CHAPTER XXXV

THE SYSTEM OF DYKES--_continued_

Direction--Termination upward--Known vertical Extension--Evidence as to the movement of the Molten Rock in the Fissures--Branches and Veins--Connection of Dykes with Intrusive Sheets--Intersection of Dykes--Dykes of more than one infilling--Contact Metamorphism of the Dykes--Relation of the Dykes to the Geological Structure of the Districts which they traverse--Data for estimating the Geological Age of the Dykes--Origin and History of the Dykes.

9. DIRECTION

Another characteristic feature of the dykes is their generally rectilinear course. So true are the solitary dykes to their normal trend that, in spite of varying inequalities of surface and wide diversities of geological structure in the districts which they traverse, they run over hill and dale almost with the straightness of lines of Roman road. In the districts where the gregarious type prevails, the dykes depart most widely from the character of the great solitary series, but still tend to run in straight or approximately straight lines, or, if wavy in their course, to preserve a general parallelism of direction.

Yet even among the great persistent dykes instances may be cited where the rectilinear trend is exchanged for a succession of zig-zags, though the normal direction is on the whole maintained. In such cases, it is evident that the fissures were not long straight dislocations, like the larger lines of fault in the earth's crust, but were rather notched rents or cracks which, though keeping, on the whole, one dominant direction, were continually being deflected for short distances to either side. As a good illustration of this character, reference may be made to the Cheviot and Hawick dyke. In Teviotdale, this dyke can be followed continuously among the rocky knolls, so that its deviations can be seen and mapped. From the median line of average trend the salient angles sometimes retire fully a quarter of a mile on either side. Some examples of the same feature may be noticed in the Eskdale dyke. The large dyke which runs westward from Dunoon has been observed by Mr. Clough to change sharply in direction three times in four miles, running occasionally for a short distance at a right angle to its general direction (see Fig. 257).

Among these solitary dykes also, though the persistence of their trend is so predominant, there occur instances where the general direction undergoes great change. Some of the most remarkable cases of this kind have been mapped by Mr. B. N. Peach and Mr. R. L. Jack, in the course of the Geological Survey of Perthshire. Several important dykes strike across the Old Red Sandstone plain for many miles in a direction slightly south of west. But when they approach the rocks of the Highland border in Glen Artney, they bend round to south-west, and continue their course along that new line.

Many years ago I called attention to the dominant trend of the dykes from north-west to south-east.[185] Subsequent research has shown this to be on the whole the prevalent direction throughout the whole region of dykes. But the detailed mapping, carried on by my colleagues and myself in the Geological Survey, has brought to light some curious and interesting variations from the normal trend. In the districts where dykes of the gregarious type abound there is sometimes no one prevalent direction, but the dykes strike to almost all points of the compass. Of the Arran dykes, so carefully catalogued by Necker, only about a third have a general north-westerly course. But in Eastern Argyleshire the abundant dykes mapped by Mr. Clough trend almost without exception towards N.N.W. In the North of Ireland, Berger found the direction of thirty-one dykes to vary from 17° to 71° W. of N., giving a mean of N. 36° W.[186] In Islay, Jura, Eigg, Mull, and Skye the mean of several hundred observations has given me similar results. Among the Inner Hebrides, however, though the general north-westerly trend is characteristic, many of the later dykes show marked departures from it. Thus in Strath, Skye, some of the youngest follow a nearly north and south direction (Fig. 253). In the Blath Bhein hill-range, Mr. Harker has found that the latest dykes cut the gabbro at right angles to the prevalent trend and are further distinguished by their low hade.

[Footnote 185: _Trans. Roy. Soc. Edin._ xxii. (1861), p. 650.]

[Footnote 186: _Trans. Geol. Soc._ iii. p. 225.]

It appears, therefore, that though there is sometimes extraordinary local diversity in the direction of the dykes in those districts where they present the gregarious type, the general north-westerly trend can usually still be recognized. But when we turn to the long massive solitary dykes, we soon perceive a remarkable change in their direction as we follow them northward into Scotland. I formerly pointed out how the general north-westerly trend becomes east and west in the Lothians, with a tendency to veer a little to the south of west and north of east.[187] This departure from the normal direction is now seen to be part of a remarkable radial arrangement of the dykes. Beginning at the southern margin of the dyke-region, we have the notable example of the Cleveland dyke, which in its course from Cleveland to Carlisle runs nearly W. 15° N. The Eskdale dyke has an average trend of W. 32° N., and the same general direction is maintained by the group of dykes which run from the Southern Uplands across the south-west of Lanarkshire and north-east of Ayrshire. But proceeding northwards we observe the trend to turn gradually round towards the west. The dyke that runs from near the mouth of the Coquet across the Cheviot Hills to beyond Hawick has a general course of W. 8° N. In the great central coal-field of Scotland the average direction may be taken to be nearly east and west, the same dyke running sometimes to the north, and sometimes to the south of that line. But immediately to the north a decided tendency to veer round southwards makes its appearance. Thus the long dyke which runs from the Carse of Stirling through the Campsie Fells to the Clyde west of Leven, has a mean direction of W. 5° S. This continues to be the prevalent trend of the remarkable series of dykes which crosses the Old Red Sandstone plains, though some of these revert in whole or in part to the more usual direction by keeping a little to the north of west. Even as far as Loch Tay and the head of Strathardle, the course of the dykes continues to be to the south of west. Tracing these lines upon a map of the country we perceive that they radiate from an area lying along the eastern part of Argyleshire and the head of the Firth of Clyde (see Map I.).

[Footnote 187: _Trans. Roy. Soc. Edin._ xxii. p. 651.]

10. TERMINATION UPWARDS

It was pointed out many years ago by Winch that some of the dykes which traverse the Northumberland coal-field do not cut the overlying Magnesian Limestone. The Hett dyke, south of Durham, is said to end off abruptly against the floor of the limestone.[188] Here and there, among the precipices of the Inner Hebrides, a dyke may be seen to die out before it reaches the top of the cliff. But in the vast majority of cases, no evidence remains as to how the dykes terminated upwards. I have referred to the occasional interruptions of the continuity of a dyke, where, though the rock does not reach the surface, it must be present in the fissure underneath. Such interruptions show that, in some places at least, there was no rise of the rock even up to the level of what is now the surface of the ground, and that the upward limit of the dykes must have been exceedingly irregular.

[Footnote 188: This is expressed in the Geological Survey Map, Sheet 93, N.E.]

Excellent illustrations of this feature are supplied by sections on the line of the Cleveland dyke. Towards its south-easterly extremity, this great band of igneous rock ascends from the low Triassic plain of the Tees into the high uplands of Cleveland. Its course across the ridges and valleys there has been carefully traced for the Geological Survey by Mr. G. Barrow, who has shown that over certain parts of its course it does not reach the surface, but remains concealed under the Jurassic rocks, which it never succeeded in penetrating. But that in places it comes within a few feet of the soil is shown by the baked shale at the surface, for the alteration which it has induced on the surrounding rocks only extends a few feet from its margin. These interruptions of continuity show how uneven is the upper limit of the dyke. The characteristic porphyritic rock may be observed running up one side of a hill to the crest, but never reaching the surface on the other side. At Cliff Ridge, for example, about three miles south-west of Guisbrough, Mr. Barrow has followed it up to the summit on the west side; but has found that on the east side it does not pierce the shales, which there form the declivity. This structure is represented in Fig 241. The vertical distance between the summit to the left, where the dyke (_b_) disappears, and the point to the right, where the Lias shale (_a_) of the hillside is concealed by drift (_c_), amounts to 250 feet, the horizontal distance being a little more than 900 feet. But as the shale when last seen at the foot of the slope is quite unaltered, the dyke must there be still some little distance beneath the surface, so that the vertical extension of this upward tongue of the dyke must be more than 250 feet. Mr. Barrow, to whom I am indebted for these particulars, has also drawn the accompanying section (Fig. 242) along the course of the dyke for a distance of nearly 11 miles eastward from the locality represented in Fig. 241. From this section it will be observed that in that space there are at least three tongues or upward projections of the upper limit of the dyke. Several additional examples of the same structure are to be seen further east towards the last visible outcrop of the dyke.

Another feature connected with the upward termination of the dyke is well seen in some parts of the ground through which the two foregoing sections are taken. Mr. Barrow informs me that at Ayton a level course has been driven into the hill for mining operations, at a height of 400 feet above sea-level, and the dyke has there been ascertained to be 80 feet broad. Higher on the hill, close to the 750 feet contour--line, its breadth is only 20 feet, so that it narrows upward as much as 60 feet in a vertical height of 350 feet. Its contraction in width during the last twenty feet is still more rapid, and in the last few yards it diminishes to two or three feet, and has a rounded top over which the strata are bent upward. The accompanying section (Fig. 243) across the upper part of the dyke will make these features clear.

Further to the west an exposure of the upper limit of the dyke has been described and figured by Mr. Teall. In 1882, at one of the Cockfield quarries (Fig. 244), the dyke was "seen to terminate upwards very abruptly in the form of a low and somewhat irregular dome, over which the Coal-measure shales passed without any fracture, and only with a slight upward arching."[189]

[Footnote 189: _Quart. Jour. Geol. Soc._ xl. p. 210.]

Near the other or north-western termination of this great dyke, similar evidence is found of an uneven upper limit. After an interrupted course through the Alston moors, the dyke reaches the ground that slopes eastward from the edge of the Cross Fell escarpment. Its highest visible outcrop is at a height of 1700 feet. But westwards from that point the dyke disappears under the Carboniferous rocks, and does not emerge along the front of the great escarpment that descends upon the valley of the Eden, where among the naked scarps of rock it would unquestionably be visible if it reached the surface. Its upper edge must rapidly descend somewhere behind the face of the escarpment, for the igneous rock crops out a little to the west of the foot of the cliff, about 1000 feet below the point where it is last seen on the hills above. Here the top of the dyke has a vertical drop of not less than 1000 feet, in a horizontal distance of five miles, as shown in Fig. 245, which has been drawn for me by Mr. J. G. Goodchild.

