CHatnell Httioetatta Siibrarg Stljaca, Neto ^nrk ' BOUGHT WITH THE INCOME OF THE SAGE ENDOWMENT FUND THE GIFT OF HENRY W. SAGE 1891 Cornell University Library QE 264.C97C7 1911 The geology of Colonsay and Oronsay, wit 3 1924 003 892 639 Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924003892639 MEMOIRS OF THE GEOLOGICAL SURVEY. SCOnAJNFD. THE GEOLOGY OF COLONSAY AND ORONSAY, WITH PART OF THE ROSS OF MULL. (EXPLANATION OF SHEET 35, WITH PAKT OF 27.) BY , E. H. CUNNINGHAM CKAIG, B.A. ; W. B. WEIGHT, B.A. ; AND E. B. BAILEY, B.A. / WITH NOTES BY G. T. CLOUGH, M.A.; aJto J. S. FLETT, M.B., CM., D.Sc. PUBLISHED BY ORDEE OF THE LOKDS OOMMISSIONEES OF HIS MAJESTY .S TKBASURY. EDIKBURGH : PRINTED FOB HIS MAJESTY'S STATIONERY OFFICE By MORRISON & GIBB LIMITED, At Tanfield. And to be purchased from. E. STANFORD, 12, 13 and 14 Long Acee, London ; W. & A. K. JOHNSTON, LIMITED, 2 St. Anbkbw Squaee, EoiKBrRGH ; HODGES, FIGGIS & Co., Gbafton Steeet, Dublin. From any Agent for the sale of Ordnance Survey Maps; or through any Bookseller from T. FISHER UNWIN, 1 Adelphi Tbebaoe, London, ■W.C, who is the sole Wholesale Agent to the Trade outside the County of London. 1911. Price Tivo 'Shillings and Threepence. lilST OF PUBIilCATIONS OF THE GEOLOGICAL SURVEY OF SCOTLAND. Gaological Map (25 miles to the iach) of the British Islands. Price, . . Coloured, 2s. ; Unooloured, Is. Maps on One-inch Scale. (Colour-printed.) Arran. (New edition.) 2s. (See references to Sheets 13 and 21 under Hand-coloured Maps. ) 32. Edinburghshire, Linlithgowshire, Fifeshire and Peeblesshire (parts of). (New edition.) (See Hand-coloured Maps. ) 2s. 6d. 33. Haddiitgtonsliire and parts of Edinburghshire and Berwickshire. (New edition. ) 2s. Sd. (See reference to first edition under Hand-coloured Maps. } 35. Colonsay and Oronsay, with part of the Boss of Mull. 2s. 6d. 71. Glenelg, Lochalsh and the South-East part of Skye. 2s. 6d. Maps on One-inch Scale. (Hand-coloured.) 1. Wigtownshire, South-Western Districts. 2s. Sd. 2. Wigtownshire, South-Eastern Districts. 2s. Sd. 3. Wigtownshire, Western Districts. 5s. '9c?. 4. Wigtownshire, East Part ; Kirkcudbrightshire, portion of S.W. Division. 9s. 6d. , 5. Kirkcudbrightshire, Southern Districts. 13s. Zd. 6. Kirkcudbrightshire, E. margin; Dumfriesshire, S. margin. 3s. Qd, 7. Ayrshire, South-Western Districts. 7s. Zd, 8. Kirkcudbrightshire, Ayrshire, and Wigtownshire (parts of). 16s. Zd. 9. Kirkcudbrightshire, N.E. ; Dumfriesshire, S.W. lis. 9d. 10. Dumfriesshire. 10s. Zd. 11. Roxburghshire and Dumfriesshire (parts of). 5s. Qd. 12. Argyllshire (Kintyre, S. half of), lis. 13. Ayrshire, Turnberry Point, and S. part of Arran (Solid, third edition, and Drift, first edition). Is.Zd. (See Colour-pitinted Map of Arran.) 14. Ayrshire, Central Distiicts. 14s. 15. Dumfriesshire, N.W. ; Ayrshire, S.E. ; and Lanarkshire, S. lis. IS. Dumfries, Selkirk, Peebles, Lanark, and Eoxburgh shires (parts of), lis. 17.vKoxburghshire, Selkirkshire, and Dumfriesshire "(parts of). 10s. Zd. 18. Roxburghshire, E. part. 2s. Qd. 19. Argyllshire (S. part of May). 10s. Zd. • 20. Argyllshire (Kintyre, Gigha L, part of Islay). 9s. M. 21. Argyllshire ; Arran, Central and N. part ; Bute, S. part ; Cumbraes, Ayrshire (part of N.W.) Solid and Drift, first editions. 16s. 3d. (See Colour-printed Map of Arran. ) 22. Ayrshire, Renfrewshire, Lanarkshire (parts of). 14s. %d. 23. Lanarkshire, Central Districts ; Ayrshire (part of W.). 19s. Zd. 24. Peeblesshire, Lanarkshire, Edinburghshire, Selkirkshire (parts of). 8s. 9d. 25. Berwick, Eoxburgh, Selkirk, and Edinbm-gh shires (parts of). 8». M. 26. Berwickshire and Roxburghshire (parts of). 3s. Zd,. 27. Argyllshire ; parts of Islay, Jura and Oronsay. 8s. 29. Argyllshire, Ayrshire, Buteshire, Dumbartonshire, and Renfrewshire (paits ot). 21s. U. , , r 30. Renfrewshive ; jjarts of Dumbarton, Stirling, Lanark, and Ayr. 16s. Zd. 31. Lanarkshire, Stirlingshire, Linlithgowshire, Dumbartonshire, Edinburgh- shire (parts of). 16s. Zd. 32. Edinburghshire, Linlithgowshire, Fifeshire, Peeblesshire (parts of). 13s. Zd. (See Colour-printed Maps. ) 33. Haddingtonshire and parts of Edinburghshire apd Berwickshire (first edition). Ws.Zd. (See Colour-printed Maps.) 34. Eastern Berwickshire. 3s. %d. 36. Seaboard of Mid Argyll (Solid and Drift editions). 16s. Zd. 37. Mid Argyll (Solid and Drift editions). 26s. Zd. 38. Perthshire, Stirlingshire, Dumbartonshire, Argyllshire (parts of). 14s. M. 39. Perthshire, Clackmannanshire, Stirlingshire, and Fifeshire (parts of). 