201

THE GEOLOGICAL STRUCTURE OF THE COUNTRY ROUND COMBE MARTIN, NORTH DEVON.• By J. W. EVANS, D.Se., F.R.S. Read March 3rd, 19". y primary object in conducting an excursion of the M Geologists' Association to this neighbourhood is to make known the result of my investigation ofits geological structures, as I believe that they are of interest and importance not only for their own sake but because foldings and dislocations of a similar character are extensively developed in the south-west of England. Much still remains to be done with regard to the correlation of the strata represented both in the special area in question and elsewhere in the south-west, but I thought that it might be useful if at the same time I stated what progress I have up to the present been able to make in this direction and what problems still remain to be solved. The foundations of our knowledge of the area were laid by Sir Henry de la Beche in his official report on the " Geology of Cornwall, Devon and West Somerset," the first publication of the Geological Survey, which was issued as early as 1839, and by the classical work of Sedgwick and Murchison, published in the succeeding year. These can only be regarded as pre­ liminary surveys, but subsequent observations have served to confirm the substantial accuracy of their general conclusions. In 1866, however, J. B. Jukes attempted to show that the existence of fundamental strike faults had been overlooked, and that in consequence the accepted succession included a triple recurrence of the same strata. In the following year R. Etheridge, who made a rapid traverse of the area and had the assistance of the fine collections and local knowledge of R. H. Valpy of Ilfracombe, conclusively showed on palreontological grounds and by comparison with the Continental Devonian that jukes's contentions were erroneous and re-established the sue­ cession originally laid down. Unfortunately, Valpy's collec­ tions have been lost and there is now no means of confirming Etheridge's and his determinations. It is at the same time apparent that in some instances Etheridge has, in consequence of insufficient familiarity with the country, misunderstood Valpy's statements as to the localities from which fossils were reported.]

• It is useful to remember that at the present time (1g••) the local magnetic deviationis about 16- W. All directions are, however, given in this paper by reference to the, true astronomical meridian. t A brief account by Valpy of the geology of Ilfracombe was issued anonymously. The 4th edition which I possess contains numerous obvious misprints, most of which were not in the original edition. 202 J. W. EVANS,

Ln , Hicks, who visited the area in the nineties, contributed some interesting observations on the geological structures, but, as will be seen, I am unable to accept his views on the strati­ graphical position of the Morte Slates. I cannot close this brief historical review without reference to the excellent work of Mr. J. G. Hamling, of Barnstaple. It is a matter for congratulation that his valuable collections, which relate chiefly to the Baggy and, Marwood and Pilton Beds and provided the materials for Mr. Whidborne's description of their fauna, are safely housed in the Natural History Museum. BIBLIOGRAPHY. Geol. Surv. Map: Sheet 27 (Old Series). Geol. Surv. Index Map: Old Series, Sheet 10 ; New Series, Sheet 18. DE LA BECHE, SIR HENRY.-" Report on the Geology of Corn­ wall, Devon and West Somerset." Geol, Suruev, SEDGWICK, A., and MURCHISON, SIR R. 1. :_.i On the Physical Structure of Devonshire." Trans, Geol, Soc., Ser. 2, vol. 5, pp. 633'-652. 1866. JUKES, J. B.-" On the Carboniferous Slate (or Devonian Rocks) and the Old Red Sandstone of South Ireland and North Devon." Quart. JOU1'n. Geol, Soc., vol. 22, p. 320. JUKES, ]. B.-" Additional Notes on the Grouping of the Rocks of N. Devon and West Somerset," Dublin. HALL, T. M.-" On the relative distribution of Fossils throughout the N. Devon Series." Quart. [ourn, Geol, Soc., vol. 23, p. 371. ETHERIDGE, R-:-" On the Physical Structure of West Somerset and N. Devon." Quart. Journ. Geol, Soc.,vol. 23, p. 568. 186). VALPY, R H.-" Notes on the Geology of Ilfracombe and the Neighbourhood." Ilfracombe. 1878. BONNEY, T. G.-" Note on the Felsite of Bittadon, N. Devon." Ceol. Mag., pp. 207-209. 1881. USSHER. W. A. E.-" On the Palseozoic Rocks of N. Devon and W. Somerset." Ceol. Mag., pp. 441-448. 1888. MARR, J. E.-" On some effects of pressure on the Devonian Sedimentary Rocks of N. Devon." Geol, Mag., PI" 218-221. HICKS, H.-" Some Examples of Folds and Faults m the Devonian Rocks at and near Ilfracombe." Geol, Mag., pp. 2-9. HICKS. H.-" On the Morte Slates and Associated Rocks in N. Devon and W. Somerset." Quart. Jour". Geol, Soc., vol, 52, p. 254. 1896. HICKS H.-" The Paleeozoic Rocks of West Somerset and North Devon." Proc, Geol, Assoc., vol. 14, pp. 357-370. 1908. HAMLING, J. G.-" Recently discovered fossils from the Lower and Upper Devonian Beds of North Devon." T1'"tls. Devon. Assoc., vol. 40, pp. 276-280. 19 10. HAMLING, ]. G. and ROGERS, 1.-" Excursion to North Devon." Proc, Geol, Assoc., vol. 21, p. 457 (with map on a scale of lin. to the mile). EVANS, ]. W. and POCOCK, R W.-" The Age of the Morte Slates." Ceol.Mag., pp. 113-115. EVANS, J. W.-" Excursion to West Somerset," Proc, Geol. Assoc., vol. 25, pp. 100, 102. EVANS, ]. W.-" Correlation of the Devonian Rocks of North Devon," Ceol. Mag., pp. 147-9. EVANS. J. W.-" Devonian of Great Britain (Sedimentary Rocks)," Handb, del' regionale1f Geologie, vol. 3, part I, The British Isles, pp. 104- 137. STRUCTURE OF COUNTRY ROUND CO~IBE MARTIN. 203

The dominant feature of the Palreozoic rocks which emerge in the south-west of England from beneath the Mesozoic strata is the great synclinorium that traverses the County of Devon from east to west. The centre is occupied mainly by Carbon­ iferous rocks and the northern and southern margins by tracts of Devonian strata. These form the coast line of West Somerset and North Devon from Minehead to Barnstaple and extend some fifteen miles to the southward from the north coast. They also appear further to the east in the Quantock Hills as an inlier which is separated from the main mass by a trough of Permian and Triassic rocks. The succession of the rocks in North Devon, as I under­ stand it, and their correlation with rocks of other areas, is given on pages 204 and 205, and .in the course of my description of the rocks I shall indicate how far my views differ from those put forward by other authors. I have provisionally called the minor divisions of the rocks near Combe Martin by local names, as their correlation with similar divisions elsewhere is still uncer­ tain. Except in the case of these minor divisions the estimates of the thickness of the strata are intended merely to give some idea of the probable thickness. The actual amount is quite uncertain. On the one hand the faulting, as will be seen, tends to omit portions of the succession, and on the other hand local folding increases the apparent thickness. Any calcula­ tion based on the dip of the cleavage, or even the average apparent dip of the strata, would be far too high. In the accompanying sketch map (Fig. 54, p. 209) I have confined myself to indicating the positions of the principal faults and the lithological characters, as the information available at present is not sufficient to justify the insertion of definite boundaries between the different divisions. Except for local variations and folds the rocks have a dip -directed approximately due south and averaging about 35°, but they bend over- in an anticline in the extreme north at the Foreland east of Lynmouth. From that point southward there is an upward succession, and a similar, but more gradual, suc­ cession can be followed along the coast to the west-south-west, oblique to the- dip, as far as Morte Point, where the coast line changes direction and continues nearly due south, approximately parallel to the dip, to Bideford Bay. The general distribution -of the beds is shown in the accompanying diagrammatic map (Fig. 53, p. 206) taken from :\Ir. Etheridge's paper.* It must be noted, however, that the trend of the strata, over the greater part of the area, of about 13c north of west, is not the actual • 0..11" /014",. Geol. Soe., vol. niii., 1867. p. s80. A more detailed map. e:ltendiog fr_ B irnstaple or Bideford Bay to a point east of the Foreland on a scale of three quart.... of an iRCII 10 the mile. was prepared by Mr. Hamling, and published in P'OI. Gtol. A.soc.• nl. xxi., 1910. 1'. 472 , 204- J. W. EVANS,

