CHAPTER XLVII

THE ACID BOSSES OF MULL, SMALL ISLES, ST. KILDA, ARRAN AND THE NORTH-EAST OF IRELAND

ii. THE ACID BOSSES OF MULL

Though of comparatively small extent, the granophyre bosses of the island of Mull afford to the geologist a large amount of instruction in regard to the relations of the different members of the volcanic series to each other. Especially important is the evidence which they contain of the connection between the acid and basic groups of rocks. They have been laid bare in many natural sections, some of which, forming entire hillsides, are among the most astonishing in the whole wonderful series which, dissected by denudation, reveal to us the structure of these volcanic regions. They lie in two chief areas. One of these extends along the northern flanks of the mountainous tract from the western side of Beinn Fhada across Loch Ba' to the west side of Glen Forsa. The other occupies for over three miles the bottom of Glen More, the deep valley which, skirting the southern side of the chief group of hills, connects the east side of the island by road with the head of the great western inlet of Loch Scridain. There are other minor areas. One of these extends for about a mile along the declivities to the south of Salen, across the valley of the Allt na Searmoin; another occurs at Salen; a third runs along the shore at Craignure. In the interior also, many isolated areas of similar rocks, besides thousands of veins, occur in the central group of hills and valleys which form the basins of the Glencannel and Forsa rivers (Map VI.).

The chief northern boss, which for the sake of convenience of reference may be called that of Loch Ba', has a length of nearly six miles, with a breadth varying from a quarter of a mile to about a mile and a quarter. It descends to within 50 feet of the sea-level, and is exposed along the crest of Beinn Fhada at a height of more than 1800 feet. It chiefly consists of a grey crystalline rock which might readily be identified as a granite, but which when examined microscopically is found to possess the granophyric structure. With this distinctly granular-crystalline rock are associated various porphyritic and felsitic masses, which pass into it, and are more specially observable along its border. An exceedingly compact black quartz-felsite or rhyolite forms its southern boundary, runs as a broad dyke-like ridge from the head of the Scarrisdale Water north-eastward across Loch Ba' (Fig. 352), and spreads out eastward into a mass more than a mile broad on the heights above Kilbeg in Glen Forsa. The sharp line of demarcation of this felsite, and its mass and extent, point to a different period of extravasation.

[Illustration: Fig. 352.--View of the hills on the south side of the head of Loch na Keal, showing the junction of the granophyre and the bedded basalts.

One bird, the bedded basalts of the Gribon plateau; two birds, the bedded dolerites and basalts of Beinn a' Chraig adhering to the northern slope and capping the hill; three birds, summit of Ben More, with A'Chioch to the left and the top of Beinn Fhada appearing in the middle distance between them; four birds, the granophyre slopes of Beinn a' Chraig with the great dyke-like mass of felsite on the left. ]

The geologist, who approaches this district from the north-east, has his attention arrested, even at a distance of several miles, by the contrast between the outer and inner parts of the hills that lie to the south-west of Loch Ba'. He can readily trace from afar the dark bedded basic rocks rising terrace above terrace, from the shores of Loch na Keal, to form the seaward faces of the hills along the southern side of that fjord. But he observes that immediately behind these terraces the mass of the rising ground obviously consists of some amorphous rock, which weathers into white debris. Nothing can be sharper than the contrast of colour and form between the two parts of the hills. The bedded plateau-rocks lie as a kind of wall or veneer against a steep face of the structureless interior (Fig. 352). Seen from the other or hilly side, the contrast is perhaps even more striking. But the astonishment with which it is beheld at a distance becomes intensified when one climbs the slopes, and finds that the sheets of dolerite and basalt (which from some points of view look quite level, yet dip towards the north-east at a gentle angle) are immediately behind the declivity abruptly truncated by a mass of granophyre. Of all the junction-lines between the acid bosses and the lavas of the plateaux, those exposed on these Mull hillsides are certainly the most extraordinary. So little disturbed are the lavas, that one's first impulse is to search for pebbles of the granophyre between the basalts, for it seems incredible that the inner rock should be anything but a central core of older eruptive material, against and round which the younger basic rocks have flowed. But, though the granophyre is so decomposing and covers its slopes with such "screes" of debris, that had the basalts been poured round it, they must infallibly have had some of its fragments washed down between their successive flows, not a single pebble of it is there to be found. This might not be considered decisive evidence, but it is extended and confirmed by the fact that the acid rock gives off veins which ramify through the basalts.

Before examining the actual contact of the two rocks, however, the geologist will not fail to observe here an admirable example of the gradual change which was described in the foregoing chapter as coming over the bedded basalts near the acid bosses. As he approaches the nucleus of white rock, the basalts assume the usual hard indurated character, not decaying into brown sand as on the plateaux, but often standing out as massive crags with vertical clean-cut joint-faces. This metamorphosed condition extends in some cases to a considerable distance from the main body of acid rock, especially where knobs of that material, protruding through the more basic lavas, show that it must extend in some mass underneath. Thus along the shore at Saline the bedded basalts succeed each other in well-defined sheets, some being solid, massive and non-amygdaloidal, others quite vesicular, and recalling the black scoriform surfaces of recent Vesuvian lavas; yet they are all more indurated than in the normal plateau-country, and they break with a hard splintery fracture. Immense numbers of dykes cut these rocks, and they are likewise pierced by occasional felsitic intrusions.

If we cross to the other side of the island and trace the bedded basalts away from the central masses of acid rock we meet with so gradual a diminution of the induration that no definite boundary-line for the metamorphism can be drawn. As we recede from the centre of alteration, the rocks insensibly begin to show brown weathered crusts, with spheroidal exfoliation, the reticulations of epidote and calcite become much less abundant, the amygdaloids gradually assume their normal earthy character, and eventually we find ourselves on the familiar types of the plateau. This transition is well seen along the shores of Loch na Keal.[401]

[Footnote 401: Some of the thick massive sheets of basic rock along the south side of this inlet may possibly be altered sills.]

These proofs of the alteration of the plateau-basalts are accompanied in Mull as in Skye by further abundant evidence that the acid rocks are of younger date than the basic. In particular, dykes and veins may be traced proceeding from the former and intersecting the latter. Thus, in the bed of the south fork of the Scarrisdale stream, a separate mass of granophyre (which under the microscope exhibits in perfection the characteristic structure of this rock) protrudes through the basalts in advance of the main mass, and a little higher up on the outskirts of that mass narrow ribbons of the granophyre run through the basic rocks. The contrast of colour between the pale veins of the intrusive rock and the dark tint of the basalts is well shown in the channel of the water. Similar sections may be seen on the flanks of Beinn Fhada, especially in the great corry north of Ben More, where the granophyre sends a tongue of finer grain between the beds of basalt. On the east side of Loch Ba' numerous proofs of similar intrusion may be observed. Thus at the east end of Loch na Dàiridh, where the granophyre has been intruded into the basalts, hand-specimens may be obtained showing the two rocks welded together. On the slopes of Cruach Tòrr an Lochain, where the granophyre has a felsitic selvage, the bedded basalts are traversed by veins of the latter material (Fig. 353). A little further east, at the head of the Allt na Searmoin, the bedded basalts, some of which are separated by slaggy scoriaceous surfaces, are intersected by another protrusion from the compact felsitic porphyry (Fig. 354).[402] A mile lower down the same valley a separate mass of granophyre sends out veins into the basalt, which as usual is dark bluish-grey in colour, indurated and splintery.

[Footnote 402: This rock appears to the eye as a black finely crystalline-granular felsite. Under the microscope, it was found by Dr. Hatch to "present a markedly granulitic structure, consisting mainly of small rounded grains of dirty brown turbid felspar, with isolated granules of colourless quartz. Scattered through the rock, or accumulated in patches, are small spherical or drop-like granules of a bright green augite (coccolite)."]

[Illustration: Fig. 353.--Section on south side of Cruach Tòrr an Lochain, Mull.

_a_, bedded basalts and dolerites; _b_, granophyre; _c_, marginal finer-grained band; _d_ _d_, veins from the granophyre traversing the basic rocks.]

[Illustration: Fig. 354.--Section at head of Allt na Searmoin, Mull.

_a_, basalts and dolerites, with slaggy upper surfaces; _b_, felsite.]