17s. 9d' 40. Fife and Kinross. 14s. M. 41. Fife, East part ; Haddingtonshire, North part. 5s. M. 45. Argyllshire, Country near Oban and Dalmally (Solid and Drift editions). 25s. Zd. 46. Perthshire, Argyllshire (parts of). 19s. Zd. 4tl. Perthshire. 16s. 3d. ' 48. Perthshire, Forfarshire, and Fifeshire (parts of), lis. 49. Forfarshire and Fifeshire (parts of). 6s. 65. Perthshire (Solid and Drift editions). 22s. Zd. and 25s. Zd. 66. Perthshire, ]Forfarshire (parts of). 22s. Zd. 57. Forfarshire and Kincardineshire (parts of). 10s. Zd. 57a. Kincardineshire, S.E. corner. Is. M. 60. Rum, Canna, Eigg, Muck. 10s. Zd. 65. Aberdeenshire, Forfarshire, Perthshire (parts of). 25s. 3d. 66. Kincardineshire, Forfarshire, Aberdeenshire (parts of). 23s. 9d. 67. Kincardineshire and Aberdeenshire (parts of). 5s. 9d. 70. Inverness-shire (West-Central Skye, with Soay), 19s. 3d. M&GLtd Wt 36454/37 5-H 500 PLATK I. (Frontispiece.) coLO]vs.«r MILLBVIE GROUP KILCHATTAN, GROUP ^"^^ MILLBUIE GROUP olMILLBUlE CROUP JSich.aMdiii MACHRINS GROUP. STAOSUNAIG PHYLLITES fcOLONSAY LIMESTOnt: i^MLORAK FLASS '^MILLBUIE. GROUP O/^ONSAY MUOSTONES (ORONSAY SANDSTONES i" Sketch-map of the stratigraphy of Colonsay and Oronsay. (Scale ; 2 miles = 1 inch.) KILCHATTAN PHYLLITESSSANOSTONES STAOSUNAIG DARK PHYLLITES | COLONSAY LIMESTONE I.'.' ■! MACHRINS GRITS JJ^;;*J i^X Xy, "^xtx p™.«jjj.^==e==«=y^=^*=#^ SCALASAIG HARBOUR *MONUMENT \ ^^-^ jO \ ^. tC^ ^^S^i. T DYKES ■m \ i\ ^% J^rf iJlMPfiOfifrfiE AND FOUATED 1 jr--^^^ lAMPfiOPffY/tE SMBEiTS kA> 1 y^^t^^A^^ OlOBtTE ^4 VVv'^K jQS^^^^^t-OED AND eiMAVEO -or^y — >i y^^ BRECC/A AT MABSIN OF OIORITE -rrt^^^SMEARED MjtBtW OF Oiefl/T£ V'^J LOCH STAOSNAIG ^'.'yy-'ABTIALLY meceiATED FLACS INJECTED WITH DiOR/TC SCALE OF FEE 7 Mw _ » S4M> WM moo Fig. 5. — Sketch-map of the Scalasaig diorite intrusion, showing the associated breccia and the localities where evidence has been obtained as to the relation of the diorite to the second movement. Some of the foliated lamprophyi-e sheets may be apophyses of the diorite ; the one on the north shore of Loch Staosnaig contains many conspicuous green pseudomorphs after augite, like those of the neighbouring diorite edge. notice. That which occurs along the margins of the Scalasaig diorite is by far the most extensive, covering about 15 acres of groimd. It Associated Breccias. 35 is often fairly well exposed and to a large extent consists of the local flags, phyllites and grits of the Torridon ; blocks foreign to the island have not been found here, but rocks not found in contact with the breccia, e.g. Umestone, are occasionally represented. The Kiloran Breccia. — -The breccia associated with the Kiloran syenite, though less extensive, is much better exposed. At its extreme outer edge, where it comes in contact with the phyUites, the breccia is composed of angular fragments of the latter interlocked with one another in the elose-fitting manner shown in Plate V. A few feet in from the marj^ boulders and great angular blocks of quartzite foreign to the island begin to make their appearance and soon form a large part of the mass. Blocks of whitish limestone, probably the adjoining Torridon hmestone in a baked condition, and fragments of decomposed lamprophyre also occur. Some of the quartzite boulders, even when as much as 2 ft. in diameter, are so remarkably well rounded as to be suggestive of water action. The quartzite of which they and the much larger angular masses consist, is frequently pebbly and, as far as can be seen, indistinguishable from that which, in Islay, is thrust forward on the top of the Torridonian rocks. It is therefore similar also to the quartzite occurring as small pebbles at two horizons in conglomeratic beds of the Torridonian sequence of Colonsay. This breccia, on account of the foreign material it contains, must either have fallen from above into an open fissure or pipe, or else it must have been brought up from below by some sort of volcanic or plutonic action. The presence of large rounded boulders without adherent matrix is more in consonance with the former alternative, since it seems more probable that these boulders have been derived from a loose surface formation than that they have been separated from a deeply buried and consolidated conglomerate. It may be supposed, for instance, that the breccia owes its origin to the down- fall, of material from the sides of a vent blasted out by steam ; such falling in of material from above has been proved in the case of the Arranneck described by Messrs. Peach and Gunn.