I. Introduction.

In a former paper which we communicated to the Society on "The Glaciation of the Shetland Isles," we endeavoured to show how the evidence supplied by the striated surfaces, the roches moutonnées, and the dispersal of the stones in the Boulder-clay points to the conclusion that Shetland had been glaciated by Scandinavian ice. It was further argued that during the climax of glacial cold the Scandinavian and Scotch ice-sheets coalesced on the floor of the North Sea, and that the great outlet for the combined ice-sheets was towards the north-west by the Pentland Firth and the Orkney Islands.

In the course of the autumn of 1879 we visited nearly all the Orkney Islands for the purpose of continuing our researches with reference to the extension of the ice in the North Sea in the Glacial period. In the paper now presented to the Society we purpose to give a summary of the results of our observations. At the outset we may state that they furnish a remarkable confirmation of the conclusions already arrived at regarding the westerly and northwesterly movement of the ice. Moreover, the presence of stones in the Boulder-clay, which must have been derived from the mainland of Scotland, and the discovery of abundant remains of marine shells in the same deposit, though in a fragmentary state, are of the utmost importance in guiding us to a satisfactory solution of the question.

No description of the glacial phenomena of Orkney has hitherto been published^*. Some references were made by Professor Geikie to

^* Since this paper was written, our friend Mr. Amund Helland has sent us a copy of his paper "Ueber die Vergletscherung der Färoër Inseln," which appeared in the 'Zeitschrift der deutschen geologischen Gesellschaft,' 1879. We are glad to see that Mr. Helland has arrived at the same conclusions as ourselves regarding the north-westerly movement of the ice in Orkney, from independent observations made in the course of last year.

page 649 the existence of roches moulonnées with striated surfaces, Boulder clay, and valley moraines in the islands in an article which appeared in 'Nature'^*. This article was written in reply to a letter by Samuel Laing, Esq., M.P., in which he asserted that there is no evidence that Orkney had participated in the general glaciation of Britain^†. So far from there being any lack of evidence regarding the glaciation of these islands, we hope to show that they contain abundant proofs of having undergone severe glacial conditions. Our observations, however, completely confirm Sir. Laing's statement that there are no raised beaches in the islands indicating changes of the relative levels of sea and land since glacial times.

II. Geological Structure.

The geological structure of the islands is comparatively simple. From Stromness on the Mainland northwards to Inganess there is an axis of ancient crystalline rocks on which the representatives of the Old Red Sandstone rest uncouformably. These crystalline rocks consist of a fine-grained granite and a grey micaceous flaggy gneiss, which occupy a strip of ground about four miles in length and about a mile in breadth. They arc prolonged southwards in the island of Graemsa. With this exception the whole of the Orkney-Islands are occupied by Old Red Sandstone strata. In the island of Hoy representatives of both the upper and lower divisions of this formation are met with, and here they are separated by a marked unconformity; but in all the other islands the beds belong to the lower division.

Throughout the islands there is a remarkable uniformity in the character of the strata belonging to the lower division. They consist mainly of hard blue and grey calcareous flagstones, which are so typically developed in Caithness. Fortunately, however, the highest beds of the Orcadian flagstone scries are totally different in character from those just described, being composed of coarse siliceous red and yellow sandstones and marls. The sandstones are full of false-bedding, and frequently conglomeratic, containing pebbles of granite, quartzite, gneiss, and other crystalline rocks.

The distribution of this arenaceous series has an important bearing on the question of the ice movement. On referring to the map accompanying this paper (Pl. XXVI.), it will bo seen that it forms a well-marked zone, running nearly north and south through the centre of the group. The relations which these siliceous sandstones bear to the flagstones are best seen in Eda, where they cover the greater part of the island, and where they form smooth flowing hills upwards of 300 feet in height. The sandstones lie in a syncline, the axis of which runs north and south, and on both sides of the island they rest conformably on the flagstones. In the islands which lie to the west and north-west of Eda, viz. Fara, Westra, Papa Westra, Egilsha, and Rowsa, the strata consist wholly of blue and grey flagstones, which are inclined at gentle angles. Though there are page 650 many minor undulations, yet on the whole there is a gradually descending scries towards the western headlands of Rowsa and Westra.

In Stronsa and Sanda the arenaceous series and the underlying flagstones are repeated by a scries of faults, which are laid down on the map.

