Content and licensingview original scan buy a printed copy

Geology of Bideford and Lundy Island. Memoir for 1:50 000 sheets 292, with 275, 276, 291 and part of 308 (England and Wales).

By E. A. Edmonds, B. J. Williams, and R. T. Taylor

Bibliographical reference: Edmonds, E. A., Williams, B. J. and Taylor, R. T. 1979. E. A. Geology of Bideford and Lundy Island. Memoir for 1:50 000 geological sheet 292, New Series, with sheets 275, 276, 291 and part of sheet 308 (England and Wales).

Authors and Contributors

• E. A. Edmonds, MSc, B. J. Williams, BSc and R. T. Taylor, BSc, PhD Institute of Geological Sciences, St Just, 30 Pennsylvania Road, Exeter EX4 6BX
• A. J. Burley, MSc, PhD, D. E. Butler, BSc, PhD J. Dangerfield, BSc, J. R. Hawkes, BSc, PhD, R. J. Merriman, BSc and Mrs A. E. Tresham, BSc Institute of Geological Sciences, London

Geological Survey of Great Britain England and Wales. Institute of Geological Sciences. Natural Environment Research Council

London Her Majesty's Stationery Office 1979. © Crown copyright 1979. ISBN 0 11 884120 3.

Typeset for the Institute of Geological Sciences by Art Reprographic (London) Ltd, London. Illustration films by UDO Jenn Ltd, London. Printed in England for Her Majesty's Stationery Office by Ebenezer Baylis and Son Limited, The Trinity Press, Worcester, and London. Dd 696507 K.16

Contributors

• A. J. Burley
• D. E. Butler
• J. Dangerfield
• J. R. Hawkes
• R. J. Merriman
• Mrs A. E. Tresham

Other publications of the Institute dealing with this and adjoining districts

Books

• British Regional Geology
• South-West England, 4th Edition, 1975
• Memoirs
• Geology of the country around Okehampton, 1968
• Geology of the country around Boscastle and Holsworthy, 1972
• Geology of the country around Bude and Bradworthy, 1979
• Geology of the country around Chulmleigh, 1979
• Mineral Dossier
• Ball Clay, 1975

Maps

• 1:625 000
• Sheet 2 Geological
• Sheet 2 Quaternary
• Sheet 2 Aeromagnetic
• 1:1 584 000
• Tectonic map of Great Britain and Northern Ireland
• 1:50 000
• Sheet 292 and others (Bideford and Lundy) 1977
• Sheet 307 and 308 (Bude and Bradworthy) (in press)
• Sheet 309 (Chulmleigh) 1979
• Sheet 310 (Tiverton) 1969 Provisional edition
• Sheet 323 (Holsworthy) 1974
• Sheet 324 (Okehampton) 1969
• 1: 250 000 (Diagram Edition of Aeromagnetic Map)
• Sheet 1 South-Western Approaches
• Sheet 2 Southern England and English Channel

Preface

New Series Geological Sheet 292 takes in Woolacombe, Bideford and Hartland Point. The addition of Morte Point (Sheet 276), Lundy Island (Sheets 275 and 291) and the southernmost coast of Bideford Bay (part of Sheet 308), has allowed the whole north-western corner of Devon to be portrayed on a single map and described in one Memoir. The Upper Devonian and Upper Carboniferous rocks, which are well exposed along an imposing coastline, have attracted geologists for many years. Their stratigraphy and structure are interpreted, and new fossil evidence is related to the classification of the Carboniferous. The Tertiary igneous centre of Lundy is described and the Pleistocene history of the district is considered in a regional context.

Sir Henry T. De la Beche's one-inch-to-one-mile survey of the district was published on Old Series Sheets 26, 27, 28 and 29 between 1835 and 1839. In the latter year his Report on the Geology of Cornwall, Devon and West Somerset was issued. Minor additions were made to the maps later in the 19th century, and the Bideford district was resurveyed by Ussher in 1878, again at one inch to one mile.

The six-inch geological survey was carried out by Mr E. A. Edmonds, Mr B. J. Williams and Dr R. T. Taylor in 1968–70. Mr Edmonds mapped the Devonian and Lower Carboniferous rocks and associated drifts around the Taw–Torridge estuary and to the north, and also Lundy Island. Mr Williams mapped the Upper Carboniferous, Permian and drift deposits of the area between Bideford and Buck's Mills, and Dr Taylor the Upper Carboniferous country around Hartland and Clovelly.

A chapter describing geophysical investigations in the Bideford–Lundy district and in the Bude district farther south was contributed to this Memoir by Dr A. J. Burley. Fossils were collected by Dr D. E. Butler, Dr D. E. White and by the surveyors, and identified by Dr M. A. Calver, Dr W. H. C. Ramsbottom and Dr Butler. The account of the palaeontology of the Devonian and Pilton Shales was contributed by Dr Butler, who gratefully acknowledges helpful discussions with Dr C. H. C. Brunton of the British Museum (Natural History). Petrological work was carried out by Dr J. R. Hawkes and Mr J. Dangerfield on rocks collected by themselves, by Mrs A. E. Tresham and by the surveyors. Mrs Tresham, Mr R. J. Merriman, Mr G. E. Strong, Dr R. Dearnley and Mr P. H. A. Nancarrow also made petrographical and mineralogical contributions.

New chemical analyses were made by Messrs G. A. Sergeant, J. M. Murphy and J. I. Read, with spectrographic work by Messrs R. G. Burns and D. R. Powis, at the Laboratory of the Government Chemist. Mr D. J. Bland made analyses for beryllium in the Lundy Granite, and radiometric age determinations on dyke rocks from Lundy were made by Mrs D. P. F. Darbyshire. Photographs listed in Appendix 3 were taken by Messrs J. Rhodes and C. J. Jeffery. The Memoir was compiled by Mr Edmonds and edited by Mr G. Bisson, District Geologist.

The authors wish to thank Mr H. St. L. Cookes, of Bideford, for his help concerning aspects of the economic geology of the district.

The 1:50 000 geological map which this Memoir describes was published in 1977.

Austin W. Woodland Director, Institute of Geological Sciences, Exhibition Road, South Kensington, London SW7 2DE, 31 May, 1979

List of six-inch maps

The following is a list of six-inch National Grid maps included, wholly or in part, in the 1:50 000 sheet described in this memoir, with the names of the surveyors and dates of survey. The officers are: E. A. Edmonds, R. T. Taylor and B. I. Williams.

Notes

National Grid references are given in square brackets throughout the memoir. Unless otherwise stated all lie within the 100-km square SS (or 21).

Numbers preceded by A refer to photographs in the Institute's collections.

Letters preceding specimen numbers refer to Institute collections as follows: E English sliced rocks MR Museum Reserve collection DX X-ray diffractometer charts X X-ray powder films

The authorship of fossil species is given in the Index.

Geology of Bideford and Lundy Island–summary

Sands and muds of seas, deltas and lakes that existed 300 million years ago and more have been compacted, folded and fractured, uplifted and eroded. A later cover of red desert sediments has been almost wholly removed. The rocks have been torn by wrench faults and invaded by granite now visible on Lundy, and sometime between 100 000 and 200 000 years ago an ice sheet advancing south across the Irish Sea pushed a tongue up the Taw–Torridge estuary to leave a thin mantle of glacial debris.

This is the history of north-west Devon told here. It accounts for the form of the coastline and the hinterland, for the distribution of stone and minerals, for the evolution of countryside of high amenity value, and for the pattern of land use which has led on the one hand to forestry on the acid clays of Hartland and on the other to intensive cropping of the rich silty loams of Braunton Great Field since the early Middle Ages.

Geological sequence

(Geological succession) Geological sequence

The rock formations of the district may be summarised as follows:

Chapter 1 Introduction

Position, climate and physiography

This Memoir describes the Bideford district of north-west Devon, stretching west and north from the Taw–Torridge estuary. The map to which it relates extends north and south beyond the boundaries of the New Series 1:50 000 Bideford (292) Sheet to include Morte Point and the coastline along the southern side of Bideford Bay (Figure 1). Thus the district stretches from Hartland Point in the west to about 5 km east of Bideford. Braunton lies just within its eastern margin and Ilfracombe about 2 km outside it. Also included is the island of Lundy, situated about 18 km north-north-west of Hartland Point and 31 km west of Morte Point.

The climate owes much to the proximity of the sea, which moderates both winter cold and summer heat. Average January temperatures, reduced to sea-level, are around 4°C, those for July around 19°C. Snow seldom lies for long, and air frosts rarely occur before November or after March. Rainfall is about one metre a year around the estuary and in the country to the north, but exceeds this figure farther west in the Hartland area; October to December are the wettest months and June is the driest. Fog is rare but sea mists can roll up rapidly to blanket the coast while the higher ground inland remains clear. Prevailing winds blow from the south-west and much of the district is acutely exposed; tree growth near the coast is stunted and almost entirely leeward, and the long sandy beaches contribute to landward-encroaching sand dunes.

Rugged cliffs of the Hartland area rise 140 m or more and are some of the most spectacular in the west country; and jagged reefs running seaward from them have accounted for the total loss of many ships. Beyond the low ground of the estuary lie the smaller but no less striking crags of Baggy Point and the cliffs of the north coast. Magnificent stretches of beach sand occur near the mouth of the estuary; Saunton Sands stretch for 5 km and are probably the finest in southwest England. Farther north the headlands of Saunton Down and Baggy Point are separated by the sands of Croyde Bay, while Woolacombe Sand extends 3 km from the northern side of Baggy Point to Woolacombe. Sands around the estuary have given rise to the dunes of Northam Burrows, Instow and the 8 km² or so of Braunton Burrows; those farther north have yielded the dunes of Croyde Burrows and Woolacombe Warren.

Away from the coast the highest ground, which rises to over 213 m OD, is that bordering the A39 road south and south-west of Clovelly; the high ground of Bursdon and Welsford moors is open moorland. Inland from Woolacombe, heights between 198 and 213 m occur around the old Mortehoe and Woolacombe Station.

The countryside is deeply dissected, and steep, narrow, commonly wooded valleys abound. Drainage is dominated by the Taw–Torridge estuary, to which flow all but a few coastal streams. Curiously, the River Torridge rises on the southern edge of the district, one of the headwaters being within 2.4 km of the coast near Clovelly. The river follows a long looped course, swinging from a south-easterly through an easterly to a north-westerly flow to re-enter the district south of Landcross and continue through Bideford to join the Taw. Its present lower reaches probably underwent reversal in early Tertiary times, disrupting drainage which flowed south and south-east to the River Exe (Edmonds, 1972). A few deeply-incised streams in the Hartland area flow west to the sea. That flowing through the grounds of Hartland Abbey was once joined by the Titchberry Water from the north and the Milford Water from the south, but eastward recession of the cliff line has destroyed the lower reaches of this system, leaving hanging valleys. Between Hartland Point and Westward Ho! a watershed runs roughly parallel to, and rarely more than 1+ km from, the coast, and north-flowing streams in this area are short and swift. Otherwise, all drainage south of the estuary is to the Torridge. The River Yeo rises within 1+ km of Buck's Mills and flows east to the Torridge at Landcross. North of the estuary the River Caen and its tributaries drain southwards from the area of Mortehoe and Woolacombe Station to enter the Taw through Braunton Pill. West and north of this drainage system, most of the small streams converge westwards to Croyde Bay and Woolacombe and northwards to Lee Bay.

Lundy Island rises to over 140 m OD. Its climate is more equable even than that of the mainland but exposure to Atlantic gales is extreme. Vegetation is sparse, but much more dense and varied on the east side than on the west.

The whole district is one of great scenic attraction. The coastline, which is followed by the south-west peninsula long-distance coastal footpath, and the countryside around Hartland have been declared Areas of Outstanding Natural Beauty. Country at Westward Ho!, Baggy Point, Woolacombe Sand, Morte Point and between Lee Bay and Ilfracombe is owned by the National Trust, as are the island of Lundy and the farms of East Titchberry and Brownsham on the coast east of Hartland Point. A 600-ha nature reserve exists in Braunton Burrows and bird sanctuaries at Chapel Wood (Spreacombe) and on Lundy.

Of the larger wild mammals, red deer are occasionally seen and roe deer are more plentiful. Only sika deer remain on Lundy, of the original three species introduced, but feral goats thrive. The grey seal is present along all the coastline. Many species of sea birds may be seen, but the puffins of Lundy are diminishing in number. Kestrels and buzzards are numerous and a few peregrine falcons breed on the cliffs. Uncommon plants occurring within the district include the Clustered Club Rush and Sand Toadflax of Braunton Burrows, and the unique Lundy Cabbage, Brassica wrightii, discovered in 1936.

Man and industry

Flint implements attest the presence of Stone Age man; many of Mesolithic age have been recovered from the fields around Croyde Hoe near Baggy Point and also from Puts-borough, Saunton Down End and Hartland. Some faience (glazed coloured earthenware) dating from the Bronze Age has also been found in the district, and the round barrows of Bursdon and Welsford moors date from this period. There are a few iron age earthworks, but only one, Clovelly Dykes, of multivallate type. Little evidence remains of Roman occupation, although the name Herculis Prom, for Hartland Point, has survived.

The Saxons may have entered the district along the northern ridgeway across Exmoor. One of the very few open fields of their farming pattern remaining in England is Braunton Great Field, which may have been one of three fields serving the village. Traces of strip cultivation survive near Croyde and North Buckland. On Lundy traces of hut circles have been found, and the Widow's Tenement ruin represents a medieval farm of Long House type; the famous stone kist which yielded the skeletons of an 8 ft 6 in (2.45 m) man and a 7 ft 8 in (2.23 m) woman is of doubtful age.

The economy of the district, in common with that of most of south-west England, is based mainly on agriculture and tourism. Thin soils overlying Devonian and Carboniferous slates, shales and sandstones afford grazing to sheep and beef cattle. Some dairy farming is carried on, especially on the lower ground bordering the estuary. Arable farming is less important, but where it is practised, as on some lighter well-drained soils or on the terrace silts of Braunton Great Field, barley is an important crop.

Several blocks of coniferous woodland, together comprising Hartland Forest, have been established on poorly drained soils developed on Carboniferous shales with sandstones in the headwaters region of the River Torridge. The main plantings have been of Norway Spruce and Sitka Spruce, with some Douglas Fir and Scots Pine. Otherwise there has been little afforestation of the district, and it seems likely that any large new plantations would take the form of additions to Hartland Forest. A few smaller strips and patches of natural woodland elsewhere have been replanted with conifers, but nowhere has this replanting been on a scale sufficiently large to have an appreciable impact on either the regional economy or the landscape.

Sea and river fishing are available to tourists, and salmon are taken in season, but the fishing industry is no longer of major importance.

Small mines were opened for iron at Spreacombe and Buckland, north of Braunton. A few trials were made between Bideford and Hartland and there has been mention of lead, silver, copper and iron. Similarly there are reports of traces of copper and molybdenum in Lundy. The only mining enterprise to survive into recent years was the working at Bideford of a culm seam which extends from Greencliff through Bideford to Alverdiscott and beyond the district to Hawkridge Wood, Chittlehampton. Dug at first for fuel and later for use as the pigment 'Bideford Black', the material has not been worked since 1969.

Sand and gravel are dug from the estuary at low tide, particularly from Broad Sands and the Shrarshook, and in the past pebbles from the Pebble Ridge, Westward Ho!, have been taken for roadmaking. Many quarries have been opened over the years, mainly in Devonian and Carboniferous sandstones, for building stone and roadstone; only two remain active.

Bideford was a busy port in Elizabethan times and now has a population of almost 11 000. Increasing trade brought tobacco from Virginia and Maryland, cod from Newfoundland and wool from Ireland. In the early 18th century there were only three important ports in Devon, Exeter (including Topsham), Plymouth and Bideford. However, ships became larger and Bideford declined; trade with America ceased in the 1770s. During the 19th and early 20th centuries small ships, commonly less than 100 tonnes, maintained a reduced foreign trade, many of them plying to Spain. They carried ball clay, sand and gravel, barley and wheat. Since the Second World War a slightly increased number of ships, individually up to some 500 tonnes, has used Bideford and there has been a striking increase in the number of small pleasure craft on the estuary. Clay was also shipped from Fremington Quay until very recently, but none now leaves the area by sea.

Shipbuilding has been carried on at Appledore (pop. 3000) for centuries, but despite a revival in 1939–45, when small naval craft were built, there is now only one sizeable boatbuilding firm among many small ones. Braunton (pop. 4303) stands on the River Caen, named after Caen in Normandy. The village itself takes its name from St Brannock, who is said to have landed on Saunton Sands from South Wales about AD 550. It was once one of the largest basket-making centres in the country and quite a busy small port, ships of up to 200 tonnes sailing up Braunton Pill to the quay. Coal was brought from South Wales and the Forest of Dean, timber from Cardiff and limestone from Caldy Island, and the ships left carrying iron ore from Spreacombe to Swansea for smelting, sand and gravel and potatoes. Now the quay is used only to unload sand and gravel from the estuary.

Hartland was formerly a thriving market town with its own mayor; in Elizabethan times its population exceeded that of Bideford. Freedom from magnetic disturbance has made the village a suitable location for an Institute of Geological Sciences observatory monitoring the geomagnetic field.

The importation of limestone, mainly in the 18th and 19th centuries, is reflected by numerous limekilns scattered around the coast–as on both banks of the estuary, at Buck's Mills, Clovelly, Mouthmill, Velator, Croyde, Vention and Woolacombe and just east of the district near Ilfracombe. A limekiln near the landing beach on Lundy was destroyed by a landslip in 1954. Local culm and brushwood were used to burn the stone for lime, which was spread to 'sweeten' acid soils, and labourers bought lime-ash from the emptied kilns to make hard beaten floors in their cottages.

Small craft have always been able to land on the sandy beaches, when those carrying limestone simply dumped their cargoes overboard, and few harbours exist outside the estuary. That at Ilfracombe lies just outside the district but has never enjoyed much trade. A small harbour existed at Hartland Quay, although the quay itself is now destroyed, and the construction of a stone pier at Clovelly in the late 16th century changed the whole character of the community from a farming to a fisherman's way of life. Ships frequently seek shelter in Lundy Roads, but landings are made on the beach.

Local buildings are commonly of stone or cob, roofed with slate or thatch. Perhaps the most famous structure in the district is the 226-m-long, 24-arch Bideford Bridge. Much of the original stone bridge, which replaced a wooden one in 1699, remains, and although the exact source of the stone is not known it is probably local. Quaint, close-knit fishing village architecture is well shown by Appledore and Clovelly. Heanton Court is of Norman origin and Tawstock church is a fine stone building dating mainly from the 14th century. Of five Lady Chapels attached to St Brannock's Church, Braunton, there remain only the ruins of St Michael's, of Tudor age; the Peel Tower, on a nearby hill, was erected to commemorate the repeal of the Corn Laws.

Hartland Abbey dates from 1169 and was the only monastic foundation in north Devon. Following the dissolution of the monasteries it passed through various hands and is now the property of the Stucley family. The present house was built in 1769 on the site of the Abbey, of which little now remains. The fine church of St Nectan at Stoke, west of Hartland, was built in 1360, replacing a Saxon church of 1055; its tower, 39 m high, is visible over a wide area.

Communications are_. poor, and have if anything worsened in recent years. Apart from the sea-borne trade mentioned above, there is a regular service to Lundy from Ilfracombe. Transport of limestone, coal and sand upstream to Torrington from a sea lock over 3 km south of Bideford Bridge, and of agricultural produce downstream, was facilitated by construction of the Rolle Canal in 1823–25. Subsequently, and by amicable agreement between Mr Rolle and representatives of the London and South Western Railway Company, the canal was replaced in 1870–72 by a railway which followed much the same route.

The railways had arrived late on the scene, and trains proved ever slow and infrequent. Horse-drawn light-goods traffic working between Barnstaple and Fremington Quay since 1848 was discontinued in 1855 when the railway was extended from Fremington to Bideford. Services from Bideford to Torrington commenced in 1872 and the Ilfracombe line was opened in 1874. The Bideford, Westward Ho! and Appledore railway reached Northam in 1901 and Appledore in 1908, but thereafter survived for only nine years. More recently passenger services through Bideford were withdrawn and much of the clay from Petrockstow now travels by road. Finally the Ilfracombe line closed in 1969. There are now no passenger services in the district, and there may well soon be no railways of any kind.

Main roads have been improved, and a proposed trunk road skirting the south side of Barnstaple will probably continue westwards into the district. Chivenor Airfield was last regularly in civil use when Atlantic Coast Airlines Ltd flew light aircraft to Lundy, up to 1947. Subsequent attempts to provide a civil airline service in 1950 and 1952–55 both failed, and since then the airfield has been used by the Royal Air Force.

Water supplies depend mainly on surface water, drawn from reservoirs or river intakes. Many small local supplies are obtained from wells and boreholes, but prospects of large yields from such sinkings are poor. Any increased demands are therefore likely to be met from reservoirs. These may be small and at high altitude, or larger and in the middle reaches of rivers, and the possibility also exists of an estuarial barrage. Failing a giant scheme for the Severn estuary, which could supply south-west England, the Bristol area and South Wales, a Taw–Torridge barrier would impound a vast quantity of water. Such a barrage, if proved feasible, would, of course, have its own considerable impact on the urban environment which the Bideford–Barnstaple–Braunton triangle seems destined to become. In long-term consideration of land use the choice will probably be between large reservoirs and desalination. Use of the estuary for water storage would benefit tourism by providing water for consumption and recreation. It could, however, have adverse effects, notably on the shipping industry and the ecosystems of the mudflats, and it would be impracticable while present methods of sewage disposal obtain.

The future of the district will probably rest with agriculture and tourism for some years. Forestry may be expected to expand on the poor clayey acid soils overlying Carboniferous shales with sandstones in the high ground east and south of Hartland. There is little prospect of development of any of the metalliferous ores, or of the Bideford culm. However, light industry will continue to stretch out westwards from Barnstaple, Chivenor Airfield may be redeveloped, possible coastal sites for nuclear power stations exist, and the potential is present for considerable development.

Geological history

The generalised geology of the district is shown in (Figure 2). The oldest rocks exposed are the Morte Slates, of Upper Devonian age. They comprise greenish grey slates, generally smooth and locally much veined with quartz, and succeed a shallow-water succession of shales, siltstones, sandstones and rare limestones of the upper Ilfracombe Slates. Land lay a short distance to the north, over what is now South Wales, and the fine sediments which became the Morte Slates were laid down in a shelf sea unruffled by major disturbances.

The northern margin of the basin of deposition was subject to change, migrating slightly to north and south, and the shallow marine environment of the Morte Slates muds gave way to variable near-shore and deltaic conditions during deposition of the Pickwell Down Sandstones, Upcott Slates and Baggy Sandstones. A change from deltas to shallow sea is marked by the Pilton Shales, and the overlying Lower Carboniferous shales formed in generally undisturbed water. The accompanying cherts are assumed to have originated in shallow water (Edmonds and others, 1968), and the associated limestone lenses in shelf lagoons. Subsequent deposition took place in deeper water, generally still hut periodically disturbed by sediment-laden currents coursing down the sides of the basin and along the sea floor.

Under these conditions accumulated the shales and fine-grained turbidite sandstones of the Upper Carboniferous Crackington Formation. Overlying this formation in the Bideford area is a near-shore, shallow-water, cyclic facies of shales, siltstones and sandstones, well exposed in the cliffs south of Westward Ho! These rocks constitute the Bideford Formation. East of the district they pass info Bude Formation strata, which also overlie the Crackington Formation. The Bude Formation is characterised by sandstones which are generally massive and show little internal structure. They probably accumulated in shallow water and have been identified in the present district north and west of Hartland and near Buck's Mills.

Earth movements of the Variscan orogeny crumpled the Palaeozoic sediments along east–west axes into open folds locally disrupted by strike faults. The only representative in the district of the succeeding New Red rocks is a small outlier on the coast at Portledge. However, a sparse scattering of small pebbles possibly derived from conglomerates of this age suggests that such fragmental rocks may have been widespread in the east–west valleys of the arid post-Variscan landscape.

Possibly the district was land during the Jurassic, and any Cretaceous sediments which may have been deposited have been eroded away. The granite and dykes of Lundy Island are of Eocene age; so, probably, is a small patch of gravel at Orleigh Court. By analogy with similar structures elsewhere (Edmonds and others, 1968, 1975) it seems likely that the NW-SE faults, which are very common, moved mainly in Tertiary times during the Alpine orogeny.

In Middle Pleistocene times the Anglian ice sheet advanced southwards. Evidence of its extent is slender, but it was probably the source of the large erratics found on the north Devon coast. Ice of the next (Wolstonian) glaciation deposited boulder clay in the Taw estuary, and the first river terrace and the raised beach of Saunton appear to date from the succeeding interglacial period. Periglacial conditions prevailed in the district during the final (Devensian) glaciation, the ice reaching no farther south than South Wales; much Head was formed at this time.

With the final retreat of the ice rising sea level flooded the river mouths. Forests growing on alluvial soils near the coast were submerged, and one of the many which became partly fossilised may now be seen on occasion at low tide at Westward Ho!

History of research

Some of the earliest work on rock classification in southwest England was that of Conybeare (1814, 1823), who worked at Clovelly. Subsequently De la Beche surveyed the present district and published some notes on the culm at Bideford (1834) before the appearance of his great Report on the geology of Cornwall, Devon and west Somerset (1839). Simultaneous research by Sedgwick and Murchison helped to elucidate the stratigraphy and structure of the Palaeozoic rocks of Devon (1840) and included a study of the raised beach of the Bideford area (1843). Knowledge of the general structure was extended by Etheridge (1867), while Maw (1864) recognised the glacial origin of the clays at Fremington. A series of papers by Hall discussed fossil finds in north Devon (1867, 1876), the geology of Lundy (1868, 1871), the culm seams of the Bideford area (1875) and the fossil forest of Barnstaple Bay (1879a).

By 1878 Ussher's official revision of De la Beche's maps, begun at Wellington in 1870, had reached the Bideford district. His results appeared in a number of regional papers dealing with the Devonian and Carboniferous rocks of south-west England (1881, 1887, 1892, 1900, 1901); the latest of these included a map of the Carboniferous rocks, and Ussher also contributed a coloured geological map to the Devonshire volume of the Victoria History of the Counties of England (1906). Contemporaries of Ussher included Hicks (1891, 1893, 1896), who studied Devonian rocks around Mortehoe and Ilfracombe, Hinde and Fox (1895), who worked on Lower Carboniferous cherts, and Whidborne (1896, 1896–1907), who considered fossils within the age range Pickwell Down Sandstones to Pilton Shales.

Arber (1907, 1911) discussed Upper Carboniferous strata and the coastal geology of north Devon, and Rogers (1907, 1908, 1909, 1910) the Carboniferous fauna and flora. A coloured map accompanying an excursion report by Hamling and Rogers, but apparently owing much to Ussher, has remained the most detailed published geological map of most of north Devon since its appearance in 1910. In that same year Dewey described glacial erratics from the Bideford district, and followed this with a paper about the raised beach (1913); subsequently he wrote the first edition (1935) of the regional handbook 'South-West England'. Dollar's (1942) account of the Lundy complex includes much useful petrological detail and is the most comprehensive description of the geology of the island to date.

The relationship between the Pilton Shales and equivalent strata on the continent was examined by Paul (1937), and the presence within these shales of the Devonian-Carboniferous boundary was confirmed by Goldring (1955a, 1962), who has also published (1971) a detailed study of the Baggy Sandstones. Carboniferous stratigraphy and structure have been investigated by Prentice (1960a, b), on the coast south of Westward Ho! and inland to Barnstaple, and by De Raaf and others (1965), Reading (1965), Burne and Moore (1971) and Freshney and Taylor (1971, 1972). A review of Devonian and Carboniferous palaeontology and stratigraphy in Cornwall and Devon was compiled by House and Selwood (1966), and contemporary ideas on the classification of the Carboniferous were summarised in the Okehampton Memoir (Edmonds and others, 1968) and by Edmonds (1974).

Early work on the drifts of the district was continued by Taylor's (1956) descriptions of erratics. There is now general agreement on the Wolstonian age of the boulder clay, but Mitchell (1960, 1972) and Stephens (1966) placed the raised beach in the preceding interglacial period, while Bowen (1969) and Edmonds (1972) favoured the succeeding one. Mitchell (1968) identified remnants of glacial gravel on Lundy Island and supposed them to be of Wolstonian age.

Research results, ideas and theories relating to the geology of north Devon were summarised within a regional framework by Edmonds and others (1969, 1975). EAE, RTT

Chapter 2 Upper Devonian

General account

A shelf sea extended across north Devon in Givetian–Frasnian times. Land lay to the north, over the present South Wales, and deeper water over what is now Cornwall and south Devon. Southward-draining rivers from the 'Welsh' coastal plain brought muds, silts and sands, which later became the slates, siltstones and sandstones of the Ilfracombe Slates. Limestones formed from coral banks and from the detritus of abundant fossils. Subsequently the waters became shallower, perhaps by the building of a great delta, and the upper Ilfracombe Slates contain much arenaceous material, including some massive sandstones.

The shallow seas of upper Ilfracombe Slates times probably deepened slightly, and the predominantly argillaceous sediments of the Morte Slates accumulated in the undisturbed waters of a shelf sea. This marked the beginning of the Upper Devonian–Lower Carboniferous northward marine transgression into Wales and central England. The Upper Devonian rocks of the present district were mostly laid down not far from the fluctuating northern margin of the basin of deposition; they reflect shallow marine, deltaic and some fluviatile sedimentation, with no deep-water deposits.

Hamling's map of 1910 shows the outcrop of the Ilfracombe Slates extending westwards to beyond Shag Point [SS 487 471], but it is now clear that this is not so and that the north coast of the present district is entirely composed of Morte Slates, of probable Frasnian–Famennian age.

Morte Slates

The Morte Slates comprise smooth greenish grey and purplish grey slates, locally silty and with a few fine-grained sandstones and scattered thin discontinuous limestones. Traces of limestone are most common in the lower part of the formation, and the Morte Slates at Morte Point (Plate 1), and for about 500 m to the east-north-east, are noticeably silty and contain many sandy streaks and beds of siltstone and sandstone. However, these distinctions are not such as can be mapped inland within the present district.

The formation appears to show much folding, including northward overturning, and is cut by a multitude of generally small NW–SE faults. It has accommodated much of the movement of the Variscan orogeny, in contrast to the overlying and more competent Pickwell Down Sandstones. Open folding at Rockham Beach was recorded by De la Beche (1839, p. 49, fig. 2 and pl. 3), but the pattern of the overfolding is best brought out diagrammatically; (Figure 14) suggests northward overturned folds whose limbs dip south-south-westward at around 84° and 50°.

The slates of Lundy Island, with their silty bands and thin limestones, most nearly resemble the lower parts of the Morte Slates. Dollar (1942) noted that although field evidence suggested predominantly slaty rocks, thin sections showed a surprisingly high arenaceous content. He identified sericite-chlorite-slates, arenaceous chlorite-slates, arenaceous biotite-tourmaline-hornfelses, various other sandy and limy rocks and two altered rocks of possibly igneous or pyroclastic type, one of them a chlorite-schist. No indisputable fossils, much less identifiable forms, have been recovered.

The fauna of the Morte Slates is poorly known. Owing in part to the strong cleavage, fossils are difficult to obtain and they are commonly highly distorted. Only the spiriferaceans have provided useful stratigraphic information.

Ussher (1881) considered the Ilfracombe Slates and the Morte Slates to be the lower and upper parts of a single series, and clearly had some difficulty in defining a line between them. Hicks (1896) provided the first descriptions of fossils from the Morte Slates and revived earlier doubts about the orderly succession of the north Devon strata by declaring that the group was the oldest in the region, fault-bounded on both sides, and that it contained a fauna in part Silurian and in no case indicative of an age younger than the Lower Devonian. He recorded Lingula from Lee Bay, Rockham Bay and Morte Point, and from the area between Grunta Beach and Barricane Beach he obtained specimens then identified as Rhynchonella lewisii? and Orthis along with spiriferaceans, bivalves and other remains. Hicks's stratigraphical and palaeontological evidence was contested by Gregory (1897). Evans and Pocock (1912) reported the discovery of Cyrtospirifer verneuili (Plate 2), associated with rhynchonellaceans resembling known Upper Devonian species, from Barricane Beach. This material was loose but considered to be of very local derivation and the evidence suggested that the Morte Slates were of Upper Devonian age and in sequence. During palaeontological investigations in connection with the present survey a similar fauna to that recorded by Evans and Pocock was obtained from in situ sandstone lenses from the north-east side of Barricane Beach (p. 16). Webby (1965) also noted Cyrtospirifer sp.from the lower Morte Slates east of Cold Harbour [ST 013 357] in the Brendon Hills.

Folded, faulted, lacking palaeontological subdivision and without mappable internal lithologies, the thickness of the Morte Slates cannot be estimated with accuracy but they are probably no more than 1500 m thick.

Pickwell Down Sandstones

Overlying the Morte Slates are some 1200 m of current-bedded and ripple-marked purple, red, brown and locally greenish grey sandstones with some red and grey shales (Figure 3). They mark a change from shallow quiescent seas to more disturbed and variable, and in part continental, conditions of near-shore, deltaic, lacustrine and fluviatile sedimentation.

The outcrop of the formation is 2.5 to 3 km wide. At the base is the so-called 'Bittadon Felsite', one or more bands of keratophyric tuff. This has been described (Rogers, 1926) from the beach exposure of Mill Rock [SS 4553 4312] and occurs as fragments some 730 m inland. Bonney (1878) examined the rock at its type locality, east of the present district, and concluded from the nature of exposed if weathered contacts that it was intrusive. However, he noted what he considered an accidental resemblance to streaky ashes of Borrowdale, Cumbria. Bonney described a finely granular, locally banded, matrix studded with broken iron-stained feldspar crystals (orthoclase with some plagioclase) and a few quartz grains. Ussher (1879, 1881) concurred, rather tentatively, with the concept of an intrusive relationship.

The usefulness of the tuff bands as a stratigraphic marker was emphasised by Rogers (1926), who believed that they could be traced throughout the outcrop of Upper Devonian rocks in north Devon. He returned to a suggestion of Etheridge (1867) that the rock was contemporaneous, noted microscopic evidence of flow structure in specimens from Mill Rock, and concluded that volcanic dust and ash had been stratified in water. Rogers's main evidence of such an origin, however, was his discovery of fish remains in the tuff bands interstratified with slates and sandstones at Mill Rock. He obtained fragments of Holonema, Bothriolepis, Holoptychius, Polyplocodus and Coccosteus. He also collected plant stems, probably Archaeopteris, from basal Pickwell Down Sandstones exposed in a quarry [SS 4610 4306] 550 m inland of Mill Rock, and A. hibernica? from the upper surface of the tuff near Fox Hunters Inn [SS 508 419], just east of the present district. No undoubted tuff was identified at Mill Rock during the recent survey.

Higher in the sequence the sandstones and subordinate shales of the formation are unfossiliferous except for scattered fragments of fossil wood. Their strength has prevented the development of tight folding, but some fracturing has occurred and locally these faults are characterised by iron mineralisation (p. 115). The uppermost beds of the Pickwell Down Sandstones are exposed at the southern end of Woolacombe Sand. They show the incoming of more fine-grained sandstones, micaceous siltstones and slates, pale green and grey colouration and the steepening of dips towards a fairly rapid passage into overlying slates. In the vicinity of Black Rock [SS 4505 4134] small fractures in sandstones are lined with quartz and show traces of manganese oxides.

Upcott Slates

Massive thickly bedded and false-bedded sandstones pass upwards into cream, green, grey and purple slates and siltstones with scattered thin fine-grained sandstones. These are the Upcott Slates (Hull, 1880), which crop out eastwards to beyond Dulverton and are best exposed in near-strike section along the northern side of the Baggy Point headland between Napps Cliff and Whiting Hole. They are probably about 250 m thick. As with the Morte Slates, vertical and steep dips are common and once again it seems that an argillaceous formation, in this case lying between the Pickwell Down Sandstones below and the Baggy Sandstones above, has been deformed more readily and therefore more intensively than its neighbours.

Goldring (1971) considered the conditions of deposition to have been a continuation of those of the Pickwell Down Sandstones, and visualised sedimentation in a back-swamp, alluvial environment or shallow fresh-water lakes. His examination of the top of the formation showed it to be a sharp but conformable junction between slates and the basal turbiditic sandstone of the Baggy Sandstones.

Baggy Sandstones

The return of some off-shore but shallow-water sedimentation, interrupted at times by penecontemporaneous erosion, marked a spasmodic renewal of the marine transgression. Some 450 m of sandstones and shales with a few scattered thin limestones form Baggy Point and the greater part of the promontory of which it is the tip (Figure 4). They have been the subject of long and detailed investigations by Goldring (1971). The studies of Austin and others (1970) suggest that the Upper Devonian conodont faunas of the uppermost Baggy Sandstones are younger than the uppermost Wocklumeria faunas of the German type section.

The beds dip at up to 75° in directions around south-south-west. Small-scale folding and crumpling occur locally but there is no evidence of large folds. Groups of sandstones form small headlands and have been named and designated as members by Goldring (1971). Between them lie mixed assemblages of grey shales, siltstones and sandstones. Sole marks occur on the basal sandstone and sporadically elsewhere, and ripple marks are fairly common. Goldring (1971, fig. 2) identified 11 facies, more than one of which is usually present in each member. He considered two of them to be non-marine, one probably non-marine, one brackish water and the rest (some 80 per cent of the formation) marine, and found evidence of organic colonisation of the sediment of three of the facies.

The most southerly, and therefore stratigraphically highest, group of sandstones [SS 4195 4057] at Baggy Point, Goldring's Mackerel Member, is the topmost of a fairly closely spaced sequence of sandstone members. It may be considered as defining the top of the lower, sandier Baggy Sandstones. Indeed there is some case for taking its top as the top of the formation. However, scattered sandstones characteristic of the Baggy Sandstones do crop out within the overlying shales and siltstones on the south side of the promontory. These beds extend upwards to a slumped bed [SS 4244 4015] (Plate 3) overlain by current-bedded sand stones, and these sandstones are taken as the top of the Baggy Sandstones, as suggested by Goldring (1971).

No inland exposure of the slumped bed has been noted, and mapping of the formation boundaries commonly depends upon the presence in rubbly exposures or brash of buff-brown fine-grained to medium-grained, commonly feldspathic and micaceous sandstone. These rocks are mostly resistant to weathering, and the Baggy Sandstones generally form higher ground between the older Upcott Slates to the north and the overlying Pilton Shales to the south.

The presence of 'Cucullaea' and Lingula in the Baggy Sandstones of the coast has long been known, the former generally in sandstones, the latter usually in shales. Ussher (1879, 1881) remarked on inland occurrences, using the term 'Cucullaea grits', and Lingula was collected from the coast by Hamling and Rogers (1910). Plant remains have been found near Croyde Hoe (Arber and Goode, 1915; Rogers, 1926). In his monograph on the Devonian fauna of southern England, Whidborne (1896–1907) preferred Salter's (1863) term Marwood Beds to the 'Baggy Beds' of Ussher and Champernowne (1879), and included strata ranging from the Pickwell Down Sandstones to the Pilton Shales. Of fossils collected during the recent survey cf. Lingula squamiformis [SS 4204 4065] may confirm a record of Whidborne, and cf. Diplocraterion yoyo [SS 4221 4031] recalls the characteristic trace-fossil of Goldring's D. yoyo facies, shales, siltstones and thin sandstones making up the commonest facies of the Baggy Sandstones. The discoveries of a fish plate, Bothriolepis? [SS 4278 4033], and of a tooth nearby [SS 4278 4026], are interesting in view of the single early record of a fish scale from Baggy Point (Phillips, 1841). The quarry from which the tooth was obtained also yielded cf. C. verneuili and the phyllocarid Echinocaris whidbornei (Plate 2); the latter has been recorded from Sloley, to the east of the present district (Whidborne, 1896–1907; Partridge, 1902). EAE

Petrography and mineralogy

Mainland

The Devonian rocks of the area consist of slates, sandstones and siltstones of shallow-water origin, ranging from the Morte Slates to the lower Pilton Shales, which now show evidence of low-grade regional metamorphism.

Samples from each of the main formations were examined by means of thin sections, in order to ascertain to what extent compositional variation accounts for differences in lithology. Point count analysis was confined to arenaceous rocks, while powdered argillaceous material was examined on a Phillips diffractometer using CuKχ radiation, 40 kV, 30 mA, scanning over the range 3°–45° 2θ.

Arenaceous rocks

Siltstones have a maximum grain size of 0.06 mm, and fine-grained sandstones have mean grain sizes between 0.06 and 0.12 mm. In thin section both consist of subangular to sub-rounded quartz grains together with scattered lithic fragments, feldspar and detrital muscovite in a matrix of finely divided quartz, muscovite and chlorite.

The quartz grains commonly show strain extinction and, more rarely, bubble trains and inclusions of minute rutile needles; one or two grains exhibit Boehm deformation lamellae which have clearly developed prior to deposition. Lithic fragments occur in both siltstones and sandstones but are most plentiful in the latter. They include chert, siltstone, greywacke, quartz-sericite-schist and quartz-chlorite-schist, quartzite and acid igneous rock. Two specimens of siltstone contain scattered books of penninite, which possibly represent altered basic or intermediate volcanic glass (E42727); (E46122). Detrital feldspar usually occurs as little-altered, subangular grains of lamellar-twinned sodic plagioclase, but rare grains of microcline are also present in the sandstones. Flakes of detrital muscovite are particularly abundant in some of the sandstones; in one sample they are accompanied by detrital biotite, now heavily chloritised (E9252); (E42715); (E42733)). Secondary overgrowths of fine-grained chlorite can be seen on some muscovite flakes.

The matrix consists of an intergrowth of very fine-grained quartz, muscovite and chlorite and probably represents recrystallised clay-grade material. It is variable in amount, being most abundant in the siltstones and rather sparse in some sandstones ((E9256); (E9257); (E42722), (E42723), (E42724)). The latter commonly possess a silica-cemented interlocking framework in which many of the quartz grains show optically continuous outgrowths. In some siltstones, especially siltstone lenses in the Morte Slates and Pilton Shales ((E42712); (E46122)), a patchy carbonate cement is present, usually consisting of clear sparry calcite, but including in one specimen (E42726) scattered granules of siderite. The carbonate cement is clearly of late origin, and in places cuts tiny quartz veinlets ((E42731);(E46122)). The opaque mineral dispersed throughout these rocks is mainly pyrite, locally weathered to limonite; it is accompanied by much secondary hematite in some sandstones ((E9256); (E9257)). Other accessory minerals include zircon, apatite, tourmaline and ilmenite; they form local concentrations in some siltstones from the Morte Slates and Pilton Shales ((E30442);(E42713)).

Modal analyses of 29 samples from the arenaceous beds are shown in (Table 1). They indicate that the fine-grained sandstones contain more quartz and lithic fragments and less matrix and carbonate cement than do the siltstones. Lithic fragments are particularly abundant in the Pickwell Down Sandstones; up to 30 per cent of the sample from Putsborough Sand (E42724) consists of fragments, including basic igneous rock. The Baggy Sandstones are richest in feldspar and in detrital muscovite and chlorite. Carbonate-rich samples occur mainly in siltstones of the Morte Slates and Pilton Shales, a specimen (E30442) from Fairlinch Quarry containing up to 38 per cent carbonate. The siltstones of the Morte Slates, Baggy Sandstones and Pilton Shales have fairly consistent modal compositions. Of the coarser sandstones, those of the Pickwell Down Sandstones range from protoquartzite to subgreywacke, whereas those of the Baggy Sandstones are protoquartzites (Pettijohn, 1957, p. 291).

Argillaceous rocks

Ten powdered samples of slates and shales from the Devonian and Pilton Shales were examined by X-ray diffractometry ((MR32600), (MR32601), (MR32602), (MR32603), (MR32604), (MR32605), (MR32606), (MR32607), (MR32608), (MR32609), (DX567), (DX568), (DX569), (DX570), (DX571), (DX572), (DX573), (DX574), (DX575), (DX576)). Their localities are listed in (Table 2). All consist of about 90 per cent quartz, muscovite and chlorite, the remainder being mainly albite with minor microline and traces of anatase and brookite. Quartz forms from 25 to 55 per cent of the samples, averaging about 35 per cent. Muscovite is generally more abundant than chlorite. These argillaceous Devonian rocks differ from most of the Upper Carboniferous shales of the district in the presence of traces of anatase, or in one case ((MR32601); (DX573)) brookite. RJM, AET

Lundy

The predominantly argillaceous sediments of Lundy contain thin, impersistent arenaceous and calc-arenaceous horizons and resemble the Morte Slates of the mainland, although they are not necessarily equivalent in age.

Thin bands of arenaceous material typically no more than a centimetre or so thick can be distinguished in most outcrops. Such rock generally has a mean grain-size in the siltstone range, with an upper limit of about 0.07 mm: that is fine sandstone. Subangular to subrounded quartz grains, showing various degrees of recrystallisation, commonly to the point of being augen-shaped, are the most abundant framework constituent. Some samples contain detrital sodic plagioclase (E4618)<span data-type="footnote">The localities from which Lundy specimens mentioned in the text were collected are listed in Appendix 2.</span>; (E9234A); (E9235A); (E11288); (E42769); (E42802)), which in one case (E42799) shows partial alteration to secondary clay minerals. Other framework constituents include detrital muscovite and small numbers of lithic fragments, chiefly chert (E9235A); (E11285); (E42769); (E42793)) and silty material (E9235A); (E11285); (E42789); (E42800)), with some acid igneous rock ((E42793); (E42799); (E42800)), quartzite (E42800) and quartz-sericite-rock (E42800). Chlorite is the dominant matrix component in the samples examined, which accords with observations on apparently similar arenaceous rocks from the Morte Slates. The matrix may also contain a little very fine-grained muscovite (sericite) ((E4618); (E9234A); (E11285); (E42769); (E42789); (E42793); (E42802)). However, in certain rocks, notably the fine-grained sandstones, mica is either absent ((E11288); (E42800)), or present only as detrital flakes of muscovite which are partly replaced by, and bear overgrowths of, chlorite (E9235A). Several of the samples lacking in sericitic mica contain epidote in amounts varying from about 1 to 5 per cent by volume (E9235A); (E11288); (E42769); (E42799); (E42800)). Accessory detrital minerals include schorlite tourmaline, zircon, magnetite, ilmenite, probably much of the epidote, carbonaceous material and, in one case (E4618), possibly an opaque, granular manganiferous mineral. Ilmenite and magnetite commonly show partial alteration to limonite. Local concentrations of heavy minerals (tourmaline, zircon, opaque oxide and ?apatite) like those recorded in siltstones from the Morte Slates (E30442) and Pilton Shales (E42713) were noted in specimens (E11288) and (E42789). Pyrite occurs in two rocks ((E4618); (E42799)). Specimen (E42799) also contains thin chlorite-sericite-veinlets cut by a minor quartz-sericitealbite-vein. Euhedral pyrite cubes are scattered throughout the specimen and as one crystal lies across a contact between the matrix and the quartz-sericite-albite-vein, it seems probable that growth of the sulphide post-dated development of all the vein-minerals.

The distribution of the chlorite and sericite matrix in the arenaceous rocks is erratic and relates to local variations in the mean grain-size. A condensed framework is evident where concentrations of quartz fragments are predominantly of sand grade, the matrix minerals being sparse compared with their relative abundance in adjacent silty areas. Some silty parts of specimens also contain irregular bands and patches of sericite-rich argillaceous material ((E4618); (E42789); (E42793)). These textural features may reflect poor sorting, although they have been modified by later shearing and recrystallisation during low-grade regional metamorphism.

Calcareous material is most easily identified in the sea-washed rocks cropping out around Hell's Gates [SS 1456 4372]. There, impersistent silty bands and shearaugen a few centimetres thick consist of 20 to 50 per cent calcite, 10 to 15 per cent chlorite and sericite, some quartz, rare plagioclase and 1 to 2 per cent granular opaque oxide which appears to be a randomly sorted mixture of ilmenite and magnetite ((E26219); (E42803)). Low-grade metamorphism has caused extensive recrystallisation, a feature which may have prompted Dollar's statement (1942, p. 45) that impure arenaceous marbles and relatively pure limestones occur among the Lundy sediments.

Examination of two poorly-localised Geological Survey specimens led Dollar (1942) to suggest the presence of an ash horizon (E4618) and of metamorphosed basic igneous rock (E4629). The former is a silty rock with irregular patches of finely divided sericite but no obvious pyroclastic fragments. The latter is chlorite-magnetite-schist bearing accessory sericite, epidote, quartz, ?sphene and an unidentified, colourless, high-relief mineral. Both it and irregular patches of sericite-hornblende-epidote partly leucoxenised ilmenite rock found in another poorly-localised specimen (E4623) of banded quartz-epidotesericite-siltstone may be altered volcanic, basic igneous material.

Dollar (1942) also referred to a group of quartz-breccias without specifying their location. Perhaps they equate with the fault breccia in the cliff at Ladies Beach [SS 1402 4426], which occupies a vertical, 0.2 to 0.3-m-wide zone separating argillaceous sediments from severely shattered granite. Two dolerite dykes cut the sediments adjacent to the zone. The breccia consists of abundant, shattered, partly re-cemented, generally angular, quartz fragments (up to 10 mm across but typically 0.3 to 0.6 mm), scattered angular pieces of quartz-chlorite-epidote-sandstone, similar siltstone, sericitic and chloritic argillaceous material, granite, plagioclase, potassium-feldspar and biotite, all set in matrix composed chiefly of finely divided quartz and sericite with ramifying growths of calcite crystals ((E4592); (E4624); (E42783); (E42784); (E42787)). Minor chlorite occurs locally in the matrix, which in places is cut by anastomosing veinlets of quartz and sericite. Locally the fragments of quartz are displaced by comminuted silty sandstone and argillaceous rock. One specimen of this rock type (E9236A) contains 1 to 5 per cent of sulphides, mainly in the form of irregular crystals and cubes of pyrite but also as anhedral chalcopyrite. Dispersed cubes of pyrite up to 3 mm across occur in another sample (E4617), which may not have come from Ladies Beach. This rock is of interest because the matrix consists predominantly of fine-grained bluish green tourmaline, with some quartz and scattered crystals of colourless fluorite. Tourmaline-veining is present in the quartz-sericite matrix of a quartz-breccia (E4600) collected along the 'new road to villa'; this rock probably came from the vicinity of Millcombe [SS 1393 4409], near the faulted granite-killas junction.

The argillaceous rocks of Lundy consist of thinly interbedded medium grey and pale greenish grey phyllites and slates. X-ray diffractometer analysis (CuKα; 40 Kχ, 30 mA; 3°–45° 2θ) of three typical samples ((MR31933); (MR31935); (MR31936); (DX 590), (DX591), (DX592) shows that about 90 per cent of their bulk composition is accounted for by quartz, chlorite and muscovite (sericite). Quartz content ranges from 25 to 30 per cent and chlorite and muscovite occur in approximately equal proportions. The remaining 10 per cent of the samples is mainly albite, with minor amounts of microcline and traces of anatase, epidote, calcite and ?pyrolusite.

Thin sections show that finely divided chlorite and muscovite flakes with a pronounced planar orientation parallel to a cleavage form the bulk of the argillaceous rocks. In places individual muscovite flakes have grown to about 0.1 mm in size. The abundant chlorite-muscovite intergrowth, presumably representing an original clay matrix, encloses mainly silt-size, strained quartz grains which because of recrystallisation are commonly augenshaped and display diffuse sericite- and chlorite-laden margins ((E11286); (E11287); (E26220)). Coarser, quartz-rich silt to fine sand grade material occurs locally in- bands or patches in predominantly pelitic rocks ((E11283); (E11284)), suggesting poor sorting. Chlorite also occurs as books of penninite flakes 0.01 to 0.2 mm in size dispersed throughout the rocks. The habit of the books is irregular, tending in numerous cases towards augen shapes. Many have been partially disrupted by shearing, and replacement by muscovite along cleavage planes is common. In one specimen (E11287) chlorite occurs as small flakes associated with clots of granular magnetite. Impersistent bands in another rock (E26220), 0.5 to 1.0 mm thick and consisting mainly of penninite with anhedral calcite, minor plagioclase, quartz and scattered small flakes of a phlogopitic mica, may be sheared-out relics of basic igneous volcanic material.

The argillaceous rocks all possess accessory quantities of finely divided, opaque, granular material distributed along the principal planes of cleavage. X-ray diffractometry suggests that it is mainly anatase; some may be pyrolusite. Optical examination indicates that magnetite is present in some specimens (E11287), and perhaps oxidation of magnetite is responsible for the presence of the thin limonitic coatings commonly seen on cleavage surfaces in the field. Other accessory minerals noted in thin sections, chiefly in the more silty bands, include epidote, tourmaline and calcite. Albite and microcline were not identified optically, although X-ray diffractometry reveals combined amounts of up to 7 or 8 per cent. JRH, RJM

Regional metamorphism

Most of the protoquartzites of the mainland show no development of a metamorphic fabric; recrystallisation is confined to the thin films of fine-grained chlorite and muscovite which surround many of the grains and presumably represent the original clay matrix. More matrix is present in many of the siltstones, and evidence of recrystallisation is stronger; specimens from the Morte Slates show the development of an incipient schistosity enveloping deformed quartz grains whose margins are now obscured by the growth of secondary muscovite and chlorite ((E46122); (E46123)). Fairly extensive recrystallisation has taken place in the argillaceous rocks; clay-minerals are completely lacking, and the persistence of albite as the dominant feldspar is probably due to its stability under greenschist facies conditions of metamorphism. The presence of anatase and brookite in the Devonian pelites, and their absence from most of the similar Carboniferous rocks, may reflect the increased degree of crystallinity attained by TiO₂-phases in sediments which have undergone more advanced, but not necessarily higher grade, metamorphic recrystallisation.

All the rocks examined clearly belong to the chlorite zone; they probably lie mainly in the chlorite I subzone, with one or two specimens from the Morte Slates representing the chlorite II subzone (Reed, 1958). RJM, AET

The texture and mineralogy of the Lundy sediments indicate some degree of recrystallisation during low-grade regional metamorphism. This effect is most pronounced in the argillaceous rocks; original clay-rich sediment has been converted to sericite- and chlorite-rich pelite which commonly shows a well-developed slaty cleavage ((E11283); (E11284); (E26219)). Localised growth of coarser flakes of muscovite along this cleavage has given rise to typical phyllites ((E11287); (E26220)). Specimen (E26220) also displays a microscopic flexing of the first schistosity by a later crenulation cleavage. The phyllitic fabric and the augen shapes of the scattered silt-size quartz grains indicate that minor shear stresses were operative during recrystallisation. The presence of a dominantly albitic feldspar in these rocks, revealed by X-ray diffractometry, is due almost certainly to its stability under greenschist facies conditions.

Evidence of metamorphic recrystallisation is less obvious in the arenaceous rocks of the island. In some sandstones chlorite and sericite occur as thin films coating the interlocking grains of the framework. Clearly, both the phyllosilicates and the texture could be the result of normal diagenetic processes. However, local shears in several specimens (E9234A); (E11285)) bear abundant metamorphic muscovite which penetrates the framework quartz grains. In addition arenaceous rocks rich in detrital epidote contain associated granular epidote which may be of metamorphic origin (E9235A).

In terms of metamorphic grade, the Lundy sediments can be classified as greenschists, with the pelitic rocks placed more precisely in the quartz-albite-muscovite-chlorite subfacies. According to Barrovian terminology the Lundy rocks belong to the chlorite II subzone (Reed, 1958), and generally are a little higher in grade than the Morte Slates of the mainland coast.

Contact metamorphism

Sedimentary xenoliths in the Lundy Granite have been converted to quartz-biotite-hornfelses and quartz-biotitemuscovite-hornfelses containing accessory opaque oxide, tourmaline and topaz (E42765). A silty specimen (E9221A) collected at the granite contact below Benjamin's Chair about [SS 138 437] has similarly been converted to a tourmaline-bearing, quartz-biotite-muscoviterock. However, the main granite-killas junction is a fault, adjacent sediments showing no evidence of contact metamorphism. The main block of sediments may have been situated several tens of metres above the granite at the time of intrusion and subsequently downfaulted or, alternatively, the present juxtaposition of rocks could have resulted from lateral movement. Consequently sediments on either side of the fault in the southern cliffs may differ in age, although they appear similar lithologically.

The quartz, sericite, calcite and localised tourmaline which form the matrix of the fault-breccia at Ladies Beach are not products of contact metamorphism. Presumably they were deposited, along with sporadic pyrite and chalcopyrite crystals, by Palaeogene hydrothermal fluids associated with the faulting. Hydrothermal activity may have been responsible also for the dispersed pyrite, and the chlorite-sericite-veining and quartz-sericite-albite-veining in specimen (E42799), from an outcrop [SS 1424 4384] by the road to the Landing Beach some 400 m from the faulted granite junction.

Contact metamorphic effects associated with the basic and trachytic dykes are negligible. No modification of the regional metamorphic fabric and mineralogy is evident in three phyllitic specimens (E4619), (E4620) collected from dyke contacts. In a fourth sample, a siltstone (E9234A), quartz fragments have a clarity which suggests the adjacent 6-m-wide trachyte dyke caused some recrystallisation. JRH, RJM

Details

Morte Slates

Coastal sections

Between Flat Point [SS 495 472] and Shag Point [SS 487 471] grey slates with lenses, wisps and bands of siltstone and silty sandstone, and traces of limestone, dip 50°–80°/185°–210°. They are cut by scattered NW–SE faults, which here and elsewhere control the alignment of the reefs of the foreshore and facilitate erosion by the sea along SE-trending guts. The cliffs are capped by collapsed slates disposed in a variety of attitudes but generally dipping at angles less than those seen in the cliff face. The strata hereabouts appear to contain more siltstone and silty sandstone, and to show less quartz veining, than is evident farther west, and this may account for the base of the Morte Slates being taken just west of Shag Point on Hamling's map (1910). There is, however, little validity in such a line on the coast and no possibility of following it inland.