STRATIGRAPHICAL SUCCESSION IN NORTH DEVON Recent Alluvium Quaternary { Pleistocene Head Stl'tlJigl'aphical lweak. Upper Culm [C] LC I {Pendleside Series with Coddon Cherts Carboniferous ewer u m Cyathaxonia Zone. 1M] Stl'atigl'aphical Brea», 1Upper Pilton Beds (M] Feet Middle Pilton Beds } Lower Pilton Beds (M] 2000 Baggy and Marwood Beds (L] 1200 Pickwell Down Sandstones [e] 30 0 0 Marte Slates [M] 15°0 Upper Woolscott Cleave Slates. Coarse Devonian arenaceous slates [M] 250 Slates with subordinate grits and lime­ stones [M] 1000 David's Hole Beds. Striped calcareous slates with intercalated limestones [M] (thickness uncertain on account of folding and faulting) 25° Red Limestone Series. Striped calcareous slates, bituminous limestone, massive limestone weathering red, and other slates and limestones (thickness on- certain on account of faulting) [M] 355 Jenny Start Beds. Flaggy limestone,. [M] 30 Slates and Limestones of Oakestor Bay •• [M] 135 Newberry Beds. Slates and limestones (M] 520 Slates with grits and lenticular Lester limestones.. •.•••• soo Middle Series { Calcareous slates and limestones I [M] Devonian Grits and slates j SSG Wild Pear { Calcareous slates and len• •- l [M] Beds ticular Iimestoaes r Little Hangman { Grits and slates •• [L] 4°0 ; { Beds. ~~ Rawn's {Shales with plant and fish c (,:) 1Shales remains. ••• [C] 200 \ ~ \ Trentishoe {Grits and slates . [e] 3°00 Grits. Lynton {. .{ [M] and Beds. Slates, limestones and gnts [L] 2000- Lower ( Foreland grits. Grits and slates (base Devonian not seen) .. [C] 15°0 ? Dartmouth Slates (possibly present, but not seen) ..•••• [C] 15°0. (Making an estimated t otal for the Devonian of North Devon of over 20,000 feet. The figures are only intended to give some idea of the possible thickness of the different beds.) The strata between the Hangman Grits and Morte Slates (both ex­ clusive), are usually included in the Ilfracornbe Bed s. [M] = Marine [L] = Littoral. [C] = Continental. STRUCTURE OF COUNTRY ROUND COMBE MARTIN. 2°5 CORRELATION WITH OTHER AREAS. l~ ,R Combe Rock of South-East of England-Glacial Beds. Pleistocene \ ~ i~ Coal Measures. o Pendleside Series. tCarboniferous Zone D3. j Zone K to base of ZI. ?Calcaire d'Etroeungt, Marine bands in the Upper Devonian of Pembrokeshire and Coomhola Beds in Ireland. Psammites du Condroz, Pon Sandstein, Gres de Fiennes, I-GI­ torcan Beds. Assise de Matagne, Schistes de Fiennes, Delabole Slates. Upper Petit Tor Combe. Devonian Saltern Cove. Calcaire de Ferques, Upper Torquay Limestones.

Calcaire Lower

de l:l ~ Torquay South -.;:::., Devon :> Blacourt, i3 Middle Devonian Limestones.

Hope's Nose Limestone l:l"H ~.!!! ~ .'""rd ....c: =..... Gres de Caffiers Limestone and Slates Oi :>'

Baggy Point •.. Qnantock Head Croyde Bay -'-'- Doddington

BlDEFORD BAY

• Igneous Rocks:, [Zd Pickwell Down Sandstones. I_ Foreland Grit•. mJ Baggy and Marwood Beds III Lynton Beds. iii Lower mHangman Grits. !1!iJlIMiddle } Pilton Beds. ~ Ilfracombe Beds. Bl Upper 1• Morte Slates. ..Culm. Later rocks -are nn'shaded: Limestones in the Ilfracombe Beds are shown in black. FIG. 53.-GEOLOGICAL MAP 0]1' NORTH DEVON AND WEST SOMERSJ}T. R. Etheridge. Reproduced [rem Quart. Iou m, Gcol, Soc., vol. 23, 1867, p. 580, by permission of the Council. STRUCTURE OF COUNTRY ROUND COMBE MARTIN. 207 strike of the beds, but the joint effect of a nearly east and west strike and the repeated occurrence of comparativelysmall faults. At the same time the continuity of the stratigraphical divisions. is not infrequently interrupted by more important displace­ ments. STRATIGRAPHICAL DESCRIPTION. The lowest beds exposed at the surface, the Foreland Grits, form the anticline at the Foreland to which allusion has already been made. They are of the typical Old Red Sandstone facies, consisting of coarse arenaceous rocks with subordinate shaly or slaty beds. The only fossils that they have yielded are badly preserved plant remains, which have been referred to Psilophyton, a Lower Devonian (Old Red Sandstone) form, com­ mon to the Midland Valley of So tland and Eastern Canada, and, I believe, obscure fish fragments. It has been contended that these rocks are identical with the very similar Hangman Grits, described later, and that they have been brought up by the faults that bound them on the south and south-west. According to this view they are later. not earlier, than the Lynton Beds. It may, however, be urged against this interpretation (1) that the highest strata north of the faults are closely similar to the Lynton rocks and are probably to be referred to their basal beds, (2) that the Hangman Grits contain pebbles of older grits indistinguishable from the Fore­ land Grits, and (3) that the fossil .remains found in the Hang­ man Grits point to a Middle Devonian age, while those of the Foreland Grits apparently belong to the Lower Devonian. I am personally inclined to correlate the Foreland Grits with the Cosheston Group of Pembrokeshire, which, in the neighbourhood of Neyland and Cosheston, appear to represent the upper part of the Red Marls of South Wales and the Welsh Border. The rocks of the Cosheston group overlie the local Red Marls and consist of Green Sandstones below and Red Sandstones above. They have yielded plant remains, which have been referred to Psilophyton. The Red Marls in their full development appear to correspond to the Dartmouth Slates of South Devon and the Schistes d'Oignies and Schistes de St. Hubert of Belgium, the Gedinnian of Maillieux. According to some authorities the Falmouth Slates of West Cornwall are of the same age. It is not improbable that a boring in the Foreland Grits would disclose the existence below them of red and greenish argillaceous beds, similar to the Red Marls and Dartmouth Slates. The Lynton Beds, which are believed to succeed the Fore­ land Grits, consist of bluish-grey irregular slates, with intercala­ tions of grits and more or less decalcified calcareous bands in 208 J. W. EVANS, which the fossils are usually represented by the spaces from which the calcareous matter of the shells has been removed. They extend from their faulted junction with the Foreland Grits, east of Lynmouth, to Lee Bay (near Lynmouth*) and Wooda Bay.] and also occur at Heddon's Mouth. Though the three latter localities will be visited on the Excursion, the Lynton Beds do not corne within the limits of the task I have undertaken. They appear to represent the Siegenian and Emsian of M. Maillieux. It is very much to be wished that some one would undertake the investigation of these beds and correlate their different horizons with those of the rocks of the Ardennes and the Rhine. On account of the bad .state of pre­ servation of the fossils and the difficulty of extracting them, he would do well to familiarize himself in the first place with the corresponding Continental rocks and their fossils. At Heddon's Mouth, five miles east-north-east of Combe Martin, the deep valley of the River Heddon reaches the sea between lofty sea cliffs. These are formed mainly by the Hang­ man Grits, but the Lynton Beds are seen at their base on the western side of the river, being apparently brought up by a north and south fault, along which the valley has been eroded. An incidental result of the movement seems to have been the separation of two successive beds on the east, leaving a space which has been filled by a breccia of irregular fragments cemented by quartz. Its maximum thickness is about a foot and it thins out to the east. The exact relations between the Lynton Beds and the Hangman Grits remain uncertain. The Hangman Grits extend along the coast to the westward as far as the promontory of the Little Hangman. They consist of massive or fissile sandstones with subordinate shales and form cliffs five or six hundred feet in height. These cannot, as a rule, be descended and it is only possible to land on the rocks of the foreshore in smooth weather. There are, however, a fair number of inland exposures in quarries and road-cuttings. Inlithological character these rocks are substantially identical with those which, in other areas, are known as the Old Red Sandstone. The lower portion, which may be referred to as the" Trentishoe Grits," are, as far as I have been able to ascertain, absolutely unfossiliferous, but there can be little doubt that they are of freshwater origin. On the west of Sherry Combe, another deep val ey to the west, which is on a smaller scale than that of the Heddon, and in the cliffs still further west, fissile beds are met with containing plant remains which are usually linear in form, frequently in parallel aggregates, but sometimes show -The"e is another Lee Bay, west of Ilfracombe. t This is the correct spelling, not" Woody" Bay, as ,rven on the Survey Maps. In the dialect of this part of Dev nshire the genitive is frequently fCllJDed by adjlng the Iuffix -a, Thus •• Smithia Parkt. meaUi the park of the IlDithy. STRUCTURE OF COUNTRY ROUND COMBE MARTIN. 209 dichotomous or lateral branching. A scale of Coccosteus, a genus of fish that ranges through the Middle and Upper Devonian, was also found. These beds may be referred to as the Rawn's Beds from the name given to the upper slopes of the cliff, east of the Little Hangman, which are composed of them. They contain deposits of iron ore which was at one

