As the distribution of the rock-formations has an important bearing on the question of the glaciation, it will be desirable to give a brief outline of the nature and respective limits of the various formations, so far as these have been already determined. The stratified rocks belong to two periods:—(a) the Old Red Sandstone; (b) the great series of metamorphic crystalline rocks on which the representatives of the Old Red Sandstone rest unconformably. To what precise part of the crystalline rocks of the Highlands the metamorphic series of the Mainland and the north isles belongs, we do not at present presume to say^*.

There are also associated with the metamorphic series some intrusive igneous rocks, and certain masses which may be viewed as products of extreme metamorphism. These may probably be relegated to the time when the metamorphism of the ancient stratified rocks took place. At least some of the igneous rocks now referred to must be older than the basement breccias of the Old Red Sandstone, inasmuch as the latter in certain localities are composed of angular fragments of the former.

But, farther, there are abundant proofs of volcanic activity during the Old-Red-Sandstone period, as is evident from the great development of contemporaneous and intrusive igneous rocks on the Mainland. Similar phenomena are met with in the isles of Papa Stour, Bressay, Noss, the Holm of Melby, and Meikle Rooe; but the magnificent sections on the western shores of Northmavine justify the conclusion that the proofs of volcanic activity on the Mainland surpass in grandeur and extent those of the other Shetland islands.

The Metamorphic Series.

On the Mainland these may be grouped in two divisions, which are clearly marked off from each other by distinct lithological characters.

(a)	Dark blue, green, and grey schists and clay-slates, with bands of quartzite and limestones.page 781
(b)	Coarse-grained micaceous and hornblendic gneiss, with associated limestones, bands of quartzitc, talcose and micaceous schists.

These subdivisions are peculiarly serviceable to the glacialist, as they help him to determine the different movements of the ice during successive phases of the ice age. The representatives of the former series extend from Fitful Head northwards by the Bonxie and Cliff Hills to Laxfirth Voe; while the members of the gneissose series lie to the north-west of the area just described. They occur in the districts of Tingwall, Weesdale, Nesting, Lunnasting, Delting, and along the eastern seaboard of Northmavine. The strike of these motamorphic rocks is generally N. 10°–20° E.; and though opposing dips are frequently met with, indicating repetitions of the strata, they usually dip to the north of west at high angles. Hence we have a gradually ascending series from the schists and clay-slates of the Cliff Hills to the coarse micaceous gneiss west of the vale of Tingwall, and the massive limestones of Whiteness and Weesdale.

To the persistent trend of the metamorphic rocks must be ascribed the remarkable ridge-shaped contour of the ground in the centre of the Mainland. The coincidence between the trend of the strata and that of the parallel ridges seems to indicate a direct relationship between the two, the denuding agents being guided in their operations by the relative hardness and softness of the materials exposed to their influence. Hence it follows that we have a scries of intervening hollows running parallel with the ridges, which usually terminate seawards in long narrow voes or sea-lochs. The erosion of these hollows has doubtless, in some instances, been duo to the partial removal of the bands of limestone by the chemical action of carbonated waters, inasmuch as the outcrop of the limestones coincides with the course of a longitudinal hollow.

The coarse-grained gneiss of Whalsey and the Outskerries, with the associated limestones, is merely the prolongation of the Mainland series; and the same remark is applicable to the gneiss occuring in Yell.

The structure of Unst and Fetlar is somewhat different, inasmuch as these isles contain well-marked zones of serpentine and gabbro, the distribution of which has an important bearing on the question of the dispersal of the stones in the Boulder-clay. In the island of Unst, the Vallafield ridge which flanks the western seaboard, whose highest elevation is about 697 feet, is mainly occupied by coarsegrained gneiss, dipping to the south of cast at comparatively high angles. On the eastern slopes of the ridge the gneiss is succeeded by grey mica-schists and green chloritic schists, and these are overlain in turn by black graphitic schists. These dark schists seem to form a reliable horizon with reference to the masses of serpentine and gabbro, as they usually crop out along the margin of the areas oeeupied by these masses and generally dip underneath them. Though these schistose rocks form but a narrow band from Belmont Bay northwards to Baliasta, they occupy a much broader area to the north of the latter point, constituting, in fact, the group of hills page 781 round Saxavord. They reappear again in the south-east corner of the island, where they cover a strip of ground about a mile in breadth between Skuda Sound and the ruins of Muness Castle.