From Shag Point to Lee Bay [SS 479 466] purple-tinged grey slates dip 55°–75°/190°–210° and are cut by many small NW–SE fractures. The slates are locally silty and carry much quartz along the cleavage, both as single stringers and as groups of veinlets. Black Pit [SS 478 466] marks one of the few NE–SW faults noted.

Farther west smooth grey slates dipping 50°–80°/205°–220° are capped by collapsed slates which have locally [SS 4775 4665]–[SS 4755 4677] developed gentle folds during superficial movement. Quartz veining is common both along and across the cleavage and is particularly marked to the east and west of Pensport Rock [SS 4741 4693]. Towards Bull Point [SS 462 469] smooth greyish green, brown and purple slates show local dips as low as 40°/205°. In some W-facing or E-facing cliffs steep south-south-westerly dips diminish upwards into slipped material showing inclinations of around 10° in similar directions [SS 4676 4676]–[SS 4648 4685]. At Bull Point southerly dips of 70°–75° are common in greyish green and purple slates, locally slightly silty, but traces of north-northwesterly dips suggest the presence of upright folding, in contrast to the general pattern of tight overfolds to the east. A short distance south-west of the lighthouse dips of 70°/180° flatten progressively upwards to 20°/180° at the top of the cliff [SS 4614 4667]. Again this is probably a superficial effect, collapsed rock being common along the cliff top to the south. From Bull Point to Rockham Bay [SS 458 461] the slates are cut by many small NW–SE faults and a few older faults trending between WSW and WNW; the coastline faces west-north-west and is finely indented by erosion along fault lines.

De la Beche (1839, pp. 44, 49; fig. 2, pl. 3) described open upright folding in Rockham Bay. This was also recorded by Hicks (1896), who obtained Lingula there, and by Arber (1911). Probably it is due to the presence among the slates of strong sandy beds which resisted the fold movements. A good deal of vein quartz occurs in the slates hereabouts, locally concentrated along the cleavage. Near the foot of the steps leading down to Rockham Bay [SS 4587 4607] vein quartz is seen as a series of discontinuous boudin-like developments associated with a fault plane.

The N-facing cliffs between Rockham Bay and Morte Point [SS 442 456] are especially deeply indented, being cut by a multiplicity of small NW–SE faults and a few trending E or NE e.g. at [SS 4522 4571]. Grey slates strike generally between 095° and 120° and are inclined vertically or steeply SSW. Vertical dip joints may carry vein quartz [SS 4505 4577]; other joint planes dipping 20°–40° northerly are strikingly displayed on the north side of the point [SS 4449 4567]–[SS 4442 4567]. The rocks hereabouts about [SS 4472 4568]–[SS 4422 4555] are locally silty and carry more streaks and beds of fine-grained sandstone than do those to the north and south.

At Morte Point (Plate 1) grey slates strike 115°. Hicks (1896), who collected large Lingula mortensis there and noted the presence of numerous minute specimens of that genus, observed that the bedding is nearly vertical and largely obscured by cleavage, and the folding acute. Weathering of silty material has locally penetrated 100 mm or more into the rocks along cleavage planes, minor joints or fractures. The occurrence of vertical and some north-north-easterly dips, 50°–70°/010°–030°, on the southern side of the point suggests the presence of close upright folds or possibly folds slightly overturned southwards. Collapsed strata at the cliff top are locally up to 6 m thick and horizontal [SS 4451 4541]. Joints are common and well exposed, dipping 25°/345° [SS 4487 4521].

Grey slates and silty slates of the cliffs and foreshore around Grunta Pool [SS 452 447] are disposed in tight folds whose axial planes are commonly nearly vertical but locally dip steeply southwards. Slates near Combesgate Beach contain silty nodules and a 1.2-m-wide quartz vein [SS 4543 4445]. Immediately to the south, quartz veins up to 1 m wide are common in slates and slaty siltstones. Barricane Beach [SS 454 442] has long been known as a fossil locality. Hicks (1896) identified his specimens from hereabouts (p. 7) as Lingula mortensis, Rhynchonella lewisii?, Spirifer hamlingii, Orthis rustica, Modiolopsis barricanensis and Avicula sp.,and he also mentioned 'encrinites'. Collecting during the recent survey yielded rhynchonellaceans? from near the bottom of the path down to Barricane Beach [SS 4542 4422] and the following fossils from sandstone lenses in slates near high-water-mark [SS 4539 4422]: Cyrtospirifer cf. verneuili, cf. Ripidiorhynchus ferquensis and cf. R.? boloniensis. The foreshore reef immediately south of Barricane Beach yielded C. cf. verneuili, rhynchonellaceans? and 'Modiolopsis' barricanensis, which may be a synonym of Leptodesma (Leptodesma) spinigerum (Conrad).

Folded and faulted grey slates and silty slates at [SS 4527 4395] near Shellsborough Cove show quartz veining associated with the fault, and contain sheared boudins of siltstone with pyrite cubes. Grey slates at the northern end of Woolacombe Sand [SS 453 439] weather brown and friable and contain flattened silty nodules with pyrite cubes. The topmost Morte Slates exposed on the coast are grey slates, vertical strike 110° or dipping 80°/005°, on the eastern edge of Woolacombe Sand [SS 457 436].

Inland sections

East of Rockham Bay small jagged outcrops of green, grey, brown and purple slates are numerous but rarely significant. Dips are generally 50°–80° around SSW, but locally southerly and at angles as low as 25°. Slight flexures, perhaps ripple marks, at one locality [SS 4603 4630] are aligned with their axes running down-dip (190°). Hicks (1896) recorded Lingula from several inland exposures in the neighbourhood of Lee. Immediately east of that village steep dips to north-north-east and south-south-west suggest the presence of upright folds [SS 487 464], and still farther east, the area inland of Flat Point is underlain by slates with siltstones and thin sandstones.

The high ground of the Morte Point promontory shows slates with siltstones, locally sandy. Dips are 50°–85°/195°–225°, with some vertical and a few north-north-easterly. These last are more common eastwards towards Mortehoe, immediately beyond which an abundance of quartz veins has given rise to rough humpy field surfaces. Two small disused pits [SS 4655 4584]; [SS 4657 4560] were apparently opened in slates. A stream section in Borough Valley [SS 4819 4600]–[SS 4830 4515] reveals silty slates tightly overfolded northwards. High on the eastern banks an old quarry [SS 4847 4527] shows grey slates, but surface brash hereabouts includes sub-rounded fragments of grey and white limestone and black crystalline limestone with crinoid debris.

The coast road through Woolacombe cuts through slates similar to those in the nearby low cliffs. Northerly, southerly and vertical dips are visible, but readily discernible folds [SS 4547 4466] are rare. North-north-easterly dips are commonest in the southern part of the village. Exposures along a track to Twitchen show fine-grained quartzitic sandstones up to 100 mm thick in slates [SS 4590 4460]. Small exposures, including some old pits, of quartz-veined slates and silty slates are plentiful in this area. Locally the slates have a lustrous greenish red sheen. Roadside sections near Oussaborough show shiny grey quartz-veined slates with one set of joints dipping steeply southwards and another vertical trending N–S [SS 4798 4351] and quartz-veined slates and siltstones showing the effects of superficial creep [SS 4707 4366]; [SS 4794 4347]. Old railway cuttings both down-line and up-line from Mortehoe and Woolacombe Station [SS 4970 4463]–[SS 4926 4458]; [SS 4850 4413]–[SS 4842 4396]; [SS 4843 4323]–[SS 4866 4311]; [SS 4920 4272]–[SS 4937 4261] are cut in slates and silty slates showing overfolding to the north. Dips lie within the range 55°–85°/186°–210° or are vertical. The southernmost cutting lies only 60 m from the base of the Pickwell Down Sandstones but the proximity of these more competent beds appears not to have affected the style of folding.

Lundy Island

The south-eastern tip of Lundy Island consists of a series of grey and greenish grey neritic slates (Plate 4), locally containing silty and sandy nodules with pyrite and chalcopyrite, with some siltstones and fine-grained sandstones and scattered thin calcareous beds and lenses. Quartz veins and stringers are common and locally carry pyrite and chalcopyrite. The rocks have been intruded by granite, a faulted contact emerging at the coast near Ladies Beach [SS 1402 4425] and Rattles Anchorage [SS 1370 4355], and by dykes (Chapter 6). Dollar (1942) identified arenaceous chlorite-slates, chlorite-epidote-sandstones. calcareous sandstones, calc-flintas, impure arenaceous marbles and a few re-crystallised limestones. Two specimens collected by Etheridge (1867) proved to be of chlorite-schist and possibly altered volcanic tuff, but no trace of such rocks has been found in the field.

Structurally the rocks are locally less intensely folded than are the Morte Slates with which they are correlated. Close upright folding is evident around Hell's Gates [SS 1456 4372] and the pattern of dips above The Rattles suggests more open folds. However, the main mass of slates which underlies Castle Hill [SS 140 438] may contain folds overturned to the south-west. Local contortion is common, as are minor folds on the main fold limbs; amplitudes of these structures are commonly of the order of 0.3 m. It is difficult to ascertain the thickness of the sediments of Lundy, but when allowance has been made for the folding it seems possible that no more than 100 m of strata are present.

Dips in the neighbourhood of Ladies Beach are commonly 20°–35°/025°–060°, and the slates here lie within the northeastern limb of an anticline. Miller's Cake [SS 1412 4412] displays a joint plane dipping about 15°E. Steeper dips (40°–65°/032°–045°) prevail around Mill Combe where silty slates show bedding plane faults [SS 1409 4400] and joints inclined 50°/180° [SS 1414 4402]. Shales sporadically exposed higher up the track dip 55°–60°/035°–040°, are locally contorted [SS 1394 4389] and may be overfolded.

Slates, siltstones, sandstones and thin impersistent limy beds in the neighbourhood of the Landing Beach dip around 60°–75° NNE or NE, but locally are inclined as gently as 35° in similar directions. Steep dips characterise the high ground around Lundy South Lighthouse. At the eastern end of the Lametry peninsula silty slates dip 50°–60° NE in the southern limb of a syncline; the northern limb dips 75°–80° SSW in the south-western part of Rat Island. The axis of the fold runs through Hell's Gates trending ESE, is characterised by twisted and contorted strata, and may be faulted.

Above The Rattles slates with fine-grained sandstones, dipping 30°/030° near the cliff top but at moderate angles south-southwestwards just below, are disposed in open folds. Tiny quartz veinlets are common. A short distance to the west, in the western cliffs above Rattles Anchorage, the slate-granite contact rises from the base of the cliffs [SS 1370 4355] trending NE and NNE and dipping very steeply north-westwards. The contact is cut by two dolerite dykes and appears to have been invaded [SS 1381 4372] by a third. Some disruption of the sediments has occurred, with brecciation and quartz veining, the latter carrying tourmaline and locally small xenoliths. No veining by granite is evident either here or at the less accessible contact at Ladies Beach. The contact high in the western cliffs of Rattles Anchorage is difficult of access and may be normal; in this connection it is interesting to note that the specimen of schist referred to above (p.17) came from this area.

Pickwell Down Sandstones

Coastal sections

The base of the Pickwell Down Sandstones is marked in places by the development of tuffaceous bands, first named 'Bittadon Felsite' by Bonney (1878). Mill Rock [SS 4553 4312], which projects through Woolacombe Sand just above high-water-mark, shows bands of strong cleaved granular rock, one of them up to 0.6 m wide, but microscopic examination has revealed them to be metamorphosed sandstones. The bands lie concordantly within greyish green, grey and purple slates, siltstones and sandstones dipping at up to 60°/190°. Traces of current bedding show the strata to be right way up, and vertical or near-vertical cleavage trends 110°. Rogers (1919) collected remains of the following typical Upper Devonian fish from what he described as a single lenticle of tuff: Holonema cf. ornatum, Bothriolepis, Holoptychius, Polyplocodus, Coccosteus. He noted that thin sections showed distinct flow structure and postulated slight re-working after deposition of the volcanic dust and ash in water. There is no readily discernible tuff in the present outcrop, which suggests that Rogers's identification was uncertain or that some exposures have been buried by sand.

Near Vention, towards the southern end of Woolacombe Sand, fine- and medium-grained grey feldspathic sandstones in adjacent exposures are vertical, strike 090°, and dip 45°/195°. Immediately to the south, Black Rock [SS 4505 4134] shows fine-grained thickly bedded purple sandstones dipping 50°/185°. Current bedding in the sandstones shows the southerly dipping strata to be right way up. A delicate reticulate pattern of fine fractures carries vein quartz and traces of manganese oxides. A set of similar fractures indicates jointing inclined east-south-east, and strong hard quartzitic purple and buff feldspathic sandstones which crop out about 230 m to the SW [SS 4494 4113] are cut by quartz-coated joints dipping east. These sandstones are current bedded, right way up and dip 50°/200°. Similarly disposed current-bedded purple and brown sandstones farther south-west [SS 4490 4104]; [SS 4485 4095] are generally micaceous.

Towards the southern end of the sands, and about 100 m from the outcrop of the top of the formation, massive and thickly bedded fine-grained current-bedded buff, grey and purple sandstones dip 65°/195° [SS 4465 4080]. They are locally micacous and show the incoming, towards the top of the formation, of micaceous siltstones. Low cliffs below Putsborough Sands Hotel display hard fine-grained sandstones dipping steeply southwards beneath thinner-bedded fine-grained sandstones, siltstones and silty slates [SS 4460 4072]. The upward passage into Upcott Slates occupies about 9 m of strata, with progressively higher sandstones becoming thinner and finer grained and giving way to predominant siltstones, silty slates and slates.

Inland sections

Surface brash of feldspathic tuff with ferruginous cavities and spots and scattered quartz grains [SS 4625 4325] marks the base of the Pickwell Down Sandstones 750 m inland from Mill Rock. Basal strata of the formation are exposed on Potter's Hill [SS 459 430] as buff, purple and grey sandstones and slates. A quarry to the east [SS 4610 4303] shows purple and greyish purple slates and sandstones striking 090°, dipping generally southwards at various angles and cut by strike faults filled with ferruginous clay. Typical purple and brown fine-grained sandstones are similarly accompanied by a good deal of slate and silty slate in two old quarries [SS 4565 4253]; [SS 4562 4247] adjoining the coastal Marine Drive.

The base of the formation makes a good topographic feature for 2 km inland from the coast to where [SS 476 430] it is displaced about 200 m S by a fault trending S and SE towards Spreacombe. Quarries on the west side of this fault show red, grey and brown thinly and thickly bedded sandstones with silty slates dipping 65°/200° [SS 4752 4286], and thickly bedded purple ferruginous sandstones and grey sandstones dipping 20°/215° [SS 4773 4247] and accompanied by fragments of brecciated ironstone which are probably fault rock. Just south of the latter quarry two small flooded adits [SS 477 424] in ferruginous Head probably mark trials for iron ore (p. 115). To the east of the fault the basal feature continues to Buttercombe [SS 4992 4227], just outside the district. No good exposures of the lower strata occur in this area, but surface brash of purple and brown fine-grained sandstone with some slate is abundant. The site of a possible old shaft [SS 4926 4237] may mark another trial for iron. A well at Spreacombe Lodge [SS 492 421] was recorded as having passed through 52 m of 'rock'.

The western slopes of Pickwell Down contain an old quarry [SS 4537 4151] in rubbly, red, purple and brown thickly bedded fine-grained sandstone dipping 30°/180°, and another [SS 4566 4142] in which purple and grey fine-grained sandstones are cut by a steep fault trending 320° and containing a 1-m-wide ferruginous gouge rock. West of Pickwell itself similar small disused quarries show the incoming of pale grey and green sandy siltstones among massive and thickly bedded fine-grained feldspathic and micaceous sandstones near the top of the formation [SS 4493 4081]; [SS 4503 4082]. The high ground of Pickwell Down, as that of Woolacombe Down to the north, carries no good exposures but traces of old pits and abundant brash of fine-grained sandstone.

Farther east a quarry [SS 4776 4137] near Spreacombe Mine (p.115) was opened in thickly bedded fine-grained sandstones disposed both horizontally and with dips around 25°/190°. Three N–S faults exposed in the quarry face are up to 0.9 m wide and filled with ferruginous fault breccia and hematite. Rubble of similar sandstones is common in the fields to the south along a line of old shafts (p. 115) and around the remains of old pits and collapsed adits [SS 4753 4106]; [SS 4760 4105]; [SS 4820 4110].

A borehole at Spreacombe about [SS 4855 4155] was recorded as having been sunk through drift and weathered rock 5.5 m, clay with sandstone bands 11.3 m, brown sandstone 13.1 m, on shattered sandstone 9.8 m. Vyse Quarry, still active as a source of roadstone and hardcore, comprises four bays. In the northernmost [SS 4933 4133] fine-grained sandstones dipping 10°/220° are cut by five small faults, three trending with the strike and two between NW and NNW. The western wall of the west bay is almost coincident with a fault trending 175°, on the eastern side of which (Figure 3) brown, grey and purple sandstones show a general dip of 45°/190° but are thrown into minor folds, commonly cut by small strike faults [SS 491 412]. Thickly bedded and massive grey fine-grained sandstones dipping 20°/195° in the east bay [SS 4920 4116] are cut by several strike faults filled with rock debris and are locally sheared and slaty in the vicinity of fractures. Sandstones of the south bay [SS 4916 4110] dip 20°/200°. Rubbly brown and purple fine-grained sandstone is poorly exposed in several small disused pits in this area, as that [SS 4932 4054] near Buckland Mine (p. 115), and debris of brown and purple ferruginous sandstone and hematitic fault rock is common at this last locality.

Farther south, in Stony Bridge Quarry [49353945], a steep fault trending around SE or ESE roughly with the strike, carries a few centimetres of clay. On the north side of the fracture 12 m of grey fine-grained sandstones with some slates dip 80°/200°. South of the fault about 40 m of strata are exposed, their dip increasing from 70°/220° near the fault to 85°/220° on the south side of the quarry. This south face lies little more than 100 m from the outcrop of the top of the formation; it shows more slates and micaceous siltstones than occur elsewhere in the quarry and the beds are locally vertical, strike 130°. The valley of the River Caen hereabouts coincides with a NNE–SSW fault, on the east side of which the top of the Pickwell Down Sandstones is displaced about 130 m S. Small quarries alongside the A361 road [around 4941 3923] display purple fine-grained sandstones, siltstones and silty slates striking 100°–110° and dipping steeply southerly.

Upcott Slates

Coastal sections

The passage beds at the top of the Pickwell Down Sandstones, at the southern end of Woolacombe Sand, dip steeply south-southeastwards beneath purple, green and grey slates and silty slates. The exact position of the junction is arbitrary, but it is here considered that Napps Cliff [SS 444 407] is cut in Upcott Slates, which also make up the whole of the northern side of the Baggy Point promontory. In Napps Cliff, and in foreshore exposures immediately to the west [SS 4416 4081]; [SS 4408 4086], purple, grey, green and yellow slates, silty slates, siltstones and scattered thin fine-grained sandstones dip 55°–60°/200°–210°. Similar rocks crop out westwards to Bloodhill Cliff [SS 4338 4085] and Cox Cliff [SS 4321 4085], with slates and silty slates predominating, and are cut by several small faults trending around NW–SE [SS 4379 4082]; [SS 4373 4084]; [SS 4347 4083]; [SS 4329 4083]; [SS 4311 4084]. Current bedding in the scattered thin sandstones commonly indicates that the strata are right way up, younging south-south-westwards, but vertically disposed slates and silty slates suggest the possible presence of some fairly tight folds whose axial planes dip steeply SSW. Some indeterminate shell fragments are present. Goldring (1971) recorded sinuous sandstone-filled grooves in the Upcott Slates of the coast; he noted that they commonly had a greater depth than breadth (e.g. depth 250 to 300 mm, breadth 200 mm).

Between Cox Cliff and Whiting Hole [SS 4217 4102] the foreshore is not readily accessible. Evidence from the cliff top points to a succession of strata overlying but similar to that exposed farther east. The slates and silty slates strike around 110°, but superficial movement of the rocks capping the cliffs renders dip measurements of doubtful value.

About 160 m south of Whiting Hole slates, silty slates and siltstones dipping around 70° SSW are overlain conformably by sandstones of the Baggy Sandstones [SS 4215 4086]. The junction is sharp, but appears to be slightly displaced locally by minor faults. The topmost slates of the Upcott Slates accommodate load casts of the overlying sandstone.

Inland sections

Exposures on the northern slopes of the Baggy Point promontory show green, grey and brown slates and silty slates with some siltstones and sandstones. The beds generally strike 100°–110° and are vertical or steeply dipping SSW. Surface debris includes a few fragments of shelly limestone, but none was found in situ.

Putsborough Sands Hotel [SS 4480 4065] stands on Upcott Slates near the base of the formation, and exposures nearby [SS 4475 4065] show purple silty slates weathering pale greenish grey. Vention Lane leads south-westwards thence, roughly across the strike of the rocks, and contains sporadic exposures of greenish grey and yellowish brown silty slates and siltstones with thin fine-grained sandstones, striking around 100°–105° and commonly vertical or steep. Putsborough [SS 449 403] lies athwart the Upcott Slates–Baggy Sandstones junction, although mainly on the former formation. Exposures of topmost beds of the Upcott Slates hereabouts show siltstones, slates and fine-grained sandstones.

Inland the Upcott Slates generally occupy a depression between the more competent Pickwell Down Sandstones to the north and the Baggy Sandstones to the south. Their junction with the former is not readily discernible, but topography suggests that it is displaced north-eastwards to Pickwell [SS 459 409] by a NE–SW fault. This doubles the width of outcrop, as compared with the formation at the coast, to about 1 km in the area north-west of George-ham. The cemetery [SS 4690 4035] north-east of Georgeham appears to lie on basal Upcott Slates. A small old pit [SS 4648 4006] between there and the village shows rubbly slates and fine-grained sandstones which match the surface brash of the fields. Numerous small exposures in and around Georgeham show slates, silty slates, siltstones and fine-grained sandstones striking 105°–115° and vertical or steep. Near Crowborough [SS 4678 3954] similar strata dip 50°–55°/360°–005° and 80°/310°, apparently in fairly upright or southerly overturned folds disrupted by NNE–SSW and NW–SE faulting.

A NNW–SSE fault passing just east of Crowborough has produced a major lateral displacement, the Upcott Slates–Baggy Sandstones junction having been moved some 700 m dextrally. Between Darracott [SS 472 392] and Upcott [SS 482 393] mapping depends on surface brash and one or two small rubbly exposures. Farther east roadside outcrops in Nethercott [SS 485 393] show slates and silty slates vertical, strike 110°, and dipping steeply south-south-west. A series of laneside exposures in similarly disposed beds [4857 3924–4863 3912] are of silty slates, siltstones and grey quartzitic fine-grained sandstones near the top of the formation.

The outcrop of the Upcott Slates in the neighbourhood of Nethercott is reduced to around 300 m in width. To the east it is displaced dextrally by a fault running along the valley of the River Caen. The A361 road along the eastern side of the valley cuts through basal Upcott Slates [SS 4935 3915] which show a southerly (upward) passage from slates and siltstones with fine-grained sandstones to predominantly slaty rocks. Farther south alongside the same road the topmost beds of the formation are seen to be mainly slates with thin sandstones striking 105°–110° and dipping steeply southerly beneath more massive sandstones of the Baggy Sandstones [SS 4927 3879]. Eastwards from the road the Upcott Slates occupy a strip of relatively low-lying ground. Grey and greenish grey slates and silty slates in scattered small exposures strike 100°–110° and show dips around vertical and 40°–80° NNE.

Baggy Sandstones

Coastal sections

The coastal section of the Baggy Sandstones is one of a normal stratigraphical succession in which the beds generally dip fairly steeply around SSW, the direction of younging. Examination confirms the detail of Goldring's (1971) work (Figure 4). However, the nature of the sediments suggests that individual beds are unlikely to persist far laterally.

The fine-grained sandstone which sharply succeeds the Upcott Slates [SS 4215 4086] forms the basal bed of a 45-m series of grey shales, siltstones and thin sandstones which crops out in the cliffs [SS 4217 4084] west of the Coastguard Lookout and dips around 70° SSW. Sandstones become more numerous towards the top of the series, which is overlain by 7 m of fine-grained sandstones, locally feldspathic, micaceous and shelly, which form Long Rock [SS 4205 4082]. Baggy Hole [SS 4212 4078], a cave accessible at very low tides and containing good examples of ripple marks, is cut approximately along the strike of the succeeding shales, siltstones and thin sandstones, which yielded a lingulid? fragment and bellerophontids ? from their topmost beds. They are themselves overlain by 5 m of feldspathic and calcareous fine-grained sandstones, with burrows of Arenicolites curvatus and Diplocraterion yoyo, forming a small reef on the foreshore [SS 4210 4076]. Argillaceous sediments above these sandstones have been more easily cut back by the sea [SS 4210 4075]; they comprise about 19 m of greenish grey shales and siltstones with scattered fine-grained sandstones which are most common towards the base, and again towards the top, of the shales. A few very thin shelly limestones are present. Goldring (1971) recorded the following fossils from this shaly sequence: Lingula squamiformis, Echinocaris sloliensis, E. whidbornei, Prothyris contorta, Edmondia spp., Sanguinolites complanatus, Bellerophon subglobatus and Actinoceras sp. A small promontory [SS 4207 4074] separating the two main coves between Long Rock and Baggy Point is composed of 18 m of thickly bedded and massive feldspathic and calcareous current-bedded sandstones with some platy micaceous sandstones. Bands, lenses and wisps of argillaceous material are present, and conglomeratic bands characterised by shale pellets occur in the lower part.

Overlying the strong sandstones are shales and siltstones similar to those below and containing a small cave. They pass upwards into similar rocks with scattered thin sandstone bands which are in turn capped sharply and somewhat irregularly by about 17 m of thinly bedded calcareous micaceous and feldspathic sandstones and fine-grained current-bedded sandstones with shale-pellet conglomerates [SS 4208 4071]. Plant debris is common and Goldring (1971) visualised sands filling a plant-carrying distributary channel; he argued that the flora of the Baggy Sandstones grew in fresh-water swamps whence debris was carried down river to the mouth. The succeeding 12 m of shales, siltstones and thin sandstones, reminiscent of similar sediments separating the larger sandstones lower in the formation, become finer upwards, dip 85°/210°, show much bioturbation by D. yoyo and appear to be cut by a bedding-plane fault [SS 4210 4068]; ferruginous staining in the vicinity of the fault extends through up to 2 m of strata. Thickly bedded and massive sandstones in the cove to the south [SS 4209 4067] are locally micaceous and flaggy and separated by 1 m of shales from overlying current-bedded sandstones with plant debris; the total thickness of this sandy sequence is about 8 m. A strike fault appears to repeat these beds in a small headland [SS 4203 4067], and in the beds between Goldring noted the greatest penetration (150 mm) of sediment by D. yoyo burrows. Beds on this headland [SS 4204 4067] have yielded a bellerophontoid?, Ptychopteria damnoniensis and 'Cucullaea'unilateralis.

The path which descends the north side of Baggy Point reaches the foreshore [SS 4203 4065] in a series of shales and siltstones with feldspathic and micaceous current-bedded sandstones which have yielded [SS 4204 4065] cf. Lingula squamiformis and Palaeoneilo sp.These strata form the northern side of the point and Goldring (1971) has designated one of the coarser-grained sandstone assemblages, whose upper part contains burrows of Monocraterion sp.,his Monocraterion Member [SS 4204 4067]; it is overlain by about 5 m of thinly bedded shales, siltstones and sandstones, and these in turn and irregularly by a similar thickness of fine-grained current-bedded sandstones with load casts and flute casts and plant debris. Following the succession upwards the next prominent sandstones are those exposed on the north side of the extremity of Baggy Point. They are about 15 m thick and show [SS 4190 4067] planty sandstones with an irregular base, separated by 2 m of silty shales with plant fragments from lens-bedded fine-grained feldspathic sandstones and more thickly bedded micaceous sandstones with shale pellets and a thin calcareous band. Some pebbles of calcareous sandstone or sandy limestone occur in the overlying shales.

Three other units of current-bedded sandstones strike WNW through the point itself and are separated by shales with a few siltstones and fine-grained sandstones. The lowest [SS 4190 4065] is about 6 m thick and consists of current-bedded locally micaceous fine-grained sandstones, feldspathic and carrying plant fragments in the higher beds and with shale pellets in the lower. The middle unit [SS 4189 4064] is thicker, about 16 m; fine-grained sandstones with an irregular base pass up into current-bedded sandstones with plant fragments and subordinate shales and with some shale pellets and 'Cucullaea'unilateralis towards the top. The highest unit [SS 4187 4061], perhaps about 17 m thick, shows fine-grained current-bedded sandstones conglomeratic towards the base, some with plant debris and others thinly bedded and bioturbated, and a thin sandy limestone.

A steep track [SS 4192 4057] down the south side of Baggy Point affords access to sections [SS 4195 4057] in which shales and siltstones are overlain by up to 8 m of current-bedded sandstones with load casts and containing rounded fragments of calcareous sandstone and, according to Goldring (1971), calcarenite in the basal part. The cove at the foot of the steep track is eroded in shales and siltstones with sporadic fine-grained sandstone bands. Similar strata crop out to the south-east, commonly dipping around 65°–75°/185°–195° and becoming slightly sandier upwards. They probably exceed 80 m in thickness, and include some calcareous feldspathic sandstones [SS 4207 4042], and thin current-bedded sandstones at the top pass upwards into an 11-m sandstone unit. Strikingly current-bedded shelly sandstones are overlain by micaceous sandstones with intraformational conglomeratic bands and lenses and a few shales, and these in turn are capped by more current-bedded sandstones containing shell debris in their lower part [SS 4215 4035]. Above the sandstones thin discontinuous lenses and bodies of crinoidal limestone lie within silty shales and siltstones. The sandstone assemblage is succeeded by the shallow-water marine shales, siltstones and thin sandstones with D. yoyo [SS 4221 4031] so characteristic of the sort of background sedimentation within which the sandy units have developed.

Overlying these more argillaceous strata in the cliffs [SS 4220 4030], with an irregular junction, are a few metres of current-bedded fine sandstones, locally micaceous, with lenses of shale pellets and shell fragments and with a limy conglomeratic band at the top. Goldring (1971) collected Whidbornella caperata, Hamlingella goergesi and Ptychopteria damnoniensis here; he noted that the sandstone beds were absent from the corresponding approximate dip section in shales, siltstones and sandstones near and within a cave [SS 4234 4025] to the SSE, but he located a 1.5-m boulder bed in this section comprising angular and rounded blocks of fine-grained sandstone in a silty matrix. The following fossils were collected from the cliff near by [SS 4228 4023]: 'Bellerophon' sp.,gastropod indet., 'Orthoceras' sp., Palaeoneilo?, Ptychopteria damnoniensis, myophoriid bivalve and crinoid debris.

A slumped bed (Plate 3) crops out prominently on the foreshore [SS 4221 4024]–[SS 4245 4015], running about WNW–ESE and wedging out eastwards. It is irregular, but around 6 m thick, and comprises large balled-up masses of fine-grained sandstone and siltstone showing traces of current bedding. The underlying shales and siltstones are much deformed in some places, little disturbed in others. Goldring (1971) quoted Professor W. D. Gill's opinion that the material originated as sand volcanoes from vents or fissures; however there is no direct evidence of this, and an explanation in terms of gravity-induced submarine slumping seems adequate. Some 5 m of sandstones immediately overlie the slumped bed. They are fine grained, current bedded, feldspathic and locally calcareous or micaceous; some conglomeratic developments with locally derived shale pellets occur at the top, suggesting slight erosion. This horizon is taken as the top of the formation.

Inland sections

Buff current-bedded sandstone [SS 4224 4081] 120 m WSW of the Coastguard Lookout probably corresponds to the lowest group of sandstones seen in the cliffs [SS 4205 4082]. Farther south a small quarry [SS 4207 4061] has been cut in current-bedded sandstones exposed at sea level on the south side of the point [SS 4195 4057]. Rocks probably belonging to the same group of sandstones crop out in a second quarry nearby [SS 4204 4060] as pale brown micaceous medium-grained sandstones with wisps of pale green and brown soapy shales. Sandstone has been dug [SS 4238 4072] from a field north-west of Croyde Hoe Farm and exposures alongside the coastal footpath yielded Palaeoneilo? [SS 4208 4047] and cf. Ptychopteria damnoniensis [SS 4223 4036]. Several old pits lie near the track leading from the farm towards Croyde Bay. The westernmost [SS 4268 4035] is overgrown, but 70 m to the east [SS 4275 4034] buff fine- to medium-grained sandstones are exposed in rubbly outcrops. An adjoining old pit [SS 4279 4033], in sandstones with subordinate shales dipping about 50°/210°, is probably that from which Arber and Goode (1915) and Rogers (1926) collected the plants Knorria sp., Sphenopteridium rigidum and Xenotheca devonica. During the present survey Dr D. E. Butler recorded ?camarotoechiids, rhynchonellaceans indet., 'Cucullaea' unilateralis, lingulid indet. and a loose fish plate, Bothriolepis?, from this quarry. Brown and grey fine-grained sandstones, pale buff micaceous sandstones and silty shales with a thin weathered limestone dug alongside the track [SS 4279 4026] dip 40°/220°. Goldring (1971) collected Cyrtospirifer verneuili, Sanguinolites sp.and Echinocaris sp.here and noted that the age suggested by this assemblage is the same as that of the lower Pilton Shales. Dr Butler obtained the following fossils from this locality in 1970: a plant fragment, cf. Cyrtospirifer verneuili, spiriferaceans indet., strophomenacean indet., cf. 'Macrochilina' pusilla, ?Vaticopsis], bellerophontoid [juv.], gastropods indet., orthocone indet., Prothyris contorta, nuculanids?, bivalves indet., Echinocaris whidbornei, ostracods, crinoid columnals, and a fish tooth.

Near the boundary of the National Trust land [SS 4284 4014] brown and buff fine-grained sandstones dip 65°/200°. Middleborough Hill [SS 432 399] is formed by a south-faulted block of Baggy Sandstones and bears several rubbly exposures of grey and buff fine-grained sandstones and shales with scattered shell fragments. East of this block the outcrop of the formation is narrow, around 400 m, in the area of Middlehill Lane [SS 436 402], and friable buff micaceous sandstones dipping 70°/020° [SS 4394 4021] suggest the presence of upright folds adjoining a NW–SE dextral wrench fault. This^. fault has displaced the Baggy Sandstones southwards almost to the outskirts of Croyde, where they form Ora Hill [SS 447 396]. Small outcrops in and around Putsborough [SS 449 403] trace a junction between slates, siltstones and fine-grained sandstones to the north and the incoming of more massive and thicker bedded sandstones to the south, and this is taken as the base of the formation.

To the east the NE–SW valley from Georgeham to North Hole [SS 4575 3943] is probably fault-induced. On its western banks silty shales and buff feldspathic micaceous medium-grained sandstones crop out in the western end of an old pit [SS 4575 3962], and flaggy feldspathic ferruginous-spotted micaceous medium-grained sandstones in the eastern end [SS 4580 3962]. Immediately east of the fault buff banded fine- to medium-grained sandstones are exposed [SS 4619 3964] near the base of the formation, and buff current-bedded micaceous medium-grained sandstones with silty shales near the top [SS 4588 3945].

An old quarry [SS 4672 3940] 150 m SSW of Crowborough has exposed a fault trending 030° and inclined steeply NW. On its north-western side thickly bedded ferruginous-spotted micaceous medium-grained sandstones, siltstones and shales are folded to display dips of 70°–90°/210° and 5°/335°. Somewhat more silty and shaly strata east of the fault are tentatively regarded as Pilton Shales (pp.24–25). The valley trending SW and S near Crow-borough appears to follow a fault, to the east of which ferruginous-spotted fine- to medium-grained sandstones [SS 4689 3932] are cut off by a NNW–SSE dextral wrench fault passing just west of Darracott. This fault has displaced the formation 700 m southwards to the vicinity of a crossroads, Lobbthorn Stile [SS 4723 3874]. Farther east the Baggy Sandstones extend to Knowle in a fairly narrow outcrop disrupted by several minor faults. Debris of fine-grained sandstone and of ferruginous-spotted feldspathic medium-grained sandstone is common, and some fragments of shale and grey crinoidal limestone occur. Rubbly buff and yellow medium-grained sandstones crop out [SS 4888 3883] in a lane west of the River Caen. On the eastern side of the river, a quarry [SS 4924 3872] alongside the A361 road at Knowle shows fairly massive fine-grained sandstones, and a septic tank [SS 4947 3863] in the village has been excavated in buff sandstones. The southern boundary of the formation in the neighbourhood of Knowle is poorly defined but appears to be a fault trending ESE through the southern part of the village and throwing out spring water [SS 4948 3825] near the edge of the district. EAE

Chapter 3 Transition Group: Pilton Shales

General account

The name Pilton Beds was introduced by Hall (1867) and adopted by most subsequent writers. It is now replaced by Pilton Shales, a lithological connotation being preferred where possible. The deltaic conditions which prevailed over north Devon in Baggy Sandstones times gave way to those of a shallow sea, and there is continuity between Baggy Sandstones and neritic Pilton Shales which renders arbitrary the selection of a boundary (p. 10). The corresponding rocks of the Transition Group between Boscastle and Launceston, on the southern side of the Carboniferous synclinorium, have yielded fossils of the Upper Devonian Wocklumeria Stufe and of the Lower Carboniferous Gattendorfia Stufe (Freshney and others, 1972; McKeown and others, 1973), and comprise slates with calcareous and siliceous bands, nodules and lenses in which originally-calcareous patches have commonly become decalcified into earthy brown friable material. The lithologies indicate deposition in a quiet shelf sea, somewhat deeper in latest Devonian times than in earliest Carboniferous. In contrast, the brachiopod-bivalve assemblage of the lower part of the Pilton Shales indicates deltaic or near-shore deposition, and the trilobite-goniatite fauna of the higher beds reflects deepening of the water, probably coinciding with transgression elsewhere over the Old Red Sandstone continent.

The Pilton Shales crop out on the coast from the southern side of the Baggy Point promontory to the northern end of Saunton Sands, and on both banks of the estuary in the neighbourhood of Chivenor Airfield. Inland they underlie Croyde, Saunton, Braunton and most of the Braunton Burrows and Braunton Great Field. The rocks are grey shales with bands and lenses of fossiliferous limestone, and with thin sandstones and calcareous sandstones both scattered sporadically throughout the sequence and locally aggregated to form sandy units whose resistance to weathering is reflected in the topography. Their petrography and mineralogy are referred to on pp. 11–15. There is some suggestion in the coastal outcrops that the arenaceous content of the rocks decreases in the upper part of the formation. Whether or not this is true in general, it does not permit the mapping of major lithological subdivisions. Although sandy units may be picked out within the Pilton Shales, and followed across country, once they become appreciably displaced by faulting correlation is impossible. The thickness of the formation is not known, but rough calculations based on the width of outcrop and the structural pattern point to a tentative estimate of about 500 m.

The most comprehensive palaeontological study of the beds is contained in Whidborne's (1896–1907) monograph, which supplemented the work of Phillips (1841); more recently, the trilobites and certain of the brachiopods have been reinvestigated (Reed, 1943, 1944; Goldring, 1955a, b, 1957). A number of workers have published the results of collecting, mainly in the 19th century (e.g. De la Beche, 1839) but including Prentice (1960a, b). Paul (1937) discussed correlation of the Pilton Shales with the Etroeungt Beds of the Ardennes and other European strata near the Devonian–Carboniferous boundary; he noted the general similarity between the uppermost Pilton Shales and strata immediately overlying Etroeungt Beds at Ratingen, north of Cologne.

Goldring (1955a, 1957, 1971) has contributed greatly to our knowledge of the biostratigraphy of the Pilton Shales. He confirmed that they span the Devonian–Carboniferous boundary, remarking (1971) on the similarity between a fauna he collected from near the top of the Baggy Beds to that from the well known Laticosta Cave [SS 4270 4004] near the base of the Pilton Shales. On faunal criteria he recognised three divisions of which he correlated the lowest, Pilton A, with the Wocklumeria Stufe, considering the remainder to correspond to the succeeding Gattendorfia Stufe and possibly in the upper parts to the Ammonellipsites Stufe. Pilton A is characterised by the trilobite Phacops (Omegops) accipitrinus accipitrinus and yields a number of species of the productellid genera Whidbornella, Steinhagella and Hamlingella which appear to become extinct in Europe before the Carboniferous. Using the productellids and other brachiopod species Goldring subdivided Pilton A into three units, A1–3. Austin and others (1970) stated that the conodont faunas of the lowest Pilton Shales, along with those of the uppermost Baggy Sandstones, appeared to be younger than the uppermost Wocklumeria faunas of the German type section but older than the uppermost Famennian faunas of Belgium.

The Devonian part of the Pilton Shales yields a varied fauna (Plate 2) in which the brachiopods are particularly diverse and abundant. The Strophomenida are well represented and include Leptaena sp.,productellids, Mesoplica praelonga, Chonetes sauntonensis and Plicochonetes margaritaceus; Spiriferida are varied, among them being Athyris concentrica which is locally abundant, Crurithyris unguiculus and spiriferaceans which are common throughout and include Cyrtospinier verneuili and other cyrtospiriferids; rhynchonellaceans occur commonly, among them being Centrorhynchus letiensis and Sinotectirostrum? laticosta, and orthaceans such as Aulacella interlineata are also met with; inarticulate brachiopods are rare. Among the molluscs, bivalves are numerous, including species of Palaeoneilo, Prothyris and several pteriacean and pectinacean genera; gastropods are much less common and are mostly indifferently preserved; 'Orthoceras'is occasionally found. Crinoid debris is abundant and a few unfragmented crinoids and asterozoans occur locally; bryozoans include species of Fenestella and other, poorly-known forms; Pleurodictyum and small solitary corals are present, but rare; Phacops and ostracods are fairly common, the latter particularly so at certain levels. Dineley and Rhodes (1956) recorded conodonts from Saunton, Gnathodus and Spathognathodus as common and Lonchodus and Ozarkodina as rare, and Austin and others (1970) noted forms including Pseudopolygnathus vogesi and Polygnathus sp.from the lowest Pilton Shales.

Goldring's Pilton B is characterised by the trilobites Cyrtosymbole (Macrobole), C. (Waribole) and Brachymetopus, Pilton C by Piltonia and Brachymetopus, Cyrtosymbole apparently being absent. Goldring has listed the following fossils from the Carboniferous part of the Pilton Shales: Cyrtosymbole (Cyrtosymbole) hicksii, C. (Macrobole) hercules, C. (M.) duodecimae, C. (M.) mulesi, C. (Waribole) porteri, Brachymetopus woodwardii, Piltonia salteri and a few goniatites from outside the present district, viz. Gattendorfia crassa and Imitoceras sp.Prentice (1960a) considered unfossiliferous silty shales above the highest fossiliferous strata to constitute a fourth division of the Pilton Shales which he named Pilton D.

Folding within the Pilton Shales is well displayed on the foreshore at Down End. Folds there are generally fairly tight with axes trending E–W and near-vertical axial planes; locally puckered limbs dip steeply to north and south, the northern limbs commonly being the longer. Amplitudes and wavelengths of 4 to 5 m are typical. Slight overturning to the south occurs rarely. It seems probable that these exposures reflect the general pattern of folding and that the Pilton Shales are much more crumpled than the underlying, more competent Baggy Sandstones, although not usually intensely contorted or overthrust. This is confirmed by mapping farther east in the Barnstaple area, where folds of a few hundred metres wave length are upright and fairly close. EAE, DEB

Details

Coastal sections

It is probable that the entire Pilton Shales coastal outcrop, from its junction with the Baggy Sandstones in the north to the northern end of Braunton Burrows in the south, is of Devonian (Pilton A) age, almost all of it falling in Goldring's two lowest subdivisions A1 and A2. A brief palaeontological investigation conducted in connection with the present survey yielded the fauna detailed in Appendix 1, from 25 coastal localities (A–Y). DEB

Croyde Bay

The basal strata of the Pilton Shales–shales, silty shales, thin fine-grained sandstones and thin limestones which weather brown and friable–are exposed along the coast from 500 m WNW of Baggy House [SS 4289 3998] to the northern edge of Croyde Sand. They are disrupted to only a minor extent by small-scale faulting. Immediately above Laticosta Cave [SS 4270 4004] such strata dip 60°/205°. False bedding in the sandstones hereabouts shows the sequence to be right way up. Weathered brown calcareous beds in the cave have yielded Sinotectirostrum? laticosta, Hamlingella goergesi and Athyris sp.No collection was made from the cave itself during the present survey but localities A, B and C (Appendix 1) are situated among the reefs close to the mouth of the cave. At 200 m W of Baggy House shales with thin sandstones and limestones are thrown into generally open folds whose axes trend WNW [SS 4267 3998].

Towards Croyde Sand similar strata dip mainly between S and SSW at up to 75° and are locally vertical. They show ripple marking and traces of load casts. Open folds are present on the foreshore and in low cliffs [SS 4295 3965]; [SS 4323 3961] with axes trending around 105°. Near the old limekiln [SS 4341 3954] shales with thin sandstones are overfolded; axial planes and both normal and inverted limbs dip steeply to slightly W of S [SS 4335 3955]. Nearby [SS 4339 3951], these beds dip 60°/215° and are cut by cleavage inclined 85°/360°.

Down End to Saunton

Slightly papery shales with thin'sandstones and limestones at the southern end of Croyde Sand [SS 4339 3882] lie at the northern end of a well-exposed dip section of folded strata extending round the promontory of Down End (Plate 5). The folding is commonly close (p. 93), but only rarely overturned, and there is constant repetition of strata. Unusually thick sandstones, or groups of several strong beds, occur locally, and either of these occurrences might give rise to a topographic feature inland. Thus a hard massive sandstone up to 1.2 m thick is displaced 1.8 m by an ENE-trending fault at [SS 4318 3870], four massive knobbly sandstones each 0.3 to 0.6 m thick dip steeply S at [SS 4324 3867] and three 0.6-m sandstones lie in tight upright folds at [SS 4307 3850]. Thin weathered bands [SS 4318 3860]; [SS 4314 3832]; [SS 4322 3824] were considered by Goldring (1957) to be tuffaceous. The section is cut by generally small NW-trending faults. Sheared and broken quartz veins occur in the neighbourhood of a 0.6-m sandstone band running slightly N of W along the floor of a small cove [SS 4314 3837].

The coast along the southern side of Saunton Down affords a strike section in shales with thin fine-grained sandstones and fossiliferous limestones. Ripple lineation is common. Locally the sandstones carry traces of load casts and in places sandstone boudins are present. The section is cut by several small faults with attendant distortion of nearby strata. A long E–W gullet [SS 4347 3804], cut along a fault line, exposes small-scale tight over-folding and shearing and twisted ovoids of sandstone. At the northern edge of Saunton Sands boudins of hard quartzitic sand stone in shales are associated with faulting [SS 4420 3780]. Silty nodules occur within the shales [SS 4439 3776] and thin sandstones nearby and farther east [SS 4447 3773] exhibit traces of load casts.

Inland sections

Croyde Bay to Braunton

Scattered exposures of silty shales occur in the roadsides around Croyde Bay and Croyde. Two boreholes about [SS 4394 3882] at the NALGO Holiday Centre reached Pilton Shales beneath 14.5 and 18.3 m of drift (p. 110). No detailed records of the strata exist but about 33 and 36 m of shales with harder bands were penetrated.

Rubbly shales with sandstones are abundantly exposed alongside the road skirting Down End and also on the steep heathy southern slopes of Saunton Down. Debris on the higher ground suggests the presence of some stronger beds of sandstone or calcareous sandstone striking E–W; two overgrown pits [SS 4446 3852]; [SS 4447 3850] are small but disproportionately deep, suggesting the working of fairly thin beds of stone. Near Saunton Sands Hotel are: a strike section mainly in shales [SS 4465 3785] from which Goldring (1957) obtained Hamlmgella goergesi; another in shales with sandstones and fossiliferous limy bands [SS 4452 3806]; grey shales containing silty ferruginous nodules and inclined steeply NNE [SS 4487 3795]; and an old quarry in shales with sandstones [SS 4480 3786]. A borehole, whose exact site is uncertain but which is described as at Saunton Sands Hotel at 61 m above OD, and therefore was probably sunk on the north side of the road, yielded the following (summarised) log:

The limestones and calcareous sandstones were commonly fossiliferous and the strata dipped at 55°–65°.

At Saunton, vertical fine-grained sandstones with fragments of shells and crinoids strike 100° [SS 4551 3782], shales with fine-grained sandstones and fossiliferous limestones are folded along axes trending 100° [SS 4558 3769], and grey shales are crumpled on E–W axes [SS 4584 3796]. About 1.6 km to the N, near the base of the formation at Forda [SS 457 392], shales, siltstones and fine-grained sandstones, some purple and brown in colour, are exposed in roadsides, farmyards and two old quarries [SS 4569 3907]; [SS 4579 3888]. Similar strata on the east side of a fault in a quarry [SS 4672 3940] near Crowborough may be Pilton Shales; they dip 60°–75° between NE and NNE, in contrast to the Baggy Sandstones west of the fault (p. 21).

Two boreholes at a pumping station [SS 4649 3748] sited on estuarine alluvium east of Saunton probably entered Pilton Shales at just below 8 m depth. Another, 3.2 km to the SSW in Braunton Burrows, proved 'hard blue rock' at 10.9 m.

Small exposures are common in the neighbourhood of Lobb [SS 473 380], where shales with micaceous siltstones and false-bedded sandstones strike just S of E and are vertical or dip steeply but right-way-up around 010°. A nearby disused quarry [SS 4742 3767], in silty shales and ripple-marked micaceous fine-grained and medium-grained sandstones with traces of thin limestones, displays right-way-up dips of around 30°/230°.

The valley of the River Caen cuts a dip section through the lower part of the Pilton Shales and shows the formation to contain a good deal of sandstone. Sporadic small exposures of shales, silty shales and thin fine-grained sandstones lie south of the probably faulted junction with the Baggy Sandstones, and massive fine-grained sandstones have been cut alongside the railway [SS 4897 3810] below The Castle earthwork. Shales and sandstones bordering the massive sandstones dip steeply NNE. A fault runs along the valley hereabouts and passes close to Buckland House [SS 484 378], where brown fine-grained sandstones strike parallel to the fracture and dip 20°/340° [SS 4852 3777]. East of the valley a fairly large disused quarry [SS 490 378] exposes fine-grained sandstones and shales dipping around 50° NNE but showing evidence of some minor tight to isoclinal overfolding. This group of sandy strata may be a displaced equivalent of the massive sandstones alongside the railway noted above. Farther east along the strike the pattern of folding within these beds is shown in roadside exposures [SS 4942 3769] by vertical dips, strike 090°, and inclination of 40°/360°.

To the south, on the outskirts of Braunton, a similar group of sandy strata trends slightly S of E from near Buckland Cross [SS 483 376] to Chapel Hill [SS 492 374]. Exposures on the west side of the valley alongside the railway [SS 4872 3750] show thin- and medium-bedded buff micaceous fine- and fine to medium-grained sandstones dipping 10°/240° and 50°/200° beneath steeply inclined shales with fine-grained sandstones to the south. Fine-grained sandstones with shales crop out in a laneside east of the valley [SS 4904 3745]–[SS 4910 3747] and are accompanied by some buff medium-grained sandstone on the high ground beside the chapel ruin [SS 4914 3733]. Shales with thin sandstones which overlie these sandy beds are well exposed for over 200 m along the A361 road below the south-western slopes of Chapel Hill and sporadically within Braunton village.

Braunton Down Quarry [SS 493 369] is heavily overgrown but appears to have been dug in folded sandstones with shales. Traces of calcareous sandstone occur, and the possible presence of beds sufficiently calcareous and in sufficient quantity to be burnt for lime is suggested by a nearby structure reminiscent of a derelict limekiln [SS 4902 3679]. Rogers (1926) recorded the fish Holoptychius from a quarry at Braunton Down and during the present survey the following were collected from neighbouring track exposures [SS 4902 3678]: Pleurodictyum sp., Fenestella plebeia, Aulacella interlineata, Steinhagella steinhagei, indeterminable spiriferaceans and rhynchonellaceans, Pleurotomaria' sp., Leptodesma sp.and crinoid columnals. Goldring (1955a) recorded Phacops (O.) accipitrinus accipitrinus nearby [SS 4902 3678] and therefore referred the beds here to the Wocklumeria Stufe; during the present survey the following were recorded from [SS 4911 3677], close to Goldring's locality: Fenestella umbrosa?, F. cf. plebeia, ?Rhabdomeson gracile, an acanthocladiid bryozoan, A. interlineata, Leptaena sp.,Chonetes sauntonensis, ?Plicochonetes margaritaceus, indeterminable athyridids and productellids?, Pleurotomaria'aff. aspera, indeterminable trilobite fragments and crinoid columnals.

Several shallow boreholes through alluvial deposits between Braunton and Velator (p. 112) have proved Pilton Shales at depths between 5.3 and 7.7 m but three failed to reach solid deposits at 9.3 to 10.8 m. S-dipping shales and silty shales in a roadside [SS 4980 3570] 0.4 km E of Wrafton have yielded Brachymetopus woodwardii and Piltonia salteri to Goldring who (1970) regarded strata bearing these forms as belonging to either the Gattendorfia Stufe or the Ammonellipsites Stufe of the Lower Carboniferous.

The Taw–Torridge estuary

Vertical and steeply dipping shales, mudstones and fine-grained sandstones exposed at low tide at the mouth of the River Caen [SS 4835 3344]; [SS 4858 3369] strike slightly S of E and are probably referable to the Pilton Shales. On the opposite (southern) bank of the estuary [SS 496 334] similar strata dip southwards at 40°–80°. Shales and mudstones with thin, rust-spotted, locally slightly cherty, fine-grained sandstones [SS 491 333] show some inversion and faulting but pass southwards beneath strata which have yielded Posidonia becheri (p. 99). EAE

Chapter 4 Carboniferous

General account

Lower Carboniferous

The lower part of the Lower Carboniferous sequence falls within the Pilton Shales and has been described in the previous chapter. The muds, silts and sands continued to be deposited in the upper part of the sequence. The shallow marine waters remained generally still, supplies of sandy detritus diminished as erosion of the South Wales landscape continued, and black muds accumulated. With the grey and black shales which have formed from these muds are associated thin cherts and limestones. To the east of the present district thick developments of chert, and large lenticular limestone masses of a size sufficient to have justified quarrying in the past, occur within these shales.

Hinde and Fox (1895) considered the cherts to indicate extremely slow and long-continued tranquil deposition and consolidation. They noted that radiolarian ooze is at present characteristic only of great oceanic depths, and deduced that Lower Carboniferous cherts and intimately associated shales were laid down in the deep water of the open sea at some distance from a shoreline. This conflicted with the general view of the time that the deposits were of shallow-water origin. Subsequently Dixon and Vaughan (1911) postulated a shallow-water, possibly lagoonal, origin for cherts in the Millstone Grit of South Wales; these beds are locally conspicuously wedge bedded, yet contain only small quantities of extremely fine-grained detrital sediment. Comparable shallow-water radiolarian rocks are known from the Precambrian of Brittany (Cayeux, 1929), the Carboniferous of Germany (Wilckens, 1908), the Upper Palaeozoic of New South Wales (David and Pittman, 1899), the Jurassic of California (Reed, 1933) and the Pliocene of Rotti, in the Timor Strait (Tan Sin Hok, 1927). It seems likely that the Lower Carboniferous cherts of north Devon originated in quiescent shallow seas with lagoons, transitional between the shallow waters of Pilton Shales times and the deeper-water, but far from abyssal, environment of the Upper Carboniferous. The associated limestones accord well with such an environment.

Exact correlation is not possible, but it seems probable that the Lower Carboniferous beds of Isley Marsh (p. 35) correspond to the Meldon Chert Formation (P_1c–P_2a), Edmonds and others, 1968; McKeown and others, 1973), plus perhaps some of the underlying shales and sandstones, of the southern side of the basin of deposition. Prentice (1960a) regarded them as extending down possibly to the base of the P₁ Zone. EAE

Upper Carboniferous

Crackington Formation

The Crackington Formation crops out around Instow and Appledore, in a narrow belt extending eastwards from the coast across the south-eastern part of the district, and over a large tract around Hartland. It includes (Freshney and Taylor, 1972; Edmonds, 1974) the Welcombe Measures of Ashwin (1958), the Welcombe Formation of Edmonds and others (1969), the Welcombe Beds of Moore (1968), the Instow Beds of Prentice (1960a), the Westward Ho! Formation of De Raaf and others (1965) and the Appledore Formation of Money (1966).

The basal beds, seen only at Instow [SS 473 303], are shales, silty shales, siltstones and fine-grained, generally turbiditic, sandstones. They accumulated in water deeper than that of Lower Carboniferous times and spasmodically invaded by turbid bottom-hugging currents thick with sediment, and they are now disposed in open or close upright folds. The sandstones, commonly graded and bearing ripple marks and sole structures, are locally aggregated into mappable groups which have on occasion been quarried. The Instow Fish Bed (p. 28) has been well known since the work of Hall (1876), Hinde and Fox (1895), Arber (1904) and Rogers and Arber (1904). It equates with the Gastrioceras listeri horizon of the Westphalian A.

