There are five districts in North-western Europe where the original widespread Tertiary lava-fields have been less extensively eroded than elsewhere, or at least where they have survived in larger and thicker masses. Whether or not each of them was an isolated area of volcanic activity cannot now be determined. Their several outflows of lava within the area of the British Isles may have united into one continuous volcanic tract, and their present isolation there may be due entirely to subterranean movements and denudation. There is a certain convenience, however, in treating the districts separately. They are—1. Antrim; 2. Mull, Morven and Ardnamurchan; 3. Small Isles; 4. Skye; 5. The Faroe Islands.

i. ANTRIM[230]

The largest of the basalt-plateaux of Britain is that which forms so prominent a feature in the scenery and geology of the North of Ireland, stretching from Lough Foyle to Belfast Lough, and from Rathlin Island to beyond the southern margin of Lough Neagh. Its area may be roughly computed at about 2000 square miles. But, as its truncated strata rise high along its borders, and look far over the surrounding low grounds, it must be regarded as a mere fragment of the original volcanic plain. It may be described as an undulating tableland, which almost everywhere terminates in a range of bold cliffs, but which, towards the centre and south, sinks gently into the basin of Lough Neagh. The marginal line of escarpment, however, presents considerable irregularity both in height and form, besides being liable to frequent local interruptions. It is highest on the west side, one of its crests reaching at Mullaghmore, in County Londonderry, a height of 1825 feet. It sinks down into the valley of the Bann, east of which it gradually ascends, forming the well-known range of cliffs from the Giant's Causeway and Bengore Head to Ballycastle. It then strikes inland, and making a wide curve in which it reaches a height of more than 1300 feet, comes to the sea again at Garron Point. From that headland the cliffs of basalt form a belt of picturesque ground southwards beyond Belfast, interrupted only by valleys that convey the drainage of the interior of the plateau to the North Channel. Above the valley of the Lagan the crest of the plateau rises to a height of more than 1500 feet.

Throughout most of its extent the basalt-escarpment rests on the white limestone or Chalk of Antrim, beneath which lie soft Lias shales and Triassic marls. Here and there, where the substratum of Chalk is thin, the action of underground water on the crumbling shales and marls below it has given rise to landslips. The slopes beneath the base of the basalt are strewn with slipped masses of that rock, almost all the way from Cushendall to Larne, some of the detached portions being so large as to be readily taken for parts of the unmoved rock. On the west side also, a group of huge landslips cumbers the declivities beneath the mural front of Benevenagh.

I have found some difficulty in the attempt to ascertain what was the probable form of surface over which the volcanic rocks of this plateau began to be poured out. The Chalk sinks below the sea-level on the north coast, but, in the outlier of Slieve Gallion, three miles beyond the western base of the escarpment, it rises to a height of 1500 feet above the sea. On the east side also, it shows remarkable differences of level. Thus, below the White Head at the mouth of Belfast Lough, it passes under the sea-level, but only 16 miles to the south, where it crops out from under the basalt, its surface is about 1000 feet above that level. If these variations in height existed at the time of the outpouring of the basalt, the surface of the ground over which the eruptions took place was so irregular that some hundreds of feet of lava must have accumulated before the higher chalk hills were buried under the volcanic discharges. But it seems to me that much of this inequality in the height of the upper surface of the Chalk is to be attributed to unequal movements since the volcanic period, which involved the basalt in their effects, as well as the platform of Chalk below it. Had the present undulations of that platform been older than the volcanic discharges, it is obvious that upper portions of the basalt-series would have overlapped lower, and would have come to rest directly on the Chalk. But this arrangement, so far as I am aware, never occurs, except on a trifling scale. Wherever the Chalk appears, it is covered by sheets of the lower and not of the upper of the two groups into which the Antrim basalts are divisible. We have actual proof of considerable terrestrial disturbance, subsequent to the date of the formation of the volcanic plateau. Thus, near Ballycastle, a fault lets down the basalt and its Chalk platform against the crystalline schists of that district. On the east side of the fault, the Chalk is found far up the slope, circling round the base of the beautiful cone of Knocklayd—an outlier of the basalt which reaches a height of 1695 feet (Fig. 263). The amount of vertical displacement of the volcanic sheets is here 700 feet.[231] Many other displacements, as shown by the mapping of my colleagues in the Geological Survey, have shifted the base of the escarpment from a few inches up to several hundred feet. Besides actual dislocations, the Antrim plateau has undergone some marked subsidences of which the most notable is that of Lough Neagh.[232]

