosay Slide.

(Table 22) Analyses of late intrusive rocks.

(Table 23) Analyses of minerals from lamprophyres.

(Table 24) Average mode of eight Graven granodiorite samples.

(Table 25) Average analysis of Graven granodiorite INAA and RNAA.

(Table 26) Analyses of minerals from the Graven Complex.

(Table 27) Radiometric ages: K-Ar ages – IGS; Radiometric ages: K-Ar ages - University of Cambridge; Radiometric ages: ⁴⁰Ar/³⁹Ar ages - University of Cambridge; Radiometric ages: ⁴⁰Ar/³⁹Ar step-heating ages.

Tables

(Table 1) Analyses of biotites from metasedimentary rocks and paragneisses

• FeO* & Fe"* - total Fe SD - standard deviation

BGS - BGS specimen number, GR - Grid reference, N - number of analyses

(Table 2) Analyses of minerals from metasedimentary schists and paragneisses

• FeO* & Fe"* - total Fe SD - standard deviation

• BGS - BGS specimen number, GR - Grid reference, N - number of analyses

(Table 3) Analyses of garnets from various gneisses

• FeO* & Fe"* - total Fe SD - standard deviation

BGS - BGS specimen number, GR - Grid reference, N - number of analyses

(Table 4) Garnet-biotite geothermometry of various schists and gneisses

Calculated for 6kb; temperature in °C; increase by 4–8 °C per lkb increase Av = average; Gross average 661°C

• BGS—BGS sample number; GR—Grid reference
• N = number of biotite–garnet-rim temperature estimates averaged Geothermometers:
• 1.Perchuk and Lavrent'eva, 1983
• 2..Hoinkes, 1986
• 3. Ferry and Spear, 1978
• 4. Hodges and Spear, 1982

(Table 5) Garnet-plagioclase-kyanite-quartz geobarometry of Yell Sound Division rocks

• N number of garnet-rim-plagioclase comparisons averaged
• Ti temperature in °C after Perchuk et al. 1985
• P3 pressure in kb after Perchuk et al 1985
• T2 temperature in °C after Hodges and Spear 1982
• Pl pressure in kb after Dachs 1990
• P2 pressure in kb after Dachs 1990

BGS - BGS sample number; GR - Grid reference

(Table 6) Analyses of metasedimentary rocks

N - number of analyses averaged; SD - standard deviation

(Table 7) Analyses of minerals from clasts in the Gossaburgh metadiamictite

• FeO & *Fe"* - total Fe SD - standard deviation

• DM - Moffat 1987; DF - D Flinn
• BGS - BGS sample number; GR - Grid reference; N - number of analyses averaged

(Table 8) Analyses of early basic and ultrabasic rocks

Fe₂O₃^*- total Fe where no FeO quoted; n.d. - not determined

• 1 and 2 recalculated to 100% without LOI (Cr and Ni omitted).
• DM - Moffat, 1987; MPA - M P Atherton
• SN - sample number
• GR - Grid reference

(Table 9) Analyses of minerals from garnet-hornblende schists, Yell Sound Division, Central Shetland Sheet

Fe0* & Fe"* - total Fe; SD - standard deviation

• SN - sample number; GR - Grid reference;
• N - number of analyses averaged
• MPA - M P Atherton; DF - D Flinn

(Table 10) Geothermometry of hornblende schists

• estimated temperature in CC;

sample number; GR—Grid reference

(Table 11) Geobarometry of hornblende schists

• Temperature after Powell, 1985
• Pressure after Kohn & Spear, 1990

• SN—sample number, GR—Grid reference

(Table 12) Analyses of minerals from ophitic metadolerites

• Fe"* & FeO* - total Fe; SD - standard deviation

BGS - BGS sample number; GR - Grid reference; N - number of analyses averaged

(Table 13a) Analyses of orthogneisses

Samples

SN - sample number; GR - Grid reference

(Table 13b) Instrumental neutron activation analyses of Othogneisses

• Values in ppm

• UL—University of Liverpool specimen number; GR—Grid reference Analyst: A S Gamil (1991)

(Table 14) Variation of rock composition with interstitial microcline content

• r = correlation coefficient; sig. = probability of r not significant;± error = 95% confidence interval; y = a + (b± error)x is the line of best fit (not the regression line), where x = volume per cent microcline and y= associated oxides in weight per cent and elements in parts per million; incr. = increment associated with a 30% change in volume of microcline.