The remainder of the formation consists predominantly of thinly bedded sandstones (beds up to about 0.3 m thick), with varying proportions of siltstones and dark grey to black shales. Medium-bedded sandstones (beds 0.3 to 1.0 m thick) occur sporadically. Thickly bedded sandstones (beds over 1.0 m thick) are rare. These definitions of bed thickness apply where reference is made to thinly bedded, medium-bedded and thickly bedded sandstones throughout this chapter. The Crackington Formation sandstones are mainly medium grey, hard, tough, fine-grained greywackes, exhibiting a great variety of sole markings. Load casts and flute casts are most common on the bases of the thicker homogeneous sandstones, whilst small prod, bounce and groove casts are found on the thinner sandstones and siltstones. Measurements of the directions of orientation of sole markings such as flute and groove casts (Ashwin, 1957, 1958; Moore, 1968) indicate that the flow of the depositional currents was predominantly from the west ((Plate 7); (Figure 5)). Thinly bedded silty sandstones with beds less than 0.1 m thick, and some very dark grey silty sandstones, are commonly graded. However, graded bedding appears to be absent from many of the sandstones and these are described as homogeneous. Graded bedding is one of the features most commonly associated with deposition of greywackes from turbidity currents. Its apparent absence from the homogeneous sandstones may result from the fine-grained texture of those rocks concealing any grading present.

The upper surfaces of the sandstones commonly show ripple marks, and ripple-drift cross-bedding is seen in cross sections. Rare small sand volcanoes occur on the upper surfaces of homogeneous sandstones (Burne, 1970).

The beds between the sandstones are usually laminated siltstones, and sequences of laminated siltstone and shale several metres thick occur at intervals. The siltstones vary considerably in thickness but seldom exceed the adjacent sandstones. Ripple-drift cross-bedding is common. These siltstones are probably in part the reworked fine-grained deposit from underflows or turbidity currents and partly represent fine-grained background sedimentation. The most distinctive of the shale horizons are the dark grey to black shales which contain calcareous nodules. Both the shales and the nodules may be fossiliferous. The nodular shales are commonly pyritous and characterised by a yellow-weathering crust and a sulphurous smell. Some nodules have a coating of gypsum. The nodules vary in type; some are diagenetic calcareous concretions in which lamination is continuous between the enclosing shale and the nodule, while others have a more homogeneous granular texture and may be concretions formed on the sea floor. The nodular shales were given local names by Freshney and Taylor (1972).

Slumped beds of various thicknesses are found at intervals in the succession. They comprise dark homogeneous silty mudstone containing balled up and contorted masses of sandstone and more thinly bedded silty sediments. A major development at Gallant Rock [SS 3170 2519], probably made up of two separate slumps, is 27 m thick. Slumped beds appear to have been laid down by flows of more or less liquefied mud moving across the sea floor carrying masses of semi-consolidated sediment. Burne (1969, 1970) preferred the name 'slurried bed' and suggested that some beds of similar appearance might have been formed by the rapid dewatering and disruption of fluidised sands following the deposition of overlying sediment.

The nodular shales persist laterally and can be used as marker horizons over distances of several kilometres. Other shales and the slumped beds are of value for local correlation, but the thick sandstones are too impersistent even for this. The Crackington Formation in the Hartland area (Figure 6) and (Figure 11) includes sediments of Namurian (R₂ and G₁) and basal Westphalian age. Fossils, including goniatites, bivalves, fish and conodonts, are mainly confined to the distinctive nodular shales; plant debris is more widely spread through the succession. The nodular shales can be correlated with the marine bands of coalfield areas, and represent marine incursions into a semi-marine environment. On the evidence of the identified goniatite horizons the Crackington Formation is about twice the thickness of the equivalent succession in South Wales and evidently represents deposition farther off-shore.

The most complete and continuous section is found along the coast between Clovelly [SS 3180 2485] and Windbury Point [SS 2872 2682] (Figure 7). The lowest fossiliferous shale, the Clovelly Court Shale, has yielded a goniatite fauna correlated with the Gastrioceras? sigma horizon of the Marsdenian (R₂) Stage. The nodular Skittering Rock Shale and Deer Park Shale are correlated with the horizons of Gastrioceras cancellatum and Gastrioceras cumbriense respectively, and indicate that a complete Yeadonian (G₁) succession is present. These beds are followed by the prominently nodular Embury Shale, correlated with the Gastrioceras subcrenatum horizon and marking the boundary between Namurian and Westphalian. This horizon, together with its inland exposures, is notable in containing three new species of Gastrioceras. The Gull Rock Shale also forms a distinctive nodular band, and yields an abundant fauna including Gastrioceras listeri and associated goniatites such as Gastrioceras circumnodosum. Matthews and Moore (1967) obtained conodonts from this shale. Abundant anthracoceratid goniatites occur characteristically at the base of the Hartland Quay Shale throughout the area. This shale is well exposed at Windbury Point. In other parts of the area the same horizon has also yielded Gastrioceras amaliae. This horizon can be equated with the Tonge's Marine Band of Lancashire and possibly the Margam Marine Band of South Wales (Ramsbottom and Calver, 1962). This succession is the only one known at present in the Cornish and Devonshire Carboniferous which includes a complete sequence between the upper Namurian and lower Westphalian. No G₁ goniatites have been recorded elsewhere in south-west England.

The Namurian succession west of Clovelly is faulted out east of the village, only the highest beds being present. On the west coast to the south of Hartland Point the lowest Crackington Formation strata found are at Upright Cliff [SS 2290 2705]; sheared Gull Rock Shale, with nodules containing G. circumnodosum, lies in the core of a major anticline.

The boundary between the Crackington Formation and the overlying Bude Formation is mapped as the base of the lowest of the thick massive sandstones characteristic of the Bude Formation. Freshney and Taylor (1972) considered that in the coastal sections west of Clovelly and south of Hartland Point the marker horizon closest to the junction was their Hartland Quay Shale, yielding G. amaliae; however they recognised that there was some local deviation, for example the lowest Bude Formation sandstones at Cow and Calf [SS 2272 2713] actually lie below the Hartland Quay Shale.

Folding causes multiple repetition of the Hartland Quay Shale and the Longpeak Shale (p. 35) on the coast to the south of Upright Cliff. Although parts of this succession lie above the Hartland Quay Shale, sandstones of Bude Formation type appear to be absent and the beds are included in the Crackington Formation. To the south, between Hartland Quay [SS 2225 2478] and Childspit Beach [SS 2240 2425], up to 128 m of strata are exposed below the Hartland Quay Shale. The fossiliferous Embury Shale and Gull Rock Shale crop out in stream sections at a number of localities around Hartland and define the northern limb of a major anticlinorial structure extending 6 km W from the coast west of Clovelly.

The narrow outcrop of Crackington Formation extending eastwards from the coast near Babbacombe Mouth [SS 397 260] to the edge of the district south of Gammaton Moor [SS 491 244] is bounded to the north by a northerlydownthrowing major normal fault (p. 92). The beds are Westphalian, G. listeri having been recorded from three localities in the cliff section at Cockington Cliff [SS 398 261] and in a road cutting 1 km S of Bideford Bridge [SS 453 255].

The area around Appledore and Instow is underlain by the Instow Beds of Prentice (1960a). The Instow Fish Bed has yielded G. listeri and fish fragments at Instow and at other places in the south of the district. In Hubbastone Quarry [SS 464 298], Appledore, a fauna characteristic of the G. amaliae horizon was identified. Prentice (1960a, p. 267) classified the beds in this quarry as Northam Beds (Westward Ho! Formation of De Raaf and others, 1965) resting on Instow Beds. They are now regarded as Crackington Formation. The rocks of the Instow–Appledore–Westward Ho! area all appear to young southwards, and the G. amaliae horizon may lie about 150 m above the Instow Fish Bed (G. listeri horizon) as in the Hartland area. However, the G. amaliae fauna, present here within a sequence of turbidite Crackington Formation strata, also occurs in Cycle 6 of the Bideford Formation (see below) only 5 km to the southwest.

Thus in the Appledore–Instow area we have a southerly-younging Crackington Formation succession up to the G. amaliae horizon (which lies near the base of the Bude Formation in the Hartland area), and to the south an apparently southerly-younging Crackington Formation and Bideford Formation sequence containing G. amaliae near its top (Figure 8). Possibly the tract of country north of Westward Ho!, where exposures are rare, contains a major east–west fault separating these two successions, or more likely the facies change to Bideford Formation is very sudden towards the south and at least one major syncline and one major anticline intervene. RTT, BJW, EAE

Bideford Formation

The Bideford Formation comprises the Northam and Abbotsham formations (Bideford Group) of De Raaf and others (1965), and thus the Abbotsham Beds and part of the Northam Beds of Prentice (1960b). It lies conformably between De Raaf, Reading and Walker's Westward Ho! Formation, now included in the Crackington Formation, and the overlying Bude Formation (Greencliff Beds of Prentice, 1960b), though there is extensive faulting near the latter junction at the coast. The Bideford Formation occupies an east–west strip 2 to 3 km wide across the southeastern part of the district through Bideford. Farther east, in the Barnstaple district, it passes laterally into typical Bude Formation strata.

The characteristic topography is a series of steep-sided east–west sandstone ridges, displaced by wrench-faults, with intervening low-lying areas underlain by shales, siltstones and silty and sandy mudstones. Thus the formation can be readily traced across country and the sandstones accurately mapped. Names of sandstones proposed by De Raaf and others are adopted here.

Lithologically the Bideford Formation is the most reminiscent of Coal Measures sequences of any of the Culm subdivisions, and is considered to have been deposited in a deltaic environment (Prentice, 1960b; Elliott, 1976). De Raaf and others (1965) distinguished three sedimentation cycles in their Northam Formation, and six in their overlying Abbotsham Formation, in the coastal section. These cycles may be regarded as Cycles 1 to 9 of the Bideford Formation (Figure 9). Each passes upwards from black mudstone through grey silty and sandy mudstones, siltstones and sandstones into a major sandstone, though many are incomplete. De Raaf and others identified eleven facies comprising the cycles, and supported Prentice's (1960b) conclusion that the deposits were deltaic. The cycles were taken to represent successive south-easterly advances of a delta-fronted shoreline into a moderately deep basin, alternating with periods when basin conditions returned.

During the recent survey G. amaliae was recovered from a mudstone [SS 4115 2815] at the base of Cycle 6. Dr M. A. Calver comments: 'The identification of G. amaliae places this horizon as the equivalent of Tonge's Marine Band of Lancashire and the Plasshofsbank marine horizon of the Ruhr (Ramsbottom and Calver, 1962). The stratigraphical position is thus above that of the G. listeri Marine Band and lies near the top of the 'Anthraconaia' lenisulcata Zone as defined in the Pennine Province. It is not possible to correlate this horizon in terms of the South Wales or Bristol sequences, as G. amaliae has not been recorded from these areas. However, the equivalent is possibly the Margam Marine Band of South Wales'. This means that the Bideford Formation is younger than Reading (1965) thought and refutes the observation that bivalves from Cycle 4 were representative of an horizon only a short distance below the G. subcrenatum Marine Band. The Instow Fish Bed fauna (G. listeri Marine Band) lies an estimated 1200 m below the G. amaliae horizon in this area; in the Hartland–Bude area and in the Appledore–Instow area these two horizons are only about 150 m apart. It is suggested that local thickening of the Bideford Formation accounts for this difference (Freshney and Taylor, 1972).

The presence of lenisulcata Zone bivalves (Simpson, 1933) at the base of Cycle 8 is confirmed (p. 47), and new finds of bivalves within Cycle 3 are the lowest fossils yet recorded in the formation. BJW

Bude Formation

The Bude Formation occupies the coastal area to the immediate south and east of Hartland Point (Figure 6) and (Figure 11), a fault-bounded tract east of Longpeak [SS 2212 2304], and much of the south-eastern portion of the district inland from the coast between Clovelly [SS 318 248] and Greencliff [SS 410 268]. It includes Ussher's (1906) Middle Culm or Morchard-type Culm, which he placed below his Eggesford-type Culm. Work by Burne and Moore (1971) and Freshney and Taylor (1972) has shown that the Bude Formation lies at the top of the Westphalian succession, overlying the Crackington Formation (the equivalent of Ussher's Eggesford-type Culm in the upper parts). Moore (1968) called these rocks around Hartland Point the Hartland Beds. The Greencliff Beds of Prentice (1960b) are also included within the Bude Formation, as is the major part of his Cockington Beds.

Thinly-bedded and medium-bedded greywackes reminiscent of the Crackington Formation contain distinctive coarser grained sandstones usually more than one metre thick. Weathering of these sandstones produces friable, pitted and honeycombed surfaces. The weathered sandstones are usually pale buff, but some at Hartland Point and Longpeak are stained reddish brown and purple, possibly reflecting an association with faulting. The characteristic friable weathering may result from the nature of the matrix, as may a greater degree of porosity which has allowed the penetration of staining fluids.

Channelling of thick sandstones into underlying beds is also typical of the Bude Formation (Figure 10). These channels may show erosive features at the base or contain shale-flake conglomerates. The orientation of the channels varies. Those at Hartland Point appear to run east–west, while other examples are aligned approximately north–south.

In the upper part of the succession at Hartland Point a number of thick shale and mudstone horizons are developed. Slumped beds in the Bude Formation are litho-logically similar to those in the Crackington Formation but occur more commonly. They are usually more than 3 m thick, the thickest exceeding 22 m. There are indications that slumped beds formed by local mobilisation of horizons of shale or mudstone, together with included or overlying sandstones. Beds beneath the slumped strata appear only slightly disturbed.

As in the Crackington Formation, the value of the various lithologies for correlation is variable. The typical sandstones are impersistent and of little use. The slumped beds vary in lateral persistence; one can be traced for about 2.5 km along the strike east of Shipload Bay [SS 246 275], and those around Hartland Point extend over at least 3 km². The thick shales around Hartland Point are fairly persistent, although a thick group of sandstones on the west side of Shipload Bay is partly the lateral equivalent of siltstones and shales ((Figure 11), columns 3 and 4). The change takes place over a distance of about 1 km and may represent a broad sandstone-filled channel or washout in the shales. A persistent shale horizon with calcareous nodules, the Longpeak Shale of Freshney and Taylor (1972), named after the occurrence at Longpeak Beach, lies some 30 m above the Hartland Quay Shale and has yielded pectinoid bivalves, fish remains and mollusc spat, but no diagnostic adult goniatites.

The Bude Formation strata immediately south of the Bideford Formation outcrop were named Greencliff Beds by Prentice (1960b). They young generally southwards, apparently succeed the Bideford Formation, and are composed of dark grey shales and pale grey siltstones, with grey sandstones of Bude Formation aspect. These beds have not yielded fossils, but a belt of siltstones and dark grey shales with sandstones immediately to the south of the Bideford Formation has been traced eastwards into the adjacent Chulmleigh (309) and Barnstaple (293) districts. Near the local base of the Bude Formation (in Prentice's Greencliff Beds) are the well-known 'culm' seams. There appear to be two main seams, one of soft anthracite, about 1 m above the Cornborough Sandstone (p. 47) on the coast, the other, the 'paint' seam (p. 116), to the south and apparently some 100 m higher in the succession. However, evidence from workings at East-the-Water (p.116), which ceased production in 1969, suggests that there are four anthracite seams in that area; possibly three of these are faulted out in the coastal section, or the single seam may split eastwards into four.

Prentice's Cockington Beds, lying south of his Greencliff Beds and conformably succeeding them, include thickly bedded sandstones and slumped horizons. They are now mostly included in the Bude Formation, which extends south to the major northerly-downthrowing normal fault which is the northern boundary of the Crackington Formation outcrop trending eastwards from the coast near Babbacombe Mouth. South of this Crackington Formation tract the Bude Formation (Prentice's Cockington Beds) is of typical aspect and has yielded G. amalzae from four horizons in Westacott Cliff [SS 396 258], immediately south of the Sticklepath Fault Zone. Farther southwards and westwards towards Clovelly the succession is commonly strongly folded and faulted, and obscured by landslip and the Permian outlier at Portledge [SS 394 248]; it youngs generally southwards, up to an horizon some 150 m above the Longpeak Shale. BJW, RTT

Petrography and mineralogy

Arenaceous rocks of the Crackington Formation commonly appear more resistant to weathering than similar beds in the Bude Formation. Modal information gives some indication as to why this may be so. On the other hand preliminary analytical work failed to reveal significant differences between three argillaceous horizons.

Arenaceous rocks examined in thin section fall into two groups: siltstones with a mean grain size of up to 0.06 mm, and fine-grained sandstones with a mean grain size between 0.06 and 0.08 mm. The siltstones contain a higher proportion of matrix, but the actual components in both groups are similar. Subangular and subrounded, strained, quartz grains are generally the main constituent. Lithic fragments include chert, greywacke, siltstone, volcanic and plutonic igneous material (mainly silica-saturated), quartzite, quartz-chlorite-schist and quartzsericite-schist, feldspar, and detrital muscovite and chlorite. All are set in a matrix composed of detrital, authigenic and metamorphic mica and chlorite, clay constituents and opaque oxide ((E38118), (E42672), (E42673), (E42674), (E42675), (E42676), (E42677), (E42678), (E42679), (E42680) (E42681), (E42682), (E42683), (E42684), (E42685)). Accessory detrital pale olive-green tourmaline is common ((E42672), (E42673), (E42679), (E42681)) and zircon (E42675) and biotite (E42675) were noted. Many samples contain wisps of carbonaceous material ((E42675), (E42676), (E42678), (E42684), (E42685)). Secondary carbonate (calcite) is also commonly present in the matrix, though rarely in amounts greater than 1 per cent; two samples ((E42677), (E42678)) from Wood Rock [SS 3143 2550]; [SS 3144 2552] contain respectively 8 and 13 per cent (see p. 34). Some of the opaque oxide is detrital ilmenite/magnetite, but most occurs as limonite alteration of these grains and as an irregularly dispersed secondary limonite-coating on quartz and various clay-grade constituents.

Sporadic suturing of contacts between quartz grains and out-growths on the grains may be evidence of diagenesis, whereas the growth of secondary chlorite and muscovite from original matrix constituents is probably due in part to low-grade regional metamorphism. However, there is no sign of deformation of the original sedimentary bedding, and considered purely in terms of the compositional classification of Pettijohn (1957) the rocks are greywackes.

Averaged modal analyses of 14 samples are shown in (Table 3). The results are set out irrespective of grain size. However, the compilation does suggest that the Bude Formation rocks contain on average 7 to 8 per cent more lithic fragments and are about 4 per cent poorer in clay-grade constituents. Combined data from the Bideford district and the Chulmleigh (309) district to the southeast reveal two sorts of greywacke; one has upwards of 15 per cent lithic fragments and is more common in the Bude Formation, the other typically has 10 per cent or less of lithic fragments and occurs in the Crackington Formation. Possibly a high content of lithic fragments, which include a good deal of unstable, feldspar-bearing igneous material, leads to low resistance to weathering. It is also possible that high and low contents of lithic fragments correlate with top-set and fore-set beds respectively.

Four samples of argillaceous rocks from the Clovelly-Hartland coast (Table 4) were powdered and examined by means of X-ray diffractometry using CuKχ radiation, 40 Kv, 30 mA and scanning over the range 3°–45°2θ. Two specimens ((MR32589), (MR32590)) came from the G. listeri horizon in the Crackington Formation, one (MR32591) from the ?G. cancellatum horizon, also in the Crackington Formation, and the fourth (MR32592) from low in the Bude Formation. Muscovite, chlorite and quartz make up about 90 per cent of the four samples, with albite and subordinate microcline the only other principal constituents. The Bude Formation sample contains in addition minor amounts of anatase and a little illite. Muscovite and chlorite are present in all samples in approximately equal amounts. The estimated quartz content is in the range 20 to 35 per cent, several per cent less than in nine samples of similar material from the Chulmleigh district. There is a similar difference in clastic quartz content of arenaceous rocks of the two districts, but it is not yet possible to say whether these differences reflect sampling bias or regional variation.

Regional metamorphism

Despite the well-preserved sedimentary texture of the greywackes, the limited degree of recrystallisation of matrix constituents may indicate regional metamorphism characteristic of the Barrovian chlorite I subzone (Reed, 1958). Classification in terms of facies is complicated by the lack of indicator minerals other than secondary chlorite, and by the presence in the Bude Formation shale sample of illite, a mineral which occurs with kaolinite in similar shales from the Chulmleigh district. Possibly the grade of metamorphism is variable, being in general transitional into the lower part of the greenschist facies.

With rocks of so low a metamorphic grade some variation is to be expected, particularly as there are marked contrasts in mean grain size. Thus the predominantly albitic composition of feldspars in the shales may indicate more recrystallisation than is evident in the arenaceous rocks. The latter certainly contain sodic plagioclase, but most of it is clearly detrital.

The present patchy distribution of calcite in the arenaceous rocks may reflect redistribution during metamorphism of matter contained originally in calcareous nodules or introduced during lithification. Equally, however, it could be due, at least in part, to post-metamorphic groundwater circulation. In Permo-Triassic times, for example, conditions probably favoured carbonate deposition from groundwater, in which case the mineral's distribution may be linked rather more closely with the evolution of a mature erosion surface than with the Variscan metamorphism. It has been noted that the habit of the calcite crystals suggests replacement of both detrital and matrix minerals, which tends to support this view. The proposition is open to test in that significantly higher carbonate contents might be expected to correlate with regions where the strata are reddened or where outlying patches of New Red Sandstone remain. JRH. RJM

Details

Lower Carboniferous

Lower Carboniferous strata crop out at only one locality in the district, on the W-facing shore of the Taw estuary at Isley Marsh [SS 491 331], between 600 and 900 m N of Lower Yelland. At low tide, and dependent on the• disposition of mud and shingle, sporadic exposures of dark grey shales, mudstones and siliceous mudstones may be seen to strike slightly north of west. The thickness of these beds is difficult to estimate because exposure is poor, their limits are ill defined and, by analogy with the Upper Carboniferous Crackington Formation rocks to the south, the strata are probably disposed in upright folds. A figure of 100 m is suggested.

Moore (1929) obtained 'Goniatites spiralis'from this locality, indicating the Goniatites granosus (P_2a) Zone of the upper Visean. Prentice (1960a) concluded that the 'G. spiralis'Bed marked the top of the Lower Carboniferous. He also collected Posidonia becheri from the foreshore here. EAE

Upper Carboniferous

Crackington Formation

Coastal sections

Brownspear Point [SS 2240 2355] to Hartland Quay [SS 2230 2475]

The strata in this section lie below the Hartland Quay Shale and are disposed in a broad anticlinorial structure. The lowest bed occurs in the cores of anticlines at Childspit Beach [SS 2246 2424] and below St Catherine's Tor [SS 2240 2412], some 124.5 m below the Hartland Quay Shale, and is a massive sandstone<span data-type="footnote">In the details of Chapter 4 of this Memoir, all sandstones are grey unless otherwise stated.</span> nearly 4 m thick. A group of thinly bedded dark grey silty sandstones with well developed graded bedding is found in the lower part of the succession at Screda Cove [SS 2234 2443], Childspit Beach and at the northern end of Speke's Mill Beach [SS 2255 2393]. The remainder of the sequence is thinly bedded and medium-bedded sandstones interbedded with siltstones. Ripple-drift cross-bedding is found on the surfaces of some sandstones, while a thin sandstone on the southern promontory of Screda Cove has small sand volcanoes 25 to 30 mm across on its upper surface.

On the southern flank of the anticlinorium the Hartland Quay Shale is exposed in a partly overturned syncline in the cliff top [SS 2263 2374]. It is weathered to a medium grey colour and contains decalcified nodules, possibly with goniatite spat, scattered in a band about 0.6 m above the base. A narrow band of black shale about 0.3 m above the base yields abundant flattened anthracoceratid goniatites. Films of plant debris and fusiform coprolite pellets<span data-type="footnote">Mr Merriman reports that X-ray powder photography of fusiform pellets from the Hartland Quay Shale ((E38044), (X5539) at [SS 2242 2524] Indicates that they are formed mainly of carbonate and apatite with a little quartz.</span> are common in the rest of the shale. A thin decalcified nodular band occurs in the base of a thin sandstone 7.4 m below the Hartland Quay Shale.

On the northern flank of the anticlinorium the Hartland Quay Shale reappears in a syncline in the cliff north of Screda Point [SS 2234 2456]; it is inaccessible but fusiform coprolite pellets have been found in fallen debris. The shale is exposed on either side of a syncline at the northern end of Cowhouse Race [SS 2220 2479] at Hartland Quay, where it contains small nodules and fusiform pellets. It also occurs in the limbs of a syncline in the fault complex immediately north of Hartland Quay. In the southern limb the shale, black and grey laminated and 1.47 m thick with some silty bands, has been eroded to form a cave [SS 2242 2484]. Plant debris and coaly smears occur throughout the shale, coprolite pellets in the upper part and on the base of the overlying sandstone, and calcareous nodules with goniatite spat sporadically in the middle. Poorly preserved anthracoceratid goniatites have been obtained from a narrow band about 0.3 m above the base. A 0.15-m sandstone 2.21 m below the Hartland Quay Shale carries shale flakes and coprolite pellets concentrated in its lower part. About 5 m below the shale a 0.3-m black shale carries calcareous nodules. Calcareous concretions also occur in the overlying sandstone. This horizon may correlate with the decalcified nodular band found beneath the Hartland Quay Shale at the southern end of the anticlinorium.

Warren Beach, Hartland Quay, to the Abbey River [SS 2254 2568]

The Hartland Quay and Longpeak shales are repeated in a series of folds. The Longpeak Shale appears in the cliff top north of Hartland Quay and passes beneath a synclinorial structure to the north of the mouth of the Abbey River. This section includes strata stratigraphically equivalent to Bude Formation farther north but without typical Bude Formation sandstones.

The Hartland Quay Shale is exposed at beach level on the northern limb of the large open syncline at the southern end of Warren Beach [SS 2252 2487]. It is typical but without nodules. Coprolite pellets are present and a few poorly preserved crushed anthracoceratids have been obtained from a band about 0.3 m above the base. To the north the shale rises to the cliff top, undulates in the cliff face and descends again to beach level [SS 2248 2513]. It reappears at beach level in the hinges of anticlines at the northern end of Warren Beach [SS 2241 2523], where nodules are found and coprolite pellets are abundant on the base of the overlying sandstone, and again on Broad Beach [SS 2245 2535]. Development of fracture cleavage has destroyed any fossils.

The Longpeak Shale and the thick underlying sequence of siltstones lie in a syncline at the top of Warren Cliff [SS 2261 2520]. The shale reaches beach level in a syncline at the northern end of Warren Beach [SS 2247 2520] and undulates between beach level and cliff top across Broad Beach to Dyer's Lookout [SS 2237 2562]. The Longpeak Shale in this section is fairly typical. At the northern end of Broad Beach [SS 2238 2542] the uppermost 1.52 m of black shales with thin silty and sandy bands, contain a scatter of fusiform coprolite pellets and overlie 0.91 m of shales with a nodular horizon at their base. Beneath are 7.21 m of flaggy, laminated siltstones with sideritic bands and thin ripple-drift cross-bedded sandstones. The thin sandstones which elsewhere lie between the Longpeak Shale and the underlying siltstones are absent. The Longpeak Shale is also exposed on the foreshore and in the low cliff north of the mouth of the Abbey River in the cores of two anticlines. Caneyella sp.and mollusc spat were collected from below Dyer's Lookout [SS 2242 2558] and conodonts, including Hzndeodella sp.,from the anticline north of the Abbey River [SS 2259 2573].

Berry Cliff [SS 2256 2577] to Upright Cliff [SS 2290 2700]

Two small synclinoria centred on the mouths of the Blegberry Water [SS 2258 2593] and the Titchberry Water [SS 2284 2664] are separated by an anticlinorial rise south of Damehole Point [SS 2227 2629]. The strata are stratigraphically equivalent to the lower part of the Bude Formation north of Upright Cliff but lack typical Bude Formation sandstones.

Northwards from the outcrop of the Longpeak Shale in Berry Cliff younger beds come on in a sequence of sandstones and laminated siltstones. Locally horizons of shale and siltstone are up to 3 m thick. Some thinner sandstones are internally laminated or ripple-drift cross-bedded. This section extends 92 m above the Longpeak Shale to a slumped bed 3.35 m thick in a syncline at the mouth of the Blegberry Water ((Figure 11), column 3). Plant remains have been obtained at [SS 2260 2591].

To the north of the Blegberry Water successively older beds crop out and the Longpeak Shale reappears in a series of anticlines at the northern end of Blegberry Beach. It contains coprolite pellets and nodules, and also Caneyella sp.,mollusc spat and Hindeodella sp. [SS 2247 2619]. Between Damehole Point and Smooth-lands a thin black shale with coprolite pellets occurs above the Longpeak Shale [SS 2242 2637], while in the anticline at the western tip of Smoothlands [SS 2238 2646] the upper part of the Longpeak Shale bearing coprolites is exposed but not the nodular band. Some of the pellets here rest in the centre of the upper surface of button-shaped nodules of black, probably phosphatic, material to which they seem to have acted as nuclei.

To the north of Smoothlands some mineralisation is associated with jointing and brecciation in the fold hinges near to a small wrench fault. Quartz, yellow to brown-weathering rhombohedra of ankerite, a little pyrite and traces of chalcopyrite and green copper secondary minerals occur on joint surfaces. There is a good deal of red hematitic staining of the sandstones in the brecciated areas.

Between Smoothlands and Upright Cliff the second synclinorium shows a similar succession to that on Blegberry Beach. A 3-m slumped bed containing a few sandstone clasts is exposed in the limbs of the syncline at the mouth of the Titchberry Water; it probably correlates with the slumped bed at the mouth of the Blegberry Water.

Older beds to the north include a further slumped bed about 3.35 m thick at the southern end of Upright Cliff [SS 2288 2687]. This bed contains large sandstone clasts in its upper part but appears to pass down into undisturbed laminated siltstones at its base. It forms an upstanding feature on the foreshore.

The Longpeak Shale and underlying sandstones and siltstones which lie beneath the slumped bed at Upright Cliff crop out over a considerable area of the foreshore. The upper part of the shale is here represented by thinly bedded sandstones with shale and siltstone partings; load casts and groove casts are common and coprolite pellets are rare. Below are 1.06 m of yellow-weathering, black, laminated shales with calcareous nodules containing juvenile goniatites, wood fragments and Caneyella sp. [SS 2283 2692]. Dr Calver noted several small specimens of Caneyella ranged along either side of a piece of fossil wood on a bedding plane, and suggested that in life the shells were attached to the wood by their byssal strands and had used it as a floating anchorage.

The Longpeak Shale is underlain by about 3 m of thin sandstones resting on about 6.4 m of laminated siltstones and shales. The silty laminae are graded, with abundant prod casts and groove casts on their bases. About 10 m of thinly bedded and medium-bedded sandstones with silty and shaly partings lie between the siltstones and the Hartland Quay Shale. At [SS 2291 2698] the latter comprises 1.37 m of yellow-weathering black shales with some fusiform coprolite pellets and sideritic nodules. Small calcareous nodules 10 to 25 mm in diameter occur about 0.75 m above the base and larger ones 50 to 100 mm in diameter 0.3 to 0.5 m above the base. A band containing abundant crushed anthracoceratid goniatites occurs about 0.3 m above the base. On drying, shale from this horizon is rapidly disintegrated by the growth of fibrous gypsum crystals in the shale foliation. This suggests that the shale is both pyritous and calcareous. The nodules are also commonly coated with gypsum.

About 21 m of thinly bedded and medium-bedded sandstones lie between the Hartland Quay Shale and the fault zone at the northern end of Upright Cliff.

Upright Cliff to Cow and Calf [SS 2272 2713]

This section continues the succession from Cow and Calf to Hartland Point (pp.51–52; (Figure 11), column 3); 105 m of strata are exposed. Immediately north of the strike fault at Upright Cliff [SS 2290 2704] an intensely sheared black shale horizon is exposed in the core of the major anticline. A few nodules from the top of the shale have yielded Gastrioceras circumnodosum together with Dunbarella sp.,indicating that this is the Gull Rock Shale. Shearing has removed all traces of fossiliferous nodules from the shale on the foreshore. The beds between the Gull Rock Shale and Cow and Calf resemble those above the Gull Rock Shale in the Black-church Rock–Clovelly section (below), consisting of thinly bedded and medium-bedded sandstones with sporadic thick sandstones interbedded with siltstones and shales. Lamination and ripple-drift cross-bedding are fairly common in the thinner sandstones.

Beckland Bay [SS 2810 2690] to Mouthmill Beach [SS 2979 2655]

To the east of the main fault through Beckland Bay thinly bedded sandstones are exposed. Overturned beds dipping 60° S at the base of the cliff, and truncated at each end by faults, include at [SS 2842 2673] a 1-m black shale with fusiform coprolite pellets and with a band of nodules about 0.3 m from its base containing anthracoceratid goniatites. Preservation is poor, as both the internal moulds of the goniatites and the fissures in the nodules have been filled with gypsum. Caneyella sp.also occurs at this locality and the shale, by comparison with the sequence at Windbury Point [SS 2872 2682], is correlated with the Hartland Quay Shale.

The Longpeak Shale crops out on the west side of Windbury Point [SS 2862 2678] and continues through to the east side. A massive sandstone forming the northern face of the point is unlike the Bude Formation sandstones. The eastern outcrop of the Long-peak Shale is a black shale 1.07 m thick with fusiform coprolite pellets in the uppermost 0.38 m and with a band of nodules 0.5 m from the top. The nodules have yielded juvenile goniatites and mollusc spat. The thick siltstones normally found beneath the shale on the west coast are here reduced to 1.91 m, with 1.37 m of thin sandstones lying between.

About 30 m of thin- and medium-bedded sandstones, some graded and laminated, lie between the Longpeak Shale and the Hartland Quay Shale. The latter crops out in the cliff at [SS 2877 2675] and comprises 1 m of black, laminated shale with silty bands. The overlying thin sandstone has coprolite pellets concentrated in flute casts on its base. Calcareous nodules in a band about 0.3 m above the base of the shale contain abundant anthracoceratid goniatites. About 1 m of flaggy laminated siltstone lies beneath the shale.

At least 77 m of Crackington Formation strata underlie the Hartland Quay Shale at Windbury Point. The succession has been measured as far as the first anticline south-east of the point, beyond which there is much repetition by folding. A small outcrop of shale with decalcified nodules and fusiform pellets in a stream section near the cliff top [SS 2903 2655] is probably the Long-peak Shale; no goniatites have been found.

The Hartland Quay Shale and the Longpeak Shale provide the link between the successions to the east and west of Mouthmill.

Blackchurch Rock, Mouthmill [SS 2989 2664], to Clovelly [SS 3180 2485]

The coast between Blackchurch Rock and Clovelly (Plate 6); (Figure 7) and (Figure 11) displays a complete succession between the Hartland Quay Shale and the Clovelly Court Shale. Most of the beds are exposed in the vertical or slightly overturned northern limb of a major anticlinorial structure between Gallantry Bower [SS 3047 2624] and Wood Rock [SS 3143 2553]; further asymmetrical folds north of Gallantry Bower repeat the Gull Rock Shale. The uppermost beds, including the Hartland Quay Shale, occur at Blackchurch Rock and are truncated by faults at Mouthmill.

The maximum thickness of. strata present is some 513 m, and the succession is as follows, measured downwards from the youngest bed:

The Gull Rock Shale is exposed in the hinge zones of two anticlines 300 and 400 m SE of Blackchurch Rock (Figure 7) and has yielded in the western anticline [SS 3005 2648] Caneyella cf. mullirugata, Dunbarella papyracea, Gastrioceras coronatum and G. listeri and in the eastern anticline Gastrioceras circumnodosum, G. coronatum and Dunbarella sp.The following downward sequence is that of the northern limb of the eastern anticline:

Folding has repeated part of the succession and the section is continued from the repetition of the Gull Rock Shale at [SS 3076 2605] (Figure 7):

The lowest beds of the section occur in the core of an anticline [SS 3120 2566] 250 m WNW of Wood Rock. To the south-east folding and faulting produce repetitions of parts of the succession, including some of the fossiliferous and slumped horizons. On the southern limb of the anticline thinly bedded sandstones and siltstones overlying the Clovelly Court Shale show signs of contemporaneous disturbance, with the thin beds streaked out and disrupted. They pass into a slumped bed about 7.5 m thick which is repeated by folding at Wood Rock [SS 3143 2553]. The upper surface of this bed on the south side of Wood Rock carries well-formed sand volcanoes.

South of Wood Rock a major strike fault [SS 3143 2550] down-throwing to the south causes repetition of the Embury Shale (p. 100). The shale is exposed at the north-western end of the shallow bay below Cliff Walk Wood, immediately south of the fault. It shows small-scale internal folding and its exact thickness cannot be determined. Near the base of the shale a laminated sandstone about 0.3 m thick is overlain and underlain by bands of intensely black shale. The upper band contains lensoid calcareous concretions with goniatites, the lower band nodules with goniatites. Fragmentary goniatites also occur on the base of the sandstone. The base of the Embury Shale carries large unfossiliferous calcareous nodules. Gastrioceras subcrenatum and G. sp. nov.with coarse lirae have been identified from this locality. The relationship of the Embury Shale to the thin- and medium-bedded sandstones to the south is not clear. The sandstones are thrown into a series of small folds plunging east. At the south-eastern end of the bay [SS 3155 2531] a NW-trending fault is probably present although not exposed. Further thin- and medium-bedded sandstones crop out in the cliff south of this fault and an E–W-trending strike fault of unknown magnitude occurs 100 m NW of Gallant Rock [SS 3170 2519]. At Gallant Rock a massive 30-m slumped bed is exposed in an anticline and syncline. This bed is correlated with the thick slumped beds described in the main part of the succession (p. 38). A thick shale horizon overlying it immediately adjacent to the strike fault is probably the Deer Park Shale. In the appropriate position beneath the slumped bed the Skittering Rock Shale occurs in both limbs of a major northerly overturned anticline. The shale is best exposed at Skittering Rock [SS 3174 2508] in the southern limb. It is medium to dark grey, laminated, about 3.5 m thick, and with a yellow weathering crust in the upper part. Large laminated calcareous nodules occur in the middle of the shales and smaller nodules with small goniatites and spat in the upper part. Casts of goniatites are present on the base of the overlying sandstone. The presence of Gastrioceras cf. cancellatum at this horizon confirms correlation with the shale horizon to the west [SS 3106 2575] (p. 38).

Scattered exposures of thin- and medium-bedded sandstones occur between Skittering Rock and Clovelly. RTT

Cockington Cliff

Exposure between the Sticklepath Fault [SS 3969 2597] and the major normal fault at [SS 3974 2609] is poor; a landslipped sequence of thinly bedded and sporadic medium-bedded sandstones with siltstone and shale bands is apparently overlain to the south by a folded slipped mass of black carbonaceous shale [SS 3971 2602]. A black nodular shale immediately south of the normal fault yielded Gastrioceras cf. circumnodosum, an orthocone nautiloid, and Cypridina? This is the G. listeri horizon, also recorded at the cliff edge of the foreshore here, where nodules yielded Gastrioceras coronatum and Anthracoceratites cf. reniformis. The same bed in a foreshore outcrop [SS 3968 2610] yielded G. circumnodosum and Cypridina?

Mermaid's Pool to Westward Ho!

Between the Mermaid's Pool Sandstone outcrop [SS 4186 2902] on the foreshore, [SS 4197 2904] in the cliff and Westward Ho! [SS 4308 2938] the beds dip 65°–75° SSW and are cut by minor wrench faults. From the top of the formation (base of the Mermaid's Pool Sandstone) downwards the composite succession is as follows:

Appledore

At the northern end of the Appledore promontory, foreshore [4609 3108–4609 3101] and cliff exposures show periclinal folds and the following composite succession:

South-west of here the foreshore, from the lifeboat slipway [SS 4592 3098] to a place [SS 4579 3078] 230 m along the coast, is occupied by about 100 m of grey silty mudstones with dark ferruginous bands and a few turbiditic sandstones, dip 80°–85°/170°, cut by minor wrench faults. These beds closely resemble those at the northern end of the coastal section at Westward Ho! BJW

Instow area

Shales and fine-grained sandstones lie in upright folds on the foreshore [SS 478 324] immediately west of East Yelland Power Station; similar but poorer foreshore exposures occur to the north [SS 4820 3274]. Shallow boreholes at the power station proved these rocks at depths of 1.5 to 3.2 m beneath made ground and Head; another [SS 4823 3214], at the nearby oil depot, passed through 6.4 m of made ground and alluvium on 30.8 m of shales with sandstones.

Shales and thin sandstones at the north end of Instow Sands display upright folds, and quartz veins which commonly trend NW. The Instow Fish Bed lies in the flanks of such folds [SS 4712 3158] and extends a few tens of metres along strike, slightly north of west. Many of the fossil remains occur in silty and calcareous nodules. According to Prentice (1960a, p. 278) the horizon yielded Gastrioceras cf. carbonarium, 'Orthoceras'sulcatum, Dunbarella papyracea var. A, Rhabdoderma elegans, Elonichthys aitkeni and Calamites sp.,but goniatites from it have been identified as G. circumnodosum by Dr M. A. Calver. EAE

Inland sections

Bideford to Gammaton Moor

Inland to the east of Cockington Cliff there are few exposures, but the road-cutting on the A386 extending from 750 to 1150 m S of Bideford Bridge reveals the G. listeri horizon. The succession from south [SS 4532 2525] to north [SS 4530 2575], starting with the youngest bed, is as follows:

BJW

Lymebridge to Ackworthy

Along the Milford Water [SS 2344 2276]–[SS 2374 2269] thinly bedded strata are disposed in a series of folds. Dark grey siltstones with some thin sandstones are exposed in the hinge of a syncline [SS 2345 2275] whose northern limb dips 70°/169° and southern limb 55°/347°. Farther upstream [SS 2350 2274] the hinge of an anticline is offset by a small WNW–ESE dextral fault. A synclinal hinge is exposed in two places [SS 2353 2275]; [SS 2359 2277]; this fold has a calculated plunge of 5° E. Above the roadbridge at Lyme-bridge thinly bedded sandstones with shale and siltstone partings lie in a series of folds. An anticline at [SS 2372 2270] has limb dips of 89°/358° and 30°/187° and a calculated plunge of 5°/269°.

A quarry at [SS 2370 2285] shows an asymmetrical anticline, northern limb dip 80°/352°, southern limb 51°/175°, calculated plunge 1° W. The succession exposed is:

Thinly bedded sandstones and Siltstones and one thick mudstone are exposed in the small stream running along the east side of the road north of Lymebridge [SS 2366 2280]–[SS 2379 2313]. A synclinal hinge [SS 2368 2288] has limbs dipping 58°/178° and 77°/347°, calculated plunge 10°/270°. An anticlinal hinge is exposed [SS 2373 2291] near the confluence of two streams. Reddened and rotten sandstones and shales at [SS 2379 2305] may indicate the proximity of a fault.

At the roadside at [SS 2374 2296] an anticline with limbs dipping 62°/002° and 37°/217°, calculated plunge 13°/280°, contains the following succession:

St Catherine's Tor to Little Barton

A single thick sandstone bed underlain by siltstone and shale dips 25°/172° in a quarry at [SS 2310 2419]. The dip shallows north of the quarry, which may therefore lie near the hinge of an anticline which follows the feature running east from this locality.

In the stream east of Kernstone [SS 2326 2369] dark grey silty-laminated mudstone over 10 m thick is overlain by sandstones dipping 41°/182°. Red crush breccia is exposed in the stream at [SS 2322 2381] and a 0.25-m zone of red fault gouge at [SS 2320 2386] truncates an outcrop of thinly bedded sandstones dipping 60° S. At [SS 2318 2390] sandstones are reddened and cut by minor crush zones

Along the Wargery Water thin sandstones interbedded with siltstones and shales dip 66°/352° [SS 2321 2392] and 54°/170° [SS 2343 2389] and lie in a synclinal hinge [SS 2350 2388]. Vertical thin sandstones and shales striking 087° [SS 2352 2389] abut against another synclinal hinge, possibly at a fault; the limbs of the syncline dip 42°/184° and 90°/354° and the calculated plunge is 9° E. Sandstones dipping 34°/007° at [SS 5358 2388] are faulted against vertical sandstones striking 092°. An anticline at [SS 2371 2391] has limb dips of 80°/176° and 28°/036°. Another anticline is exposed at [SS 2390 2398] and to the east dips are northerly. A thin bed of sandstone [SS 2393 2399] has small sand vol canoes on its upper surface. Around [SS 2417 2400] dips are near vertical and the beds young to the north.

Valley of the Abbey River, Blackpool Mill [SS 2267 2565] to Pattard Bridge [SS 2614 2495]

Fold hinges are commonly exposed in the lowest reaches of the Abbey River in sandstones, siltstones and shales, but the Long-peak Shale, exposed on the coast to the west, has not been recognised. An anticline at [SS 2262 2565] plunges 21°/072°. About 4 m of black shales are exposed at [SS 2267 2561]. A 1.3-m shale [SS 2268 2557] with fusiform coprolite pellets lies in the northern limb of an anticline; reference to the coastal section suggests that this shale is about 30 m above the Longpeak Shale. A gorge south of the footbridge [SS 2268 2555] shows grey medium-bedded sandstones with silty and shaly partings disposed in an anticline and syncline; one limb of the syncline is vertical, striking 079°, the other dips 48°/182°.

A thick sandstone exposed in the river at [SS 2276 2543] dips 85°/161° in a crag on the valley side to the south [SS 2270 2545] and 55°/167° on the north side of the valley, where it is also exposed in a quarry [SS 2298 2546] which shows this bed to be at least 1.5 m thick.

Thin- and medium-bedded sandstones with silty partings form a series of folds [SS 2306 2519]–[SS 315 2510]; the limbs of one anticline [SS 2315 2510] dip 80°/360° and 41°/172°. Synclinal and anticlinal folds [SS 2363 2494] show limbs dipping 55°/352° and 50°/177°. To the north of Bow Bridge thin sandstones in a trackway [SS 2465 2486] show ripple-drift cross-bedded tops. A quarry [SS 2470 2495] exposes the following beds dipping 69°/355°:

Black shales in the north bank of the Abbey River 15 m upstream of Bow Bridge [2461 2475] contain decalcified nodules, one of which has yielded the fish Rhabdoderma elegans. Another contained a fusiform coprolite pellet. This horizon is probably the Gull Rock Shale. On the opposite bank thinly bedded sandstones and ripple-drift cross-bedded siltstones overlie black shales and dip 65°/177°. An anticlinal hinge [2468 24771 in thin- and medium-bedded sandstones with siltstones plunges 13°/260°. The Abbey River to the east follows a shale horizon, probably the Gull Rock Shale, on the northern flank of the anticline. The anticline is exposed at a bend in the river [SS 2478 2479] and shows sandstones in the core and the nodular Gull Rock Shale on both limbs. Decalcified nodules [SS 2476 2480] beneath cross-laminated siltstone have yielded Gastrioceras sp.The shales are finely inter-laminated with silty bands and are 5.5 m thick. On the southern limb of the anticline the nodular horizon is overlain by thin- and medium-bedded sandstones.

The bridge across the river at Glen Cottage [SS 2481 2480] may conceal a small dextral fault; the base of what is probably the Gull Rock Shale appears to be offset about 8 m southwards and the underlying sandstones crop out on the north bank. The nodular horizon at the top of the shale has yielded fusiform coprolite pellets [SS 2485 2478]; some of the nodules preserve tubular cavities which may have been worm burrows. The shale is overlain by thin ripple-drift cross-bedded sandstones with one or two reaching 0.25 m in thickness. The dip is 72°/177°.

Upstream [SS 2509 2473] a syncline has limbs dipping 35°/178° and 44°/325° and plunges 12° E. Another easterly plunging synclinal hinge in a meander [SS 2531 2477] may be the same fold. Black silty shales cropping out across the whole width of the river [SS 2547 2478] show fracture cleavage dipping 80°/177° perpendicular to bedding. Farther upstream a synclinal hinge plunges 20°/252° [SS 2550 2478] and an anticlinal hinge plunges about 12°/257° [SS 2552 2478]. Brown-weathered micaceous silty sandstones with plant remains crop out at [SS 2569 2480].

Tributary streams south of Hartland Mill

The Gull Rock Shale is exposed in a series of anticlines and synclines [SS 2483 2461]–[SS 2490 2444] with steep or slightly overturned northern limbs. Black shales in an anticline [SS 2483 2459] contain calcareous nodules in their uppermost 1 m. The topmost nodules are of black silty limestone with numerous adult and juvenile Gastrioceras sp.Some nodules and fusiform coprolite pellets have been caught up in the base of the overlying thin sandstone which is somewhat conglomeratic and also contains some goniatites. Farther upstream [SS 2486 2455] shales are exposed in another anticline. On the northern limb nodules occur in the upper 1 m of shale; Gastrioceras sp.is found as moulds on the base of the overlying sandstone, which is overturned and dips 60°/179°, but not in the nodules. The shales crop out for about 10 m upstream and show a well-developed fracture cleavage which dips steeply south. Pyrite is present on small joint surfaces and as irregular aggregates, a feature of the Gull Rock Shale elsewhere. In the southern limb of the fold a layer of nodules about 0.3 m thick is overlain by a thin grey micaceous sandstone. Some of the nodules preserve tubular cavities, possibly worm burrows, normal to the bedding. Shales above the sandstone are nodular in their basal 1 m; the uppermost nodules are decalcified, only the outer crust being preserved, and contain G. circumnodosum. Coprolite pellets are scattered within the nodular shale and some of the nodules, and a fish fragment was obtained from a nodule in the lower part of the band. Two thin sandstones, separated by shale and dipping south, overlie the nodular shale. Thin sandstones beneath the Gull Rock Shale are exposed in an anticlinal hinge [SS 2488 2455]; the fold limbs dip 74°/357° and 43°/167° and the plunge is 8°/085°. The shale reappears in the next syncline upstream. In the hinge zone about 0.6 m of thin sandstones overlie decalcified nodules. The shale in the southern limb is about 7.5 m thick. Thin sandstones are exposed in an anticlinal hinge [SS 2486 2451]. Thin- and medium-bedded sandstones in the limbs of an anticline are exposed for 21 m downstream of a weir [SS 2490 2444].

A small quarry on the valley side [SS 2497 2441] reveals medium-to thick-bedded sandstones with silty partings dipping 45°/180°. Others show thin- and medium-bedded sandstones interbedded with siltstones dipping 64°/164° [SS 2505 24351], and 3 m of thin-and medium-bedded sandstones with some silty partings dipping 38°/187° [SS 2506 2419]. A single thick sandstone bed in the stream [2508 24231 dips 35°/172°.

Thin- and medium-bedded sandstones interbedded with siltstones crop out in anticlinal hinges [SS 2508 2413]; [SS 2501 2396] along the stream north-east of Newton. A trackside section [SS 2475 2363] 60 m SSE of Newton shows soft greyish brown siltstones and silty sandstones with plant fragments, 2.5 m, overlain by black shales weathering to grey, 4 m, and thin sandstones and siltstones, 3 m. About 3 m of thin- and medium-bedded sandstones with silty partings dip about 55° S in the quarry at [SS 2525 2316].

Along a stream [SS 2542 2397]–[SS 2581 2336] a thick sandstone [SS 2543 2384] dips 42°/172° and thinly bedded sandstones interbedded with dark grey siltstones [SS 2544 2378] lie in an anticline and dip 20°/344° and 40°/180°. A small quarry [2545 23621 exposes 3 m of medium- and thick-bedded sandstones dipping 48°/177°. Black siltstones and shales with some thin sandstones dip 47°/187° [SS 2551 2352], and thin sandstones overlie 3 m of siltstones in a trackway [SS 2537 2348]. The limbs of an anticline [SS 2558 2342] dip 78°/186° (overturned) and 35°/177^°. Dark grey to black siltstones with scattered sandstones occur in an anticline with a vertical northern limb [SS 2579 2336]; they are overlain by thin sandstones interbedded with siltstones to north and south.

In a quarry [SS 2530 2419] medium-bedded sandstones 1 m thick are overlain by a 2.4-m sequence of thinly bedded sandstones with silty partings passing up into pale grey banded siltstones or mudstones with graded beds up to 50 mm thick. The beds dip 17°/027°.

A stream section [SS 2621 2413] shows thinly bedded silty sandstones and siltstones disposed in a series of small folds. Plunges for two adjacent synclines and the intervening anticline are 9°/085°, 0°/088° and 10°/256°, calculated from dip measurements.

Hartland Village area

Shales, siltstones and thin sandstones, locally stained pink [SS 2632 2430], are exposed sporadically.

Abbey River and tributaries, Pattard Bridge [SS 2614 2495] to Rosedown Mill [SS 2764 2484]

Immediately east of Pattard Bridge medium-bedded sandstones and one thick sandstone are overlain by 1.8 m of siitstones and thinly bedded sandstones and siltstones dipping 79°/002°. About 12 m E of the bridge the dip is 27°/172°, and about 36 m farther upstream the medium- and thick-bedded sandstones are repeated. Southerly dipping thinly bedded sandstones with some medium-bedded sandstones are exposed at intervals upstream to [SS 2650 2498], where thin sandstones and siltstones are folded into an anticline with limbs dipping ^400/3570 and 20°/212° and a plunge of 8°/277°. Upstream from [SS 2704 2487] black mudstones crop out for about 6 m, and most of the exposures as far as Rose-down Mill [SS 2764 2482] are of shales. The Abbey River evidently follows the strike of a belt of shales hereabouts. Dips are mainly southerly but locally vertical or north-westerly. In one place [SS 2720 2487] shales dipping 20°/172° are cut by a vertical fracture cleavage striking 082°. Some thinly bedded sandstones have been affected by superficial disturbance 40 m downstream of Rose-down Mill. A quarry section [SS 2658 2500] shows thinly bedded sandstones with wavy silty partings 1.8 m, overlain by thin- and medium-bedded sandstones with silty partings 3.7 m; the beds dip 43°/171°.

Thin- and medium-bedded sandstones with silty partings and some interbedded siltstones at the southern end of the stream section north of Mount Pleasant [SS 2630 2506]–[SS 2660 2560] form a syncline [SS 2631 2510] with limbs dipping 73°/338° and 67°/178° and a calculated plunge of 26°/256°, and an anticline [SS 2632 2514] showing dips of 80°/348° and 63°/172°. A very tight syncline [SS 2531 2530], whose limbs dip 78°/164° and 82°/352°, shows the following section on its northern limb:

This fossil horizon does not resemble the Gull Rock Shale, in which G. circumnodosumis commonly found in association with G. listeri, and no equivalent horizon has been found on the coast so far. Dr Calver suggests that it could be the equivalent of a marine band such as the Lower Foot of Lancashire, which lies between the C. subcrenatum and G. listeri horizons.

At and immediately upstream from [SS 2643 2533] black shales with some thin sandstones and siltstones show southerly and northerly dips with some disturbance and slight overturning. Nodules [SS 2644 2538] at the top of black shales are up to 150 mm across and mainly decalcified. Some contain goniatites and fusi form coprolite pellets. The surrounding shale is sheared and brecciated but yielded one coprolite pellet. The dip is 48°/353°, but 10 m downstream the shales are overlain by thin sandstones and shales dipping 68°/187°; thus an anticline appears to be present. Dr Calver has identified G. listeri and G. cf. coronatum from the nodules and the horizon is typical of the Gull Rock Shale.

Thinly bedded sandstones with shaly partings at [SS 2644 2541] dip 70°/004° and are cut by two crush zones 0.3 to 0.6 m wide containing red and grey clay gouge and sandstone fragments. These faults trend 317° and 322° and dip 80° NE. The Head is reddened for about 23 m downstream.

In the stream east of Norton [SS 2668 2493]–[SS 2673 2506], thin sandstones are disposed in an anticline with dips of 70°/347° and 40°/154° and a calculated plunge of 9°/090°. The following beds dip 88°/327° in the east bank [SS 2703 2547]:

Dr Calver comments that this goniatite fauna is typical of the G. listeri horizon. The horizon does not altogether resemble the Gull Rock Shale, in which the nodules occur at the top of a shale band, but the rocks may be overturned. Possibly it is the same horizon as that north of Mount Pleasant (above).

Farther upstream thin- and medium-bedded sandstones with silty partings and some interbedded siltstones show two anticlinal hinges. In one [SS 2707 2555] the limbs dip 50°/337° and 38°/182° and the plunge is 21°/252°; the other [SS 2708 2557] shows limbs dipping 72°/348° and 41°/197° and has a calculated plunge of 18°/265°.

Shales crop out for 19 m upstream from [SS 2749 2499], beyond which for 13 m they are overlain by siltstones and shales with thin sandstones dipping 83°/154° and possibly overturned. About 14 m farther upstream very thin sandstones interbedded with siltstones and shales dip at 90° and strike 078°; bottom structures indicate younging to the north and some dips of 87° S indicate slight overturning. Medium-bedded sandstones farther upstream exhibit large load casts and dip 77°/356° [SS 2750 2506] and 75°/342° [SS 2752 2514].

Thinly bedded sandstones and shales overlain by grey shaly siltstones and thin sandstones dip about 50° S in the stream at [SS 2704 2468]. A zone of buff brecciated sandstone or siltstone [SS 2705 2464] with a central band of orange clay may mark a strike fault. The sandstones below the crush zone dip 35°/202°. About 11 m upstream thin sandstones with ripple-drift cross-bedded tops dip south.

In the corner of a field above a stream at [SS 2758 2467] the following beds dip 45°/177°: black shales, 2 m, overlain by thin sandstones with siltstones and shales, 3 m, and thin- and medium-bedded sandstones, 1.2 m. Exposures in and alongside the stream show shales, thin sandstones and siltstones with southerly and northerly dips.

Abbey River and tributaries, east of Rosedown Mill [SS 2764 2484]

Shales and thin- and medium-bedded sandstones generally dip steeply southwards and are locally overturned [SS 2822 2462]. An anticline in thin sandstones [SS 2836 2403] plunges 9° E with limbs dipping 70°/017° and 46°/182°. Thin sandstones and siltstones are exposed [SS 2869 2372] in the hinge of a syncline whose limbs dip 45°/177° and 90° on a strike of 097°. An anticline in thin sandstones and siltstones [SS 2882 2344] plunges 4° E, with limbs dipping 75°/003° and 45°/178°.