The masses of serpentine and gabbro in Unst lie in a trough formed by these schists. They may be said to form two parallel zones crossing the island from south-west to north-east, the serpentine lying to the west of the gabbro. The serpentine area is the larger of the two, though somewhat irregular in outline; at the northern limit between Baliasta Kirk and Swena Ness, the mass is nearly two miles in breadth, but as it is traced southwards it diminishes to half a mile in breadth. Another patch of gabbro is to be met with on the promontory east of the ruin of Muness Castle.

It seemed to us that the serpentine has resulted from the metamorphism of the gabbro. Hero and there in the gabbro area, as, for instance, on the west side of Uya Sound, lenticular patches of serpentine occur, as if the transmutation had partly begun and had been interrupted. The gradual transition from the one rock to the other is well seen in the promontory on the south side of Balta Sound. Professor Heddle, who advocates this view, states that the gradual passage can be seen in hand specimens on Swena Ness.

The structure of the northern portion of Fetlar is comparatively simple. The central hollow coincides with a low anticlinal axis of black graphitic schists and chloritic schists similar to those in Unst, and apparently occupying the same horizon with reference to the gabbro and serpentine. These rocks throw off on both sides of the arch beds of gabbro and serpentine, forming the elevated ground round the Yord Hill on the east and the hills near Urie on the west. At Urie the serpentine which overlies the gabbro is immediately succeeded to the west by coarse-grained gneiss, the perfectly conformable junction between the two being distinctly visible on the shore west of the promontory of Urie. The broad mass of serpentine which stretches from the Vord Hill eastwards to Grating Bay is thrown into a synclinal trough, which is nowhere deep enough to bring in the overlying gneiss to the west of Urie. To the east of Gruting Bay occur the micaceous and chloritic schists as well as the graphitic schists, which contain in minor folds small patches of serpentine.

Intrusive Igneous Rocks in the Metamorphic Series.

Under this heading we shall only indicate those intrusive rocks and those products of extreme metamorphism which are probably older than the Old-Red-Sandstone period. We have already referred to the areas of gabbro and serpentine in Unst and Fetlar; but in addition to these there are certain masses on the Mainland deserving special notice.

Of these by far the largest is the mass of diorite occurring in the districts of Delting and Northmavine on the Mainland. It is upwards of ten miles in length, and in places it exceeds two miles in breadth; but it ought to be borne in mind that the whole of the area now described is not occupied by the diorite, nor is the boun- page 783 clary line so uniform as we have represented. A minute examination of this tract convinced us that the groundwork of the area, so to speak, is formed of metamorphic schists, which are traversed in all directions by largo and small veins of this rock. Both the diorite and the schists are intersected by innumerable veins of quartz-felsite which were injected at a more recent date, the whole series of rocks forming a complicated network.

Again, in Dunrossness, between Quendale Bay and Loeh Spiggie, there is a mass of intrusive rock termed by Hibbert epidotic syenite, which is traceable northwards through the islands of Oxna, Hildasay, the Sandistura rocks, the Channes, and part of Papa west of Scalloway, to the Mainland in Bixetter Voe and onwards to Aith Voe. This rock varies considerably in character throughout its course; in some places it is a quartz-felsite, while in the neighbourhood of Bixetter and Aith Voes it is a true porphyritic granite, with large crystals of orthoclase. There can be no doubt that it is an intrusive mass, because it crosses obliquely the strike of the metamorphic rocks on Fitful Head and the Wart of Skewsburgh; and it is equally clear that the eruption was prior to the Old-Bed-Sandstone period, as the basement breccias of that formation rest unconformably on this rock, and are largely made up of angular fragments of the subjacent mass.