(Table 15a) Analyses of metasedimentary rocks and paragneisses with and without interstitial microcline

• 1. Average of 30 microcline-containing meta-sedimentary rocks and paragneisses
• 2. Average of 65 microcline-free meta-sedimentary rocks andparagneisses

• SD - standard deviation

(Table 15b) Modes of metasedimentary rocks and paragneisses with and without interstitial microcline

• 1. Average of 30 microcline-containing meta-sedimentary rocks and paragneisses
• 2. Average of 65 microcline-free meta-sedimentary rocks and paragneisses
• SD - standard deviation

(Table 16) Analyses of microporphyroblastic rocks.

• 1 (S90835) [HU 519 793] microporphyroblastic psammite (semigneiss)
• 2 (S90902) [HU 531 806] microporphyroblastic psammite with microcline (semigneiss)
• 3 Average of two analyses of microporphyroblastic gneisses – (S90732) [HU 516 787] & (S91013) [HU 5245 8265]

(Table 17) Analyses of metasedimentary rocks and paragneisses

• 1 Average of 26 metasedimentary rocks - microcline-free
• 2 Average of 14 para-semigneisses - microcline-free
• 3 Average of 25 paragneisses - microcline-free
• 4 Average of 4 microcline-paragneisses

(Table 18) Analyses of minerals from the Hascosay Slide

FeO* & Fe"* - total Fe

• BGS - BGS sample number; GR - Grid reference; N - number of analyses averaged

(Table 19) Geothermometry of rocks from G-cpx (A) the Hascosay Slide

• estimated temperature in °C
• A. Powell (1985) garnet-clinopyroxene geothermometer
• B. Powell (1985) garnet-hornblende geothermometer
• C. Perchuk et al (1985) garnet-hornblende geothermometer
• D. Perchuk et al (1985) garnet-biotite geothermometer
• E. Ferry & Spear (1978) garnet-biotite geothermometerm
• F. Dachs (1990) garnet-biotite geothermometer

• BGS—BGS sample number; GR—Grid reference

(Table 20) Geobarometry of rocks from the Hascosay Slide

Temperature from Powell, 1985 Pressure from Kohn & Spear, 1990

• BGS—BGS sample number, GR—Grid reference

(Table 21) Analyses of rocks from the Hascosay Slide

• Fe₂O₃* - total Fe; n.d. - not determined

• DM - Moffat, 1987
• BGS - BGS sample number; GR - Grid reference

(Table 22) Analyses of late intrusive rocks

• Fe₂O₃^* - total Fe

• BGS - BGS sample number; GR - Grid reference

(Table 23) Analyses of minerals from lamprophyres

FeO^* & Fe"^* = total Fe

• BGS - BGS sample number; GR - Grid reference; N - number of analyses averaged

(Table 24) Average mode of eight Graven granodiorite samples

• Analyst: A S Gamil (Gamil, 1991)

(Table 25) Average analysis of Graven granodiorite INAA and RNAA

• Mean and standard deviation (SD) of 4 analyses
• INAA - L70504, L75463, L75464, L75469
• RNAA - L75463, L75465, 1.75467, L75468
• INAA - Instrumental Neutron Activation Analysis
• RNAA - Radiochemical Neutron Activation Analysis
• L70504 etc - University of Liverpool specimen numbers
• Analyst A S Gamil (Gamil, 1991)

(Table 26) Analyses of minerals from the Graven Complex

FeO* & Fe"* - total Fe

• UL - University of Liverpool specimen number, GR - Grid reference Analyses by A S Gamil (Gamil, 1991)

(Table 27) Radiometric ages

K-Ar ages - IGS

• Brook, 1977; location [HU 506 801]; hornblende gneiss

Radiometric ages: K-Ar ages - University of Cambridge

Radiometric ages: 40Ar/39Ar ages - University of Cambridge

• 1 & 2 Biotite from biotite-garnet schist, [HU 363 708]

40Ar/39Ar step-heating ages

• Hornblende separated from (S92154); Cambridge number - 59967Hb; = 0.007500

• Biotite separated from (S94728); Cambridge number - 68306Bi; J =0.007279