The south-east corner of Shapinshay is occupied by these sandstones, where they are associated with a dark green slaggy diabase, which forms part of an ancient lava-flow. They reappear on the south shore of Shapinshay Sound, and cross the Mainland in a narrow strip from Inganess Head to Scapa Bay. They are continued also along the north-west shore of Scapa Flow as far as Orphir Kirk, and they likewise extend along the eastern shore to Howquoy Head, near St. Mary's. These sandstones and marls are brought into conjunction with the flagstones of the Mainland by two great faults, which we have traced on the ground; but in Cava, Fara, Flota, South Ronaldshay, and Burra they graduate downwards into the flagstones, and are regularly interbedded with them. As the result of careful mapping of the coast-sections in the southern islands, we have come to the conclusion that Scapa Flow occupies the centre of a geological basin, towards which the strata are inclined on almost every side, and round whose shores the highest members of the Lower Old Red Sandstone in Orkney are to be found. We have elsewhere stated our reasons for believing that the Orcadian flagstones, with the conformable sandstones and marls, are the equivalents of the higher subdivisions of the Caithness series^*.

It ought to be clearly borne in mind that to the north-west of the great fault which extends from Houton Head eastwards by Scapa to the bay east of Work Head, the Old Red strata consist wholly of flagstones, save the conglomeratic beds, which repose unconformably on the crystalline axis, north of Stromness.

The physical features as well as the geological structure of Hoy are somewhat different from those which obtain in the other islands. Instead of a low undulating tableland, terminating seawards in a bluff cliff or sloping downwards to a sandy beach, which is the dominant type of Orcadian scenery, the island of Hoy forms a prominent tableland, trenched by a series of deep narrow valleys, which arc occasionally flanked by conical hills upwards of 1500 feet high. These narrow valleys must have been admirably adapted for nourishing a series of local glaciers towards the close of the Glacial period, as is evident from the long moraines now strewn over the hill-slopes.

The greater portion of the island is occupied by coarse false-bedded sandstones, which are but the counterpart of the Upper Old Red Sandstones at Dunnet Head, in Caithness. Near the base of this division occur some contemporaneous volcanic rocks, which are admirably exposed on the noble cliff in the north-west of the island

^* " The Old Red Sandstone of Orkney," by B. N. Peach and J. Horne Trans, of the Phys. Soc. Edinb. vol. v. 1880.

page 651 and at the base of the Old Stan of Hoy. The whole series rests unconformably on the flagstones; and in the south-west portion the upper division is brought into conjunction with the lower by a fault which extends from Melsetter to the coast-line opposite Risa island^*.

III. Glaciation.

The glacial phenomena of Orkney completely establish the double system of glaciation which we found to obtain in Shetland. There is satisfactory evidence for maintaining that during the primary glaciation the Orkney Islands must have been overridden by a mass of ice which moved from the North Sea to the Atlantic; but towards the close of the Glacial period, when the great mer de glace had retreated from the Orcadian coast-line, local glaciers must have lingered for a time in the valleys of Hoy and in some of the more elevated parts of the Mainland.

Though these islands do not comprise any districts that might be compared with North Mavine or the promontories of Lunnasting in Shetland, which are dotted all over with finely preserved roches moutonnées and rock-basins, nevertheless a careful search along the cliff-tops reveals numerous instances of glaciated surfaces and ice-markings. The latter, however, are not so abundant as we found to be the case in Shetland, which may be satisfactorily explained by the rapid disintegration of the flagstones when long exposed to atmospheric waste.

In the island of Westra the average direction of the striae in the eastern part of the island is W. 20°–30° N. Close by Noltland Castle, at the roadside, the trend is W. 20° N., on the north-west face of Cleat hill N.W., and immediately to the east of the same hill W. 30° N. At Rackwik, on the eastern shore, the ice-markings vary from W. to W. 20° N., while in Tuquoy Bay they point W. 10° S.

A careful examination of the striated surfaces on the hills west of Pierowall proves that the ice must have been slightly deflected as it impinged on the eastern slopes, the lower portion moving in the direction of the northern coast-line, while the higher strata streamed westwards over the hill-tops towards Nonp Head and Russitaing. On the north-eastern face of the hill south of Ourness several examples were noted pointing N. 30°–35° W., but in the gap between the hills the direction is W. 5° S.

Perhaps one of the most interesting features connected with the glaciation of Westra is the freshness of the ice-markings on Nonp Head (240 feet) and along the cliff-tops to the south. A few yards to the north and south of the highest point of this bold headland, finely preserved stria; were observed on grey flags, where the thin Boulder-clay had been recently removed by the action of the sea, trending W. to W. 3° N. Above Ramna Gio the direction page 652 varies from W. 10° N. to W. 10° S.; at Russitaing, W. 20° S.; near the Red Hare, W. 10° S.; near Inganess, W. 15° 8. to W. 18° N.; and again, in the bay south of Inganess, a well-marked instance points W. 12° N.

In some parts of the island of Eda the proofs of glaciation are marvellously fresh, more especially on the surfaces of the harder sandstones. From the finely glaciated surfaces and numerous roches moutnnnées in the centre of the island north of Lonton it is evident that the ice must have overtopped the hills in its northwestward march. On the east slopes of the Stennie hill the direction of the striæ is W. 20°–25° N., and not far to the south W. 40° N.