* In the latter case, however, igneous agglomerate is associated with the sedimentary material of the breccia, a link in the evidence which is absent in the Colonsay examples. Volcanic vents filled solely with sedimentary breccias are however known, such as some of the smaller necks of the Dunbar shore described by Sir A. Geikie and Mr. Maufe.t The association of plutonic igneous rocks with the breccias in Colonsay is in agreement with the explanation suggested here, since both in Arran and Skye | it has been shown that plutonic rocks are apt to find their way up pipes filled by agglomerate and breccia. Some of the lamprophyre dykes of the island contain fragments of quartzite, similar to that occurring in the breccias associated with the plutonic rocks, but, so far as has been observed, always isolated in the igneous matrix. It seems unprofitable at present to discuss whether these fragments have been carried forward from the parent magma mass, or picked up incidentally from Torridonian conglo- merates, or derived in some other manner. It is important to note, * " The Geology of North Arran," 3Iem. Oeol. Survey, 1903, chaps, viii. andix. t " The Geology of East Lothian," 2nd od., Mem. Geol. Survey, pp. 90, 92. t " The Tertiary Igneous Rocks of Skye," Mem. Geol. Survey, 1904, p. 16, etc. 36 Plutonic Intrusions of Colonsay. however, tliat the dykes in question were probably injected later than the period of earth-movement which followed the upwelUng of the plutonic rocks associated with the breccias of Kiloran and Scalasaig. THE AGE OF THE PLUTONIC EOCKS. In the first place it appears certain from the field evidence that the Kiloran syenite and the Scalasaig diorite belong to one and the same phase of igneous activity ; each is associated with a breccia of most exceptional type, and each was intruded during a pause between movements of compression which, as wiU be shown in Chapter VII., gave rise to important cleavage structures. In the second place on petrological grounds it is difficult to believe that the kentallenite of Balnahard can belong to a distinct period. It would be a strange coincidence if rocks so nearly allied, and at the same time so peculiar as the Scalasaig augite diorite and the Balnahard kentallenite, should have had an independent origin within so narrow a compass as the island of Colonsay. It must be remembered that, not only has this association of augite diorite and kentallenite been recognised by Messrs. Hill and Kynaston as characteristic of the main- land occurrences, but also that the Scalasaig augite diorite itself, at one point, merges into true kentallenite. This view, ranging the various plutonic intrusions of Colonsay alongside of one another, does not necessarily preclude the possibility that the kentallenite may be later than the second cleavage movement. The relation of the kentallenite intrusion to the period of second cleavage can therefore only be settled on direct evidence, and this unfortunately ofiers grave difficulties of interpretation. The suggestion, offered above, however, that the second cleavage movement of Colonsay may have been but an interlude during a single period of intermittent igneous activity, receives some support when a comparison is instituted between the history of Colonsay and that of the mainland. Many will unhesitatingly place the Colonsay kentall- enite and augite diorite, along with their analogues on the mainland, in the suite of Newer Granite intrusions. This suite is extensively developed both in the Highlands and Southern Uplands of Scotland, and is of Lower Old Red Sandstone age. No later age can be postulated either for the second cleavage movement or for the later uncleaved lamprophyres of Colonsay. The conception outhned above has the advantage of regarding all the lamprophyres of Colonsay as a single extended suite ; it also ofiers an explanation of the marked petrological affinity of the later lampro- phyres and the Kiloran syenite. It may be added that cleaved lamprophyres are known in many parts of the Highlands, and have been generally regarded as early offshoots of the Newer Granite magma. The well-known foliation in the Crifiel granite of the Southern Uplands may also be cited as an instance of a demonstrable Old Red Sandstone granite affected by intense movement. In this latter case, however, it is not clear from the descriptions whether the movement was regional, as in