It will be observed that in these sections (Figs. 241, 242 and 245) there is a curiously approximate coincidence between the inequalities in the upper surface of the dyke and those in the form of the overlying ground. The coincidence is too marked and too often repeated to be merely accidental. Whether the ancient topographical features had any influence in determining, by cooling or otherwise, the limit of the upward rise of the lava, or whether the dyke, even though concealed, has affected the progress of the denudation of the ground overlying it, is a question worthy of fuller investigation.

11. KNOWN VERTICAL EXTENSION

Closely connected with the determination of the upper limit reached by the dykes, is the total vertical distance to which they can be traced. Of course, the depth of the original reservoir of molten rock which supplied them remains unknown, and probably undiscoverable. But it is possible, in many cases, to determine at least the inferior limit of the thickness of rock through which the molten material of the dykes has ascended. Along the great basalt-escarpments of Mull and Skye, the ascent of dykes from base to summit may often be observed. Thus, on the cliffs of Dunvegan Head, on the west coast of Skye, which rise out of the sea to a height of about 1000 feet, several dykes may be observed rising through the whole series of basalts up to the crest of the precipice. In the dark gabbro hills of the same island, numerous dykes may be seen climbing from the glens right up the steep rugged acclivities and over the crests, through a vertical thickness of more than 3000 feet of rock (Fig. 333). The dykes which cross Loch Lomond, and ascend the hills on either side of that deep depression, must rise through at least as great a thickness. But where a knowledge of the geological structure of the ground enables us to estimate the bulk of the successive rock-formations which underlie the surface, it can be shown that the lava ascended through a much greater depth of rock. Measurements of this kind can best be made towards the eastern end of the Cleveland dyke, where the different sedimentary groups have not been seriously disturbed, and where, from natural sections and artificial borings, their thicknesses are capable of satisfactory computation. The highest bed of the Jurassic series anywhere touched by the dyke is the Cornbrash. It is certain, therefore, that the igneous rock rises through all the subjacent members of the Jurassic series up to that horizon. There can be no doubt also that the Trias and Magnesian Limestone continue in their normal thickness underneath the Jurassic strata. To what extent the Coal-measures exist under Cleveland has not been ascertained; possibly they have been entirely denuded from that area, as from the ground to the west. But the Millstone Grit and Carboniferous Limestone probably extend over the district in full development; and below them there must lie a vast depth of Upper and Lower Silurian strata, probably also of still older Palæozoic rocks and beneath all the thick Archæan platform. Tabulating these successive geological formations, and taking only the ascertained thickness of each in the district, we find that they give the results shown in the subjoined table.[190]

[Footnote 190: Drawn up for me by Mr. G. Barrow.]

STRATA CUT BY THE CLEVELAND DYKE

Cornbrash-- Feet. Lower Oolite and Upper Lias, as proved by bore-hole on Gerrick Moor, 950 Middle and Lower Lias, ascertained from measurement of cliff-sections and from mining operations to be more than 850 New Red Sandstone and Marl, found by boring close to the Tees to exceed 1,600 Magnesian Limestone, at least 500 Coal-measures, possibly absent 0 Millstone Grit, not less than 500 Carboniferous Limestone series at least 3,000 Silurian rocks, probably not less than 10,000 ------- 17,400

There is thus evidence that this dyke has risen through probably more than three miles of stratified rocks. How much deeper still lay the original reservoir of molten material that supplied the dyke, we have at present no means of computing.

12. EVIDENCE AS TO MOVEMENT OF THE MOLTEN ROCK IN THE FISSURES

It is usual to speak of the molten material of the dykes as having risen vertically within the fissures. And doubtless, on the whole, the expression is sufficiently accurate. In the case of such long dykes as those of Central Scotland and the North of England, where the petrographical character of the material remains so uniform throughout, it is obvious that the andesite or dolerite cannot have come from a mere single pipe like a volcanic orifice. Nor can we easily understand how it could have been supplied even from a series of such pipes. The general aspect and structure of the dykes suggest that the fissures were rent so profoundly in the crust of the earth as to reach down to a reservoir of molten rock which straightway rose in them. The roof of such a reservoir, however, may have been irregular and uneven, so that a fissure need not have traversed it continuously, but may have only touched its upward projecting vaults. Hence gaps would arise in the continuity of the dyke-material.

The ascent of lava from a line of such separate openings along a fissure would necessarily involve lateral as well as vertical movements in the molten mass which would be forced along the open rent until the several streams united and filled it up. We might therefore expect somewhere to find instances of flow-structure in the dykes pointing to these movements. I have already referred to the lines of amygdales frequently noticed in dykes, especially towards the centre. Occasionally these steam-vesicles may be observed to be drawn out in one general direction indicative of the trend of motion of the molten rock.

Some of the best examples of this feature which have come under my observation occur among the trachytic dykes of the south-east coast of Skye between Kyle Rhea and Loch na Daal, where they have been mapped and carefully investigated by Mr. Clough, who has conducted me over the sections. In some of these dykes, as already narrated, the marginal portions display a finely spherulitic structure, the small pea-like spherulites being grouped into fine ribs or rods. It is also observable that the steam-vesicles which may retain their spherical forms in the centre are elongated in the same direction as the rows of spherulites. Where this lineation is developed vertically, it no doubt points to the vertical ascent of the lava between the two walls of the fissure.

But in other examples, the elongation is nearly horizontal, and between the two positions Mr. Clough has registered many intermediate trends. It would thus appear that in some places the lava has certainly flowed laterally between the fissure-walls. Moreover, the trend of the spherulitic rods and of the amygdales is found to vary in closely adjoining planes at different distances from the margin, as if after the outer portions of the dyke had consolidated into position, there was still movement enough to drag the rows of spherulites and vesicles up or down along the trend of the fissure.

Mr. Clough has observed that in some dykes, while the amygdaloidal vesicles are large and undeformed in the centre, they become elongated and inclined downward in the direction of the margin, as if the central portions had not only remained fluid longer than the rest, but had a tendency to rise upwards in the fissure, though there was obviously less motion after these central vesicles appeared than in the marginal parts where the vesicles are so much drawn out.

13. BRANCHING DYKES AND VEINS

It might have been anticipated that the uprise of such abundant masses of molten rock, in so many long and wide fissures, would generally be attended with the intrusion of the same material into lateral rents and irregular openings, so that each dyke would have a kind of fringe of offshoots or processes striking from it into the surrounding ground. It might have been expected also that dykes would often branch, and that the arms would come together again and enclose portions of the rocks through which they rise. But in reality such excrescences and bifurcations are of comparatively rare occurrence. As a rule, each dyke is a mere wall of igneous rock, with little more projection or ramification than may be seen in a stone field-fence. Among the short, narrow and irregular dykes of the gregarious type branchings are occasionally seen, and in some districts are extraordinarily abundant. But among the great single dykes such irregularities are far less common than might have been looked for. A few characteristic examples from each type of dyke may here be given.

The Cleveland dyke, which in so many respects is typical of the great solitary dykes of the country, has been traced for many miles without the appearance of a single offshoot of any kind. Yet here and there along its course, it departs from its usual regularity. As it crosses the Carboniferous tracts of Durham and Cumberland, there appear near its course lateral masses of eruptive rock, most of which doubtless belong to the much older "Whin Sill." But there is at least one locality, at Bolam near Cockfield, in the county of Durham, where the dyke, crossing the Millstone Grit, suddenly expands into a boss, and immediately contracts to its usual dimensions. Around this knot several short dykes or veins seem to radiate from it. The dyke has been quarried here, and its relations to the surrounding strata have been laid bare, as will be again referred to a little further on.[191]

[Footnote 191: This locality was well described by Sedgwick, in his early paper on Trap-Dykes in Yorkshire and Durham, _Trans. Cambridge Phil. Soc._ ii. p. 27.]

Among the great persistent dykes of Scotland the absence of bifurcation and lateral offshoots offers a striking contrast to the behaviour of the dykes in those districts where they are small in size and many in number. But exceptions to the general rule may be gathered. Thus the Eskdale dyke is flanked at Wat Carrick with a large lateral vein, which is almost certainly connected with the main fissure. The Hawick and Cheviot dyke splits up on the hill immediately to the east of the town of Hawick, sends off some branches, and then resumes its normal course (Fig. 246). Again, one of the two nearly parallel dykes which run from Lochgoilhead across Ben Ledi into Glen Artney bifurcates at the foot of that valley, its northern limb (about two miles long) speedily dying out, and its southern branch throwing off another lateral vein, and then continuing eastward as the main dyke (Fig. 247).

In the districts of gregarious dykes, however, abundant instances may be found of dykes that branch, and of others that lose the parallelism of their walls, become irregular in breadth, direction, and inclination, so as to pass into those intrusive forms that are more properly classed as veins. Excellent illustrations of bifurcating dykes may be observed along the shores of the Firth of Clyde, particularly on the eastern coast-line of the isle of Arran. The venous character has become familiar to geologists from the sketches given by Macculloch from the lower parts of the cliffs of Trotternish in Skye.[192] Still more striking examples are to be seen in the breaker-beaten cliffs of Ardnamurchan. The pale Secondary limestones and calcareous sandstones of that locality are traversed by a series of dark basic veins, and the contrast of tint between the two kinds of rock is so marked as even to catch the eye of casual tourists in the passing steamboats. The veins vary in width from less than an inch to several feet or yards. They run in all directions and intersect each other, forming such a confused medley as requires some patience on the part of the geologist who would follow out each independent ribbon of injected material in its course up the cliffs, or still more, would sketch their ramifications in his note-book. A good, though perhaps somewhat exaggerated, illustration of their general character was given by Macculloch.[193] The accompanying figure (Fig. 248) is less sensational, but represents with as much accuracy as I could reach, the network of veins near the foot of the cliffs. One conspicuous group of veins, which, seen from a distance, looks like a rude sketch of a lug-sail traced in black outline upon a pale ground, is known to the boatmen as "M'Niven's Sail." Another admirable locality for the study of dykes and tortuous veins is the northern coast of the Sound of Soa, where an extraordinary number of injections traverse the Torridon Sandstones on which the plateau-basalts rest (Fig. 323).