As the posteriority of the Mull granophyre and felsites to the basalts is thus proved, the further question remains as to their manner of intrusion. Here and there, especially on the south-eastern side, between the head of the Scarrisdale river and Loch Ba', the line of junction between the two rocks is nearly vertical, but a body of black felsite intervenes as a huge wall between the ordinary granophyre and the basalt. On Beinn Fhada and Beinn a' Chraig the line of separation, as I have above remarked, is inclined outwards, and plunges under the basalts at an angle of 30° to 40°. The terraced basalts and dolerites are not sensibly disturbed, but end off abruptly against the steep face of intrusive rock. We might suppose that in this case the younger rock had merely carried upward the continuation of the beds that are truncated by it, as if an orifice had been punched out for its ascent. But on the top of the ridge of Beinn a' Chraig we find that the outliers which there remain are not portions of the lower basalts, but of the upper "pale group" of Ben More. The same rocks are prolonged on the other side of the Scarrisdale Glen, sweep over the summit of Beinn Fhada, and run on continuously into the crest of A'Chioch and the upper part of Ben More. The granophyre has usurped the place of the lower dolerites and basalts, but has left the more felspathic lavas of the "pale group" in their proper position. And to make this remarkable structure still more clear, sections may be seen on the southern flanks of Beinn Fhada, where the upper surface of the granophyre comes down obliquely across the edges of the lavas, and allows the junction of the basalts and the "pale group" to be seen above it (Fig. 355). As in the case of Beinn an Dubhaich, it is as if the granophyre had eaten its way upward and dissolved the rocks which it has replaced.

[Illustration: Fig. 355.--Section on south side of Beinn Fhada, Mull.

_a_, bedded basalts and dolerites; _b_, "pale group" of Ben More; _c_, granophyre.]

The usual kind of contact-metamorphism has been produced around this intrusive boss. It is most marked in the outliers that cap Beinn a' Chraig and on the two ridges to the south-west, where it is seen to consist in a high degree of induration, the production of a shattery, irregularly-jointed structure, and the effacement of the obvious bedding which characterizes the unaltered rocks.

The position of this eruptive mass, quite a mile broad, breaking through, without violently tilting, more than 1800 feet of the bedded basalts, and then stopping short about the base of the "pale group," presents a curious problem to the student of geological physics. It at once reminds him of many sections among Palæozoic granites where an eruptive boss has ascended and taken the place of an equivalent volume of the surrounding rocks, which, though more or less metamorphosed, are not made to dip away from it as from a solid wedge driven upwards through them. In this Mull case, however, there are some peculiar features that deserve consideration, for they seem to show that here, as elsewhere, passages for the uprise of the intrusive rock were already provided by the presence of volcanic pipes, which, even if filled up with fragmentary materials, would no doubt continue to be points of weakness. Round the flanks of the Loch Ba' boss, and here and there on its surface, patches of intensely indurated volcanic agglomerate may be detected. A little to the south of the tarn called Loch na Dàiridh, the granophyre is succeeded by the black, flinty felsite or rhyolite already referred to. This rock in some places exhibits a beautiful flow-structure, with large porphyritic felspars, and encloses a great many fragments of dolerite and gabbro, varying from the size of a pea up to blocks several inches in diameter. Lying on its surface are detached knolls of much altered dolerite, basalt, and coarse breccia or agglomerate. On its southern margin one of these patches of agglomerate contains abundant fragments of various felsitic rocks, among which are pieces of a compact rock with flow-structure like that found in place immediately to the north; also rounded pieces of quartzite, and of compact and amygdaloidal basalt wrapped up in a very hard matrix which seems to consist largely of basalt-dust. No bedding can be made out in this rock, and the mass looks like part of a true neck. Further down the slope the bedded basalts appear. The actual junctions of the different rocks cannot be satisfactorily traced, but the structure of the ground appears to me to be as shown in Fig. 356. A patch of similar agglomerate appears a little to the south-west of the last section in front of a cliff of the felsite, and seems to be enclosed in the latter rock, and other exposures of agglomerate, underlain and intensely indurated by the felsite, may be noticed on the ground that slopes towards Loch Ba'.

[Illustration: Fig. 356.--Section to south of Loch na Dàiridh, Mull.

_a_, basalts; _b_, dolerite; _c_, volcanic agglomerate; _d_, black felsite; _e_, granophyre.]

That these agglomerates do not belong to the period of the eruption of the granophyre and felsite, but to that of the bedded basalts, may be inferred from their intense induration next the acid rocks, and also from the fact that similar breccias are actually found here interposed between the bedded basalts. This is well shown on the hill above the Coille na Sròine, where the accompanying section can be seen (Fig. 357). The broad dyke-like mass of black flinty felsite already referred to runs as a prominent rib over the southern end of Beinn a' Chraig into the head of the Scarrisdale glen (see Fig. 352). It cuts across the bedded basalts, and immediately to the south of where these appear, a thin intercalated bed of breccia crops out, of the usual dull-green colour, with abundant fragments of basalt and many of yellow and grey felsite.

From these various facts we may, I think, conclude that along the strip of ground now occupied by the Loch Ba' boss of granophyre and felsite, there once stood a line or group of vents, from which, besides the usual basalt-debris, there were ejected many pieces of different felsitic or rhyolitic rocks, and that these eruptions of fragmentary material took place during the accumulation of the plateau-basalts. These volcanic funnels occasioned a series of points or a line of weakness of which, in a long subsequent episode of the protracted volcanic period, the acid rocks took advantage, forcing themselves upwards therein, and leaving only slight traces of the vents which assisted their ascent. The mingling of acid and basic fragments in the material ejected from these vents is another proof of the existence of acid rocks in the volcanic reservoirs before the advent of the great granophyre intrusions. The evidence thus entirely confirms the conclusions deduced from the Skye area.

[Illustration: Fig. 357.--Section of junction of south side of Loch Ba' granophyre boss, with the bedded basalts, Mull.

_a_, bedded basalts; _b_ _b_, basalt-tuff and breccia; _c_, granophyre; _d_, black felsite; _e_, coarse dolerite dyke, 30 or 40 feet wide.]

The second or Glen More boss, instead of rising into hilly ground, is confined to the bottom of the main and tributary valleys, and has only been revealed by the extensive denudation to which these hollows owe their origin. It begins nearly a mile below Torness and extends up to Loch Airdeglais--a distance of almost four miles. Though singularly devoid of topographical feature, it exhibits with admirable clearness the relation of the granophyres to the gabbros, and thus deserves an important place among the tracts of acid rocks in the Western Islands. Its petrographical characters change considerably from one part of its body to another. For the most part, it is a true granophyre, sometimes with orthoclase, sometimes with plagioclase as its predominant felspar. At Ishriff, as already stated, it is sprinkled with long acicular decayed crystals of hornblende; but at the watershed the ferro-magnesian mineral is augite. The surrounding rocks are mainly the plateau-basalts, with their sills of dolerite and gabbro.

[Illustration: Fig. 358.--Mass of dark gabbro about two feet in diameter traversed by pale veins of granophyre, lying on north slope of Creag na h-Iolaire, Mull.]

[Illustration: Fig. 359.--Section at Creag na h-Iolaire, Glen More, Mull, showing basalts and gabbros resting on and pierced by granophyre.

_a_, much indurated and altered basalts and dolerites; _b_ _b_, gabbro; _c_, granophyre; _d_ _d_, basalt dykes.]

This strip of granophyre sends abundant apophyses from its mass into the dark basic rocks around it. Some of the best sections to show the nature of these offshoots are to be found on the steep hillslope which mounts from the watershed in Glen More southward into the Creag na h-Iolaire (Eagle's Crag), and thence up into the great gabbro ridge of Ben Buy. From the main body of granophyre a multitude of veins ascends through the basalts and gabbros from two feet or more in breadth down to mere filaments (Fig. 358). Even at a height of 300 feet up the hill some of these veins are still three inches broad, and present the usual granophyric structure, though rather finer in grain than the general mass of the boss, and sometimes assuming a compact felsitic and spherulitic texture at the immediate contact with the surrounding rock. One of the most striking proofs of the posteriority of these veins is furnished by the perfect flow-structure they not infrequently exhibit along their margins, their long felspar crystals being arranged parallel to the walls in lines that follow the sinuosities of the boundary between the two rocks. Patches of gabbro and of the indurated basalts may be seen lying on the granophyre, from which veins and strings ramify through them (Fig. 359). Similar veins can be traced upward into the main body of coarse gabbro, forming the ridge of Ben Buy. Some of them are of the usual granular granophyric texture, others are dull and fine-grained (claystones of the older authors).