Colonsay, or purely local and conditioned by the circum- stances of the igneous intrusion. Age. 37 There is of course a fundamental weakness in the argument de- veloped above. Petrological correlations, when applied to regions some little distance apart, are apt to prove misleading. It can scarcely be doubted that the Scalasaig augite diorite and the Balnahard kentallenite belong to the same igneous period, and also that Messrs. Hill and Kynaston are correct in grouping the similar intrusions of the mainland with the Newer Granites of their neighbourhood ; but it is a larger step fo extend the generahsation and include the Colonsay masses in the same great suite. It is a larger step, and ia view of the consequences entailed, it is one that we cannot accept with entire confidence. CHAPTER VI. MiNOE Intrusions of Colonsay and Oeonsay. The great majority of the minor intrusions of Colonsay and Oronsay are either lamprophyres or basalts, but there are also a few felsites. The lamprophyres attain their maximum development in the north and east of the island, and the basalts in the west and south. The latter occur as vertical dykes which are probably of Tertiary age and have the familiar north-westerly trend characteristic of similar intrusions ia the West Highlands. The lamprophyres on the other hand dip at aU angles and have commonly a sill-like character, their direction being largely controlled by the structure and bedding of the sediments. These difEerences suggest a striking, if as yet ill understood, contrast between the conditions of orustal stress which regulated the two types of intrusion. It may be regarded as certain from analogy with other parts of the West Highlands that all the lamprophyric intrusions are not only earUer but much earlier than any of the basaltic series. LAMPROPHYRE DYKES AND SHEETS. The Colonsay rocks which have been grouped together as lampro- phyres possibly include many types essentially distinct in petrological character and in age. These types are, however, difficult to distinguish in the field ; only the more recent are sufficiently fresh to render identification under the microscope at all possible ; the remaiader are composed largely of secondary minerals, principally chlorite, epidote and calcite, and effervesce freely in acid. Such an advanced state of decomposition is frequently met with among lamprophyres elsewhere. They possess, moreover, in common with the fresher types, the extreme intractability to the hammer which characterises this group of rocks. As regards distribution, the lamprophyre intrusions seem to be absent or at least very rare in Ardskenish and south of a line drawn thence to Loch Staosnaig. Northwards they increase in abimdance and are remarkably developed in the whole of the country north-east of Kiloran. They appear to be wanting in Uragaig. All the lamprophyres appear to be later than the movement which caused the first cleavage (see Chap. VII.). The greater number of them are earher than the second. They are afiected by the cleavage and folding produced during the second movement (see Fig. 6), and tms has largely contributed to their advanced state of decomposition. Some are, however, distinctly later than the second movement, the folds of which they truncate (see Fig. 7). They may thus at once be divided into two classes clearly separated from one another in point of age, and contrasting strikingly in several important respects. Lampro-phyre Dykes and Sheets. 39 shore north of Port an Obain, northern pnd of Colonsay, in- truded along the cleavage of the primary movement, and involved in the folding of the secondary movement. The dykes of the first group are generally intruded along the bedding planes or along the first cleavage, being thus, in some sense, of the nature of sills, often dipping at low angles and pursuing a sinuous course across the country. Occasionally, however, they are exceedingly irregular, and may change suddenly from dyke form to sill form and vice versa. They often show considerable variation of character within a short distance, but are on the whole more homo- geneous than the later dykes. None are sufficiently fresh for identifi- cation, and many of them exhibit both in the hand specimen and under the microscope a pronounced foliation. The proportion of cleaved to uncleaved lamprophyres is much greater on the east side of Colonsay than on the west. The dykes and sills of the second group are much less numerous, but considerably more massive than those of the first, from which they are also easily distinguished by their freshness, even in cases where the relation to F^g- 6.