[Footnote 192: _Western Islands_, plate xvii.]

[Footnote 193: _Op. cit._ plate xxxiii. Fig. 1.]

As a general rule, the narrower the vein the finer in grain is the rock of which it consists. This compact dark homogeneous material has commonly passed by the name of "basalt." Its minuteness of texture probably in most cases arises from local rapidity of cooling, and it is doubtless the same substance which, where in larger mass in the immediate neighbourhood, has solidified as one of the other pyroxene-plagioclase-magnetite rocks.

With regard to the places where such abundant tortuous veins are more especially developed, I may remark that they are particularly prominent under a thick overlying mass of erupted rock, such as a great intrusive sheet, or the bedded basalts of the plateaux, or where there is good reason to believe that such a deep cover, though now removed by denudation, once overspread the area in which they appear. It will be shown in the sequel that such horizons have been peculiarly liable to intrusions of igneous material of various kinds, and at many different intervals, during the volcanic period. A thick cake of crystalline rock seems to have offered such resistance to the uprise of molten material through it, that when the subterranean energy was not sufficient to rend it open by great fissures, and thus give rise to dykes, the lavas were either forced into such irregular cracks as were made partly in the softer rocks underneath and partly in the cake itself, or found escape along pre-existing divisional planes. In Ardnamurchan, round the Cuillin Hills of Skye, and in Rum, the overlying resisting cover now consists mainly of gabbro sheets. In the east of Skye, in Eigg, and in Antrim, it is made up of the thick mass of the plateau-basalts.

14. CONNECTION OF DYKES WITH SILLS

Every field-geologist is aware how seldom he can actually find the vent or pipe up which rose the igneous rock that supplied the material of sills and laccolites. He might well be pardoned were he to anticipate that, in a district much traversed by dykes, there should be many examples of intrusive sheets and frequent opportunities of tracing the connection of such sheets with the fissures from which their material might be supposed to have been supplied. But such an expectation is singularly disappointed by an actual examination of the Tertiary volcanic region of Britain. That there are many intrusive sheets belonging to the great volcanic period with which I am now dealing, I shall endeavour to show in the sequel. But it is quite certain that though these sheets have of course each had its subterranean pipe or fissure of supply, they can only in rare instances be directly traced to the system of dykes. On the other hand, the districts where great single dykes are most conspicuous, are for the most part free from intrusive sheets, except those of much older date, like the Carboniferous Whin Sill of Durham and those of Linlithgowshire, Stirlingshire and Fife.

Yet a few interesting examples of the relation of dykes to sheets have been noticed among British Tertiary volcanic rocks. The earliest observed instances were those figured and described by Macculloch. Among them one has been familiar to geologists from having done duty in text-books of the science for more than half a century. I allude to the diagram of "Trap and Sandstone near Suishnish."[194] In that drawing seven dykes are shown as rising vertically through the horizontal sandstone, and merging into a thick overlying mass of "trap." The author in his explanation leaves it an open question "whether the intruding material has ascended from below and overflowed the strata, or has descended from the mass," though from the language he uses in his text we may infer that he was inclined to regard the overlying body as the source of the veins below it.[195]

[Footnote 194: _Western Islands of Scotland_, pl. xiv. Fig. 4.]

[Footnote 195: _Op. cit._ vol. i. pp. 384, 385.]

The section given by Macculloch, however, does not quite accurately represent the facts. The narrow dykes there drawn have no connection with the overlying sheet, but are part of the abundant series of basaltic dykes found all over Skye. The feeder of the gabbro sill was presumably the broad dyke which descends the steep bank immediately on the southern front of Carn Dearg (636 feet high). The accompanying figure (Fig. 249) shows what seemed to me to be the structure of the locality, but the actual junction of the dyke and sheet is concealed under the talus of the slope.[196] I shall have occasion in a later Chapter to refer again to this section in connection with the history of intrusive sheets, and also to cite from the neighbouring island of Raasay another good example of the same relation between dyke and sill.

[Footnote 196: In more recently surveying this ground, Mr. Harker has been led to regard the coarse sill as independent of the other intrusions, and as almost certainly later than the basalt-sheets of the same locality. When it reaches the base of these sills it turns so as to pass beneath them as a gabbro-sill, which is conspicuous near the summit of Carn Dearg. It runs westward for some distance, almost immediately breaking across the bedding so as to leave the basalt, and rapidly tapering until it dies out.]

Sedgwick, in the paper above quoted, gave an account and figure of the expansion of the Cleveland dyke at Bolam, to which allusion has already been made. He showed that from a part of the dyke which is unusually contracted a great lateral extension of the igneous rock takes place on either side over beds of shale and coal. While in the dyke the prisms are as usual directed horizontally inward from the two walls, those in the connected sheet are vertical, and descend upon the surface of highly indurated strata on which the sheet rests.

The most important examples known to me are those which occur in the coal-field of Stirlingshire. In that part of the country, the remarkable group of dykes already referred to, lying nearly parallel to each other and from half a mile to about three miles apart, runs in a general east and west direction. From one of these dykes no fewer than four sills strike off into the surrounding Coal-measures. The largest of them stretches southwards for three miles, but the same rock is probably continued in a succession of detached areas which spread westwards through the coal-field and circle round to near the two western sheets that proceeded from the same dyke. Another thick mass of similar rock extends on the north side of the dyke for two and a half miles down the valley of the river Avon. These various processes, attached to or diverging from the dyke, are unquestionably intrusive sheets, which occupy different horizons in the Carboniferous series. The one on the north side has inserted itself a little above the top of the Carboniferous Limestone series. Those on the south side lie on different levels in the Coal-measures, or, rather, they pass transgressively from one platform to another in that group of strata.

No essential difference can be detected by the naked eye between the material of the dyke and that of the sheets. If a series of specimens from the different exposures were mixed up, it would be impossible to separate those of the dyke from those of the sheets. A microscopical examination of the specimens likewise shows that they are perfectly identical in composition and structure, being chiefly referable to rocks of the dolerite, but partly of the tholeiite type. I have therefore little doubt that these remarkable appendages to this dyke are truly offshoots from it, and are not to be classed with the general mass of the sills of Central Scotland, which are of Carboniferous, partly of Permian, age. The accompanying diagrammatic section (Fig. 250) explains the geological structure of the ground.

An interesting and important fact remains to be stated in connection with these sheets. They are traversed by some of the other east and west dykes. This is particularly observable in the case of the sheet which extends northwards from the dyke through the parish of Torphichen. Two well-marked dykes can be seen running westwards among the ridges of the sheet. It is obvious, therefore that these particular dykes are younger than the sheet. But, as will be shown in the sequel, there is abundant evidence that all the dykes of a district are not of one eruption. The intersection of one eruptive mass by another does not necessarily imply any long interval of time between them. They mark successive, but it may be rapidly successive, manifestations of volcanic action. Hence the cutting of the sheets by other dykes does not invalidate the identification of these sheets as extravasations from the great dyke by which they are bounded.

15. INTERSECTION OF DYKES

Innumerable instances may be cited, where one dyke, or one set of dykes, cuts across another. To some of these I shall refer in discussing the data for estimating the relative ages of dykes. In considering the intersection from the point of view of geological structure, we are struck with the clean sharp way in which it so generally takes place. The rents into which the younger dykes have been injected seem, as a rule, not to have been sensibly influenced in width and direction by the older dykes, but go right across them. Hence the younger dykes retain their usual breadth and trend (Fig. 251). In trying to ascertain the relative ages of such dykes we obtain a valuable clue in studying the respective "chilled edges" of the two intersecting masses, as has already been pointed out.

Not only do dykes cross each other, but still more is this the case among the narrower tortuous intrusions known as Veins (Fig. 252). Among the illustrations which the dykes of the Inner Hebrides supply of these features one further characteristic example may be culled from the shore of Skye, near Broadford, where the gently-inclined sheets of Lias limestone are traversed by three systems of dykes (Fig. 253). One of these systems runs in a N.W. or N.N.W. direction, a second follows a more nearly easterly trend, while the third and youngest runs nearly north and south.

16. DYKES OF MORE THAN ONE IN-FILLING

The intersections of dykes prove that the process of fissuring in the earth's crust took place at more than one period, and prepare us for the reception of evidence that the same line of fissure might be again re-opened, even after it had been filled with molten material. Numerous instances have now been accumulated in which dykes are not single or simple intrusions, but where the original dyke-fissure has been re-opened and has been invaded by successive uprisings of lava.[197] Compound dykes have thus been formed, consisting of two or more parallel bands of similar or dissimilar rock.

[Footnote 197: See an example figured by Macculloch, _Western Isles_, plate xviii. Fig. 1.]

While it is not difficult to conceive of the re-opening of a vertical fissure during terrestrial strain, and the injection into it of later intrusions of a volcanic magma, it is not so easy to understand the mechanism where the line of weakness has been slightly inclined or horizontal, and where, consequently, there has been the enormous superincumbent pressure of the overlying part of the earth's crust to overcome. Yet gently inclined compound dykes exhibit their parallel bands with hardly less regularity than do those that are vertical. The difficulty of explanation is felt most strongly in the attempt to realize the origin of the compound sills described in Chapter xlviii.

In the re-opening of dyke-fissures the later intrusions have generally taken place along the walls, or where the dykes were already compound, between some of the component bands. Less frequently the first dyke has been split open along the middle, and a second injection has forced its way along the rent.