It is evident, therefore, that the present position of the Chalk platform is far from agreeing with that which it presented to the outflow of the sheets of basalt. But, on the other hand, there can be no doubt that its surface at the beginning of the volcanic outbursts was not a level plain. It was probably a rolling country of low bare chalk-downs, like parts of the South-east of England. The Irish Chalk attains its maximum thickness of perhaps 250 feet at Ballintoy. But it is liable to rapid diminution. On the shore at Ballycastle about 150 feet of it can be seen, its base being concealed; but only two and a half miles to the south, on the outlier of Knocklayd, the thickness is not quite half so much. On the west side of the plateau also, there are rapid changes in the thickness of the Chalk. Such variations appear to be mainly attributable to unequal erosion before the overflow of the basalts. So great indeed had been the denudation of the Cretaceous and underlying Secondary formations previous to the beginning of the volcanic eruptions, that in some places the whole of these strata had been stripped off the country, so that the older platform of Palæozoic or still more ancient masses was laid bare. Thus, on the west side of the escarpment, the basalt steals across the Chalk and comes to rest directly upon Lower Carboniferous rocks.

The authors who have described the junction of the Chalk and basalts in Antrim have generally referred to the uneven surface of the former rock as exposed in any given section. The floor on which the basalt lies is remarkably irregular, rising into ridges and sinking into hollows or trenches, but almost everywhere presenting a layer of earthy rubbish made of brown ferruginous clays, mixed with pieces of flint, chalk, and even basalt.[233] The flints are generally reddened and shattery. The chalk itself has been described as indurated, and its flints as partially burned by the influence of the overlying basalt. But I have not noticed, at any locality, evidence of alteration of the solid chalk, except where dykes or intrusive sheets have penetrated it.[234] There can be no doubt that the hardness of the rock is an original peculiarity, due to the circumstances of its formation. The irregular earthy rubble, that almost always intervenes between the chalk and the base of the basalt, like the "clay with flints" so general over the Chalk of Southern England, no doubt represents long-continued subærial weathering previous to the outflow of the basalt. Even, therefore, if there were no other evidence, we might infer with some confidence from this layer of rubble, that the surface over which the lavas were poured was a terrestrial one. Here and there, too, we may detect traces of the subsidence of the basalt into swallow-holes dissolved in the chalk subsequent to the outflow of the basalt-sheets.

The Antrim plateau is not only the largest in the British Islands, it is also the most continuous and regular. It may be regarded, indeed, as one unbroken sheet of volcanic material, not disrupted by any such mountainous masses of intrusive rock as in the other plateaux interrupt the continuity of the horizontal or gently inclined sheets of basalt. Around its margin, indeed, a few outliers tower above the plains, and serve as impressive memorials of its losses by denudation. Of these, by much the most picturesque and imposing, though not the loftiest, is Knocklayd already referred to, which forms so striking a feature in the north-east of Antrim (Fig. 263).

The total thickness of volcanic rocks in the Antrim plateau exceeds 1000 feet; but, as the upper part of the series has been removed by denudation, the whole depth of lava originally poured out cannot now be told. A well-marked group of tuffs and clays, traceable throughout a large part of Antrim, forms a good horizon in the midst of the basalts, which are thus divisible into a lower and upper group (Fig 264).