Hence it is evident that the granophyres of Mull have been protruded not only after the accumulation of the plateau-basalts, but after these were traversed by the sheets and veins of gabbro. The amount of acid rock injected into these older rocks over the mountainous part of the island is enormous; but I reserve further reference to it for the section on acid Dykes and Veins, for these are the forms in which it chiefly occurs in that region. It should be added, that in the localities here referred to basalt-veins and dykes are generally abundant, cutting through all the other rocks (Fig. 359). So numerous are they that the geologist ceases to take note of them when his thoughts are engaged upon the problems presented by the masses through which they rise.

iii. THE ACID BOSSES OF SMALL ISLES

In the island of Eigg three small bosses or sheets of acid rock occur. That at the northern end rises through the Jurassic sedimentary rocks, and forms a bold cliff from 150 to 200 feet high. It is a light grey granophyric porphyry, with rounded blebs of quartz in a micropegmatic base of quartz and felspar. The other two masses, of smaller size, cut through the bedded basalts[403] (Map VI.).

[Footnote 403: _Quart. Journ. Geol. Soc._ xxvii. (1871) p. 294.]

In the opposite island of Rum, the acid protrusions play a much more important part. On the east side of the hills, they occur in sheets at the base of the gabbros; on the west side, they form a large tract of hilly ground, which, stretching along the coast-line for about three and a half miles from the headland of A' Bhrideanach to Harris, forms there a range of shattered sea-cliffs, that tower for 1000 feet above the Atlantic breakers that beat about their base. The area extends inland to the slopes on the west side of Loch Sgathaig, a distance of about three and a half miles, descending in a range of precipices along its northern front, and reaching in its culminating summit, Orval, a height of 1868 feet above the sea. The rocks of which this triangular area consists resemble those of the Mull bosses. They are chiefly quartz-porphyries, becoming felsitic in texture towards their contact with adjacent rocks. In some places, as was noticed by Macculloch on the sea-cliffs,[404] they have a rudely bedded structure. Thus on the north-west front of Orval, this structure is shown by parallel planes that dip outwards or north-west at 30° to 40°, and which are made still more distinct by an occasional intrusive dyke or sheet of basalt between their surfaces. I have already alluded to indications of an internal arrangement in the granitoid bosses of Skye (p. 381).

[Footnote 404: _Western Islands_, vol. i. p. 487.]

[Illustration: Fig. 360.--Section on north side of Orval, Rum.

_a_, Torridon sandstones; _b_, bedded basalts of Fionn Chro; _c_, dolerite; _d_, quartz-porphyry.]

[Illustration: Fig. 361.--Junction of Quartz-Porphyry (Microgranite) and Basic Rocks, south-east side of Orval, Rum.

_a_, basalts and dolerites; _b_, dolerite and gabbro veins; _c_, quartz-porphyry cutting _a_ and _b_.]

As in the other islands, the granophyres, porphyries and felsites of Rum have been intruded at the base of the volcanic series. Over much, if not all, of their area they lie directly on the red Torridon sandstone. That the bedded basalts once covered them is indicated by the position of the three outliers of the basalt-plateau already noticed. But a fourth outlier still lies upon the porphyry of Orval as a cake that dips gently northward. It consists of a bedded, dark, finely-crystalline, ophitic dolerite, porphyritic in places, with a rudely prismatic or columnar structure (Fig. 360). It has undergone contact-metamorphism, and tongues from the underlying rock project up into it. On the south-eastern side of the same hill, still more striking evidence is presented of the posteriority of the acid to the basic rocks. The porphyry shows here the same tendency to assume a bedded structure, the parallel "beds" again dipping outward or south-east at 40°. They plunge under the body of gabbro, dolerite and other intrusive masses which from this point stretch eastward into the great cones of Allival and its neighbours. The rock at the junction is a fine microgranite with traces of micropegmatite. It is composed of a holocrystalline base of quartz and orthoclase, with porphyritic crystals of microcline, blebs of quartz and scattered granules of augite. The rocks that rest immediately next it are basalt and dolerite, into which it has sent an intricate network of veins (Fig. 361).[405] It has also pushed long tongues down the slope into them, which may be seen traversing the dolerite and gabbro veins that cut the basalts. The basic rocks next the porphyry have been intensely altered. They seem in places as if they have been shattered by some explosive force, and had then been invaded by the mass that rushed into all the rents thus caused. This remarkable structure is still better displayed on St. Kilda, and is more fully described in the following account of the geology of that island.

[Footnote 405: In a thin slice cut from a specimen showing the junction, there is a minute vein of the porphyry penetrating the basalt which is much altered, while the porphyry becomes much finer in grain than at a distance from the contact.]

iv. THE ROCKS OF ST. KILDA

Brief allusions to St. Kilda and its rocks have already been made (pp. 173, 358). We may now enter more fully upon the consideration of its geological structure and history.

When the weather is clear there may be seen from the western headlands of the Outer Hebrides a small blue cone rising above the Atlantic horizon at a distance of about 60 miles. As the voyager approaches this distant land it gradually shapes itself into a group of islets of which St. Kilda, the largest and only inhabited, has an extreme length of about four miles, a breadth of less than two miles, and a height of 1262 feet above the sea. Four miles to the north-east Borrera, about one square mile in extent, rises with precipitous sides to a height of 1000 feet. Off the north-western promontory of St. Kilda the huge rock of Soay, half a square mile in area, towers from 600 to 800 feet above the waves. Borrera has two attendant rocks--Stack Li and Stack an Armin--huge pyramidal masses several hundred feet high, and the home of thousands of gannets. St. Kilda possesses two less imposing islets between its north-western headland and Soay, and a third to the south-east known as Levenish.

The scenery of this picturesque group affords a good indication of its geological structure. It displays two distinct types of topographical form. In Borrera the marvellous combination of spiry ridges, deep gullies and clefts, notched crests and splintered pinnacles, at once reminds the visitor of the outlines of the Cuillin Hills of Skye. The same features are repeated on a less magnificent scale in Soay and along the whole of the south-western precipitous coast-line of St. Kilda.

In marked contrast to these varied outlines, the eastern half of St. Kilda rises with a smooth green surface, varied with sheets of grey screes, up to the rounded summit of Conagher, the highest point in the island. If the dark crags of the rest of the island group remind one of the Cuillins, this eastern tract recalls at once the form and colour of the Red Hills of Skye. A closer examination shows that in each case the topography arises from the influence of the very same rocks and geological structure as in that island.

There is, however, one aspect in which St. Kilda has no rival throughout the Western Isles. Its russet-coloured cone, though rising on the west side with gentle green slopes from the central valley, plunges on the eastern side in one vast precipice from a height of 1000 feet or more into the surge at its base. Nowhere among the Inner Hebrides, not even on the south-western side of Rum, is there any such display of the capacity of the youngest granite to assume the most rugged and picturesque forms. It is hardly possible to exaggerate the variety of outline assumed by the rock as it yields along its system of joints to the influence of a tempestuous climate. It has been carved into huge projecting buttresses and deep alcoves, the naked stone glowing with tints of orange and fawn colour, veiled here and there with patches of bright green slope, or edged with fringes of sea-pink and camomile. Every outstanding bastion is rent with chasms and split into blocks, which accumulate on the ledges like piles of ruined walls. To one who boats underneath these cliffs the scene of ceaseless destruction which they present is vividly impressive.

The geology of St. Kilda was sketched by Macculloch, who recognized the close resemblance of its two groups of rock to the "augite-rock" (gabbro) and "syenite" (granophyre) of Skye and other islands of the Inner Hebrides. But he left the relations of the two groups to each other undetermined.[406] Professor Heddle has published a brief reference to the rocks of St. Kilda, without, however, offering any definite opinion as to the geological structure of the islands.[407] The best account of the geology has been given by Mr. Alexander Ross, who obtained evidence that the acid sends veins into the basic rock. He brought away specimens clearly showing this relation, but in his description left the question open for further inquiry.[408] To some of the observations in these papers reference will be made in the sequel. The following account is based on the results of two visits paid by me to St. Kilda in the summers of 1895 and 1896, during which I was enabled to examine the rocks on land, and to sail several times round the islands, boating along those parts of the cliffs which presented features of special geological importance.

[Footnote 406: _Description of the Western Isles_, vol. ii. p. 54.]

[Footnote 407: In an article on the general geological features of the Outer Hebrides contributed to _A Vertebrate Fauna of the Outer Hebrides_, by J. A. Harvie-Brown and T. E. Buckley, 1888.]

[Footnote 408: _British Association Report_, 1885, p. 1040, and a much fuller paper in the _Proceedings of the Inverness Field Club_, vol. iii. (1884), p. 72.]