D.H. . DAVIDS HOLE N.B. . NEWBERRY Bf~CH C.~l.B COMBE MARTIN BfIICH P. .PARADE - - ··..Sea LeveJ}F~ULT5 --- ··..Oulcrcpe ~ Scal., lIMESTONES a k.:;,.·::·.'1 ARHIIlCEOUS BEDS !

FIG. 54.--COMBE MARTIN. time mined and exported to South Wales. It probably repre­ sents ferruginous material that was precipitated in a shallow lake or marsh. Bands of more arenaceous character also occur containing fragments and pebbles of white quartz, similar to those seen in the Upper Old Red Sandstone of South Wales. The Little Hangman Beds that succeed consist of arenaceous rocks, with calcareous beds, and afford evidence of a renewal of marine conditions. They occur high up on the western slopes of Sherry Combe, some 600 feet above the sea, and are also

PROC. GEOL. Assoc., VOL. XXXII!., PART 3, 1922. 14 2IO J. W. EVANS, found near the summits of Holdstone Down, the Girt Hang­ man, and the Little Hangman at a still greater height, but are best examined at the base of the Little Hangman Cliffs on the east of Wild Pear Beach. At least three fossiliferous bands occur at this place, but the fossils are very difficult to get out of the matrix and are seen as a rule only in section. The higher bands show sections of a large brachiopod, which was formerly known merely from loose blocks found on the beach, but I was fortunate enough to find it in situ. It has been referred to Stringocephalus burtini, an attribution that is probably correct, though the internal structure differs some­ what from sections of a German specimen, made at the Natural History Museum, through the kindness of Dr. Smith Woodward. If this identification is correct, these beds must form part of the Givetian, and the continental beds below will be either Lower Givetian or Eifelian (Couvinian of Maillieux). The incoming of marine conditions in the Givetian in succession to continental or littoral conditions is not confined to North Devon. It is seen along the northern border of the Namur synclinorium in Belgium, where a series of littoral beds, the Assise du Poudingue d'Alvaux, contains plant remains and Spirifer mediotextus , Macrocheilus arculatus, and an imprint of Strvngocephalus burtini.* In the Boulonnais a similar succession is met with, the Caffiers Conglomerate and the Caffiers Sandstones, containing plant remains, being suc­ ceeded by the Blacourt Limestone with Stringacephalus burtini, Atrypa aspera, species of Athyris and Spirifer and other fossils.t The Hangman Grits do not seem to be the exact equivalents of the Staddon Grits of South Devon and of Cornwall, which are somewhat similar lithologically but apparently entirely marine. These include the Spirifer cultrijugatus zone at the base of the Middle Devonian, and beds indistinguishable in appearance contain fossils that would indicate an Emsian or Upper Lower Devonian age. The Staddon Grits are succeeded near Torquay by dark grey slates and shaly limestones with Calceola sandalina, representing higher portions of the Lower Middle Devonian, but there is no evidence of the existence of marine beds of this age in North Devon. It is probable there­ fore that the Hangman Grits are, as a whole, younger than the Staddon Grits. The Little Hangman Grits appear to pass upwards into the Wild Pear Beds, consisting of alternations of shales and thin ferruginous limestones weathering brown, and occasional thicker lenticles. Faulting, however, occurs at or near the junction. In the cove immediately east of the Little Hangman there is a normal fault striking north-east and south-west, with • L. D. Stamp, .. Long Excursion to Belgium," Proe, Geol. Assoc., vol. xxxii i., 1922, pp. 10-11. t J. W. D. Robinson, Qua". /0«,... Geol. Soc., vol. lxxvi., 1921, pp. 230, 231. STRUCTURE OF COUNTRY ROUND COMBE MARTIN. 2II a downthrow to the south-east, but the movement is probably not very considerable, though it must result in the omission of part of the succession. On the south-east the grits are bounded by another fault at a distance of a few yards from the cliffs and striking parallel to them at about 30° west of north. The Wild Pear Beds, which form the north-east part of the foreshore, west of the last-mentioned fault, and the south-eastern hill­ slopes, have yielded few fossils except joints of crinoid stems. They have been considered to be an outlier of later rocks, but though they have suffered both folding and faulting, this is not, in my opinion, the case, and they appear to pass up com­ formably into a succession of grits, hard shales and occa­ sional limestones, which I propose to call the "Lester Series." These form the Lester cliffs on the south of Wild Pear Beach and the foreshore at their foot as well as Lester Point and the north-east side of Combe Martin Beach. They are repeatedly folded into synclines and anticlines, and show in places evidence of pneumatolytic action by the presence of pyrites and at Lester Point quartz veins containing galena. A conspicuous limestone band in Combe Martin Beach contains numerous shells, apparently brachiopods and gastropods which can, however, only be seen in section, as they cannot be separated from the matrix. It has been termed the" holey" limestone because it is weathered into hollows by the sea, in contrast to the grits which stand out in bold relief. The upper portion of the series in Combe Martin Beach consists of a sue­ cession of somewhat calcareous slates, containing numerous cylindrical forms, sometimes referred to as "fucoids," which are more arenaceous than the surrounding slates, and appear to represent worm-burrows, but there are branched types, which may be of different origin. They are usually from 3 to 4 mm. in diameter, but sometimes measure only a millimetre across. A little south of Lester Point forms about a centimetre in diameter are sometimes found. These cylindrical bodies are flattened by the cleavage and appear to be parallel to it even where it does not coincide with the stratification. Apparently they originally extended in all directions so that along whatever line the pressure was applied the network as a whole, as well as each stem, was compressed at right angles to it. They also occur at other horizons but are especially numerous here. The only fossil that could certainly be recognized generically in the Lester series was Tentaculites at Lester Point. The calcareous slates are succeeded by a number of lime­ stone beds, varying from 6" to 2' 6" in thickness, followed by soft weathered slates. The Lester series is lost to sight at the south­ east corner of Combe Martin Beach, as the position of the great fault is approached which follows the line of the Combe Martin Valley in a north-west and south-east direction, but they appear 212 J. W. EVANS, to be continued to the south-east on the same (north-east) side­ of the valley in a series of slates with small lenticles of limestone and well-marked gritty bands, giving rise to a succession of minor valleys, separated by ridges which slope obliquely down towards the main valley. The most northerly of these ridges appears to consist of the same grits that form the Lester Cliffsand Lester Point. Itwas in the rocks of this ridge that the silver-lead mines were sunk which were worked from about the twelfth century and only abandoned less than sixty years ago. Even allowing for folding, the total thickness must be considerably greater than that seen in the cliffs and foreshore. On the south-western slopes of the valley, on the other hand, there is an alternation of shales and thick limestones which have been extensively quarried in the past to be burnt for lime, mostly for agricultural purposes, though work has now been almost entirely abandoned. Further reference to these rocks will be made in the course of the description of the coastal section, to which I now return. Immediately south-west of the mouth of the stream draining the Combe Martin Valley. a series of calcareous slates and lime­ stones, with occasional bands of infiltrated quartz, are seen in the foreshore. and they are exposed in the cliffs and foreshore for nearly a third of a mile to the westward. The lowest beds are only visible at low-water of spring tides* north of the double promontory that forms the western boundary of Newberry Beach, between the " L" and "W" of L.W.M.O.T. on the 6-inch maps. They consist of thick limestones, with an estimated thickness (including some inter­ calated slaty beds) of 45 feet. They cannot be recognized on the north-east side of the Combe Martin fault and probably repre­ sent a higher horizon than anything found there. Some 175 feet vertically above these limestones is an easily recognizable bed of limestone, about two feet thick. It is highly ferruginous, contains a considerable amount of pyrites. and is associated both above and below with green calcareous shales, the colour of which appears to be due to a chlorite allied to thuringite. Above this massive limestone, at a vertical dis­ tance of 85 feet, is a thin highly-contorted limestone-a " crinkled" or " wriggly " band-which forms a definite horizon. Another occurs ISS feet still higher, and 13 feet above it is a well­ marked limestone, about IS inches thick, usually thrown into rounded folds. This series of rocks may be termed the New­ berry Series as they are typically developed in Newberry Beach. They do not appear to be ever folded on a large scale, though frequently faulted. The different horizons can be recognized and followed past the great natural arch, known as the Parlour, as far as the ravine, immediately to the east of the promontory, • These occur On this coa.t early in the afternoon. STRUCTURE OF COUNTRY ROUND COMBE MARTIN. 213 terminated by the rock known as David's Stone. Here there is a powerful fault, or rather several closely associated faults, striking about 30° west of north and possibly related to the Salt Lane Valley, near Newberry. This dislocation cuts out a considerable portion of the succession and brings in a much higher series, consisting of thick white limestones and calcareous slates, which may be termed the David's Hole* Limestones. They contain numerous fossil remains, especially crinoid stems and worm-burrows, similar to those found in the Lester series but somewhat smaller than most of them. These limestones are thrown into numerous major and minor folds; the latter have a curious cylindrical appearance, which has caused them to by popularly regarded as fossil trees. The same structure, probably in the same beds, is seen in a quarry near the road above, disused since the landslip of Christmas, 1919. Above these folded limestones there is a well-marked" crinkled" band, some three inches thick, and another smaller one about twenty feet above it. Beyond David's Hole the coast-line continues for about .a hundred yards in a westerly direction, parallel to the strike, up to a point where what is apparently an important north-east and south-west fault, with a down-throw to the south-east, meets a south-south-east and north-north-west fault. The coast then turns abruptly in the last-mentioned direction and continues for about half a mile, with the same general trend more or less modified by faults approximately parallel to the shore and by marine erosion. As a result of the southerly dip the succession is a descending one and this is accentuated by the strike faults which have a down-throw on the dip side. This causes the omission of beds; but what is the thickness lost it is, at present, impossible to say. Separated, therefore, by a gap of unknown amount, we have below the David's Hole Beds a descending series of striped calcareous slates, bituminous limestones, massive limestones weathering red, and further calcareous slates and limestones. These beds, which may be referred to as the" Red Limestone Series," are apparently some of those cut out to the east on the shore by the David's Stone faults, but they are seen in some of the quarries on the south­ west of Combe Martin Valley. They also appear, as one would expect, in the quarries above the beach a little to the north, where they are carried up by the dip. Continuing along the beach we reach a well-marked pro­ montory, Jenny Start Point, forming the southern boundary of Oakestor Bay (wrongly called Sandy Bay on the 6-inch map. Here the dip-slope is formed of a shaly limestone with well­ preserved corals determined by Dr. Stanley Smith as Endophyllum (Spongophyllum) bowerbanki, Cyathophyllum • David's Hole is a perforation in a spur of the cliffs some 400ft. to the west of David's Stolle. r. W. EVANS, caspitosum and Pachypora cervicornis, besides small flatly­ coiled univalves, which are probably gastropods. This horizon also has not been recognized on the shore to the eastward, being cut out in the same manner by the David 's Stone faults. In Oakestor Bay there is a succession of shales and limestones, but so far these have yielded no fossils. North-west of Oakestor Bay, as far as Broad Strand,* the cliffs are formed of slaty beds with limestone bands thrown into frequent folds. They resemble the Newberry and Lester series. and probably can be correlated with them. High up­ in the cliffs limestone with Cy athophyllum. caspitosum and Endophyllum. appears to occur and may be correlated with the J enny Start Limestone, as may also the