The Lower Basalts have a thickness of from 400 to 500 feet. But, as already mentioned (p. 194), they die out in about six miles to no more than 40 feet at Ballintoy. They are distinguished by their generally cellular and amygdaloidal character, and less frequently columnar structure. The successive flows, each averaging perhaps above 15 feet in thickness, are often separated by thin red ferruginous clayey partings, sometimes by bands of green or brown fine gravelly tuff. The most extensive of these tuff-bands occurs in the lower part of the group at Ballintoy, and can be traced along the coast for about five miles. In the middle of its course, near the picturesque Carrick-a-raide, it reaches a maximum thickness of about 100 feet and gradually dies out to east and west. The neck of coarse agglomerate at Carrick-a-raide, is doubtless the vent from which this mass of tuff was discharged (see Fig. 301). Owing to the thinning out of the sheets of basalts, as they approach the vent, the tuff comes to rest directly on the Chalk, and for some distance westwards forms the actual base of the volcanic series.[235] Occasional seams of carbonaceous clays, or of lignite, appear in different horizons among the basalts. Beneath the whole mass of basalt, indeed, remains of terrestrial vegetation here and there occur. Thus, near Banbridge, County Down, a patch of lignite, four feet ten inches thick, underlies the basalt, and rests directly on Silurian rocks. Such fragmentary records are an interesting memorial of the wooded land-surface over which the earliest outflows of basalt spread.

In looking at the great basalt-escarpments of Antrim, the Inner Hebrides or the Faroe Islands, and in following with the eye the successive sheets of lava in orderly sequence of level bands from the breaking waves at the base to the beetling crest above, we are apt to take note only of the proofs of regularity and repetition in the outflows of molten rock and to miss the evidence that these outflows did not always rapidly follow each other, but were separated by intervals of varying, sometimes even of long duration. One of the most frequent and conspicuous proofs of such intervals is to be found in the red layers or partings above referred to which, throughout all the basalt-plateaux, so commonly intervene between successive sheets of basalt. These red streaks cannot fail to arrest the eye on the coast-precipices where by their brilliant contrast of colour, they help to emphasize the bedded character of the whole volcanic series.

Examined more closely, they are found to consist of clay or bole which shades into the decomposed top of the bed whereon it lies, and is usually somewhat sharply marked off from that which covers it. This layer has long, and I think correctly, been regarded as due to the atmospheric disintegration of the surface of the basalt on which it rests, before the eruption of the overlying flow. It varies in thickness from a mere line up to a foot or more, and it passes into the tuffs and clays which are sometimes interposed between the sheets of basalts. It may be looked upon as probably furnishing evidence of the lapse of an interval sufficiently extended to permit a considerable subserial decay of the surface of a lava-sheet before the outflow of the next lava. But an attentive study of the plateaux discloses other and even more remarkable indications that the pauses between the consecutive basalt-beds were frequently so prolonged as to allow extensive topographical changes to be made in a district. Nowhere is the long duration of some of these intervals more impressively taught than in the central zone of sedimentary strata in Antrim.

This persistent group of tuffs, clays, and iron-ore is generally from 30 to 40 and sometimes as much as 70 feet thick. From the occurrence of the ore in it, it has been explored more diligently in recent years than any other group of rocks in the district, and its outcrop is now known over most of the plateau. The iron-ore bed varies from less than an inch up to 18 inches in thickness, and consists of pisolitic concretions of hæmatite, from the size of a pea to that of a hazel nut, wrapped up in a soft ochreous clayey matrix.[236] Where it is absent, its place is sometimes taken by an aluminous clay, worked as "bauxite," which has yielded stumps of trees and numerous leaves and cones. Beneath the iron-ore or its representative, lies what is called the "pavement,"—a ferruginous tuff, 8 to 10 feet thick, resting on "lithomarge,"—a lilac or violet mottled aluminous earth sometimes full of rounded blocks or bombs of basalt. The well-known horizon for fossil plants at Ballypallidy is a red tuff in this zone. The section of strata between the two basalt-groups at this locality may serve as an illustration of the nature and arrangement of the deposits.[237]

In some of the Ballypallidy tuffs the most frequent lapilli are pieces of green and brown glass, which Mr. Watts compares with the pitchstone of Sandy Braes, though rarely containing phenocrysts as that rock does. He has found also in these strata a smaller proportion of lithoidal rhyolites and occasionally fragments of basic rock.