Of the first of these two types, numerous instances have now been observed in the West of Scotland. If the portion of a compound dyke exposed at the surface be limited in extent, we may be unable to determine which is the older of two parallel bands of igneous rock, though the fact that they present to each other the usual fine-grained edge due to more rapid cooling, shows that they are not one but two dykes, belonging to distinct eruptions. So far as I have noticed, where one of the dykes can be continuously traced for a considerable distance, the other is comparatively short. I infer that the shorter one is the younger of the two.

In the Strath district of Skye, Mr. Harker has recently observed that many of the basic dykes, both those older and those younger than the granophyre protrusions, are double, triple or multiple. Thus in a conspicuous dyke, more than 100 feet wide, to the south-east of Loch Kilchrist, belonging to the older series, he has detected at least six contiguous dykes which as they are traced south-eastward, in spite of their interruption by the Beinn an Dubhaich granite, can be seen to separate and take different courses, or successively die out. He remarks, further, that "many cases of apparent bifurcation of dykes are really due to the separation of distinct dykes which have run for some distance in one fissure. Sometimes apparent variations in the width of a dyke are to be explained by this dying out of one member of a double dyke. These multiple dykes are less easily detected in the newer than the older set, owing to greater uniformity of lithological type in the prevalent kinds and to the frequent absence of chilled selvages."[198] An example of a compound basic dyke cutting the crest of the gabbro-mass of the Cuillin Hills is shown in Fig. 333.

[Footnote 198: MS. notes supplied by Mr. Harker.]

Instances of the second type of compound dykes are less common. Here, instead of being re-opened along one of the walls, the fissure has been ruptured along the centre of the dyke, and a second injection of molten material has then taken place. This structure may be observed where the materials of the compound dyke are on the whole similar, such as varieties of dolerite, basalt, diabase or andesite. In these cases the rock of the central dyke is generally rather fine-grained, sometimes decidedly porphyritic, and often a true basalt. Where broad enough to show the difference of texture between margin and centre, it exhibits the usual close grain along its edges, indicative of quicker cooling. The older dyke presenting no such change at its junction with the younger, was obviously already cooled and consolidated before its rupture.

Whilst the centre of a dyke has occasionally proved to be a line of weakness which has given way under intense strains in the terrestrial crust, this rupture and the accompanying or subsequent ascent of molten material in the re-opened fissure may sometimes have been included as phases of one connected volcanic episode. In those instances, for example, which have been above described, where a central vitreous band has risen along the heart of a dyke, the petrographical affinities of the rocks may be so close as to suggest that although the main dyke had consolidated and had subsequently been ruptured along its centre by powerful earth-movements, these changes all belonged to the same period of dyke-making, and the subsequent uprise of glassy material was merely a later phase in the movements of the same subterranean magma.

But where, as probably happens in the large majority of compound dykes, there is a strongly marked difference between the respective bands of rock, we must either infer that two essentially different magmas co-existed in the volcanic reservoirs underneath, and were successively injected into the same fissures, or that a sufficient lapse of time occurred to permit a total renewal of the nature of the magma, and an uprise of this changed material into fissures which sometimes coincided with older dykes. If any interlocking of the crystals of the several bands of a compound dyke could be detected, we might suppose that the first-injected material had not become consolidated and cold before the uprise of the newer rock. But in general it would seem that so sharp a line of demarcation can be drawn between the two rocks as to indicate that their protrusion was due to two distinct and perhaps widely-separated volcanic paroxysms.

Compound dykes of basic material occur not only among the ordinary straight north-westerly series, but also among the less regular gregarious dykes and veins, such as abundantly intersect the gabbro bosses. Moreover they are to be found among the youngest intrusions, for they traverse the masses of granophyre. Conspicuous examples of such late compound dykes are displayed along the cliffs of St. Kilda, as will be more particularly described in a later Chapter. These St. Kilda dykes often occupy not vertical fissures but parallel rents with a gentle inclination (see Figs. 367, 368).

The Tertiary volcanic series of Scotland furnishes many examples of compound dykes of a much more complex character where parallel bands of some acid (granophyre, felsite, quartz-porphyry) or intermediate (andesite) rock is associated with others of the more usual basic material (dolerite, basalt, diabase). As the acid intrusions belong to a comparatively late part of the volcanic history, their modes of occurrence will be discussed in Chapters xlvi., xlvii. and xlviii. But no account of the general system of dykes would be complete without some reference to these compound examples, which will therefore be briefly described in the present section of this work.

Early in this century some striking illustrations of the association of acid and more basic rocks within the same fissure were noticed by Jameson in the island of Arran. He described and figured instances at Tormore, on the west side of that island, where a group of pitchstones and "basalts" or andesites have been successively protruded into the same fissures in the (probably Permian) red sandstones of that district.[199]

[Footnote 199: _Mineralogy of the Scottish Isles_, 1800.]

In some instances the more basic rock has been first injected, and has subsequently been disrupted, by the more acid pitchstone. In other cases the order has been the reverse. The successive ruptures have taken place sometimes along the centre, sometimes at the margins, and sometimes irregularly along the breadth of the dykes. Professor Judd has recently studied these rocks, and has given descriptions of their chemical composition and microscopic characters. He regards them as having been successively injected into the fissures from the same subterranean reservoir, in which two magmas of very different chemical constitution were simultaneously present.[200]

[Footnote 200: _Quart. Jour. Geol. Soc._ vol. xlix. (1893), p. 536. Full details of the compound dykes of Tormore and Cir Mhor in Arran, and references to previous writers will be found in this paper. The probable age of the youngest eruptive rocks of this island will be discussed in Chapter xlvii. p. 418.]

Nowhere in the Tertiary volcanic regions of Britain do compound dykes appear to be so abundant as in the centre and southern part of the island of Skye. During the progress of the Geological Survey in that district, Mr. Clough and Mr. Harker have mapped a large number in the ground between the Sound of Sleat and the Red Hills. With regard to these dykes Mr. Harker observes that the several members are generally petrographically different, some being basic, others intermediate, and others acid. "There is usually," he remarks, "a symmetrical disposition, two similar and more basic dykes being divided by a more acid one; for example, two andesites separated by a pitchstone. Thus at the mouth of the little stream which runs from Torran into the bay east from Dùn Beag a dyke, apparently 18 feet wide, is found on examination to consist of a central dyke (specific gravity 2·86) flanked by two more basic dykes (specific gravity 3·02)."

In the great majority of examples hitherto observed in Skye the two lateral dykes consist of some basic rock (diabase or basalt), while the central and thickest band is of some acid material (granophyre or quartz-felsite). This triple arrangement occurs both in dykes and sills.

As an illustration of the association of the two kinds of rock in dykes I may cite an example which appears on the southern edge of the Market Stance of Broadford (Fig. 254). Here the characteristic triple arrangement is typically developed. A central light-coloured band, about eight to ten feet broad, consists of a spherulitic granophyre in which the spherulites are crowded together and project from the weathered surface like peas, though they do not here show the curious rod-like aggregation so marked in some other dykes. On either side of this acid centre a narrow basalt dyke intervenes as a wall next to the Torridon sandstone which here forms the country-rock. Such compound dykes have sometimes a total width of 100 feet or more.

In this instance, and generally throughout the district, there is nothing to indicate that the different bands of the dyke have any relation to each other as connected uprises of material from the same original magma which was either heterogeneous or was undergoing a process of differentiation beneath the terrestrial crust. On the contrary, the several parts of each dyke are as distinctly marked off from each other as they could have been had they been injected at widely separated intervals of volcanic activity.

Mr. Harker, in the course of his survey of this Skye ground, has observed that "where evidence is available, the central acid dyke is found to be newer than the basic ones. It has not split a single basic dyke, but has insinuated itself between the two members of a double dyke. This is more clearly seen when the acid magma has been forced into a triple or multiple basic dyke; the perfect symmetry of arrangement may in this case be lost. For instance, on the shore north-east of Corry, Broadford, a 13 feet dyke of granophyre occurs in a multiple dyke of basalt, but it has taken its line so as to leave only a one-foot dyke on one side, and a group with a total width of 12 feet on the other. Also it has not accurately kept its course, but has cut obliquely across one of the group of dykes alluded to. In some cases it is certain that the acid magma has to some extent dissolved a portion of the wall of a basic dyke with which it has come in contact. This may account for the magma finding its easiest path along, and especially between, pre-existing more basic dykes." This subject will be again referred to in Chapter xlviii., when the phenomena of compound sills are discussed.

Before closing this account of compound dykes, I may remark that no examples have yet been observed among the ordinary Tertiary dykes of Britain where, by a process of differentiation between the walls of a fissure, successive zones have been developed in the dyke, differing from each other in structure and composition, but becoming progressively and insensibly more acid towards the centre, such as have been described from the older rocks of Norway and Canada. Among the Tertiary gabbro bosses, indeed, there occur sheets or dykes which present a remarkably banded structure, to which full reference will be made in later pages. But I have never seen anything at all resembling such a structure among the dykes of andesite, dolerite, or basalt.

17. CONTACT-METAMORPHISM OF THE DYKES

A geologist might naturally expect that such abundant intrusions of igneous rock as those of the dykes should be accompanied with plentiful proofs of contact-metamorphism. But in actual fact, evidence of any serious amount of alteration is singularly scarce. A slight induration of the rocks on either side of a dyke is generally all the change that can be detected.

Some of the larger dykes, however, show more marked metamorphism, the nature of which appears in many cases to be chiefly determined by the chemical composition of the rock affected. Thus a considerable alteration has been superinduced on carbonaceous strata, particularly on seams of coal. In the Ayrshire coal-field the alteration of the coal extends sometimes 150 feet from the dyke, the extent of the change depending not merely on the mass of the igneous rock, but on the nature of the coal, and possibly on other causes. Close to a dyke, coal passes into a kind of soot or cinder, sometimes assumes the form of a finely columnar coke (Fig. 255), and occasionally has become vesicular after being fused.[201] Shales are converted into a hard flinty substance that breaks with a conchoidal fracture and rings under the hammer. Fireclay is baked into a porcelain-like material. Limestone is changed for a few inches into marble. As an illustration of this alteration, I may cite a dyke ten feet broad which cuts through the chalk in the Templepatrick Quarry, Antrim. For about six inches from the igneous rock the chalk has passed into a finely saccharoid condition, and its organisms are effaced. But beyond that distance the crystalline structure rapidly dies away, the micro-organisms begin to make their appearance, and within a space of one foot from the dyke the chalk assumes its ordinary character.