— Lamprophyre dyke on the the secondary folding is not at once obvious. Petrographically they are all vogesites, but, as is usual with rocks of this class, they vary considerably in character, not only from place to place, but also in the same exposure, or even in the same hand specimen. They have generally, but not exclusively, an east and west trend, and are in their manner of intrusion much more inde- pendent of the structure of the rocks than their predecessors. Two remarkable examples cross the north end of the island in an approxi- mately east and west direction, dipping south at a considerable angle. On the coast, about a quarter of a mile north of Port na Cuilce, the relation of one of these to the secondary folding is well seen. It cuts across and truncates the folds of the secondary system in such a manner as to leave no doubt that it is subsequent to them (see Fig. 7). East of Port na Omlce the other is very well ex- posed, having a thickness of 20 Fig. 7.— Margin of an ejst and west vogesite f^^ .r. r, j, • hetero- dyke, north of Port na Cuilce, truncating ^° *" ^\ " ^^ ^^^J Jietero the folds produced by the second move- geneous m- character, and is ment in the first cleavage. Scale = about charged with numerous frag- one-third of the natural size. ^^^^^ ^^ vein-quartz. A few pieces of phyUite, showing the crinkling produced in the cleavage by the secondary movement, can be seen enclosed in the dyke here and there. The crinkling could not in this case have been produced after the fragment had been incorporated in the dyke, for the numerous structural hues close by in the dyke show no such puckering. This latter dyke is for the most part exceedingly coarse-grained, and has a general resemblance to the syenite of Kiloran Bay, being in places very hornblendic, and in others very felspathic. The resem- blance is heightened by the numerous inclusions of vein-quartz. These, 40 Minor Intrusions of Colonsay. however, have no appearance of having undergone solution, such as has been described in the case of the similar fragments in the horn- blendite margin of the Kiloran Bay intrusion. Under the miscroscope (13533 and 13534), the hornblendes can be seen to be mostly idioT morphic, but some of the larger crystals are corroded. They are set in a coarse-grained matrix of orthoclase, plagioclase and micropegmatite, the orthoclase predominating. The plagioclase, which is not abundant, is probably andesine. There are a few quartz grains which, however, may possibly be derived. The quartz of the micropegmatite is some- times in optical continuity with these grains. Of the secondary minerals the most striking is the colourless hornblende, which is grown on to the brown hornblende in optical continuity, even the most dehcate twinning passing from one to the other. Chlorite and calcite are largely developed, while the felspars are considerably micacised. The rock is a typical vogesite. In the eastern part of its course however near Leac Bhuidhe, it contains a fair amount of biotite (13539). A 7-foot vogesite dyke, of irregular east and west trend, occurring on the shore north of Port Mor, Kilchattan, is, judging from its remarkable freshness, probably of this age. It is composed (13537) of strikingly idiomorphic lath-shaped hornblendes of a yellow to green.- ish brown colour, and colourless or very slightly coloured augites, not so well formed as the hornblende, but still somewhat idiomorphic. These are surrounded more or less ophitically by orthoclase and a little soda-lime felspar of intermediate composition, possibly andesine. Apatite occurs as needles penetrating the other minerals. Pale green secondary hornblende is sometimes grown upon the original mineral. Bpidote occurs in secondary grains and strings, forming pseudomorphs with chlorite. These latter are occasionally fibrous, and may in this case be after biotite. The rock compares remarkably well with an augite vogesite from Schiehalhon (11271), described and figured by Dr. Flett.