The pale and coloured clays that occur in this marked sedimentary intercalation have doubtless been produced by the decomposition of the volcanic rocks and the washing of their fine detritus by water. Possibly this decay may have been in part the result of solfataric action. From true bauxite or aluminium-hydrate, the sediments vary in composition and specific gravity and pass into aluminous silicates and iron-ores. They seem to indicate a prolonged interval of volcanic quiescence when the lavas and tuffs already erupted were denuded and decomposed.[238]

The area over which this interesting series of stratified deposits now extends is obviously much less than it was originally. It has indeed been so reduced by denudation into mere scattered patches that it probably does not exceed 170 square miles. But the group can be traced from Divis Hill, near Belfast, to Rathlin Island, a distance of 50 miles, and from the valley of the Bann to the coast above Glenarm, more than 20 miles. There can be little doubt that it was once continuous over all that area, and that it probably extended some way further on each side. If the so-called Pliocene clays of Lough Neagh be regarded as parts of this group of strata, its extent will be still further increased. Hence the original area over which the iron-ore and its accompanying tuffs and clays were laid down can hardly have been less than 1000 square miles. This extensive tract was evidently the site of a lake during the volcanic period, formed by a subsidence of the floor of the lower basalts. The salts of iron contained in solution in the water, whether derived from the decay of the surrounding lavas or from the discharges of chalybeate springs, were precipitated as peroxide in pisolitic form, as similar ores are now being formed on lake-bottoms in Sweden. For a long interval, quiet sedimentation went on in this lake, the only sign of volcanic energy during that time being the dust and stones that were thrown out and fell over the water-basin, or were washed into it by rains from the cones of the lava-slopes around.

It may here be remarked that the tendency to subsidence in the Antrim plateau seems to have characterized this region since an early part of the volcanic period. The lake in which the deposits now described accumulated was entirely effaced and overspread by the thick group of upper basalts. But long after the eruptions had ceased, a renewed sinking of the ground gave rise to the sheet of water which now forms Lough Neagh.[239]

Nowhere else among the Tertiary basalt-plateaux of Britain has any trace been found of so marked and prolonged a pause in the volcanic activity as is indicated by the Antrim zone of tuffs and clays. Throughout the Inner Hebrides, indeed, numerous intercalations of sedimentary material occur among the basalts, but these consist mainly of tuffs and volcanic conglomerates with less frequent shales and coal-seams, and they never suggest so distinct and lengthened an interval as is indicated by the Antrim deposit.

It is not improbable that this interval was marked by the outbreak of rhyolitic eruptions somewhere in the region. The abundance of rhyolite fragments in some of the tuffs is striking evidence that acid rocks were in one way or other brought to the surface at this time. At Glenarm one of the members of the stratified series is a marked rhyolitic conglomerate, composed of rounded pebbles of a rock not unlike the well-known rhyolite of Tardree and Carnearny. These fragments, obviously of local origin, must either have been derived from a surface of acid rock laid bare by denudation, or from rhyolite ejected in lapilli or poured out in streams. I formerly believed that all the Antrim rhyolites had been injected into the basalts after the close of the plateau-period. But the proved abundance and wide extent of the rhyolitic detritus among the sediments associated with the iron-ore point to a possible outflow of acid lavas with accompanying tuffs during the sedimentary interval between the two groups of basalt. The characters of the Antrim rhyolites, however, will be more particularly discussed in Chapter xlvii., in connection with the acid rocks of the Tertiary volcanic series.