[Footnote 201: Explanation of Sheet 22, Geological Survey of Scotland, p. 26.]

Sandstones are indurated by dykes into a kind of quartzite, sometimes assume a columnar structure (the columns being directed away from the dyke-walls), and for several feet or yards have their yellow or red colours bleached out of them. The granite of Ben Cruachan where quarried on Loch Awe, as I am informed by Mr. J. S. Grant Wilson, is traversed by a basic dyke, and for a distance of about 20 feet is rendered darker in colour, becomes granular, and cannot be polished and made saleable.

Where many dykes have been crowded together, their collective effects in the alteration of the strata traversed by them have sometimes been strongly developed. One of the most remarkable illustrations of this influence is presented by the district of Strathaird, which was cited by Macculloch for the abundance of its dykes. In recently mapping this ground for the Geological Survey, Mr. Harker has observed in some places a score or more dykes in actual juxtaposition, while over considerable distances he found it difficult to detect any trace of the Jurassic strata, through which the igneous rocks have ascended. As might be expected under these circumstances, such portions of the strata as can be seen display an altogether exceptional amount of contact-metamorphism. Mr. Harker has noticed some limestones at Camasunary which have been changed into very remarkable lime-silicate rocks, with singular bunches of diopside crystals.

These, however, are the extremes of contact-metamorphism by the Tertiary basic dykes. A geologist visiting the Liassic shores of Strath in Skye will not fail to be surprised at the very slight degree of alteration in circumstances where he would have expected to find it strongly pronounced. The dark shales, though ribbed across with dykes, are sometimes hardly even hardened, and at the most are only indurated from an inch or two to about two feet. These baked bands project above the rest of the more easily denuded shales, and so adhere to the dykes as almost to seem part of them. Again the limestones, where traversed by dykes some distance apart, are not rendered in any appreciable degree more crystalline even up to the very margin of the intrusive rock. Where the igneous material has been thrust between the strata in sills, it has produced far more general and serious metamorphism than when it occurs in the form of single dykes. The famous rock of Portrush, already referred to as having been once gravely cited as an example of fossiliferous basalt, is a good illustration of the way in which Lias shale is porcellanized when the intruded igneous material has been thrust between the planes of bedding.

In the West of Scotland, where dykes are so abundantly developed, considerable differences can be observed between the amount of metamorphism superinduced by adjacent dykes which may be of the same thickness, and cut through the same kind of strata. Such variations have not probably arisen from differences in the temperature of the original molten rock. Perhaps they are rather to be assigned to the length of time occupied by the ascent of the lava in the fissure. If, for instance, the fissure opened to the surface and discharged lava there, the rocks of its walls would be exposed to a continuous stream of molten rock as long as the outflow lasted. They would thus have their temperature more highly raised, and maintained at such an elevation for a longer time than where the magma, at once arrested within the fissure, immediately proceeded to cool and consolidate there. It would be an interesting and important conclusion if we could, from the nature or amount of their contact-metamorphism, distinguish those dykes which for some time served as channels for the discharge of lava above ground.

Some dykes which have caught up fragments of older rocks in their ascent have exercised a considerable solvent action on these inclusions. Examples of this feature have already been cited from Skye, where they have been studied by Mr. Harker (pp. 129, 163).

In connection with the metamorphism superinduced by dykes, reference may again be made to the alteration which they themselves undergo where they have invaded a carbonaceous shale or coal. The igneous rock, as we have seen, loses its dark colour and obviously crystalline structure, and becomes a pale yellow or white, dull, earthy substance, or "white trap." The chemical changes involved in this alteration have been described by Sir J. Lowthian Bell.[202] Dr. Stecher has also discussed the alterations traceable by the aid of the microscope.[203] Though most of the instances of such transformation in Britain occur in the Carboniferous system, and have taken place in intrusive rocks of probably, for the most part, Carboniferous or Permian age, yet they are not unknown in the Tertiary volcanic series. Some of the "white trap" of the Coal-measures may indeed belong to the Tertiary period, but the coals and carbonaceous shales interstratified in the Tertiary basalt-plateaux have reacted on both the superficial lavas and the sills, and have given rise to the same kind of alteration as in the Carboniferous system, as will be shown in a later Chapter.

[Footnote 202: _Proc. Roy. Soc._ xxiii. (1875), p. 543.]

[Footnote 203: Tschermak's _Mineralogische Mittheilungen_, ix. (1887), p. 145, and _Proc. Roy. Soc. Edin._ 1888.]

Some marked examples of this alteration of intrusive igneous material are to be observed among the basalt dykes which cut the Lower Lias Shales of Skye. These shales, where black and carbonaceous, as in the island of Pabba, have exercised an unmistakable influence on the abundant dykes which intersect them. The chilled selvage of each dyke has assumed the dull earthy pale-grey or yellowish aspect, which extends for a few inches from the wall into the interior, where it rapidly passes into the ordinary black crystalline basalt. These features will be readily understood from the accompanying diagram (Fig. 256). Where the dykes give off narrow veins a few inches broad, these consist entirely of the "white trap." The shales are often traversed with strong joints parallel to the walls of the dykes, and the transverse joints of the dykes are sometimes prolonged into the bands of indurated shale.

18. RELATION OF DYKES TO THE GEOLOGICAL STRUCTURE OF THE DISTRICTS WHICH THEY TRAVERSE.

In no respect do the Tertiary dykes of Britain stand more distinguished from all the other rocks of the country than in their extraordinary independence of geological structure. The successive groups of Palæozoic and Mesozoic strata have been so tilted as to follow each other in approximately parallel bands, which run obliquely across the island from south-west to north-east. The most important lines of fault take the same general line. The contemporaneously included igneous rocks follow, of course, the trend of the stratified deposits among which they lie, and even the intrusive sills group themselves along the general strike of the whole country. But the Tertiary dykes have their own independent direction, to which they adhere amid the extremest diversities of geological arrangement.

In the first place, the dykes intersect nearly the whole range of the geological formations of the British Islands. In the Outer Hebrides and north-west Highlands, they rise through the most ancient (Lewisian) gneisses, through the red pre-Cambrian (Torridon) sandstones, and through the oldest members of the Cambrian system. In the southern Highlands, they pursue their course across the gnarled and twisted schists of the younger crystalline (Dalradian) series. In the South of Scotland and North of England, they traverse the various subdivisions of the Lower and Upper Silurian rocks. In the basins of the Tay, Forth, and Clyde they cross the plains and ridges of the Old Red Sandstone, with its deep pile of intercalated volcanic material. In Central Scotland, and the northern English counties, they occur abundantly in the Carboniferous system, and have destroyed the seams of coal. In Cumberland and Durham, they traverse the Permian and Trias groups. In Yorkshire, and along the West of Scotland, they are found running through Jurassic strata. In Antrim, they intersect the Chalk. Both in the North of Ireland, and all through the chain of the Inner Hebrides, they abound in the great sheets and bosses of Tertiary volcanic rocks. These are the youngest formations through which they rise. But it is deserving of note, that they intersect every great group of these Tertiary volcanic products, so that they include in their number the latest known manifestations of eruptive action in the geological history of Britain.[204]

[Footnote 204: They have not been found cutting the pitchstone-lava of the Scuir of Eigg.]

In the second place, in ranging across groups of rock belonging to such widely diverse periods, the dykes must necessarily often pass abruptly from one kind of material and geological structure to another. But, as a rule, they do so without any sensible deviation from their usual trend, or any alteration of their average width. Here and there, indeed, we may observe a dyke to follow a more wavy or more rapidly sinuous or zig-zag course in one group of rocks than in another. Yet, so far as I have myself been able to observe, such sinuosities may occur in almost any kind of material, and are not satisfactorily explicable by any difference of texture or arrangement in the rocks at the surface. No dyke traverses a greater variety of sedimentary formations than that of Cleveland. In the eastern part of its course, it rises through all the Mesozoic groups up to the Cornbrash. Further west it cuts across each of the different subdivisions of the Carboniferous system; and, of course, it must traverse all the older formations which underlie these. But the occasional rapid changes noticeable in its width and direction do not seem to be referable to any corresponding structure in the surrounding rocks. The Cheviot dyke crosses from the Carboniferous area of Northumberland into the Upper Silurian rocks and Lower Old Red Sandstone volcanic tract of the Cheviot Hills. It then strikes across the Upper Old Red Sandstone of Roxburghshire, and still maintaining the same persistent trend, sweeps westward into the intensely plicated Silurian rocks of the Southern Uplands. Its occasional deviations have no obvious reference to any visible change of structure in the adjacent formations. Again, some of the great dykes at the head of Clydesdale furnish striking illustrations of entire indifference to the nature of the rock through which they run. Quitting the Silurian uplands, they keep their line across Old Red Sandstone and Carboniferous rocks, and through large masses of eruptive material.

In the third place, not only are the dykes not deflected by great diversities in the lithological character of the rocks which they traverse, they even cross without deviation some of the most important geological features in the general framework of the country. Some of the Scottish examples are singularly impressive in this respect. Those which strike north-westward from the uplands of Clydesdale cross without deflection the great boundary-fault which, by a throw of several thousand feet, brings the Lower Old Red Sandstone against Silurian rocks. They traverse some large faults in the valley of the Douglas coal-field, pass completely across the axis of the Haughshaw Hills, where the Upper Silurian rocks are once more brought up to the surface, and also the long felsite ridge of Priesthill. The dykes in the centre of the kingdom maintain their line across some of the large masses of igneous rock that protrude through the Carboniferous system. Further north, the dykes of Perthshire cut across the great sheets of volcanic material that form the Ochil Hills, as well as through the piles of sandstone and conglomerate of the Lower Old Red Sandstone, and then go right across the boundary-fault of the Highlands, to pursue their way in the same independent manner through grit, quartzite, or mica-schist, and across glen and lake, moor and mountain.