Thin- and medium-bedded sandstones with siltstones exposed in a stream gully pass from moderate northerly dips to vertical, strike 052° [SS 2788 2490], with southerly younging indicated by ripple-drift cross-bedding. Farther up the gully [SS 2789 2492] the dip is 13°/212°, suggesting an anticline; the following section is seen:

Dr Calver has identified G. subcrenatum, together with three new species of Gastrioceras, from this locality, which represents part of the Embury Shale. It was probably here that Rogers (1910) obtained goniatites from thin beds of soft sandstone.

At the top of Colpit (Hescott) Quarry [SS 2792 2495], 0.9 m of Head, formed of buff sandstone rubble in a silty matrix, overlies about 3 m of thinly bedded sandstones and siltstones. The main quarry section shows 5.6 m of thin- and medium-bedded, dark grey micaceous sandstones interbedded with shales and some siltstones in the ratio 2:1. The beds are folded into two open anticlines separated by a small syncline and appear to lie stratigraphically above the fossiliferous horizon in the gully below the quarry.

A quarry at [SS 2831 2460] beside a stream exposes purple-stained flaggy and fissile micaceous sandstones dipping 40°–50°/128°. Reddened and disturbed thin sandstones and shales beside the stream [SS 2863 2452] probably lie in a fault zone. A slumped bed [SS 2867 2451] of dark grey, poorly bedded, silty mudstone with contorted sandstone clasts up to 1 m long is at least 6 m thick.

Laminated grey and black shaly mudstones crop out in scattered exposures [SS 2899 2357]; [SS 2934 2365]; [SS 2937 2347] and in a well [SS 2942 2361]. Black shale fragments were ploughed up in a field to the west. Thin sandstones crop out [SS 2936 2328] on the track south of Sowden.

Blegberry Water [SS 2265 2595]–[SS 2411 2575]

The slumped bed shown in the coastal section (Figure 6) is exposed at the mouth of the stream. Elsewhere outcrops show thin- and medium-bedded sandstones, shales and mudstones. A small tributary stream [SS 2361 2593] exposes an 8.5-m section of sandstones up to 0.15 m thick interbedded with shales and siltstones. An anticlinal hinge at [SS 2385 2584] shows black silty shales overlying thin sandstones dipping 81 °/008° in the northern limb; the southern limb dips 73°/174°.

Titchberry Water [SS 2318 2675] to Moor [SS 2640 2632]

Thinly bedded, flaggy and micaceous buff-weathering sandstones and siltstones [SS 2318 2675]; [SS 2322 2677]; [SS2326 2683]–[SS 2335 2683]; [SS 2342 2681]–[SS 2358 2681] show moderate southerly dips. Black shales crop out in the stream bed [SS 2372 2678], and southerly dipping thinly bedded sandstones with shale bands up to 1.2 m thick extend upstream to Shamley Bridge [SS 2463 2678]. Shamley Quarry [SS 2490 2685] contains thin- and medium-bedded sandstones dipping 47°/184°. Between [SS 2480 2681] and [SS 2489 2681] the stream runs parallel to the hinge of an anticline. The northern limb is vertical or slightly overturned and shows [SS 2480 2681] thin sandstones and shales, 3 m, overlain by black shales, 2.4 m, and thin sandstones and shales, 2.4 m. A thick mudstone band crops out in the stream bank at [SS 2523 2665]. Immediately downstream of Gawlish Bridge sandstones dip 76°/359°, and 20 m upstream of the bridge 4 m of shale are exposed.

At the north-west corner of the farm buildings at Laburnum Cottage [SS 2521 2681] an anticlinal hinge shows thin sandstones with black shales in the core; the limbs dip 40°/196° and 68°/104°. A stream to the north-east has exposed an anticline [SS 2520 2689] whose limbs dip 45°/358° and 84°/187°. A vertical fault [SS 2530 2690], marked by 0.6 m of red clayey gouge, trends 155° through partly reddened mudstones and silty sandstones. Just downstream medium- and thick-bedded sandstones are cut by a small fault trending 137°. Further stream exposures [SS 2538 2661]–[SS 2680 2657] show thinly bedded sandstones, siltstones, mudstones and shales dipping southwards, northwards and locally between north-east and east-south-east. A small fault trends NW-SE at [SS 2563 2677]. Grey to black laminated mudstones [SS 2680 2657] dip 48°/192° with a fracture cleavage dipping 73°/172°. The bedding/cleavage intersection indicates that they lie on a fold plunging 38°/247°. Similar rocks in another stream course [SS 2593 2619]–[SS 2653 2618] are locally reddened and disturbed [SS 2604 2617], probably indicating proximity to a fault, and contain plant fragments [SS 2653 2618]. In a stream farther west [SS 2535 2631] an anticlinal hinge with limbs dipping 58°/352° and 35°/197° has a calculated plunge of 11°/270°.

Brownsham-Hescott-Clovelly

A stream section [SS 2825 2593]–[SS 2834 2605] exposes disturbed black shales [SS 2825 2593] and the following section [SS 2829 2599]:

The dip is 90° on a strike of 078° and the beds young to the north.

Beckland Water [SS 2827 2615]–[SS 2852 2614] exposes thin- and medium-bedded sandstones with siltstones and shales dipping to north and south. A small outcrop of black shale at [SS 2903 2655] contains decalcified nodules and fusiform phosphatic pellets. A stream flowing north to Mouthmill [SS 2981 2510]–[SS 2980 2655] cuts through similar strata. The forestry track beside the stream [SS 2984 2597] appears to cross an anticline trending E–W. A small reddened crush zone is exposed [SS 2986 2620] and also a syncline [SS 2985 2634] with limbs dipping 30°/177° and 48°/347°.

Near Brownsham, at the confluence of two small streams [SS 2868 2583], the following beds are exposed, dipping 26°/202°:

Dr Calver has identified G. subcrenatum with typical lirae from this locality, together with three new species of Gastrioceras. The fauna clearly equates with that of the Embury Shale near Colpit (Hescott) Quarry, and on the coast between Mouthmill and Clovelly. A stream section [SS 2897 2588]–[SS 2905 2587] contains a vertical shale band at least 10 m thick, probably on the northern limb of an anticline.

At [SS 2872 2497] about 3 m of laminated dark grey silty shales with some thin sandstones in the upper part dip 15°/202°. Temporary exposures in drainage ditches at Yapham [SS 2886 2545]; [SS 2870 2527] revealed large thicknesses of shales. Sandstones, siltstones and shales around Hescott show an anticlinal hinge [SS 2939 2495] with northern limb inverted and dipping 58°/184° and southern limb dipping 35°/192°.

A stream east of Higher Velly and Nether Velly [SS 2956 2408]–[SS 2977 2451] shows thin-, medium- and a few thick-bedded sandstones dipping north and south. Northerly dips are steep. A 14-m black shale with silty bands and some thin sandstones [SS 2962 2420] dips 90° on a strike of 097°; it is underlain by thin sandstones and youngs northwards. Thick laminated shale [SS 2969 2438] is underlain by sandstones dipping 52°/177°.

Stream sections near West Dyke [SS 3075 2370], in thin- and medium-bedded sandstones interbedded with siltstones and shales, include an exposure [SS 3024 2317] of 8 m of mudstones and another of an anticline [SS 3026 2408] whose limbs dip 74°/356° and 40°/187°. At [SS 3011 2430] thin- and medium-bedded sandstones dip 80°/162° but are overturned. At [SS 3011 2438] 3 m of thin sandstones and shales are overlain by laminated shales which pass up into 2.5 m of flaggy sandstones and siltstones overlain by thin sandstones. The dip is 48°/186°. About 40 m downstream silty mudstone with included sandstone clasts is probably a slumped bed.

A small quarry [SS 3091 2577] reveals thin- and medium-bedded sandstones with silty or flaggy partings dipping 83°/350°. Similar rocks with interbedded siltstones have been quarried nearby [SS 3165 2500]; [SS 3146 2482]. Exposures in Hobby Drive show [SS 3177 2411] synclinal and anticlinal hinges; the limbs of the anticline dip 85°/014° and 45°/354°. A stream cuts through an anticline plunging 6°/256° [SS 3210 2389] and another plunging 2°/069° [SS 3209 2393]. RTT

Instow-Yelland-Eastleigh

In the neighbourhood of Yelland, a shallow borehole proved shales at 2.9 m [SS 4939 3256] but two others of 6-m and 10-m depths sited on boulder clay (p. 108) failed to enter solid rock. Many roadside exposures in the vicinity of Instow Town [SS 482 311] suggest the presence of groups of sandstones trending slightly S of E, and thickly bedded fine-grained sandstones have been quarried 1 km to the E [SS 4912 3100]. Most exposure in Instow is of shales and siltstones with subordinate thin turbidite sandstones, but the higher ground inland, on which stand Worlington [SS 4810 3056] and Fullingcott [SS 4890 3038], contains several sandstone groups which show widespread evidence of small-scale quarrying [SS 4796 3030]; [SS 4834 3070]; [SS 4852 3063]; [SS 4826 3011]. Sporadic exposures of shales, siltstones and thin sandstones occur in roadsides and farmyards, and old pits mark the positions of several sandstone belts. In one small quarry [SS 4914 2980] fine-grained, locally quartzitic sandstones and siltstones with some silty nodules and organic fragments are disposed in an anticline (p. 98) trending N of W. Another quarry [SS 4834 2904], near Tapeley, shows 6.1 m of massive and thickly bedded fine-grained feldspathic sandstones, grey weathering to brown, reminiscent of the Bideford Formation (p. 30). About 1 km S of E from Treyhill a 6.7-m band of fine-grained thickly bedded sandstones dips steeply between south and south-south-west within a series of shales, siltstones and fine-grained sandstones [SS 4977 2844]; the beds are right way up. South of Eastleigh a line of old pits and patches of abundant sandstone debris extend along strike for about 2 km [SS 479 276]–[SS 496 275]. EAE

Northam and the River Torridge

About 3 m of grey silty mudstone in a road cutting [SS 4487 2917] at Northam dip 70^°/175^°, and scattered exposures of turbiditic sandstones occur in the lane banks hereabouts.

Hubbastone Quarry, Appledore [SS 464 298], contains thinly bedded and a few medium-bedded turbidite sandstones which have yielded two faunas. A 0.3-m dark grey nodular shale in the south-west corner of the quarry [SS 4638 2975] encloses weathered rusty nodules with antracoceratids, Gastrioceras amaliae, Caneyella sp. [Juv.], Posidonia?, mollusc spat and conodonts including Hindeodella sp.This is the G. amaliae horizon. On the north side of the quarry [SS 4637 2982], dark grey shaly mudstone 15 m below the G. amaliae horizon yielded Gastrioceras?, Caneyella sp.,mollusc spat, and abundant conodonts in aggregates, including Hindeodella sp.and platformed types. Dr Calver comments on the abundance of conodonts and on the vague resemblance of one of the two poorly preserved goniatites to G. listeri. He notes that if it were the G. listeri horizon, the persistent nodular G. circumnodosum band should be present. Its absence, together with the proximity of the G. amaliae horizon, suggests that the specimen is not G. listeri.

On the west bank of the Torridge [SS 4589 2867]–[SS 4601 2889] a succession similar to that on the coast is exposed. Some 250 m of beds which correlate well with the coastal section are overlain by 14 m of graded siltstones which Walker (1970, p. 57) equated with turbidite sandstones of the coastal section. These are succeeded by 23 m of cross-laminated siltstones in which cross-laminated sandstone beds become thicker and more common upwards, with the Mermaid's Pool Sandstone at the top.

On the east bank of the Torridge [SS4692 2933]–[SS 4687 2926], and probably lower in the formation, sandstone 1 m, is overlain by dark silty mudstone 0.75 m, two 0.5-m graded turbidite sandstones, and laminated silty mudstone with siltstone bands 3 m. The beds dip 85°/176°. Farther north [SS 4728 2987]–[SS 4724 2980] the following beds dip 70°–80°/177°:

Bideford Formation

Coastal sections

The Bideford Formation forms the cliffs between the northern end of Cornborough Cliff [SS 4197 2904] and the southern end of Abbotsham Cliff [SS 4064 2726]. The beds are folded into simple open anticlines and synclines, and the pattern of conjugate north-westerly dextral and north-easterly sinistral wrench faults is well seen on the foreshore. There is much repetition of outcrop. In the composite section below, the top 246.28 m (from the culm bed above the Cornborough Sandstone down to the base of the turbidite group in Cycle 6) were measured in the cliffs between [SS 4064 2726] and [SS 4082 2749]; the 116.42 m below were measured in the cliffs between [SS 4137 2842] and [SS 4142 2850].

Simpson (1933, p. 434) collected a bivalve assemblage from near the base of this last bed and identified 'Carbonicola aff. pseudacuta', 'Carbonicola aff. recta', 'Carbonicola cf. ornata', and 'Anthracomya cf. bellula. Dr Calver and Dr R. M. C. Eagar consider that Simpson's illustrations (fig. 1, 1 and pl. xxiii, figs. 1, 3 and 5) show that this fauna would now be referred to an assemblage with Carbonicola extenuata and C. crispa, such as characterises the C. proxima-extenuata belt of the top of the lenisulcata Zone. The bed recurs in the northern limb of a syncline in Cornborough Cliff [SS 4134 2831], where Carbonicola cf. extima and C. proxima were found 0.5 m above the base; the faunal evidence from both localities suggests the horizon lies in the upper part of the lenisulcata Zone.

Cycle 3 (179.00 m)

The Raleigh Sandstone and immediately underlying beds, comprising the top 42 m of Cycle 3, were measured in the cliff section [SS 4100 2771]–[SS 4106 2779] at the northern end of Abbotsham Cliff.

The basal mudstones are best seen in cliff and foreshore [SS 4157 2863]–[SS 4167 2873] at the north end of Cornborough Cliff. Grey slightly silty mudstone 82 m below the base of the Raleigh Sandstone yielded Curvirimula cf. scotica [juv.] at [SS 4158 2863], and 47 m below this horizon Carbonicola cf. pseudacuta [juv.], C. cf. lenicurvata [juv.] and Curvirimula cf. scotica were found [SS 4166 2869]. Dr Calver comments that these are juvenile or stunted forms of high Namurian or possibly low Westphalian age. They are the earliest bivalves found in the Carboniferous rocks of the district.

Cycle 2 (48.15 m)

Cliff and foreshore exposures [SS 4167 2873]–[SS 4171 2879] show:

Cycle 1 (158.20 m)

The top 5.8 m of the Rock Nose Sandstone were measured in the cliff face [SS 4176 2888], the next 9.2 m on the foreshore [SS 4180 2890] and the basal 4.3 m in the cliff [SS 4170 2887] as follows:

The beds below the Rock Nose Sandstone are exposed on the foreshore [SS 4170 2889]–[SS 4177 2902] and in the cliff [SS 4181 2892]–[SS 4191 2902] as follows:

The beds below comprise the Mermaid's Pool Sandstone, 26.9 m thick as measured in the cliff face [SS 4194 2904] and readily traced westwards across the faulted foreshore:

Inland sections

Abbotsham to Bideford

A small quarry [SS 4142 2825] contains the following section, dip 85°/182°:

A small quarry [SS 4172 2820] in the Cornborough Sandstone 100 m to the west shows the following beds dipping 80°/190°:

In a trackside exposure [SS 4195 2792] near Cornborough 1.35 m of grey silty mudstone with harder sandy bands dip 85° N at the top of a sandstone feature. An exposure at the base of a similar feature [SS 4163 2794] revealed 3.55 m of cleaved buff to grey silty mudstone and siltstone, with nodular sandy bands towards the top, overlain and channelled by 0.35 m of buff to brown cross-bedded sandstone. Dark grey cleaved silty shale 1 m thick lies at the top of the feature nearby [SS 4161 2796]. About 260 m to the west, at the base of another feature, the following beds dip 70°–75°/010°:

About 3 m of buff current-bedded sandstone dip 80° N at Pusehill [SS 4268 2829]. In a quarry [SS 4285 2863] at Buckleigh 4.5 m of buff ripple-marked thinly bedded sandstone, probably the Raleigh Sandstone, dip 80°/180°. The same sandstone behind the Raleigh Garage [SS 4481 2753] comprises 6 m of coarse-grained, well-bedded and thinly bedded sandstone overlain by 5.1 m of muddy fine-grained sandstone with coarser-grained patches and cross-bedded bands.

Almost 19 m of the Raleigh Sandstone dip 54° S in a large flooded quarry [SS 441 276] (Plate 8). The uppermost 6 m, of muddy cross-laminated fine-grained sandstone, show small-scale sedimentary structures. A quarry [SS 4346 2764] at Kenwith is probably also in Raleigh Sandstone; some 4 m of buff thinly bedded sandstone with ripple-drift cross-bedding and carbonaceous streaks dip 72°/010°.

In a quarry near Abbotsham [SS 4157 2708] 12.2 m of massive brown feldspathic sandstone (Cornborough Sandstone) dip 84°/180°; the sandstone becomes more flaggy upwards and contains bands rich in plant fragments and mudstone pellets. Flaggy feldspathic sandstone 5.22 m thick dips 75°/176° in a quarry [SS 4204 2804] 450 m to the east.

East-the- Waler to Little Pillhead

East of the River Torridge the first sandstone ridge south of Westleigh is probably the Mermaid's Pool Sandstone; fragments of fine-grained cross-laminated bioturbate sandstone occur in a disused quarry [SS 4665 2854]. Rubble of muddy cross-laminated fine-grained sandstone in a quarry [SS 4624 2820] 500 m to the SW probably represents the Raleigh Sandstone. Flaggy muddy sandstone 1.82 m thick dips 75°/010° in a road cutting [SS 4600 2787]. Near Pillhead 7.62 m of fine- to medium-grained thickly bedded sandstone with a pinkish grey tinge in a quarry [SS 4786 2736] are probably the Raleigh Sandstone; the beds are vertical, strike 102°. In cuttings alongside the Old Barnstaple Road 3 m of grey silty mudstone with thin sandy bands dip 60° S [SS 4738 2703], and 4.1 m of silty mudstone and siltstone dip 80°/102° [SS 4732 2696]. Another cutting 540 m to the west [SS 4681 2682] contains about 14 m of landslipped dark grey laminated shale and silty shale.

Bude Formation

Coastal sections

Cow and Calf [SS 2272 2713] to Hartland Point [SS 2297 2779]

The highest part of this succession is found in the syncline at Hartland Point. The thick sandstones of Bude Formation type forming part of the projecting rock called Cow and Calf are taken to be the base of the formation. They lie 56 m below the Hartland Quay Shale. The succession described below in downward order, is summarised in (Figure 11), column 3, and is some 375 m thick:

Hartland Point to Shipload Bay [SS 246 275]

The upper part of the sequence at Hartland Point can be traced eastwards to Shipload Bay, the thick shale unit above the uppermost slumped bed providing a marker. A channel cut into this shale, and filled with massive sandstone of Bude Formation type with shale-flake conglomerates at the base and within the mass of the sandstone, is exposed on the approach road to the lighthouse [SS 2318 2763]. This channel trends E–W, is visible in the shale on the east side of Barley Bay [SS 2360 2759] and is probably present high in the cliff on the west side of Hartland Point. An unusual imbricate structure, in which prisms of sandstone have subsided into the underlying shale (Plate 9) is visible at the foot of the cliff on the west side of Hartland Point [SS 2303 2769] and on the west side of Barley Bay. Ashwin (1957) interpreted this structure as a form of tectonic imbrication, but the presence of lenses of thin sandstone filling the hollows between the displaced prisms of sandstone indicates movement shortly after deposition. The cause of the displacement is not clear. The sandstone prisms may represent rotational slips, perhaps implying a sea floor sloping northwards. Alternatively they may reflect disturbance of the sandstones above the shale during excavation of the channels described above.

The major slumped beds of Hartland Point continue eastwards across Barley Bay (Plate 10), where they are displaced by a NW-SE dextral fault. They are particularly well developed on the northern limb of the synclinorium, where huge sandstone clasts are enclosed in the silty matrix. On the southern limb, close inshore in Barley Bay, the slumped beds appear to be thinner, more shaly and with few clasts. Measurement of the axial direction of a folded sandstone clast within the upper slumped bed in Barley Bay indicates that the folding was produced by eastward movement; this may reflect the direction of movement of the slumped bed. Correlation with the section in West Titchberry cliff on the west side of Shipload Bay ((Figure 11), column 4) shows that the slumped beds of Barley Bay have died out and part of the sequence is represented by massive sandstones of Bude Formation type, some over 4 m thick. Two massive slumped beds occur lower in the sequence. The lower one is some 22 m thick, the thickest slumped bed in the Hartland area.

Shipload Bay to Beckland Bay [SS 281 269]

The stratigraphical equivalence of parts of the succession on either side of the fault through Shipload Bay has been established by the measurement of a series of short sections along the strike as far as Beckland Bay [SS 2810 2690], where the Longpeak and Hartland Quay shales are exposed ((Figure 11), columns 5 and 6). A slumped bed about 3 m thick which crops out at Eldern Point [SS 2484 2763] can be traced, together with the two overlying siltstone-shale horizons, for 3 km to the east. A thin slumped bed 4 m above this horizon dies out just east of Eldern Point. The upper of the thick siltstone-shale horizons, about 52 m above the persistent slumped bed on the east side of Shipload Bay, shows joint surfaces with traces of green copper secondary minerals. There is no evidence of external mineralisation and probably the rocks contain some syngenetic copper. Traces of copper secondary minerals have been noted in some other shales in the Bude Formation. This thick siltstone-shale horizon is of further interest in that its upper part is a slumped bed and the junction between the slumped bed and the underlying shale is planar. The slumped horizon becomes thinner eastwards and dies out between the measured section at Blue Mellem [SS 2645 2755] and that at Chapman Rock [SS 2677 2764]. A similar relationship of a slumped bed resting on shale is found on the west side of Shipload Bay ((Figure 11), column 4), but the two horizons do not appear to be related.

The rocks exposed at the base of the cliff and on the foreshore between Shipload Bay and Beckland Bay are not typical of the Bude Formation, being mainly thin- and medium-bedded sandstones with siltstones and shales. More typical thick sandstones and shales crop out higher in the cliff and at the top of the Blue Mellem landslip. Massive blocks of silty rock with sandstone clasts at the foot of the landslip indicate the presence in the cliff of a major slumped bed. This may be a continuation of the massive slumped bed which forms an isolated outcrop just east of the fault in Shipload Bay.

Farther east [SS 2713 2751], a 3.5-m sandstone, with a shale-flake conglomerate at its base, is channelled into underlying thin sandstones. The sandstone thins out to east and west over a distance of 60 m but the trend of the channel is not clear. Eastwards from the major synclinal axis at [SS 2753 2728] successively older beds comprise thin- and medium-bedded sandstones, siltstones, and shales characteristic of the lower part of the Bude Formation. A nodular shale horizon correlated with the Longpeak Shale is exposed at [SS 2820 2690] within thick sandstones of Bude Formation type. It comprises 4.19 m of black shale with grey silty bands and has developed a yellow weathered crust. Fusiform coprolite pellets occur in the uppermost 0.45 m. A narrow band of decalcified nodules containing goniatite spat occurs 0.5 m from the top. Another, in which the nodules are coated with gypsum, lies 1.27 m from the top and is underlain by a few thin silty sandstones. Fossils from this shale include Caneyella sp.and Rhabdoderma sp.

The base of the Bude Formation is taken beneath the thick sandstones below the Longpeak Shale at Beckland Bay. However, strata above the shale at Windbury Point to the east, although they include some thick sandstones, are more characteristic of the Crackington Formation.

Longpeak Beach [SS 2228 2280] to Brownspear Beach [SS 2245 2345]

At the northern end of Brownspear Beach, south of Speke's Mill Mouth, major strike faults with southerly downthrows repeat the lower part of the Bude Formation. The base is taken below the thick sandstone immediately above the Hartland Quay Shale at Hole Rock [SS 2214 2263], just south of the district. Some 68 m of strata ((Figure 11), column 1) comprise thick sandstones scattered throughout a succession of shales, siltstones and thin sandstones.

The Longpeak Shale occurs 33 m above the base of the sequence [SS 2221 2267]. It comprises shales containing calcareous nodules with goniatite spat and fish remains, 0.75 m, overlain by dark grey shale with coprolite pellets, 1 m, thin laminated or ripple-drift cross-bedded sandstones, 2.5 m, and shale with fusiform coprolite pellets, 1 m. A thick sequence of siltstones and thin sandstones occurs beneath the Longpeak Shale, as at Hartland Point. Multiple repetition of the shale occurs on Longpeak Beach as far north as Longpeak [SS 2211 2301]. Also on Longpeak Beach [SS 2218 2278] a group of four sandstones up to 6 m thick is channelled into the siltstones underlying the Longpeak Shale (Figure 10). These sandstones lie in the southern limb of an anticline and pass into thin sandstones on the foreshore to the west. They reappear closer to the cliff in the northern limb of the next anticline to the south. An approximate NNW–SSE trend for the channel is inferred.

A 1.2-m shale with fusiform coprolite pellets lies 40 m above the Longpeak Shale. A similar bed occurs at this level at Hartland Point and also in the cliff between the two strike faults at Brown-spear Beach, the latter occurrence suggesting that the main movement was on the northern fault.

The top of the succession is marked by a 3.5-m slumped bed in the core of a syncline at the southern end of Brownspear Beach. RTT

Clovelly to Buck's Mills

From the eastern end of Clovelly Bay to 165 m SE of Black Rock the cliff section [SS 3240 2437]–[SS 3302 2411] is obscured by landslip. At Black Rock 1 m of massive sandstone is overlain by thinly bedded sandstone 2 m and sandstone 2 m, dipping 58° N. Eastwards from 165 m SE of Black Rock the beds in the cliff are inverted and dip approximately 70°/170°. The section is as follows, described in order from north-west [SS 3302 2411] to south-east [SS 3320 2397], beginning with the youngest bed:

The beds to the south-east of the thrust fold-axis dip about 70°/170° in normal succession and are described in order from east [SS 3322 2397] to west [SS 3320 2397], beginning with the youngest bed:

Beyond a steep-sided stream valley [SS 3322 2397], where about 10 m of strata are obscured by slip, the beds are inverted and are described in order towards the south-east, beginning with the youngest bed:

The beds beyond the slip retain the same dip but are in normal succession; the sequence is as follows, from south-east [SS 3333 2388] to north-west [SS 3320 2390], beginning with the youngest bed:

Eastwards to the Lower Bight of Fernham [SS 3333 2388]–[SS 3343 2385] landslip obscures the section. Between the Lower Bight of Fernham and Barton Wood the cliff section shows a northerly-over-turned anticline whose axis falls from 15 m up the cliff in the west [SS 3343 2385] to the base of the cliff in the east [SS 3416 2384]. The axis is a high-angle reverse fault or thrust in most places. The beds in the S-dipping normal southern limb are poorly exposed in the upper part of the cliff, but the northern inverted limb contains the following succession in the west [SS 3343 2385], described from the base of the cliff upwards (stratigraphically downwards):

The beds around the Longpeak Shale are best seen at the eastern end of this section [SS 3416 2384], where the northern inverted limb of the anticline reaches the foreshore immediately adjacent to the cliff. The section here measured towards the cliff face (stratigraphically downwards) is as follows:

From here [SS 3416 2384] to West Buck's Point the cliff section shows beds in normal succession (the southern limb of the anticline), dipping at up to 60°/165°–170°. The succession apparently starts at or just above the 2-m sandstone at the top of the previous inverted section and is described in order from east [SS 3473 2583] to west [SS 3416 2384], beginning with the youngest bed:

At West Buck's Point [SS 3473 2383], the cliff section forms the southern limb of a W-pitching pericline, and the beds described above are repeated in the foreshore in the northern limb of the fold, in normal sequence. The sequence of beds in the cliff continues upwards, in normal succession with a steepening dip, towards Buck's Mills, and is described in order from east [SS 3490 2379] to west [SS 3473 2383], beginning with the youngest bed:

The section from [SS 3490 2379] to [SS 3507 2375] is obscured by landslip. The beds to the east lie in the northern overturned limb of an anticline whose southern limb is probably high in the cliff. The succession is as follows (stratigraphically downwards):

The section shows steep dips at its eastern end, near the core of the anticline, but a gentler dip of about 20° at its western end [SS 3522 2371]. About 60 m W of the foreshore end of the cliff road at Buck's Mills, inverted beds [SS 3539 2369] in the lower part of the cliff and the foreshore may lie stratigraphically above the preceding section. The succession is as follows, from foreshore towards cliff, beginning with the youngest bed:

Buck's Mills to Peppercombe

About 30 m E of the bottom of the cliff road at Buck's Mills the cliff section [SS 3548 2366]–[SS 3552 2366] shows a normal succession of beds apparently stratigraphically above the last section, and this succession is repeated on the foreshore in the northerly overturned or vertical limb of an anticline. The higher beds crop out towards the east and the section is as follows, from the youngest bed downwards:

Section obscured by landslip and debris [SS 3567 2367]–[SS 3571 2367]. The beds to the east may be stratigraphically above the previous section. They show little way-up evidence but the eastward succession is thought to be stratigraphically downwards as follows:

The first headland east of Buck's Mills [SS 3577 2370]–[SS 3579 2371] shows a normal succession, of unknown relationship to the previous section, as follows from the top downwards:

About 180 m of landslipped and debris-covered cliff separate this exposure from the next [SS 3600 2374], which shows an anticline overturned to the north. From the axis of the fold upwards, the southern limb in normal succession comprises broken shale 1 m, massive sandstone 1 m, shaly broken rock 1.3 m, overlain by variably bedded sandstones in the cliff as in the last section. A further 180 m to the east, beyond more land-slipped cliff, the axis of probably the same fold is exposed in stratigraphically higher beds. The section in the southern limb, normal succession measured stratigraphically downwards, is as follows:

At 200 m to the east [SS 3634 2388] the inverted limb of the fold, from the edge of the foreshore up into the cliff, shows the following section (stratigraphically downwards):

At 180 m to the ENE, a section on the foreshore [SS 3649 2396] opposite Gauter Point shows inverted beds, stratigraphically above the last section but with an unknown thickness of strata between them. The section stratigraphically downwards from the foreshore towards the cliff is as follows:

About 460 m of landslipped cliff, apparently covering a series of thickly bedded sandstones, separate the last exposures from the following section [SS 3700 2395]–[SS 3708 2397], in which dips are south-westerly at various angles and which is given from the top downwards:

About 20 m of broken and contorted thickly bedded sandstones here [SS 3708 2397] lie near the axis of an anticline. Some faults may be present; the beds to the north-east of the axis do not correlate with those described above and comprise, from the top [SS 3747 2406] downwards:

For 110 m NE from here the cliff shows intermittent exposures of NE-dipping thickly bedded sandstones with shale and siltstone bands. Presumably these beds lie stratigraphically above the previous section; their top comprises the following steeply dipping succession from the top [SS 375 2410] downwards:

For 80 m NW the cliff section is obscured by slip but must contain a synclinal axis or a fault since the beds between here and the Permian rocks at Peppercombe dip south-west at various gentle angles. These beds are probably at least in part equivalent to those farther south-west; the succession from the top [SS 3763 2414] downwards, is:

Peppercombe to Portledge

Immediately north-east from the Peppercombe Fault [SS 3817 2430]–[SS 3825 2436] strata dip 25°/175° in the following succession, measured stratigraphically downwards:

Landslip and rock debris mask the cliff for 40 m NE, beyond which [SS 3829 2438] the following section extends to a near-vertical fault [SS 3831 2439]; the beds are recorded in descending stratigraphical order:

The cliff between the fault and the Permian unconformity [SS 3837 2443] shows a contorted near-vertical succession younging to the north-east and described stratigraphically downwards as follows:

Portledge to Babbacombe Mouth

Exposure is poor for 1 km NE from Portledge Mouth but at 60 m NE of Portledge Mouth the following beds dip 45°/100° [SS 3872 2475] and are recorded stratigraphically downwards:

At 180 m NE of Portledge Mouth 2 m of red-stained silty shale rest on 2.45 m of massive red sandstone, in beds up to 1 m thick with sole marks and muddy nodular inclusions, on 1.6 m of red broken silty shale with thin sandstone beds, on 0.5 m of red sandstone, dip 40°/195°. A further 100 m from Portledge Mouth 1 m of massive pink sandstone rests on 1 m of sandstones in shale, on 0.5 m of massive sandstone, on 2 m of red sandstones in beds up to 0.3 m thick with striped silty mudstone partings [SS 3885 2489]. The dip is 70°–90°/195° but may have been affected by landslip. Another section [SS 3890 2495]–[SS 3893 2497], dip 80° NE, shows in descending sequence:

Thickly bedded, broken and landslipped sandstones crop out in the cliff to the north-east [SS 3893 2497]–[SS 3899 2504]. Beyond them strata striking E–W are vertical or slightly overturned and terminated to the north-east by a near-vertical fault [SS 3900 2508]; the beds are described towards the north-east and stratigraphically downwards as follows:

A further 30 m to the NE, beyond a section of cliff occupied by landslipped broken sandstone, the following inverted beds, striking E–W and stratigraphically above the previous section, were recorded [SS 3902 2507]–[SS 3903 2509] and are described towards the north-east and stratigraphically downwards as follows:

About 55 m of landslipped and overgrown cliff separate the last exposure from a complex section [SS 3906 2514]–[SS 3907 2516], of uncertain relationship to the preceding sections. The beds are inverted and dip 65°/165°; the northward succession is as follows, beginning with the youngest bed:

An isolated exposure [SS 3909 2519] in Higher Rowden cliff shows 3.4 m of pink sandstones, in beds up to 0.9 m thick, in the hinge region of an anticline. At 30 m farther NE [SS 3911 2523] another exposure reveals 1.6 m of red-banded silty mudstone, on 1.0 m of sandstone with shale partings, on 3.5 m of massive pink sandstones with striped silty mudstone partings, in a close syncline.

A further 35 m to the NE [SS 3912 2527], N-dipping beds in the cliff comprise part of the northern limb of the anticline which lies north of the syncline seen in the last exposure; the northward succession is:

An anticline [SS 3923 2534] near the northern end of Rowden Cliff brings up 1.5 m of massive pink sandstone, on 1.5 m of sandstone in 0.5-m beds with striped sandy mudstone partings, on 1.5 m of massive sandstone.

Babbacombe Cliff is almost entirely obscured by active landslip, but an inverted sequence [SS 3942 2556] shows 2.0 m of thinly bedded sandstones stratigraphically overlying 1.6 m of thickly bedded orange sandstones with bottom structures, on 0.6 m of striped sandy mudstone, on about 7.0 m of orange-red thickly bedded sandstone. Immediately north-east of the outfall of the Babbacombe stream [SS 3948 2565] southerly-dipping beds in the southern limb of an anticline comprise, in descending stratigraphical sequence:

The northern limb of the anticline is truncated by a fault [SS 3950 2570], and between it and another fault [SS 3954 2575] 60 m of cliff section are occupied by inverted beds dipping 60°–70°/175°. These beds occur in normal succession north-east of the latter fault, forming the southern limb of a northerly-overturned anticline [SS 3954 2575]–[SS 3956 2578]; the section is as follows, in northward succession beginning with the youngest bed:

North-eastwards from here [SS 3956 2578]–[SS 3960 2585] a sequence involving a slumped bed is repeated and disrupted by complex minor faulting; the succession is as follows in descending order:

Between this fault and the next [SS 3964 2588] lies a tectonically complex succession, in descending order:

Between this fault and the Sticklepath Fault [SS 3969 2597] lies a S-dipping Bude Formation sequence with probably the same slumped bed near the top, described in descending order northwards:

Cockington Cliff to Greencliff

Just north of the normal fault [SS 3924 2609] bounding the Crackington Formation in Cockington Cliff 200 m of the cliff show overturned beds dipping 75°–80°/175° as far as [SS 3985 2627]. This succession is repeated by folding and faulting to the north as far as [SS 3987 2632], and farther north to [SS 3996 2656] the same strata are exposed, dipping southwards at various angles and folded into tight periclinal folds on the foreshore; the composite succession is as follows, beginning with the youngest bed:

Northwards along the coast from here to the Cornborough Sandstone [SS 4069 2723] the cliff is reduced in height and the section much interrupted by landslips. The succession youngs southwards and the discontinuous sequence, stratigraphically downwards, is as follows:

The following 5.75-m sequence is exposed in a prominent crag on the foreshore [SS 4033 2605]:

This [SS 4046 2687] is the base of Prentice's (1960b) Cockington Beds, where they rest conformably on his Greencliff Beds.

Northwards from here to the base of the Bude Formation the succession comprises dark grey to black shales, with pale grey siltstones. The sequence is difficult to piece together because of folding and faulting, but generally dips and youngs southwards. About 60 m of beds crop out northwards to a stream [SS 4050 2694], and are separated by a fault from another southward-younging sequence consisting of up to 120 m of black shales and pale grey siltstones to the north. This latter sequence is fault-bounded at its northern end [SS 4060 2710], beyond which the dip reverses to northwards and 32 m of dark grey cleaved shales with ironstone nodules are overlain by laminated pale grey siltstones with slump structures, the boundary [SS 4062 2712] being 28 m S of a limekiln. Traces of a soft coaly culm bed can sometimes be seen at the mouth of the stream [SS 4063 2713] 5 m S of the limekiln when a stormy high tide has removed parts of the storm beach. This may be the Paint Seam (p. 116). At the top of the siltstones are 2 m of thin- to medium-bedded turbiditic sandstones, which are faulted [SS 4067 2722] against 55 m of S-dipping dark grey to black laminated shales, silty in parts and with ferruginous nodules. At the base of these shales is the well-known culm bed, here squeezed and contorted. Fragments of soft pale grey silty seatearth, with rootlets, occur below the culm bed, but have not been found in situ. Some 1 to 2 m of black shale yielding Planolites sp.separate the culm bed from the top of the Cornborough Sandstone. B JW

Inland sections

Blagdon Farm to Exmansworthy Cliff

A small cutting beside the farm road [SS 2352 2726] exposes 24.8 m of thin- and medium-bedded sandstones with groups of siltstones and shales. The dip is 62° N, becoming vertical at the northern end of the section.

Outcrops near the car park at the cliff top [SS 2345 2749] show sandstone and a thick horizon of black shale; the latter has been flexed by a dextral wrench fault. Opposing dips of the sandstones indicate a synclinal structure.

At West Titchberry Farm temporary exposures in trenches [SS 2430 2734] showed 3 m of massive sandstone and 3 m of black shale with thin sandstones. At the top of the eastern end of East Titchberry Cliff [SS 2496 2747] a southerly dipping succession includes a thick black shale ((Figure 11), column 5) which can be traced eastwards in the top of the cliff to the eastern end of Gawlish Cliff [SS 2550 2745]. Parallel exposures south of the cliff edge suggest repetition of the same shale in the southern limb of a large syncline, possibly the major syncline which trends through Hartland Point and Barley Bay. Thick shale at the cliff top at the northern end of Exmansworthy Cliff [SS 2751 2728] also lies in a large syncline and is probably the same horizon.

Swansford Hill to Milford Mill

A small quarry [SS 2264 2306] facing Swansford Hill exposes 2.4 m of medium to pale grey, laminated silty shales and thin sandstones dipping 30°/213°.

The northern fault boundary of the Bude Formation crosses the Milford Water [SS 2261 2352] south-east of Speke's Mill Mouth. Several small faults occur upstream in a succession of thin- and medium-bedded sandstones with predominantly southerly dips. A waterfall [SS 2266 2346] owes its origin to a thick composite sandstone which dips 84°/356°. The same sandstone is overlain by shales with thin sandstones in the side of the trackway 40 m E of the fall. Thin- and medium-bedded sandstones and occasional thickly bedded sandstones, with silty partings and with some thicker shales and siltstones, extend upstream for 450 m to [SS 2294 2317]. The beds in this section have been thrown into a series of anticlines and synclines, but in general probably young from south to north like the strata on the coast. Isolated exposures immediately upstream show thin- and medium-bedded sandstones with silty partings dipping 50°–70° S. RTT

Buck's Cross to Horns Cross

Some 1.95 m of thickly bedded turbiditic coarse-grained sandstone with bottom structures were seen in an old pit [SS 3295 2292] at Downland. Farther east about 10 m of thickly bedded sandstone dipping north were poorly exposed in a large overgrown quarry [SS 3542 2316] by the side of the road leading to Buck's Mills, and in a roadside exposure [SS 3567 2346] at Buck's Mills 1 m of medium-grained sandstone was seen to underlie 2 m of grey shale with thin siltstone wisps dipping 70°/010°. At Holwell a disused quarry [SS 3785 2300] showed about 5 m of thickly bedded sandstone dipping 65°/110°

Fairy Cross to Littleham

At Fairy Cross a stream section [SS 4095 2456]–[SS 4105 2451] revealed thickly bedded massive sandstone, about 3 m, successively overlain by grey silty shale with thin sandy bands, 2.5 m, grey thinly bedded silty sandstone 1.5 m, dark grey banded shale, 0.25 m, and massive sandstone, 0.3 m, all dipping 55°/185°. Sporadic exposures of thickly bedded sandstone crop out in the valley of the River Yeo for 2 km upstream of Yeo Vale [SS 4217 2344].

A stream bed [SS 4263 2490] at High Park shows 1.2 m of dark grey contorted shale underlying 3.5 m of thickly bedded sandstone. In a cart track [SS 4338 2317] running south-west from Littleham thickly bedded sandstone was recorded overlain by 8.6 m of dark grey shaly mudstone dipping 55°/170°. A drainage trench [SS 4514 2470]–[SS 4520 2469] revealed a succession dipping 70°/355° and comprising 3 m of thickly bedded crumbly sandstone, successively overlain by 2.1 m of dark and pale grey banded silty shale, 1.4 m of thinly bedded sandstone, 0.1 m of black shale and 0.3 m of graded sandstone with flute casts on the base.

Landcross to Huntshaw Mill

A cutting on the west side of the A386 Torrington to Bideford road, between [SS 4581 2323] and [SS 4580 2343], exposed Bude Formation strata dipping 50°/180°, as follows:

A roadside quarry [SS 4588 2356] 150 m to the north contains the following sequence dipping 55°/195°:

A further 100 m to the NE, along the roadside [SS 4597 2364] a 2-m brown-weathered massive sandstone, dipping 20°/215°, is overlain by 2 m of grey silty shale with thin sandstone beds.

Farther east, near Venton, 2 m of massive thickly bedded sandstone underlie 1 m of thinly bedded sandstone and 1.25 m of grey silty mudstone, in a stream section [SS 4883 2276].

Greencliff-Bideford-Gammaton

In this area a belt of Bude Formation rocks lie south of and above the Bideford Formation. There. are many small exposures of soft pale grey siltstone, but there are few quarries owing to the predominance of soft siltstones and shales over sandstones. A roadside cutting [SS 4141 2669] at Rixlade contains about 2 m of dark grey shaly mudstone below 1.35 m of massive graded sandstone with bottom structures, dipping 45°/185°. In temporary excavations [SS 4366 2625] at Moreton House dark shales dipping 55°/180° are overlain by 2.25 m of coarse-grained medium-bedded sandstone below about 8 m of soft pale grey siltstone. East of the River Torridge, foundation excavations for new houses [SS 4650 2607] revealed a thickness of about 18 m of soft pale grey siltstones, disposed in irregular contorted folds. A roadside cutting [SS 4846 2579] 400 m NW of Gammaton exposes 5.5 m of grey shaly and silty mudstone with some grey siltstone beds up to 90 mm thick, dipping 60°/185°. BJW

Chapter 5 Permian

General account

Rocks of presumed Permian age occur in a small faulted outlier on the coast, near Portledge and Peppercombe, the outcrop extending only about 2 km inland. The rocks consist of breccias with beds of sand and sandstone, and scattered calcareous bands and nodules ((Plate 11) and (Plate 12)). They rest unconformably on red-stained Bude Formation strata, and are faulted against them at the northern margin of the outlier. The unconformity can be seen to truncate the folding and quartz-veining in the Bude Formation rocks.

There is no direct evidence as to the age of these deposits. A sample from the coastal section [SS 3853 2457] was found by Dr G. Warrington to be barren of palynological material. There is, however, a general resemblance between these breccias and many of the Permian breccias of south and central Devon. Ussher (in manuscript) noted the similarity to the 'basement breccias of south Devon'; he also said that some of the Portledge breccias resembled 'brecciated Keuper Sandstones locally present in the Bridgwater area'.

Red staining of the Upper Carboniferous rocks is intense around the outlier, especially near the basal unconformity, and is also much in evidence along fault-zones and joints. Reddened areas around and east of the Permian inlier were noted and mapped by Prentice (1960b, p. 403). He considered that the Portledge outlier, and the red-stained area north of it, were bounded to the north by strike faults which were responsible for the preservation of the Permian rocks, and suggested that faults might bound other areas of reddening. South and west of the Portledge–Peppercombe area, red staining is visible in the upper part of the cliffs to the east of Clovelly, as far as Freshwater Fall (Arber in discussion of Prentice, 1960b, p. 405). The staining is evidently superficial and does not greatly affect the rock at the base of the cliff. From Clovelly westwards a belt of reddened soil and Head can be traced as far as the south-western outskirts of Hartland village. In this Clovelly–Hartland area there is no evidence of a strike fault, though it is possible that one exists along the line of the Abbey River. It is noteworthy that the Clovelly–Hartland belt of staining lies near to the line of a notional westerly continuation of the Portledge outlier and might reflect an original extension. The slight offset could be accounted for by wrench faulting, and the difference in height between the base of the Portledge Permian and the probable base in the Clovelly–Hartland area, some 120 to 150 m, could be attributed to faulting or to the form of the Permian land surface.

Evidently the Permian cover was once much more extensive, as shown by the reddened areas, and Prentice (1960b, p. 403) thought that Permo-Triassic or reddened Upper Carboniferous rocks might be present beneath Northam Burrows. He cited records of red sandstones from a well at Westward Ho! and of deep red water from a borehole on the Pebble Ridge. Also a seismic refraction survey by McFarlane (1955, p. 426) has suggested a possible east–west trough of Permian sandstone about 550 m wide beneath Northam Burrows. BJW, RTT

Details

At the western (coastal) boundary of the Portledge outlier the Permian rocks rest unconformably on Bude Formation strata [SS 3802 2425]. The breccias dip 25°/022°; the basal 2 m are very pebbly, and are overlain eastwards by 15 m of breccias with sandstone partings, the angular to subangular fragments being composed of Carboniferous sandstones, siltstones and silty mudstones. The sandy matrix is usually bright red and fairly coarse grained. Sporadic calcareous concretionary nodular bands occur up to 0.1 m thick. The succeeding 2 m consist of very coarse breccia, with fragments up to 0.25 m across; they are in turn overlain by 14 m of breccia with sandstone bands, with a 0.5-m calcareous nodular band 3 m above the base. At [SS 3817 2430] the breccias are thrown against Bude Formation strata by the Peppercombe Fault, but on the foreshore they can be traced across the fault to where the unconformity is again visible in the cliff [SS 3836 2443].

There is uninterrupted exposure between here and the Portledge Fault [SS 3870 2472]. The beds dip around 25°/020° and the succession is as follows:

Inland exposures are very poor, though 1.5 m of red breccia with bright red sand lenses were seen in the stream bed [SS 3930 2467] 140 mat 250° from Portledge House. BJW

Chapter 6 Igneous rocks

General account

Greenstone dykes have been recorded in Lee Bay, 2 km east of Bull Point. Arber (1911) noted the reported presence of three, to be seen only at extremely low tides. Blundell (1957) located one and concluded from measurement of remanent magnetisation that the material represented Tertiary intrusion. It seems probable that the dykes are visible only rarely, when a very low tide coincides with suitable disposition of shingle; one, seen near low water mark [SS 478 467], strikes east.

All the other igneous rocks of the district, if we exclude coastal erratics, occur on Lundy Island, most of which consists of granite. Both the granite and the adjoining slates have been intruded by many small dykes, among which the main rock types are olivine-dolerites, but some fine-grained basaltic variants occur. The more acid dykes range from glassy to fine grained. Dollar (1942) distinguished pitch-stones, trachytes and orthophyres. Lundyite (Hall, 1915) is a soda-rich orthophyre characterised by soda-amphibole.

The age of the granite has long been in dispute. Petrological arguments have been advanced to support a Permo-Carboniferous age (Williams, 1849; Hall, 1871; Harker, 1929; Dollar, 1942) and a Tertiary age (Judd, 1874; Thomas and Richey, 1930; Guppy and Thomas, 1931). More recently Blundell (1957) thought the magnetisation of the granite not inconsistent with a Permo-Carboniferous origin. However, isotopic age determinations (Dodson and Long, 1962; Miller and Fitch, 1962), using the potassium-argon method on the minerals biotite, feldspar and quartz, have shown the Lundy Granite to be around 52 million years old and hence probably middle Eocene in age. Most workers in recent years have favoured a Tertiary age for the dykes (Guppy and Thomas, 1931; Dollar, 1942; Blundell, 1957) and radiometric dating of the granite has established this as correct.

Dollar (1942) distinguished two main granite types, both usually medium- to coarse-grained biotite-muscovitegranites, which he believed to represent separate intrusions, the earlier type (G1) being lifted and fractured by the magma of the later (G2). He described his G1 as showing large single or twinned perthitic orthoclase crystals with a little albite-oligoclase, interstitial quartz and quartz-feldspar intergrowths, and with large plates of muscovite and laths of brown biotite held in a mesh of phenocrystic albite. Accessory zircon, apatite, pyrrhotite and iron oxides occur, together with possible monazite and sphene. Secondary minerals include garnet, sericite, chlorite, cordierite, anatase, epidote, zoisite, tourmaline, topaz, fluorite, pinite, beryl, rutile, cassiterite, copper minerals, calcite and kaolin. Dollar (1942) described his G2 as containing large phenocrysts of feldspar and smaller ones, up to 4 mm across, of bipyramidal quartz, plates of intergrown biotite and muscovite supported in a mesh of perthite, and subordinate albite and quartz. Accessory zircon and monazite are present, and secondary garnet, chlorite, kaolin, sericite, tourmaline, topaz and fluorite.

The main criterion for distinction is thus the presence or absence of fairly small crystals of bipyramidal quartz, and the evidence presented for two separate major periods of intrusion seems inadequate. Sharply defined small xenoliths occur, and such apparent contact lines as may be traced within the main granite are most likely to reflect the fragmentation and engulfing of a consolidating crust (p. 71). Comparisons may be made with the interpretations of the history of the Dartmoor Granite (Edmonds and others, 1968) and the probably similar origins of the Cornish granites (Edmonds and others, 1975).

Thin veins and dykes of fine-grained granite are fairly widespread. They have invaded the main granite as late-stage intrusions but are nowhere of mappable extent. Dollar (1942) distinguished a 'Central and Southern Micro-granite (G3)', of aplitic type, and a 'North-Western Micro-granite (G3a)', usually rich in biotite. It seems scarcely necessary to force a variety of small fine-grained late-stage granitic intrusions into a dual classification, particularly one with a possible implication of age relationships. Nevertheless Dollar's descriptions of microgranitic rocks are a valuable record. He noted that all types locally developed pegmatitic margins containing druses. His G3 is described as a nearly equigranular intergrowth of albite, orthoclase and quartz with pale and dark micas. Rare phenocrysts of albite occur, as do large irregular masses of quartz, variable amounts of topaz, sericite as an alteration product of feldspar, and local tourmaline. Fine banding by felsic and dark mica-rich layers is present in places Marginal pegmatitic developments locally include leaves of molybdenite. Dollar (1942) described his G3a as a dark-coloured streaky fine-grained mosaic intergrowth of quartz and feldspar with bands and patches of biotite. He considered it richer in biotite than his G3 and noted that it contained smaller amounts of muscovite, orthoclase, albite and quartz. A little cordierite occurs sporadically in the quartz, and zircons in the biotite. Topaz is associated with mica and feldspar, and some of the latter shows alteration to sericite.

It seems likely that all the granitic rocks of Lundy shared a common source. Analyses quoted by Dollar (1942) point to slight changes taking place in the composition of the magma, such as increasing soda content and falling lime content, but the chemical and mineralogical similarities of the rocks are more striking than the differences. Thus the only major point at issue is whether two distinct main granites may be distinguished or whether the evidence points to a continuous process of successively formed, wholly or partly consolidated crusts being overtaken, broken, engulfed, and more or less assimilated by later magmas whose slightly changed composition reflected the slow processes of differentiation. Field evidence suggests that no valid boundary can be drawn. The presence of xenoliths near the slates attests contamination by country rock, and it seems most likely that dissimilar rock types within the granite, if we exclude the small late-stage intrusions, owe their origin to absorbed xenolithic material, either sedimentary or igneous.

Dollar (1942) recognised the possibility that the Lundy Granite had been produced in Tertiary times by palingenesis at depth near the northern margin of the Cornubian batholith. However, problems of accounting for the heat-source, and his conclusion that stress directions related to flow matched those of the petrographically similar mainland Permo-Carboniferous granites, led him to abandon the idea in favour of the concept that a differentiated partial magma was squeezed off northwards from the main parent magma in Variscan times. In the discussion of the paper both A. K. Wells and O. T. Jones pressed the likelihood of a Tertiary age for all the igneous rocks of the island, a view first advanced by Judd (1874) and now confirmed radio-metrically.

Dollar considered that the magma may have risen along fissures running north-north-eastwards from the parent source. No firm evidence exists, but the shape of the island and of the sea floor in its vicinity suggests that there is no great submarine outcrop of granite. On geophysical maps the granite shows up as a residual gravity 'low', and prominent aeromagnetic anomalies immediately to the south and west of Lundy confirm that the granite does not extend far beyond the island. The margins of the upper part of the intrusion are thus probably steep, but the relationship of the visible granite to its feeder channel is unknown; it could be almost dyke-like, or the magma could have spread laterally from a narrow fracture.

Numerous dykes are mainly radially disposed around the granite. They range in thickness up to 2.3 m and intrude both slates and granite, but are generally visible only in coastal cliffs. In the field it is scarcely practicable to establish any finer division than that between basic dykes and the rest. The former are about 200 in number, typically around 1 m thick and consist of dark grey or black ophitic dolerites, commonly showing spheroidal weathering and shedding a brittle brown oxidised crust. They are characteristically readily eroded by the sea and thus generally marked by narrow gullies in the cliffs. The remaining 32 dykes are fine-grained to medium-grained pale grey rocks weathering to buff and characteristically carrying small rectangular feldspar crystals in a fine-grained matrix. They are of acid or intermediate composition and broadly of trachytic rock types. Both sorts of dyke are sharply defined, commonly with glassy margins, and both are locally vesicular in their interiors. The little evidence that exists of their mutual relationships suggests that the more acidic dykes generally cut the more basic but that the reverse also occurs.

The nature of the magmas which gave rise to the granite and to the two types of dyke, is open to conjecture. Dollar (1942) opted for an acid sodipotassic, perhaps originally sodic, granitic material, but a marginal differentiate of a parent magma could well have been of more basic composition. Evidence from other Tertiary igneous provinces ranges from the simple sequence of magma types deduced by Harker (1904) on Skye and Rum, ultrabasic grading to acidic during the plutonic phase and reverting to ultrabasic at a later stage and lesser depth, to the confusing multiplicity of alternating magma types on Mull and Ardnamurchan. Whatever the fundamental petrogenesis, however, an association of Tertiary granites with late-stage minor basic dykes is common. In western Scotland and the adjacent part of northern Ireland the dyke swarms show a striking NW–SE alignment; on Lundy they show a slight west-north-westerly regional trend, but are typically radially arranged and thus less obviously related to the main lines of Tertiary faulting. They appear to have risen rapidly from a basic magma lying at fairly shallow depth, since little contamination or layering is evident. Dykes of trachytic type may reflect changing magmatic compositions or injection of liquors more nearly related to the granitic magma, and it is relevant to recall the presence of trachytes and microsyenites among the minor Tertiary intrusives of western Scotland, and the identification of orthophyres on Skye.

The genetic sequence of igneous rocks on Lundy accords best with the coexistence at some stage of two separate liquors, one acidic, the other basic. This could have been effected by the separating off, and subsequent differentiation, of a basic magma from the Cornubian reservoir source. But this implies the existence at depth of a major basic pluton in the vicinity or south of Lundy of which there is no direct evidence. However, Brooks and Thompson (1973) interpreted a positive Bouguer anomaly centred 10 km WNW of Lundy as indicating just such a body, and Tertiary dyke material has been inferred in a seaward extension of the Sticklepath Fault Zone (see below). Perhaps the Cornubian reservoir was tapped twice, the interval between being of either space or time, and separate acid and basic magmas resulted. Whatever the origins, subsequent events resulted in the emplacement of granite, locally contaminated by assimilation, and injection by its late-stage acidic fluids, and finally of the many dykes, the doleritic perhaps mostly preceding the trachytic.

There remains the question of why the Lundy intrusion is where it is. Dollar (1942) conjectured that the locality marked the intersection of two zones of weakness in the Palaeozoic crust, a N–S Tenby–Fowey zone and an E–W Salisbury–Ilfracombe zone. He envisaged the magma rising via a NNE–SSW fracture, up-arching and assimilating sediment, forcing apart the walls of the fracture, and consolidating as a thick, possibly laccolithic, sheet which was not less than 8 km long and 5 km wide. It now appears (Owen, 1971) that Lundy is on a NW–SE upstanding block or horst, bounded on its north-eastern side by the undersea extension of the Sticklepath Fault Zone and on its southwestern side by a fault of similar trend, possibly the Cambeak Fault (Dearman, 1964). Thus the island does not lie at a Tertiary focus of crustal weakness, although such a conjecture relating to an original magmatic separation in Variscan times cannot be excluded. Brooks (1973) related Tertiary igneous activity in western Britain to a hypothetical underlying hot spot in the Earth's mantle.

The most likely routes for rising Tertiary magmas were those channels provided by the NW–SE fault lines.