Along the eastern coast, between Calf Sound and Lonton Bay, the ice-markings point N. 20°–30° W., while between the Kirk of Skail and the Veness promontory the average direction is W. 35°–40° N. In one remarkable instance, on the shore about a mile to the south of the Kirk of Skail, striæ were observed on a highly inclined rock-face trending north and south, while on the cliff-top the direction is W. 35° X., the former being evidently due to local deflection. Along the western coast the general direction of the ice-movement is in perfect harmony with that just described. In the neighbourhood of Warness, which forms the south-west promontory of the island, the trend is W. 13° N., while to the west of the Wart of Eda, on the cliff-tops, it varies from W. 28° N. to W. 43° N.; and again, to the north of Seal Skerry, W. 40° N. One of the best examples to be met with in the island occurs in the bay east of Fara's Ness, where a small stream enters the sea. This burn has cut down through a deposit of shelly Boulder-clay to the polished pavement on which it rests; and along the stream-course the firm lines produced by the ice-chisel may bo seen to advantage on the glaciated surfaces of the sandstones. The direction of these instances is N. 27° W., but on the shore, close by the mouth of the stream, the trend is W. 38° N.

Notwithstanding the widespread covering of blown sand which envelops the greater portion of the island of Sanda, we succeeded in finding abundant traces of the ice-movement. In the Burness peninsula striated surfaces are numerous along the coast-lino, about a dozen instances occurring between Hermaness Bay and the Holms of Eyre, which, with one or two exceptions, point W. 10°–15°N. To the west of Loch Roo the direction is W. 25° N.; and not far from the Saville boulder, on the eastern shore of the peninsula, the trend is N.W.

On the shores of Kettletoft Bay the average direction is W. 10°X.; inland from this bay towards the Free Church it varies from W. 20°–40° N., while in Bacaskeal Bay it is N. 32° W. This north-westerly movement is equally borne out by the evidence obtained in the southern part of the island; for in the bay west of Hack Ness the ice-markings point N. 30° W., and on the western shore between Spur Ness and Stranquoy N. 8°–17° W.

The island of Stronsa likewise supplies conclusive evidence page 653 regarding the direction of the ice-movement; for in Odin Bay, where an important section of Boulder-clay occurs, which we shall describe presently, the striæ point W. 15°–35° N.; between Kirk-buster and Finga the trend is W. 10°–40° N., at Burgh Head W. 40° N., and north of Holland W. 40° N. On the western coastline, on the shores of Rousholm Bay, the direction varies from W. 41° N. to N. 40° W.; and on the shores of Linga Sound it is W. 40° N. It is of the utmost importance to note the perfect agreement in the trend of the ice-markings in different parts of this island, because it indicates a persistent movement in one determinate direction.

A careful examination of the striated surfaces on Shapinshay confirms in a remarkable manner the evidence regarding the ice-flow during the primary glaciation in the northern islands. Along the west coast, between Galtness and Stromberryness, the direction varies from W. to N.W., while on the shores of Veantro Bay, which indents the northern part of the island, the markings point N.W. and N. 35°–40° W. Further along the eastern coast-line, between Gioness and the school-house, the direction is W. 30° N., and the same trend is observable southwards towards the church; near Foot stria: were noted pointing N. 30° W., and close to Haco's Ness N. 20° W.

It is impossible, within the limits of this paper, to describe the various instances we met with in the Mainland, and we will therefore merely indicate the general trend in different parts of the island. On the glaciated surfaces of granite and gneiss north of Stromness numerous examples occur trending W. lo°–20° N. and W. 12° S. Immediately behind the town the direction varies from W. 8°–40° N., while on the moorland between Yesnabae and the Loch of Stennis, as well as at the Ring of Brogar, the same variation is observable from W. 12° N. to N.W.

On the hill-slopes overlooking Gorsness and the island of Gairsa the average direction of several examples is N. 25°–30° W., and along the coast-lino from Irland Bay to Houton Head the trend varies from W. 12°–42° N. One instance occurs in Irland Bay pointing W. 32° S., which probably belongs to the later glaciation.

In Kirkwall Bay, a short distance to the east of the pier, beautifully striated flagstones may be seen where Boulder-clay has been recently removed by the action of the sea, the striæ running N. 6° W. and N.N.W.; and so also on the surfaces of the flagstones in the Scapa Quarry the direction is N. 8° W. Along the shore from Scapa to Howquoy Head the average direction of several examples is N. 30°–35° W., and near St. Mary's the trend varies from N.W. to N. 25° W.