limestone­ with Cyathophylliom caspitosum seen in the disused road above Broad Strand. The outlying rocks on the other hand are separated as a rule from the cliffs by important faults and appear to belong to the Lester Series. Still further to the north-west th e rocks of the Warren, a peninsula which forms the northern boundary of the inlet of Watermouth, are very similar to those of the upper portion of the Lester Series. Beyond the Widmouth Cha nnel the limestones continue along the shore, but I have not yet given them the same atten­ tion as tho se nearer Combe Martin. The J enny Start Lime­ stone is seen on Widmouth Head, and the quarries west of Sam son's Beach appear to be of the same age, while the rocks of Hele Bay and at the foot of Hillsborough present con­ siderable resemblance to the David 's Hole Lim estones. West of Hele the great succession of limestone comes to an end, being succeeded by higher and less calcareous beds. Returning to the Combe Martin Valley we find the limestone qu arries arranged in east and west lines, echeloned one behind the other, between the stream and the high ground to the south-west. The same beds seem to be brought up again more than once, probably by folding, but th ere is little actual evidence of this except in Harris's Quarry, where a well-marked syncline pitches to the west. It is possible, however, that in some cases the outcrops represent successive lenticular formations of limestones at different horizons, as Maillieux has shown to be the case in the Frasnian of Belgium. In Orchard's Quarry (445 in the I : 2 ,500 map) mu ch of th e limestone has been convert ed into partially oxidized iron carbonate, which has been min ed for umber. In the Knapp quarries, above Jenny Start, limestones similar to tho se in the Combe Martin Valley appear some two thousand feet further north, as the result, no doubt, of faulting, and are continued westward through Bramant's Wood into Northfield. Wood. It is difficult to correlat e them with rocks on the shore. • Not to be confounded wit h" Broad Strand ," near H illsbcr ough, STRUCTURE OF COUNTRY ROUND COMBE MARTIN. 215 but it is probable that they all form part of the succession which has been described. To the south-east of Combe Martin the limestones are everywhere replaced along the line of strike by grits and shales, no doubt as the result of faulting. The high ground from Stony Corner to Kentisbury Down appears to be occupied by distinctly arenaceous beds, the stratigraphical position of which is at present uncertain. Still further eastward, in the neighbourhood of Blackmore Station and the Rowleys, they are replaced by slates with small lenticular limestones, similar to those of the higher parts of the Lester Beds, but there is nothing corresponding with the thick limestones, which have been described between Combe Martin and Ilfracombe. The latter appear to be everywhere succeeded by a thick succession of slates, occasional gritty beds and lenticular or shaly lime­ stones, such as those at the quarries near Henstridge (south­ east of Combe Martin), Goosewell (west of Berrynarbor), and Champernowne (south of Hillsborough). They form the shore­ line from east of Capstan Hill, Ilfracombe, westward nearly as far as Lee (the western locality of that name). In the Goose­ well Quarry, south of Watermouth Castle, I have found Spirifer verneuili (also known as Spirifer disjunctus). This is, there­ fore, in the Upper Division of the Devonian and is, no doubt, of Frasnian age. By the kindness of Professor Marr I learn that the same fossil was found in a quarry above West Hagginton Bay, that must be near the summit of the massive limestones. There is no definite evidence at present as to where the line should be drawn between the Middle and Upper Devonian. I have provisionally placed it below the David's Hole Beds, on account of the great similarity of the massive limestones that underlie them to the Blacourt Limestone of the Boulonnais, which appears to be definitely in the Middle Devonian, but close to the summit of that division. It is possible, however, that the division should be made at a lower horizon, below the Jenny Start Beds. The highest Ilfracombe Beds are the Woolscott Cleave Slates, which are coarse and hard, and somewhat arenaceous. They have been quarried for roofing purposes in Wools­ cott Cleave Plantation and near Harper's Mill, to the south of Berrynarbor. The arenaceous bands are sometimes thrown into small folds and the mean dip of the strata appears to be very low, say 6°, though the cleavage dips at about 50°. These beds were believed by Hicks to be a repetition of the arenaceous Lester Beds brought up On the south side of a great syncline, but there seems no foundation for this contention. They appear to pass upwards into the Morte Slates, fine, smooth shining grey slates, showing occasional arenaceous or calcareous bands, and frequently traversed by quartz veins. These have taken the 216 J. W. EVANS, cleavage so perfectly that any fossils they contain are very much obscured, but in some more arenaceous patches, near Woolacombe, Spirifer uerneuili, var. hamlingi, besides Rhynchoneltas have been found and in some quarries an obscure Lingula occurs. There can be little doubt that the Morte Slates are to be referred to the Frasnian and may be the equivalents of the Assise de Matagne of Belgium. The Delabole Slates of North Cornwall are probably of the same age. At their summit the Marte Slates become more arenaceous and are succeeded by the Pickwell Down Sandstones, which appear to have been laid down under conti­ nentalconditions andcorrespond to the Psammites du Condroz and Pon Sandstein of the Continent. The junction is usually faulted and Hicks believed the Marte Slates to be Silurian or Lower Devonian rocks which had been brought up by strike faults. There seems, however, to be no strike fault of any consequence at the base of the Morte Slates, and the faulting that occurs at the junction of the Morte Slates and Pickwell Down Sandstones is only what is usually present in disturbed areas at the junction of rocks of different physical characters. Near the junction of the Morte Slates and Pickwell Down Sandstone, a band of igneous rock is frequently met with. At Bittadon (Bonney, 1878) and Smithia Park it has been de­ scribed as intrusive; but at Woolacombe Bay fish remains iHoloptychius and Bothriolepis) were discovered in it by Mr. Inkerman Rogers and determined by Dr. Smith Woodward. The Pickwell Down Sandstone has also yielded indeterminable plant remains, probably Archceopteris. The Baggy and Marwood Beds that succeed the Pickwell Down Beds represent littoral deposits similar to those of the Little Hangman Beds, which follow the Middle Devonian continental deposits, but the fossil species are quite distinct. They are overlain by the Pilton Beds, mostly argillaceous and calcareous deposits, the lower and middle beds of which are Devonian, while the upper are Carboniferous, extending upwards to the base of the Zaphrentis zone. Then a break in the succes­ sion apparently occurs, above which deposition seems to have recommenced with the higher portion, D3, of the Dibunophyllum zone, about the horizon of the Cyathaxonia Beds, followed by the Pendleside Series, including the Coddon Cherts. A similar break, terminating at the same point, appears to occur in North Cornwall, South Devon, Belgium, and still further to the east. It is interesting to note that the gap increases from north to south, being apparently non-existent in Gower, while in South Devon it extends downwards into the Upper Devonian and in Northern Brittany it appears to commence in the Middle Devonian. In other respects the North Devon Devonian is intermediate between that of South Wales and the Devonian of South Devon STRUCTURE OF COUNTRY ROUND COMBE MARTIN. 2I7 and North Cornwall on the other side of the synclinorium. In North Devon there are three continental periods succeeded by as many marine phases. South of the synclinorium in North Cornwall there is only one, the Dartmouth Slates, and in South Devon and South Cornwall even these contain occasional marine forms. In South Wales, on the other hand, there is no marine Devonian, except near the summit of the Devonian, where a fauna corresponding to the Baggy and Marwood and Lower Pilton Beds has been found in Pembrokeshire. The difference is really, however, of less importance than that between the strata north and south of the synclinorium, for it mainly consists in the omission in South Wales of strata I.om the Lynton Beds to the Morte Slates inclusive, and later on the omission in North Devon of the Carboniferous Limestone, from just above the base of the Zaphrentis zone to a little below the top of the Dibunophyllum zone. Although Carboniferous and Permian rocks are not repre­ sented in the immediate neighbourhood of Combe Martin, there can be little doubt from their distribution in adjoining areas that they at one time existed here and have .since been removed. Mesozoic and Tertiary strata are also absent, but as Triassic and Jurassic deposits occur in the low ground adjoining the Bristol Channel on the north, and that near Glastonbury, which forms its true continuation to the east, as well as in the smaller pre-Triassic valleys near Minehead and the Quantocks, it is not improbable that these rocks also occur off the coast near Combe Martin. There is an important difference between the state of aggrega­ tion of the last Devonian strata of the Old Red Sandstone type, the Pickwell Down Beds, and the two preceding occurrences of a similar character, the Hangman Grits and Foreland Grits. The former is much less consolidated and apparently yields much more easily to marine action, as witness the wearing back by the sea of Woolacombe Bay, though it is comparatively resistant to subaerial agencies. This difference between strata that must have been originally very similar lithologically may be the result of the greater superincumbent load on the older rocks. If this be so the mass of any later rocks must have been small compared to that of the Ilfracombe Beds and Morte Slates and the Pick­ well Down Beds, and consequently the Carboniferous, Permian, Mesozoic and Tertiary rocks laid down must have been com­ paratively unimportant from the point of view of their total thickness. Mr. Lamplugh has shown in his Presidential Address,* that the geological formations that have survived the destructive forces of erosion probably occur in the major­ ity of cases in regions where they were originally accumu­ lated to a great thickness. • Qu..ri, [our», Geol. 50<., vC)J. lxxv., P, xciit, 218 J. W. EVANS, The latest formation, other than recent alluvium, is the " Head" which represents the downward creep of the subsoil, probably under rigorous conditions of climate, in which the water filling the interstices of the subsoil was frozen during a great part of the year, so that the whole moved forward with what was essentially glacier-movement, though, no doubt, the effect was intensified by repeated alternations of melting and freezing at certain periods. This action affected rocks of all kinds, even hard sandstones, but it is most marked in the case of the slates. The movement is exactly that of a flow. It is usually, of course, downhill, but near the valley bottom, it may occasionally be seen in transverse sections to have apparently changed its direction, and this was, no doubt, due to the fact that the descending material would sometimes sweep toward one side of the valley and sometimes toward the other. ROCK STRUCTURES. I now proceed to deal with the structural features of the rocks of the Combe Martin area. It is situated on the northern or under limb of the great Devon synclinorium, which is inclined downward to the south or a little west of south and at the same time pitches slightly to the west. The dip is therefore approximately to the south. It is frequently, however, modified in direction as well as in amount by minor folding or by the proximity of faulting. The amount of the dip usually varies from about 30° to 50°, but naturally, when folding occurs, it may be anything. A strong cleavage is developed in argillaceous rocks, which has usually the same strike as the stratification, but a greater dip to the south, varying from 50° to nearly vertical. In cal­ careous slates the cleavage is seen to consist of a series of small lenticles, a few millimetres or centimetres in extent, formed by innumerable small thrusts from the south. As shown by Pro­ fessor Marr (1888) a similar lenticular structure on a somewhat larger scale, also due to over-thrusting, can sometimes be observed. Overfolding on a small scale is fairly common except in the harder and more massive grits and in the argillaceous beds. The latter, when in the condition of clay, seem to have flowed under heavy directed pressure-that is to say the excess of pressure in one direction over that in another-like a highly viscous liquid. As demonstrated by A. S. E. Ackermann,* if the directed pressure exceeds a certain maximum, it creates. a condition of apparent fluidity, t and this seems to have been at one time the case in these rocks and to have resulted in a flow obliquely upwards to the north, and compression