A similar mass of porphyritic granite occurs in Unst on the bluff headland of Lambaness and on the rocky promontory north of Skaw Bay, which likewise bears important testimony regarding the direction of the ice-movement. In addition to these masses there are minor veins of granite, gabbro, and serpentine, some of which are indicated on the map. There is one fact bearing on the age of the veins of serpentine on the Mainland which is worthy of note; and that is, the occurrence of fragments of this rock in the basement breccias of the Old Red Sandstone in Dunrossness. This circumstance plainly indicates that the formation of the serpentine veins in that neighbourhood preceded the formation of the breccias.

The Old Red Sandstone.

A glance at the map will show the various areas occupied by the members of this formation in Shetland. Beginning with the irregular areas on the east side of the Mainland, the succession may be most readily grasped by means of the following section (fig. 1, p. 784)^*.

Owing to a scries of faults which form the boundary-line between the metamorphic rocks and the Old Red Sandstone, over a great part of the districts of Lerwick, Quarff, Conningsburgh, and Dunrossness, it so happens that different zones in this vertical section are brought into conjunction with the schistose rocks. The true base of the series, however, is exposed in the neighbourhood of page 784

Fig. 1.—Vertical Section of Old Red Sandstone strata on east side of Shetland.

East Quarff, on the hills to the north of the bay and to the south towards Fladabister; while still another locality is met with near Loch Spiggie in Dunrossness. In each of these localities the breccia varies in character according to the nature of the underlying rock.

In the bay west of Brenista Ness, the overlying series of the Brenista Flags is thrown against the breccias and underlying schists by a fault which is traceable inland in a N.N.W. direction. This series consists of well-bedded rod flags, which persistently dip to the east till Gulberwick Bay is reached. The fault just referred to, when traced inland, always throws the flags down against the basement-breccia, and hence the actual superposition is not satisfactorily seen in the neighbourhood of Brenista. Between East Quarff and Fladabister, however, the one group may be seen resting conformably on the other; and, in addition to this, we find that the basal breccia, which forms vertical cliffs on the coast-line about 200 feet high, thins out inland till there is only about 3 feet of breccia between the underlying schists and the overlying Brenista Flags. In some instances the breccia disappears altogether, and the Brenista Flags rest directly on the schists, a fact which points to the gradual submergence of the area.

Returning to the shore-section north of East Quarff, there is a gradually ascending series from the Brenista Flags to certain coarse conglomerates seen in a small stream at the head of the bay of Gulberwick, which are totally different from the basal breccias already described. Not only are the enclosed pebbles well rounded, but to a large extent the stones are composed of different materials. These beds are traceable up the slope of the Gulberwick hollow, to the road between Lerwick and Scalloway, where they form crags on the hill face, and where they may be seen in small quarries by the roadside. They may be followed also across the hills northwards to Rovey Head, about two miles north of Lerwick, where they are brought into conjunction with the schists by a fault which is well seen on the shore. From Rovey Head southwards to the ridge overlooking the head of Fitch Dale, this fault forms the boundary- page 785 line between the conglomerates and the metamorphic rocks. At this point it dies out, and the boundary-line southwards towards Fladabister is formed by the basement breccia already described.

At Rovey Head the conglomerates are thrown into synclinal and anticlinal folds; but eventually they dip to the south-east, and are succeeded immediately by grey sandstones, with blue and grey flags passing upwards into the series of the Lerwick Sandstones. The dominating members of this series are coarse grits, frequently conglomeratic, with partings of fine red shales.