In the St. Kilda islets three groups of rock differing from each other in age may be recognized. 1st, A series of gabbros, dolerites and basalts which have been intruded through and between each other as sills; 2nd, a mass of granophyre which invades these sills; and 3rd, abundant dykes and veins of basalt which occur both in the basic and acid masses.

From the extension of the basalt-dykes across the Outer Hebrides it is clear that the Tertiary volcanic region reached at least to within 60 miles of St. Kilda. Whether or not it stretched over the intervening space now overflowed by the Atlantic must be matter for conjecture. There can be no doubt that the intrusive rocks of St. Kilda are in age and origin the equivalents of those of the Inner Hebrides. The remnants left of them were assuredly not superficial extrusions, but are characteristic examples of the more deep-seated intrusions of the Tertiary volcanic period. Down to the most minute details of structure they reproduce the features so well displayed by the gabbros and granophyres of Skye, Rum and Mull. If it is demonstrable in the case of these islands that the intrusions have taken place under a deep cover of basalt-sheets, now in large part removed, the inference may legitimately be drawn that at St. Kilda a basalt-plateau once existed which has been more completely destroyed than in the other regions. Not a fragment of such a plateau has survived, unless we may perhaps be allowed to recognize it in some of the basalts enclosed among the gabbro-sills. Placed far amid the melancholy main and exposed to the full fury of the Atlantic gales, these islets must be regarded as the mere fragmentary cores of a once much more extensive volcanic area. The geologist who visits them is deeply impressed at every turn by the evidence of the active and unceasing destruction which their cliffs are undergoing. Nothing now remains save the deep-seated nucleus of intrusive sills, bosses and dykes.

1. _The Gabbro Sills._--The rudely-bedded arrangement of these rocks is conspicuous along the west side of St. Kilda, in Soay and in Borrera. They consist of coarse and fine varieties disposed in successive sheets which dip at angles varying from as little as 15° up to as much as 60° or even more. In St. Kilda they form the picturesque promontory of the Dune, and extend thence along the western side of the island to its extreme northern end. Their escarpments face the ocean, and their dip-slopes descend towards the north-east in grassy declivities to the south bay and the long verdant glen which runs thence across to the north bay. The same strike is prolonged into Soay, but further east in Borrera the direction curves so as to present vast escarpments towards the west and shelving sheets of rock towards the east.

None of the gabbros seen by me are as coarse as the large-grained varieties of Skye, nor does there appear ever to be such a marked banded structure among them as that displayed by the Cuillin rocks. Faint banding, however, may be noticed. A series of specimens which I collected from the west side of the island has been sliced for microscopic examination, and Mr. Harker has furnished me with the following notes regarding them.

"An olivine-gabbro from the west side of St. Kilda [7107] is a dark, heavy, medium-grained rock, in which augite and felspar are conspicuous. The microscope shows, in addition, plentiful grains of olivine, with but little original iron-ore, and some apatite-needles. The structure is ophitic, the plates of pale-brown augite enveloping both olivine and felspar. A little brown hornblende and red-brown mica are probably original, the rock showing little sign of alteration. The felspar is labradorite, with albite- and Carlsbad-twinning, and forms elongated rectangular crystals.

"Another specimen [7108] is a rock of similar appearance but somewhat coarser texture, and structurally is a more typical gabbro than the preceding, the felspar having little of the 'lath' shape, while the augite, though still moulded on the felspar, scarcely assumes an ophitic habit. A striking feature in this rock is the way in which the augite is crowded with 'schiller'-inclusions, in places so closely as to be almost opaque. A high magnification shows that these inclusions are dark, linear in form, and disposed along two directions intersecting at a high angle. The labradorite has unusually close twin-lamellation on both albite and pericline laws, and it is possible that this is a strain-effect.

"A third specimen [7109] is from a rock in every respect identical with the preceding, except that the olivine is rather more plentiful, and in some grains is partially serpentinized."

While the gabbros of St. Kilda are not a mere uniform boss, but a series of sills and irregular masses which have been successively injected into each other, they have subsequently been cut through by many basalt-dykes and veins. These, which are sometimes as abundant as in the gabbro of the Cuillin Hills, traverse the rocks both in the line of bedding and also at many different angles across it. As they generally weather faster than the gabbros, they give rise to deep narrow clefts which may be traced up the whole height of the precipices, occasioning sea-caves below and sharp notches on the crests above.

These scenic features, so indicative of the geological structure that causes them, are specially well seen on the western face of the Dune or south-western promontory of the island, and likewise in the strangely rifted precipices further north and in Soay. They are, however, most impressively displayed around the naked walls of Borrera, which in their marvellous combination of spiry ridges, deep straight gullies, and splintered crests, remind one at every turn of the scenery of Blaven and the Cuillin Hills.

2. _The Granophyre Boss and its Apophyses._--The eastern half of the island of St. Kilda consists of a pale rock which Macculloch long ago identified with the granophyre of Skye, and which, as he pointed out, has much resemblance to parts of the granite of Arran.[409] Not only does it give rise to topographical forms like those of the Red Hills, but it weathers, like the Skye granophyre and the Arran granite, into thick bed-like sheets divided by transverse joints into large quadrangular blocks. On closer inspection it is found to resemble still more precisely the acid rocks of the Inner Hebrides. It possesses the same drusy micropegmatitic structure as the granophyres of Skye, Rum and Mull. The ferro-magnesian constituents are present in small quantity, hence the pale hue of the stone. The quartz and felspar project in well-terminated crystals into the drusy cavities, which are sometimes further adorned with delicate tufts of clear crystallized epidote. In these and other respects the rock displays the familiar external forms of the younger or Tertiary granites of Britain.

[Footnote 409: _Description_, vol. ii. p. 54.]

Mr. Harker's notes on the microscopic structure of this granophyre are as follows:--"The prevailing felspar is orthoclase, often very turbid from secondary products. Even what appear to be distinct crystals are sometimes seen in the slices to be invaded on the margin by quartz in rough micrographic intergrowths, and much of the finer intergrowth occurs as a fringe to the crystals. In this case the felspar of the micropegmatite can often be verified to be in crystalline continuity with the crystal which has served as a nucleus [6624]. Quartz occurs in distinct crystals and grains as well as in the micropegmatite. There is a more granitoid variety of the rock, in which only a very rude approach to micrographic intergrowths is seen [6623]. In both varieties there is but little trace of any ferro-magnesian mineral; the more typical granophyre has what seems to be destroyed augite, while the granitoid rock contains a little deep-brown biotite. Scattered crystal-grains of magnetite occur in both."

Narrow ribbon-like veins of a finer material, sometimes only an inch in breadth, traverse the ordinary granophyre. Similar veins run through the rock of the Red Hills in Skye; they are sharply defined from the enclosing rock, as if the latter had already solidified before their intrusion. With regard to the microscopic structure of some thin slices prepared from these veins, Mr. Harker remarks that "the material of the veins is of a type intermediate between granophyre and microgranite [6622, 6623]. The chief bulk is a finely-granular aggregate of quartz and felspar, the latter very turbid; but in this aggregate are imbedded numerous patches of micropegmatite, often of perfect and delicate structure. These areas of micropegmatite show some approach to a radiate or rudely spherulitic structure, and, in some cases, are clustered round a crystal of felspar or quartz. Some granules of magnetite and rare flakes of brown biotite are the only other constituents of the rock. Although they must be of somewhat later date, there is evidently nothing in the petrographical characters of these fine-textured veins to separate them widely from the ordinary granophyres of the region."

These veins may be compared with the spherulitic dyke that traverses the granophyre of Meall Dearg at the head of Glen Sligachan (described at p. 381), which, though undoubtedly somewhat younger than the rock that contains it, yet presents the very same structures as are visible at the margin of that rock.[410] The material of this dyke and of the finer veins of St. Kilda and the Red Hills probably belongs to a later period of protrusion from a deeper unconsolidated portion of the same acid magma as at first supplied the general body of granophyre.

[Footnote 410: _Quart. Journ. Geol. Soc._ vol. 1. (1894), p. 220.]