Hamilton (1971) reported that seismic profiles indicated strong faulting of Mesozoic rocks to the north and west of Lundy. Probably the main intrusions were emplaced by middle Eocene times, before the NW–SE Tertiary movements reached their maximum. The horst and graben structure would then be ill defined, and the probable presence of igneous rocks other than those of Lundy on the floor of the Bristol Channel has been demonstrated in recent years. Donovan and others (1971) noted that dredged samples from the Horseshoe Rocks, 5 km NW of Ilfracombe, included Devonian rocks and dolerite; they suggested that the sheared and slightly metamorphosed character of the latter pointed to an Upper Palaeozoic age, but Brooks and Thompson (1973) favoured a Tertiary origin. More striking is the presence of an elongate negative aeromagnetic anomaly extending north-westwards from a point some 10 km NW of Lundy. This has been interpreted as indicating a 30-km-long Tertiary dyke or dyke swarm dipping steeply south-westwards and occupying the Sticklepath Fault Zone (Cornwell, 1971). The problem of deriving separate acidic and basic magmas directly, or at one stage removed, from a Cornubian reservoir remains. But it is suggested that their subsequent rise was controlled by the NW–SE Tertiary fault lines which so rent south-west England. EAE

Dead Cow Point [SS 1272 4519], where deep clefts have formed and huge blocks have slipped seawards.

Horizontal and low-angle joints parallel the topography and are probably due to load-relief during erosion (Figure 12). They are well displayed on the crags above the upper parts of the cliffs on the west side of the island between The Old Lighthouse [SS 1320 4429] and North West Point [SS 1300 4820].

Lundy Granite

The plane of junction between granite and sedimentary rocks on Lundy trends NNE–SSW. Although not exposed inland the boundary seems to be straight or only slightly curved. The contact in the cliffs at the western side of Rattles Anchorage [SS 1370 4355] may be normal, but at [SS 1381 4370] the junction has been invaded by a basic dyke and some disruption and brecciation of both sediments and granite suggest that the dyke followed a fault. In a gully at the back of Ladies Beach [SS 1402 4426] two dolerite dykes occupy the junction zone; to the south-east the slates are almost unaltered, except for baking due to the dykes, but to the north-west the granite is shattered and crushed for a distance of about 10 m from the dykes. Thus it is possible that, from the south coast to Ladies Beach, the junction is faulted.

Such faulted relationships accord with the absence of granite veins from the slates and with the lack of evidence of contact metamorphism. The granite shows little sign of having been chilled but, as some granites maintain coarseness of grain-size close to their margins, this does not necessarily indicate great distance from the contact. Xenoliths are present in exposures on the south and south-west coasts, which may indicate proximity to the margin or the roof.

Dollar (1942) found that trends of steep joints showed a concentration between north-west and north-north-east. Measurements during the present survey indicated some concentration between 110° and 125° and between 170° and 180°, but there is much variation. Joints with a northerly trend are particularly well developed on the cliffs between Pilots' Quay [SS 1300 4395] and Goat Island [SS 1315 4375] and between Needle Rock [SS 1292 4557] and The Lundy intrusion is composed mainly of one rock type, a coarse-grained megacrystic granite. This Dollar (1942) designated 'G1' and he also noted a syenitic variant. The proportion of feldspar megacrysts varies from 1 to 20 per cent, although locally the granite is non-megacrystic. Megacryst length ranges from 15 to 70 mm, the average being about 25 mm. The groundmass is coarse, having an average grain size of 1 to 2 mm.

Second in terms of areal extent, and probably also in terms of volume, is a fine-grained megacrystic granite which crops out in many places but is characteristic of certain areas. Dollar (1942) named this type 'G2'. It occurs as irregular pods and patches within the coarse granite and may represent fragments of a consolidating crust engulfed by advances of the same magma. Alternatively this xenolithic material may be of sedimentary origin and result from the incorporation and granitisation of argillaceous and silty-argillaceous material (Edmonds and others, 1968, 1975).

Dollar (1942) recorded the presence of a certain amount of 'mixed' rock on North West Point, which he ascribed to the invasion of solid G1 granite by the magma of G2 granite. A similar rock type can be seen in association with fine-grained megacrystic granite 200 m SW of The Old Lighthouse and on the upper cliffs above the Mousehole and Trap. It seems likely that this mixed rock arose from the contamination of the main granite by almost completely assimilated argillaceous sediment.

Bodies of fine-grained poorly-megacrystic granite are fairly widespread–the G3 and G3a intrusions of Dollar (1942). They occur as vertical, inclined or horizontal sheets, rounded pods or irregular masses. Commonly their margins are sharp but irregular, although locally they grade into both coarse-grained and fine-grained megacrystic granite. Inclined and horizontal sheets average 0.5 m in thickness but may reach 3 m. Vertical sheets occur singly or in swarms, and normally they are no more than 1 m thick. In several instances a vertical sheet is seen to cut a horizontal sheet. Pods range up to 0.3 m across.

Pegmatites are generally associated with horizontal or gently inclined sheets, but some combined bodies pass laterally into pegmatite alone. They are commonly developed at the upper margins of sheets but also occur within them.

Main granite

The feldspar megacrysts are euhedral or subhedral, flattened parallel to the (010) crystallographic plane, and commonly Carlsbad-twinned. Sodium-cobaltinitrite staining of flattened surfaces indicates that almost all the megacrysts are potassium feldspar, and staining of thin sections shows that these are perthites. The albite phase takes the form of equant and tabular patches (E42763), irregular veins or stringers (E42770) or both (E42776). The margins of the perthite megacrysts are usually very irregular as the crystals have outgrown the areas of their original plagioclase hosts to incorporate feldspars of the groundmass. Commonly the megacrysts abut against irregular quartz and mica crystals which have limited their growth. Plagioclase crystals have in some instances had the same effect. Inclusions of groundmass minerals within the megacrysts are common.

Some feldspar megacrysts ((E42763); (E42770); (E42776)) show zones of quartz blebs, as described by Shelley (1966). The quartz in these zones may be accompanied by muscovite ((E42780); (E42806)) or plagioclase ((E42770); (E42780)). In one example (E42780) up to four zones are present. Most of the zones are composed of rounded or vermicular quartz, but innermost zones are commonly continuous (E42805). The inner margins of zones are normally even, whereas outer margins are irregular. In one specimen (E42763) a zone outlines the original plagioclase crystal within a potassium feldspar megacryst.

The quartz megacrysts occur in similar abundance in all samples of the coarse megacrystic granite, ranging in size from about 3 to 10 mm, and averaging between 4 and 6 mm across. Typically the quartz megacrysts consist of several smaller anhedral crystals aggregated together and joined along sutured margins by the addition of silica. It seems that original groundmass quartz acted as a nucleus for subsequent growth. The additional quartz is strained (E42763) and internally sutured (E42738) but in optical continuity with the original crystal. Generally however each individual part of a megacryst has a different orientation and a very irregular margin. Clearly the quartz megacrysts are of late formation. Inclusions of biotite, plagioclase, potassium feldspar and muscovite occur but are uncommon.

The principal minerals of the groundmass are potassium feldspar, quartz, plagioclase, biotite and a colourless or pale-coloured mica, possibly phlogopite. Potassium feldspar is present in subhedra and anhedra which range from 2 to 15 mm. They are patch- and string- or veinperthites and are commonly Carlsbad-twinned ((E11308); (E42770); (E42801)). Sodium-cobaltinitrite-stained thin sections reveal that these patch-perthites result from the replacement of pre-existing plagioclase crystals (E42801). Specimen (E42805) illustrates that replacement commenced in the centres of crystals and proceeded outwards. In some cases (e.g. (E42806)) replacement stopped just short of the crystal boundary but generally perthites have grown beyond the margins of the partially replaced plagioclase crystals, forming outgrowths and linking up with other groundmass feldspars.

Quartz occurs in the groundmass as anhedral crystals ranging in size from 0.1 to 3 mm, and commonly having undulose extinction (E42805). Commonly the margins of the quartz crystals are very irregular and much of the quartz is clearly interstitial. Clusters of quartz grains may be fused together forming incipient megacrysts. Typically the groundmass quartz is free from inclusions, rare biotite flakes being an exception. Quartz commonly occurs as inclusions in feldspar megacrysts and groundmass minerals. It can be in the form of blebs and vermicules in potassium feldspar megacrysts and ragged inclusions in groundmass plagioclase.

Plagioclase in the groundmass normally ranges from oligoclase to sodic andesine, is commonly Carlsbad- and lamellar-twinned and rarely pericline- or chess-boardtwinned (E42770). In a few specimens (E42750) the plagioclase is compositionally zoned. Many crystals are tabular subhedral, ranging in length from 0.6 to 10 mm; some are lath-shaped and up to 5 mm long. Commonly plagioclase shows some alteration to sericite, normally in the central regions of crystals where alteration may be intense, but also towards the margins. In zoned crystals the alteration may be confined to the central zones (E42776) and can be heavy (E42806). No myrmekitic intergrowths were noted but rare crystals (E42806) show a zone of quartz blebs surrounding a heavily sericitised core. A few plagioclase crystals show partial patchy replacement by potassium feldspar. Plagioclase commonly occurs as inclusions in potassium feldspar megacrysts and groundmass crystals.

Biotite is a common groundmass constituent forming subhedra and anhedra between 0.3 and 5 mm across. Crystals occur singly or in clusters (E42770). Biotite locally shows alteration to muscovite (E42801), generally along cleavage traces, or to chlorite (E42776). It is commonly intergrown with quartz (E42806). Pale brown faintly pleochroic mica, probably phlogopite, occurs in some specimens. Muscovite is present in many specimens, commonly as an alteration product of biotite, where it may retain the haloes of radioactivity damage caused by included zircons ((E42738); (E42744); (E42801)).

Dollar (1942) listed the following accessory minerals: anatase, apatite, beryl, cassiterite, cordierite, epidote, fluorite, garnet, monazite, pinite, pyrrhotite, rutile, sphene, topaz, tourmaline, zircon, zoisite. An additional mineral, belonging to the columbite-tantalite series (X7005), was identified by Mr R. J. Merriman, who also confirmed the presence of monazite (X7003) and topaz (X7002). Apatite (E42748), garnet (E42763), topaz (E9247) and zircon are visible in thin section and garnet and tourmaline can be seen in some hand specimens. Tourmaline forms small clusters. Sparse opaque iron oxide occurs in the ground-mass as anhedra about 0.1 mm in size and as inclusions in biotite ((E42763); (E42776)). Rastall and Wilcockson (1915, p. 617) stated that pyrrhotite comprised the whole of the magnetic residue.

Zircons are commonly present in biotite ((E42771); (E42776); (E42780); (E42805)), attaining a size of 0.15 x 0.06 mm (E42770). Opaque iron oxide inclusions within biotite are rare (E42763). Garnet, pale pink in thin section and red in hand specimen, is probably almandine. Commonly the mineral is altered to muscovite along cracks and at the margin.

Two analyses of the coarse granites-one nonmegacrystic (E42748) and the other having 11.7 per cent megacrysts (E42806)—are given in (Table 5).

Six samples of coarse granite, ranging in megacryst content from less than 1 per cent to 11.7 per cent, were analysed for beryllium by Mr D. J. Bland. The gamma-neutron method described by Bowie and others (1960) was used and the results show a range of values from 10 to 29 mg/kg Be. The mean of 18 mg/kg Be is considerably more than the mean for all granites of 5 ppm Be (Beus, 1961). It is slightly higher than that of the Dartmoor Granite (Edmonds and others, 1968).

Modal analyses were made on nine stained specimens, using a macro point counting technique and also (by Dr R. Dearnley) using an image analysing computer. Results show that the principal rock-type of the Lundy intrusion is an adamellite, only one specimen having a ratio of potassium feldspar to plagioclase of more than two to one. Mineralogical means and ranges for the nine specimens are as follows:

The mean ratio of potassium feldspar to plagioclase is 1.3:1 and the ratios range from 2.3:1 (granite) to 1:1.5 (adamellite).

Syenite

Two specimens from Shutter Point ((E9230); (E11309)) and one from North West Point (E11303) are syenites. These are hypidiomorphic granular rocks composed mainly of feldspar with a little mica. The potassium feldspars are similar to those of the main granite. Plagioclase is in the albite range. Generally the feldspars are sericitised. Micas include pale brown pleochroic biotite and colourless varieties. The specimen from North West Point contains a few very small quartz crystals. One specimen (E9230) has an average grain size of 2 mm with a few potassium feldspars up to 10 mm in length, but the other specimens are finer with an average grain size of 0.5 mm.

Fine-grained megacrystic granite

The mineralogy of the fine-grained megacrystic granite is similar to that of the main coarse granite, but the grain size and texture of the groundmass differ appreciably. The feldspar megacrysts are almost all potassium feldspars with features similar to those of the perthite megacrysts of the coarse granite. Megacryst lengths average 20 mm and range from 15 to 50 mm. One or two zones of blebby quartz are present in many crystals ((E42740); (E42781)). Quartz megacrysts average 3 mm across but some attain 10 mm. Each is composed of several crystals which commonly show undulose extinction and are joined with sutured margins (E42781).

The groundmass is composed of potassium feldspar, quartz, plagioclase, biotite and a little muscovite. The potassium feldspar is perthite and the plagioclase is oligoclase. Strongly pleochroic biotite is scattered throughout most specimens. In a few cases the mineral forms aggregates (E42762), and rare single crystals reach several millimetres across ((E42760); (E42781)). Some specimens have very little biotite and one (E9246) has abundant muscovite.

Accessory minerals noted were garnet, zircon, tourmaline, topaz, apatite and opaque iron oxide. Dollar (1942) recorded monazite and fluorite in addition. Garnet occurs embedded in potassium feldspar megacrysts ((E9243); (E42740); (E42781)) or associated with aggregates of chloritised biotite (E42756). One crystal (E9243) is cracked, partly altered to sericite and has a rim of quartz. The pale pink colour of these garnets in thin section suggests that they are almandine. Zircon commonly occurs as inclusions in biotite where they have caused radioactive damage. Apatite and opaque iron oxide are both rare in this rock type and are usually included in or associated with biotite ((E42756); (E42762)). Brown and blue tourmaline is present in several specimens ((E9246); (E11302)), forming a fine network around and through groundmass quartz and feldspar.

The groundmass texture is xenomorphic granular. The quartz, micas and most feldspars are anhedral. Only a few feldspar crystals are subhedral. Typically the groundmass is even grained with an average size of 0.1 mm. A few specimens show a range from about 0.03 to 0.3 mm.

Fine-grained poorly megacrystic granite

This rock type, which forms sheets, pods and irregular shaped masses, carries rare potassium feldspar megacrysts up to 20 mm long and rare quartz megacrysts up to 10 mm in diameter (E42761). The feldspar megacrysts are perthites which have features similar to those of the other granite types except that the zones of quartz and mica crystals have not been observed.

These fine-grained rocks are composed mainly of potassium feldspar, quartz and plagioclase. In most specimens potassium feldspar is slightly kaolinised and in a few, many crystals exhibit more intense alteration. Plagioclase crystals range from albite to oligoclase, show Carlsbad-, lamellar- and pericline-twinning, and commonly have kaolinised cores.

Pleochroic biotite occurs in some specimens (E42745), commonly having minute inclusions of zircon. Muscovite may be present with biotite (E42761) or as the only mica, where it is commonly accompanied by topaz ((E9248); (E42749); (E42767)). One specimen (E42778) has much fine granular topaz and interstitial pale blue tourmaline, with a little pale brown biotite and muscovite.

The rocks have xenomorphic granular texture. Generally the feldspars are anhedral, but some of the smaller plagioclase crystals are subhedral. Quartz crystals are rounded and granular, much of the mineral being interstitial. Some specimens are evenly granular with grains 0.2 and 0.3 mm across ((E42749); (E42767); (E42811)); others show a range from 0.1 to 0.6 mm ((E9229); (E42753)). The remainder have an average grain size of 0.4 mm. Where associated with pegmatites the fine-grained granite commonly exhibits grain-size and mineralogical banding.

Pegmatite

Generally the pegmatites consist of quartz-feldspar inter-growths, with varying amounts of biotite and muscovite. In a typical specimen, from Battery Point (MR31893), 70 per cent of the rock is composed of euhedral to subhedral potassium feldspar having an average crystal length of 15 mm. Most of the remainder is subhedral to anhedral quartz in aggregates and single crystals up to 10 mm across. On crags above Jenny's Cove [SS 1335 4570] and in Gannets' Combe [SS 1345 4735] pegmatites consist mainly of quartz in aggregates and single crystals up to 50 mm in size, with some feldspar and sheaves of biotite and muscovite plates up to 10 mm across.

Dollar (1942) recorded greenish apatite, lithium-bearing white mica and rare molybdenite from these rocks. Euhedral beryl crystals were extracted from a drusy cavity in pegmatite at Battery Point [SS 1277 4490], and acicular black tourmaline occurs as inclusions in quartz from the same body.

Lundy Dyke Suite

The dykes on Lundy may be a south-easterly extension of the 30 km long dyke or dyke swarm postulated by Cornwell (1971) to stretch in a north-westerly direction from a point 10 km north-west of the island (p. 121). If this is so, the swarm would appear to be analogous to those associated with the ring complexes in the Hebrides.

The basic dykes usually have well-developed, close-spaced joints parallel to their walls, with subordinate joints at right angles, while in the trachytic dykes the dominant jointing is transverse and widely spaced. Generally the dykes seem to have been emplaced easily, their verticality and general straightness indicating a lack of resistance to intrusion. Only a few show any marked disruption of the granite into which they are intruded. Where brecciation and granulation do occur, the country rock is affected for no more than a few centimetres from the contacts.

Mrs D. P. F. Darbyshire has made determinations of potassium and argon on whole-rock samples from three dykes. She notes that the ages of the basic dykes accord with palaeomagnetic evidence (Blundell, 1957) and are indistinguishable, within the bounds of experimental error, from isotope ages quoted for the granite (Dodson and Long, 1962; Miller and Fitch, 1962). Her results are given in (Table 6), the errors quoted being estimates of analytical precision only. Mussett and others (1976) reported potassium/argon datings for six dykes as ranging from 44.6±±1 to 54.3±1 Ma. The authors stated that their most reliable date was the earliest, determined on a basalt.

Dolerite dykes

About 100 specimens, representing about half the basic dykes, were examined. They include olivine-dolerites (45 per cent), analcime-bearing olive-dolerites (35 per cent), olivine-free dolerites (10 per cent) and a few quartzdolerites. Some specimens from dyke margins are highly amygdaloidal, others are glassy. There is considerable variation in grain size and texture. Most specimens contain a few scattered phenocrysts and some of each of the main types are porphyritic. Many of the dykes are altered, so that olivine is rarely fresh but commonly replaced by serpentine as pseudomorphs. These last are difficult to distinguish where pyroxene is extensively replaced by chloritic material. Unfortunately localisation is poor for some specimens from early collections.

About 20 specimens were described by Sir John Flett (unpublished notes), who recognised ophitic olivinedolerites, crinanites and quartz-dolerites. Dollar (1942) described these rocks briefly, noting the presence of tholeiitic rocks in addition to Flett 's types. Modal analyses are shown in (Table 7).

Mr R. J. Merriman examined two specimens of weathered dolerite by X-ray diffraction and reports that one of them ((MR31891); X 6856) is composed of plagioclase with a minor amount of a smectite mineral and the other ((MR31909); X 6857) is mainly quartz with minor amounts of feldspar, kaolinite and mica.

Olivine-dolerites

The olivine-dolerites include porphyritic and aphyric types, both showing variation in groundmass grain-size. Typically the rocks are compact and dark grey to greyish black, but some of the porphyritic specimens have an olive-grey or olive-black groundmass.

Phenocrysts are generally of olivine and plagioclase or plagioclase alone, although specimen (E9227) has olivine alone and a few have microphenocrysts of plagioclase. The plagioclase is greenish labradorite in laths 1 to 10 mm long which are euhedral to subhedral, commonly zoned (E11313) and with some inclusions. Some crystals are veined and patchily chloritised (E42786), others relatively fresh (E9202). Three specimens ((E9231); (E11278); (E11279)) contain a few microphenocrysts of labradorite averaging 0.5 mm in length. Olivine phenocrysts ((E9203); (E9222); (E42786)) occur as euhedra and subhedra up to 2 mm across (E9227). Where freshest (E9222) the olivine shows unaltered cores, but in most specimens it is replaced by serpentine or less commonly by iddingsite (E42786). One specimen (E9231) has altered olivine and a few partly altered titanaugite phenocrysts.

The groundmass is composed of plagioclase, pyroxene, olivine, opaque iron oxide and chlorite. The plagioclase, normally labradorite but in some cases andesine or bytownite, is commonly chloritised, some quite extensively (E42786). Crystals are subhedral and generally lath-shaped, ranging in length from 0.2 mm in the finer-grained rocks ((E9231); (E11278)) to 1.5 mm or more in the coarser specimens ((E4608); (E9225)).

The pyroxene is a pale purplish brown titanaugite, except that colourless to very pale green augite occurs in one specimen (E9202). The titanaugite commonly forms equant subhedra or anhedra, ranging from 0.1 to 0.6 mm. Pyroxene also forms plates 0.2 to 0.4 mm across in the porphyritic rocks (E9227) and 0.9 to 1.5 mm in aphyric specimens (E9225). Normally titanaugite shows very little alteration, but in one specimen (E9202) many of the augites are marginally replaced by a serpentine mineral and a few of the larger crystals are more extensively altered.

Olivine occurs as subhedra averaging 0.08 mm in the finer- and 0.5 mm in the coarser-grained porphyritic rocks, but ranging from 1.5 to 2.0 mm in aphyric specimens. Fresh olivine is rare. Normally the mineral is altered to serpentine but in one specimen (E9227) it is partly replaced by carbonate, and in a few cases alteration is to iddingsite (E9223), chlorite, or chlorite with carbonate ((E9231); (E11278); (E11279)).

Opaque iron oxide is common but sporadic throughout the olivine-dolerites. In the finer-grained rocks it occurs as subhedra and anhedra averaging 0.03 to 0.04 mm, generally as discrete grains but also in aggregates. In other specimens (E9227) it ranges from 0.1 to 0.3 mm and is commonly enclosed within titanaugite. In a few specimens ((E9222); (E9223)) larger irregularly shaped frameworks of iron oxide, showing corrugated or serrated edges, appear to end of Quarterwall Cottages have resulted from the alteration of olivine; in one specimen (E9223) they have developed from iddingsite.

In the aphyric rock-types particularly, chlorite forms primary irregular to angular interstitial patches ((E9225); (E11282)). In all olivine-dolerites it may occur as an alteration product of plagioclase, in some chloritisation is extensive (E9233). Many specimens show amygdales up to 2 mm in diameter filled with chlorite or calcite and chlorite (E4616).

In some specimens (E9202) a little potassium feldspar is present. Acicular crystals of apatite up to 2 mm long and about 0.006 mm wide occur in clusters within plagioclase (E4611).

In general the aphyric specimens are coarser than the groundmass of the porphyritic types, and show a more pronounced subophitic texture. Porphyritic specimens normally have an intergranular texture ((E4611); (E11292); (E11300)), although plagioclase may be subophitically related to titanaugite ((E9222); (E11280); (E42757)). Several of the finest-grained specimens are highly vesicular ((E9224); (E11289)) and are probably from dyke-margins.

A few specimens contain xenocrysts of plagioclase and augite, accompanying normal phenocrysts in a normal groundmass. Some augite xenocrysts are deeply corroded (E9220), and the well-rounded plagioclase crystals have ill-defined margins. Most plagioclase xenocrysts are crowded with opaque inclusions but some are only marginally affected. In one specimen (E42741) plagioclase crystals could be either xenocrysts or phenocrysts; they have the dusty band characteristic of some of the obvious xenocrysts, but also outgrowths reforming them into euhedral crystals. Another rock (E11268) contains ill-defined concentrations of titanaugite, chlorite pseudomorphs after olivine and opaque iron oxide needles.

Olivine-analcime-dolerites (crinanites)

Some of the basic rocks contain analcime and other zeolites. Olivine-dolerites with appreciable quantities of these minerals have been termed crinanites. In unpublished notes on analcime-bearing rocks from Lundy, Flett noted resemblances to rocks described in the Knapdale Memoir (Peach and others, 1911). With certain exceptions ((E9216); 9240). the rocks he described were non-porphyritic, but subsequent collection has yielded both aphyric and porphyritic varieties. Walker (1934) provided a useful summary of the use of the name crinanite, which had been defined by Fleti (1910) as a rock transitional between camptonite anc dolerite, rich in analcime. In the Knapdale Memoir, Fleti described such rocks as fine grained, with perfect ophitic texture, and seldom porphyritic. He added that corrodec feldspar phenocrysts, mainly of bytownite, occurred in a few specimens, and that the teschenites were less commonly porphyritic than the crinanites. It is proposed to cal] the relevant Lundy rocks crinanites and to include both aphyric and porphyritic types.

The crinanites are typically dark grey or greenish grey. fine-grained rocks, which weather to rusty brown. They car, be distinguished in hand specimen from the normal olivinedolerites by the presence of analcime-filled amygdales. A few specimens ((E4613); (E11319); (E11321)) have many amygdales 3 to 4 mm across and some up to 10 mm. In others the amygdales are smaller and scattered.

Plagioclase and olivine occur as phenocrysts. Generally one or the other is found in a coarse- or medium-grainec groundmass but in the finer rock-types the two commonly occur together. The plagioclase is in the labradorite range. concentrically or irregularly zoned (E42794), and crystals are from 1.5 to 6 mm long. Olivine phenocrysts occur as euhedra and subhedra, commonly about 1.5 mm across (E9240) but up to 3.0 x 4.5 mm; they are generally serpentinised, although in one specimen (E11314) the cores of the crystals remain fresh.

The groundmass consists typically of laths and zoned or unzoned euhedra of plagioclase, subhedral to anhedra] pyroxene, opaque iron oxide, serpentine pseudomorphs after olivine, and interstitial analcime. The plagioclase is. andesine-labradorite, much of it fresh but commonly altered patchily or along cleavage traces to chloritic material. The pyroxene is slightly pleochroic purplish brown titanaugite, rarely showing traces of hour-glass zoning. Opaque iron oxide, probably ilmenite, forms crystals (0.05 to 0.10 mm) commonly enclosed in titanaugite. Olivine is generally altered to serpentine minerals but rarely (E11315) the larger crystals have fresh cores.

Analcime occurs in irregular interstitial patches and in amygdales of 1 mm average diameter; it constitutes 1 to 7 per cent of the rock. Normally clear, it is locally pale brown and turbid with weak birefringence. Some analcime is accompanied by a fibrous radiating zeolite, probably natrolite (E11296). The latter also occurs alone and fills the centres of some chlorite-lined amygdales. Thomsonite also occurs in cavities as radiating crystals up to 10 mm long (E4615) and was identified by Mr P. H. A. Nancarrow (X7123).

Apatite is present in sheaves of fine needles, generally embedded in plagioclase ((E9206); (E11301); (E11325)), and specimens (E11296) and (E11301) contain small scattered biotite flakes, generally associated with opaque iron oxide.

The grain size of these rocks is governed by the sizes of plagioclase and pyroxene crystals. The porphyritic rocks have an average grain size of about 0.5 mm (range 0.1–0.7 mm). Typically the aphyric types are a little coarser, the plagioclase averaging 0.7 (0.4–1.8 mm) and the pyroxene 0.6 mm (0.1–1.2 mm). The porphyritic rocks are dominantly subophitic (E9240), with some specimens showing intergranular texture ((E9219); (E11274); (E11314)) or both textures ((E9216); (E42790); (E42797); (E42809)); the aphyric specimens have mainly intergranular texture ((E4610); (E9207); (E11275); (E11325)) or are intergranular and subophitic ((E9204); (E9206); (E11298); (E11315); (E42747)).

Olivine-free dolerites

A few dolerites appear to be free from both quartz and olivine. Rocks of this type are dark greenish grey or black, generally non-porphyritic, ranging in average grain size from 0.1 mm (E9228) to 0.5 mm (E11322). Xenocrysts of zoned plagioclase occur singly and in aggregates (E9241), and one specimen (E9218) has a few plagioclase phenocrysts.

In some specimens the groundmass plagioclase is oligoclase to andesine, in others labradorite with an outer zone of andesine. The pyroxene, occurring in laths 0.3 mm in length and in granules, is purplish brown titanaugite ((E9226); (E11312); (E11322); (E42772)). Euhedra and subhedra of opaque iron oxide, evenly scattered throughout these rocks, are plentiful in specimen (E42772). Chlorite is common in most specimens; it occurs interstitially ((E9241); (E11322)), filling vesicles (E9241) and replacing plagioclase and pyroxene (E9218). Accessories include interstitial analcime ((E9226); (E11312)), apatite and biotite (E9226). The rocks generally have an intergranular texture but one specimen (E42772) is subophitic.

Quartz-dolerites

The quartz-dolerites are fine grained, olive-grey and mostly non-porphyritic, with sparse xenocrysts of rounded quartz, corroded plagioclase and pyroxene. Specimen (E11269) contains xenoliths of fine-grained basaltic material. The groundmass plagioclase occurs as laths 0.15 to 0.20 mm in length ((E9205); (E9237)), or as tabular crystals (E11269). Pyroxene normally occurs in granules 0.1 mm across, but also forms larger plates and is generally altered to carbonate and chloritic material. Where fresh it appears to be augite (E9237). Quartz is present as scattered interstitial anhedral grains. Opaque iron oxide is a common accessory in subhedra 0.01 mm across. The texture of these rocks is generally even grained and intergranular.

Trachyte dykes

The trachytic dykes are medium to fine grained or glassy, and commonly carry small rectangular or acicular feldspar phenocrysts. They are pale to greenish grey in colour and weather to pale brown or buff.

Sir John Flett (unpublished notes) recognised variolitic and holocrystalline (orthophyric) types and spherulitic pitchstones. One of the orthophyric types, containing a sodium-rich amphibole, was isolated by Flett and given the name lundyite' by Hall (1915). Dollar (1942) briefly described the group, distinguishing pitchstones, potash trachytes and orthophyres.

Of 42 specimens representing about 20 dykes, most are quartz-trachyandesites (quartz-latites) or quartz-trachytes, containing various accessory mineral assemblages and showing a variety of textures. There are in addition a small number of glassy spherulitic rocks, a few cryptocrystalline 'felsites' which probably form a link between the main types and the glassy rocks, and several specimens exhibiting basic affinities. Some of the dykes show flow-banding at their margins.

Many of the intrusions are difficult to reach, and use has been made of material collected by R. Etheridge between 1857 and 1863, and by T. C. Hall, around 1915. Unfortunately these specimens are not all well localised, and localities and grid references are approximate (Appendix 2).

Quartz-trachyandesites and quartz-trachytes

Rocks of this type are composed essentially of oligoclase, potassium feldspar and minor quartz. It appears that they crystallised as mainly oligoclase and that subsequently they suffered potassium metasomatism which resulted in the partial conversion of the plagioclase to potassium feldspar and the consequent formation of a little quartz.

Modal analyses (Table 8) indicate a total feldspar content of more than 70 per cent. Four analyses of stained sections show ratios of potassium feldspar to oligoclase which indicate a range in rock-type composition from trachyte to trachyandesite. Quartz content ranges from 2.7 to 8.6 per cent, averaging about 4.2 per cent.

Most of the specimens contain a few phenocrysts but none can be termed porphyritic. The phenocryst content of one specimen (MR14873) was estimated at about 7 per cent. The phenocrysts are generally potassium-feldspars whose glassy appearance in hand specimen suggests sanidine; most are clear and fresh and some are twinned (E9210). They range in size from 1.5 to 6 mm. In two specimens ((E9232); (E42792)) the potassium-feldspars show extensive replacement by calcite. In a few cases corrosion suggests that some are xenocrysts.

The oligoclase of the groundmass occurs mainly as elongated laths, typically 0.1 to 0.25 mm long and 0.05 or 0.06 mm wide. In a few coarser-grained specimens ((E11290); (E11306); (E42792); (E42798)) average feldspar lengths range from 0.5 to 1.0 mm. Equant crystals rarely predominate (E42774) but are scattered in many specimens. In about half the specimens the groundmass feldspar is altered, in varying degrees, generally to chloritic material but in specimens (E42774) and (E42792) to calcite. The stained specimens show extensive replacement of plagioclase by a ramifying network of minute potassium-feldspar crystals ((E42743), (E42792), (E42798)). Hall (1915) drew attention to the high content of K₂O in a specimen (E9213) from the glassy selvedge of a dyke, which suggests that potassium was concentrated in the late-stage fluids of the magma and was therefore available for the conversion of plagioclase.

Much of the quartz is present as interstitial anhedra, but subhedral crystals occur locally ((E11267); (E11305); (E11306)). In about half the specimens the crystals range from 0.01 to 0.05 mm, in others from 0.1 to 0.2 mm and in a few they reach 0.25 mm.

Most samples contain about 1 per cent of opaque iron oxide; specimens (E4627) and (E9210) contain more and specimen (E42798) carries over 6 per cent. The oxide grains are usually from 0.01 to 0.05 mm across, but range from 0.004 mm (E9215) to 0.10 mm ((E4593); (E11290)).

About half the specimens contain scattered crystals of a blue and green sodic amphibole described by Sir John Flett (unpublished notes) and first reported by Hall (1915). In some specimens the mineral is moulded on the feldspars but in others it is associated with glass or interstitial quartz. Evidently the amphibole formed late in the crystallisation sequence. Crystals are variable in size but generally small. Good prismatic sections up to 0.08x 0.03 mm are present in one specimen (E9232), although most of the crystals are subhedral and considerably smaller. In specimen (E4593) a few prisms reach 0.14 mm in length. The best examples are in specimen (E9244), where the mineral is plentiful in crystals up to 0.12 mm long. The amphibole is markedly pleochroic in deep inky blue, bottle-green and brown, the colours commonly masking the polarisation tints. Extinction angles are wide; maximum values of 74° and 75° in specimen (E9244) indicate the arfvedsonite range. Mr R. J. Merriman examined a separation from this specimen by X-ray diffraction (X6792) and confirms that the mineral is a sodic amphibole. Mrs A. E. Tresham has provided a quantitative analysis using the electron microprobe (Cambridge Instruments Geoscan). Four crystals from specimen (E9244) were analysed and the means and standard deviations for eight elements, together with the unit cell contents, are given in (Table 9). Dr M. T. Styles has re-examined the analysis by Mrs Tresham and has estimated the Fe₂O₃ content by adjusting the Fe₂O₃:FeO ratio to give a charge balanced formula. The amphibole is arfvedsonite according to the classification compiled by Leake (1978). It contains too little calcium for a kataphorite (Hall, 1915).

In several specimens fine acicular needles cross the sections in all directions and may form part of radiate or variolitic structures. They are best seen in specimen (E9245), where they range in length from 0.1 to 1.3 mm and are about 0.02 mm wide. Generally occurring in specimens containing the sodic amphibole, the needles appear to be of early formation. The crystals are entirely altered to secondary minerals, principally calcite, but were possibly of hornblende.

Many thin sections show interstitial glass, in amounts varying from less than 1 per cent to as much as 20 per cent. The glass is normally pale green but may be brown and turbid or tinged with yellow. The pale green variety is presumably now cryptocrystalline chlorite material. Generally the glass forms small angular areas measuring about 0.08 by 0.10 mm, but it also occurs in irregular patches up to 0.4 mm across.

Pyroxene occurs in specimens (E9210) and (E42792), and minute flakes of brown mica and needles of apatite are present in accessory amounts. Small brown rectangular prisms (E9214) about 0.05 mm across and brown acicular crystals may be pseudomorphs after pyroxene or amphibole. The pyroxene shows alteration to green chloritic material and calcite (E42792), but where fresh appears to be augite.

The main types of intermediate rocks display inter-granular, orthophyric, variolitic and trachytic textures. About half of the specimens exhibit one texture alone; the remainder show two textures with one usually dominant. The variolitic group is the largest ((E9214); (E42798)), with some of the specimens tending towards being intergranular ((E9209); (E9215); (E9245)) or orthophyric ((E4587); (E42792)). Specimens with typical intergranular texture are (E4595), (E4627) and (E9232), and several specimens ((E4589); (E9210)) are intergranular and variolitic. Orthophyric texture is well exhibited by specimens (E4585), (E11267) and (E11306), and trachytic texture by specimens (E26222) and (E42743).

Glassy spherulitic trachytes

Three specimens ((E9211); (E9212); (E9213)) from one dyke and one specimen (E26221) from an associated dyke illustrate a glassy facies into which the main group of dykes may pass. They are streaky, greenish grey or bluish grey rocks, consisting mainly of brown, yellowish brown or buff glass with microlites and spherulites, together with a few phenocrysts of feldspar, pyroxene and opaque iron oxide.

In two specimens ((E9213); (E26221)) the buff glass is crowded with minute spherulites less than 0.01 mm in diameter, probably of feldspar. Yellowish brown elongate patches have a central core of quartz or opaline silica and are surrounded by radiating crystalline areas of chloritic material. Pale green or pale brown augite laths, about 0.05 mm in length and having wispy terminations, are scattered throughout the glassy groundmass, locally forming radiating aggregates. Specimen (E9212) has unfilled and chlorite-filled cavities giving the rock a spongy texture. Specimen (E9211) shows microlites of feldspar and honey-coloured pyroxene arranged in radiating aggregates. The yellowish brown glass is full of curved trichites (hair-like crystallites) and in some areas, particularly around the feldspar phenocrysts, axiolites (crystalline fibres radiating from an axis) are common.

Feldspar phenocrysts are euhedra or subhedra of sanidine up to 3 mm long and commonly rounded or deeply corroded. Some of the crystals appear to have been broken. Phenocrysts normally occur singly but may be in small clusters (E9213). Many are twinned, with cores of ex-solved albite perthite which are commonly altered. Mr R. J. Merriman reports that X-ray examination (X6855) of a phenocryst from specimen (E9212) shows the whole crystal to be composed of low-albite.

Pyroxene phenocrysts are green to yellowish green pleochroic aegirine or aegirine-augite. They occur as euhedra and subhedra about 0.2 mm across, are commonly rounded or corroded and are normally associated with opaque iron-oxide crystals.

Felsitic trachytes

The fine-grained felsitic trachytes seem to form a textural group transitional between the trachytes and the glassy rocks. However they may have formed from a more acidic fraction of the intermediate magma. Several have been silicified, which makes it difficult to determine their original mineralogy. They are pale buff or grey rocks in hand specimen and all carry a few phenocrysts of potassium feldspar, probably sanidine, which is commonly altered and corroded (E42777).

The groundmass is composed of feldspar, quartz and a little opaque iron oxide. In specimen (E11307) the feldspar appears to be oligoclase, but a stained section (E42777) reveals partial replacement of the plagioclase by potassium feldspar. Quartz is anhedral and interstitial, and in two specimens ((E4591); (E11307)) fills cavities. Pale green acicular crystals which may be amphibole are scattered throughout specimens (E4591) and (E4594). Minute flakes of biotite are present in specimen (E4590), and section (E11307) contains a few crystals of the arfvedsonitic amphibole described above (p. 78). Several specimens ((E4590); (E4591); (E4630); (E42777)) show original glass replaced by chloritic material.

The groundmass grain size is generally 0.04 to 0.08 mm, but grains of opaque iron oxide and other accessories are normally about 0.006 mm across. The texture of these rocks, especially the silicified specimens, is felsitic, but specimen (E4590) has a mainly intergranular texture with a few incipient varioles and sheaves of feldspars.

Trachytes with basic affinities

A few specimens, although essentially intermediate igneous rocks, exhibit some basic igneous features not shown by most of the trachytes, and may therefore provide a link with the basic group of dyke rocks. They are either non-porphyritic ((E4612); (E11277); (E11328)) or very poorly porphyritic ((E11291); (E11311); (E11329)). Phenocrysts are potassium feldspar or plagioclase up to 2.5 mm in length. One specimen ((E11311)) contains several small phenocrysts of fresh augite.

The groundmass feldspar is oligoclase-andesine in laths and tabular crystals 0.2 to 0.5 mm long. Pyroxene, mainly titanaugite, is common in most specimens, forming pale brown elongate laths and colourless granules ((E4612); (E11311)) or plates up to 0.8 mm across (E4612). Brown titanaugite, rarely with a green core, occurs in specimen (E11291), and pseudomorphs of calcite and chlorite, probably after pyroxene, are present in other specimens. Serpentinised olivine is common in one specimen (E4612), as euhedra 0.3 mm across accompanied by smaller anhedra, and is rare in another (E11277). Opaque iron oxide is common in some specimens ((E4612); (E11277); (E11291); (E11311)), generally as small euhedra. Apatite needles and rare brown and green flakes of mica are present as accessory minerals. Arfvedsonitic amphibole is another accessory ((E11277); (E11291); (E11329)), closely associated with and possibly replacing pyroxene (E11291), or associated with interstitial chlorite. Calcite fills amygdales and in some specimens is an alteration product of pyroxene. Chlorite is an alteration product of feldspar and pyroxene and may replace glass.

The textures of these rocks are variable. In grain size they range from relatively coarse (E4612) to fine (E42808). Several specimens ((E11329); (E42808)) have strongly flow-aligned feldspar laths. Intergranular texture is exhibited by specimen (E4612), and specimen (E11311) is orthophyric.

Chemistry of dykes

Five analyses of basic dykes and four of trachytic dykes are given in (Table 10) and (Table 11) respectively. Three of the analyses were made for Hall (1915), and the remainder are of specimens collected during the recent survey. Additionally, minor constituents have been determined on Hall's specimens and CO₂ values re-determined; for one sample (E9206) sulphur has been re-determined and reported as FeS₂, an adjustment being made to the value for Fe₂O₃.

Dolerites

The analyses show clearly that the basic rocks have been derived from an alkali olivine-basalt magma. In four cases the SiO₂ contents lie towards and at the lower end of the range of 52 British olivine-basalts, olivine-dolerites and crinanites quoted by Guppy and Thomas (1931) and Guppy and Sabine (1956), but are well outside the range of 31 analyses of quartz-gabbros, quartz-dolerites and tholeiites given by the same authors. The higher percentage of SiO₂ in the fifth specimen reflects its xenolithic character. Manson (1967) gave average SiO₂ percentages as 47.2 for alkalic basalts and dolerites (745 analyses) and 51.2 for tholeiitic basalts and dolerites (1228 analyses). The percentages for the Lundy rocks clearly fall within the alkaliolivine-basalt field. The SiO₂ contents of the two Lundy crinanites are comparable with those of other British Tertiary crinanites.

Basic rocks of Lundy generally show Al₂O₃ values close to the average value for all basic rocks used by Manson (1967), and even the 18.11 per cent Al₂O₃ of sample (E42786) is within his range. Average values for Fe₂O₃ and FeO in alkalic and tholeiitic rocks are very similar, and compared with the general data the Lundy dykes are quite high in the range of Fe₂O₃ values and very variable with respect to FeO.

Percentages of MgO are rather below the average for tholeiitic rocks and appreciably below the figure for alkalic rocks. However, available data indicate that MgO values for British Tertiary basic rocks are a little lower than the world-wide averages of Manson. CaO values for tholeiitic and alkalic rocks are very close, and with one exception the Lundy rocks are rather poor in this constituent.

Values for Na₂O are well above Manson's average for tholeiitic rocks but closely comparable with his figure for alkalic rocks. The K₂O contents of the Lundy basic rocks ((E42788) excepted) are well below the alkalic average of 1.4 per cent and also below the average for tholeiitic rocks (0.7 per cent). Nevertheless they do follow a general tendency for Tertiary olivine-dolerites to contain less K₂O (0.8 per cent calculated from 20 analyses) than do Tertiary quartz-dolerites and tholeiites (1.2 per cent based on 15 analyses).

The values of P₂O₅ lie between Manson's averages for the two groups of basic rocks. Percentages of MnO are considerably higher than Manson's averages, although generally within the ranges of values given. Lundy basic rocks have TiO₂ contents close to Manson's average for alkalic rocks and, except for one value, distinct from his average for tholeiitic rocks.

Eleven of the trace elements determined for the Lundy basic rocks were compared with median and arithmetic mean data compiled by Prinz (1967) in a study of 257 analyses of basaltic rocks from world-wide locations. Values for barium are low compared with basic rocks as a whole and particularly low for alkaline basic rocks except in the case of sample (E42788). This single high value correlates with the relatively high figure for K₂O in this rock. However, as Prinz noted, basaltic provinces are known to have a very wide variation in barium content.

Median and mean figures for cobalt show little variation throughout the analyses used by Prinz, and the values for the Lundy basic dykes depart very little from these. However, chromium values for the dykes are all below mean figures except for the amount in sample (E42786). The values for samples (E42747) and (E42772) are particularly low but, as chromium usually shows the greatest dispersion of all the spectrographically determined elements, the significance of an average value is reduced. Copper too is variable in these dykes. The values for samples (E42772) and (E42788) are just below the median and mean values given by Prinz, and those for samples (E42747) and (E42786) are considerably lower. The value of 200 mg/kg for sample (E9206) is almost twice the mean for alkaline basic rocks but well within the range quoted for all basic rocks. This high copper value may reflect the presence of sulphide.

Gallium values from Lundy lie within the limits given in Prinz's study, although scattered in relation to his mean (17–21 mg/kg) and median (19–20 mg/kg) values for all basaltic rock types.

Only one of the lithium values compares closely with the mean (12 mg/kg) and median (10 mg/kg) values given by Prinz for alkaline basic rocks. The other values are all relatively high and one, of 40 mg/kg, for sample (E42786), is at the extreme range of Prinz's analyses. Lithium can substitute for sodium to a limited extent in plagioclase, but that element is present in the Lundy rocks in only average amounts. There is no apparent reason for these high values unless the lithium is present in abnormally high quantities in olivine and pyroxene.

All the values for nickel are low when compared with Prinz's mean of 101 mg/kg, although the lowest values are within the range he quoted. Various studies point to Nienrichment in early olivine and to some concentration in magnetite and pyroxenes. In the Lundy rocks it is therefore most probably concentrated in olivine and its alteration products.

Rubidium values are generally well below the mean (41 mg/kg) and median (30 mg/kg) values quoted, and the high figure for specimen (E42788) reflects the relatively high content of K₂O in this rock. Strontium also is well below the median (700 mg/kg) and mean (774 mg/kg) values; this correlates with the low barium figures for the Lundy basic rocks.

The mean and median values for vanadium given by Prinz are fairly close for both tholeiites and alkali basalts, and the figures for the Lundy basic dykes could fit within either range. Vanadium is most abundant in magnetite, less common in pyroxene and scarce in olivine. It is perhaps significant that the greatest amount is present in the olivine-free dolerite.

The Lundy basic rocks are anomalously high in zirconium in comparison with mean and median values for 70 alkali-basalts quoted by Prinz. Specimens (E9206), (E42747) and (E42772) all have values near the high end of the range, and specimen (E42788) contains twice that amount. Prinz's study indicates that alkali-basalts contain more zirconium than do tholeiites. Caho and Fleischer (1960) found evidence for regional variation of zirconium in basaltic rocks, which may be relevant in the case of Lundy. The element is chiefly concentrated in pyroxenes.

Values of boron and fluorine in the Lundy rocks were compared with data quoted by Goldschmidt (1954). He gave values of 1 to 2 mg/kg of boron in basalts from the USA and Germany, and an average of 3 mg/kg in gabbros from Germany. Values quoted for fluorine vary from 0.01 per cent (100 mg/kg) for basalts and andesites (Shepherd, 1940) to 0.031 per cent (310 mg/kg) for German gabbros (Koritnig, 1951), while a value of 0.08 per cent (800 mg/kg) (Fersman, 1934) is accepted for the average fluorine content of the upper lithosphere. The results for the Lundy basic dykes are comparable with these figures for both elements.

It is worth noting that specimen (E42788) is a xenolith-bearing type, and that the relatively high values for barium, rubidium, zirconium and fluorine may reflect this fact.

Trachytes

The four analyses of the trachytic dykes were compared with analyses of 15 Palaeozoic and four Tertiary trachytes, orthophyres and trachyandesites quoted by Guppy and Thomas (1931) and Guppy and Sabine (1956), and with the averages for 24 trachytes and 42 lathes. given by Nockolds (1954). Silica percentages for the Lundy rocks fall within the range of the Tertiary specimens and towards the top of the range shown by Palaeozoic analyses, but they are 5 to 10 per cent higher than the average values given by Nockolds. Alumina content on the other hand is below the values for the Tertiary rocks, well below the Nockolds averages and low in the range of Palaeozoic analyses. Values for both Fe₂O₃ and FeO are similar to the range for Tertiary rocks but spread on either side of Nockolds's averages and fall well within the range of values for Palaeozoic rocks. Percentages of MgO and CaO are well below Nockolds's averages, and MgO is low when compared with the other data. However, contents of CaO generally fall within the ranges shown by other trachytic rocks.

Both Na₂O and K₂O percentages are comparable with the Palaeozoic and Tertiary data, apart from one high K₂O figure for a glassy trachyte (E9213). However, the latter value is close to the Nockolds average for trachytes, while the values for the other three Lundy rocks compare more closely with his value for latites. Values of TiO₂ are lower than Nockolds's averages but fall within the ranges of the other two sets of data.

Of the trace-elements, barium, zirconium and caesium are particularly high and boron and strontium are fairly high in the glassy trachyte. Sulphur is relatively high in specimens (E9213) and (E42774), and barium fairly high in the latter. Zirconium is particularly high in specimens (E9232) and (E42792) and rubidium is high in all four trachyte samples. JD

Details

Rattles Anchorage to Little Shutter Rock

A dolerite dyke trending NNE from the beach at Rattles Anchorage [SS 1380 4365] runs up the cliff to meet the junction plane between slates and granite; this plane dips steeply NNW in the cliffs to the west. Brecciation of sediment and granite is evident, suggesting that thr dyke has invaded a fault plane. A cave or adit penetrates the cliff at the fault line, and traces of sulphides indicate iron and copper mineralisation. Immediately to the west, two dolerite dykes in the cliff [SS 1379 4367] and [SS 1377 4364] cut the slate-granite contact which, over this short distance west of the boundary dyke/fault at Rattles Anchorage, may be a normal contact. The generally medium-grained granite appears to be slightly finer grained near the contact, and the presence of some small xenoliths and a slight colour change point to contamination by country rock.

Between the beach at Rattles Anchorage and Little Shutter Rock [SS 133 433] medium-grained muscovite-biotite-granite with small feldspar megacrysts has been invaded by dolerite dykes, which trend between NW and WNW and may follow joints in the granite. In two localities the dykes appear to bifurcate [SS 1366 4355]; [SS 1362 4348] and three of them may be traced across the southern tip of the island [SS 1344 4339]–[SS 1333 4343]; [SS 1342 4337]–[SS 1332 4341]; [SS 1339 4334]–[SS 1332 4339]. The Devil's Limekiln [SS 1338 4347] is a 90 m-deep blowhole in this locality. It has been formed by erosion along two dolerite dykes about 1.5 m thick; one trends 310° from the south coast [SS 1345 4343] and the other 060° from the west coast [SS 1331 4344]. Dollar (1942) noted that at this south-western tip of the island pneumatolytic agencies had reddened feldspars, and that schorlite and fluorite occurred on joint planes and within the granite.

Little Shutter Rock to Pilots' Quay

Medium-grained granite immediately north of the Devil's Limekiln displays joints dipping 30°/335° and 30°/340°. Two dolerite dykes on the coast here, each less than 1 m thick, may be traced north-westwards for 220 m [1329 4346–1316 4362]; [1329 4349–1318 4363]. Another trends NE from beach level [SS 1321 4365] and splits into 5 apophyses within medium-grained muscovite-biotite-granite containing megacrysts of alkali feldspar up to 20 mm long. The granite cliffs between Montagu Steps [SS 1317 4361] and Goat Island [SS 1315 4375] contain a number of dark xenoliths, presumably of country slates.

Northwards from Goat Island to Pilots' Quay [SS 1300 4395] several dolerite dykes crop out [SS 1315 4383]; [SS 1314 4385]; [SS 1319 4388]; [SS 1310 4387]; [SS 1305 4393], around 0.6 to 0.9 m in thickness. The granite hereabouts is characterised by joints which trend between NW and NNE and dip westerly at 60° or more. This inclination has resulted in much seaward slipping of the granite. Pilots' Quay is one of the localities which Dollar (1942) instanced as illustrating the sharp intrusive nature of the contact between his two main granite types; it is possible to scramble from the 'quay' to the summit plateau of the island, and there is no mappable line to be seen.

Pilots' Quay to Battery Point

To the north of Pilots' Quay, joint-induced instability of the cliffs remains in evidence. A quarry [SS 1305 4427] at the cliff top just west of The Old Lighthouse shows medium- to coarse-grained biotitemuscovite-granite with megacrysts of alkali feldspar and some reddish brown garnet. Pods of fine-grained granitic rock occur totally enclosed by coarser granite. Some are small, others up to

0.6 m thick and 1.5 m long, and their form suggests that they originated as inclusions. Joints in the granite hereabouts are steep or vertical, trending NW, NE and E [SS 1305 4436] and N and E [SS 1301 4443]. A dolerite dyke 0.6 m wide crops out in the lower cliffs [SS 1296 4412] 150 m N of Pilots' Quay and similar thin dykes are common farther north [SS 1296 4423]; [SS 1294 4427]; [SS 1293 4430]; [SS 1293 4432]; [1292 4444 bifurcating; 1291 4445]; [SS 1288 4448]; [SS 1287 4456]; [SS 1287 4463]; [SS 1288 4465]. The cliffs above the last-, mentioned dyke show [SS 1303 4464] medium-grained biotitemuscovite-granite in which the megacrysts of alkali feldspar range up to 25 mm in length. Megacrysts up to 30 mm long occur above Battery Point in similar granite showing near-vertical joints trending 190° and 285°–290° [SS 1296 4478]; [SS 1292 4480] and N and E [SS 1288 4487]. The lower cliffs around Battery Point [SS 127 449] contain several doleritic dykes. To the south two thin ones trend E [SS 1281 4476]; [SS 1280 4477]; a third of similar trend [SS 1277 4482] is 1.2 m wide and crops out immediately north of a much split dyke rarely exceeding 0.3 m in width. Dolerite dykes exposed at the ruins of the old battery [SS 1276 4488] are 1.2 to 1.4 m wide, trend just E of N and slightly N of E, and show marked spheroidal weathering.

Battery Point to St James's Stone

Between Battery Point and Dead Cow Point [SS 127 452] split dolerite dykes converge at three localities [SS 1280 4496]; [SS 1285 4503]; [SS 1285 4510]. Granite near the last of these is reported to have been investigated during the Second World War as a possible source of molybdenum, which occurs as molybdenite within a pegmatite. A thin trachytic dyke crops out nearby [SS 1286 4508].

The granite cliffs above Dead Cow Point, in the area known as Earthquake, are deeply cleft by steep joints trending slightly E of N, roughly parallel to the coastline. Northerly trending dolerite dykes [SS 1277 4516]; [SS 1277 4525]; [SS 1284 4530]; [SS 1286 4531] which cut across the point probably invaded similar joint planes. Other thin dolerite dykes trend seawards in the roughly radial pattern characteristic of Lundy [SS 1290 4535]; [SS 1292 4537]; [SS 1293 4540]; [SS 1294 4554], as do two trachytic dykes [SS 1281 4528]; [SS 1292 4539]. Medium- to coarse-grained biotite-granite and biotite-muscovitegranite at the northern end of Earthquake [SS 1305 4545] is cut by vertical N–S joints which have dictated massive collapse; huge blocks of granite have slipped and tilted seawards leaving deep N–S clefts up to 30 m deep.

Trachytic [SS 1305 4561]; [SS 1306 4563] and doleritic [SS 1318 4564] dykes on the southern side of Jenny's Cove are aligned with this northerly joint direction. Thin dykes of doleritic [SS 1308 4564]; [SS 1319 4564]; [SS 1327 4572]; [SS 1325 4583]; [SS 1328 4591] and trachytic [SS 1326 4579]; [SS 1325 4584] type crop out in the cove. The granite in cliffs above, medium grained with feldspar megacrysts up to 30 mm long, is cut into beautifully tiered pinnacles by vertical (trend N and E) and horizontal joints [SS 1333 4582]; [SS 1332 4586] and contains pod-like microgranitic inclusions [SS 1334 4591]. Joint planes lower in the cliffs [1323 4587–1318 4594] dip 65°–80° W and are locally curved.

Northwards to St James's Stone numerous dykes occur, commonly 0.2 to 0.6 m thick; they are mainly of dolerite [SS 1314 4600]; [SS 1314 4604]; [SS 1315 4608]; [SS 1313 4611]; [SS 1311 4618]; [SS 1306 4626]; [SS 1316 4633]; [SS 1318 4633]; [SS 1321 4632]; [SS 1321 4636]; [SS 1322 4643]; [SS 1315 4661]; [SS 1311 4673] but three are trachytic [SS 1317 4631]; [SS 1321 4638]; [SS 1311 4668]. The granite of this stretch of coast is of medium-grained texture. A microgranite (?granitised inclusion) shows megacrystic feldspar extending into it from the margin, suggesting that the feldspar has grown secondarily in situ. Smooth granite outcrops [SS 1325 4664]–[SS 1327 4674] above the Devil's Slide are traversed by veins and stringers of microgranite 70 to 100 mm wide picked out by differential weathering.

St James's Stone to North West Point

Immediately west of the old field enclosure in which stands the Widow's Tenement coarse-grained biotite-muscovite-granite [SS 1320 4681]; [SS 1331 4687] contains alkali feldspar megacrysts up to 40 mm long. To the north similar rock contains veins and pods of microgranite [SS 1328 4703]–[SS 1321 4711]. Granite outcrops are locally smoothly rounded [SS 1313 4715]. They display joints trending just W of N, some vertical, others dipping steeply westwards (Plate 13) and a few inclined steeply to the E, and microgranite veins standing out resistant to weathering [SS 1313 4715]–[SS 1301 4725]. Farther north the granite is well jointed; one set dips about 60° W, another is vertical, trends E–W and has given rise to deep clefts, and a third is horizontal or very gently inclined eastwards [SS 1289 4725]–[SS 1291 4737].