In Bressay, however, these arenaceous and conglomeratic strata are overlain by a more flaggy series, which is more or less persistent till Noss Head is reached. We were struck with the resemblance which some of these grey flaggy bands bear to the calcareous flags of Orkney and Caithness containing the fish-remains; but a careful search failed to bring any to light. Numerous plant-remains have long ago been detected, not only in these strata but also in some of the other groups on the eastern shore of the Mainland.

In the peninsular tract of country which lies to the west of the Weesdale district there is a great series of rocks which, with the exception of a small tract at Melby, have been hitherto considered as forming part of the metamorphic series. The small strip of Old-Red-Sandstone rocks at Melby, measuring about a mile and a half in length, has been referred to by previous observers. They are separated by a fault from the red quartzites and shales of Sandness Hill; and on approaching the fault it is observable that the beds are much shattered on account of this dislocation. They consist of reddish sandstones with dark blue flags and shales, dipping to the east of south and south-east, from Sandness to near Melby.

The great series of rocks which occupies almost the whole of the remainder of this peninsular tract, and which by their fossil contents we have proved to be of Old-Red-Sandstone age, has a somewhat different lithological character. Over a great part of this area the beds consist of grey and blue altered sandstones, with green and pale shales. The altered sandstones are usually traversed in every direction by joints, which are coated with peroxide of iron; and in places the beds have a marked schistose character. Sometimes the sandstones are converted into genuine quartzites, and the shales inter-bedded with them are distinctly cleaved. The strata lie in a trough the axis of which runs approximately from Fontabrough Voe eastwards by the village of Walls to the head of Bixetter Voe. On the north side of the syncline we have a gradually ascending series exposed on the coast-lino from the cliffs of Sandness Hill southwards towards Fontabrough Voe, the average strike of the beds being E. 20° N.

We discovered the plant-remains on the hills north of Walls, and subsequently in quarries by the roadside east of the village, and on the hills between Gruting and Bixetter Voes. They have been examined by Mr. C. W. Peach, who has kindly furnished the notes on the specimens embodied in the Appendix. He is of opinion that the plants are identical with those found in the Old-Red formation of page 786 Caithness and Orkney; and the strata in which they are imbedded, altered though they be, must be relegated to that period.

This conclusion is still further strengthened by the occurrence in these rocks of interbedded porphyrites and tuffs in a highly altered form, which we detected on the headlands between Aith Ness and Clouster, and on the western shore south of Dales Voe, resembling in many respects the contemporaneous volcanic rocks to be described presently. Further, we are inclined to believe that the series of altered thick-bedded sandstones and shales which occupy the greater portion of this peninsular tract are on the same horizon with the Lerwick Sandstones on the eastern side of the Mainland.

It is not improbable that the alteration of the strata in this wide area may be duo to the existence of a mass of granite underneath these rocks. We shall have occasion to refer to the mass of granite in the heart of these beds in Sandsting, and to similar intrusive masses of Old-Red-Sandstone age to the north. The extent of ground occupied by these acidic rocks indicates the great volcanic activity which prevailed during that period; and though these are now isolated at the surface, it is highly probable that they may be connected underneath.

These altered fossiliferous strata are brought into conjunction with the gneissose rocks to the cast and north by two great faults which we have traced on the ground, the one running north and south, and the other approximately cast and west. Usually the altered strata are terribly shattered and baked close to the lines of dislocation, and are likewise injected with numerous veins of very fine-grained felsite.

Contemporaneous Igneous Rocks.

In the western district of Northmavine, between Stennis and Ockren Head, at the mouth of Roeness Voe, there is an important development of ancient lavas and ashes, associated at certain localities with ashy sandstones and red flags belonging to this period. The tract of ground occupied by these rocks measures about six miles in length, and varies in breadth from one to two miles. The structure of this narrow tract is comparatively simple, as the strata form a flat syncline, in the centre of which lies coarse ash, and underneath a series of slaggy porphyrites with occasional beds of red ashy sandstones and flags.