Undoubtedly the most interesting feature in the granophyre of St. Kilda is its junction with the mass of basic rock to the west of it. This junction-line runs from about the middle of the chief or south bay (where, however, its precise position is concealed under detritus) across the island to the north shore, where it descends the face of the precipice and plunges under the sea. Important as the actual contact of the two rocks obviously is in regard to their relative date, it has not hitherto been observed or described. Macculloch noticed "numerous fragments of trap penetrated by veins of syenite," but he did not see these rocks in place, and, in spite of their apparent testimony to the posteriority of the acid intrusions, he was inclined to believe that the veins were not real veins, but that the "trap" and "syenite" had a common origin and would be found to pass into each other, as he thought also occurred in Mull and Rum. In recent years Mr. Alexander Ross, during his visit to St. Kilda, collected specimens illustrating the varieties of gabbro, dolerite and basalt, and showing the intrusion of the acid into the basic rocks. As already stated, he was disposed to regard the "granite" as of younger date than the gabbros, but left the question undecided.[411]

[Footnote 411: In his paper, _Proceed. Inverness Field Club_, vol. iii. (1884), p. 78, Mr. Ross quotes a letter from Prof. Judd, who there states that the rock supposed to be granite "is seen under the microscope to be a quite different rock--a quartz-diorite." Some of the specimens from St. Kilda collected by Mr. Ross were exhibited at a meeting of the Geological Society on 25th January 1893. With regard to these Prof. Judd, in the course of the discussion on his paper on "Inclusions of Tertiary Granite in the Gabbro of the Cuillin Hills," remarked:--"They show a dark rock traversed by veins of a light one, but the dark rock is not a gabbro and the light one is not a granite" (_Quart. Journ. Geol. Soc._ vol. xlix. (1893), p. 198).]

The best locality for the examination of the junction of the main granophyre mass with the gabbros is inaccessible save by boat, and only in the calmest weather. It occurs in the great cliff on the northern side of the island between the north bay and the sea-stack known as the Bragstack. The line of contact emerges from below the sea-level, and ascends the cliff with a westward inclination of from 60° to 80°. Here, as in Skye, the acid rock underlies the basic masses, which are rudely bedded and much jointed. About 150 feet above the sea-level, the nearly vertical cliff breaks up into an exceedingly rocky and rugged acclivity, across which the junction seems to slope at a lower angle. But the place is hardly reachable, save perhaps by the intrepid, barefooted cragsmen of St. Kilda.

[Illustration: Fig. 362.--Junction of granophyre and gabbro, north side of St. Kilda.]

Along the sharply defined line of contact the granophyre is close-grained, and sends a network of veins into the dark sheets of gabbro. The general features of the junction are represented in Fig. 362. The veins are narrow, those nearest the main body of granophyre diverging from it at a still more acute angle than those from the mass of Meall Dearg (Fig. 376), and then branching so as to enclose masses of the gabbro and to run across them in long parallel veins. A characteristic feature of many of these veins, besides their narrowness, is their tendency to split up at the ends into mere fingers and threads as represented in Fig. 363.

Owing to the depth of soil on the cultivated land, and of boulders and sand on the beach, the actual junction of the main body of granophyre with the gabbro is not seen on the southern shore. But a few yards to the westward of where it must lie, the beach is cumbered with large blocks of rock broken up from the mass, which can be seen _in situ_ a little further south forming a line of low cliff with a rugged foreshore. These rocks consist of various gabbros and basalts of rather fine grain, profusely traversed with veins of white granophyre. Some of these veins are two feet or more in breadth, and, when of that size, show the distinctive granular texture and drusy structure of the main part of the acid rock. But from these dimensions they can be traced through every stage of diminution until they become mere threads. When they are only an inch or two broad, they assume a finely granular texture like that of the veins that run through the body of the granophyre.

[Illustration: Fig. 363.--Veins of granophyre traversing gabbro and splitting up into thin threads, north side of St. Kilda.]

The amount of injected material in the dark basic rocks is here and there so great as to form a kind of breccia (Fig. 364), which, from the contrast of tone between its two constituents, makes a conspicuous object on the shore. Here, as in the example already cited from Rum, the basic rocks seem to have been shattered into fragments, and the acid material to have been injected into the minutest interstices between them. The enclosed fragments are of all sizes from mere grains up to blocks a foot or more in length. They are generally angular, like rock-chips from a quarry. Moreover, they are not all of the same kind of material. While at this locality most of them consist of basalt, they include also a few large and small pieces of rather coarse gabbro. There has evidently been a certain amount of transport of material, as well as an extensive disruption of the rocks _in situ_. The granophyre here and there assumes a darker or greener tint, as if it had dissolved and absorbed some portion of the older rock.

Still more astonishing are the sections to be seen on the western cliffs and rocky declivities of the ridge to the north of the Dune, at a distance of perhaps 500 or 600 yards westwards from those of the South Bay. Here the gabbro-sheets are traversed by a number of conspicuous white bands, which on examination prove to be veins or dykes of granophyre. As viewed from the sea, the general disposition of the two groups of rocks is represented in Fig. 366. The broadest mass of granophyre breaks out towards the bottom of the precipice, and slants upward as a sheet intercalated between the gabbro sills, with a breadth of about 40 or 50 feet, but rapidly thinning away in its ascent. One of the bands below it has a breadth of about 15 feet. The material of these intrusions is a pale fine-grained granophyre like that of the South Bay, I did not detect, either here or anywhere else in St. Kilda, a definite spherulitic structure such as is so common in the granophyre dykes of Skye.

[Illustration: Fig. 364.--Pale granophyre injected into dark basalt, South Bay, St. Kilda.

The crags on the further side of the bay are the gabbro sheets of the Dune. (From a photograph by Colonel Evans.)]

Though the acid intrusions are somewhat irregular both in thickness and direction, they lie generally parallel to each other in the line of strike of the bedding of the gabbros. They are no doubt apophyses from the main body of granophyre, which emerges to the surface about a third of a mile to the eastward, but may of course be at no great depth underneath.

[Illustration:

Fig. 365.--Veins of granophyre traversing a line-grained gabbro and scarcely entering a coarse-grained sheet, west side of Rueval, St. Kilda. ]

Besides the broader bands of acid rock, and diverging from them, a complicated network of veins ramifies in all directions through the gabbros, as at the South Bay. The extraordinary degree to which the basic rocks have been shattered into fragments is strikingly displayed here, likewise the extreme liquidity of the acid magma, whereby it was able to insinuate itself into every chink and cranny. But the observer notices that this condition of excessive disruption is not shared by all the basic sills, and is not attendant upon all the acid dykes. As an example of this irregular distribution of the structure, I give the accompanying sketch (Fig. 365), where a fine-grained gabbro has been completely broken up and intersected with granophyre veins, while the coarser sheet overlying it has almost entirely escaped. The dark basalt-like sheets appear generally to have been much more disrupted than the more largely-crystalline varieties. It is noticeable here, also, that the fragments entangled in the network of granophyre veinings do not entirely belong to the rock that has been shattered, but sometimes include large and small lumps of different gabbros, showing some transference of material with the inrush of the acid magma.

Though closer in grain where it comes in contact with the gabbro, the granophyre never assumes any vitreous texture along its margin. A series of thin slices, prepared from specimens collected by me in the South Bay in the summer of 1895, was examined by Mr. Harker, who furnished the following notes regarding them:--"The basalt traversed by the granophyre is a fine-textured variety with small porphyritic felspars. These latter seem to be usually unaltered, retaining the glass cavities which in some of the crystals are abundant. The groundmass, however, shows minerals of metamorphic origin which must be derived mainly from the original augite. A brown mica is the most conspicuous; but with it are associated some brownish-green hornblende and certain chloritic and perhaps serpentinous substances. It is chiefly near the margin of a fragment of basalt that the mica gives place to these minerals. The basalt still retains plenty of unaltered granules of augite in the central parts of a fragment. It is not certain that the secondary minerals named come exclusively from the augite of the basalt; from their form and mode of occurrence they may in part have replaced olivine or even rhombic pyroxene.

[Illustration: Fig. 366.--View of sills and veins of pale granophyre traversing dark sheets of gabbro, west side of St. Kilda.

(From a photograph by Colonel Evans.)]

"The acid rock, though styled granophyre above, belongs to a granitoid variety of that group of rocks, and has but little indication of micrographic structures. Compared with the other granophyres from St. Kilda, sliced and examined, these examples show a less acid composition. This is expressed mineralogically in the presence of a somewhat larger proportion of ferro-magnesian minerals and of soda-lime felspar. These features might indeed be matched in many normal granophyres among the Western Isles, but in the present case it can hardly be doubted that they are to be explained, at least in some degree, by the acid magma having taken up a certain amount of material from the basalt. Many of these Tertiary granophyres have undoubtedly been modified by the incorporation of pieces of basalt and gabbro, and a collection made in the Strath district of Skye will furnish examples for future study. Professor Sollas's description of similar phenomena in the Carlingford district has already proved the importance of this kind of action.[412] In the present instance, both brown mica and hornblende occur plentifully in the granophyre, and especially round the basalt fragments. This latter point is conclusive as to the derivation of the basic material, and further proves a certain degree of viscosity in the acid magma at the time of its intrusion."