No one can contemplate these repeated examples of an entire want of connection between the dykes and the nature and arrangement of the rocks which they traverse without being convinced that the lines of rent up which the material of the dykes rose were not, as a rule, old fractures in the earth's crust, but were fresh fissures, opened across the course of the older dislocations and strike of the country by the same series of subterranean operations to which the uprise of the molten material of the dykes was also due.

In the fourth place, the dykes for the most part are not coincident with visible lines of fault. After the examination of hundreds of dykes in all parts of the country, and with all the help which bare hillsides and well-exposed coast-sections can afford, the number of instances which have been met with where dykes have availed themselves of lines of fault is surprisingly small. Some of these cases will be immediately cited. To whatever cause we may ascribe the rupture of the solid crust of the earth, which admitted the rise of molten rock to form the dykes, there can be no doubt that it was not generally attended with that displacement of level on one or both sides of the dislocation, which we associate with the idea of a fault. Nowhere can this important part of dyke-structure be more clearly illustrated than along the Cleveland dyke, where the igneous rock rises through almost horizontal Jurassic strata and gently inclined Coal-measures (Figs. 241, 242, 243, 244). Besides the localities already cited, mining operations both for coal and for the Liassic ironstone have proved over a wide area that the dyke has not risen along a line of fault. Again, in Skye, Raasay, Eigg, and other parts of the west coast, where Jurassic strata and the horizontal basalts of the plateaux are plentifully cut through by dykes, the same beds may generally be seen at the same level on either side of them.

In the fifth place, while complete indifference to geological structure is the general rule among the dykes, instances do occur in which the molten material has found its way upward along old lines of rupture. Most of such instances are to be found in districts where previously existing faults happened to run in the same general direction as that followed by the dykes. These lines of fracture might naturally be re-opened by any great earth-movements acting in their direction, and would afford ready channels for the ascent of the lava, as we have seen to have not infrequently happened in the case of dyke-fissures, which are shown by compound dykes to have sometimes been re-opened several times in succession even after having been filled up with basalt. Yet it is curious that, even when their trend would have suited the line of the dykes, faults have not been more largely made use of for the purpose of relief. Some of the best examples of the coincidence of dykes with pre-existing faults in the same direction are to be found in the Stirlingshire coal-field. The dyke that runs from Torphichen for 23 miles to Cadder occupies a line of fault which at Slamannan has a down-throw of more than 70 fathoms. The next dyke further south has also risen along an east and west fault.

But other examples may be observed where pre-existing fissures have served to deflect dykes from their usual line of trend. Thus the Cleveland dyke, after crossing several faults in the Coal-measures, at last encounters one near Cockfield Fell, which lies obliquely across its path. Instead of crossing this fault it bends sharply round a few points south of west, and after keeping along the southern flank of the fault for about a mile, sinks out of reach. Some of the Scottish examples are more remarkable. One of the best of them occurs in the Sanquhar coal-field, where a dyke runs for two miles and a half along the large fault that here brings down the Coal-measures against the Lower Silurian rocks. At the north-western end of the basin, this fault makes an abrupt bend of 60° to W.S.W., and the dyke turns round with it, keeping this altered course for a mile and a half, when it strikes away from the fault, crosses a narrow belt of Lower Silurian rocks, and finds its way into the parallel boundary fault which defines the north-western margin of the Southern Uplands.

Some of the Perthshire dykes, where they reach the great boundary-fault of the Highlands, present specially interesting features. There can be no doubt that this dislocation is one of the most important in the general framework of the British Isles, though no definite estimate has yet been formed of how much rock has been actually displaced by it. The fact that in one place the beds of Old Red Sandstone are thrown on end for some two miles back from it, shows that it must be a very powerful fracture. Here, therefore, if anywhere, either an entire cessation of the dykes, or at least a complete deflection of their course might be anticipated. It would require, we might suppose, a singularly potent dislocation to open a way for the ascent of the lava through such crushed and compressed rocks, and still more to prolong the general line of a fracture across the old fault. Two great dykes, about half a mile apart, run in a direction a little south of west across the plain of Strathearn. Passing to the south of the village of Crieff, they hold on their way until they reach the highly-inclined beds of sandstone and conglomerate which here lean against the Highland fault in Glen Artney. They then turn round towards south-west, and run up the glen along the strike of the beds, keeping approximately parallel to the fault for about three miles, when they both strike across the fault, and pursue a W.S.W. line through the contorted crystalline rocks of the Highlands. About two miles further south, another dyke continues its normal course across the belt of upturned Old Red Sandstone; but when it reaches the fault it bends round and follows the line of dislocation, sometimes coinciding with, sometimes crossing or running parallel with that line, at a short distance (see Fig. 247).

Some remarkable examples have been mapped by Mr. Clough in Eastern Argyleshire, where broad bands of basalt or other allied rock run in a N. and S. direction, and are formed by the confluence of N.W and S.E. or N.N.W. and S.S.E. dykes, where these are drawn into a line of fault (Fig. 257). These broad bands, he has found to be not usually traceable for more than a mile or so, for the dykes of which they are made up will not be diverted from their regular paths for more than a certain distance, so that one by one the dykes leave the compound band to pursue their normal course. He has observed that the occasional great thickness of these compound bands depends partly on the size and partly on the number of separate dykes that are diverted into the line of transverse fissure; for, where the fissure crosses an area with fewer north-west dykes, the band becomes thinner or ceases altogether.

In some rare cases, the dykes have been shifted by more recent faults. I shall have occasion to show that faults of more than 1000 feet have taken place since the Tertiary basalt-plateaux were formed. There is therefore no reason why here and there a fault with a low hade should not have shifted the outcrop of a dyke. But the fact remains, that, as a general rule, the dykes run independently of faults even where they approach close to them. Mr. Clough has observed some interesting cases in South-eastern Argyleshire, where the apparent shifting of a dyke by faults proves to be deceptive, and where the dyke has for short distances merely availed itself of old lines of fracture. One of the most remarkable of these is presented by the large dyke which runs westward from Dunoon. No fewer than three times, in the course of four miles between Lochs Striven and Riddon, does this dyke make sharp changes of trend nearly at right angles to its usual direction, where it encounters north and south faults (Fig. 257). It would be natural to conclude that these changes are actual dislocations due to the faults. But the careful observer just cited has been able to trace the dyke in a very attenuated and uncrushed form along some of these cross faults, and thus to prove that the faults are of older date, but that they have modified the line of the long east and west fissure up which the material of the dyke ascended.

19. DATA FOR ESTIMATING THE GEOLOGICAL AGE OF THE DYKES

I have already assigned reasons for regarding the system of north-west and south-east or east and west dykes as belonging to the Tertiary volcanic period in the geographical history of the British Islands. But I have no evidence that they were restricted to any part of that period. On the contrary, there is every reason to consider the uprise of the earliest and latest dykes to have been separated by a protracted interval. That they do not all belong to one epoch has been already indicated, and may now be more specially proved.

The intersection of one dyke by another furnishes an obvious criterion of relative age. Macculloch drew attention to this test, and stated that it had enabled him to make out two distinct sets of dykes in Skye and Rum. But he confessed that it failed to afford any information as to the length of the interval of time between them.[205] It is not always so easy as might be thought to make sure which of two intersecting dykes is the older. As was explained in Chapter vi. (vol. i. p. 81), we have to look for the finer-grained marginal strip at the edge of a dyke, which, where traceable across another dyke, marks at once their relative age. The cross joints of the two dykes also run in different directions. Reference may again be made to the illustration given in Fig. 253 where three distinct groups of dykes intersect each other as they traverse the Lias limestones of Skye. The chilled edges and the different arrangement of joints mark these dykes out from each other, while the order in which they cross each other furnishes a clue to their relative age. If from such sections, repeated in different parts of a district, certain persistent petrographical characters can be ascertained to distinguish each particular system of dykes, a guide may thereby be obtained for the chronological grouping of the intrusions even where evidence of actual intersection is not visible. In the case just cited from Skye, the later north and south dykes are characterized by their lines of vesicular cavities and by the large porphyritic felspars which they contain.

[Footnote 205: _Trans. Geol. Soc._ iii. p. 75.]

It is obvious, however, that although sections of this kind suffice to prove the dykes to belong to distinct periods of intrusion, no longer interval need have elapsed between their successive production than was required for the solidification and assumption of a joint-structure by an older dyke before a younger broke through it. They may both belong to one brief period of volcanic activity. But when we pass to a series of dykes traversing a considerable district of country, and find that those which run in one direction are invariably cut by those which run in another, the inference can hardly be resisted that they do not belong to the same period of eruption, but mark successive epochs of volcanic energy. An excellent example of this kind of evidence is furnished by Mr. Clough from Eastern Argyleshire. The east and west dykes in that district are undoubtedly older than those which run in a N.N.W. direction (Fig. 257).[206] The latter are by far the most abundant, and are on the whole much narrower, less persistent, and finer in grain. On the opposite coast of the Clyde, a similar double set of dykes may be traced through Renfrewshire, those in an east and west direction being comparatively few, while the younger N.N.W. series is well developed. The great sheets or "sills" connected with one of the Stirlingshire dykes, already described, appear to me to furnish similar evidence in the younger dykes which run through them. And this evidence is peculiarly valuable, for it shows a succession even among adjacent dykes which all run in the same general direction.

[Footnote 206: As already stated, Mr. Clough and also Mr. Gunn are inclined to separate these older east and west dykes from the Tertiary series and to regard them as probably of late Palæozoic age.]