Extensive areas of bare medium-grained megacrystic granite occur on the north-western part of the island plateau, where the thin peat cover has been burned away. In descending the steps towards Lundy North Lighthouse it is possible to discern a change within the granite towards a finer matrix and the presence of larger crystals of quartz. Megacrysts of feldspar remain characteristic, commonly up to 20 mm long but locally twice as large. This corresponds to the passage from Dollar's (1942) G1 granite to his G2. It is possible to assign an approximate position to the change say [SS 1316 4802] but the choice is arbitrary.

The following thin dykes crop out at high water mark along this stretch of coast: doleritic [SS 1318 4682]; [SS 1320 4684]; [SS 1316 4702]; [SS 1311 4706]; [SS 1306 4706]; [SS 1307 4710]; [SS 1307 4711]; [SS 1305 4715]; [SS 1302 4717]; [SS 1292 4724]; [SS 1291 4725]; [SS 1290 4727]; [SS 1293 4736]; [SS 1294 4738]; [SS 1290 4747]; [SS 1290 4750]; [SS 1293 4756]; [SS 1300 4763]; [SS 1301 4764]; [SS 1302 4768]; [SS 1306 4770]; [SS 1308 4779]; [SS 1311 4787]; [SS 1311 4793]; [SS 1310 4796]; [SS 1309 4797]; [SS 1307 4797]; [SS 1306 4799]; trachytic [SS 1316 4700]; [SS 1287 4723]; [SS 1288 4727]; [SS 1290 4738]; [SS 1308 4778].

North West Point to Gannets' Bay

The cliffs of North West Point are of biotite-granite and biotitemuscovite-granite with a fairly fine-grained matrix surrounding quartz crystals up to 10 mm across and alkali feldspar megacrysts typically up to 20 mm long but rarely up to 80 mm. Two dykes trend NNE or NE across the point from the west coast to the north coast, one trachytic [SS 1304 4805]–[SS 1311 4813] and the other 0.75 m thick and of dolerite [SS 1302 4815]–[SS 1306 4821]. Alongside the north-coast outcrop of the former are traces of a fine-grained dolerite apparently trending NW, but the relationships of the dykes one to another are not clear. At the landing quay [SS 1319 4810] four dolerite dykes trend around NE and a fifth appears to run E–W. None is more than about 0.6 m thick, some sections converge and others are joined by stringers. Other basic dykes crop out on the north coast in Kittiwake Gully [SS 1321 4809] and in and near Puffin Gully [SS 1339 4806]; [SS 1344 4806] and a trachytic dyke strikes NE [SS 1330 4811].

Trachytic intrusions [SS 1351 4797] occur immediately south of North East Point and between there and Gannets' Rock [SS 1370 4757] a number of thin dolerite dykes cut the cliffs [SS 1351 4781]; [SS 1360 4777]; [SS 1359 4776]; [SS 1358 4774]; [SS 1359 4769]; [SS 1361 4768]; [SS 1362 4765]; [SS 1361 4763]; [SS 1364 4759]. Gannets' Rock itself is bounded to the north by an E–W dolerite dyke, and its separation from the main island at high water has been facilitated by erosion along a similar intrusion.

The granite of the cliffs is typically a medium- to coarse-grained megacrystic biotite-muscovite-granite, but local textural variations occur. Thus in Gannets' Bay [SS 136 474] may be noted biotiterich areas, pegmatitic patches and some fairly sharply defined fine-grained pods which may be granitised inclusions.

One dolerite dyke [SS 1354 4744] in Gannets' Bay trends WSW up South Combe to a rubbly exposure of fine-grained basaltic material [SS 1324 4737] on the edge of the central plateau and thence probably due W down a further gully to the west coast. Several other dolerites crop out in the bay [SS 1361 4752]; [SS 1356 4749]; [SS 1354 4737]; [SS 1355 4734]; [SS 1357 4732]; [SS 1359 4731]; [SS 1361 4728].

Gannets' Bay to Halfway Wall Bay

Medium- and coarse-grained megacrystic biotite-muscovitegranite of the cliffs may be examined at Brazen Ward [SS 1393 4685]. It is cut by near-vertical joints running about NNW and ESE and by horizontal joints the weathering of which produces a misleading impression of erosion platforms (Figure 12). The presence of a sea cave, Queen Mob's Grotto (p. 111), however, is clear evidence of a higher sea level. Some exposures inland of the cliffs show finer grained granite [SS 1385 4676] but at the eastern end of Three-quarter Wall [SS 1368 4659] medium-grained granite contains larger quartz crystals and feldspar megacrysts up to 40 mm long. Granite at the coast to the east of the latter locality shows the flat-lying joints to be inclined about 10°/160° [SS 1386 4661].

Eight thin dolerite dykes crop out at the coast near Brazen Ward [SS 1373 4713]; [SS 1376 4708]; [SS 1379 4705]; [SS 1382 4700]; [SS 1390 4674]; [SS 1388 4671]; [SS 1384 4662]; [SS 1386 4659].

Granite at the top of the coastal bevel near Tibbett's Lookout [SS 1382 4629] varies from fine or medium grained to coarse grained, some of the changes being fairly sharply defined. Farther south microgranite pods [SS 1384 4619] in medium-grained granite show local development of megacrystic feldspar and quartz at their margins [SS 1383 4615]. Nearby massive outcrops show medium-grained biotite-muscovite-granite with a fairly fine-grained matrix surrounding megacrysts of feldspar and quartz [SS 1388 4616] and similar granite with joint planes dipping 45° NNW and 60° SSE [SS 1391 4608]. Medium-grained biotite-muscovite-granite [SS 1380 4590] at the eastern end of Halfway Wall has larger and more numerous feldspar megacrysts than has the rock immediately to the north. It is cut by vertical joints trending ESE and SSW, and by very well developed horizontal joints which have facilitated the development of a logan stone.

Between Tibbett's Point [SS 1410 4637] and Halfway Wall four thin trachytic dykes trend between NW and WNW [SS 1402 4626]; [SS 1400 4623]; [SS 1392 4594]; [SS 1390 4586] and dolerite dykes between NW and WSW [SS 1400 4619]; [SS 1399 4617]; [SS 1402 4605]; [SS 1400 4603]; [SS 1399 4601]; [SS 1394 4596].

Halfway Wall Bay to Ladies Beach

At 90 m S of the eastern end of Halfway Wall fine- and medium-grained biotite-muscovite-granite is seen to be rather poorly megacrystic [SS 1379 4578]. Similar rock at sea level has been invaded by two basic dykes [SS 1390 4575]; [SS 1400 4547].

A track which once carried the tramway of the quarry company is cut into the coastal bevel between the southern end of Halfway Wall Bay and the eastern end of Quarter Wall, and adjoining it are old granite quarries. The northernmost and largest [SS 138 456] is in somewhat poorly megacrystic medium-grained biotitemuscovite-granite, with one set of joints vertical and trending ESE and another inclined 60°–80° W. Pegmatitic developments at the margin of a microgranite vein, and in the main granite, consisting of quartz and feldspar with beryl, apatite and topaz, were noted by Dollar (1942). Farther south the granite of V. C. Quarry [SS 1385 4536] shows vertical joints trending NNE and ESE and others inclined 45° to slightly S of E, and has been intruded by a basic dyke 0.3 m wide. Granite nearby [SS 1386 4528] contains traces of small brown crystals, possibly garnets. Medium-grained biotitemuscovite-granite at Quarry Pond [SS 1375 4504] contains feldspar megacrysts up to 25 to 30 mm long, and medium- and coarse-grained biotite-muscovite-granite near the ruins of Quarterwall Cottages shows similar megacrysts 20 to 25 mm long [SS 1389 4495].

The granite of Quarry Beach [SS 140 451] is medium to coarse grained with feldspar megacrysts rarely more than 25 mm long. Joints dip 12°/030° and 60°/200° and a third set is vertical, trend 130°, or dips steeply SW. Veins up to 0.3 m wide of quartz rock with a little mica and feldspar are common, and some pods of microgranite show penetration at their margins by feldspar megacrysts. A fine-grained pale grey somewhat glassy trachytic dyke [SS 1395 4519], 0.75 m wide and trending 110°, exhibits very slight marginal chilling but has had little effect on the surrounding granite. A thicker (1.6 m) similar orthophyric dyke [SS 1393 4512] trends around 120° but follows a zigzag course. It is grey, weathering to brown with a ferruginous crust, carries laths of feldspar, and shows 50-mm-wide chilled margins. Three doleritic dykes at Quarry Beach follow roughly parallel trends. The northernmost [SS 1394 4518] is 0.76 m wide, splits into two at the foot of the cliff and shows spheroidal weathering. Another [SS 1394 4515], up to 1 m wide and also spheroidally weathered, contains vesicles locally carrying quartz. The third is exposed only as basaltic rubble [SS 1395 4510]. A fourth basic dyke crops out [SS 1400 4497] below the ruins of Quarterwall Cottages.

Between Quarter Wall and Ladies Beach rather poorly megacrystic medium-grained biotite-muscovite-granite is cut by micro-granite in N–S veins [SS 1385 4473], or thin ramifying discontinuous veins [SS 1386 4453]; [SS 1392 4442]. Feldspar megacrysts commonly do not exceed 20 mm in length. Medium- to coarse-grained biotite-muscovite-granite near the contact with slates contains megacrysts up to 25 to 30 mm long [SS 1390 4426]; [SS 1389 4421]; [SS 1396 4415].

The Lametry peninsula

This is the south-eastern tip of Lundy; it is formed of slates, which have been invaded by numerous dykes. A weathered brown friable medium-grained trachytic dyke with blebs of quartz and with vesicles in the interior [SS 1415 4399] trends just N of E below Mill Combe, and a basaltic dyke appears to cut a trachytic dyke nearby [SS 1422 4396]. The Landing Beach is traversed by a buff ferruginous dyke of orthophyric type with some quartz trending 150°. It appears to truncate a vertical dolerite dyke [SS 1424 4385], which trends 082° and is typically 1.2 m wide but locally split into two, and meets a medium-grained buff trachytic or orthophyric dyke which is well exposed [SS 1428 4379] at the foot of the steps leading up to Lundy South Lighthouse. This last dyke is up to 2.3 m wide and carries a little quartz; it trends NE or NNE and is in part adjoined on its southern side by a fault. A dolerite dyke exposed in the foreshore [SS 1432 4381] splits east-south-eastwards to pass on either side of the jetty; the northern section passes through Hell's Gates between the main island and Rat Island and may have facilitated the erosion which created this channel, and the southern section is vesicular and contains quartz. Two dolerite dykes [SS 1476 4372] up to 1.2 m wide and trending NNE lie to the east of Rat Island.

Five parallel dykes run WNW on the south side of Lundy South Lighthouse. The two outermost are trachytic and visible in the foreshore and lower cliff [SS 1428 4357]; [SS 1430 4361]. The others are doleritic [SS 1452 4362]; [SS 1433 4364]; [SS 1432 4367] and are cut by the trachytic dyke seen at the foot of the lighthouse steps (above) where it continues south-south-westwards to cross the western side of Lametry Beach [SS 1423 4370]; this relationship is well displayed by the view north-west to Marisco Castle from the headland immediately south of the lighthouse (Plate 14). Several dolerite dykes between 0.3 and 1 m in width trend between NW and NNE in the cliffs south of Castle Hill [SS 1417 4367]; [SS 1415 4365]; [SS 1414 4364]; [SS 1414 4361]; [SS 1413 4358]; [SS 1404 4355]; [SS 1400 4356]; [SS 1388 4366]; [SS 1386 4365]; [SS 1383 4365]. EAE

Chapter 7 Structure

General account

Folds

The Devonian rocks of the district lie on the southern limb of a large anticlinal structure whose axis trends approximately west-north-west through Lynton, and on the northern limb of the major synclinorium which contains most of the Palaeozoic rocks of Devon. The regional structure is simple (Figure 13). Progressively younger formations dip southwards and succeed each other in that direction without large-scale repetition. Fold attitudes change from northward-overturned on the north coast to upright in the Pilton Shales of the Taw–Torridge estuary. Fold axes trend between E–W and ENE–WSW and are sub-horizontal.

Simpson (1969, 1970, 1971) defined this area as his structural Zone I. He noted that the progressive northward overturning was best illustrated by the change in attitude of the slaty (axial planar) cleavage within the more argillaceous formations. These, the Morte Slates, Upcott Slates and Pilton Shales, exhibit intensive small-scale folding. In striking contrast is the competence of the Pickwell Down Sandstones and the Baggy Sandstones, which are but slightly folded except in a regional sense. This contrast led Holwill and others (1969) to suggest that regional folding had produced differential movement between arenaceous formations, and that the Hangman Grits and Pickwell Down Sandstones had acted as a couple on the Ilfracombe Beds and Morte Slates between, the relative movement being an overriding northwards by the Pickwell Down Sandstones. The concept, if valid, would be equally applicable to the Pickwell Down Sandstones–Upcott Slates–Baggy Sandstones 'sandwich' farther south. However, there is no evidence of major strike thrusting. The relevant formational boundaries within the present district are commonly gradational, even though the sedimentological change may occur within only a few metres of strata, and it seems most likely that fold movements were accommodated mainly by crumpling within the argillaceous strata without any great inter-formational differential movement. The widespread presence of slickensides points to much minor slipping on bedding planes.

Shearman (1962) considered that folds in the Ilfracombe Beds and Morte Slates were of three orders of magnitude, and Holwill (in Holwill and others, 1969) defined their sizes in numerical terms more precise than the evidence justifies. However, it is broadly true to say of the Devonian of north Devon that geological mapping has demonstrated large-scale folds which are rarely apparent in exposure. This is best seen between Fremington and Swimbridge, east of the present district, where the amplitudes<span data-type="footnote">Amplitude, as used here, is the height of a fold from crest to trough; wavelength is measured from one crest to the next.</span> of the larger folds are generally measured in hundreds of metres. The foreshore and cliffs, and a few inland sections, display folds typically of a few tens of metres amplitude, and upon the limbs of these folds are developed small parasitic flexures commonly less than a metre in height.

Holwill (1963) mapped folds in the Ilfracombe Beds, between Ilfracombe and Combe Martin, with axial planes inclined around 50° S. Similar and tight overfolding is visible at Ilfracombe, and Simpson (1971) inferred axial plane dips of about 30° S at Lynmouth. Within the Morte Slates it is commonly difficult to distinguish bedding, but slaty cleavage suggests southward steepening of axial planes until in the Morte Point area they dip steeply south or approach the vertical. Kink bands are common. Tight to close folding towards Ilfracombe gives way to close to open folding at Morte Point. A contoured stereogram of poles to bedding planes of the whole of the Morte Slates outcrop within the district (Figure 14) shows a maximum of 19 per cent, which corresponds to fold limbs dipping 84°/198°, and a lesser peak of 9 per cent corresponding to fold limbs inclined 50°/202°.

The Pickwell Down Sandstones dip in a generally southerly direction. Angles are commonly steep towards the base and top of the formation, near the passages into the Morte Slates and the Upcott Slates, and probably some close folds are present. But within the rest of the outcrop gentle dips, some as low as 10°–20°, are common, and such minor folding as occurs takes the form of undulations on these gently inclined strata.

Close to open upright folds, similar to those of the Morte Slates, characterise the Upcott Slates. Inland exposure is poor, but this structural pattern appears to persist to the eastern edge of the district.

The stronger and more competent Baggy Sandstones dip uniformly southwards on the coast, except for a few minor parasitic folds. Some contrary dips were noted inland, but most of these occur in the vicinity of fault lines or near the edge of the outcrop and it seems likely that in general little minor folding is present.

The Pilton Shales show much near-upright open and close folding ((Figure 13)b). Such structures occur near the base of the formation, pointing the contrast with the more competent sediments to the north. They are well exposed in dip sections, and in plan on the foreshore, around Down End. These folds have amplitudes and wavelengths of a few metres. It is possible that the thin weathered bands thought by Goldring to be tuffaceous (p. 24) may be correlated and, if so, that they reflect large-scale folds with dimensions of some hundreds of metres ((Figure 13)a); but this is largely conjecture.

Simpson (1971) appears to take the Lower Carboniferous rocks above the Pilton Shales as lying roughly along the southern edge of his structural Zone I. However, any zonal boundary here is arbitrary. Small-scale upright folds within the basal Upper Carboniferous may be measured in metres or a few tens of metres, and much resemble those in the Pilton Shales, while mapping to the east of the district has shown the Pilton Shales and overlying Lower Carboniferous and Crackington Formation strata to be disposed in large regular upright folds. EAE

Folding in the Upper Carboniferous rocks in the south of the district is spectacularly displayed in the cliffs (Figure 15) and (Figure 20). This area lies in Simpson's (1969, 1971) structural Zone II, corresponding to Zone I of Dearman (1969a) and Zone II of Sanderson and Dearman (1973). Upright E–W-trending folds with sub-horizontal axes are typical, their axial planes tending to dip steeply southwards. Stratigraphical marker horizons are repeated by both large-scale and small-scale folding.

The establishment of the stratigraphic succession for the coastline between Bude, Hartland Point and Clovelly (Freshney and Taylor, 1972) has enabled an interpretation of the general structural pattern to be made. The fold envelope is a gently undulating surface with major anticlinoria and synclinoria, those at Wood Rock and Hartland Point having amplitudes of the order of 700 m. RTT

In the rocks of the Bideford Formation on the coast between Cockington and Westward Ho!, Prentice (1960b) and De Raaf and others (1965) have demonstrated a fold pattern of large, more or less open anticlines and synclines with amplitudes of 400 to 450 m. Minor parasitic folds are common on the flanks of these larger structures and much movement has been accommodated by squeezing and distortion of the shaly beds between the thicker sandstones. BJW

The fossiliferous marker horizons in the Upper Carboniferous rocks are found only sporadically inland. However, the occurrences are sufficient to indicate that the pattern of major folds persists across the countryside as broad periclinoria, producing inliers and outliers of the Crackington and Bude formations whose boundaries are too complex to be mapped in detail with the limited exposure available. In (Figure 16) the inferred trends of the major folds in the western part of the area are shown schematically.

The smaller folds vary considerably in scale and form. They may be cylindrical or conical, with parallel, similar, zigzag or chevron profiles. Some are box folds (Figure 17).

Interlimb angles vary from open to tight, rarely being greater than 90° or less than 45° and commonly in the range 50°–60°. Many of the folds in the Crackington and Bude formations in the southern part of the district are asymmetrical; the dip of their axial planes is between 45° and 90° southward, most commonly between 60° and 80°.

Locally axial planes are curved, generally steepening upwards. The dip of the north-facing limbs is commonly of the order of 70°–90°, and locally they are inverted. Fold axes are generally sub-horizontal but also show easterly and westerly plunges of up to 30°, probably associated with periclines. (Figure 18) shows the range of fold plunges in the Hartland area. (Figure 19) illustrates the style of folding.

A few overturned folds occur, as at Speke's Mill Beach (Figure 20), which appear to have been subjected to a second phase of folding. They resemble folds described by Dearman (1967a, b) from the Bude Formation at Duck-pool and the Crackington Formation at Welcombe Mouth, on the coast to the south of the district, where there is clear evidence that the first upright folds have been refolded. The nature of these later movements is uncertain. It seems unlikely that a uniform second phase of deformation occurred in the district, as was suggested by Zwart (1964), since the second folds are overturned to both north and south and are localised in occurrence, although polyphase folding is the rule farther south (Dearman and Freshney, 1966; Dearman, 1969b, 1971). Dearman (1967a) suggested that such folds were generated with opposed sense of overturning during the upward tilting of the flanks of an open synclinorium which developed after the main phase of folding. An overriding southerly movement, directed down-dip on one flank of a syncline and up-dip on the other, has been suggested at Welcombe Mouth (Dearman, 1967b). The folds at Speke's Mill Beach lie in the centre of a shallow synclinorium, but there is no evidence of the development of other folds of this type to the north or south and there is no appreciable difference in the attitude of the folds on the two flanks of the synclinorium. On the other hand an open upright fold with a southerly overturned hinge [SS 2284 2673], seen on the southern flank of a major anticline to the north at Upright Cliff, could have been formed by the slipping of strata down-dip. It seems possible that all these anomalous folds could have been produced by local sliding movement and the collapse of fold hinges under gravity (Freshney and Taylor, 1971), following the relaxation of forces which produced the initial phase of folding. Such effects could have been localised and related to the attitude of the existing structures, and therefore would be variable in sense of overturning. Dearman (1967a) also noted the possibility of a second phase of deformation genetically related to Tertiary wrench faulting.

Much of the folding affects predominantly sandstone successions within which axial-planar cleavage is not usually developed. Shales and siltstones, and more rarely the sandstones in the hinge zones of folds, may develop a fracture cleavage.

The development of prominent slickensides on the bedding planes of some sandstones, usually in the form of a striated film of quartz, results from bed-over-bed slip which is an essential feature in the development of the zigzag and chevron folds found in the district. The striae are invariably at about 90° to the hinge line of folds.

Multiple joint sets are developed normal to the bedding planes in the sandstones, causing the rocks to break into blocks. The spacing of the joint planes varies with the thickness of the individual beds, the thinner the beds the closer the joints. Multiple phases of joint formation have been postulated (Hancock, 1971, 1972), as some joints are open fractures and others have been invaded and sealed by quartz veins. (Plate 7) and (Plate 8) show typical joint patterns in sandstones. RTT

Faults

Prentice (1960a) believed the area immediately south of the Taw–Torridge estuary to be characterised by extensive major thrusting, and Reading (1965) considered the theoretical possibility of much movement along gently dipping thrusts, but there is little evidence to support such interpretations. Indeed the striking aspects of the fault pattern within the northern part of the district are the preponderance of NW–SE fractures (generally dextral with a vertical component), the characteristic subordinate set trending around NE–SW (mainly sinistral with a vertical component), and the virtual absence of E–W dislocations of appreciable size.

In this northern area diminutive bedding-plane slips are universally present, but the few strike faults which have been developed are mostly small. Many of the NW–SE fractures too show only limited movement, locally as little as a few tenths of a metre. But this set of faults includes the major fractures of the district, commonly of the order of several hundred metres and in the case of the Sticklepath Fault measurable in kilometres. A negative aeromagnetic anomaly (Cornwell, 1971) may indicate Tertiary dyke material with a strong reversed remanent magnetisation which has been injected along this fault zone north-westward from Lundy (p. 121). A major Tertiary basin east of Lundy may reflect trough faulting similar to that of the Petrockstow and Bovey basins, Its presence has been proved by boreholes and sampling and its outline delineated by interpretation of sparker records (p. 124). The NE–SW fractures appear to link NW–SE faults, but generally not to cut them. Their typically sinistral sense of movement has in part accommodated the relative stresses between strata on the north-eastern side of one dextral wrench fault and the same beds on the south-western side of a similar fault farther north-east.

Lundy lies on the western side of the dextral Sticklepath Fault, and if the slates of the island are correctly identified as Morte Slates then their position relative to this formation on the mainland suggests a sinistral displacement. This points to the possibility of a major sinistral off-shore fault, or more probably a group of faults, trending around north or north-north-east parallel to the Westward Ho!–Morte Point coastline and cut off by the Sticklepath Fault. Such a structure could go some way towards explaining the anomaly of the Morte Slates occupying a headland while the Pickwell Down Sandstones, the toughest rocks of the district, front on to a bay. Not only might the general line of the coast be fault-controlled, but its present outline may be superimposed from a time when headlands on strike with the present Morte Point and Baggy Point owed their development to the resistance of Hangman Grits and Pick-well Down Sandstones respectively.

Fault control of the north coast has been suggested by a number of workers. The extension of the presumed Cannington Thrust has been inferred (Falcon, 1952; Bott and others, 1958) to lie within a few kilometres of the shore. If present, it probably represents Variscan movement, although the geophysical evidence would admit of an explanation in terms of facies changes (Cornwell, 1971; Donovan and others, 1971; Brooks and Thompson, 1973). The possible influence of Tertiary movements on the form of the northern coast has been mentioned by Donovan (1963) and Owen (1971). EAE

Faulting in the south of the district follows the same general pattern as in the north, and three main sets of faults can be recognised. The commonest are those with a NW–SE alignment, a dextral sense of movement, and a steep westerly inclination; displacements vary between tens of millimetres and hundreds of metres.

The Sticklepath Fault reaches the coast in the vicinity of Babbacombe Mouth, with associated faults to the north towards Greencliff. It gives rise to a series of prominent valleys inland, and has been traced via the upper reaches of the River Yeo, southwards to the east of Orleigh Court. Although hillside wash may be misleading, it appears that the Eocene outlier of Orleigh Court is not truncated by this fault.

In some cases the movement on these NW–SE faults is not a simple one, inclined slickensides on fault planes indicating that both vertical and horizontal components are involved. Barley Bay [SS 235 276] and Shipload Bay [SS 246 275] have been excavated along the lines of NW–SE dextral fractures.

The faults with a NE–SW alignment and sinistral displacement are usually found as secondary fractures associated with larger dextral movements. They are most common in the Bideford Formation where the folds are cut by a conjugate set of north-westerly dextral and northeasterly sinistral faults.

A group of strike faults trends approximately E–W. Examples at Upright Cliff [SS 229 270] and Wood Rock [SS 314 265] have normal southerly downthrows of the order of 200 m. Burne and Moore (1971) postulated a major strike fault with a northerly downthrow along the southern margin of the Bideford Formation. The recent mapping suggests the existence of such a fault at the southern boundary of the belt of Bude Formation which crops out south of the Bideford Formation and the position of this fault can be seen on the coast in Cockington Cliff. A further concealed strike fault with a southerly downthrow is thought to cross the coast at Clovelly; this might account for the absence east of the village of the upper Namurian and lower Westphalian strata found on the coast to the west.

Two major normal faults throw down the Permian rocks against Bude Formation at Peppercombe and Portledge. The fault at Peppercombe strikes NNW–SSE, while the Portledge Fault trends E –W. Both appear to be steeply inclined.

A few E–W-trending faults, mainly seen on the coast around Babbacombe Mouth, dip about 45° S and show evidence of a small amount of thrusting, usually less than 5 m; there is no evidence of any major thrust faults. Small fractures, some showing reverse movement, are found in the hinge zones of some folds. These dislocations are not usually continuous and result from accommodation in the hinges of the tighter folds.

The mapping of faults inland is commonly a matter of conjecture, but is assisted by the use of aerial photographs. Many streams appear to follow the general alignment of faults, but only rarely is evidence of faulting found in stream sections. The E–W faults are almost impossible to trace RTT, BJW

Age of structures

The ages of the structures of the district are known only in broad terms. It seems likely that the folding was effected by a single tectonic phase. Radiometric dating of slates by the potassium-argon method (Dodson and Rex, 1971) gives ages of metamorphism ranging from 330 to 270 million years (late Carboniferous) for the Devonian and Carboniferous in the north of the district. The rocks from the southern part of the Carboniferous outcrop are insufficiently recrystallised for dating. The significance of these ages is uncertain as they probably relate to some extent to the period of cooling associated with uplift and erosion following the folding. The wide scatter of dates may be due to the presence of variable amounts of detrital mica, which would tend to increase the ages obtained.

The paralic sediments of the Bideford area were suggested by Walker (1970) to extend to the top of Westphalian C, but the present palaeontological evidence indicates a limit within Westphalian A. The highest known beds of the Bude Formation lie just above the Warren Gutter Shale [SS 201 110] on the adjacent Bude Sheet. This shale, which is some 1000 m above the top of the succession at Hartland Point, has yielded a fauna indicative of the 'Anthracoceras'aegiranum horizon at the base of Westphalian C (Freshney and others, 1979). Permian strata rest unconformably on folded Carboniferous sediments. In the absence of direct local evidence, such a major break in the succession is best taken as the Carboniferous–Permian boundary, although a Step hanian age has been postulated for the lowest of the New Red rocks in south Devon. A late Westphalian or early Stephanian age for the folding in this region would seem to be most likely (Simpson, 1962).

The pattern of the faulting in the district is symmetrically arranged in relation to the fold axial traces, and is evidently related to the same stress field. Field evidence indicates that the faults post-date the folding. The occurrence of two or more sets of slickensides suggests that in some cases multiple movements have taken place. Faults affect the Permian rocks as well as the older sediments. Most of the fractures were probably initiated during the Variscan earth movements, but many were spasmodically reactivated, especially during the Tertiary Period (Ussher, 1913; Blyth, 1957; Dearman, 1964; Shearman, 1967; Edmonds and others, 1975) and may still move occasionally (Edmonds and others, 1968).

The emplacement of the Lundy igneous mass is discussed on p. 70. EAE, RTT, BJW

Details

Folds

Morte Slates

Between Flat Point [SS 495 472] and Bull Point [SS 462 469] folds are close and axial planes dip about 65° SSW. In many localities along the cliffs slates are superficially folded owing to creep, and in places the rocks have collapsed to lie approximately horizontally. At Bull Point some steep north-north-westerly dips suggest the presence of upright close folds with vertical axial planes. Upright open folds occur in Rockham Bay (p. 16), although the coast hereabouts remains characterised by folds slightly overturned to the north.

At Morte Point close to open folds have sub-vertical axial planes, some dipping steeply southwards, others steeply northwards. To the south, in the neighbourhood of Grunta Pool [SS 452 447], fold limbs dip steeply southwards or are vertical, indicating tight folds very slightly overturned to the north. Traces of an anticlinal axis on the foreshore nearby [SS 4529 4469] suggest a trend to 095° and dips around 70° to N and S. A tight syncline at the side of the coastal road [SS 4547 4466] trends E–W and contains quartz-filled fractures. Around Barricane Beach [SS 454 442] fold axes trend 110° and steep northerly and southerly dips occur. Folding is exposed in plan on the foreshore [SS 4528 4396] at the northern end of Woolacombe Sand.

No inland sections show exposed fold hinges, but some inferences may be drawn from dips (pp.16–17).

On Lundy Island (pp.17–18) an anticline between Ladies Beach [SS 140 443] and Mill Combe [SS 141 440] shows its long limb dipping around 30° NE and passes to the south-south-west into close folds overturned to the south-west, an attitude contrasting with the northward overturning of the northern Morte Slates of the mainland (Figure 13). Similar overfolds persist beneath Castle Hill [SS 140 438] and above the Landing Beach, but pass into open upright folds in the cliffs [SS 140 437] above The Rattles. The only well-exposed fold is of close upright type, a syncline whose axis runs east-south-east through Hell's Gates [SS 1456 4372], between the main island and Rat Island. Limbs of the syncline dip 50°–60° NE and 75°–80° SSW, and the slates in the synclinal core are twisted and contorted.

Pickwell Down Sandstones

A quarry [SS 4776 4137] near Spreacombe Mine shows horizontal strata in the northern part and beds dipping 25°/190° in the western part. N–S faults carrying iron ore (pp.18–19) disrupt the beds but it seems likely that the sediments of this area are disposed in gentle upright folds.

Low dips to slightly west of south persist throughout Vyse Quarry, except in the western bay [SS 491 412]; here the sandstones in the main (northern) face show a general dip of 45°/190° into the quarry, but within that prevailing dip they are disposed in minor close folds whose axes trend E–W with the strike and whose axial planes are mainly steep or vertical (Figure 3).

Upcott Slates

The presence of close or tight folds slightly overturned to the north-north-east in the slates of the coast has been inferred from dips (p. 19). Similarly, in the neighbourhood of Crowborough [SS 4678 3954], northerly dips of 50°–55° and north-easterly dips of 80°, the latter possibly affected by nearby faulting, suggest close folds sub-vertical or overturned southwards.

Farther east, in Nethercott, folds appear to be very slightly overturned northwards. In contrast, exposures around Winsham [SS 499 388], just east of the district, suggest close folds whose axial planes dip steeply south-south-west. The general structural pattern across the outcrop is therefore one of close folding, the folds commonly being nearly upright but locally slightly overturned towards the north or south.

Baggy Sandstones

Strata near the track leading to Croyde Hoe Farm show dips of 40°/220° [SS 4279 4026] and 65°/200° [SS 4283 4015]; these may indicate local tight folds overturned to the north-north-east, but the two localities are separated by a fault. About 1 km to the east dips of 70°/020° [SS 4394 4021] suggest the possibility of close upright folding, but a nearby dextral wrench fault trends parallel to the strike of the sandstones and the beds have been disturbed after folding.

In the northern face of an old quarry [SS 4672 3940] south-southwest of Crowborough medium-grained sandstones with subordinate siltstones and shales lie in an open fold, the axial plane of which is inclined towards the north, and whose limbs dip 70°–90°/210° and 5°/335°.

Pilton Shales

The foreshore due west of Baggy House shows open folding exposed in plan [SS 4267 3998]. A syncline and complementary anticline, with some minor crumpling, trend WNW; the wavelength is a few tens of metres and amplitude possibly similar. Open folds are exposed 450 m farther SE [SS 4295 3965] and are seen nearby to trend to 105° [SS 4323 3961]. Northward overfolding is present locally [SS 4336 3956] near an old limekiln (p. 24).

Folds are well displayed on the rocky shore around the promontory of Down End. They are smaller than those on the northern side of Croyde Sand, perhaps because the latter adjoin the Baggy Sandstones. Much of the folding is close, but several open folds are present. Near the southern edge of Croyde Sand an anticlinal axis trends to 115° [SS 4339 3882]. The northern limb of the fold dips 20° NNE and passes northwards into crumpled and faulted strata in which load casts indicate some inversion; the southern limb dips slightly more steeply south-south-west and shows minor crumpling. Farther south fairly close folds trend between 105° and 115° [SS 4310 3865]. Their axial planes are vertical or dip very steeply around NNE or SSW. Amplitudes are up to 6 m. Crumpled beds hereabouts show some gentle southerly dips [SS 4312 3861]. Open folds are present in which the longer limbs dip northwards; other folds show some southerly overturning [SS 4305 3862]. Immediately to the south, close folds slightly overturned southwards [SS 4303 3853] give way to close upright folds [SS 4307 3850], which are succeeded to the south by vertical beds in which false bedding shows them to young southwards [SS 4309 3848]. The upright folds have amplitudes up to 5 m, and similar folds farther south [SS 4309 3843] trend to 105°. In general the fold limbs hereabouts are roughly equal in length, but where asymmetry exists it is commonly the N-dipping limbs which are the longer, as at [SS 4313 3835].

Open to close anticlines and synclines exposed in plan on the foreshore at the southern end of Down End [SS 4320 3824] show fold axes trending E–W. They are succeeded to the south-east by folds trending between 105° and 115°, although severe crumpling, shearing and small-scale overfolding occurs locally near faults [SS 4347 3804]. Slight southward overturning of folds trending to 105° occurs on the south side of Saunton Down [SS 4361 3790], and overturned folds are associated with faulting and the development of boudins farther east [SS 4420 3780]. A nearby anticline [SS 4427 3782] shows a vertical northern limb and steeply S-dipping southern limb.

Inland sections at Down End show the close near-upright folds of the shore trending to 095° [SS 4332 3845] and small-scale tight folds, trend 120°, associated with faulting [SS 4331 3831]. Traces of fold axes discernible at Saunton trend about 100° [SS 4558 3769] and easterly [SS 4584 3796]. A quarry [SS 490 378] on the extreme northern outskirts of Braunton shows evidence of tight to isoclinal folding overturned southwards, but farther east dips suggest fold limbs vertical and inclined around 40°/360° [SS 4942 3769]. The excavations of Braunton Down Quarry [SS 493 369]; [SS 4946 3695] are much overgrown but there is some evidence of close folds with steep or vertical axial planes. EAE

Upper Carboniferous

Longpeak Beach to Brownspear Beach

The folds exposed in the cliff in this section (Plate 18) are fairly typical of those seen along the coast northwards, with planar limbs and more or less angular hinges, and with N-facing limbs generally more steeply dipping than S-facing limbs, although more symmetrical folds are seen in the middle of Longpeak Beach [SS 2228 2280]. Plunges are mainly easterly at up to 18°. Anticlines show limbs dipping 69°/355° and 43°/167°, plunge 5°/083° [SS 2226 2272], 75°/347° and 50°/180°, plunge 117260° [SS 2280 2282], and 79°/360° and 50°/172°, plunge 9°/089° [SS 2217 2299]. To the north of Longpeak the folds on the foreshore have gentle plunges to west and east of 2°–3°. The slumped bed in the hinge of a syncline at [SS 2235 2328] runs right across the foreshore, indicating a nearly horizontal fold. On the foreshore to the north two anticlinal hinges merge to the west and the syncline between them dies out. Beds at the foot of the cliff at [SS 2241 2329] are overturned to the north but their attitude may in part reflect landslip. An anticline and syncline at the top of the cliff on Swans-ford Hill [SS 2242 2311] plunge 13°/269° and 18°/269° respectively, markedly more steeply than the folds on the foreshore below.

Brownspear Point to Hartland Quay

The prominent anticline forming Brownspear Point [SS 2240 2355] has planar limbs which dip 82°/352° and 58°/005°; the plunge is 16°/265°. At Speke's Mill Mouth [SS 2233 2363] a synclinal box fold on the foreshore has limbs dipping steeply north and south; the two hinges merge eastwards and the fold dies out upwards in the cliff.

Two unusual overturned folds in the cliff [SS 2262 2378] above Speke's Mill Beach (Figure 6) and (Figure 20) are similar to folds described by Dearman (1967a, b). The northern anticline resembles a N-facing box fold in the upper part of the cliff, with one axial plane nearly horizontal and the other nearly vertical. At the foot of the cliff the same fold is upright but the southern limb has developed a second open hinge and steepens southwards to dip 85°/180°. The anticline to the south also has a hinge zone overturned to the north in the upper part of the cliff. The face of the cliff is partly concealed by landslip, but this fold is not represented at the foot of the cliff and seems to have been formed during overturning. The general appearance of these folds gives the impression that the hinge zones of the anticlines have been pushed over to the north. It is not certain that this reflects a distinct later phase of deformation, although the cleavage in the Hartland Quay Shale appears to be warped. The folds lie at the centre of a synclinorial downwarp. At the northern end of Speke's Mill Beach [SS 2237 2397] several small periclines are developed en echelon.

The folds on the foreshore beneath St Catherine's Tor [SS 2251 2412] lie near the centre of an anticlinorium. Steeply dipping northern limbs characterise the anticline. An anticline and syncline on the foreshore [SS 2230 2406] show respectively an undulating crest and a gentle easterly plunge. Folds immediately to the north plunge 3°/254° and 13°/254°.

Between Childspit Beach [SS 2240 2425] and Hartland Quay [SS 2225 2478] the folds are upright, open and more or less symmetrical. On Childspit Beach they plunge west. Those in Screda Cove plunge gently east, one anticline [SS 2233 2442] displaying dips of 52°/350° and 46°/166° and plunging 2°/079°. Between Screda Point and Hartland Quay the folds plunge 2°–4° E; an anticline [SS 2232 2466] having limbs dipping 57°/004° and 51°/178° plunges 4°/091°. The syncline immediately to the north at Cow-house Race has a shallow westerly plunge. At sea level it is asymmetrical, with planar limbs dipping 60°/002° and 36°/185°, but it passes upwards into a symmetrical open concentric fold.

Hartland Quay to Hartland Point

Folds in the near-vertical Warren Cliff to the north of Hartland Quay are shown in ((Figure 6) and (Figure 15). An open symmetrical syncline in the upper part of the southern end of the cliff [SS 2255 2454] appears to correlate with the like fold at Life Rock [SS 2225 2493] indicating the displacement on the intervening faults. To the north of the open fold upright, close, chevron and zigzag folds are developed (zigzag folds, as defined by Fleuty in 1964, are chevron folds with unequal limb lengths). Amplitudes are of the order of 140 m. The zone of chevron folds passes abruptly into an upright open anticline, the northern limb of which steepens sharply to near vertical, giving the fold some resemblance to a box fold tilted to the north. There is evidence of disharmonic folding in the cliff above, where the relatively incompetent Longpeak Shale and underlying siltstones pass over this structure.

On Broad Beach [SS 2245 2535] the close, upright chevron folds have northern limbs of anticlines steeper than southern. This style of folding continues northwards to the Abbey River [SS 2254 2568]. Fold plunges are to east or west. An anticline [SS 2240 2547] showing dips of 82°/347° and 36°/178° plunges 7°/257°. The variability of plunges around Dyer's Lookout and the mouth of the Abbey River suggests that folds there are periclinal, a view confirmed by the disposition of the Longpeak Shale on the foreshore.

Chevron folds are exposed to the north at Blegberry Beach, with a synclinorium centred on the syncline at the mouth of the Blegberry Water [SS 2258 2593]. Fold axes are virtually horizontal. An anticline [SS 2257 2587] shows dips of 80°/350" and 55°/169°. Just south of Damehole Point [SS 2225 2631] a group of periclines is picked out on the foreshore by the crop of the Longpeak Shale and associated siltstones. Typical chevron and zigzag folds occur north of Damehole Point. The anticline at the western end of Smooth-lands has a horizontal axis with limbs dipping 74°/349° and 39°/170°, while the next syncline to the north plunges 8°/075°. The large syncline at the mouth of the Titchberry Water [SS 2284 2664] marks the centre of a synclinorial downwarp. Its limbs dip 80°/004° and 42°/172° and it plunges 10°/093°. The northern limb of the synclinorium is warped by a few minor open folds, including an anticline whose hinge is overturned to the south, and passes into a major faulted anticline in Upright Cliff [SS 2290 2703]. This anticline plunges 10°/253° and raises the Gull Rock Shale to beach level. It and the syncline at Hartland Point are of much greater amplitude than the largest folds described to the south and are essentially single folds with some superimposed minor folds. They share a vertical or locally overturned N-facing limb. Strata at Cow and Calf [SS 2272 2713] dip 87°/342° and at Blagdon Cliff [SS 2298 2748] overturned beds dip 52°/174°, although the overturning may have been produced, at least in part, by superficial collapse. Southerly dipping beds in the southern flank of the Upright Cliff anticline extend inland along the Titchberry Water as far as Shamley Bridge [SS 2463 2678].

The Hartland Point syncline [SS 2301 2761] (Plate 17) has an amplitude of about 700 m. It contains the highest part of the succession found in the Hartland area. The syncline plunges 9°/080° at Hartland Point, in a direction opposite to that of the Upright Cliff anticline.

Hartland Point to Mouthmill

The Hartland Point syncline crosses the point as a single fold, but between Barley Bay and Shipload Bay (Plate 15) and (Plate 16) additional folds are superimposed on the main structure and the easterly plunge becomes less steep. An upright, parallel anticlinal fold developed in the centre of the main syncline is well displayed in the cliff on the east side of Barley Bay. In Shipload Bay the axis of the syncline has been offset by dextral faulting; it continues beyond the fault at the eastern end of East Titchberry Cliff [SS 2495 2738], and can be traced in the cliff top to the east.

Between the Shipload Bay fault and Eldern Point [SS 2482 2760], on the northern limb of the syncline, a tight anticline and syncline show axial planes dipping about 45° S; fold style is between parallel and similar. More open asymmetrical parallel folds can be traced along the strike to the east for 850 m; the limbs dip about 70° N and 40° S and the axial traces remain virtually the same distance apart over this distance. The fold axes are about horizontal. On the foreshore between [SS 2580 2764] and Chapman Rock [SS 2677 2764] the anticlines and synclines occur in related pairs which develop from a point, with the axial traces diverging as the folds increase in wavelength and amplitude. In one case the axial traces diverge from one point and converge on another; two other anticlinal/synclinal fold pairs diverge near the point where the first pair dies out, but their axial traces continue to diverge to the east of Chapman Rock. These folds are open and upright. The anticlines generally plunge towards the point of convergence of the axial traces while the synclines plunge away from it.

To the east of Chapman Rock the coast runs south-east across the strike and successively younger beds are exposed. A syncline [SS 2721 2748] develops a central anticline as it passes up the cliff. The next anticline to the south has limbs dipping 62°/352° and 42°/182° and plunges 6°/265°. A major syncline [SS 2753 2728], containing the highest part of the succession present hereabouts, appears to be the eastern continuation of the Hartland Point syncline. Its limbs dip 88°/359° and 23°/178° and there may be a very gentle westerly plunge. The hinge is angular and the limbs are planar, with a slight undulation on the near-vertical southern limb. A series of anticlines and synclines of zigzag profile [SS 2677 2715]–[SS 2785 2701] have easterly plunges decreasing from 10°/111° to 0°/092°, the latter in the southernmost syncline. The southern limb of this syncline dips 88°/002° close to the hinge; to the south it is overturned to dip 64°/187° [SS 2790 2696]. Overturned beds crop out south-eastwards towards the hinge of an overturned anticline which contains the Longpeak Shale and is exposed at beach level [SS 2801 2690]. The upper limb dips 45°/004°, the inverted lower limb is poorly exposed but dips fairly steeply south and is nearly vertical on the foreshore. The axial plane dips about 40° S. Faulting to the east may have affected the attitude of this fold, although a similar flexure occurs at Warren Beach (p. 94). A less overturned anticline on the foreshore [SS 2830 2678] may be the continuation of this fold.

Inverted beds at the foot of Beckland Cliff may have had their attitude accentuated by toppling (p. 113) associated with faulting. An anticlinal/synclinal fold pair on the foreshore east of the faults in Beckland Bay converges and dies out eastwards. The anticline has its northern limb dipping 807356°, but the syncline to the south is open and upright with both limbs dipping at about 40°.

South of Windbury Point [SS 2872 2682] the northerly dip steepens from 40° to 86° and successively lower beds are exposed towards Mouthmill. A group of open upright folds [SS 2910 2665] have near horizontal axes. At the western end of Mouthmill Beach [SS 2925 2657] a syncline at beach level shows a box-like profile but passes up the cliff into a close parallel fold with a rounded hinge (Figure 17). The adjacent anticline to the south is a zigzag fold with planar limbs and an angular hinge. The axial traces of these folds pass across Mouthmill Beach to the fault at Mouthmill [SS 2979 2655].

Mouthmill to Clovelly

The generalised profile of the folding in this section is shown in (Figure 7). To the north of Blackchurch Rock [SS 2989 2664] a small anticline plunges 10° W, the adjacent syncline shows dips of 46°/002° and 49°/182° and a horizontal axis. Successively lower beds are exposed southwards from the Hartland Quay Shale at Blackchurch Rock to the Gull Rock Shale, which crops out in an anticline at [SS 3006 2658]. The northern limb of this fold steepens until slightly overturned, dipping 84°/186°; the southern limb dips 317187' and the axis is about horizontal. The large anticline and syncline to the south-east, beneath Gallantry Bower [SS 3047 2624], plunge 10°/263° and 57261° respectively. They have planar limbs and fairly rounded hinges. The single hinge of the anticline breaks up into a complex of small folds in the incompetent thin sandstones and shales exposed in the core of the fold on the shore platform.

Strata south-east of the syncline at Gallantry Bower lie in the slightly overturned northern limb of the Wood Rock anticline, a major anticlinorial structure comparable in amplitude to the folds at Hartland Point. Typical dips are 50°/169° [SS 3083 2601] and 75°/167° [SS 3110 2573]. The lowest beds of the succession occur in an anticline [SS 3120 2566] north of Wood Rock, whose southern limb dips 19°/207° and which plunges 9° W. The syncline to the south is a typical planar-limbed fold with a small strike fault along its axis.

The simple form of the Wood Rock anticlinorium is complicated by a large southerly downthrowing strike fault south of Wood Rock [SS 3145 2550] and by the smaller folds developed on the main structure to the south-east. However, the main fold can be traced westwards in broad outline by means of fossiliferous nodular shale horizons. Two occurrences of the Embury Shale carrying G. subcrenatum, at [SS 2868 2583] near Brownsham Farm and at Colpit (Hescott) Quarry [SS 2789 2492], together with the overlying shales, appear to define the northern and southern flanks. The Gull Rock Shale, with G. circumnodosum, south of Hartland Mill [SS 2483 2461] seems to lie near the western end of the fold which plunges gently westwards. South of Wood Rock, towards Gallant Rock [SS 3170 2520], small folds plunge at the relatively high angles of 15°–23° E; they maintain the typical style of steep northern limbs to the anticlines. Just north of Skittering Rock a large close anticline [SS 3173 2512] with a vertical to slightly overturned northern limb has an angular hinge and planar limbs. It plunges 17°/090° and a minor strike fault runs parallel and close to the axis. An upright close anticline just north of Clovelly plunges west.

Clovelly to Peppercombe

To the east of Clovelly a syncline and anticline [SS 3200 2466] have planar limbs and angular hinges. The limbs of the syncline dip 25°/006° and 44°/192° and the plunge is 4°/278°. Just south of Devil's Kitchen [SS 3220 2455] is a series of small tight northerly-overturned folds, some of them nearly isoclinal. The cliff face between Devil's Kitchen and Black Rock [SS 3286 2421] is obscured by slip, but evidence from boreholes sunk in connection with a proposed hydroelectric pumped storage scheme suggests the presence of folds quite strongly overturned to the north.

Between Black Rock and the Lower Bight of Fernham two near-isoclinal anticlines show northern inverted limbs dipping 80° and southern limbs 70°–75°; the folds strike 080°. Farther east, towards the cliff below Barton Wood [SS 3416 2394], the coast section follows the strike of a northerly-overturned anticline whose hinge falls eastwards from 15 m above the base of the cliff at [SS 3343 2385] to foreshore level at [SS 3416 2384]; the northern limb is vertical, the southern dips at 70°, and the axis trends 080°. At West Buck's Point [SS 3473 2383] an anticline plunges 10°/260°; the northern limb dips at up to 70°, the southern at 50° steepening eastwards to 75°.

The foreshore at Buck's Mills [SS 353 237]–[SS 360 238] is occupied by a large pericline with axial trend 082°. The northern limb dips at 80°–90°, the southern at 60°–70°. A nearby complementary syncline [SS 3594 2386] plunges 10°/262°. In the cliff at [SS 3600 2374] a northerly-overturned anticline shows a northern inverted limb dipping at 85°, a southern limb at 75°–80°; the axial trend of this fold is 078°, and the hinge may be traced eastwards in several cliff exposures to [SS 3634 2388]. On the foreshore at Gauter Point [SS 3635 2404] an anticline, plunge 10°/279°, shows both limbs dipping at up to 75°.

Portledge to Greencliff

On the foreshore west of Portledge [SS 385 248] a rather box-shaped syncline plunges 8°/095°; its southern limb is attenuated by the major Portledge Fault. Dips are up to 85°. A similarly plunging syncline crops out on the foreshore [SS 3874 2490] 180 m to the NE. Two minor synclines occur [SS 388 250] 365 m NE of Portledge Mouth; the northern one plunges 10°/090°, the southern one 8°/280°. The intervening anticline is steeply dipping and its hinge region broken by a fault. Below Higher Rowden cliff [SS 390 252], two synclines on the foreshore plunge 6°/085°; the limbs of these folds and the anticline between dip at up to 70°. In the cliff at [SS 3909 2519] an E–W-trending anticline in massive sandstones has a vertical northern limb and a southern limb dipping at 50°. Farther north, in Higher Rowden cliff [SS 3913 2524], the northern limb of this anticline becomes inverted; it forms the southern limb of a tight E–W syncline, the northern limb of which is normal, dipping 50° S. Massive sandstone at [SS 3923 2534] near the northern end of Higher Rowden cliff is disposed in an anticline trending 075°, with its northern limb dipping at 85° and its southern at 45°. A fracture in the hinge region has the appearance of a thrust, but there is little evidence of movement. On the foreshore 90 to 180 m NW of this anticline two eastward-plunging synclines are separated by a highly fractured zone.

In Westacott Cliff [SS 3953 2574] the northern vertical limb of an E–W-trending anticline is truncated by a fault; the southern limb dips at 65°. About 150 m to the NE [SS 3962 2586] a similar anticline trends E–W. In Cockington Cliff [SS 3985 2630] an anticline exposed in cliff and foreshore is unusual in that the southern limb is steeper (70°) than the northern (45°). The fold plunges 6°/278°. Some 35 m to the NE [SS 3988 2632], an anticline of unusual (north-westerly) trend is truncated by a fault in its southern limb. Northwards, for nearly 200 m, as far as [SS 3992 2652], the foreshore reveals a series of periclinal anticlines and synclines, all trending to 265°. Dips are up to 50°, and the folds plunge at angles up to 5°. An anticline in thickly bedded sandstones [SS 4031 2665] forms a prominent reef on the foreshore at the southern end of Green Cliff, whilst a symmetrical open anticline plunges 5°/272° on the foreshore at the northern end of Green Cliff [SS 404 270]; both limbs dip at 50°.

Greencliff to Westward Ho!

The Bideford Formation rocks of this coast lie in large open anticlines and synclines with amplitudes of 400 to 450 m. A major anticline trends 288° between Abbotsham and Cornborough cliffs, but its hinge region is obscured by faulting in the cliff [SS 4112 2790] and foreshore. A small anticlinal flexure on the southern limb of this fold is exposed in the cliff [SS 4085 2752] and shows the mudstones of Cycle 7 of the Bideford Formation squeezed into irregular folds within its hinge zone. The Cornborough Sandstone is brought down to the base of the cliff [SS 4133 2825] in the large syncline to the north; this fold trends 278°, with limbs dipping at up to 70°. The next anticline to the north [SS 415 286] is broken by wrench faulting. The northernmost pair of major folds in this foreshore area comprises a syncline [SS 4129 2870] and a complementary anticline 50 m to the north. The anticline has the Raleigh Sandstone as its core.

Appledore

Mudstones with thin sandstones of the Crackington Formation lie in tight periclinal anticlines and synclines on the foreshore [SS 461 311] at Appledore. The sandstones are well-jointed. Dips are up to 85° and the axial trend 260°. The amplitude of these folds is about 20 m. BJW

Instow area

Open to close upright folds are exposed on the estuary shore to the north [SS 4820 3274] and west [SS 478 324] of East Yelland Power Station. Wavelengths are of the order of 50 m and axes trend E–W. Similar folds characterise the rocky foreshore at the northern end of Instow Sands; there are indications here of steep (65°–85°) southerly-dipping limbs and more gently inclined, but variable (15°–55°), northerly-dipping limbs.

Roadside exposures around Instow town [SS 482 311] show dips of 50°–70° to slightly E of N and 65°–70° around SSW. In a stream 300 m SSE of Raddy, fold limbs dip 75°/010° and 607190° [SS 486 301]. The fold style displayed on the foreshore, with S-dipping limbs inclined more steeply than N-dipping limbs, is repeated in a small disused quarry [SS 4914 2980] near Huish; an anticline here shows its southern limb inclined 70°/200° and its northern limb 15°/010°. In contrast, an old quarry [SS 4977 2844] 1.5 km farther SSE, in which southerly-dipping strata prevail, shows in its northern wall traces of an anticline whose northern limb is vertical and whose southern limb dips 80° S. EAE

Faults

Morte Slates

Although many small faults, mainly of NW–SE trend, have provided lines of weakness which have been etched out by the sea, few are well exposed. One [SS 4742 4676] occurs just west of Lee Bay and two [SS 4589 4630]; [SS 4587 4623] on the north side of Rockham Bay, all trending NW–SE. The cliffs at Rockham Beach show collapsed slates between two near-strike faults, one trending WNW [SS 4590 4619], the other W [SS 4591 4617]. Near the foot of the steps leading down to the beach a low-angle strike fault dips around 30° SSW, flattening higher in the cliff, and is accompanied by disrupted vein quartz [SS 4587 4607]. Immediately to the south are two faults dipping steeply south [SS 4586 4603] and south-south-west [SS 4584 4601]. Strike faulting [SS 4527 4395] near Shellsborough Cove is accompanied by sheared boudins of siltstone containing pyrite cubes, and immediately to the south occur numerous minor faults trending about N–S.

Inland exposure is poor; some faults have been mapped on evidence of topography, but numerous small fractures must be present although obscured. At Lee a strike fault is inclined at about 70°/230° [SS 4817 4626] and another about 80°/210° [SS 4816 4623].

A faulted junction between slates and granite is exposed on the south coast of Lundy Island (p. 83). Near Mill Combe several faults parallel to bedding dip 55°/045° [SS 1409 4400]. Immediately south of the trachytic dyke [SS 1428 4379] at the foot of the steps leading up to Lundy South Lighthouse (p. 85) a fault trends to 020°. At the eastern end of the Landing Beach a small gully [SS 1446 4370] has been cut by the sea along a fault trending just E of N; the fracture zone is 0.15 m wide and filled with clay. Nearby [SS 1449 4369] slates have collapsed north-eastwards down faults parallel to bedding.

Pickwell Down Sandstones

Slates and sandstones in a quarry [SS 4610 4305] east of Potter's Hill are cut by strike faults trending E–W and filled with ferruginous clayey gouge. Iron-rich fault rock is exposed at the coast [SS 4470 4081] near Putsborough Sands Hotel and in a quarry [SS 4566 4142] just over 1 km to the ENE, where it fills a 1-m-wide fracture zone trending to 320°; similar rock is found as fragments in a quarry [SS 4773 4247] over 2 km farther ENE.

The last-mentioned locality lies 80 m N of two old adits (p. 18) and 1100 m N of the quarry [SS 4776 4137] in which is exposed the orebody once worked in Spreacombe Mine. Three near-vertical faults in the quarry strike N–S, they are up to 0.9 m wide and filled with hematitic breccia. Faults cutting the sandstones in the workings of Vyse Quarry trend to 175° [SS 4905 4118], 130° [SS 4934 4131] and north-west or north-north-west [SS 4933 4135] and are accompanied by minor fractures. The presence of ferruginous breccia on the dumps at Buckland Mine (p.115) suggests that the north-easterly-trending workings probably followed a fault [SS 4905 4029]–[SS 4922 4051]. In Stony Bridge Quarry a fault [SS 4933 3949] containing a few centimetres of ferruginous clay dips about 80° SW; it rises up the face on strike with steeply inclined sandstones and slates and becomes coincident with bedding near the top of the quarry.