In the southern islands striæ are not so abundantly found, owing to the readiness with which the soft yellow sandstones and marls crumble away when long exposed to the denuding agencies. In South Ronaldshay several examples occur, the general direction of which is W. 20° N. These may be seen on the cliff-tops near Stow Head and Halcro Head by removing the coating of Boulder- page 654 clay. Even on the cliffs of the island of Hoy, overlooking the Atlantic, striated surfaces have been observed by Professor Geikie at a height of 600 or 700 feet above the sea-level.

The evidence now adduced regarding the ice-movement proves beyond all doubt that the islands have been glaciated in one determinate direction, independently of their physical features. A glance at the striæ map accompanying this paper (PI. XXVII.) shows the remarkable uniformity of the ice-flow in the different islands. Here and there, where local causes interfered with the general movement, slight deflections are met with; but, on the whole, the prevalent direction varies from W.N.W. to N.N.W. A careful examination of the numerous striated surfaces convinced us that the ice-sheet must have crossed the islands from the North Sea to the Atlantic. Indeed no one who reflects for a moment on the physical features of the islands could reasonably attribute the striations to a local radiation of the ice. If we except Hoy, these scattered islands contain no mass of elevated ground which is capable of giving rise to a local ice-sheet. So far from this being the case, we shall have occasion to refer to the absence of any indications of the existence of local glaciers in most of the islands towards the close of the Glacial period, a phenomenon which is doubtless due to this very cause. On the contrary, when we view the persistent northwesterly trend of the striations in connexion with the physical features, when we consider that the glaciated surfaces along the cliff-tops, as well as the roches moutonnées on the hill-slopes, prove that the islands must have been overflowed by the ice, we cannot resist the conclusion that the ice-movement during the primary glaciation originated beyond the limits of Orkney.

Fortunately the dispersal of the stones in the Boulder-clay amply confirms the foregoing conclusions regarding the north-westerly movement of the ice, while the presence of Scotch rocks in the same deposit enables us to demonstrate that the ice-sheet which crossed this group of islands must have radiated from the mainland of Scotland.

IV. Boulder-clay.

This deposit is not spread over the general surface of the low undulating tablelands in the form of a more or less continuous covering. It occurs mainly round the bays, whore it frequently attains a considerable depth, while the inland districts are covered with a thin clayey soil, duo to the decomposition of the underlying flagstones. We shall have occasion to describe one or two sections of Boulder-clay which may be traced continuously along the shore for half a mile, and which are quite undistinguishable from the Lower Boulder-clay of Scotland. Occasionally thin patches of this deposit are to be found on the cliff-tops, containing well-striated stones and foreign rocks, clearly indicating that the islands must have been overflowed by the ice.

In the island of Westra the Boulder-clay is sparingly distributed, but some excellent sections are to be met with round the bays in page 655 the southern districts. At Rackwik, near Stangar Head, it consists of a tough tenacious gritty clay, which is chiefly made up of red sandstone fragments, about 80 per cent, of the larger blocks being composed of sandstones which are foreign to the island. Some of these blocks, which arc finely smoothed and striated, measure six feet across. The deposit rests on the grey flagstones, and some small subangular fragments derived from the underlying rocks are likewise included; but the great majority of the stones consist of sandstones which we identified as belonging to the island of Eda. Crossing the peninsula to the shores of Tuquoy Bay, similar sections arc presented, resting on grey flagstones, the included blocks being composed of the underlying rocks, red sandstones, quartzites, and chalk-flints. It is important to note that the red sandstone blocks do not form such a large percentage at this locality, but that they gradually diminish in number as wo recede towards the north-west. On the slopes of Cleat hill fragments of granite, quartzite, diorite, and dolerite are associated with the flagstones in this deposit; while still further north, near the church, blocks of red and white honeycombed sandstone and small pink granite stones were observed in the Boulder-clay in addition to the local rocks.

Along the west coast hardly any Boulder-clay is to be met with; but some thin patches arc to be seen on the cliffs at Nonp Head, containing well-striated stones derived from the flagstones of the island. Occasional smooth blocks of Red Sandstone occur in the hollows amongst the debris of the underlying rocks, which arc, in all likelihood, the relics of the onco existing Boulder-clay.

Now it is evident, on a moment's consideration, that the gradual decrease in number of the red sandstones in the Boulder-clay, as we traverse the island from the south-eastern headlands towards the west coast, indicates that the ice-flow must have been towards the Atlantic; and when wo consider that these sandstones nowhere occur in situ in Westra, and that they could only have been derived from the adjacent islands of Eda and Sanda, we are forced to conclude that the ice-movement must have been altogether independent of the islands.