Immediately above the iron-ore of Antrim, or separated from it in places by only a few inches of tuff, comes the group of Upper Basalts, which varies up to 600 feet in thickness, though as the upper portion has been everywhere removed by denudation, no measure remains of what may have been the original depth of the group. The general character of these basalts is more frequently columnar, black and compact, and with fewer examples of a strongly amygdaloidal structure than in the lower group. But this distinction is less marked in the south than in the north of Antrim, so that where the intervening zone of tuffs and iron-ore disappears, no satisfactory line of division can be traced between the two groups of basalt. The occurrence of that zone, however, by giving rise to a hollow or slope, from which the upper basalts rise as a steep bank or cliff, furnishes a convenient topographical feature for mapping the boundary of these rocks. Among the upper basalts, also, there is perhaps a less frequent occurrence of those thin red partings of bole between successive flows, so conspicuous in the lower group. But the flows are not less distinctly marked off from each other. Nowhere can their characteristic features be better seen than along the magnificent range of cliffs from the Giant's Causeway eastwards. The columnar bed that forms the Causeway is the lowest sheet of the upper group, and may be seen resting directly on the zone of grey and red tuffs. It is about 60 or 70 feet thick; and, while perfectly regular in its columnar structure at the Causeway and the "Organ," assumes further eastward the confusedly starch-like arrangement of prisms already referred to. But in the great cliff section of the "Amphitheatre," the more regular structure is resumed, the bed swells out to about 80 feet in thickness, and columns of that length run up the face of the precipice, weathering out at the top into separate pillars, which, perched on the crest of an outstanding ridge, are known as the "Chimneys." The basalt-beds that succeed the lowest one are each only about 10 to 15 feet thick (Fig. 265).

Between the successive sheets of the Upper Basalts thin seams of red ferruginous clay though, as I have said, less frequent perhaps than in the lower group, continue to show that the intervals between successive eruptions were of sufficient duration to admit of some subærial decay of the surface of a lava before the outflow of the next bed. Occasional thin layers of tuff also, and even of pisolitic iron-ore, have been observed among these higher basalts. But the most interesting and important intercalations are inconstant seams of lignite. One of the most conspicuous of these lies immediately above the basalt of the "Causeway," where it was long worked for fuel, and was found to be more than six feet thick. But it is quite local, as may be seen at the "Organ" over which it lies, having a thickness of only 12 inches and rapidly dying out so as to allow the basalts above and below it to come together. The removal of the upper portion of the basalts by denudation has destroyed the records of the latest part of the volcanic history of the Irish plateaux.

It is obvious that nowhere in Antrim does any trace exist of a central vent or cone from which the volcanic materials were discharged. There is no perceptible thickening of the individual basalt-sheets, nor of the whole series in one general direction, in such a manner as to point to the site of some chief focus of eruption. Nor can we place reliance on the inclination of the several parts of the plateau. I have pointed out that the varying dip of the beds must be attributed mainly to post-volcanic movements, or at least to movements which, if not later than all the phases of volcanic action, must have succeeded the outpouring of the plateau-basalts. There has been a general subsidence towards the central and southern tracts now occupied by the valley of the Bann and Lough Neagh. But nowhere in the depression is there any trace of the ruins of a central cone or focus of discharge.

The Antrim plateau, in these respects, resembles the others. But as has already been remarked, it differs from them in one important particular. It has nowhere been disrupted by huge bosses of younger rocks, such as have broken up the continuity of the old lava-fields further north. Yet it also is not without its memorials of younger protrusions. It contains not a few excellent examples of true volcanic vents, and, as above stated, it includes some small acid bosses that may represent the great protrusions of the Inner Hebrides, and may have been connected with superficial outflows of rhyolitic lava and showers of rhyolitic tuff.