Upcott Slates

Several small gullies on the coast have been excavated along fractures but no good exposure of a fault was noted. Inland, the presence of dislocations is largely inferred from topography. An anomalous dip of 80°/310° [SS 4675 3954] at Crowborough may be attributed to a fault coursing northwards and north-eastwards along a small valley immediately to the east.

Baggy Sandstones

In the small cove on the northern side of Baggy Point a fault [SS 4210 4068] dips 85°/210° parallel to bedding; brecciated rock is present and ferruginous staining extends over a width of 2 m. A parallel strike fault [SS 4207 4065] is exposed 40 m to the SW in the same cove. Baggy Point is pierced by a N–S cave [SS 4193 4065]–[SS 4193 4058] which appears to have been excavated along a fracture. A major NW–SE fault extends from the north coast of the promontory [SS 4330 4082] almost to Croyde. Inland, a fault exposed in a quarry near Crowborough dips about 80° NW [SS 4671 3938].

Pilton Shales

Two of the several small faults which cut the basal Pilton Shales on the coast are exposed on the north side of Croyde Sand [SS 4334 3957]; [SS 4342 3951]. On the foreshore at Down End faults with displacements of up to 3 m are exposed [SS 4338 3881]; [SS 4332 3882]; [SS 4330 3879]; [SS 4317 3872]; [SS 4318 3870], as are rucking and crumpling associated with a fault [SS 4320 3865]. Contorted, crumpled and sheared strata with distorted sandy ovoids mark an E–W fault on the south side of Down End [SS 4347 3804], and a number of small faults occur farther east e.g. [SS 4394 3787]; [SS 4420 3780]. An old quarry [SS 4671 3938] south-south-west of Crowborough contains a fault dipping about 80° NW (pp. 24–25). On the southern side of the Taw–Torridge estuary faults exposed on the shore trend about 110°–120° [SS 491 333]; [SS 496 334]. EAE

Upper Carboniferous

Brownspear Beach [SS 2245 2345]

Two strike faults trending E–W and with a total southerly down-throw of about 200 m bring the lower part of the Bude Formation against the upper part of the Crackington Formation. Zones of brecciation cross the beach and pass up the cliff, the fault planes dipping steeply to the south. The northernmost fault can be traced in the stream section of the Milford Water, and both faults are associated with springs and gullies on the eastern side of the valley of the Milford Water. Reddening of the soil east of Little Barton [SS 2410 2365] may indicate the general line of their continuation.

A small NW–SE fault with a dextral displacement of about 7.5 m crosses Brownspear Beach but there is no clear evidence that this fault displaces the strike faults. A further small strike fault to the south of those described above appears to terminate against the dextral fault.

Childspit Beach [SS 2235 2433]

A small NE–SW-trending sinistral wrench fault with a lateral displacement of about 9 m crosses the northern end of Childspit Beach and is associated with a dolomite-cemented breccia. Slickensides pitch 15° SW on the sub-vertical fault plane in the cliff. The fault can be traced a short distance inland by means of a spring line and it appears to intersect the Hartland Quay–Kernstone fault line (below).

Hartland Quay [SS 2225 2478] to Damehole Point [SS 2225 2631]

At Hartland Quay and the adjacent Warren Beach, a group of NW–SE-trending dextral wrench faults with sub-vertical planes shows a total lateral movement of some 200 m. Small faults west of the main group include one trending NE–SW with a sinistral displacement of about 4 m. Inland the general line of the faults can be traced by means of topographic features. Reddened fault breccia is found in the stream section north of Kernstone [SS 2320 2365]. Between Damehole Point and the mouth of the Abbey River four small dextral faults occur. The eastern two can be traced southwards towards Blegberry.

Upright Cliff [SS 2290 2700] to Hartland Point

At Upright Cliff a major normal strike fault dips about 60° S and throws down about 200 m, the Gull Rock Shale and the Hartland Quay Shale being nearly juxtaposed. About 1 m of breccia is present on the foreshore, and slickensides indicate a small sinistral component of movement.

North of Cow and Calf [SS 2272 2715] several small NW–SE-trending dextral faults occur, together with one small conjugate sinistral fault. The total dextral displacement is of the order of 25 m. The most easterly fault of the group produced horizontal slickensides and can be traced for about 2 km SE. A small dextral fault just west of Hartland Point has a maximum lateral displacement of about 9 m but dies out south-eastwards on the foreshore; the movement may be taken up by adjustment along the strike of the bedding.

Barley Bay [SS 2350 2760] to Beckland Bay

Barley Bay appears to have been eroded along the line of a NW–SE-trending dextral wrench fault which displaces by 100 m the axis of the major syncline running E–W through Hartland Point. The fault, which dips steeply west, appears also to throw down westwards. The fault line can be traced to the south-east as far as the Titchberry Water by means of surface features.

Shipload Bay [SS 2480 2740] has a similar shape to Barley Bay and it too has been excavated along a large dextral fault. The lateral displacement is estimated to be 400 m and there has also been a westerly downthrow. The Shipload Bay fault has been traced south-eastwards by surface features for about 2 km. A stream section [SS 2644 2541] reveals two zones 0.5 to 0.7 m wide of clayey fault gouge and crushed sandstone trending 320° and dipping 80° SW; reddened head extends for 20 m downstream and may conceal other fracture lines.

Two small dextral wrench faults west of Chapman Rock [SS 2677 2764] show lateral displacements of 11 and 15 m and can be tentatively traced inland for 1 km by surface features. A minor sinistral fault occurs at Little Chapman Rock to the east.

Beckland Bay [SS 2810 2690] contains two dextral wrench faults. The western one is vertical with horizontal slickensides and a thick crush breccia, while a minor dislocation just west of the main fault dips west and has slickensides which indicate some normal movement. The lateral displacement on the main fault is uncertain but possibly 150 m. The eastern fault has a dextral displacement of about 35 m.

Mouthmill Beach [SS 2950 2655]

Minor NW–SE-trending faults at the far western end of the beach have a total apparent dextral displacement of about 10 m. Two minor dextral faults occur near the centre of the beach; the western one has moved obliquely, with a vertical component throwing down west. At Mouthmill the main line of dislocation is marked by a distinct gully on the foreshore and the fault continues inland along the valley to the south-east. A maximum lateral movement of about 250 m is suggested by displacement of the Hartland Quay Shale between Windbury Point and Black-church Rock.

Wood Rock [SS 3145 2554]

About 100 m N of Wood Rock an E–W-trending fault in the axis of a syncline has a small southerly downthrow and dips steeply south. A dextral wrench fault with about 5 m lateral displacement passes east of Wood Rock. Immediately south of Wood Rock an E–W strike fault dips south; normal movement is indicated by vertical slickensides and terminal flexuring. This fault repeats the outcrop of the Embury Shale and has a vertical displacement of about 200 m.

Clovelly to Westward Ho!

Two NE–SW faults occur south-east of Clovelly [SS 3205 2463]. That the northern one is a wrench fault is indicated by sub-horizontal slickensides. These faults are probably sinistral, and both lie on the line of the fault in the NE–SW valley of the Clifford Water to the south-west. RTT

In the cliff at Lower Bight of Fernham [SS 3345 2384] a normal fault trends north-easterly, dips 30° NW and throws down 3 m. At 670 m to the east [SS 3416 2384] a fault in the hinge region of a northerly overturned fold dips 35° S and shows evidence of reversed movement of less than 1 m. Minor north-westerly dextral wrench faults intersect the large pericline on the foreshore at Buck's Mills [SS 365 237].

There are two major faults in the region of the coastal Permian outlier. The Peppercombe Fault dips 80°/240° and cuts the cliff at Giffard's_.Jump [SS 3817 2430], where Permian rocks to the southwest are thrown down against Crackington Formation. The fault displaces the basal Permian unconformity by at least 200 m in a sinistral sense, though the major displacement is the normal south-westerly downthrow. The northern boundary fault (Portledge Fault) of the outlier throws down southwards, trends E–W, and is visible as 1 m of fault gouge in a poorly exposed section in the cliff [SS 3870 2472]. A few normal faults occur within the Permian coastal section; they throw down only 2 or 3 m southwards and dip about 10° to 15°.

In the cliff [SS 3906 2514] near Rowden Gut, two steeply inclined faults, separating cleaved shales from shales with sandstone bands, appear to have both wrench and normal components. Rowden Gut itself is an embayment in the foreshore probably controlled by a north-westerly wrench fault associated with these faults.

In Westacott Cliff vertical or near-vertical wrench faults occur at [SS 3951 2570], [SS 3952 2573], [SS 3954 2575] and [SS 3956 2578]; the first and last of these attenuate the northern vertical or overturned limbs of asymmetric anticlines. In the cliff near Paddon's Path [SS 3969 2597] the rocks are broken and squeezed; landslip obscures much of the succession but it is thought that the Sticklepath Fault crosses the coast hereabouts. Rocks containing G. listeri occur immediately to the north, and the G. amaliae Marine Band is found just to the south.

An E–W fault [SS 3974 2609] in Cockington Cliff is thought to be the northerly-downthrowing major fault originally postulated by Burne and Moore (1971). The present southerly dip may be the result of slip. A black nodular shale found in the fault gouge yielded G. listeri, as did a nearby locality [SS 3968 2610] in the Crackington Formation to the south of the fault. North of the fault the rocks belong to the Bude Formation.

Farther south in Cockington Cliff [SS 3994 2636] an E–W fault inclined at 45° appears to be reversed, with a displacement of about 5 m. Scattered dextral wrench faults occur on the foreshore as far north as Green Cliff. North of Green Cliff, Bideford Formation rocks are transected by a series of conjugate wrench faults, almost all the north-westerlies having dextral movements and most of the north-easterlies being sinistral. The larger faults can be traced inland by displacement of sandstone features.

Instow area

Faults trend around E–W on the foreshore [SS 4785 3248] west of East Yelland Power Station and north of Instow Sands [SS 4747 3217]. In both places the fractures are on strike with fold limbs. Displacements are small, commonly of the order of a few metres. EAE

Chapter 8 Eocene

The outlier of flint-gravel at Orleigh Court [SS 4296 2227] extends into the southern part of the present district, and covers an area some 300 m by 400 m, 1 km south-west of Littleham. Exposure is poor and confined to ditches and roadside banks. The deposits consist of poorly rounded nodular flint pebbles set in buff and brownish red sand. The sand matrix varies from coarse to fine and contains a variable small proportion of silt and clay. Geophysical surveys (p. 124) suggest a maximum thickness of about 50 m for the deposit.

Rogers and Simpson (1937, p. 309) described the outlier and gave an account of the mineral assemblage as well as a list of the fossils found as flint-casts. They described upper and lower portions of the gravel; the upper was red, rich in iron oxide and flints, and contained fragments of a ferruginous grit of possible Lower Greensand origin, whilst the lower gravel was buff-coloured, with fewer and smaller flints. Additionally the upper gravel contained more silt and clay, and a zircon-rutile-tourmaline heavy mineral suite; the lower gravel was characterised by a tourmaline-zirconandalusite assemblage. The fossils were listed as being numerous in the upper gravel, and included forms derived from all the zones of the Upper Chalk from Sternotaxis planus to Belemnitella mucronata. The recent survey suggests the possibility that the upper gravel of Rogers and Simpson is the original deposit and the lower gravel a downslope weathering product derived from it.

Rogers and Simpson assembled evidence indicating a likely Pliocene age for the gravels, and considered them to have been laid down on the 122-m (400-ft) erosional platform, either in a small depression or as part of a once more extensive deposit. There are several indications, however, of an older age for the deposit. The somewhat angular flints which make up most of the gravel do not appear to have been transported far, suggesting that the gravel may be a residual deposit derived from the Chalk. The Orleigh Court gravel resembles the Haldon Gravels (Hamblin, 1973), and basal gravels in the Bovey and Petrockstow basins. Its heavy mineral assemblage is similar to that of the Haldon Gravels (Selwood and others, in prep.). The south and central Devon gravels are generally regarded as Eocene (Hamblin, 1973, p. 468), and this age is suggested for the gravels of Orleigh Court. BJW

Chapter 9 Pleistocene and Recent

Introduction

It is possible to deduce a sequence of events in the Bideford district in Pleistocene times without being able to assign proved dates. However, there is general agreement, largely stemming from comparisons with southern Ireland, that the boulder clay of the Taw–Torridge estuary is of Wolstonian<span data-type="footnote">East Anglian and Continental terms are used here, with the relationships shown in (Table 12), to facilitate comparisons with other works. They offer a convenient scheme for the location of events within a sequence of glacial and interglacial periods.</span> age (Mitchell, 1960, 1972; Stephens, 1961, 1966, 1970), and this correlation permits the main Pleistocene deposits of the district to be fitted into a time scale (Table 12).

No erratics have been found inland to match the size of those on the coast, and this suggests that the coastal boulders represent a glaciation other (and hence earlier) than that responsible for the boulder clay. Pebbly clay and sand derived from the inland glacial drift has been planed off at 1st Terrace level, and terrace deposits of this age equate with the raised beach. Hence both raised beach and 1st Terrace post-date the Wolstonian stage, and they are assigned to the Ipswichian. The raised beach is overlain by old blown sand and this in turn by Head.

A few isolated patches of drift cannot be fitted into the scheme. They include terraces near the mouths of the Abbey River and the Milford Water, adjoining a coastline subject to vigorous marine erosion, which although of relatively recent origin can be related neither to each other nor to the terraces of the Taw and Torridge rivers. The rock floors of the valleys of the Milford Water, Abbey River and Titchberry Water are graded to a sea level lower than the present one but of unknown date.

Recent deposits include alluvium, submerged forest, storm beach (the Pebble Ridge), beach material and blown sand.

Landslips are not common inland, but the cliffs, especially those south of Westward Ho!, are capped by large patches of slipped strata. EAE, RTT

General account

Pleistocene

Probably the oldest drift deposits of the district are represented by scattered erratics which occur near beach level. The two largest, near Baggy House and Saunton, are overlain by raised beach deposits. Most workers now agree that these erratics arrived in Anglian (Elster) ice which deposited them on a wave-cut platform carved in earlier, possibly Cromerian, times (Table 12). Bowen (1969), however, although publishing his views only in abstract form, preferred a Hoxnian age for the platform, thus implying that the large erratics were emplaced by a Wolstonian glaciation. Kidson (1971) has pointed to the possibility of a composite age for the platform, a Cromerian shelf being re-trimmed by interglacial seas throughout the Pleistocene.

The size and disposition of the big coastal erratics point to a mechanism of emplacement that permitted large boulders to reach the coast but not to be transported inland. Mitchell (1960) envisaged an Anglian ice sheet which stood against the north coasts of Devon and Cornwall and extended south to Porthleven. He considered it to be the source of both the large erratic blocks of the shore and platform and of the smaller erratics of the raised beach, which he regarded as of Hoxnian age. However he later (1972) expressed doubts about his earlier views on the extent of Anglian ice. Relationships established during recent surveys have shown that the raised beach is probably of Ipswichian (Eemian) age, and therefore likely to contain reworked Wolstonian glacial material.

Thus the balance of evidence points to the large coastal erratics being the only surviving relics of Anglian glaciation, and raises the questions of the mode of emplacement of these blocks and the extent of the ice. Mitchell (1960) first thought that the blocks were placed more or less in their present positions by an ice sheet, dismissing the possibility that they travelled south in floating ice on the grounds that no agency existed for lifting the larger erratics to where they are now found. Kellaway (1971) agreed, noting that the floating ice theory would require a sea level 'as high or higher than at present; this during a major glaciation'.

As early as 1878 Ussher (MS notes) had preferred the idea of transport in ice floes, and Hughes (1887) maintained that however the erratics had travelled they were placed in their present position by waves. In recent years Stephens (1961, 1966) has favoured a floating ice mechanism and Edmonds (1972) has argued that pack ice could have drifted south not only at the height of glaciation but also during the period of rising sea level that marked the waning of the cold and the approaching warmer climate of the Hoxnian Interglacial.

Clearly, erratic material could have travelled on ice floes, as acknowledged by Mitchell (1972), in which case the ice front may have advanced no farther south than the Bristol Channel. Mitchell's (1960) alternative hypothesis requires an ice sheet pressed against the whole length of the north Somerset, north Devon and north Cornwall coastline, advancing up the estuaries and sweeping across the peninsula from St Ives Bay to Mount's Bay in the extreme southwest. Yet the only deposits attesting the presence of this ice are a few scattered blocks and perhaps some of the reworked fragments in the raised beach. Kellaway (1971) supposed that Anglian glaciation was even more extensive, with ice at well over 305 m above present sea level in the vicinity of Cheltenham, Welsh erratics carried across country to Salisbury Plain and ice advancing eastwards up the English Channel to the coastal areas of Hampshire and Sussex. The implication is that south-west England was covered by ice, with an Anglian sheet abutting against local ice caps moving outwards from high ground. The absence, or non-discovery, of undoubted inland erratics in Devon and Cornwall might be explained in part by this disposition of ice and in part by the obscuring effects of later Head deposits. Nevertheless the known erratics of probable Anglian origin in south-west England are confined to the coast and inland evidence of possible ice action is at present largely topographic. Build-up and spread of local ice could have complicated the interpretation of glacial topography. The field is currently wide open to speculation and theory, unencumbered by too rigid a framework of fact.

The low sea level of Anglian times was succeeded by rising waters with the transition to the Hoxnian Interglacial. Possibly the sea rose to 30 m OD (Orme, 1962). If, as is likely, deposits of this age were laid down in the district, it seems that none has survived, except perhaps for ice-rafted erratics of early Hoxnian times. Several writers, and notably in recent years Mitchell (1960, 1972) and Stephens (1961, 1966), have assigned the raised beach and overlying sands to this period, considering that they lay beneath the Wolstonian boulder clay at Fremington. How ever, no connection has been demonstrated between gravels at raised beach level, on the coast or in the estuary, and gravelly clays spasmodically exposed in inland pits in glacial clays (Edmonds, 1972).

At the end of the Hoxnian or Great Interglacial ice once more advanced southwards across the Irish Sea to north Devon. Mitchell (1968) noted erratics at 84 to 107 m OD on Lundy Island, and considered some granite outcrops there to show rounding due to ice action. He related this evidence to the passage of Wolstonian ice, and visualised Lundy surrounded by ice with those parts of the island above about 107 m standing up as a nunatak. Ice may have been pressed on to the plateau in the north of the island, moulding some granite into roches moutonnées and giving rise to an outwash gravel, now represented by scattered pebbles, and to the easterly drainage to Gannets' Bay.

The Wolstonian ice sheet first blocked the Bristol Channel and then advanced up it. It pushed a tongue up the Taw–Torridge estuary to Barnstaple, blocking that major drainage course and perhaps causing the waters to escape westwards via Pillhead [SS 476 270] towards Bideford and an unknown outlet to the sea along the then southern edge of the ice sheet. Mitchell (1960) suggested the name 'Lake Maw' for the water impounded in the Bristol Channel, and it is possible that this lake found an outlet from the Somerset Levels through the 'Chard Gap' to the Axe valley (Stephens, 1970). Acheulian implements from the Axe valley gravels suggest an age not younger than Wolstonian (Table 12). Additionally, the ice may have advanced across the Levels and released meltwater through the gap; the 5 to 10 per cent of 'foreign' pebbles in the gravels may, however, have come almost entirely from north Devon and Somerset (J. F. N. Green, 1947), and C. P. Green (1974) has attributed them to local plateau gravels of Tertiary age. Evidence from the east of the present district (Edmonds, 1972) points to a recession of the ice, with deposition of boulder clay, followed by readvance to Barnstaple with a temporary stand around Fremington; the 2nd Terrace, now seen at and west of Braunton, dates from around this time. Final retreat of the ice may have both released large volumes of water and inaugurated isostatic readjustment of the land mass. Boulder clay was deposited, mainly free from stones within the present district but known to have yielded a few medium-sized erratics farther east (Taylor, 1956). It is generally calcareous and contains shell fragments and bits of lignite. This material now extends as a broadening crop from just north of Instow to the edge of the district at Yelland, bordering the alluvial flat of the River Taw. It was first and rather tentatively identified as boulder clay by Maw (1864).

The possibly Cromerian wave-cut platform commonly developed along the coast lies at 1.5 to 7.5 m above highwater-mark. Of widespread occurrence upon it are well-developed raised beach deposits (Plate 19) which are probably Ipswichian in age (Bowen, 1969; Edmonds, 1972). Sedgwick and Murchison first recognised the raised beach between Baggy Point and Saunton. Their conclusions, read to the Geological Society of London in 1836 but not published until 1843, were confirmed by Williams (1837) and De la Beche (1839).

The deposit consists mainly of sands with pebble layers and some shingle. Shell fragments within the sands have provided the calcium carbonate cement which has locally converted the lower part of this drift into hard calcareous sandrock. Stratification generally appears horizontal (Figure 21), with false bedding only locally developed, and the fauna indicates warm or temperate conditions of deposition. Bate (1866) considered that even the horizontal layers of the deposit were built of finely false-bedded strata, and that the whole comprised blown sand, a conclusion which reflects the search for a single mechanism of origin to explain the whole sequence of coastal drifts below the Head. The very existence of a raised beach was again questioned by Hughes (1887), who maintained that although these coastal deposits between Saunton Down and Croyde were marine beach sediments, nevertheless they lay within reach of present-day waves and did not reflect a higher sea level. However his views received little support, then or later.

Silts of the 1st Terrace in the neighbourhood of Braunton, pebbly clay and sand bordering the estuary at Home Farm Marsh [SS 494 333], opposite Chivenor Airfield, and the raised beach of the coast appear to be equivalents of Ipswichian age (Edmonds, 1972). Mitchell (1960) and Stephens (1961, 1966) followed Maw (1864) in believing the pebbly drift to be the raised beach, but the deposits are dissimilar, the pebbly assemblage being closely akin to the stonier parts of the nearby boulder clay. Field relationships suggest that the pebbly drift which now rests on a shore platform alongside the estuary has been derived from the boulder clay. Some sorting has occurred and the deposit reflects both fluvial and solifluction processes and has not travelled far. The implications are that glacial drift became reworked as it moved as a sludge or was washed downhill, that much of the finer mud of the matrix was gently washed out and that the deposit accumulated on and around an old platform in slow-moving estuary waters which planed it off at the then alluvial (now 1st Terrace) level.

Overlying the undoubted raised beach is a variable thickness of sands. Pengelly (1867) referred to differing interpretations of their origin; he considered them to be mainly subaqueous, although acknowledging that blown sand might overlie the marine material in places. Dewey (1910, 1913) agreed that the sands were water lain. The possible presence of old blown sand was considered by Arkell (1943, 1945), who concluded that such material accumulated on the beach deposits during the Riss-Wiirm (Ipswichian) Interglacial. Stephens (1966) was in no doubt that fossil dune sands were present. It seems likely that marine shingle is overlain by beach sands which pass upwards in many places into old blown sand containing terrestrial shells. Maw (1864) observed concretion of the Recent blown sand of Braunton Burrows (Figure 22) and considered some of the material so formed to be as hard as the sandrock of the raised beach.

Capping the raised beach sands and old blown sand of the coast is a variable thickness of Head–silt, sand and clay with many angular stones of local origin. Arkell discussed the dating of the Head first (1943) as Riss (Wolstonian) and later (1945) as Warm (Devensian), and Stephens (1966, 1970) believed that Head of both these periods was present. No doubt Head deposits formed in early cold periods, and locally the material is roughly layered. Perhaps some of the Head is pre-Devensian, but there is no clear evidence of this. A mantle of weathered rock which has not undergone much lateral movement has accumulated in late Pleistocene and Recent times and is quite indistinguishable from Head in the field.

Small-scale disruption of superficial deposits by frost heaving probably dates from Devensian times.

McFarlane (1955) carried out seismic refraction surveys which suggested that the buried channel of the Taw–Torridge river system is graded to a sea level about 46 m lower than the present one. It is not possible to assign the cutting of this channel to a particular phase of the Pleistocene, and Kidson (1971) envisaged development throughout the period. Probably erosion during low sea-level (glacial) times was followed by deposition during the succeeding interglacial and removal of these deposits during the next glacial phase. McFarlane (1955) noted that the suggested lower sea level implied a shoreline about 13 km out into the present Bideford Bay.

Finally it is necessary to mention high-level erosion surfaces devoid of sediments. Balchin's (1952) Georgeham Surface ranges from about 130 m OD down to 90 m; Arber's (1960) Hele Surface ranges from about 46 to 30 m and her Croyde Surface lies at about 15 m. The surfaces are not well defined in the present district. Zeuner (1959) envisaged progressively lower sea levels in succeeding interglacials and suggested that a Lower Pleistocene (or Calabrian) sea level of 180 to 210 m had fallen to 80 to 85 m by Cromerian times. Arber favoured an Eemian (Ipswichian) age for her surfaces. But there are no adequate grounds for firm conclusions as to the age or origin of these features, and some are locally so vague as to render their very presence questionable. EAE

Recent

Narrow alluvial flats in the Hartland area (Figure 23) border streams which are actively cutting down and which commonly emerge on the coast in hanging valleys. Downcutting of the Abbey River has been accentuated by artificial straightening of the channel downstream of Hartland Abbey [SS 2450 2490]. The alluvium rarely exceeds 1 m in thickness except at the coast. In the higher reaches of the streams it is usually stony clay or silt with fine sand, locally with gravel at the base; the sandstone fragments are generally angular and the deposit differs little in appearance from Head. In the lower reaches the alluvium consists mainly of subrounded to rounded gravel with subangular to angular cobbles and blocks in a matrix of sand and silt. RTT

Wide spreads of alluvium border the River Taw upstream of its junction with the Torridge, and narrow alluvial strips extend up the small tributary valleys which feed these rivers. Extensive flats of estuarine silt, clay and sand are exposed at low tide, and Braunton Marsh is composed of similar silts and clays. Estuarine alluvium passes imperceptibly into sands of the modern beach. On-shore winds have swept the beach sands of Woolacombe, Croyde, Saunton, Instow and Westward Ho! inland to build areas of dunes, the most extensive of which is Braunton Burrows (Figure 24), where the sands rise to almost 40 m above sea level.

The Pebble Ridge which extends for about 3 km flanking Northam Burrows is a storm beach gravel deposit (Plate 20). The pebbles are mainly of sandstone and range in size up to small boulders 0.3 m or more across; they are worked and reworked by wave action at high water. Originally a pebble spit built out northwards from Westward Ho!, the ridge has been rolled inland. It now rises 2 m or more above the level of the estuarine alluvium, up to 6 m above the beach, encroaches on the sands of Northam Burrows and is occasionally breached by the sea. Rogers (1908) calculated, from a study of old maps, that the southern end of the ridge had moved about 180 m eastwards between 1863 and 1903, since when it had been fairly stationary. Sandymere [SS 439 308], now less than 2 ha in extent, was a pool of scarcely 1 ha in 1884, but invasion by the sea when the migrating ridge reached it extended the mere to form a lake of over 5 ha.

Fragments of peat are occasionally thrown up on the pebble ridge by waves, and lumps of blue clay containing rootlets have been washed up on the beaches bordering the estuary. Both materials are probably derived from the submerged forest off Westward Ho!, periodically exposed at low tide. The clay and peat there contain fallen branches, and some tree stumps in the position of growth. Oak, birch, hazel, alder and elder have been identified, together with bramble, dogwood and various herbaceous plants. Bones of red deer, ox, boar, wolf, horse, sheep, goat, pig, dog and man have been found, as have flint artefacts of Mesolithic age. The forest was submerged by the rising seas of post-glacial times and the peat is the uppermost of the materials which accumulated during the Flandrian transgression (Godwin, 1956). A sample of the peat was dated as 6585 ± 130 B.P. (early Atlantic) (Churchill and Wymer, 1965); more recent datings (p. 111) have given younger ages.

The well-known flint artefacts and flakes of the Baggy Point promontory are of Mesolithic type, shaped in post-glacial times. Their abundance suggests that flints were brought there for dressing, possibly from Orleigh Court (p. 101), the-nearest natural occurrence. EAE

Details

Glacial deposits

Near Baggy House a boulder [SS 4279 4001] of garnetiferous hypersthene-bearing gneissose granulite rests on Pilton Shales and is overlain by sands. Hughes (1887) described it as yellowish white gneissose granite and measured the exposed part as 8 x 6 x 6 ft (2.4 x 1.8 x 1.8 m). Subsequent weathering of the cliff allowed Taylor (1956) to assess the total size as 13 x 7 x 7 ft (4 x 2 x 2 m), indicating that this erratic is the largest of the north Devon coast. An estimated weight approaching 50 tonnes, based on specific gravity of 2.7, rivals that of the Giant's Rock at Porthleven. The block is now less fully exposed. Taylor's (1956) microscopic examination showed the rock to be much decomposed and of granulose structure. Quartz and feldspar, mainly oligoclase with fine lamellar twinning, form a mosaic and show strain shadows. The feldspar is generally kaolinised and stained with iron oxide, the quartz commonly cracked with a few inclusions. Greenish brown and brown biotite are present, together with a little hornblende, a few crystals of hypersthene, scattered aggregates and grains of garnet, prisms of zircon and some magnetite. The rock has not been matched exactly, but a Scottish source seems possible.

A fragment of dolerite, and another of agglomerate, were noted nearby. These smaller erratics lie within the intertidal zone, as does a granitic fragment at the south end of Croyde Bay; all are liable to movement. A larger boulder, lying on the foreshore among the folds of Pilton Shales off Down End, the promontory at the south end of Croyde Bay, was first noted by Dewey (1910). He described it as a foliated granite, of quartz, microcline, perthite, green mica and biotite, with secondary chlorite and accessory zircon and apatite, and considered it similar to certain gneisses of western Scotland. Several small granitic fragments were seen near Down End.

The famous Saunton gneissose granite erratic [SS 4401 3787] lies on the raised shore platform cut in Pilton Shales and is partly buried by raised beach deposits. It has been variously mentioned and described by Williams (1837), Bate (1866), Pengelly (1867, 1873), Hall (1879b), Hughes (1887), Hamling and Rogers (1910) and Dewey (1910, 1913). Lee (1881) sketched the boulder in position in 1871. The Saunton erratic was buried by a cliff fall towards the end of 1950 and re-exposed by digging some five or six years later, an exercise supported by the Devonshire Association and the Nature Conservancy. It is about 2 m across as at present visible and has been estimated to be up to 12 tonnes in weight. The rock is a gneissose or banded granite (Dewey, 1910), consisting of quartz, microcline, oligoclase, green and white mica, and accessory apatite, epidote and sphene. Some mortar structure is present, with small mineral grains packing the spaces between larger crystals, and it seems likely that this is of mechanical origin. Dewey (1910) considered that the erratic might have come from outcrops of similar rock in Gruinard Bay, Wester Ross.

Hughes (1887) described a 'grey porphyry' from near the Saunton erratic [SS 4388 3785] as 4 ft (1.2 m) across and resembling material from Arenig, although possibly derived from the 'felsite' at the base of the Pickwell Down Sandstones. Taylor (1956) noted that this erratic was subject to movement by waves and had been reduced in size. He identified the rock as an altered trachyte or trachyandesite, with phenocrysts of oligoclase, showing alteration at the edges and some exhibiting strain shadows, set in a fine matrix of plagioclase laths. Limonite staining was common. Amygdales housed some lamellar calcite, and green patches at their margins and elsewhere were probably secondary chlorite after original ferromagnesian mineral. No quartz was visible.

Small fragments of granitic rocks are sparsely scattered on the beaches and along the estuary.

Chunks of tenacious, sticky bluish grey clay occur on the shores of the estuary near Instow, as immediately west of the cricket ground [SS 473 314]. Similar material found on the seaward side of the nearby refuse tip below the tipped debris may be in situ [SS 4739 3147]. This may be the westernmost occurrence of the commonly calcareous boulder clay first so described from near Fremington by Maw (1864). It passes east beneath alluvium for 450 m, and emerges as an eastward-widening outcrop bounded roughly on its northern side by the A39 road.

A ditch [SS 4871 3142] near West Yelland revealed sticky silty clay with a few angular fragments of sandstone and pebbles of sandstone and quartzite. Immediately south of the farm a borehole [SS 4881 3145] proved soil on stony clay 2.35 m, unsorted stones in a clayey sandy matrix 1.05 m, sandy silty clay with rock fragments of gravel grade and some larger 4.8 m, on sandy silty clay with stones of fine gravel grade 1.8 m. An immediately adjacent borehole, sunk to 10.05 m, indicated the presence of scattered larger fragments among the lowest material. Some 660 m to the east-north-east a borehole [SS 4946 3173] near Yelland Manor passed through soil 0.5 m, into sandy and silty clay with some gravel-size stones and scattered cobbles 5.5 m. The material was virtually unsorted.

Farther east-north-east tenacious brown and grey silty clay exposed in a bank [SS 4983 3190] contains no stones, and trenches [SS 4982 3206] showed 1 m of sticky red, blue and grey clay, locally silty and apparently without any Head cover. The boulder clay hereabouts crops out over a width of 500 m or more. It is bounded by Crackington Formation rocks to the south and passes beneath alluvium to the north. The scanty information available suggests that the base of the formation may fall northwards at about 1 in 5 from the southern margin.

Mitchell's (1968) 'glacial gravel' on Lundy is a scatter of foreign pebbles, rarely more than 0.1 m in diameter, on the west side of the north end of the island between 84 and 107 m above sea level. He considered that out of 100 pebbles examined, 27 were certainly not of island origin; these were of flint, chert and various types of sandstone, three were unidentified, and there seems little doubt that all were emplaced by ice or melt water. Mitchell envisaged Wolstonian ice moving south along the west coast of Lundy and being deflected eastwards up on to the plateau by projecting ridges of granite. At the top of the coastal slope [SS 1328 4675] just north of the ridge which terminates in St James's Stone, and in a similar position [SS 1313 4715] with relation to a ridge farther north, he identified smooth granite outcrops as roches moutonnées. And he related the easterly flowing melt water to the small valley system which converges on Gannets' Bay. These last two propositions remain a visionary if plausible story woven round the tangible presence of erratics.

Pebbles (MR32866, (E47388), (E47389), (E47390), (E47391), (E47392), (E47393), (E47394)) collected by Dr Hawkes around [SS 1315 4752], 0.6 km at 170° from North Lighthouse and 87 m above present sea level, included grey quartzose sandstones, greywacke-type sandstones, hematite-stained sandstones, vein quartz, and single specimens of granophyric microgranite and chert. The chert pebble is very angular, the others being well-rounded. Mr G. E. Strong reports that the hematite-stained sandstones may be matched from the Old Red Sandstone of South Wales, or the Devonian or Permo-Triassic sandstones of north Devon and Somerset, and sandstones similar to the greywackes occur in central Wales, north Devon and Ireland. EAE

River terraces

At Blackpool Mill [SS 2260 2565] the Abbey River is incised 2.5 m into bedrock about 4.5 m below the level of a large abandoned meander (Plate 21). A ditch [SS 2370 2488] alongside the same river contains pebbles in a silty matrix about 4 m above the alluvium and 100 m downstream from a probable knick point. Benches without terrace deposits occur at [SS 2300 2515] and [SS 2265 2395]. RTT

Small silty flats of the 1st Terrace border the narrow alluvium of two stream valleys which converge on Lee [SS 483 463].

Deposits of the 2nd Terrace, probably of late Wolstonian age (Edmonds, 1972), extend from near Saunton eastwards to Braunton on a bench, generally about 150 m wide, which runs roughly parallel to, and just north of, the B3231 road. There are no sections through the terrace but the surface material indicates brown silt and clayey silt. The back of the feature stands at 23 to 25 m above sea level. East of Braunton the corresponding terrace narrows eastwards and rises to 30 m above sea level.

The B3231 road from Saunton to Braunton runs mainly on the 1st Terrace, which is commonly about 200 to 250 m wide in the west but broadens to underlie Braunton Great Field, the southeastern part of the town between the alluvial strips of the River Caen and Knowl Water, and Wrafton. The brown sand and silt of Braunton Great Field affords free-draining first-class arable land. A shallow borehole [SS 4843 3616] near the eastern edge of the field passed through 3 m of firm brown sandy clay with some gravel and scattered pieces of sandstone. Another nearby [SS 4845 3617] proved firm brown sandy clay with some gravel and scattered cobbles 1.5 m, crumbly mottled brown clayey sand with some gravel 5.8 m, dense brown silty clayey sand with some fine gravel 0.3 m, dense sandy gravel with traces of clay 0.75 m, on compact brown fine- to medium-grained sand with some fine- to medium-grained gravel 2.4 m. The back of this 1st Terrace stands variously at heights of 15 to 20 m above sea level. EAE

At Edge Mill [SS 4486 2297], 0.8 km SE of Littleham, the base of a small strip of 1st Terrace is about 3.35 m above the alluvium; the deposit is rather loamy, with small pebbles and subangular fragments of sandstone, shale and quartz, and rare granite fragments; large irregular brown-coated sandstone fragments up to 0.13 m across are common.

A patch of gravel [SS 460 230] bordering the alluvium on the west bank of the river 0.5 km NE of Annery House reaches a height of 4.8 m above the flood plain.

East of Pillmouth gravels of the 1st Terrace extend along the western flank of the alluvium, rising to about 4.5 m above the flood plain. They are exposed in the river bank [SS 4705 2396] and consist of subangular to well-worn Carboniferous sandstone and shale with quartz in a matrix of pale brown clayey coarse sand. A few subangular sandstone fragments up to 0.25 m across, and one well-worn granite pebble 0.1 m in diameter, were found. Sporadic impersistent intercalations of sand and loam occur. The upper part of the terrace is apparently obscured by Head which has flowed down over it.

Raised beach and associated deposits

At the landward end [SS 4540 4422] of Barricane Beach, Woolacombe, 0.75 m of shelly sand with quartz pebbles and a few small granitic fragments lie at about 11 m OD and are overlain by 3 m of stony Head. A well-developed shelf at about 13 m OD at the northern end of Woolacombe Sands carries no beach deposit but is probably part of the raised shore platform [SS 455 439]. To the south, the lower slopes of Potter's Hill [SS 459 430] carry a good deal of sandstone brash, but Ussher (MS notes dated 1878) noticed buff consolidated sandstone exposed on the flanks of the hill which he identified as old blown sand. This outcrop cannot have lain at much less than 40 m OD.

The southern end of Woolacombe Sands abuts against Napps Cliff, but the lower coastal stretch west and north-west of Puts-borough Sands Hotel is vaguely bevelled at about 3 m above high water. Scattered pebbles, mainly of local sandstone and vein quartz but including a few bits of flint, suggest the remnants of a raised beach.

It is between Baggy Point and Saunton Sands that the most striking developments of raised beach occur. An elevated shore platform cut as an irregular surface on steeply dipping Devonian rocks is overlain by calcareous pebbly sand and shelly calcareous sandrock which has yielded a warm/temperate marine fauna (Prestwich, 1892). In places this sandrock passes upwards into false-bedded sands containing terrestrial snails such as abound in Braunton Burrows; these are probably old dune sands and locally they extend to a considerable height above the platform. Sedgwick and Murchison (1843) noted that the raised beach and overlying sands rose to 18 to 21 m above the sea near Baggy Point.

In the neighbourhood of Pencil Rock [SS 422 403] the platform, at 14 m OD, carries over 4 m of shingle and sandrock. To the east [SS 4231 4024]; [SS 4251 4021] pebbly raised beach deposits are overlain by probable old blown sand. Near Baggy House [SS 4289 3998] the platform is at about 10 m OD and bears pebbly shelly sand in horizontal layers (raised beach) 1.5 m, overlain by brown calcareous sand and sandrock (?old blown sand) 4 m, and stony Head.

At Middleborough the platform is at about 13 m OD. Sandy shingle of the raised beach is overlain by false-bedded sands and appears locally [SS 4307 3972] to rest on thin stony clayey drift whose passing resemblance to an old Head must be disregarded because of the near impossibility of such a deposit surviving in this position. Stephens (1966) thought that two or possibly three ages of Head were present here above the raised beach, and he identified a clayey pebbly drift resting on the raised beach as weathered till; the evidence for these conclusions is slender. Sedgwick and Murchison (1843) noted shelly shingle near the limekilns [SS 4341 3954] at the north end of Croyde Bay.

Inland of the modern blown sand that forms Croyde Burrows, flats to the north and south of the stream, at about 12 m OD, are regarded as raised beach rather than 1st Terrace, largely because of their coastal disposition. Pebbly sand, overlain by 1.5 m of stony Head, was noted beneath blown sand [SS 4364 3921] at about 8 to 9 m OD, and traces of compact false-bedded sand and ?Head beneath blown sand to the south [SS 435 390]. Ussher (MS notes dated 1878) observed rounded fragments in roadside drift on the western outskirts of Croyde and on the south side of the stream [SS 4477 3918]. Two boreholes [about 4394 3882] at the NALGO Holiday Centre are recorded as having passed through 14.5 and 18.3 m of drift, meeting sand with pebbles at depths of 9.75 and 12.7 m respectively. Presumably these sequences represent Head on raised beach deposits; the latter are 4.75 and 5.6 m thick respectively.

On the south side of the Saunton Down promontory the platform ranges in height from 10 m OD at its western end to 8 m farther east. Shelly sands and sandrock, with pebbles scattered and in layers, lie on the platform and range up to 9 m in thickness. At this or lower levels they pass upwards locally into false-bedded sands, which contain scattered land-snail shells and are probably old blown sand. These in turn are capped by Head, randomly mixed silt, sand and clay with angular stones mainly of local origin. Near the western end of the section [SS 434 381] horizontally bedded hard pebbly sandrock 3 m, is overlain by pale brown sand whose bedding is locally false and contorted 15 m, capped by Head 3 m. Farther east horizontally bedded beach deposits and overlying ?old blown sand total 15 m in thickness [SS 435 380], and raised beach sands 4.5 m thick contain three layers of pebbles and lie beneath 15 m of Head. Cliffs bordering the northern end of Saunton Sands show, in their western part, about 4.5 m of pale brown coarse friable sandrock, locally false bedded and with pebbles near the base. Farther east stratified pebbly sands of the raised beach are up to 9 m thick, locally false bedded, and overlain by Head. The section is obscured by modern blown sand south of Saunton Sands Hotel.

Mr G. E. Strong and Mr R. J. Merriman report that a specimen ((MR32610); (E46486)) from the raised beach near Saunton Sands [SS 4416 3781] is a hard calcareous sandy conglomerate, with pebbles of coarse siltstone and fine-grained sandstone set in yellowish grey cemented sand. The basal part of the specimen is finely laminated calcareous sandstone, with pebbles showing imbrication. In the upper part subangular to subrounded pebbles up to 50 mm across constitute about 50 per cent of the rock. The sandstone pebbles are very fine grained, and show fine parallel lamination; the siltstones are sheared, and contain oxidised iron-stained chlorite. There are also small ferruginous pebbles. The matrix consists mainly of subangular to subrounded quartz, fragments of schist, chert, and microgranite, and some oligoclaseandesine and glauconite (X 7215). The heavy mineral suite includes yellowish brown and bottle-green tourmaline, zircon, biotite, epidote, clinozoisite, garnet, amphibole, pyroxene and opaque iron oxide. The cement is sparry calcite, which forms a network supporting individual grains and constitutes about 25 per cent of the rock.

A borehole on Braunton Burrows (p. 113) may have penetrated the raised beach. EAE

The low cliff [SS 4568 3069] on the west side of the Appledore peninsula is capped by brown fine-grained sand with subrounded Carboniferous sandstone pebbles and scattered quartz pebbles. One flint pebble was found. The base of the deposit is 2.44 m above high water.

At Westward Ho! [SS 4324 2932] an exposure seen during work on the sea wall revealed 1.8 m of brown loamy Head resting on 2.13 m of raised beach composed of ellipsoidal sandstone pebbles up to 0.2 m across closely packed in brown sand. The base of the deposit is 3.35 m above high water. Nearby [SS 4316 2926] the deposit is 3.66 m thick and contains sandstone cobbles up to 0.42 m across. Farther west [SS 4306 2922]–[SS 4300 2921] the raised beach contains a high proportion of sand, and is overlain by 1 m of brown fine-grained sand capped by 2.2 m of sandy Head.

A good section through the raised beach [SS 4228 2911]–[SS 4210 2907] is as follows:

The base of the deposit is about 4 m above high water. Scattered exposures of similar raised beach may be seen as far south-west as Cornborough Cliff [SS 4124 2816]. BJW

Horizontal jointing in the Lundy Granite gives a false impression of a multiplicity of shore platforms, and stream profiles are inadequate for a study of grading levels. However, some evidence exists in the form of a sparse scatter of pebbles on some ledges, although many of the smaller ones could have been thrown to considerable heights by waves and some others presumably have been dislodged from higher positions. Queen Mab's Grotto [SS 1381 4691] has the appearance of a sea cave and now stands 6 or 7 m above sea level, perhaps related to a nearby platform [SS 1383 4694] of similar height (Figure 12). EAE

Head

In the northern, Morte Slates, country jagged outcrops of slate and vein quartz stand out from a cover up to 3 m thick of clay and silty clay carrying slate, silty slate and quartz fragments. The stony Head noted above the raised beach at Barricane Beach [SS 4540 4422] (p. 110) shows remnants of sandy layers which may represent blown sand deposited during formation of the Head. The more rolling country of the strong Pickwell Down Sandstones carries an abundance of sandstone rubble on the uplands, suggesting that the Head cover, mainly of ferruginous stony sand, is thin, probably rarely exceeding 1.5 m. On the slopes, however, such materials have accumulated in slightly greater thicknesses. A small gorge [SS 453 415] at the coast near Vention reveals 3 m of Head, consisting of angular sandstone fragments in silty clay. A well about [SS 4855 4155] at Spreacombe passed through 3.4 m or possibly 5.5 m of drift, and another [SS 4963 3987] to the south-east proved 5.5 m of clay.

On the south side of the Baggy Point promontory stony clayey Head is exposed up to 18 m thick. It has accumulated on the coastal bevel in a much greater thickness than is generally likely to be found inland. However, a borehole [SS 440 396] at Croyde Bay passed through 10 m of clay and stones on 2 m of fine red sand before reaching rock, and two boreholes at the NALGO Holiday Centre about [SS 4394 3882] respectively proved 9.75 and 12.7 m of clay and stones with some sand. Near the mouth of the Croyde stream [SS 4364 3921] 1.5 m of silty clay packed with angular fragments of slate and sandstone and some quartz separate blown sand above from pebbly sand (?raised beach remnants) below. On the south side of Saunton Down the cliffs, especially at their eastern (Saunton Sands) end show up to 15 m of stony Head; again this thickness exceeds that likely to be present inland.

On the south side of the Taw–Torridge estuary, opposite Chivenor Airfield, a slight rise between Home Farm Marsh [SS 494 333] and the river is capped by pebbly clay and sand. Traces only are exposed along the river bank, but the surface of the fields is brown silty clay with fragments of fine-grained sandstone and pebbles up to 0.13 m across of sandstone, quartzitic sandstone and quartz.

South of the drifts bordering the estuary, on the Crackington Formation rocks between Instow and Westleigh, a widespread Head cover of yellowish brown silty clay with angular sandstone fragments generally does not exceed 3 m in thickness. EAE

Around Hartland, Head is commonly over 1 m thick and ranges up to 5 m in the banks of incised streams. The weathered mantle on steep valley slopes has been affected by hill wash, creep and slip and locally overlies alluvium. Sections 6 m high in the banks of the Abbey River [SS 2340 2495] show the remains of a soil profile between alluvium and weathered cover [SS 2292 2527]. Similar relationships are seen elsewhere [SS 2243 2638]; [SS 2258 2574]. Hereabouts the Head typically consists of yellowish buff silty clay with sandstone fragments. In waterlogged ground the iron content is usually in a reduced state and the Head is grey or bluish grey while the sandstone fragments are bleached white. Where the water level oscillates an ochreous upper layer commonly passes down through a mottled grey and brown deposit into the reduced zone at a depth of 0.3 to 0.6 m. Clay content may be greater where surface water has concentrated the clay fraction in low-lying ground or where the bedrock is predominantly shaly.

A belt of red-stained Head extending from west of Hartland village to Clovelly may have been derived from reddened Carboniferous rocks overlain by an outlier of Permian rocks (pp. 66–67). Head is also stained red locally in the vicinity of faults. Exposures show 2.4 m of yellow silty clayey Head [SS 2343 2675], 1 to 2 m of red silty Head [SS 2615 2325], 3 m of brown silty Head [SS 2642 2544] and 1.5 m of cliff-top silty Head [SS 2943 2638], all containing much sandstone debris.

Bluish black hard pan, known locally as 'black ram', is developed in some areas of poorly drained ground. It is produced by the cementation of weathered material by iron and manganese, and may form a layer up to 0.3 m thick at shallow depth. Large masses in excess of 0.5 m across have been unearthed during the reclamation of wet ground. Hard pan has been found in field drains and at surface [SS 2274 2460]; [SS 2476 2586]; [SS 2539 2585]; [SS 2721 2547]. RTT

Lundy Island is largely free from Head, except for local accumulations on the coastal bevel, between the central plateau and the cliff edge. A typical example is seen just south of the Mousehole and Trap, where 6 m of coarse rubbly bouldery Head caps a low cliff [SS 1382 4688]. Thin clayey silt (brickearth) with scattered flints, small quartzitic pebbles and Mesolithic flint flakes, occurs in Brick Field [SS 137 448]. EAE

Recent deposits

Submerged forest

In July 1878, between 60 and 90 m off the south end of the Westward Ho! pebble ridge, Ussher (MS notes) noted two flat foetid spongy peaty masses projecting through the beach sands. The vegetable stratum was 0.45 m thick, a mass of matted marsh plants and reeds with scattered woody fragments and associated with brown clay, and rested on bluish grey sticky clay containing a few twigs and fragments of reed. Borings were abundant, presumably the work of Scrobicularia, whose shells were present and which lives buried in the muds of tidal estuaries.

A record in the Royal Albert Memorial Museum, Exeter, gives a thickness of 0.6 m for the peat and notes that the blue clay rests on a kitchen midden containing flints, split bones, leg bones of ox and red deer, pottery, a stone pounder, charcoal, and cockle and periwinkle shells. Churchill and Wymer (1965) noted that the peat rested on bluish grey sandy clay containing pollen which in turn was underlain by sterile blue clay yielding no organic material except a few descending roots. They considered that sea level at the time of formation of the peat was 4 to 6 m below that at present. EAE

In March 1970 severe westerly gales removed much of the recent beach material and at a place [SS 4319 2958] 160 m at 325° from Westward Ho! slipway, an expanse of peat some 36 by 20 m was revealed. The peat was grey to brown silty and sandy, with wood fragments and roots, some in the position of growth. The deposit was between 0.45 and 0.61 m thick, and in places was seen to rest on a bed of flattened cobbles up to 0.1 m across, the bed being at least 0.3 m thick. The base of the peat was 2.74 m below highwater-mark. The surface of the peat was traversed by gullies running NW–SE, and the whole mass was tunnelled by modern boring molluscs. A sample of wood gave a radiocarbon date of 4995 ± 105 years B.P., whilst a sample of the peat, which was probably contaminated, was dated as 3075 1100 years B.P. (Williams, 1972) BJW

Modern beach deposits

Little detail is available about the modern marine sediments of the coast. In the north and south, and round the headlands, boulder beaches predominate, with scattered small isolated coves floored by shelly sand. The larger bays are bounded on their landward sides by broad beach sands. Saunton Sands were found by McFarlane (1955) to have an almost uniform thickness of 9 m along a N–S seismic traverse [SS 444 377]–[SS 444 336]. Immediately south of the present mouth of the Taw-Torridge estuary, sands and clays thickened from 9.75 m [SS 443 327] to 24 m [SS 441 323] and McFarlane estimated that the maximum thickness, of about 30 m, lay a short distance farther south. His seismic profile immediately on the seaward side of the pebble ridge [SS 439 319]–[SS 433 295] suggested a fairly uniform thickness of sand and clay of around 11 m.

Alluvium

Away from the estuary shores, the only alluvial strips of significant width are those of the River Caen and the Knowl Water, in the Braunton area, the westerly flowing tributary of the River Tor-ridge north of East-the-Water and the easterly flowing tributary of that river which passes through the northern part of Bideford. These last two stretches may lie in a single valley which accommodated westerly flowing waters when the estuary was blocked by Wolstonian ice (Edmonds, 1972).

Silts and silty clays of the River Caen north of Braunton mantle a flat commonly 100 m wide, but locally up to 180 m. Similar sediments bordering the Knowl Water immediately upstream of Wrafton crop out over a width of up to 250 m. Several shallow boreholes sunk through the alluvium on the southern outskirts of Braunton revealed the presence of sand and gravel. Selected borehole logs may be summarised as follows: at [SS 4857 3624], gravel with traces of clay 1.0 m, sandy gravel 2.0 m, sand with traces of gravel below 4 m depth 1.6 m, sandy gravel with some cobbles 2.1 m, on shales 0.3 m; at [SS 4853 3619], sandy clay and clayey sand with some gravel and cobbles 3.9 m, gravel 0.3 m, sand and gravel with traces of clay 0.9 m, gravel 0.6 m, sand with traces of gravel in lower parts 2.1 m, silty clay 0.15 m, sand and gravel with traces of silty clay 1.0 m, on gravel and sand 0.45 m; at [SS 4860 3621], sandy clay 0.6 m, clayey silt with silty clay and peat 0.75 m, gravel with some silty and sandy clay 1.8 m, sand with some gravel 1.8 m, sandy gravel with cobbles 2.1 m, on shales 1.5 m; at [SS 4858 3600], clay 0.3 m, gravelly clay 0.45 m, silty sandy clay 0.15 m, sandy gravel with traces of clay 1.3 m, fine- to medium-grained sand 1.5 m, sandy gravel 2.6 m, on shales 0.3 m; and at [SS 4864 3593], sandy clay with some gravel 1.5 m, sandy gravel with traces of clay and sandstone fragments 2.7 m, sand with traces of gravel 0.6 m, sandy gravel 0.5 m, on shales 0.7 m. These records show a range of thicknesses of alluvium overlying Pilton Shales hereabouts of 5.3 m to over 9.4 m. Other shallow boreholes nearby penetrated similar alluvial sediments of thicknesses which fell within this range.

On the south side of the River Taw a wide expanse of alluvium extends from Instow to Yelland. The surface materials are silts and clays, but records of 'stones' and stony clay penetrated by a number of shallow boreholes suggest the presence of gravels. Four such boreholes at East Yelland Power Station [SS 481 324] proved 1.5 to 3.0 m of alluvium on Crackington Formation. Another [SS 4823 3215], at the nearby oil depot, pierced 6.4 m of 'clay and stones'. A sixth [SS 4939 3257], alongside the railway north of Lower Yelland, passed through soil 0.3 m, stiff sandy and silty clay with fine and coarse gravel 2.9 m, on shale 6.0 m. LAE

In the river bank [SS 4728 2988] 200 m at 185° from Instow Station a section in the alluvium is revealed, its base being about 1 m above high-water-mark:

BJW

A borehole [SS 468 289] sited near the edge of a small patch of alluvium 400 m WNW of Westleigh passed through 6.1 m of silt and clay before entering Crackington Formation. EAE

At the mouth of the Titchberry Water [SS 2287 2668] about 1.5 m of alluvium in the cliff comprise locally derived, poorly sorted, brown sand and silt with gravel and pebbles and with angular sandstone blocks up to 0.25 m across. Patchy dark brown or purplish black manganiferous cement occurs in the upper part. Tabular pebbles are generally flat-lying with some tendency to an imbricate structure. Farther upstream [SS 2322 2676] fairly well-rounded pebbles and sand 0.56 m, are overlain by buff to grey clayey silt 0.75 m, and brown silty clay with angular sandstone fragments 0.3 to 0.6 m. A valley inland of Smoothlands [SS 225 264], now occupied by a tiny stream, contains alluvium and old meanders of a former westward extension of the Titchberry Water. The alluvium is overlain at its margins by roughly stratified angular detritus [SS 2243 2635] which has been cut into by the former stream.

Cascading waters at the mouth of the Abbey River have cut down below 4.6 m of poorly sorted alluvial clayey silt, sand and gravel with angular blocks up to 0.6 m across. Most of the pebbles are less than 0.1 m in diameter but larger than those at the mouth of the Titchberry Water, reflecting the greater carrying power of the larger stream. All the fragments are of locally derived sandstone with some vein-quartz. On the northern side of the valley the alluvium is overlain by weathered detritus into which the stream has cut.

Downstream of Hartland Abbey alluvium about 0.9 m thick contains rounded to subrounded pebbles in a matrix of silt and fairly well-rounded gravel. A section [SS 2340 2495] shows 0.61 m of orange-brown silt and gravel containing rounded to angular pebbles, cobbles and blocks, overlain by 0.15 m of similar but grey-coloured material which may represent an old soil horizon and is overlain by some 6 m of Head.

Farther upstream the river shows the following section [SS 2591 2484] in alluvium: fairly well-rounded gravel with some angular to subangular blocks of sandstone 0.23 m, overlain by orange and grey mottled silt with mainly angular sandstone fragments up to 0.15 m across 0.61 m, and brown slightly clayey silt with sandstone fragments and with some black manganese spotting at its base 0.46 m. RTT

Estuarine alluvium

Apart from the muds, clays, silts, sands and stones of the estuary shores, which are constantly reworked by river and tidal waters and therefore present no enduring mappable lines between one lithology and another, the main areas of estuarine alluvium are Braunton Marsh, the lowest reaches of the River Caen, south of Braunton, and the low ground immediately south of Northam Burrows. In all cases the surface deposits are silts and clays.