Along the east coast of Eda the Boulder-clay is not so abundant as in some of the more sheltered bays on the opposite side of the island, which is satisfactorily accounted for by supposing that the rocky slopes facing Eda Sound were exposed to the full sweep of the mer de glace. Here and there, however, patches do occur, as on the north shore of Lonton Bay, where the deposit contains smoothed and striated chalk-stones, along with blocks of red and white sandstones and grey flagstones. Its most noteworthy feature is the presence of worn fragments of marine shells, which are scattered irregularly through the stony clay. Similar sections occur in the bay of Calf Sound near the pier, and also along the west coast near the Wart of Eda, where shell-fragments were likewise observed.

Perhaps the finest section of this deposit in Eda occurs along the 'banks of a small stream which flows into the bay about a mile east of Fara's Ness on the west coast. The stream has cut down through page 656 the stony clay to the finely grooved pavement of sandstone, so that the glacialist can examine thoroughly the nature of the deposit. It consists of tough red clay, packed with smooth and striated stones scattered irregularly through the section. There is no trace of stratification in the deposit, as it retains the same tumultuous character throughout. The stones are beautifully striated along the major axis, and arc mainly composed of the underlying red and yellow sandstones, varying in size from an inch to several feet across. In addition to these wo noted smooth chalk-stones, chalk-flints, and subangular blocks of the grey flagstones. The most interesting feature, however, is the occurrence of small worn fragments of marine shells which are scattered indiscriminately through the deposit; they are smoothed and striated precisely in the same way as the stones in the Boulder-clay, as if they had been subjected to the same abrasion.

Shelly Boulder-clay was also observed on the west coast of Sanda, between Spur Ness and Stranquoy: and sections of the same deposit are to be found in Bacaskeal Bay. In the Burness peninsula, near the Holms of Eyre, the shore is bounded by low cliffs of purple shales and flags, with a coating of Boulder-clay, which is just sufficient to cover the surface of the rocks. It is chiefly composed of fragments of the underlying rocks, but likewise contains fragments of sandstone, granite blocks, and smoothed stones of gneiss and schist, all of which, except the sandstone, are foreign to the island.

In Stronsa several important sections were met with both on the east and west sides of the island. On the shores of Linga Sound, not far from the narrow isthmus of Aith, a soction of shelly Boulder-clay occurs resting on grey sandstone, the deposit being upwards of 25 foot thick, and comprising chalk, chalk-flints, and white quartz, in addition to the blocks derived from the flags and sandstones of the island. Further, on the north-east corner of Rousholm Bay a thin coating of this deposit rests on the flagstones, which are bent over to the north-west in the direction of the ice-flow.

One of the best exposures of Boulder-clay in Orkney occurs on the eastern shores of Odin Bay, in Stronsa, where it forms a continuous cliff for nearly half a mile. At intervals the section is obscured by a grassy covering, but every succeeding storm washes anew the face of the cliff, and exposes a fresh surface for examination. The deposit, which varies from 20 to 30 feet in depth, consists of a tough gritty clay of a reddish colour, full of well-smoothed and striated stones, which are mostly of small size. There are few largo boulders to be seen, the largest rarely exceeding a foot in diameter. There is not the slightest trace of stratification from one end of the section to the other, as the stones are disposed irregularly through the clayey matrix. By far the greater number of the included blocks have been derived from the flagstones and the sandstones which occur in the neighbourhood; but the following rocks are likewise represented:—granite, pink porphyritic felstone, gneiss, schist, quartzite, white quartz, dark limestone, with abundant plant-remains, which is probably of Calciferous-Sandstone age, oolitic page 657 limestone, oolitic calcareous breccia, fossil wood (probably oolitic), chalk, and chalk-flints, all of which are foreign to the island. When we come to collate the evidence regarding the primary glaciation, we shall discuss the probable localities from which these blocks were derived. At present it is sufficient to state that the evidence is clearly in favour of their having been carried from the mainland of Scotland.

Equally important is the presence of numerous fragments of marine shells throughout the deposit. Though we examined the section with the utmost care, we did not succeed in dislodging a complete shell; indeed so worn are the fragments that it was with the utmost difficulty that we obtained specimens sufficiently well preserved for determination. Nearly all the fragments are smoothed and striated, like the stones in the Boulder-clay; and there can be little doubt that these characteristics are due to the very same cause in both cases. Amongst the broken shells we detected fragments of Cyprina islandica, Mytilus, and Mya truncata; but a careful search, after severe storms, by some local collector would certainly increase this list considerably.

In the island of Shapinshay shelly Boulder-clay occurs at various localities on the east coast, as at Kirkton, where it contains finely striated chalk-stones. The best sections, however, occur along the western shore, and especially in the bay south of Galtness, where it forms a bluff cliff washed by high tides. This cliff furnishes valuable evidence regarding the ice-cam', inasmuch as we noted amongst the included stones blocks of the slaggy diabase which occurs in situ in the south-east corner of the island, along with striated fragments of the sandstones which are associated with the volcanic rocks. In this section smooth blocks of chalk and oolitic limestone, with numerous fragments of marine shells, were also observed.