• T'''.... Soc. 0/ Engineers, 1919, pp. 37-107; 1920, pp. 196-249; 19U, pp. 87-130. t General or uniform pressure 011 the other hand appears to have a colltrary e1fect. STRUCTURE OF COUNTRY ROUND COMBE MARTIN. 219 at right angles. Harder bands (whether calcareous or arenaceous) that were intercalated in the argillaceous rocks and were incapable of flow under the conditions that prevailed were thrown into undulations. The undulations in anyone such band appear to be independent of those in the others and to have been determined by its thickness. Thus, a calcareous and siliceous band, a centimetre or so thick, was thrown into numerous over-folds with about the same radius. The "crinkled" or "wriggly" bands so formed are very constant in their occurrence and are useful for local correlation. A study of the crinkled band on the west side of Newberry Beach showed that the slates had contracted at right angles to the cleavage by more than S3 per cent. of their original thickness. Limestones, or occasionally arenaceous bands, a foot thick. are thrown into correspondingly larger folds, locally known, as we have seen, as" fossil trees," and these again are apparently also characteristic of particular horizons. For instance, the tree-like folds in the quarry south-west of the road near the land­ slip, those in the quarry below the road further to the south­ east, and those at the foot of the descent to the beach east of David's Hole, are all, I believe, at the same horizon. Massive limestones, on the other hand, may be thrown into folds ten feet or more in diameter, as is the case near David's Stone. Grits may be similarly affected, as in the Little Hang­ man Grits and in the Lester Beds in Wild Pear Beach, Lester Point, Combe Martin Beach, and the Warren. In a few cases slates and intercalated limestones are folded as a whole, as in the foreshore of Wild Pear Beach and Combe Martin Beach. The bisecting planes of the folds, like that of the Great Syn­ clinorium,dip to the south, parallel to the cleavage. Occasion­ ally, as in the cove north-west of Golden Cove, minor over-folds occur in the course of major folds, thus converting them into small anticlinoria and synclinoria. These minor folds, too, have their bisecting planes parallel to the plane of the cleavage. Where an anticline occurs it is always accompanied by a syncline to the north of it, but typical isoclinal folding, con­ sisting of a number of such folds, is only of local occurrence. The axes of the folds generally pitch to the westward at angles of from ISO to 30°, but I know of two exceptions where the pitch is to the east. One is in Ilfracornbe, west of the Harbour ; and the other is the small synclinorium to the north-west of Gol­ den Cove. The occurrence of anticlines and synclines and their pitch to the west are often well seen in the level foreshore carved out by marine erosion. This is the case with those above referred to in Combe Martin Beach and Wild Pear Beach. There is a slight tendency for the dip to diminish III immediate proximity of the sea, but it is only at and to 220 J. W. EVANS, the east of the Foreland that it is actually reversed. In the neighbourhood of the pre-Triassic valleys. found further east, it is apt to swing round towards them. The faulting in the area now under consideration is of an interesting character. The most remarkable feature is the occur­ rence of innumerable faults, striking between west and north. The prevalent direction appears to vary with the locality. For instance, in the faults between the Little Hangman and the neighbourhood of Combe Martin Beach. the strike is nearer north than west. Thence, to Oakestor Bay, it is nearly north-west. and beyond that for some distance it is roughly west-north-west. In practically every case the hade is to the south-west and the strata are apparently shifted to the north-west on that side. As the dip is to the south this apparent movement could be explained either by a down-throw on the south-west or a heave to the north-west. Whenever slickensides are present they slant downwards towards the north-west at angles which are usually within the limits of ISO and 30° with the horizontal. but it may vary in different striations even in the same fault and at the same place. It will be seen therefore that the outh­ west side has moved both downwards towards the south-west and forward towards the north-west. It must be remembered. however, that all movements are not recorded by slickensides and it is possible, indeed probable, that the earlier movements were simply downwards to the south-west on the fault plane. Where anticlines and synclines are present their relative positions on the two sides of the fault afford additional evidence of the character of the movement. It is interesting to note that a similar prevalence of north­ west and south-east faulting is revealed in the Cornish mines and that there is in the majority of cases a shift to the right, that is to say. if the fault, or cross-course as it is termed, is approached from the east. the shift is to the north-west. and if from the west, to the south-east, which agrees with what is found to be the case near Combe Martin. I am aware of no evidence from slickensides or folds of the actual direction of the movement in Cornwall. The hade in these Cornish faults is said by Carne to be to the east, which is on the face of it improbable. Unfortun­ ately, full advantage does not appear to have been taken in the past of the facilities afforded by mining operations. and the statements of the older writers sometimes contradict each other.... The great faults in the south of East Cornwall also strike to the west of north and cause a shift of a similar character. and