* There is, however, somewhat less augite and more hornblende in the Colonsay rock, while the plagioclase felspar is also shghtly more basic. The 15-foot " sill " which has been traced for about three-quarters of a mile in a north-easterly and northerly direction across the hills north of Kilchattan, dipping La a westerly direction contrary to the cleavage and bedding, is also an augite vogesite (13531). It may possibly be a continuation of the above described dyke north of Port M6r. One of the most striking features of the vogesites of Colonsay is their extremely heterogeneous character. This is well shown in the series of dykes or sills to the east of the Sandhills of Kiloran Bay. These are intruded along the bedding of the flags, which, however, dips at an angle of 50°. The margin for about a foot in is fairly homogeneous, with well marked flow lines parallel to the edge, but the centre is a veritable breccia, consisting of angular fragments of dark, coarser-grained rock, set in a lighter-coloured and finer-grained matrix, recalHng the xenolithic structure of the Skye peridotites. One of these dykes is traceable for a short distance up the hill, and in that direction takes on a very coarse-grained or even plutonic aspect. * " The Geology of the Country round Blair AthoU, Pitlochry, and Aberfeldy (Explanation of Sheet 55)," Mem. Geol. Survey, 1905, p. 123, plate vii. fig. 3. North-West Dykes. . 41 This dyke here and there contains detached pieces of the surround- ing schists, as well as numerous fragments of slightly pebbly felspathic quartzite, similar to that obtained from the Kiloran syenite and breccia, and from certain horizons in the Torridon sediments. A similar quartzite has also been obtained from a lamprophyre dyke in the neighbourhood of Dun Crom, north of Balnahard. In several places along the north-west coast, the later predomi- nantly east and west vogesite dykes can be seen cutting the earlier lamprophyres, which trend in a northerly direction with the strike, and are traversed by the cleavage of the second movement. A good example can be seen on the foreshore in Port an Obain, the later east and west dyke being in this case a homblende-vogesite. The felsite dykes occur very sporadically in different parts of the island. Their period of intrusion is probably much the same as that of the lamprophyres. They are for the most part too decomposed for petrological examination. THE NOETH-WEST DYKES. From Notes supplied by J. S. Flett. The north-west dykes of Colonsay include varieties of olivine dolerite and monchiquite. They belong to a series which has a wide distribution in Argyllshire and the adjacent part of the west High- lands. In Jura and in Knapdale,* on the mainland of Scotland, there are dolerite dykes with the same strike and the same petro- graphical characters as those of Colonsay, and in Ardmucknish t mon- chiquites and camptonites occur very like those of Colonsay and the southern part of Mull. Hitherto it has been assumed that all these dykes are of Tertiary age, but if they are Tertiary they differ in some important respects from the dolerite dykes which Mr. Harker has described from Skye.f The most important distinction is the occur- rence of camptonites and monchiquites (some of which contain nephe- line) among the Argyllshire dykes. In the south of Mull no camptonites or monchiquites have yet been found cutting the Tertiary lava flows and sills, but there are a few Tertiary dykes of analcite dolerite that seem to present afiinities to the dolerites which accompany the mon- chiquites of Colonsay. Among the north-west dykes of the Beinn Dearg group in Skye Mr. Wright has recently fovmd analcite dolerites essentially similar to those of Colonsay. It is clear, at any rate, that the north-west dykes of Argyllshire are post-Carboniferous, for in Knapdale (Sheet 28) two of them have been proved to intersect the east and west quartz dolerite dykes which are known to be of late Carboniferous age.§ The olivine dolerites are fine-grained dark rocks, rarely vesicular or porphyritic. They contain much olivine, usually fresh and well preserved. The zonal felspars are lath-shaped with central areas * " The Geology of Knapdale, Jura and North Kintyre," JItm. Geol. Survey, p. 108. ■f " The Geology of the Country near Oban and Dalmally," Mem. Geol. Survey, 1908, p. 124. J '''The Tertiary Igneous Rocks of Skye,'' Mem. Geol. Survey, 1904. §"The Geology of Knapdale, Jura and North Kintyre," Mrm. Geol. Survey, p. 108. 