On the south bank of Roeness Voe, rather more than a mile from Ockren Head, in a steep grassy goo, the slaggy porphyrites are brought into conjunction with the pink quartz-felsite by a fault. In Braewick Bay, west of Hillswick, the interbedded and intrusive igneous rocks are not found in such close proximity, the junction being concealed by the sandy beach; but there can be little doubt that the same fault runs out to sea in this bay.

Crossing the coarse volcanic breccia, which forms the centre of the syncline, to Ockren Head, at the mouth of Roeness Voe, the successive lava-flows are admirably shown, piled on each other in regular succession. This headland, as well as the adjacent stack, exhibit at page 787 least four different lava-flows, capped by coarse ash. The lavas thicken and thin out rapidly, and likewise exhibit. The usual slaggy structure at the top and bottom of the flow. Some of these beds are also highly involved, and show clearly the way in which the partially solidified crust has been caught up and rolled over and over in the advancing current of still molten lava.

We have already alluded to the porphyrites and tuffs which occur in the altered rocks north of Walls. We also detected a bed of lava in the Holm of Melby, and a thin bed of tuff associated with the grey flags on the east side of Bressay, opposite the north end of the island of Noss. The contemporaneous volcanic rocks found in Papa Stour have been previously described by Professor Geikie.

Intrusive Igneous Rocks.

In the north and western portions of the Mainland there is a splendid development of highly siliceous intrusive rocks, which occupy the most elevated ground in the island. They extend from a point on the north end of the Mainland opposite the island of Uya, southwards to Roeness Voe, culminating in the dome-shaped mass of Roeness Hill. Thence they cross the peninsular tract to the Heads of Grocken, west of Hillswick, reappearing in the slender columns of the Drongs. The western portion of Meikle Rooe is formed of the same material, and likewise the north-eastern headlands of Vemeutry, while the small area of quartz-porphyry at Melby must also be included in the same great intrusive series. In addition to the areas now referred to, there are other lenticular masses varying in size down to veins a few feet across, occurring at intervals from Mavis Grind, northwards to Roeness Voe and Ollaborry. These traverse the Northmavine diorite and metamorphic rocks alike, increasing in number and extent as they approach the Roeness mass.

These rocks vary considerably in character; but they all agree in possessing a largo proportion of silica, while the felspar is usually orthoclase. As a rule, they are coarsely crystalline, the two prevalent ingredients, quartz and orthoclase felspar, being distinctly crystallized, which causes the rock to assume a marked granitoid texture. There can be little doubt that these coarsely crystalline rocks must have originally consolidated under great pressure, though the materials under which they lay buried have been wholly removed by denudation. Further, the marked columnar structure which meets the eye along both banks of Roeness Voe, and from the Heads of Grocken to Braewick Bay, as well as along the western shores of Meikle Rooe, suggest the idea of a great intrusive sheet, forced in like a wedge between the metamorphic series and the members of the Old Bed Sandstone long since worn away. A similar intrusive sheet occurs in Papa Stour, as described by Professor Geikie, where the same columnar structure is apparent, and where a fragment of the once superincumbent strata is still to be seen at the Horn of Papa. Fortunately the intrusive nature of this latter shoet is placed beyond doubt, inasmuch as the pink porphyry is seen page 788

Fig. 2. Section across Northmavine from Ockren Head to Skea Ness.

cutting across the underlying sandstones from a lower to a higher horizon.

As the result of careful mapping of the boundaries of the Northmavine mass, we are of opinion that the Roe-ness-hill plateau is a great intrusive sheet which forced its way upwards and laterally between the metamorphic strata on the one hand, and the members of the Old Red Sandstone on the other, at the time when the Mainland lay buried under the sedimentary deposits which accumulated during that period. It is not at all improbable that this immense mass may have been connected with the surface by pipes which traversed the superincumbent strata, and may have discharged volcanic materials at the surface.