ii. MULL, MORVEN AND ARDNAMURCHAN

This plateau covers nearly the whole of the island of Mull, embraces a portion of Morven on the Argyleshire mainland, and, stretching across Loch Sunart, includes the western part of the peninsula of Ardnamurchan (Map VI.). That these now disconnected areas were once united into a continuous lava-field which extended far beyond its present limits is impressively indicated by their margin of cliffs and fringe of scattered islands and outliers. The plateau went west, at least, as far as the Treshnish Isles, which are composed of basalt. On its eastern border, a capping of basalt on the top of Beinn Iadain (1873 feet) in Morven, and others further north, prove that its volcanic sheets once spread into the interior of Argyleshire (Fig. 266). On the south, its fine range of lofty cliffs, with their horizontal bars of basalt, bear witness to the diminution which it has undergone on that side; while, on the north, similar sea-walls tell the same tale. Not only has it suffered by waste along its margin, it has also been deeply trenched by the excavation of glens and arms of the sea. The Sound of Mull cuts it in two, and the mainland portion is further bisected by Loch Sunart, and again by Loch Aline. The island of Mull is so penetrated by sea-lochs and divided by deep valleys that a comparatively slight depression would turn it into a group of islands. But, besides its enormous denudation, this plateau has been subjected to disruption, and perhaps also to subsidence, from subterranean movements. In the southern portion of the island of Mull it has been broken up by the intrusion of large bosses and sheets of gabbro, and by masses as well as innumerable veins of various granitoid and felsitic rocks. In Ardnamurchan, it has suffered so much disturbance from the same cause that its original structure has been almost obliterated over a considerable area. Moreover, it has been dislocated by many faults, by which different portions have been greatly shifted in level. The most important of these breaks is one noticed by Professor Judd, and visible to every tourist who sails up the Sound of Mull. It traverses the cliffs on the Morven side, opposite Craignure, bringing the basalts against the crystalline schists, and strikes thence inland, wheeling round into the long valley in which Lochs Arienas and Teacus lie. On its western side, the base of the basalt-series is almost at the sea-level; on its eastern side, that platform rises high into the outliers of Beinn na h-Uamha (1521 feet) and Beinn Iadain. The amount of displacement is probably more than 1000 feet. Many other minor faults in the same district show how much the crust of the earth has been fractured here since older Tertiary time.

A little to the west of Mull, and belonging originally to the same plateau, lies the isle of Staffa, the famous columnar basalts of which first attracted the attention of travellers, and gave to the Tertiary volcanic rocks of Scotland their celebrity (Fig. 266a).

In spite of the extent to which it has suffered from denudation and subterranean disturbance, and indeed in consequence thereof, the Mull plateau presents clear sections of many features in the history of the basalt-outflows and of the subsequent phases of Tertiary volcanic action which cannot be seen in the more regular and continuous tableland of Antrim. Moreover, it still possesses in its highest mountain, Ben More (3169 feet), a greater thickness, and probably a higher series, of lavas than can now be seen in any other of the plateaux.

The difficulties, already referred to in regard to Antrim, of tracing the probable form of ground on which the volcanic eruptions began, are even greater in the case of the Mull plateau. We can dimly perceive that the depression in the crystalline rocks of the Highlands which had, from at least the older part of the Jurassic period, stretched in a N.N.W. direction along what is now the western margin of Argyleshire, lay beneath the sea in Jurassic time, and was then more or less filled up with sedimentary deposits. The hollow appears thereafter to have become a land-valley, whence the Jurassic strata were to a large extent cleared out by denudation before its subsequent submergence under the sea in which the upper Cretaceous deposits accumulated. Professor Judd has shown that relics of these Cretaceous strata appear on both sides of the plateau from under the protecting cover of basalt-sheets. But, before the volcanic eruptions began, the area had once again been raised into land, and the youngest Secondary formations had been extensively eroded.

In their general aspect the basalts of Mull agree with those of Antrim, and the circumstances under which they were erupted were no doubt essentially the same. But considerable differences in detail are observable between the succession of rocks in the two areas. When I first visited the island in 1866, the only available maps, with any pretensions to accuracy, were the Admiralty charts; but, as these do not give the interior except in a generalized way, it was difficult to plot sections from them, and to arrive at satisfactory conclusions as to the thickness of different groups of rock. Accordingly, as the successive nearly flat flows of basalt can be traced from the sea-level up to the top of Ben More, I contented myself with the fact that the total depth of lava-beds in Mull was at least equal to the height of that mountain, or 3169 feet. The publication of the Ordnance Survey Maps now enables us to make a nearer approximation to the truth. From the western base of the magnificent headland of Gribon, the basalts in almost horizontal beds rise in one vast sweep of precipice and terraced slope to a height of over 1600 feet, and then stretch eastwards to pass under the higher part of Ben More, at a distance of some eight miles. They have a slight easterly inclination, so that the basement sheets seen at the sea-level, at the mouth of Loch Scridain, gradually sink below that level as they go eastward. It is not easy to get a measurement of dip among these basalts, except from a distance. If we take the inclination at only 1°, the beds which are at the base of the cliff on the west, must be about 700 feet below the sea on the line of Ben More, which would give a total thickness of nearly 3900 feet of bedded lava below the top of that mountain. We shall not probably overestimate the thickness of the Mull plateau if we put it at 3500 feet.