[Footnote 412: _Trans. Roy. Irish Acad._ vol. xxx. (1894), pp. 477-572.]

Another series of specimens which I collected in the following year was submitted to Mr. Harker for petrographical determination, and his observations on two of the microscopic slices are as follow: "A breccia from the South Bay, St. Kilda [7105], consists of angular fragments up to two inches in diameter set in a matrix of grey granophyre of medium texture. The fragments belong to two types--one of very close texture (basalt), the other more evidently crystalline (diabase). Both are cut by the slice.

"The basalt shows very evident metamorphism, its augite being wholly transformed into greenish-brown hornblende. The little felspar-laths and granules of iron-ore seem to be unaltered, though the latter may perhaps have contributed to the formation of the hornblende. Another fragment of basalt has some larger crystal-grains of augite, and these are not converted into hornblende.

"The diabase shows a less marked boundary under the microscope, but otherwise has similar characters to the preceding. The striated felspar-crystals and grains of iron-ore have not been re-crystallized. A considerable amount of pale augite remains, but there is also plenty of deeply-coloured hornblende, both fibrous and compact. This diabase is certainly an intrusive rock, but the basalt, from its petrographic character, might be from a lava-flow or from a dyke.

"The granophyre is of somewhat coarse texture, the micrographic structure being only of a rude type. It is notably richer in the darker constituents than is usual in such rocks. Further, the hornblende and magnetite tend to cluster in little patches which suggest destroyed fragments of basic rocks. A grain or two of sphene occur, a mineral foreign to the normal granophyres.

"Another similar specimen [7106] from the same locality shows a basic rock of coarser texture, approaching some of the gabbros in appearance and with boundaries in places not very sharply defined. The grey matrix is again relatively rich in the dark elements, and the manner in which they occur in little patches, like nearly obliterated 'xenoliths,' points unmistakably to a certain amount of absorption of basic material by the acid magma, with consequent enrichment in the ferro-magnesian minerals.

"The slice cuts only the acid rock, which is seen to be of granitoid rather than granophyric structure, though the tendency of the felspar to enclose quartz-grains is unlike a typical granite. Oligoclase, with combined albite- and Carlsbad-twinning, is well represented in addition to orthoclase, and some zoned crystals seem to be of albite with a border of oligoclase. Brown hornblende and a little brown mica are the coloured constituents. Magnetite and apatite are also observed."

The testimony of the rocks of St. Kilda to the posteriority of the granophyre to the gabbros and basalts is thus clear and emphatic. It entirely confirms my previous observations regarding the order of sequence of these rocks in Mull, Rum and Skye. But the St. Kilda sections display, even more strikingly than can be usually seen in these islands, the intricate network of veins which proceed from the granophyre, the shattered condition of the basic rocks which these veins penetrate, the remarkable liquidity of the acid magma at the time of its intrusion, and the solvent action of this magma on the basic fragments which it enveloped.

3. _The Basic Dykes._--Reference has already been made to the numerous dykes by which the gabbros of the St. Kilda group of islets is traversed. Similar dykes occur also, though less plentifully, in the granophyre. It remains for future observation to determine whether there is one series older and another later than the intrusion of the acid rock. In any case, it is quite certain that the dykes in the gabbro do not all belong to one period of injection, for frequent examples of intersection may be noticed, especially on the cliffs of Borrera, and also cases of double and even treble dykes which have been formed by successive infillings within the same fissure. The remarkably varied precipices of that island are marked by the long narrow rifts left by the weathering of vertical dykes, which, as above remarked, may be followed with the eye from the sea-level to the sky-line, ascending obliquely across the bedding of the gabbro sheets. Another group of dykes may be traced sloping upward at low angles along the face of the cliffs and affording admirable ledges with overarching roofs for innumerable gannets, kittywakes and guillemots. Other dykes and ribbon-like veins may be seen traversing the gabbro in many different directions, precisely as among the Cuillin Hills. As no similar network of dykes and veins is to be observed in the granophyre, I am disposed to regard a large number of these intrusions as older than that rock. But I did not observe any actual example of a basic dyke truncated by the granophyre.

There can be no doubt, however, that an injection of similar dykes and veins took place after the invasion of the granophyre. These later intrusions are conspicuously displayed along the cliffs that extend from the gabbro junction on the north side of St. Kilda round the eastern coast into the South Bay. They maintain a general parallelism and ascend from the sea-level at varying angles of inclination, running up the pale sea-wall as dark bands. They consist of basalt-rocks, and may often be seen to branch and to die out. Like those in the gabbro, they are not infrequently compound, being made up of two or three or even more distinct dykes. This is well seen on the great precipice below Conacher, where the section given in Fig. 367 is displayed. Here in a vertical height of about 800 or 900 feet, there must be at least seven dykes, simple and compound. A little further south a triple dyke may be seen to be composed of a thick central zone and two thinner marginal bands, of which the lower strikes off from the others and maintains an independent course through the granophyre (Fig. 368).

[Illustration: Fig. 367.--Section of the sea-cliff below Conacher, St. Kilda, showing basic dykes in granophyre.]

[Illustration: Fig. 368.--Triple basic dyke, sea-cliff, east side of St. Kilda.]

V. THE GRANITE OF ARRAN

The northern half of the island of Arran is mainly occupied by one of the most compact and picturesque groups of granite mountains in Scotland.[413] These heights, rising out of the Firth of Clyde to a height of 2866 feet, present, in their spiry and serrated crests, a contrast to the smoother contours of the older granitic elevations of this country. The granite is surrounded by a ring of schistose rocks, belonging to the metamorphic series of the Southern Highlands, save for a short distance on the eastern margin, where it comes in contact with and indurates the Lower Old Red Sandstone. Macculloch long ago pointed out that no pebbles of the granite are to be found in the surrounding conglomerates and red sandstones of Carboniferous and younger age.[414] Geologists accordingly came to the conclusion that the protrusion of the granite took place after Carboniferous time, and hence that it had no connection with the appearance of the far older granites of the Highlands. In the year 1873 I gave reasons for believing the granite to be not only younger than the Carboniferous formations, but to be referable with most probability to the Tertiary volcanic series.[415] The progress of inquiry has tended to confirm this inference, though no direct proof of its correctness has been obtained. Two lines of investigation may be pursued, and each leads to the conclusion of the probability of the Tertiary age of the granite. One of these proceeds on a comparison of the petrographical characters of the Arran rocks with those of undoubted members of the Tertiary series among the Western Isles. The other inquiry deals with the relation of the rocks to each other in the general geological structure of Arran itself.

[Footnote 413: The rocks of Arran have often been described. Besides the work of Macculloch above quoted, reference may be made to the paper by Sedgwick and Murchison, _Trans. Geol. Soc._ 2nd Ser. vol. iii. p. 21; A. C. Ramsay's _Geology of the Island of Arran_, 1841, the paper of Necker de Saussure quoted on p. 412; J. Bryce's _Geology of Clydesdale and Arran_, 3rd edit. 1865. The island is at present being surveyed for the Geological Survey by Mr. W. Gunn.]

[Footnote 414: _Description of the Western Islands of Scotland_, vol. ii. p. 388.]

[Footnote 415: _Trans. Edin. Geol. Soc._ vol. ii. part iii.]

Macculloch first remarked the strong lithological resemblance of the Arran granite to the "syenite," or granophyre, of Skye and St. Kilda.[416] More recent petrographical investigation, as already stated, has furnished additional proofs of the connection between the acid rocks of these islands. So closely indeed are these rocks linked by megascopic and microscopic characters, that the petrologist has no hesitation in placing them together as probably products of the same period of igneous activity.

[Footnote 416: _Description_, vol. ii. p. 352.]

From the general geological structure of Arran, a further strong argument may be deduced in favour of the late date of the eruptions of granite. Good reasons have been given for classing as Permian the bright red sandstones which occupy much of the central and southern parts of this island, and include the little volcanic group already referred to. These sandstones have been invaded by a complex series of eruptive rocks which would thus be later than the Permian period. No igneous masses posterior to this period are certainly known in Britain save those of Tertiary age. The larger body of granite in the northern half of the island nowhere comes into direct contact with the newer red sandstones, but these strata are pierced by smaller bodies of granite. Hence, both by the evidence of their internal structure and by the stratigraphy of the ground, the later igneous rocks of Arran may be reasonably grouped together as one important and consecutive series, comparable in age and general characters with those of Tertiary date in the Inner Hebrides.