But in all these cases it is obvious that we have little indication of the length of time that intervened between the successive injections of the dykes. In Skye, however, more definite evidence presents itself that the interval must have been in some cases a protracted one. As far back as the year 1857,[207] I showed that the basic dykes of Strath in Skye are of two ages; that one set was erupted before the appearance of the "syenite" (granophyre) of that district, and was cut off by the latter rock; and that the other arose after the "syenite" which it intersected. Recent re-examination has enabled me to confirm and extend this observation. The younger series which traverses the granophyre is much less numerous than the older series in the same districts. In Chapter xlvi., where the relations of the granophyres to other members of the volcanic series will be discussed, further details will be given from that region of Skye to demonstrate that there is a pre-granophyre and a post-granophyre series of basic dykes. As a good illustration of the younger series I may refer to the way in which these rocks make their appearance in the island group of St. Kilda, where both the gabbros and granophyres of the Tertiary volcanic series are characteristically developed. Numerous dykes traverse both these rocks. Those in the gabbro are more abundant than those in the granophyre--a circumstance which is exactly paralleled among the basic and acid bosses of Skye. It is not improbable that in these remote islands a similar difference in age and in petrographical character may be made out between two series of dykes, one older and the other younger than the granophyre. There is ample proof, at all events, of a post-granophyre series.

[Footnote 207: _Quart. Jour. Geol. Soc._ vol. xiv. p. 16.]

The pale colour of the precipices in which the St. Kilda granophyre plunges into the sea gives special prominence to the dark ribbon-like streaks which mark the course of basalt-dykes through that rock. Moreover the greater liability of the material of the dykes to decay causes them to weather into long lines of notch or recess. Four or five such dykes follow each other in nearly parallel bands, which slant upward from the sea-level on the eastern face of the hill Conacher to a height of several hundred feet.[208] (Fig. 258, see also Fig. 367.)

[Footnote 208: This relation of the later dykes to the granophyre was observed here by Macculloch (_Western Isles_, vol. ii. p. 55).]

The acid eruptions of the Inner Hebrides are marked by so varied a series of rocks, and so complex a geological structure, that they may, with some confidence, be regarded as having occupied a considerable interval of geological time. Yet we find that this prolonged episode in the volcanic history was both preceded and followed by the extravasation of basic dykes.

Reference has already been made to recent observations by Mr. Harker, who, in mapping the Strath district of Skye for the Geological Survey, has not only confirmed the generalization as to the existence of a series of dykes earlier, and another later, than the great granophyre protrusions of the Inner Hebrides, but has made some progress towards the detection of a means of distinguishing the two series even where no direct test of their relative age may be available. He thinks that the general habit and petrographical characters of the dykes may on further investigation be found to afford a sufficiently reliable basis for discrimination. He finds that where the relative ages of the dykes with reference to the granophyre can be fixed, the earlier or pre-granophyre series is without exception basic. It consists of fine-textured basalts or diabases, without any conspicuous porphyritic crystals. Its dykes are less regular and persistent in their bearing than those of the later series; have frequently a considerable hade, even as much as 45°, and often show chilled edges with tachylitic selvages. In Skye many of these earlier dykes may be connected with the gabbro. They appear to be more basic and to have a higher specific gravity than those of the later series which most resemble them.

The later or post-granophyre dykes include several types, the relative ages of which are not yet definitely fixed. They run in straight parallel lines, and thus seldom intersect each other. They are generally vertical or highly inclined, and are much more frequently characterized by amygdaloiclal structure than the earlier series. Mr. Harker distinguishes the following varieties among them: (_a_) Quartz-felsites and other acid rocks; these are not very common. (_b_) Pitchstones and various spherulitic and variolitic rocks: the actual pitchstones observed are comparatively few in number, but it is certain that some of spherulitic varieties are devitrified pitchstones. (_c_) Basic rocks, not conspicuously porphyritic and less decidedly basic than the dykes of the pre-granophyre series; most of the later groups come into this or the next group, (_d_) Porphyritic basic dykes not infrequently carrying inclusions of gabbro, granophyre or other rocks. The porphyritic felspars seem to be in great part of foreign derivation, and the same is certainly true of the augite which occasionally accompanies them and of the quartz that appears in some examples.[209]

[Footnote 209: Annual Report of the Director-General of the Geological Survey in Report of Science and Art Department for 1895.]

In the Carlingford district of the North-east of Ireland, similar evidence has been obtained that one series of dykes preceded and another followed the protrusion of the granites and granophyre which are in all probability geologically coeval with the acid bosses of the Inner Hebrides. The distinction was observed and mapped by Mr. Traill for the Geological Survey. Professor Sollas in recently confirming these observations has not noticed any striking difference between the pre-granite and post-granite dykes, the whole appearing to consist of the same coarsely porphyritic material.[210]

[Footnote 210: See Sheets 59, 60, and 71 of the Geological Survey Map of Ireland; Professor Sollas, _Trans. Roy. Irish Acad._ vol. xxx. (1894), p. 477; and Annual Report of the Director-General of the Geological Survey for 1895.]

While the eruption of the granophyre bosses furnishes proof that the dykes are not all of the same age, other evidence may be gathered to show how much older some of the dykes are than the youngest lava-streams in the volcanic history of Tertiary time in Britain. The Scuir of Eigg, to which fuller reference will be made in Chapter xxxviii., is formed of a mass of pitchstone, which has filled up an ancient valley eroded out of the terraced basalts of the plateaux. At both ends of the ridge, these basalts are seen to be traversed by dykes that are abruptly cut off by the shingle of the old river-bed which the pitchstone has occupied (Figs. 279, 282). It is thus evident that, though these dykes are younger than the plateau-basalts, they are much older than the excavation of the valley out of these basalts, and still older than the eruption of pitchstone. The latter rock probably belongs to the close of the period of lava-eruptions. The enormous denudation of the basalt-plateaux after the injection of the dykes and before the outflow of the pitchstone affords a convincing proof of the vastness of the interval between the eruption of the two kinds of rock.[211]

[Footnote 211: _Quart. Jour. Geol. Soc._ xiv. p. 1.]

It is thus demonstrable that the dykes which in Britain form part of the great Tertiary volcanic series, were not all produced at one epoch, but belong to at least two (and probably to many more) episodes in one long volcanic history. As they rise through every member of that series of rocks (save the pitchstones), some of them must be among the latest records of the prolonged volcanic activity. But, on the other hand, some probably go back to the very beginning of the Tertiary volcanic period.

20. ORIGIN AND HISTORY OF THE DYKES

Reference has already been made to the doubt expressed by Macculloch whether the dykes in Skye had been filled in from above or from below. That the dykes of the country as a whole were supplied from above, was the view entertained and enforced by Boué. He introduces the subject with the following remarks:--"Scotland is renowned for the number of its basaltic veins, which gave Hutton his ideas regarding the injection of lava from below; but, as the greatest genius is not infallible, and as volcanic countries present us with examples of such veins arising evidently from accidental fissures that were filled up by currents of lava which moved over them, and as the Scottish instances are of the same kind, we regard it as infinitely probable that all these veins have been formed in the same way notwithstanding the enormous denudation which this supposition involves; and that only rarely do cases occur where they have been filled laterally or in some other irregular manner."[212] I need not say that this view, which, except among Wernerians, had never many supporters, has long ago been abandoned and forgotten. There is no further question that the molten material came from below.

[Footnote 212: _Essai Géologique sur l'Écosse_, p. 272.]

1. In discussing the history of the dykes, we are first confronted with the problem of the formation of the fissures up which the molten material rose. From what has been said above regarding the usual want of relation between dykes and the nature and arrangements of the rocks which they traverse, it is, I think, manifest that the fissures could not have been caused by any superficial action, such as that which produces cracks of the ground during earthquake-shocks. The fact that they traverse rocks of the most extreme diversities of elasticity, structure, and resistance, and yet maintain the same persistent trend through them all, shows that they originated far below the limits to which the known rocks of the surface descend. We have seen that in the case of the Cleveland dyke, the fissure can be proved to be at least some three miles deep. But the seat of the origin of the rents no doubt lay much deeper down within the earth's crust.

It is also evident that the cause which gave rise to these abundant fissures must have been quite distinct from the movements that produced the prevalent strike and the main faults of this country. From early geological time, as is well known, the movements of the earth's crust beneath the area of Britain, have been directed in such a manner as to give the different stratified formations a general north-east and south-west strike, and to dislocate them by great faults with the same average trend. But the fissures of the Tertiary dykes run obliquely and even at a right angle across this prevalent older series of lines and are distinct from any other architectonic feature in the geology of the country. They did not arise therefore by a mere renewal of some previous order of disturbances, but were brought about by a new set of movements to which it is difficult to find any parallel in the earlier records of the region.[213]

[Footnote 213: The only other known example of such a dyke-structure in Britain is that of the Pre-Cambrian series of dykes in the Lewisian gneiss of Sutherland, described in Chapter viii.]

We have further to remember that the fissures were not produced merely by one great disturbance. The evidence of the dykes proves beyond question that some of them are earlier than others, and hence that the cause to which the fissures owed their origin came into operation repeatedly during the protracted Tertiary volcanic period. One of the most instructive lessons in this respect is furnished by the huge eruptive masses of gabbro and granitoid rocks in Skye. These materials have been erupted through the plateau-basalts. The granitoid bosses are the younger protrusions, for they send veins into the gabbros; but their appearance was later than that of some of the dykes and older than that of others. Nevertheless, the youngest dykes generally maintain the usual north-westerly trend across the thickest masses of the granophyre. Thus we perceive that, even after the extrusion of thousands of feet of such solid crystalline igneous rocks, covering areas of many square miles, the fissuring of the ground was renewed, and rents were opened through these new piles of material. From the evidence of the dykes also we learn that some fissures were repeatedly re-opened and admitted a new ascent of molten magma between their walls. The general direction of the fissures remained from first to last tolerably uniform. Here and there indeed, where one set of dykes traverses another, as in Skye and the basin of the Clyde, we meet with proofs of a deviation from the normal trend. But it is remarkable that dykes which pierce the latest eruptive bosses of the Inner Hebrides rose in fissures that were opened in the normal north-westerly line through these great protrusions of basic and acid rock.