The base of the volcanic series of Mull can best be seen on the south coast at Carsaig, and at the foot of the precipices of Gribon. As already stated, it is there found resting above Cretaceous and Jurassic rocks. The lowest beds are basalt-tuffs, of the usual dull green colour. They are in places much intermingled with sandy and gravelly sediment, as if the volcanic debris had fallen into water where such sediment was in course of deposition. One of the most interesting features, indeed, in this basement part of the series, is the occurrence of bands of non-volcanic material which accumulated after the tuffs and some of the lavas had been erupted, but before the main mass of basalts. Those at Carsaig include a lenticular bed, 25 feet thick, of rolled flints, which, with some associated sandy bands, lies between sheets of basalt. On the opposite side of the promontory is the well-known locality of Ardtun, from which the first land-plants in the volcanic series were determined. The actual base of the basalts is not there seen, being covered by the sea. The "leaf-beds," with their accompanying sandstones, gravels, and limestone, lie upon a sheet of basalt, which in some parts is exceedingly slaggy on the top, passing down into a black compact structure, and assuming at the base of the cliff a columnar arrangement, with the prisms curved and built up endways towards each other. Some of the gravels exceed 30 feet in thickness, and consist of rolled flints, bits of chalk, and pieces of basalt and other basic igneous rocks. But some of their most interesting ingredients are pebbles of sanidine lavas, which have been recognized in them by Prof. G. Cole.[241] No known protrusions of such lavas occur anywhere beneath or interstratified with the plateau-basalts of this district. As will be afterwards shown, all the visible acid rocks, the geological relations of which can be ascertained, are here of younger date than these basalts. I am disposed to regard the fragments found in the Ardtun conglomerates as probably derived from some of the basalt-conglomerates of the plateau, in which fragments of siliceous igneous rocks do occur. Though there is no evidence that any lavas of that nature were here poured out at the surface before or during the emission of the basalts, the contents of these fragmental volcanic accumulations suggest that such lavas, already consolidated, lay at some depth beneath the surface, and that fragments were torn off from them during the explosions that threw out the materials of the basalt-conglomerates to the surface.

The succession of strata at the Ardtun headland varies considerably in a short distance, some of the sedimentary deposits rapidly increasing or diminishing in thickness. The section as measured by Mr. Starkie Gardner is as follows[242]:—

Mr. Starkie Gardner has called attention to the extraordinarily fresh condition of the vegetation in some of the layers of the Ardtun section. One of the leaf-beds he has found to be made up for an inch or two of a pressed mass of leaves, lying layer upon layer, and retaining almost the colours of dead vegetation. Among the plants represented is a large purple Ginkgo and a fine Platanites, one leaf measuring 15¹/₂ inches long by 10¹/₂ broad. The characteristic dicotyledonous leaves at this locality possessed relatively large foliage.[243]

To the early observations of Macculloch we are indebted for the record of an interesting fact in connection with the vegetation of the land-surface over which the first lava-flows spread. He figured a vertical tree trunk, imbedded in prismatic basalt, and rightly referred it to some species of fir.[244] This relic may still be seen under the basalt precipices of Gribon. Mr. Gardner found it to be "a large trunk of a coniferous tree, five feet in diameter, perhaps Podocarpus, which has been enveloped, as it stood, in one of the flows of trap to the height of 40 feet. Its solidity and girth evidently enabled it to resist the fire, but it had decayed before the next flow passed over it, for its trunk is a hollow cylinder filled with debris, and lined with the charred wood. A limb of another, or perhaps the same tree, is in a fissure not far off."[245]

At different levels in the volcanic series of Mull, beds of lignite and even true coal are observable. These seem to be always mere lenticular patches, only a few square yards in extent. The best example I have met with lies among the basalts near Carsaig. It is in part a black glossy coal, and partly dull and shaly. Some years ago it was between two and three feet thick, but now, owing to its having been dug away by the shepherds, only some six or eight inches are to be seen. It lies between two basalt-flows, and rapidly disappears on either side.