[Illustration: Fig. 369.--Jointed structure of the granite near the top of Goatfell Arran.

(From a photograph by Mr. W. Douglas, lent by the Scottish Mountaineering Club.)]

The igneous rocks of Arran, later than the probably Permian sandstones, range from acid to basic in composition. Besides the northern granite, there are in the southern part of the island acid rocks that include granite, coarse-grained quartz-porphyry and fine-grained felsite. Where the relations of these rocks to each other can be seen, the felsite is found by Mr. Gunn to be newer than the porphyry, into which it sends sills and dykes.

A feature observed by the same geologist in Arran offers a further point of resemblance to the acid sills and dykes of Skye. He has noticed that accompanying the quartz-porphyry of Drumadoon and Bennan, a mass of basic rock forms a kind of fringe or selvage round it, sometimes with what appears to be a rock of intermediate character between them. Basic sills are abundant south of Glen Ashdale, though to the west of Whiting Bay most of the intrusive sheets are of acid material.

Some of the quartz-porphyry sheets are markedly columnar. One of them, near Corriegills, displays a divergent grouping of the prisms, not unlike parts of the pitchstone sheets of Eigg and Hysgeir, and suggestive of the rock having flowed along a hollow like that of a valley. No certain trace, however, has been found of any Tertiary lava-stream in Arran, nor has evidence of tuffs been detected in any part of the younger igneous series. All the rocks appear to be intrusive, though so abundant and varied are they as to indicate that they belong to a vigorous eruptive centre, which may have poured out at the surface lavas and ashes, since entirely removed by denudation.

The numerous basic dykes for which the south end of Arran has long been celebrated have a general northerly trend, and appear to be all of the same or nearly the same age. They undoubtedly cut through the quartz-porphyries and the coarse-grained basic sills, but are less numerously visible in the finer-grained basic sills, while in the felsitic sheets they are seldom to be seen. In several places dykes running in an E.N.E. direction cut the others, and are therefore of later date.[417] The compound dykes of Tormore on the west side of the island have been already noticed (p. 161).

[Footnote 417: _Ann. Rep. of Geol. Surv._ for 1894, p. 286.]

VI. THE NORTH-EAST OF IRELAND

In the north-eastern counties of Ireland there are two regions which afford ample material for discussion in connection with the protrusion of acid rocks during the Tertiary volcanic period. One of these, which for distinction may be called the Carlingford region, embraces the tract of country which includes the Mourne Mountains on the north-east side of Carlingford Lough and the ranges of Slieve Foye and Slieve Gullion on the south-west side. The other lies mainly within the basaltic plateau, the largest of its scattered portions forming parts of the hills of Carnearny and Tardree in the county of Antrim (Map VII.).

1. The Carlingford Region

a. _The Mourne Mountains._--This compact and picturesque group of hills, about twelve miles long and six miles broad, and reaching a height of 2798 feet in Slieve Donard, presents a comparatively simple geological structure, since it consists almost entirely of granitic rocks which pierce, overlie and underlie Upper Silurian grits and shales. So far as regards the contact of these rocks with the disrupted sedimentary formations, all that can be asserted is that the granite must be later than at least the older part of the Upper Silurian period. But for at least two reasons, the eruptive rocks may be regarded with some confidence as part of the Tertiary series. In the first place, there is a strong petrographical resemblance between the Mourne Mountain granite and that of the Island of Arran and the granitic parts of the granophyre of the Western Isles. And this resemblance is so close as to furnish a cogent argument in favour of grouping all these rocks together as parts of one geologically contemporaneous series. In the second place, the Mourne Mountain granite abruptly cuts off a large number of basic dykes which, running in a general N.N.W. direction, may be looked upon as almost certainly members of the Tertiary system of protrusions.

The manner in which the granite of the district behaves towards certain detached areas of Silurian strata with their accompanying dykes is one of the most astonishing features in the whole assemblage of intrusive rocks in Britain. As has been excellently shown in the Geological Survey Map and sections by Mr. W. A. Traill, the granite has carried up on its surface broad cakes of vertical Silurian strata, together with all their network of dykes.[418] A cake of this kind, from 50 to about 200 feet thick and nearly two miles broad, has been bodily uplifted from the rest of the mass and carried upward by the granite, so that the truncated ends of the beds of grit and shale with their system of dykes stand upon a platform of granite, from which also numerous veins penetrate them. There can be little doubt that the basic dykes thus broken through are parts of the great Tertiary system, and if so, the granite which disrupts them cannot be older than Tertiary time.

[Footnote 418: See Sheets 60, 61 and 71 of the one-inch map of the Geological Survey of Ireland, and Sheets 22, 23 and 24 of the Horizontal Sections. The Explanation to these Sheets of the map was written by Professor Hull, Mr. Traill having previously retired from the service. The Mourne Mountain area is now undergoing critical revision by Prof. Sollas for the Geological Survey, and important additional material for the elucidation of this district may be expected from him.]

Besides the older basic dykes disrupted by the granite, a younger but much less abundant series traverses that rock, and also follows a general north-westerly direction. These later dykes in some cases cross more acid dykes which have risen through the granite. There is no trace of any superficial discharge from the Mourne Mountain area. But from the analogy of other districts we may easily conceive that the granite represents the underground parts of volcanic material which has now been entirely removed.

b. _Slieve Foye and Barnavave District._--This area embraces the mountainous ground lying between Carlingford Lough and Dundalk Bay, and culminating in Slieve Foye (1935 feet). It measures roughly about six miles in extreme length and four miles in breadth.

The remarkable assemblage of basic and acid materials in this area has received considerable attention from geologists. The relative order of the two groups of rocks was first clearly recognized by Griffith, who showed that the granite (granophyre) is intruded into the gabbro.[419] Professor Haughton subsequently confirmed this observation, and proved the post-Carboniferous date of the intrusive materials, which he compared with those of Skye.[420] The general distribution of the rocks was traced out in some detail by the Geological Survey, and described in the official _Memoirs_.[421] More recently the district has been examined by Professor Sollas, who, bringing the photographic camera and the microscope to the aid of field-geology, has elucidated the structure and relations of the rocks, and has obtained abundant evidence that the acid and basic rocks maintain there the same relative order as among the Inner Hebrides.[422]

[Footnote 419: _Journ. Geol. Soc. Ireland_ (1843), p. 113.]

[Footnote 420: _Quart. Journ. Geol. Soc._ vol. xii. (1856), p. 171; xiv. p. 300; and _Journ. Geol. Soc. Ireland_ (1876), p. 91.]

[Footnote 421: Sheet 71 of the Geol. Surv. Ireland, and accompanying Explanation. These were the work of Mr. W. A. Traill.]

[Footnote 422: _Trans. Roy. Irish Acad._ vol. xxx. (1894), p. 477. This is part i. of what is intended to be a series of papers.]

One of the first features in this tract of country to arrest the eye of the geologist is the situation of this centre of protrusion and that of Slieve Gullion along a north-west line, coincident with the general direction of the numerous basic dykes of the region. Whether or not the successive intrusions took place contemporaneously in the two areas, they have followed each other in the same order. In the Barnavave district the igneous rocks occupy an area of about 20 square miles. They consist of a central and chief mass composed of acid materials, which have risen through the basic rocks now found as an interrupted ring round them.

In his more recent examination, Prof. Sollas has devoted special attention to the influence of the solvent action of the acid magma upon the basic rocks and upon its own composition and structure. Besides confirming the work of previous observers as to the order of appearance of the two kinds of material, he has obtained evidence that the gabbro had not only completely solidified, but was traversed by contraction-joints, possibly even fractured by earth-movements, before the injection of the granophyric material. He found that this material, like that of the Inner Hebrides and St. Kilda, must have been in a state of great fluidity at the time of its intrusion, and made its way into the minutest cracks and crevices. In observing the solvent action of the granophyre, he ascertained that this action took place even in comparatively narrow dykes, which probably consolidated at no great depth beneath the surface.[423]

[Footnote 423: _Op. cit._]

c. _The Slieve Gullion District._--This area is separated from that just described by a narrow strip of Silurian strata, so that its isolation as a separate igneous district is complete. It will be observed from the map to continue the same north-westerly line as the Slieve Foye tract, the two together running in that direction for a distance of some 16 miles. It is interesting to note the adoption of this predominant north-westerly trend even by eruptive masses which were mainly of acid material.