42 Minor Intrusions of Colonsay. of labradorite, while the margins are sometimes oligoclase. The augite has a purple colour and slight dichxoism. Biotite occurs in small scales adherent to the iron-oxides. The ophitic structure of these rocks is exceedingly perfect ; small lath-shaped crystals of felspar are embedded in large plates of pirrple augite, and even the. olivine is interfered with in its growth by the felspar, as the edges of its grains are often moulded on felspar crystals. The irregular masses of iron oxides, also, are often penetrated by bars of plagioclase. In another group of these dykes the same minerals occur as in the olivine dolerites, but a considerable amount of analcite and zeolites is also constantly present. These rocks might be described as analcite-bearing dolerites, but as they are very common and have well-defined characters, it will be more convenient to give them a distinctive name. They are abundant among the north-west dykes of Jura and Colonsay and the adjacent parts of Argyllshire (Knapdale). Some of the dykes of the Beinn Dearg group in Skye belong to the same series. We propose to call these rocks crinanites after Loch Crinan, which lies nearly in the centre of their area of distribution. The crinanites, then, are dark-coloured, fine-grained basic rocks (Plate VI. Kg. 4) consisting mainly of oUvine, augite and plagioclase felspar, with a considerable amount of analcite and zeoKtes. Olivine is abundant in small grains, more or less altered to serpentine. The augite is always purple and is sometimes bluish or plum-coloured; it is pleochroic with X brown, Y reddish-violet, Z darker bluish-violet, but the depth of the colour varies in different crystals ; the orientation is XYX=a6c, and the extinction angle about 44°. The pyro- xene in fact belongs to the variety usually described as titaniferous and much resembles that which occurs in many basic nepheline-rocks and teschenites. Chemical analysis proves that the crinanites are rich in titanium. The augite has a marked dispersion of one of the optic axes and shows hour-glass and zonal structures. The felspar has albite (rarely Carlsbad or pericline) twinning and belongs mostly to labradorite, though the outer zones are more rich in soda and may consist of oligoclase or albite. The iron oxides form irregular plates often fringed with small scales of dark brown biotite. Most of these rocks have very perfect ophitic structure, and the augite occurs as small angular patches between the lath-shaped felspars or enclosing them. Occasionally there is a certain number of pyroxenes showing idiomorphic outlines, and in some of the rocks the ophitic character is hardly visible. In a few specimens there are large corroded felspar phenocrysts, consisting mainly of bytownite. Analcite and radiating clusters of zeolites fill up spaces between the felspars or occupy small rounded steam cavities. Perfectly trans- parent analcite is not uncommon, but often this mineral is turbid and granular with weak double refraction. The radiate zeolite appears to be mostly natroUte. Evidently these have been the last minerals to crystallise, and as the rocks are often very fresh, it is difficult to believe that they have originated from the decomposition of felspar. They are more properly a pneumatolytic infilling of interstitial spaces during a period immediately following the crystaUisation of the pyiogenetic minerals. Carbonates and chlorite are often associated Norih-West Dykes. 48 with them, and veins of analcite and zeolites, easily distinguished by their low refractive indices, often ramify through the substance of the felspar. In their composition and in the properties of their minerals these crinanites bear much resemblance to the teschenites which are so largely developed among the Carboniferous intrusive basic rocks in the central valley of Scotland. The olivine, zoned basic plagioclase and purple augite, analcite and zeolites of the two groups of rocks are very similar, but the teschenites are much coarser- grained, less frequently porphyritic and contain much more alkali felspar. The teschenites occur as large sills or laccoHtes, the criaa- nites as narrow vertical dykes which often can be followed for long distances in nearly straight lines. The crinanites in Colonsay show transitions to the camptonites, and are associated with monchiqiutes, some of which contain nepheline. These facts show their theralitic afBnities ; no camptonites have been found among the Carboniferous intrusive rocks, but there are monchiquites in this series in Fife and in Bast Lothian, and nepheline occurs in some essexites and therahtes of Carboniferous age in the west of Scotland.