If we traverse the Mainland from Scapa and Kirkwall westwards, by the Loch of Stennis, to the crystalline axis north of Stromress, similar conclusive evidence regarding the north-westerly movement of the ice is obtained from the Boulder-clay. On referring to the map of Orkney, it will bo seen that the narrow zone of red and yellow sandstones which crosses the Mainland from Inganess to Scapa extends south-westwards along the shore as far as Orphir Kirk. Now, in the shelly Boulder-clay in Kirkwall Bay, to the east of the pier, striated blocks of red sandstone are commingled with the flagstones in the clayey matrix. The latter are by far the most numerous, and arc likewise beautifully scratched along the major axis; but the sandstone blocks constitute a fair percentage of the included stones. From the lithological charactcr of these blocks, we had no hesitation in concluding that they had been derived from the sandstones to the east of Kirkwall.

Again, in the sections occurring on the coast between Houton Head and Irland Bay, the observer cannot fail to note the gradual increase in the number of the sandstone blocks in this deposit as he approaches Houton Head, a phenomenon which is quite intelligible when he remembers that the striations along the shore point W. page 658 12°–42° N., the latter being the prevalent direction. Indeed at Houton Head the ice-markings are nearly parallel with the coastline, so that the sandstone blocks could not possibly have come from Hoy. Blocks of the same rock are strewn on the hill-slopes above Gorsness, to the north-east of Maes Howe. It is a significant fact that not a single block of the granite or gneiss which occurs in situ to the north of Stromness and in the island of Graemsa is to be found in the Boulder-clay between Irland Bay and Houton Head, or anywhere to the east of the axis of crystalline rocks; but as soon as the western limit of these rocks is crossed, numerous blocks of granite and gneiss are strewn on the slopes and along the cliff-tops between Brak Ness and Inganess. Had the ice-movement been from the north-west, the phenomena would have been precisely the opposite of those we have described.

In the southern islands this deposit is not abundant; but in South Ronaldshay, on the shores of Water Sound, cast of St. Margaret's Hope, we observed patches of it containing blocks of sandstone, flags, and chalk, with comminuted shells. In this instance the shells, when being dislodged, crumble readily to a white powder.

V. Moraines.

One conspicuous feature connected with the Glacial phenomena of Orkney is the remarkable absence of any traces of local glaciers except in Hoy and the Mainland. When we consider the abundance of moraine heaps in all the more important islands of the Shetland group, this difference seems all the more striking; but when we remember the marked contrast between the physical features of the two groups of islands, the difficulty at once disappears. As we have already indicated, the only mass of elevated ground which would be capable of nourishing a series of local glaciers, after the great mer de glace had melted back from the Orcadian coast-line, occurs in Hoy. Hence we find that in the valleys which drain the group of conical hills in the north of that island moraines occur in abundance and also of great size. Professor Geikie has already described several examples which also came under our notice^*. In the valley to the page 659 east of Hoy hill a moraine mound, nearly half a mile long and from fifty to sixty feet high, runs across the mouth of the glen. It would seem that the later glacier which filled the valley did not succeed in scooping out the moraine profonde belonging to the primary glaciation, as the moraine matter rests on stiff sandy Boulder-clay. Further, in the hollow below Coulax hill several concentric heaps were observed which extend across the valley, indicating pauses in the retreat of the glacier.

In the Mainland also the moory ground between Finstown and Maes Howe is dotted all over with conical moraine heaps, evidently deposited by the glaciers which moved off the northern slopes of the Orphir hills. On the east side of the range of hills that runs north from Finstown several parallel moraine ridges may bo observed not far from Ellibister. Again, in the peninsular tract to the south-east of Kirkwall, a splendid series occurs in a valley situated about three miles north of Graemshall. At the point where the highroad from Roseness joins that from St. Mary's to Kirkwall, the concentric arrangement of the moraine heaps is admirably displayed.

VI. Erratics.

Boulders do not occur very plentifully in Orkney; but we felt convinced, from an examination of those we met with, that they must have been mainly distributed during the primary glaciation. In Westra blocks of granite and quartzite arc found on the slopes of Cleat hill; and rounded stones and boulders of red sandstone from Eda occur in the southern district as well as along the western shores.

In the north of Sanda, at Saville, a remarkable boulder of gneiss is met with, which has been described by previous observers. It measures × 6 × 2½ feet above ground, but its base is buried underneath the surface. Professor Heddle, who has made a minute examination of this boulder, states that it docs not appear to be a British rock. He gives the following description of it in a recent number of the 'Mineralogical Journal'^*:—" It consists in greatest amount of white finely striated oligoclase, the crystals of which are penetrated by fine filaments of actinolite, glassy quartz in much smaller amount, dark green finely foliated lustrous hornblende in well-marked crystals, very little of a pale-green mica, a minute amount of a pale-brown mineral, which may, but does not appear to be sphene, and a speck or two apparently of thorite. The mass also contains a single crystal of pale-green apatite four or five inches in length by over an inch in width, and this apatite contains imbedded cryptolite."