Its relations to the metamorphic series are admirably defined. Along the eastern seaboard of Northmavine it forms a mural escarpment about 200 feet high, part of which is known by the name of the Biurgs. Innumerable veins of quartz-felsite branch off from the main mass and intersect the metamorphic series. Further, it sometimes happens that portions of the adjacent rocks are enclosed in the quartz-felsite, as, for instance, near Colifirth Voe, where a fragment of serpentine is caught up in the mass. Again, on the north bank of Roeness Voe, the sheet spreads over the edges of the diorite and metamorphic rocks without producing any deflection of the strike of the metamorphic series, as shown in the accompanying section (fig. 2).

To the west of Hillswick, on the picturesque Heads of Grocken, the highly siliceous quartz-felsite is thrown against the metamorphic series by a fault, which is well seen on the cliffs. This fault passes out to sea between the little islets of Waterhouse Holm and the Drongs, and reappears in Meikle Rooe, separating the quartz-felsite from the diorite of that island. In all likelihood this dislocation is the northward prolongation of the great north-and-south fault already described, which brings the metamorphosed Old-Red-Sundstone rocks into conjunction with the gneissose rocks of Weesdale.

Our conclusion regarding the Roeness plateau is strengthened by a consideration of the relations of the granite mass of Sandsting to the altered Old-Red strata of that district. This siliceous intrusive rock has several lithological varieties; but away from the margin of the area, as, for instance, on the hills above Gruting and round Skelda Voe, it is an ordinary granite consisting of pink orthoclase, quartz, and mica.

On the shores of Gruting Voe, at the foot of Culswick Hill, the junction of the granite with the Old-Red quartzites and shales is well seen, from an examination of which it is evident that the granite is intruded along the lines of bedding of these strata. The junction-line has nearly the same inclination as that of the quartzites, which dip to the north at an angle of about 20°. The two rocks are not strictly conformable, however; for the granite here and there cuts across the bedding, indicating in an unmistakable manner the intrusive nature of the igneous rock. The junction is a sharp and well-defined line, as small hand-specimens can easily be got, 2 inches across, including the granite and the quartzite, the two being firmly welded together. Near the junction of the two rocks the quartzites are pervaded by numerous dykes of pink felstone proceeding from the main mass.

The mass of granite in the north of Delting and on the western shores of Sulem Voe was probably erupted during the same period of volcanic activity; but the evidence is not so convincing as that referring to the areas already described.

But in addition to the grand series of intrusive rocks we have just indicated, there is evidence to prove that even these quartz-felsites and granites were invaded by a still later series of dykes, of a basic character. Hibbert detected the existence of these dykes on Roeness Hill; and during our traverses in the district of Northmavine, Delting, and Meikle Rooe we came across many similar masses, varying in breadth from 2 feet to several yards. Along the cliffs of Roeness Voe, and in the island of Meikle Rooe, these dykes are strikingly exhibited, forming great wall-like masses, running generally in a north-and-south direction. Sometimes they project above the acidic rocks, while, again, they weather more rapidly, forming great clefts in the face of the cliff. They are fine-grained, and consist of a dark-green diabase porphyrite. They traverse the metamorphic rocks, as well as the porphyrites and tuffs, west of Braewick; and there can be little doubt, therefore, that they form the last indications of volcanic activity during the Old-Red-Sandstone period in Shetland.

Close by the entrance to the Noss Sound, on the Bressay shore, we detected a scries of necks arranged in a linear manner, which seem to have come to the surface along a line of fissure. Similar necks occur on Noss, on the opposite side of the Sound. It is highly probable that these volcanic orifices served merely as vents for the discharge of steam, with occasional showers of triturated materials derived mainly from the sides of the vents. The adjacent bed of tuff, associated with the grey flags, as well as the nature of page 790 the agglomerate which now fills these necks, seems to support this view.