• J. Carne, U Relative Age of the Veins of Cornwall:' T,. R.G'.S. Corn., vol. ii., 1822 ; W. ]. Henwood, " The Heave of a Copper Lode by a Flucan Vein," Trans. Roy. Geol, Soc., Corn.., 1827, p. 329 ; II On some Intersections of Veins in the Mines of Dolcoath and Wheal Prudence," Rep. Brit. Assoc., for 1837, Tr, 0/ Sections, p. 74; D. A. MacAlister, .. The Mineral Lodes of the Camborne District," Sum-me Progr, Geol, Sure. for 19°2, pp. 39, 40, 19°3; 41 Tin and Copper Deposits of Camborne," Tr, R.G.S. Corn., vol. xii., 1904; Mem. Geol.Surv., sheet 352 (Falmouth and Camborne), pp. 157. 160. 1906; sheets 351 and 358 (Lands End) pp. 99. 100, 1907. STRUCTURE OF COUNTRY ROUND COMBE MARTIN. 22r there are a number of faults in South Ireland and South Wales presenting similar features. Others have been described byJ. W. Dudley Robinson in the Boulonnais" and by E. Kayser east of the Dill between Ehrungshausen and Hohensolms.t The faults that come next in importance are those which strike in a north-east and south-west direction and appear to be of earlier date. These hade, as a rule, to the south-east and the downthrow, so far as my observations go, is likewise on that side. They have usually a greater effect on the dip of the adjoining strata than the north-west and south-east faults, as if the strata were softer and more flexible when the movement took place. In their immediate neighbourhood the argillaceous strata have a marked dip to the south-east, while in the case of the north-west faults the tendency to dip to the south-west is less marked and more local. lt will be noticed that each of these directions of faulting is oblique to the general strike of the beds, but that the apparent downthrow is in one case south-west and in the other south­ east, and therefore, in both cases more in the direction of the southerly dip than against it. The downthrow, being therefore in both cases usually on the dip side, tends to cut out the outcrop of strata. There is every reason to believe that by this means the suppression of particular beds is very common in North Devon. I have verified this in a number of cases and as similar relations appear to exist on a large scale it is possible that considerable thicknesses of beds may sometimes be omitted. For instance, in the Ilfracombe Beds, near Rodhuish, in West Somerset, an important bed of red coral limestone occurs, presenting very close resemblance to well-known limestones in South Devon. lt is remarkable that it should be found in no other locality in North Devon. This is probably not because it was never laid down, but because it has been suppressed by faulting. The fault that is believed to occur at Heddon's Mouth has probably, as we have seen, an unusual strike practically due north and south and a considerable downthrow to the east. In spite of the contentions of Jukes and Hicks, there is no evidence that the strike-faults play any very important role in determining the structure of North Devon. Normal faults, more or less parallel to the strike, are not, however, very uncommon. The hade and downthrow may be either to the north or to the south, but more frequently the latter, so that they have more or less the same general direction as the bedding and cleavage. They are sometimes associated with pneumatolytic action, resulting in the formation of metal-