Two boreholes at the pumping station [SS 4649 3748], Saunton, at about 12 m OD, proved over 8 m of drift, probably wholly estuarine alluvium although perhaps with raised beach material at the base, on Pilton Shales. One passed through soil 0.6 m, sand 1.8 m, pebbles 0.6 m, clay 3.0 m, sand 1.5 m, stones and pebbles 0.6 m, on clay (probably weathered shales). The other proved clay 1.8 m, rock fragments and clay 3.7 m, sandy clay and gravel 2.1 m, sand 0.75 m, on weathered rock and shales. A third borehole, whose position is uncertain but which may have been sunk some 2 km to the south about [SS 470 356] and at 4.6 m OD, passed through 11.7 m of sand, clay and gravel before entering shale. EAE

Blown Sand

Woolacombe Warren stretches from Woolacombe to Vention as a strip of sand dunes about 3 km long and up to 200 m wide. In the north the dunes rise to over 30 m OD, higher than the valley flats to the east between Woolacombe and Potter's Hill. Throughout most of the length of the Warren the strong Pickwell Down Sandstones rise steeply eastwards from the coast, and locally the blown sand has accumulated against this bank up to 45 m OD.

The hillocks of Croyde Burrows rise to over 20 m. They back against a raised beach flat and low-lying ground of Pilton Shales between the coast and Croyde, and their consequent tendency to extend eastwards is countered by planting with marram grass, and stabilising with wire mesh and brushwood.

The topography of Braunton Burrows comprises a number of N–S features, alternate belts of dunes and slackozvgns, shaped by the prevailing on-shore winds (Figure 24). Adjoining the beach sands is a narrow zone of low foredunes up to 5 m in height. This is succeeded by a broad zone of sand hills rising to 15 m OD and separated by a discontinuous belt of slacks from a second ridge of dunes. These slacks are mainly dry but contain temporary freshwater lakes and pools; patchy thyme sward is characteristic. The succeeding dunes contain the highest peaks of the Burrows (38 m OD). Beyond these high dunes is a broad, rather poorly defined belt of lower ground which contains scattered isolated hillocks, many temporary fresh-water pools and lakes and a few permanent ones. Landwards again lies a zone of lesser dunes, followed by flat sands with a few scattered small dunes which merge into the cultivated fields of Braunton Marsh. Oatson (1918) noted that scattered plants of Salsola (saltwort) occurred on the seaward side of the Burrows, near high-water-mark. The foredunes and higher dune ridges farther inland are colonised by marram grass (Ammophila), with a variable amount of creeping willow on the landward sides. The discontinuous slacks are characterised by the liverwort Riccia, but locally both these flats and the broader belt inland of the highest dunes are richly and variously vegetated; wet slacks commonly have the fresh-water table near ground level and carry rushes, mosses, creeping willow, alder and various other scrubby trees of stunted growth.

A borehole [SS 4525 3424] at about 16 m OD, within the high dune belt, proved grey fine-grained sand with silty patches and consolidated layers 7.3 m, soft blue silty clay 2.5 m, sand with coarse and fine gravel 0.5 m, coarse, medium and fine sand with fine-grained gravel and marine shells 0.6 m, on hard blue rock 0.6 m. It is possible that the basal drift layer is raised beach. Towards the southern end of Braunton Burrows examples of apparently horizontal bedding were noted [SS 460 330]. A borehole [SS 4579 3248] at the old Bideford lighthouse, at the southern end of Braunton Burrows, passed through 8.5 m of blown sand on 5.8 m of sand, clay and pebbles (estuarine alluvium), on rock.

Small areas of blown sand occur north of Instow. The dunes rise to about 5 m OD and are advancing slowly eastwards. EAE

Landslip and coast erosion

Coastal landslips are common in the Hartland area, where erosion of the cliffs ranges from the slow accumulation of scree to large-scale collapse of strata. Some landslips appear to be old and stabilised; others show evidence of recent movement. Most slips probably occur in the winter months when more rain water is available for lubrication.

The main factor facilitating landslipping throughout the area is the closely spaced jointing and bedding of the rocks, and masses of slipped rock tend to disintegrate into a debris slide. Landslips may occur where faults, usually wrench faults, intersect the cliff. The strata in the acute angle between the fault and the cliff face become detached from the body of the cliff and collapse. Penetration of surface water along the fault planes may assist the development of these slips, and springs are common at the seaward ends of faults. Examples of fault-controlled landslips occur at Hartland Quay [SS 2250 2485], Blagdon Cliff [SS 2285 2730], Shipload Bay [SS 2490 2735] and Beckland Bay [SS 2825 2675].

Toppling failures (de Freitas, 1972) occur where vertical or steeply dipping strata strike parallel to the cliff. The rock mass separates along bedding and jointing and topples progressively outwards from the cliff until a position of stability is reached in which the loosened blocks lock together. Weathering or wave action at the foot of the mass may lead to further periodic collapse. Such processes appear to account for areas of overturned strata and associated landslip at Longpeak [SS 2240 2327], near the coastguard station at Blagdon Cliff [SS 2300 2745], at Barley Bay [SS 2335 2760], Shipload Bay [SS 2475 2735] and Beckland Bay [SS 2845 2670]. Extensive landslips east of Clovelly may have resulted from the toppling of the steeply dipping or overturned northern limbs of folds in this section. Small areas of toppled strata are found at a number of other points along the coast.

On the west coast, at Speke's Mill Beach [SS 2260 2371]; [SS 2260 2383], small landslips have occurred where the core of a syncline has separated along joints and bedding planes and collapsed seawards.

Rotational landslips seem to have occurred at Brass Pan [SS 2575 2750] on Gawlish Cliff, and Blue Mellem [SS 2645 2750] on Fatacot Cliff. They have resulted from the failure of thick shale horizons and overlying sandstones of the Bude Formation which dip at a low angle to the south into the top of the cliff. The eastern part of the Blue Mellem slip is also associated with wrench faulting. The main areas of slip are restricted to the upper part of the cliff, but beneath the most active parts the lower cliff also has collapsed and debris slides extend on to the beach. Open tension cracks and slip scars at the cliff top indicate recent movement, and appreciable changes have taken place in the cliff line since the 1:10 560 Ordnance Survey of 1904.

There is little evidence of landslipping inland. A level feature with a steep back to the south of the Abbey River [SS 2280 2530] may be an old landslip or the remains of a river terrace. Some 6 m of Head-like debris which overlie alluvium in the bank of the Abbey River [SS 2292 2527]–[SS 2340 2495] may have been emplaced by superficial landslip or soil creep. Similarly some rock outcrops on the north side of the Abbey River valley near Hartland Abbey show signs of disturbance which could reflect toppling failure or soil creep.

Coastal erosion is particularly active on the W-facing Hartland coast. A broad shore platform is being developed as a result of recession of the near-vertical, flat-topped cliff. Some attempt is being made to control the erosion of thick shales near the lighthouse.

South of Hartland Point valleys once parallel to the coast have been breached by the sea and remain as isolated sections floored by alluvium (Plate 22). Stephens (1966) suggested that these sections might represent glacial melt-water channels related to an offshore ice sheet, but Arber (1911) was probably correct in thinking them the eroded remnants of an old river system which linked the Titchberry and Milford waters and the Abbey River ((Plate 21); (Figure 23)). The highest river deposits of the Milford Water at Speke's Mill Mouth lie at 38 m OD. Those behind St Catherine's Tor are at 30.5 m and those at Hartland Quay at about 20 m. The alluvium of the Titchberry Water falls from 38 m at the present mouth to between 15 and 23 m at Damehole Point. The top of the alluvium at the mouth of the Abbey River stands at about 15 m. Breaches in the lower valley of the Titchberry Water north of Damehole Point and north of Smoothlands are on fault lines.

On the more sheltered, N-facing coast east of Hartland Point the embankments of the Iron Age encampment on Windbury Head [SS 2865 2660] have largely disappeared as a result of recession of the cliff, and no traces of raised beach occur anywhere along this coast. The hogs-back cliff profiles developed eastwards towards Clovelly reflect the instability of cliffs parallel to the strike. However, erosion is slower than on the W-facing coast. Thus the stream entering the sea at Mouthmill is graded virtually to sea level, although smaller streams such as the Beckland Water disgorge over waterfalls at the cliff edge. Stephens's (1966) postulated glacial melt-water channel north of Clovelly Court is in line with the major strike fault at Wood Rock [SS 3145 2550], and the depression in the cliff line has probably resulted from normal erosion on this line of weakness. RTT

Chapter 10 Economic geology

Mining

Several trials and prospects have left their small marks on the district but mining has never been an important industry. The metallic minerals are mainly of low-temperature origin. Lead, silver, copper and molybdenum have been reported (Figure 25), but the only recorded metalliferous mining has been for iron ore in Devonian rocks at Spreacombe. These ores appear to follow joints or faults. Culm has been dug from Carboniferous rocks, the alignment of the workings suggesting the presence of a group of seams or a single seam split in some places.

Iron

Spreacombe Mine (Dines, 1956, p. 759) produced 779 tonnes of brown hematite in 1874–76 and 1889; the 118 tonnes extracted in 1876 were valued at £70.80. An old quarry [SS 4777 4137] 700 m west of Spreacombe exposes red sandstone, locally argillaceous. Three N–S vertical fracture zones up to 1 m thick contain brecciated red and purple sandstone cemented by hematite and limonite with some manganese oxide; this constitutes the ore which has been worked in the mine. The two westernmost fractures lie close together and trend towards a 12-m-deep shaft [SS 4776 4140] immediately north of the quarry. The easternmost fracture is in line with an adit [SS 4779 4129], now filled, on the opposite (south) side of the road; this adit trended 210° for 9 m and then turned southerly. A line of six old shafts, or depressions marking collapsed shafts, trends 175° thence to [SS 4781 4104], presumably marking the line of a lode. About 270 m to the west a N–S elongated depression [SS 4753 4107] may mark a collapsed adit. One kilometre north of the quarry, and 440 m east-south-east of Roadway, two old adits [SS 4771 4239] lie just above a stream bank. The westermost trends 280° and the easternmost 360° and both are flooded. Their entrances are in rusty sandy Head, and purple ferruginous sandstone is exposed in nearby quarries; some fragments of iron-rich fault breccia were noted.

Buckland Mine (Dines, 1956, p. 759) was a trial for iron but is not recorded as ever having gone into production. There are shafts 1 km east of North Buckland at [SS 4906 4030], and thence at 220 and 250 m at 036°, presumably the trend of the lode. Fragments of ferruginous sandstone and fault breccia occur.

Ferruginous staining is common elsewhere within the Pickwell Down Sandstones. In a small quarry [SS 4566 4142] on the western slopes of Pickwell Down an iron-rich fault zone up to 1 m wide trends to 320°. About 650 m west of Buttercombe there is a small circular depression [SS 4926 4237] whose position suggests a collapsed shaft rather than a well. It lies in line with a fault running north east from Spreacombe, a trend similar to that suggested for the lode at Buckland Mine.

Silver and Lead

Prospecting for silver and lead at Ubbastone (Hubbastone), Appledore (Slader, 1965), was reported in the North Devon Journal in October 1852. Working was said to have been suspended pending the formation of a mining company, but there is no later reference to the enterprise, whose exact location is not known.

Copper

Vancouver (1808, p. 55) mentioned a trial for copper at Alwington, 6 or 7 km south-west of Bideford, but reported it abandoned when only pyrite was discovered. EAE

Chope (1902), quoting the original minutes of the formation of the Wheal Hartland Coppermine Company, dated 9th December 1830, stated that 'this abortive search for copper' was located in the second field west of Coalpit Lane, running between the Hartland–Bideford road and Highdown. A small hollow and mound [SS 2787 2453] may be the site of the working. The minutes indicate that 79 shares were taken up, out of a total of 100; the main shareholder was Mr Arscot Tremear, who took 20 shares and may well have been the promoter. A Mr Richard Chope, the treasurer, took five shares. Slader (1965), citing Chope, placed Wheal Hartland near the road to Woolfardisworthy through Kennerland Cross, but was probably mistaken.

Traces of chalcopyrite associated with ankerite have been noted in faults exposed on the Hartland coast (p. 99), and films of green copper secondary minerals occur on the joint surfaces of some shales, as on the east side of Shipload Bay [SS 2484 2748] (p. 53). Chalcopyrite has also been found with ankerite infilling the cavities in goniatites within calcareous nodules. This association indicates that some of the chal-copyrite is likely to be of diagenetic origin. RTT

Traces of chalcopyrite on Lundy were prospected in the 19th century. Two adits driven one above the other into the granite cliff near Long Ruse [SS 1306 4776] were a trial for copper. Another horizontal adit [SS 138 437], near the granite-slate contact in the cliff below Benjamin's Chair, represents a second unsuccessful search for economic copper deposits. A little copper and some tin have been recorded from quartz veins cutting slates on Lundy.

Molybdenum

Traces of molybdenum have been noted in the granite cliffs [SS 1285 4510] of the bay north of Battery Point on Lundy. The possibility of extracting the metal was considered during the Second World War but was rejected as uneconomic. Plates of molybdenite occur locally within late-stage pegmatitic intrusions. EAE

Coal ('Culm') and pigment

Carbonaceous material within Upper Carboniferous rocks has been worked sporadically along a 20-km line trending just south of east from Greencliff to Hawkridge Wood west of Chittlehampton. Traces of old pits are common, many having been sunk for individual fuel supplies. It is probable that culm was dug for burning during the Middle Ages.

Polwhele (1797) mentioned an 'excellent coal-mine' at Tawstock, in the Barnstaple district. This mine had been the subject of a newspaper report in 1790, when it was described as being reopened, the coal being 'extremely good', burning 'remarkably clear' and being 'fit for the purpose of burning lime etc.'. Canal transport to Exeter was envisaged. Some 60 tonnes of coal a week were produced but the workings, which reached 46 m depth, were closed in 1800 because of flooding. Polwhele (1797) also mentioned culm workings at Westleigh and Bideford, within the present district. Lysons and Lysons (1822) described anthracite occurring on the coast near Bideford, and at Abbotsham, and mentioned its use as a pigment in Plymouth Dockyard.

Coal dug near Greencliff was used on the spot to burn limestone from South Wales. Early 19th-century trials for coal were made at Abbotsham and Westleigh, and Strong (1889) recorded that small quantities of coal were said to have been found at the former place. De la Beche (1834) first noted the fact that the beds extend 21 km inland from Greencliff, and Sedgwick and Murchison (1840) referred to three beds of culm at Bideford. They described the southernmost as the Paint Vein. EAE, BJW

In a detailed account of the culm beds, De la Beche (1839) listed several active workings. He stated that the bulk of the production in the East-the-Water mines came from the Middle or Great Anthracite Bed, 0.13 to 4.27 m thick, averaging 2.13 m. Workings 365 m long were entered by adit. Production was some 4500 tonnes per year for about two years prior to 1838, by which time this seam had been worked out down to about 18.3 m, the limit with the contemporary machinery. Little anthracite was worked after about 1850, but the Paint Seam was worked intermittently until 1969.

A mining engineer's report stated in 1927 that four anthracite seams were known north of the Paint Seam. A newly sunk shaft [SS 4743 2626] to 28.1 m intersected a roadway running N–S at 3.05 m from the bottom. At 52.43 m north of the shaft the Two Foot Seam of hard anthracite was 0.38 to 0.46 m thick; at 36.58 m north of the shaft the Five Foot Seam (possibly De la Beche's Middle or Great) had a maximum thickness of 1.09 m. At 38.40 m south of the shaft the No. 1 Mary Ann Seam was 0.61 m thick, and 2.44 m farther south the No. 2 Mary Ann Seam ranged from 0.61 to 0.79 m in thickness. The Paint Seam was 41.15 m farther south; in the seven years up to 1926 it yielded 150 tonnes per annum, the culm being crushed to make Bideford Black pigment. Later production was variable.

The seams dip steeply, and were worked in stopes as in metalliferous lode mining. These steep and variable dips suggest the possibility that there may be fewer than four seams, repeated by folding. Probably the Paint Seam is the highest in the coal-bearing succession, which may extend over some 175 m of the local basal Bude Formation.

Many old pits, adits and shafts were found during the present survey, and most are shown on the 1:50 000 geological map. The authors are indebted to Mr H. St L. Cookes for pointing out many of these localities and for information concerning the mining operations which were carried out at East-the-Water from the early 19th century until 1969. BJW

Chope (1902) stated that a thin seam of culm was worked near Coalpit Lane [SS 2820 2485], Hartland. Mr Heal, a local carpenter, is said to have made his black paint from this deposit, but the name Coalpit Lane is the only remaining evidence of working. RTT

Quarrying

Only two quarries are currently active in the Bideford district. Carboniferous sandstone is dug at Colpit Quarry [SS 2792 2495], near Hartland (p. 45). Pickwell Down Sandstones are worked at Vyse Quarry, North Downs [SS 491 412], where the rock, mainly massive fine-grained sandstone, is crushed on the site for use as roadstone and hardcore.

Numerous small pits throughout the district show where a variety of rocks has been dug for local use, in some cases for construction of a single building. A few examples of larger pits or quarries may be mentioned. Grey slates of the Morte Slates have been quarried [SS 4706 4366] just northeast of Over Woolacombe Barton and, together with associated siltstone, at [SS 4795 4348] near Oussaborough and at [SS 4996 4300] near Bradwell.

Pickwell Down Sandstones have been quarried more than any other formation, for their strong massive sandstone. A disused quarry [SS 4610 4303] immediately east of Potter's Hill, south of Woolacombe, shows purple and grey sandstone and slate. Other large pits are at [SS 456 425], where the sandstones are accompanied by a good deal of slate and silty slate, at [SS 4753 4287], [SS 4771 4258], [SS 4774 4247], [SS 4566 4142], where the sandstones are cut by an iron-rich fault zone, and [SS 4502 4083]. Stony Bridge Quarry [SS 4935 3945] is the largest disused quarry in the Pickwell Down Sandstones; a 24-m face shows sandstone with some slate and siltstone dipping steeply between south-east and south-south-east.

The Baggy Sandstones, generally fine to medium grained and more friable than those of Pickwell Down, have been quarried for strictly local use at and around [SS 4279 4032] near Croyde Hoe, and at [SS 4671 3940] near Crowborough.

Shales, slates and sandstones of the Pilton Shales have been dug from many small pits, as at [SS 4331 3846], [SS 4480 3785], [SS 4578 3888], and several localities at Saunton. A larger quarry [SS 474 377] near Fairlynch shows a 15-m face of silty slates, shales and fine- and medium-grained sandstones. Similar stone has been worked at Corhill Head Copse [SS 490 378], north of Braunton, and at Braunton Down Quarry [SS 493 369], and small diggings are common north of Braunton alongside the old railway and the A361 road.

Crackington Formation rocks have been dug from many small pits east and south-east of Instow. Generally the workings extend along sandstone crops as lines of shallow depressions, broken near-surface stone having been taken for a nearby farm, cottage or barn. Massive and thickly bedded fine-grained sandstones have been quarried [SS 4835 2904] near Tapeley and [SS 4977 2844] a kilometre east-north-east of Eastleigh.

Some rough walling in coastal areas shows the use of sandrock from the raised beach.

One of the major stone-quarrying enterprises in the district was the working of granite on Lundy. The Lundy Granite Company was established in 1863 and worked stone in five quarries on the eastern side of the island; all five lie at the top of the seaward slope, adjacent to the central plateau, the southernmost at [SS 1376 4504] and the northernmost at [SS 138 456]. A broad track skirting the quarries on their seaward side marks the position of an old tramway, along which granite was carried to the head of an incline [SS 1385 4508], whence it was lowered down the cliffs to a pier. The slopes are littered with granite blocks from the quarries, the incline is discernible, but the quay has been destroyed. Small vessels only were able to berth and the company ran its own steamer to Instow and Fremington Quay. Local granite was used to build Marisco Castle in the 13th century and the two currently operating lighthouses on Lundy in the 1890s, and 'exported' granite may be seen in the Victoria Embankment and the Charing Cross Hotel.

The Lundy Granite Company operated for five years, bringing over 200 labourers to Lundy; it closed down in 1868 but the last of the employees, for whom cottages and a hospital were provided, did not leave until 1887. Failure was due even more to the weather than to the distances involved. For periods of several weeks ships were prevented from berthing and the company found itself compelled to buy granite from Cornwall to fulfil its contracts. An attempt to reactivate the quarries in 1897, under the name Lundy Granite Quarries, survived for three years. Subsequent revivals, first as Lundy and Mainland Quarries Ltd and then as Lundy Island Granite Quarries Ltd, achieved very little and the industry ceased in 1911. EAE

Aggregate materials used in the construction of wearing courses on roads become polished by the passage of traffic, and this increases the likelihood of skidding. The degree of polishing varies with the rock type involved. The most polish-resistant natural materials tend to be those classified in the B. S. Gritstone Trade Group. However, a survey of pre-Permian arenaceous rocks (Hawkes and Hosking, 1972) suggests that comparatively few 'gritstone' formations in Britain will yield aggregate combining both resistance to polishing (measured as the polished stone value PSV), and adequate strength (measured as the aggregate abrasion value AAV).

Feldspathic greywacke from Stony Bridge Quarry [SS 4935 3945] has PSV 75 and AAV 19; Crackington Formation greywacke from Shamley Quarry [SS 2490 2685] near Hartland Point has PSV 68 and AAV 8. Aggregate materials for 'A' class roads should have a PSV of 62 or more and an AAV of 10 or less. Thus although the Pickwell Down Sandstones of Stony Bridge Quarry might provide an excellent skid-resistant surface, they would probably be too weak for use. The Crackington Formation greywackes, however, might locally provide aggregate strong enough for use in wearing courses and in addition they seem to possess a significantly greater resistance to polish than is recommended for 'A' class roads. It follows that the Carboniferous sandstones of north-west Devon might constitute a resource of skid-resistant road-surfacing material. JRH

Sand and gravel

The Braunton Sand and Gravel Company dig sand and gravel from the bed of the Taw–Torridge estuary by grounding barges on a sandbank on the ebb tide, dredging up a cargo and refloating at high tide. The material is rather Shelly and of mixed grade; it is unloaded at any of several quays and supplied to local contractors. Sand and gravel dredged from the Bristol Channel is also landed in the estuary. Total landings, although variable, are of the order of 20 000 to 30 000 tonnes a year. They may be increased, although the shallower submarine banks are being depleted, and remoteness from the main areas of demand argues against major development in the near future.

In the 19th century beach sand was dug from Saunton Sands, distributed by horse-drawn cart and sold to cottagers at a penny a bucketful for spreading on beaten earth floors.

Clay

Boulder clay crops out to the south of the A39 road eastwards from Venn crossroads [SS 4821 3136] to beyond Yelland Manor [SS 4935 3191]. The outcrop ranges up to 500 m in width. Farther east a 6-m thickness of lake clays within the boulder clay is dug to supply Brannam's Pottery in Barnstaple. These stoneless pottery clays are not known to extend within the present district; two boreholes [SS 4881 3145] near West Yelland, sunk to 10 and 10.05 m, and a third [SS 4946 3175] south of Yelland Manor, sunk to 6 m, are recorded as being entirely within variably stony boulder clay. However, the few exposures present indicate that a good deal of the clay contains scattered small stones only, and while perhaps unsuitable for pottery it forms a small reserve of brick clay.

Silty and sandy loams have been dug from Brick Field on Lundy and used there for roughly made bricks and tiles.

Soils

Soils (Figure 25) on the Devonian rocks to the north of the Taw estuary are generally thin, acid and damp, and are mainly devoted to the grazing of sheep and cattle. Rather more dairying is carried on on the locally calcareous silty clay loams of the Pilton Shales, and some cereal and root crops are raised. South of the estuary the Upper Carboniferous rocks mainly weather to wet clays and silty clays. The Bideford Formation gives rise locally to sandy friable soils, but most of the sandstones of the Crackington Formation contain so much argillaceous material that they weather to a silty clay. Some patches of lighter soil occur on the Bude Formation. Accumulation of fine clayey down-wash in valleys and depressions has produced low-lying expanses floored by up to 2 m of clay, bluish grey and unoxidised below and orange-brown oxidised above, with consequent intractable drainage problems. Such low wet ground encourages the development of 'black ram', a ferruginous and manganiferous hard pan commonly about 0.3 m thick and a similar distance below the surface. South of the estuary also the emphasis is on stock and dairy farming, with not much arable land. Farms of the district are commonly between 20 and 35 ha; on those which cultivate cereals, barley and oats are commonly preferred to wheat, and of the root crops good yields of potatoes have been obtained from the lighter soils.

Rich alluvial silty loams bordering the estuary support dairy herds, arable farming, market gardening and a bulb farm. Brassicas, sugar beet and salad crops are raised in addition to cereals. Braunton Marsh, largely of estuarine alluvium, affords some of the finest cattle-grazing land in Devon. Braunton Great Field represents one of the three great open village fields of Saxon times. In 1840 it extended over 200 ha, in 1889 about 160, and it now comprises about 140 ha in 200 strips ranging from 0.3 to 2.5 ha each. As the size of the open field has diminished so has the number of farmers, from 50 at the end of the 19th century to about a dozen today. Similarly the practice of dressing the land with seaweed and lime-rich (shelly) sand, brought by cart or by boat up Braunton Pill, has died out.

Hartland Forest has been created by the Forestry Commission in an area of Upper Carboniferous rocks bearing poor wet acid soils capable of supporting commercial plantings of conifers. In 1896 a group of Frenchmen began an attempt to 'reclaim' Braunton Burrows by the planting of pine trees, but they abandoned their efforts after only a few weeks. A few small pines are scattered among the dunes at present but it is not known whether any are descendants of these early plantings. Sands from the Burrows are spread extensively over adjoining fields by on-shore winds, and the dunes themselves are occasionally threatened by blow-outs. Operations to clear beach minefields after the Second World War resulted in the washing away of the seaward foredunes and wave attack upon the westernmost of the remaining three dune ridges. The main ridge was bare of vegetation and was being blown away, while the landward ridge, although blown out at two places, was migrating inland. Following advice from the Nature Conservancy the War Department began planting marram grass in 1952 and within seven seasons the blow-outs were healed, the fore-dunes restored and extensive areas of central dunes anchored. Some 600 ha of the southern end of Braunton Burrows is a National Nature Reserve.

Lundy Island is now the property of the National Trust. Its thin sandy soil, derived from granite, offers rough grazing to farm stock, deer and feral goats. It is characteristically peaty, locally so much so that accidental fires in the mid 1930s laid bare much underlying granite in the northern part of the island. However the so-called Tillage Field and Brick Field, immediately north of the settlement, have yielded cereals for use as cattle food.

Water supply

In general the district, which lies within the North Devon province of the South West Water Authority, depends for water upon surface supplies, although small quantities are drawn from wells and boreholes mainly by individual users. The slates and shales of the Morte Slates, Upcott Slates, Pilton Shales and much of the Carboniferous outcrop have low permeability. Storage and movement of ground water within them is dependent upon the presence of fissures and shatter belts. The rocks have been much broken by folding and faulting, and hence may store appreciable quantities of water. However the groundwater is likely to occupy fracture systems which may or may not be interconnected, so that supplies will in general be small and unpredictable. The Pickwell Down Sandstones and Baggy Sandstones offer better potential as aquifers, although being more competent than the argillaceous formations they have factured less during orogenic movements.

Two reservoirs supplying Ilfracombe lie about 400 to 500 m east of the district, on Morte Slates. They are constructed one immediately upstream of the other in a small valley with limited catchment area and have a combined capacity of 245 million litres. Similar limitations apply to possible additional sites in the district, and later reservoirs some kilometres to the east, at Challacombe and Wistlandpound on the western borders of Exmoor, are not of large capacity. A reservoir tank to supply Braunton was constructed in the 19th century alongside Buttercombe Lane, about 11 km east of the village and a few hundred metres east of the district. It stores water from small valleys to the west and north-west and provides a gravity-fed piped supply to Braunton. Water has been impounded on Crackington Formation strata south and south-east of Bideford. Jennett 's Reservoir [SS 442 247] is sited in an eastward-draining tributary of the River Torridge, and the two adjacent and linked Gammaton reservoirs [SS 485 250] lie east of the river and store water from a very small catchment without a permanent flow.

A 0.15-m-diameter borehole at Spreacombe, at about [SS 4855 4155], penetrated 40 m of Pickwell Down Sandstones (p. 19) and yielded 0.38 litres per second (l/s) for an 18-m drawdown, recovery taking 45 minutes. The site is in line with north-west-trending and north-east-trending faults; water was struck at 14 m, in sandstones within clays, and also at the top of and within a 10-m-thickness of shattered sandstone at the bottom of the hole. A nearby 52-m borehole at Spreacombe Lodge [SS 492 421] obtained 0.76 l/s and a 0.15-m-diameter borehole [SS 4963 3987] in this formation at Heddon Dene was sunk to 37 m and yielded 0.63 l/s from sandstones with no apparent drawdown. The only well record from the Baggy Sandstones outcrop refers to a site [SS 448 401] at Putsborough where a 27-m borehole yielded 0.51 l/s.

A 120-m borehole [SS 448 378] in Pilton Shales near Saunton Sands Hotel was dry and a combined well and borehole [SS 472 377] near Lobb yielded 0.32 l/s from 19 m of similar strata. In contrast a 65-m bore [SS 497 376] just outside the district, and about 550 m downstream of the Braunton reservoir mentioned above, supplied 4.04 l/s on test and subsequently 2.63 l/s on continuous pumping for three weeks, when the yield began to decrease. It seems possible that this borehole passed from Pilton Shales into Baggy Sandstones and that the water, recorded as having been struck at 15 m, 21 m and in the bottom 10 m, came mainly from the latter formation.

Water supplies from Crackington Formation strata are varied, but generally small. A 30-m well at Buckleigh, with 137 m of headings, gave 1.51 l/s, as did a 69-m borehole near Westleigh, while a 37-m bore at Cornborough yielded 1.26l/s. In the Bideford and Abbotsham areas supplies are correspondingly small, holes up to 46 m deep yielding between 0.76 and 1.58 l/s. In the east of the district, at East Yelland Power Station [SS 481 324], shales with sandstones of the Crackington Formation were drilled to 37 m and a supply of 1.26 l/s obtained for a 19-m drawdown. Several sinkings in this formation along the southern edge of the district obtained small supplies of up to 0.6 l/s. At Melbury Water Works [SS 383 200] a 30-m borehole drilled to supply Northam gave 6.06 l/s on test.

The Crackington and Bude formations of the Hartland area have been tapped for small supplies for individual farms. Amounts ranging from 0.25 to 1.01 l/s have been drawn from wells and boreholes between 12 and 50 m deep.

Several boreholes probably draw at least part of their supplies from drift deposits. At the NALGO Holiday Centre, Croyde, two boreholes penetrated respectively 14.5 and 18.3 m of Head and raised beach on 33 and 36 m of Pilton Shales, and yielded 2.02 and 1.39 l/s. Three bores are sited on the estuarine alluvium of Braunton Marsh. Two at a pumping station [SS 465 375] passed through 8 m of clay, sand and pebbles on 16 m and 3 m respectively of Pilton Shales and yielded 1.89 and 1.01l/s. The third, 2 km to the south, passed through 11 m of similar drift into 5 m of shales and gave 1.12 l/s with a drawdown of 9 m; water is recorded as having been struck within a metre of the surface and a further inflow was observed within the bottom metre of the hole.

Water supplies on Lundy Island depend wholly on storage of surface water, and are at present inadequate in dry summer weather when the population is much increased by holidaymakers. A covered reservoir tank constructed in 1900 lies 100 m west of the hotel. Several natural pools and springs on the island are untapped, but offer only small potential supplies, generally at a considerable distance from the main area of settlement. The stream alongside the road to the landing beach is small and impersistent, although the remains exist of the watermill which gave the name Mill Combe to the valley. A few streams emerge at the coast as small cliff waterfalls and two of these, on White Beach [SS 1398 4446] and Ladies Beach [SS 1402 4426], have in the past been used by ships taking on water. North West Point, on which stands Lundy North Lighthouse, contains a cave at sea level within which fresh water, the Virgin's Spring, bubbles up into the sea; this must be rainwater which has percolated down through weathered granite into fissures in the fresh rock below.

In the immediate future the North Devon Division of the South West Water Authority will rely largely on Meldon Reservoir on Dartmoor, which began impounding in March 1972 and supplying water in the summer of that year. Demands for water and storage capacity at Meldon are such that this source will need to be supplemented by Wimbleball Reservoir in the 1980s. It seems unlikely that a return will ever be made to the practice of constructing small reservoirs such as currently supply Bideford and Ilfracombe. Also, the available evidence offers little hope of large supplies from wells or boreholes. Additional surface storage could be provided by dams in the upper or middle reaches of the Torridge or Taw river systems. A dam near Beaford Bridge [SS 543 142], for example, would impound a considerable quantity of water on ground of low agricultural value. However, it would flood a large area including the Petrockstow ball clay basin. Fresh water could be drawn from the estuary, after sealing it off from the sea. This would have a major impact on the Bideford district in the fields of shipbuilding, sea-borne trade, fishing, sand and gravel dredging, mudflat ecology, sewage disposal and tourism, and several of these effects would be echoed in the surrounding region. Barnstaple alone discharges 7 million litres of crude sewage into the river every day. A barrage in the neighbourhood of Appledore, founded on rock at up to 25 m below OD and impounding water at about high-water-mark, would create a reservoir of some 230 000 million litres capacity. However, in March 1972 the North Devon Water Board and the Devon River Authority rejected the idea of an estuarial barrage, except as a possibility in the distant future.

Miscellaneous

Seaweed has in the past been widely gathered for use on the land, but during the Second World War this material was collected from the rocks around Croyde Bay and used as a source of iodine.

Salt was once obtained from sea water by evaporation in salt pans on Saunton Sands. The material was cleaned and prepared for sale at Saunton Court, and then stored at Lobb; the latter place, now a farm, was identified as a saltern or salt works in the Domesday Book of 1086.

Future prospects

There is little prospect of mining, for either metals or culm, having an appreciable effect on the economy of the Bideford district. On the other hand resources of stone suitable for roadstone and hardcore are large and might be exploited in the course of any major local development such as a trunk road, dam, barrage or nuclear power station. Quarrying would probably be largely from the Pickwell Down Sandstones and the Baggy Sandstones. However the amount of strong sandstone within the Pilton Shales increases eastwards, and requirements south of the estuary might be met most conveniently from sandstones of Carboniferous age.

Off-shore sand and gravel resources are considerable, but again mainly relevant to local demand. The vast quantities of beach and dune sand at Woolacombe, Saunton, Braunton Burrows and Northam are likely to be preserved untouched for reasons for amenity.

Boulder Clay north-east of Instow could be used for brickmaking but the deposit is probably too small to permit economic working.

There is no major aquifer in the district and no prospect of obtaining large supplies of groundwater. Any major reservoir scheme would necessarily take the form of impounding water in the Taw–Torridge estuary.

Thus exploitation of mineral resources is likely to have appreciable impact on the district only in the event of large-scale urban and industrial growth. EAE

Chapter 11 Geophysical investigations

This chapter summarises the results of geophysical investigations in the Bideford–Lundy district. Regional results can be presented most effectively by including also the Bude Bay (307) and Bude (308) sheets.

Regional aeromagnetic surveys

(Figure 26) shows the results of aerial surveys of the Earth's total magnetic field flown by Hunting Surveys Ltd in 1958 (over land) and 1961 (over sea). A small area in the extreme north-east was surveyed by Canadian Aero Service Ltd in 1960. The boundaries between these areas are indicated on the figure by broken lines, or by the coastline. The mismatch in the contours at the boundaries of the various surveys may result from small errors in calibration or from secular variations in the field for which full corrections were not made. At sea, the measurements were made at a height of 1000 ± 150 ft (305 ± 46 m)along north–south lines 2 km apart, and on land at a height of 500 ± 100 ft (152 ± 30 m) along north–south lines 0.4 km apart. The map shown in the figure represents departuresfrom a linear reference field having a value of 47 903 nanoteslas (nT=gamma) at the origin of the National Grid, increasing at a rate of 2.17 nT/km northwards and decreasing at 0.26 nT/km eastwards.

The regional negative magnetic gradient from south to north across the whole district is disturbed by a roughly circular 'high' about 15 km in diameter around Lundy. This is further complicated by a 'low' about 3 km to the west of the island, which divides the positive anomaly into two parts, the more prominent of which is centred just off the island's south-east corner. Blundell (1957) found the Lundy Granite to be only weakly magnetised, so that it could not be the source of the surrounding anomalies and is therefore unlikely to extend far out to sea; his measurements on samples from dykes on the island showed that they have a strong reverse magnetisation. More recent measurements by Mr S. R. Winter and Professor M. Brooks (personal communication 1977) demonstrated that submarine dykes can produce anomalies at sea level of up to 250 nT, and that some of them are normally magnetised. Thus, Tertiary dykes must contribute towards the aeromagnetic field, but the shapes of the more prominent anomalies suggest a different source. They probably arise principally from components of a large Tertiary igneous complex, the possible existence of which is discussed further in the section on gravity surveys. Inspection of the flight line profiles shows that the magnetic sources causing the more prominent anomalies around Lundy must be close to the surface and are magnetised in more than one direction. This suggests that components of the complex may have been formed at different times, as has been suggested for the Tertiary igneous ccmplex on Skye on consideration of the aeromagnetic results (Brown and Mussett, 1976).

To the north-west of Lundy, an elongated negative anomaly extends north-westwards for about 30 km. It was suggested by Cornwell (1971) that this is probably caused by a reversely magnetised Tertiary dyke, or dyke swarm, dipping steeply to the south-west.

Regional gravity surveys

A marine gravity survey of the area north of latitude 51° was carried out jointly by the University College of Swansea and the Marine Geophysics Unit of IGS in 1971 (Brooks and Thompson, 1973). Marine data south of 51° were obtained by the IGS in a later survey. On land, regional gravity surveys were carried out by the author in 1970. Most of the marine data were measured on east–west traverses 5 km apart with north–south traverses 12 km apart, and the land data at a station density a little under one per square kilometre. All results have been adjusted to NGRN 73, the 1973 National Gravity Reference Network, and the Bouguer anomalies shown in (Figure 27) have been derived using the 1967 International Gravity Formula, based on a rock density of 2.67 g/cm³ (2.70 g/cm³ for land data). Marine data were corrected for water depth and land data for topography. The accuracy of the land results is within 0.1 mGal and the average discrepancy at the intersection of marine traverses was quoted by Brooks and Thompson as 1.1 mGal.

The dominant features on the Bouguer anomaly map are the large 'high' lying to the west and south of Lundy, and the steep negative gradient to the north-north-east which forms part of a large 'low' in the Bristol Channel.

The Lundy 'high' lies at the north-east end of a zone of high Bouguer anomaly values (above 30 mGal) which runs roughly parallel to the coast of Devon and Cornwall and extends beyond Land's End (Davey, 1971). Bott and others (1958) made the first gravity measurements on Lundy and tentatively explained the high values in terms of an unexposed basic mass genetically associated with the granite. Brooks and Thompson (1973) produced a model of a large basic pluton, assumed to be contemporaneous with the granite, lying at shallow depth. Assuming a mean density of 2.95 g/cm³ for this pluton, and removing 'regional' effects by trend surface analysis, they predicted a maximum thickness of basic rock to the west of the island of between 2.5 and 4.0 km. The intrusion may partially surround, and perhaps underlie, the Lundy Granite, though the thickness of any basic rocks beneath it must be small. Bott and others measured a local residual 'low' of small amplitude on the island and concluded that the granite could have only a limited lateral extent; this view has been supported by later measurements at sea and by the aeromagnetic results.

Both gravity and aeromagnetic evidence indicate that the boundary of the postulated pluton lies close to the projected line of the Sticklepath Fault and may be controlled by it. There is no magnetic feature over the peak of the Bouguer anomaly 'high' about 12 km west of Lundy. This suggests that the basic intrusion does not have a uniform composition, and contains a large, relatively non-magnetic component. Brooks (1973) considered that the intrusion had a similar structure to the Tertiary igneous centres of western Scotland and that all of the centres might have resulted from movement over an underlying hot spot.

In Barnstaple Bay, a residual 'high' lies off Morte Point. However there is no coincident aeromagnetic anomaly and the causative body may be of different type from the conjectured basic pluton beneath the Lundy 'high'. The cause of the 'high' extending for some hundreds of kilometres south-westwards from the area of the Lundy 'high' is also speculative: it may result from crustal thinning, or possibly from an extensive zone of deep crustal basic intrusion (suggested by Professor Brooks in discussion of Brooks and Thompson, 1973).

The Bristol Channel 'low' lies to the north-east, and only its south-western part is within the area covered by (Figure 27), which shows the westerly continuation of the steep gravity gradient over Exmoor. Some of this gradient may arise from large-scale regional effects: Brooks and Thompson estimated a negative regional gradient of near 0.4 mGal/km towards the NNE, of uncertain cause. After removing this from the observed Bouguer anomaly they attributed the remaining residual gradient in the same direction partly to Mesozoic material in the Bristol Channel syncline and partly to Carboniferous rocks, which beneath Exmoor may be overthrust by Devonian strata. The idea of a thrust was introduced by Bott and others (1958) to explain the whole of the gradient (i.e. effectively with no regional contribution) before any gravity data in the Channel were available. However the existence of a thrust has still not been proved. An alternative suggestion (e.g. Donovan and others, 1971) is that sandstone beds in the Devonian become progressively thinner towards the south, while Brooks and others (1977) have produced a model in which 'a thick sequence of relatively low density Lower Palaeozoic or late Precambrian rocks approaches to within about 2 km of the surface off the north Devon coast'.

Gravity anomalies on land correlate well with the geological strike and reflect lithological changes. South-east of Morte Point a trough in the gravity contours (A in (Figure 27)) follows the outcrop of the Pickwell Down Sandstones, which are clearly less dense than the adjacent formations. Farther south, an east–west ridge of higher values lies roughly along grid line 120N, with lower values along line 110N. These features continue eastwards into the Chulmleigh district, where the ridge corresponds to the outcrop of the Crackington Formation, and the trough to the more arenaceous and therefore less dense Bude Formation. There is a less obvious correspondence on the Bude Sheet, but the explanation may be that there is only a thin layer of the Bude Formation to the north of grid line 117N.

Marine seismic surveys

Recent seismic surveys have generally confirmed the structure deduced from the gravity results and provided more detailed information. Fletcher (1975) described surveys which confirmed the extension of the Sticklepath Fault towards the north-west, passing 4.5 km east of Lundy; drilling suggested that the fault forms the western margin of a large Oligocene basin structurally similar to those at Bovey Tracey and Petrockstow. Seismic surveys by Brooks and James (1975) indicated up to 340 m of Tertiary sediments which are thickest in the west near to the fault. Brooks and James also made measurements over the Bouguer anomaly peak to the west of Lundy which provided supporting evidence for the existence of an intrusion at a depth of about 0.5 km. East of Lundy, the extent of the granite could not be determined with any certainty, but it is unlikely that it extends as far as the line of the Sticklepath Fault.

In most of the remainder of the area (i.e. the southwestern part of (Figure 27)) Dore (1976) reported poor seismic penetration and concluded that the Upper Palaeozoic rocks exposed on land continued westwards.

The Orleigh Court Outlier

In 1970 the Applied Geophysics Unit carried out gravity measurements over the gravel outlier at Orleigh Court (p. 101). A 2-mGal anomaly was defined which indicated a maximum thickness of lighter material close to Orleigh Court. If this were all due to gravel, the probable maximum thickness would be about 50 m, with an average of 20 m over most of the area. Trial seismic spreads suggested that the gravel layer may be thinner than this, and underlain by sediments lighter than the surrounding Carboniferous rocks but having a similar seismic velocity. AJB

References

ARBER, E. A. N. 1904. The fossil flora of the Culm Measures of north-west Devon. Proc. R. Soc. London, Ser. B, Vol. 74, pp. 95–99.

ARBER, E. A. N. 1907. On the Upper Carboniferous rocks of west Devon and north Cornwall. Q. J. Geol. Soc. London, Vol. 63, pp. 1–28.

ARBER, E. A. N. 1911. The coast scenery of north Devon. (London: Dent.)

ARBER, E. A. N. and GOODE, R. H. 1915. On some fossil plants from the Devonian rocks of north Devon. Proc. Cambridge Philos. Soc., Vol. 18, pp. 89–104.

ARBER, MURIEL A. 1960. Pleistocene sea levels in north Devon. Proc. Geol. Assoc., Vol. 71, pp. 169–176.

ARKELL, W.J. 1943. The Pleistocene rocks at Trebetherick Point, Cornwall; their interpretation and correlation. Proc. Geol. Assoc., Vol. 54, pp. 141–170.

ARKELL, W.J. 1945. Three Oxfordshire palaeoliths and their significance for Pleistocene correlation. Proc. Prehist. Soc., New Ser., Vol. 11, pp. 20–32.

ASHWIN, D. P. 1957. The structure and sedimentation of the Culm sediments between Boscastle and Bideford, north Devon. Unpublished Ph.D. Thesis, University of London.

ASHWIN, D. P. 1958. The coastal outcrop of the Culm Measures of southwest England. Pp. 2–3 in Abstr. Proc. 2nd Conf. Geol. Geomorphol. Southwest Engl. (Penzance: Royal Geological Society of Cornwall.)

AUSTIN, R. L., DRUCE, E. C., RHODES, F. H. T. and WILLIAMS, A. 1970. The value of conodonts in the recognition of the Devonian–Carboniferous boundary, with particular reference to Great Britain. C. R. 6e Congr. Int. Stratigr. Geol. Carbonif. (Sheffield, 1967), Vol. 2, pp. 431–443.

BALCHIN, W. G. V. 1952. The erosion surfaces of Exmoor and adjacent areas. Geogr. J., Vol. 188, pp. 453–476.

BATE, C. S. 1866. An attempt to approximate the date of the flint flakes of Devon and Cornwall. Trans. Devon. Assoc. Adv. Sci. Lit. Art, Part 5, pp. 128–136.

BEUS, A. A. 1961. Distribution of beryllium in granites. Geochemistry [Engl. trans11, No. 5, pp. 432–437.

BLUNDELL, D. J. 1957. A palaeomagnetic investigation of the Lundy dyke swarm. Geol. Mag., Vol. 94, pp. 187–193.

BLYTH, F. G. H. 1957. The Lustleigh Fault in north-east Dartmoor. Geol. Mag., Vol. 94, pp. 291–296.

BONNEY, T. G. 1878. Note on the felsite of Bittadon, north Devon. Geol. Mag., Vol. 15, pp. 207–209.

BOTT, M. H. P., DAY, A. A. and MASSON SMITH, D. 1958. The geological interpretation of gravity and magnetic surveys in Devon and Cornwall. Philos. Trans. R. Soc. London, Ser. A, Vol. 251, pp. 161–191.

BOWEN, D. Q. 1969. A new interpretation of the Pleistocene succession in the Bristol Channel area. Proc. Ussher Soc., Vol. 2, p. 86.

BOWIE, S. H. U., BISBY, H., BURKE, K. C. and HALE, F. H. 1960. Electronic instruments for detecting and assaying beryllium ores. Trans. Inst. Min. Metall., Vol. 69, pp. 345–359.

BROOKS, M. J. 1973. Some aspects of the Palaeogene evolution of western Britain in the context of an underlying mantle hot spot. J. Geol., Vol. 81, pp. 81–88.

BROOKS, M. J. BAYERLY, M. and LLEWELLYN, D. J. 1977. A new geological model to explain the gravity gradient across Exmoor, north Devon. J. Geol. Soc. London, Vol. 133, pp. 385–393.

BROOKS, M. J. and JAMES, D. G. 1975. The geological results of seismic refraction surveys in the Bristol Channel, 1970–1973. J. Geol. Soc. London, Vol. 131, pp. 163–182.

BROOKS, M. J. and THOMPSON, M. S. 1973. The geological interpretation of a gravity survey of the Bristol Channel. J. Geol. Soc. London, Vol. 129, pp. 245–274.

BROWN, G. C. and MUSSETT, A. E. 1976. Evidence for two discrete centres in Skye. Nature, London, Vol. 261, pp. 218–219.

BORNE, R. V. 1969. Observations on the deposition of the Bude Sandstones. Proc. Ussher Soc., Vol. 2, pp. 109–110.

BORNE, R. V. 1970. The origin and significance of sand volcanoes in the Bude Formation (Cornwall). Sedimentology, Vol. 15, pp. 211–228.

BORNE, R. V. and MOORE, L. J. 1971. The Upper Carboniferous rocks of Devon and north Cornwall. Proc. Ussher Soc., Vol. 2, pp. 288–298.

CAYEUX, L. 1929. Les roches sedimentaires de France: roches siliceuses. Mem. Serv. Expl. Carte Géol. Détaill. France. (Paris.)

CHAO, E. C. T. and FLEISCHER, M. 1960. Abundance of zirconium in igneous rocks. Rep. 21st Sess. Int. Geol. Congr., Part 1, pp. 106–131.

CHOPE, R. PEARCE. 1902. Mining at Hartland. Notes of the past No. 45. Hartland Chron., No. 73, p. 30.

CHURCHILL, D. M. and WYMER, J. J. 1965. The kitchen midden site at Westward Ho!, Devon, England: ecology, age and relation to changes in land and sea level. Proc. Prehist. Soc., Vol. 31, pp. 74–84.

CONYBEARE, J. J. 1814. Memoranda relative to Clovelly, north Devon. Trans. Geol. Soc. London, Ser. 1, Vol. 2, p. 498.

CONYBEARE, J. J. 1823. On the geology of Devon and Cornwall. Ann. Philos., Ser. 2, Vol. 5, p. 184; Vol. 6, p. 35.

CORNWELL, J. D. 1971. Geophysics of the Bristol Channel area. Proc. Geol. Soc. London, No. 1664, pp. 286–289.

DAVEY, F. J. 1971. Bouguer anomaly map of the North Celtic Sea and entrance to the Bristol Channel. Geophys. J. R. Astron. Soc., Vol. 22, pp. 277–282.

DAVID, T. W. E. and PITTMAN, E. F. 1899. On the Palaeozoic radiolarian rocks of New South Wales. Q. J. Geol. Soc. London, Vol. 55, pp. 16–37.

DEARMAN, W. R. 1964. Wrench-faulting in Cornwall and south Devon. Proc. Geol. Assoc., Vol. 74, pp. 265–287.

DEARMAN, W. R. 1967a. On the significance of northward facing structures at Duckpool, near Bude in north Cornwall. Proc. Ussher Soc., Vol. 1, pp. 266–269.

DEARMAN, W. R. 1967b. Structural patterns in the Upper Carboniferous rocks at Welcombe Mouth, north Devon. Rep. Trans. Devon. Assoc. Adv. Sci. Lit. Art, Vol. 99, pp. 273–286.

DEARMAN, W. R. 1969a. An outline of the structural geology of Cornwall. Proc. Geol. Soc. London, No. 1654, pp. 33–39.

DEARMAN, W. R. 1969b. On the association of upright and recumbent folds on the southern margin of the Carboniferous synclinorium of Devonshire and north Cornwall. Proc. Ussher Soc., Vol. 2, pp. 115–121.

DEARMAN, W. R. 1971. A general view of the structure of Cornubia. Proc. Ussher Soc., Vol. 2, pp. 220–236.

DEARMAN, W. R. and FRESHNEY, E. C. 1966. Repeated folding at Boscastle, north Cornwall, England. Proc. Geol. Assoc., Vol. 77, pp. 199–215.

DE FREITAS, M. H. 1972. Some examples of cliff failure in S. W. England. Proc. Ussher Soc., Vol. 2, pp. 388–397.

DE LA BECHE, H. T. 1834. On the anthracite found near Bideford in north Devon. Proc. Geol. Soc. London, Vol. 2, p. 106.

DE LA BECHE, H. T. 1839. Report on the geology of Cornwall, Devon and west Somerset. Mem. Geol. Surv. G.B.

DE RAAF, J. F. M., READING, H. G. and WALKER, R. G. 1965. Cyclic sedimentation in the lower Westphalian of north Devon, England. Sedimentology, Vol. 4, pp. 1–52.

DEWEY, H. 1910. Notes on some igneous rocks from north Devon. Proc. Geol. Assoc., Vol. 21, p. 429.

DEWEY, H. 1913. The raised beach of north Devon: its relation to others and to Palaeolithic Man. Geol. Mag., Vol. 10, pp. 154–163.

DEWEY, H. 1935. South-west England. Br. Reg. Geol., Inst. Geol. Sci.

DINELEY, D. L. and RHODES, F. H. T. 1956. Conodont horizons in the west and south-west of England. Geol. Mag., Vol. 93, pp. 242–248.

DINES, H. G. 1956. The metalliferous mining region of southwest England. Mem. Geol. Surv. G.B.

DIXON, E. E. L. and VAUGHAN, A. 1911. The Carboniferous succession in Gower (Glamorganshire), with notes on its fauna and conditions of deposition. Q. J. Geol. Soc. London, Vol. 67, pp. 477–571.

DODSON, M. H. and LONG, L. E. 1962. Age of Lundy Granite, Bristol Channel. Nature, London, Vol. 195, pp. 975–976.

DODSON, M. H. and REX, D. C. 1971. Potassium-argon ages of slates and phyllites from south-west England. Q.1. Geol. Soc. London, Vol. 126, pp. 465–499.

DOLLAR, A. T. J. 1942. The Lundy complex: its petrology and tectonics. Q. J. Geol. Soc. London, Vol. 97, pp. 39–77.

DONOVAN, D. T. 1963. The geology of British seas. (Hull: University of Hull.)

DOLLAR, A. T. J., LLOYD, A. J. and STRIDE, A. H. 1971. Geology of the Bristol Channel. Proc. Geol. Soc. London, No. 1664, pp. 294–295.

DORE, A. J. 1976. Preliminary geological interpretation of the Bristol Channel approaches. J. Geol. Soc. London, Vol. 132, pp. 453–459.

EDMONDS, E. A. 1972. The Pleistocene history of the Barnstaple area. Rep. Inst. Geol. Sci., No. 72/2.

EDMONDS, E. A. 1974. Classification of the Carboniferous rocks of south-west England. Rep. Inst. Geol. Sci., No. 74/13.

EDMONDS, E. A., MCKEOWN, M. C. and WILLIAMS, M. 1969. South-west England. 3rd Edition. Br. Reg. Geol., Inst. Geol. Sci.

EDMONDS, E. A., MCKEOWN, M. C. and WILLIAMS, M.1975. South-west England. 4th Edition. Br. Reg. Geol., Inst. Geol. Sci.

EDMONDS, E. A., WRIGHT, J. E., BEER, K. E., HAWKES, J. R., WILLIAMS, M., FRESHNEY, E. C. and FENNING, P. J. 1968. Geology of the country around Okehampton. Mem. Geol. Surv. G.B.

ELLIOTT, T. 1976. Upper Carboniferous sedimentary cycles produced by river dominated, elongate deltas. J. Geol. Soc. London, Vol. 132, pp. 199–208.

ETHERIDGE, R. 1867. On the physical structure of west Somerset and north Devon and on the palaeontological value of the Devonian rocks. Q. J. Geol. Soc. London, Vol. 23, pp. 568–698.

EVANS, J. W. and POCOCK, R. V. 1912. The age of the Morte Slates. Geol. Mag., Vol. 49, pp. 113–115.

FALCON, N. L. 1952. In discussion of COOK, A. H., HOSPERS, J. and PARASNIS, D. S. The results of a gravity survey in the country between the Clee Hills and Nuneaton. Q. J. Geol. Soc. London, Vol. 107, pp. 287–306.

FERSMAN, A. E. 1934. Geochemistry. Vol. 1. (Moscow and Leningrad.) [Reference cited by GOLDSCHMIDT, 1954.] FLETCHER, B. N. 1975. A new Tertiary basin east of Lundy Island. J. Geol. Soc. London, Vol. 131, pp. 223–225.

FLETT, J. S. 1910. P. 52 in Summ. Prog. Geol. Surv. G.B. for 1909.

FLETT, J. S. pp. 116–118 in PEACH, B. N., WILSON, J. S. G., HILL, J. B., BAILEY, E. B. and GRABHAM, G. W. 1911. The geology of Knapdale, Jura and North Kintyre. Mem. Geol. Surv. G.B.

FLEUTY, M. J. 1964. The description of folds. Proc. Geol. Assoc., Vol. 75, pp. 461–492.

FRESHNEY, E. C., EDMONDS, E. A., TAYLOR, R. T. and WILLIAMS, B. J. 1979 Geology of the country around Bude and Bradworthy. Mem. Geol. Surv. G.B.

FRESHNEY, E. C., MCKEOWN, M. C. and WILLIAMS, M. 1972. Geology of the coast between Tintagel and Bude. Mem. Geol. Surv. G.B.

FRESHNEY, E. C. and TAYLOR, R. T. 1971. The structure of mid-Devon and north Cornwall. Proc. Ussher Soc., Vol. 2, pp. 241–248.

FRESHNEY, E. C. and TAYLOR, R. T. 1972. The Upper Carboniferous stratigraphy of north Cornwall and west Devon. Proc. Ussher Soc., Vol. 2, pp. 464–471.

GODWIN, H. 1956. History of the British flora. (Cambridge: Cambridge University Press.)

GOLDRING, R. 1955a. The Upper Devonian and Lower Carboniferous trilobites of the Pilton Beds in N. Devon with an appendix on goniatites of the Pilton Beds. Senckenb. Lethaea, Vol. 36, pp. 27–48.

GOLDRING, R. 1955b. Some notes on the cardinal process in the Productidae. Geol. Mag., Vol. 92, pp. 402–412.

GOLDRING, R. 1957. The last toothed Productellinae in Europe (Brachiopoda, Upper Devonian). Paliiontol. Z., Vol. 31, pp. 207–228.

GOLDRING, R. 1962. The bathyal lull: Upper Devonian and Lower Carboniferous sedimentation in the Variscan geosyncline. Pp. 75–91 in Some Aspects of the Variscan Fold Belt. COE, K. (Editor). (Manchester: Manchester University Press.)