More frequent than these inconstant layers of fossil vegetation are the thin partings of tuff and layers of red clay, sometimes containing iron-ore, which occur at intervals throughout the series between different flows of basalt. But even such intercalations are of trifling thickness, and only of limited extent. The magnificent precipices of M'Gorry's Head and Gribon expose a succession of beds of columnar amorphous and amygdaloidal basalt, which must attain a thickness of at least 2500 feet, before they are overlain by the higher group of pale lavas in Ben More. On the east side of the island, thin tuffs and bands of basalt-conglomerate occur on different horizons among the bedded basalts, from near the sea-level up to the summit of the ridge which culminates in Beinn Meadhon (2087 feet), Dùn-da-Ghaoithe (2512 feet), and Mainnir-nam-Fiadh (2483 feet). Reference has already been made to the remarkably coarse character of some of the breccias intercalated among the basalts in this part of Mull, and to the enormous dimensions of some of the masses of mica-schist and quartzite which have been carried up from a depth of 2000 feet or more by volcanic agency (see ante, p. 196, and Fig. 262).

Above the ordinary compact and amygdaloidal basalt comes the higher group of pale lavas already referred to as forming the uppermost part of Ben More, whence it stretches continuously along the pointed ridge of A'Chioch, and thence northwards into Beinn Fhada. The same lavas are likewise found in two outliers, capping Beinn a' Chraig, a mile further north, and I have found fragments of them on some of the loftier ridges to the south-east. This highest and youngest group of lavas in the plateaux has been reduced to mere isolated patches, and a little further denudation will remove it altogether. Yet it is not less than about 800 feet thick, and consists of bedded andesitic or trachytic lavas, which alternate with and follow continuously and conformably upon the top of the ordinary plateau-basalts. These dull, finely crystalline or compact, light-grey rocks weather with a characteristic platy form, which has been mistaken for the bedding of tuffs. The fissility, however, has none of the regularity or parallelism of true bedding, and may be observed to run sometimes parallel with the bedding of the sheets, sometimes obliquely or even at right angles to it. Even where this structure is best developed, the truly crystalline nature of the rocks can readily be detected. Some of them are porphyritic and amygdaloidal, the very topmost bed of the mountain being a coarse amygdaloid. Intercalated with these curious rocks there are others in which the ordinary characters of the dolerites and basalts of the plateaux can be recognised. The amygdaloids are often full of delicate prisms of epidote.

In Mull, as in the other areas of terraced basalts, we everywhere meet with gently inclined sheets, which do not thicken out individually or collectively in any given direction, except as the result of unequal denudations. So far as I have been able to discover, they afford no evidence of any great volcanic cone from which they proceeded. Their present inclinations are unquestionably due, as in Ireland, to movements subsequent to the formation of the plateau. In Loch-na-keal they dip gently to the E.N.E.; in Ulva and the north-west coast to N.N.E.; near Salen to W.S.W. on the one side, and N.W. on the other. Round the southern and eastern margins of the mountainous tract of the island, they dip generally inwards to the high grounds.

The Mull plateau presents a striking contrast to that of Antrim, in the extraordinary extent to which it has been disrupted by later protrusions of massive basic and acid rocks over a rudely circular area, extending from the head of Loch Scridain to the Sound of Mull, and from Loch-na-keal to Loch Buy. The bedded basalts have been invaded by masses of dolerite, gabbro, and granophyre, with various allied kinds of rock. They have not only been disturbed in their continuity, but have undergone considerable metamorphism.

Again, further to the north, in the promontory of Ardnamurchan, the plateau has been disrupted in a similar way, and only a few recognisable fragments of it have been left. These changes will be more appropriately discussed in connection with similar phenomena in the other plateaux further north.