• .. The Geology of the Boulonnais," Quart. Jour» Geol, Soc., vol, lxxvi., pp. "~-236, 1921. t jaht';. d" k, p. Geol, Landesanstalt, 1900, pl. I. 222 J. W. EVANS, liferous lodes. Among the minerals observed are pyrite, mar­ casite, argentiferous galena, zinc-blende, tetrahedrite, copper pyrites, chessylite, malachite, millerite and stibnite. The lodes occur chiefly in the Lester Beds, but are not confined to them. They correspond to the lodes of Cornwall which are usually parallel to the strike of the cleavage and stratification and therefore strike east and west or deviate a few degrees on either side. In Cornwall they dip sometimes to the north and sometimes to the south, but more usually to the former. They are frequently faulted by the north-west and south-east cross courses. I have already alluded to the fact that minor over-thrusting from the south gives rise to lenticular structure. Over-thrusts on a small scale also occur in a few instances in the apices of over-folds, but I am aware of no case in which older strata are thrust for any considerable distance over younger beds. The contrast in this respect with South Devon, the Boulonnais and Belgium is very remarkable. It is due, I believe, to the presence of sufficient up-standing resistance in North Devon to prevent 'Over-riding. This resistance may be present either because the rocks are at the time of the movement buried sufficiently deeply, or because the locality is situated on the under-limb of a synclinorium or, in other words, the over-limb of an anti­ clinorium or major over-fold, so that the over-fold itself supplies the resistance. South Devon, on the other hand, was on the southern or over-limb of the synclinorium and there was little or no resistance on the north to prevent over-thrusting. On the north side of Combe Martin Beach there is evidence from slickensides of a movement in the plane of stratification, not directly downwards but with a slight inclination from west above to east below. In this case the movement may have been a backward slip, rather than a thrust forward of the super­ incumbent stratum. In Berry's Quarry, on the other hand, the slickensides appear to point to a horizontal movement along the stratum at right angles to the dip. I have observed no indication of any special occurrence of faults parallel to the coast of the Bristol Channel with a down­ throw towards the sea, such as are found on many coast-lines, but it is possible that such may occur beneath the sea. The faulting on the margin of the depressions on either side of the Quantocks lends some support to this supposition. The modifica­ tion of the dip towards the Bristol Channel and other pre-Triassic valleys, to which allusion has already been made, may be con­ nected with faulting of this description. Somewhat similar relations are met with on the north side oHthe Bristol Channel. Another phenomenon that deserves mention is the creasing that sometimes occurs of the cleavage surface of the slate. The STRUCTURE OF COUNTRY ROUND C01l1BE ]'[ARTIN. 223

<:reases are roughly parallel to the dip. The surface has a long gentle upward slope from west to east for a width of from one to twelve inches, followed by a short steep downward slope perhaps half-an-inch in width. This is repeated a number 01 times, adjoining slopes being separated by well-marked edges so that they resemble the laminations on the surface of plagio­ clase felspar. They give the impression of being due to pressure from the west. Just east of David's Hole there are similar structures on a larger scale. A few words may be added on the subject of joints. These, as a rule, are either parallel to the strike, dipping roughly at right angles to the dip of the stratification and cleavage. or they strike approximately north-east and south-west and dip steeply towards the south-east. The gash veins in limestones or sand­ stones, that is to say fissures filled with carbonate of lime or quartz, have, as a rule, the same orientation as the latter joints.