He states that the only Scotch rock resembling the Saville boulder which he is aware of is to be found in Sutherlandshire; but it has orthoclase as its felspar, and does not contain apatite. Should this boulder really prove to be of Scandinavian origin, its presence page 660 has an important bearing on the question! of the extension of the ice in the North Sea. Some smaller blocks of gneissose rocks occur in the neighbourhood. A few boulders of conglomeratic sandstones occur in Eda, which may be purely local.

On the Mainland blocks of white and reddish-grey sandstone are strewn on the hill-slopes north of Finstown and on the moory ground south of Maes Howe, which have been derived from the north-west shore of Scapa Flow; and so also along the west coast, between Brak Ness and Inganess, north of Hoy Sound, boulders of granite and gneiss are met with on the flagstone area to the west, of the axis of crystalline rocks.

VII. Conclusion.

The evidence now adduced regarding the glacial phenomena of Orkney is of the utmost importance in solving the question of the extension of the ice in the North Sea. We have already referred to the remarkable uniformity in the trend of the ice-markings throughout the islands, which, with certain exceptions, vary from W.N.W. to N.N.W. From the manner in which these striations maintain their persistent north-west trend, irrespective of the physical features of the country, it is evident that the agent which produced them must have acted independently of the islands.

Nay, more, the dispersal of the stones in the Boulder-clay leaves no room for doubt that the ice-sheet must have crossed the islands from the North Sea to the Atlantic. It is no doubt true that the lithological varieties of the Orcadian rocks are not so numerous as in Shetland, and hence the corroborative evidence of the northwesterly movement is not so abundant. Still in those cases where the geological structure of the ground permitted us to test with certainty the direction of the ice-carry, we were driven to the conclusion that the ice-flow must have been towards the Atlantic. In Westra the Boulder-clay sections contain striated blocks of red and white sandstone, which have been derived from Eda, and it is particularly observable that they diminish in number as we move towards the north-west. In Shapinshay blocks of the slaggy diabase from the south-east corner of the island occur in the Boulder-clay near Galtness; and so also in the Mainland, the red and white sandstones which cross the centre of the island are represented in the moraine profonde on the shore between Houton Head and the Loch of Stennis. Yet, again, to the west of the axis of crystalline rocks at Stromness, smoothed blocks of gneiss and granite are found in considerable abundance.

Fortunately, however, we have additional evidence which enables us to demonstrate, not only that the ice-movement must have been from the North Sea towards the Atlantic, but, what is of still greater moment, that the ice which glaciated Orkney must have come from Scotland. In the numerous sections of Boulder-clay described in this paper we have had occasion to refer to the occurrence of smoothed and striated stones of dark-grey limestones full of plant-remains, oolitic limestone, calcareous breccia, chalk, chalk- page 661 flints, fossil wood, pink granite, porphyritic felstone, &c., all of which are foreign to the islands.

According to the opinion of Mr. Carruthers, F.R.S., the blocks of dark-grey limestone with plant-remains in all probability belong to the Calciferous Sandstone series. He has identified a well-marked specimen of Lepidostrobus in one of the blocks, though it is not distinct enough to bo named specifically. Lithologically the boulders resemble some of the thin limestone bands in the Cementstone series of Central Scotland; and the nearest locality to Orkney where these rocks occur in situ is in the county of Fife. With reference to the Secondary rocks, Professor Judd, F.R.S., states that, besides the chalk and chalk-flints, he detected amongst our collection some specimens which resemble some of the Secondary rocks of Scotland. Two specimens of the calcareous breccia from the Boulder-clay in Odin Bay "very closely resemble parts of the Upper Oolites of Sutherland," and two other blocks are probably from the same locality. Moreover, he adds that the specimens of oolitic limestone very possibly come from some part of the Secondary series in Scotland.

In addition to these, wo observed, in the Odin-Bay section, large blocks of a remarkable rock which seems to be petrified wood. It has a curious fibrous structure and is very calcareous; indeed under the microscope it appears to be mainly made up of crystals of calcite, though occasionally there are portions where the structure is still retained. Blocks of the same rock, however, occur in the Caithness Boulder-clay, which show traces of organic structure under the microscope. On dissolving a small piece of the rock a large residue of coaly matter was obtained, which ignited with a strong flame. It would appear that this rock is largely burnt for lime in Sutherlandshire, where it is washed out of the Oolitic shales.