This district measures about ten miles in length by from one to five miles in breadth. The rocks are, on the whole, similar to those in the area south of Carlingford Lough, and bear the same relation to each other, the acid being intrusive in the basic series. It is worthy of remark that the Tertiary eruptive rocks have made their appearance in the midst of the older granite of Newry. This granite has been already alluded to as disrupting Upper Silurian strata, and being probably of the age of the Lower Old Red Sandstone (vol. i. p. 290). In long subsequent ages, after protracted denudation, during which its cover of Silurian and Carboniferous formations was stripped off and it was laid bare, it was broken through by the whole series of basic and acid protrusions of Slieve Gullion.

This district is portrayed on Sheets 59, 60, 70 and 71 of the Geological Survey of Ireland, which show a central core of basic and acid material piercing the Newry granite.[424] Round this core and touching it at its north-western and south-eastern end, but elsewhere separated from it by a space of several miles, runs a curiously continuous band of igneous material which is marked as "quartziferous porphyry" and "felstone-porphyry" on the Survey maps.

[Footnote 424: The ground was chiefly mapped and described by Mr. Joseph Nolan and Mr. F. W. Egan.]

The south-western portion of this elliptical ring possesses a peculiar interest from its including certain remarkable masses of breccia or agglomerate. These rocks have been mapped by Mr. Nolan, and are described by him in the official _Explanation_, but in more detail in two separate papers.[425] Having had an opportunity of paying a brief visit to the ground, I can confirm the general accuracy of his mapping and description, and am able to add a few further particulars to the facts enumerated by him.

[Footnote 425: Sheet 70 of the Geol. Surv. Map of Ireland and Explanation thereto; also _Journ. Roy. Geol. Soc. Ireland_, vol. iv. (1877), p. 233; _Geol. Mag._ 1878.]

The tract of ground where these agglomerates appear forms a prominent ridge which rises several hundred feet above the lower country on either side, and extends in a W.N.W. direction for about seven miles, nearly along the line of junction between the Newry granite and the Silurian strata. The ridge has a breadth varying from a few hundred yards to upwards of a mile. It is separated from the main igneous mass of the Slieve Gullion area by an intervening strip of lower ground from three-quarters of a mile to about a mile and a half in width, which is occupied by the Newry granite. At the north-west end of the ridge the newer eruptive rocks lie within the area of that granite, while at the south-east end they rise entirely amongst the Silurian strata.

Beginning at the south-eastern extremity, we find the agglomerate occupying several detached eminences and surrounded by altered Silurian grits and shales. Further west the rock occurs in larger and more continuous masses, appearing at intervals, especially along the southern borders of the quartz-porphyry which forms by much the greater part of the ridge. Actual junctions of the agglomerate with the older rocks around seem to be seldom visible. I found one, however, above the gamekeeper's house on the southern flanks of the hill called Tievecrom. The Upper Silurian grits and shales, in a much indurated and shattered condition, are there traceable for several hundred feet up the slope, until they are abruptly cut off by the agglomerate. The line of separation appears to be nearly vertical, the truncated ends of the strata being wrapped round by the mass of fragmental material.

The most remarkable features of this agglomerate, which has been well described by Mr. Nolan, are the notable absence of truly volcanic stones in it, and the derivation of its materials from the rocks around it. I found only one piece of amygdaloid, but not a single lump of slag, no bombs, no broken fragments of lava-crusts, and no fine volcanic dust or enclosed lapilli. The rock may be said to consist entirely of fragments of Silurian grits and shales where it lies among these strata, and of granite where it comes through that rock. Blocks of these materials, of all sizes up to two feet in breadth, are confusedly piled together in a matrix made of comminuted debris of the same ingredients.

The agglomerate on the ridge of Carrickbroad has no definite boundary, but seems to graduate into an andesitic rock, and then into a quartz-felsite or rhyolite. This apparent gradation is one of the most singular features of the ridge. The andesite resembles some of the "porphyrites" of the Old Red Sandstone. It is close-grained, with abundant minute felspar-laths, and numerous large porphyritic felspars, which latter are sometimes aggregated in patches, as in the old porphyries of Portraine, Lambay Island and the Chair of Kildare. This rock has undoubtedly been erupted at the time of the formation of the agglomerate, or at least before the loose materials were compacted together; for it is full of separate stones of the same materials, and becomes so charged with them as to become itself a kind of agglomerate, with a small proportion of andesitic matrix cementing the blocks.

A thin slice prepared from one of the specimens obtained by me from this hill has been studied by Mr. Watts, who reports that the fine-grained andesitic matrix in which the stones are imbedded has often been injected into their minute fissures, and that the minute fragments enclosed in this matrix consist here of a trachyte-like porphyry, felsite, andesites, basalts of various degrees of fineness and olivine-basalt, together with isolated grains of felspar, such as might have been derived from the breaking up of some of these fragments.

Westward from Carrickbroad, the chief eruptive rock is a dark, sometimes nearly velvet-black, flinty, occasionally almost resinous, quartz-porphyry or rhyolite, with abundant quartz and large felspars and occasional well-marked flow-structure. This material, near the much smaller protrusion of andesite, is curiously mixed up with that rock, as if the two had come up together. Sometimes they seem to pass into each other, at least the separation between them cannot be sharply drawn. There can be little doubt, however, that the acid magma continued to ascend after the other, for it sends veins and strings into the more basic material, and encloses blocks of it. This thoroughly acid porphyry plays the same part as the andesite in regard to the stones of the agglomerate. Throughout its whole extent, it is found to enclose these stones, which here and there become so numerous as to form the main bulk of the mass, leaving only a limited amount of quartz-porphyry (rhyolite) matrix to bind the whole into an exceedingly compact variety of breccia. Occasionally the acid rock cuts through the ordinary clastic agglomerate, as may be well seen on the southern face of Tievecrom.

A specimen of this porphyry with its enclosed fragments, which was collected by me from above the old tower at Glendovey, Carrickbroad, has been sliced and examined by Mr. Watts under the microscope, and is thus described by him: "The large fragment in this slide consists of ophitic olivine-dolerite full of large phenocrysts of olivine. It is broken up and penetrated by veins of quartz-porphyry, rich in quartz, which exhibits a beautiful flow-structure. The felspars and augite of the dolerite do not appear to have suffered much alteration at the margin of the fragment, but the olivines are much serpentinized, the serpentine passing into a border of actinolite which runs in veins into the neighbouring rock and even passes out into the quartz-porphyry at the junction, impregnating it with actinolite and chlorite for some distance. A few particles of basalt also occur and a portion of a granite-fragment comes into the slide, from the edge of which a piece of biotite has floated off into the quartz-porphyry."

The essentially non-volcanic material of the agglomerate shows, as Mr. Nolan pointed out, that it was produced by æriform explosions, which blew out the Silurian strata and granite in fragments and dust. These discharges probably took place either from a series of vents placed along a line of fissure running in a north-westerly line, or directly from the open fissure itself. Possibly both of these channels of escape were in use; detached vents appearing at the east end and a more continuous discharge from the fissure further west.

After the earliest explosions had thrown out a large amount of granitic and Silurian detritus, andesitic lava rose in the fissure, and solidifying there enclosed a great deal of the loose fragmentary material that fell back into the chasm. Subsequently, and on a more extensive scale, a much more acid magma ascended from below, likewise involving and carrying up a vast quantity of loose stones, among which are pieces of basalt and dolerite.

No evidence remains as to the extent of the material discharged over the surface from this fissure. Denudation has removed all the surrounding fragmental sheets as well as any lava that may have flowed out upon or become intercalated among them. There remains now only the cores of the little necks at the east end, and the indurated agglomerate and lava that consolidated along the mouth of the fissure or vents.

This is the only example of such a line of fissure-eruption which has yet been met with in the British Isles. Its connection with the eruptive masses of Slieve Gullion and Carlingford links it with the Tertiary volcanic series. But no evidence appears to remain regarding the epoch in the long volcanic period when the eruptions from it took place. They may possibly date back to the time of the plateau-basalts; but the abundant acid magma, which constitutes one of their distinguishing characteristics, suggests that they more probably belong to the later time when the main protrusions of acid material took place. They suggest that coeval with the uprise of the great domes of Slieve Gullion, Carlingford and the Mourne Mountains there may have been many superficial eruptions of which, after prolonged denudation, all trace has now been effaced.

2. The Antrim Region

Reference was made in