GEOLOGICAL HISTORY. We are now in a position to form a consistent idea of the ,geological history of North Devon and the areas adjoining it, so far as evidence is available for the purpose. For the purpose -of explaining the course of events, it is impossible to avoid referring to the different regions by the names which they now bear ; but it is most important to remember that the distance between any two points, and their orientation relatively to one another, have varied in a veryconsiderable degree in the course of geological time. During the Devonian Period the coast-line between the land -on the north and the sea on the south seems to have repeatedly shifted its position. The former idea of a barrier or ridge. separ­ .ating Old Red Sandstone from marine conditions, had as little reality as the Crete du Condroz that was supposed to separate the Bassin de Namur from that of Dinant during the same period. The surface seems to have slanted gently and con­ tinuously downward from the foot of a mountain region in Wales southward to what is now the English Channel. To the north was an area where desert conditions prevailed, varied only by rivers that flowed after heavy rainfall and by sheets of water that dried up in times of drought. In those days, when vegeta­ tion appears to have been still, at least, semi-aquatic. and to have required an ample supply of moisture in the form of standing water or marshy ground through the whole or a great part of the year, desert conditions were consistent with ample rainfall, for wherever there is no vegetation to protect the surface from the action of sun and frost and wind a desert will come into existence. 224 J. W. EVANS, Three times, commencing at the close of the Silurian Period, the waters of the ocean retired to the southward, and three times they advanced again and recovered much of the lost ground. At the beginning of the Devonian Period the sea lay to the southward, and in South Devon and South Cornwall the water in which the materials of the Dartmouth Slates were laid down was, at times, sufficiently salt for a few marine organisms to· exist, but in North Cornwall there is no trace of marine influence and this is also the case with the Foreland Grits of North Devon and West Somerset. Subsequently, in Meadfoot and Lynton times, the shore-line shifted to the northward till it lay some­ where in the region that is now the Bristol Channel, for there is no proof that it ever reached the present southern coast of Wales. It then retired again southward and continental con­ ditions are evidenced in North Devon by the Hangman Grits, but the south of the county still remained under the sea. Upper Middle Devonian time (Givetian) was marked by an important and nearly world-wide advance of the sea, but once again there is no evidence that it ever extended to South Wales, though in the absence of any rocks that can be referred to that period it is impossible to make any definite assertion on the subject. Once more in the later Upper Devonian (Famennian) the sea retreated southward and continental conditions prevailed in North Devon, but the deposits laid down south of the syncline in Cornwall and South Devon were still marine, though they show evidence in their red and greenish colouring and arenaceous composition of the probable neighbourhood of a desert land area. The clos­ ing days of the Devonian period were marked by the recurrence of marine conditions, not only in North Devon but in western South Wales and in Ireland. Yet it would seem that at about the same time a progressive movement of elevation laid bare and open to erosion, first Cornwall and the South of Devon, and then the region to the northward that forms the subject of this. paper, and it was not till nearly the close of the period of deposition of the Carboniferous Limestone that the whole was. once more covered by the sea. It would appear that the accumulation during Devonian and early Carboniferous times of some 20,000 feet of sediments over a wide tract, sometimes. submerged beneath a shallow sea and sometimes exposed by its retreat, caused a continuous depression of the older rocks. on which they were laid down and a rise of the area to the southward where the material deposited was less. It was to this depression that the Devonian rocks of North Devon owe their preservation. All these changes were of a comparatively slow and peaceful character, though further north violent earth movements were taking place from time to time during a considerable portion of the Devonian Period, but at or towards the close of the Carbon- STRUCTURE OF COUNTRY ROUND COl\IBE MARTIN. 225 iferous Period a crustal storm of the greatest violence affected the area which is now the south-west of England. A powerful thrust from the southward urged a solid mass of ancient Archsean rocks, forming, by that time, an elevated plateau and occupying approximately the present position of the English Channel, against the low-lying and still unconsolidated Devonian and Carboniferous sediments to the north. It seems to have been forced upward over their margin and at the same time it drove them before it, crumpling them into innumerable folds, including the synclinorium of Central Devon, and in all probability other complex anticlines and synclines, extending over what is now the Bristol Channel. At the same time it impressed a strong cleavage on the softer beds. The result appears to have been the piling up of a great east and west mountain range to the northward of our area, from which the ground sloped down southward to the centre of the great Devon synclinorium. That this was the case is shown by the transport of material to which the Permian breccia of West Somerset and North Devon bears witness. This was; there is every reason to believe, a deposit strictly analogous to the Pleistocene Head which has already been described, and it, too, owed its existence to the prevalence of a severe climate and can be explained by the flow of loose material cemented by frozen moisture. As in the case of the Head it .must have had a general downward movement. It is therefore interesting to note that the breccia overlying the Ilfracombe Beds, near Rodhuish in West Somerset, contains rock fragments with Carboniferous fossils, which must, in all probability, have come from north of the Bristol Channel, and that the similar breccia at Tiverton in North Devon, overlying the Culm, contains Upper Devonian fossils that must have moved from at least eight miles to the northward and probably more. The direction of the pressure that caused the folding varied in different portions of the south-west of England. In the district we are considering it was from due south or a little west of south, but in some areas further to the west it was apparently from a direction somewhat east of south. An indirect result of the great thrust from the south was the intrusion of the granite laccolites of the south-west of England. Though the nearest of these, Lundy, is some 27 miles distant, at least at the surface, their influence is seen in the metalliferous veins and the mineralization, slight though it is in many cases, of the rocks. Tourmaline, for instance, occurs in one of the limestones, and sulphides are, as we have seen, not infrequent. The period of pressure from the south was followed by one of reaction in which the pressure was replaced by tension, at first in a southerly and afterwards in a south-easterly direction. This resulted in north-east and south-west fissures, some of which were filled with the materials dissolved in the water permeating PROC. GEOL. Assoc., VOL. XXXIII.. PART 3. 1922. 15 226 J. W. EVANS, the rocks, and largely derived from magmatic sources; while others became normal faults, down which the strata slipped towards the south or south-east. Towards the close of the Permian a condition of crustal tension developed in another direction. Perhaps, as a first step in the separation of the North American Continent from Europe and the evolution of the North Atlantic, the rocks sagged to the west-south-west and numerous faults, with a north-north­ west strike and a hade to the west-south-west were formed. This first movement was probably directly down the fault planes, but subsequently-long subsequently, I believe-a new source of weakness made itself felt further north and an important movement set in obliquely down the fault slopes, that is to say, downward to the south-west and forward to the north-west, but more especially the latter. At the same time there was a general tilt to the westward of the whole country shown in the westward pitch of the majority of folds and the westward tilting of the slickensides in planes of stratifica­ tion. The earlier movements on these fault planes probably took place in the closing phases of the Permian subsidence, but many of the fault movements affected the Trias and Lias, and there can be little doubt they were connected with the Tertiary disturbances which affected the whole of Europe. However this may be, the general result has been to give the rocks the present west-north-west trend parallel to the ancient Bristol Channel valley, though the actual local strike is, as has been stated, nearly east and west. In the course of these developments the mountain region to the north of Devon would, towards the end of the Permian Period, be given a strong inclination to the west or west-north­ west, with the result that powerful streams, following in their direction the lines of weakness in the anticlinorium region, would soon excavate a great longitudinal valley comparable to those of the Brahmaputra and the Indus in the Himalayas. It is impossible now to say where it had its source, but it was probably well to the east-south-east of Glastonbury or even Shaftesbury, and we can trace important tributary valleys such as that to the south-west of the Quantocks and another between North Hill, near Minehead, and Dunkerry Beacon. All these had been already excavated by early Triassic times, for they are partly filled with sediment of that and the Jurassic periods; but much of this has been since removed, restoring to a large extent the form of the Triassic valleys. The existence of the Bristol Channel valley and its tributaries must, either during its continuance as such, or when filled with comparatively soft sediments, have definitely affected the struc­ ture of the surrounding rocks. There would be a tendency STRUCTURE OF COUNTRY ROUND COMBE MARTIN. 227 in the rocks forming the valley sides to slide or settle down them with a movement akin to that of a landslip. At the time of the first excavation of the broad valley of the Bristol Channel there would be an upward compensatory movement towards isostatic stability, but this would be to a large extent arrested by the subsequent infilling. At the same time the consolidation of the infilling sediments would give them a downward sag eventuating in a synclinal structure or trough faulting. On a renewal of pressure from the south, such as occurred in early Tertiary times, one would expect these valleys to become regions of crustal weakness. The older rocks of the slopes would be pushed for­ ward, compressing the later included rocks and throwing them into secondary folds. At the same time the structure of the older rocks would be affected, their dip would be modified in the manner which has been indicated, and renewed movements would be induced along the already existing fault planes. In Pleistocene times the Bristol Channel was probably an alluvial plain through which the Severn and other rivers flowed sluggishly towards the sea. On the northern shores the glaciers crept down the valleys, while on the south, where a rigorous climate prevailed, the debris of weathering made its way still more slowly down the slopes and it now constitutes the Head, which is so characteristic of the south-western counties and renders accurate geological mapping so difficult. I have not found in the Old Red Sandstone hills of this neighbourhood any of the cirques which are so characteristic of those of West Somer­ set, and which, following Professor Bonney's views, I have ascribed to frost action (1914). Mr. Clement Reid believed that in late Neolithic times there was a rise in the sea-level compared to that of the land of about 60 feet. The coast phenomena near Combe Martin are quite consistent with this. Since the sea has been at its present level, a distance of some 250 feet appears to have been eroded from the cliffs by marine action. Where a stream entered the sea in a broad valley this was converted into an estuary, which is still only partly filled in by alluvium. Narrow valleys, on the other hand, have long since been raised by deposition to sea-level. In the case of smaller streams, like that of Sherry Combe, which. have a rapidly descending course, the cutting back by the sea has been so much faster than the cutting down by the river that the valley definitely" hangs" and the river descends in a waterfall from a deep notch in the cliffs. The course of the Combe Martin Valley stream is, however, at first sight, difficult to reconcile with the supposition that the land recently stood at a higher level, since it flows into the sea over a gentle rocky slope, in which no appreciable channel has been excavated. It is probable, however, that it originally 228 J. W. EVANS, entered the sea a few yards further east, where the shore is now composed entirely of detritus, without rocks in situ, and that its course has since been artificially modified. In conclusion I would express my conviction that although the general direction of the Bristol Channel and its former exten­ sion eastward is approximately parallel to the dominant trend of the rocks of North Devon and South Wales from east-south­ east to west-north-west, the actual coast-line is not determined, except to a very limited amount, either by the geological structure or by marine erosion. The rocks that resist denudation by the sea project, it is true, to a certain extent, but they terminate abruptly against faults which should have brought their out­ crop still further seaward. Their disappearance could be explained on the supposition that the country was already planed down more or less to its present level before the final north-west and south-east faulting carried the continuation of these rocks downward and forward under the sea. But if this were so, we should expect the same faults to be represented by conspicuous scarps on the land, but of this we can find no trace. We are accordingly left no alternative but to believe that the present coast-line of North Devon is mainly due to fluviatile and other subaerial erosion in late Permian or possibly early Triassic times, which defined the main outlines of the great valley now occupied by the Bristol Channel and its tributaries, and has only been modified in a minor degree by subsequent earth-movements or marine or terrestrial erosion.

EXCURSION TO COMBE MARTIN.

APRIL 12TH TO 20TH, EASTER, 1922.

REPORT BY J. W. EVANS, D.sc., F.R.S., Director of the Excursion.

PLATES 6 and 7.

FIFTY-FOUR members and friends took part in the excursion. Some reached Combe Martin on the evening of the rzth and others, taking advantage of the excursion tickets issued by the London and South-Western Railway, arrived the following even­ ing. The majority were accommodated in apartments in Wood­ lands Road and the remainder at the King's Arms.

APRIL 13TH. The party assembled on the Parade and, after a few words of explanation from the Director, proceeded to Broad Strand, a sandy beach about a mile to the nor th-west. Some reached it by boat and the others accompanied the Director up Newberry Hill and along a road abandoned since the recent landslip.