Such a gigantic system of parallel fissures points to great horizontal tension of the terrestrial crust over the area in which they are developed. Hopkins, many years ago, discussed from the mathematical side the cause of the production of such fissures.[214] He assumed the existence of some elevatory force acting under considerable areas of the earth's crust at any assignable depth, either with uniform intensity at every point or with a somewhat greater intensity at particular points. He did not assign to this force any definite origin, but supposed it "to act upon the lower surface of the uplifted mass through the medium of some fluid, which may be conceived to be an elastic vapour, or, in other cases, a mass of matter in a state of fusion from heat."[215] He showed that such an upheaving force would produce in the affected territory a system of parallel longitudinal fissures, which, when not far distant from each other, could only have been formed simultaneously, and not successively; that each fissure would begin not at the surface but at some depth below it, and would be propagated with great velocity; that there would be more fissures at greater than at lesser depths, many of them never reaching the surface; that they would be of approximately uniform width, the mean width tending to increase downwards; that continued elevation might increase these fissures, but that new fissures in the same direction would not arise in the separated blocks which would now be more or less independent of each other; that subsequent subsidences would give rise to transverse fissures, and by allowing the separated blocks to settle down would cause irregularities in the width of the great parallel fissures. He considered also the problem presented by those cases where the ruptures of the terrestrial crust have been filled with igneous matter, and now appear as dykes. "The results above obtained," he says, "will manifestly hold equally, whether we suppose the uplifted mass acted upon immediately through the medium of an elastic vapour or by matter in a state of fusion in immediate contact with its lower surface. In the latter case, however, this fused matter will necessarily ascend into the fissures, and if maintained there till it cools and solidifies, will present such phenomena as we now recognize in dykes and veins of trap."

[Footnote 214: _Cambridge Phil. Trans._ vi. (1835), p. 1.]

[Footnote 215: _Ibid._ p. 10.]

The existence of a vast lake or reservoir of molten rock under the fissure-region of Britain is demonstrated by the dykes. But, if we inquire further what terrestrial operation led to the uprise of so vast a body of lava towards the surface in older Tertiary time, we find that as yet no satisfactory answer can be given.

2. In some districts the dykes can be connected with the gabbros which occur as intrusive sills and irregular bosses in the basalt-plateaux and among older rocks. The gabbros, however, are traversed by still later dykes, which must then be independent of any visible mass of these rocks. The connection of dykes with the gabbros is what we might naturally expect to find, if the more coarsely crystalline rock represents portions of the basic magma which consolidated at some depth below the surface. If we could penetrate deep enough, it is not improbable that the dykes might be found in large measure to shade downward into vast bodies of gabbro. Such a relation has been observed in the Yellowstone district, where Mr. Iddings has noticed that the centre toward which the dykes of the Old Crandale volcano converge is a large mass of granular gabbro, passing into diorite, the dykes becoming rapidly coarser in grain as they approach the gabbro-core.[216]

[Footnote 216: _Journ. Geol._ i. (1893), p. 608.]

3. The rise of molten rock in thousands of fissures over so wide a region is to my mind by far the most wonderful feature in the history of volcanic action in Britain. The great plateaux of basalt, and the mountainous bosses of rock by which they have been disrupted, are undoubtedly the most obvious memorials of Tertiary volcanism. But, after all, they are merely fragments restricted to limited districts. The dykes, however, reveal to us the extraordinary fact that, at a period so recent as older Tertiary time, there lay underneath the area of Britain a reservoir or series of reservoirs of lava, the united extent of which must have exceeded 40,000 square miles.

That the material of the dykes rose in general directly from below, and was not, except locally, injected laterally along the open fissures, may be inferred, although proof of such lateral injection on a small scale may here and there be detected. The narrowness of the rents, and their enormous relative length, make it physically impossible that molten rock could have moved along them for more than short distances. The usual homogeneous character of the dyke-rocks, the remarkable scarcity of any broken-up consolidated fragments of them immersed in a matrix of different grain, the general uniformity of composition and structure from one end of a long dyke to another, the spherical form of the amygdales, the usual paucity of fragments from the fissure walls--all point to a quiet welling of the lava upward. Over the whole of the region traversed by the dykes, from the hills of Yorkshire and Lancashire to the remotest Hebrides, molten rock must have lain at a depth, which, in one case, we know to have exceeded three miles, and which was probably everywhere considerably greater than that limit.

Forced upwards, partly perhaps by pressure due to terrestrial contraction and partly by the enormous expansive force of the gases and vapours absorbed within it, the lava rose in thousands of fissures that had been opened for it in the solid overlying crust. That in most cases its ascent terminated short of the surface of the ground may reasonably be inferred. At least, we know, that many dykes do not reach the present surface, and that those which do have shared in the enormous denudation of the surrounding country. That even in the same dyke the lava rose hundreds of feet higher at one place than at another is abundantly proved. When, however, we consider the vast number of dykes that now come to the light of day, and reflect that the visible portions of some of them differ more than 3000 feet from each other in altitude, we can hardly escape the conviction that it would be incredible that nowhere should the lava have flowed out at the surface. Subsequent denudation has undoubtedly removed a great thickness of rock from what was the surface of the ground during older Tertiary time, and hundreds of dykes are now exposed that doubtless originally lay deeply buried beneath the overlying part of the earth's crust through which they failed to rise. But some relics, at least, of the outflow of lava might be expected to have survived. I believe that such relics remain to us in the great basalt-plateaux of Antrim and the Inner Hebrides. These deep piles of almost horizontal sheets of basalt, emanating from no great central volcanoes, but with evidence of many local vents, appear to me to have proceeded in large measure from dykes which, communicating with the surface of the ground, allowed the molten material to flow out in successive streams with occasional accompaniments of fragmentary ejections.[217] The structure of the basalt-plateaux, and their mode of origin, will form the subject of the next division of this volume.

[Footnote 217: It is interesting to note that in the great paper on Physical Geology already cited, Hopkins considered the question of the outflow of lava from the fissures which he discussed. "If the quantity of fluid matter forced into these fissures," he says, "be more than they can contain, it will, of course, be ejected over the surface; and if this ejection take place from a considerable number of fissures, and over a tolerably even surface, it is easy to conceive the formation of a bed of the ejected matter of moderate and tolerably uniform thickness, and of any extent" (_op. cit._ p. 71).]

We can hardly suppose that the lava flowed out only in the western region of the existing plateaux. Probably it was most frequently emitted and accumulated to the greatest depth in that area. But over the centre of Scotland and North of England there may well have been many places where dykes actually communicated with the outer air, and allowed their molten material to stream over the surrounding country, either from open fissures or from vents that rose along these. The disappearance of such outflows need cause no surprise, when we consider the extent of the denudation which many dykes demonstrate. I have elsewhere shown that all over Scotland there is abundant proof that hundreds and even thousands of feet of rock have been removed from parts of the surface of the land since the time of the uprise of the dykes.[218] The evidence of this denudation is singularly striking in such districts as that of Loch Lomond, where the difference of level between the outcrop of the dykes on the crest of the ridges and in the bottom of the valleys exceeds 3000 feet. It is quite obvious, for example, that had the deep hollow of Loch Lomond lain, as it now does, in the pathway of these dykes, the molten rock, instead of ascending to the summits of the hills, would have burst out on the floor of the valley. We are, therefore, forced to admit that a deep glen and lake-basin have been in great measure hollowed out since the time of the dykes. If a depth of many hundreds of feet of hard crystalline schists could have been removed in the interval, there need be no difficulty in understanding that by the same process of waste, many sheets of solid basalt may have been gradually stripped off the face of Central Scotland and Northern England.

[Footnote 218: _Scenery of Scotland_, 2nd edit. (1887), p. 149. But see the remarks already made (p. 150) on the curious coincidence sometimes observable between the upper limit of a dyke and the overlying inequalities of surface.]

The association of fissures and dykes with the accumulation of thick and extensive volcanic plateaux, over so wide a region of North-western Europe as from Antrim to the North of Iceland, finds its parallel in different parts of the world. One of the closest analogies presents itself among the Ghauts of the Bombay Presidency, where vast basaltic sheets, probably of Cretaceous age, display topographical and structural features closely similar to those of the Tertiary volcanic plateaux of the British Isles. The dykes connected with these Indian basaltic outflows correspond almost exactly in their general character and stratigraphical relations to those of this country. They occur in great numbers, rising through every rock in the district up to the crests of the Ghauts, 4000 feet above the sea. They vary from 1 or 2 to 10, 20, 40, and even occasionally 100 or 150 feet in width, and are often many miles in length. They observe a general parallelism in one average direction, and show no perceptible difference in character even when traced up to elevations of 3000 and 4000 feet.[219]

[Footnote 219: Mr. G. T. Clark, _Quart. Journ. Geol. Soc._ xxv. (1869) p. 163. For remarks on the connection of dykes with superficial lavas, see _postea_, p. 268.]

Thousands of square miles in the Western States and Territories of the American Union have been similarly flooded with basic lavas. Denudation has not yet advanced far enough to lay bare much of the platform on which these lavas rest. But the dykes that traverse the rocks outside of the lava-deserts afford an example of the structure which will ultimately be revealed when the wide and continuous basalt-plains shall have been trenched by innumerable valleys and reduced to fragmentary plateaux with lofty escarpments (p. 267).

It is to the modern eruptions of Iceland, however, that we turn for the completest illustration of the phenomena connected with dykes and fissures. An account of these eruptions will therefore be given in Chapter xl. as an explanation of the history of the Tertiary basalt-plateaux of Britain.