In all probability most of the blocks of granite, felstone, gneiss, quartzite, and schist which occur throughout Orkney, save those in the Stromness district, have been derived from the northeast of Scotland, though they possess no special characteristics which might enable us to identify them with any particular locality.

Now it ought to be borne in mind that chalk, chalk-flints, and various rocks of Jurassic age are found in the Boulder-clay of Caithness, and also in the same deposit in the low grounds of Banffshire and Aberdeenshire, where it possesses the same physical characters as in Orkney, and likewise contains fragments of shells. It seems perfectly reasonable to conclude, therefore, that the Boulder-clay in these widely separated localities must be ascribed to a common cause, or, in other words, to the action of land-ice. Indeed no one who attentively examines the sections in Orkney would ascribe them to the action of icebergs or coast-ice. We have already discussed the objections to the marine origin of the Shetland till, and the very same arguments apply with equal force to the present case.

Moreover, on referring to the chart showing the probable path of the ice in the North Sea, which accompanies this paper (PI. XXVII.), it will be seen that it is impossible to escape this conclusion. The ice, which radiated from the north-east of Scotland, not only filled the basin of the Moray Firth, but likewise spread over the low grounds of Banffshire and Aberdeenshire. The researches of previous investigators point to this conclusion; and quite recently, during the prosecution of the Geological Survey of the south side of the Moray Firth, additional facts have transpired which tend to confirm it. Further, in the neighbourhood of Dunbeath, on the Caithness coast, the stria; gradually swing round till they run parallel with the shore, eventually bending inland till they point towards the northwest, in harmony with the trend of the ice-markings in Orkney. Clearly, then, the ice must have been deflected so as to override the low grounds of Caithness, as pointed out long ago by Dr. Croll. Similarly in Forfarshire and Kincardineshire, the ice which moved off the south-east slopes of the Grampians, on reaching the coastline, was bent round in a N.N.E. direction, as indicated on the chart. A glance at the chart will also show how the land-ice was deflected along the south-east coast of Scotland, as described by our colleague Dr. James Geikie, F.R.S. Now these marked deflections undoubtedly point to some opposing force which was capable of overcoming the seaward motion of the Scotch ice-sheet. Had it been allowed to follow its natural pathway then the phenomena would have been widely different.

The results of our investigations in Shetland prove that the Scandinavian mer de glace not only invaded the North Sea, but likewise overlapped that group of islands in its march to the Atlantic. The presence of this mass in the bed of the German Ocean furnishes a satisfactory explanation of the phenomena above referred to; for the two ice-sheets must have coalesced on the sea-floor, and the combined ice-field would naturally take the path of least resistance. In other words, one portion would flow north-westwards by the Orkney Islands, while the southern portion would flow in the direction of the English coast, as laid down on the chart. In all probability the dividing line would be somewhere opposite the basin of the Forth.

We can quite well understand therefore how the Scotch ice-sheet, as it crept outwards along the bed of the Moray Firth towards the North Sea, must have pushed along the marine shells and silt which it encountered on the sea-floor. These would be commingled with the moraine profonde which had gathered underneath the ice-sheet; and the shells would ultimately be smoothed and striated precisely like the stones in the bottom moraine. Hence the occurrence of Scotch rocks together with shell-fragments in the Orkney Boulder-clay is what we would naturally oxpect; and in the light of the foregoing reasoning all difficulty as to the explanation of the phenomena disappears. It is not necessary for us to assign the precise localities from which the various foreign rocks have been derived; it is sufficient for our present purpose if we show, as has been done, that page 663 they may have come from the basin of the Moray Firth or the eastern counties of Scotland lying to the north of the basin of the Forth. The presence of blocks of limestone of Calciferous-Sandstone age in the Odin-Bay section in Stronsa seems to indicate that a portion of the ice which crossed Fife was deflected to the north; and even if the Saville boulder should prove to be of Scandinavian origin, its position in the north of the group is quite in keeping with the path which would be followed by the Scandinavian ice.

It is a significant fact that nowhere in the Shetland Boulder-clay did we find a vestige of the Secondary rocks of Scotland; and though the evidence is merely negative, it nevertheless confirms the foregoing conclusions. We are inclined to believe also that the absence of marine shells in the same deposit, which we noted in our previous paper, may probably indicate that a portion of the present sea-floor round Shetland formed dry land during the climax of glacial cold. We see, therefore, how the glacial phenomena of Orkney furnish a striking confirmation of the views advocated by Dr. Croll more than ten years ago.

Though we visited nearly all the islands of the group and traversed the greater part of the coast-line, we found no trace of gravel kames or raised beaches indicating recent changes in the relative level of sea and land.

Explanation of the Plates.

Plate XXVI.

Glacial Chart of the Orkneys.

Plate XXVII.

Chart showing the probable path of the ice in the North Sea.