* Although there are no camptonites in Colonsay, one of the crinanite dykes shows the ocellar structures which are so characteristic of the camptonites. The ocelli are small rounded spots which have a central area of analcite surrounded by a radiating zone of prismatic crystals of felspar, brown hornblende and purple augite, aU of these being in very perfect crystals. In the same rock there are also rounded vesicles which contain only analcite and zeoKtes ; this dyke (.13488) occturs at Port Mor on the west side of the island. Two monchiquite dykes of Colonsay deserve special notice as they furnish petrographical material of exceptional interest. One of them occurs near Reasagbuie on the east coast, the other and more important has been traced for half a mile in a north-west direction across the hills east of Lower Kilchattan. Both are essentially of the same type and present a striking appearance in the field, as they contain large crystals of biotite and hornblende. The largest pheno- crysts occur only in the central portion, and the parallel arrangement of the biotite plates gives rise to a rude cleavage structure. The Kilchattan dyke has long been known in the island, and our attention was first drawn to it by Mr. Archibald Campbell. The Reasagbuie dyke is a typical monchiquite and contains, in addition to the large porphyritic crystals of biotite and hornblende already mentioned, many small phenocrysts of ohvine and augite. The olivine is all decomposed into serpentine. The porphyritic augite has the usual purpHsh-brown colour and sHght dichroism; many of the crystals are repeatedly twinned on the orthopinakoid and have a cleavage or parting parallel to that face. The groundmass is rich in small eumorphic prisms of augite, often much longer than they are broad. There is also a fair amount of dark brown biotite and a nearly transparent granular material which is a mixture of analcite and carbonates. Numerous rounded ocelli are visible in the groimdmass ; they consist of analcite and subradiate plates of biotite. No felspar is present in the rock, and in the groundmass hornblende is absent or occurs only in inconsiderable amount. * Summary of Progress of the Geological Survey for 1908, p. 44. 44 Minor Intrusions of Colonsay. The dyke at Lower Kilcliattan is in one respect not a typical monchiquite, as it contains no olivine and no pseudomorphs whicH would render it probable that olivine had been originally present. It is also exceptionally rich in biotite ; these features mark it as belonging to the ouachitites as defined by J. F. WilHams.* It con- tains also a considerable amount of decomposed nepheline, and consequently we may describe it as a nepheline ouachibite ; the presence of nepheline in rocks of the monchiquite group has been ascertained by Lacroix f and others, but is by no means common. The fine dark matrix of the rock is full of minute scales of black mica, and there are also rounded white spots or " ocelli " and vesicles filled with calcite and analcite. The most striking peculiarity of the rock, however, is the presence of large black phenocrysts of biotite, augite and hornblende, which make it a typical " lamprophyre." Some of the hornblende crystals are IJ in. in diameter, and the augite and biotite are not much smaller. The black mica is easily distinguished by its perfect cleavage and the six-sided outlines of its crystals, which have their angles slightly rounded by corrosion. The crystals of augite and hornblende are less perfect, and both minerals are pitch black in colour. One of the sliced specimens of this rock has a crystal of apatite of irregular form nearly half an inch across. It is dark, being filled with minute inclusions, and its size is extraordinary in a rock of this kind. In section the porphyritic biotite is dark brown, with pleochroism from yellow to reddish-brown ; it is nearly uniaxial and very free from inclusions. The hornblende is of similar colour to the biotite and intensely pleochroic with X greenish-brown, Y darker brown and Z intensely dark brown. The augite is pale green, almost colour- less in thin sections ; it contains inclusions of biotite, hornblende and iron oxides, and has many cavities which are fiUed with carbonates and analcite. The large apatite, above referred to, contains innumer- able small tubular fluid inclusions. All of these minerals are much corroded and have embayed or irregular margins, and upon the pale green augite there is anexternal zone with a brownish colour, like that of the pyroxene of later formation in the rock. A composite phenocryst of remarkable character is visible in one of the slides ; it consists of biotite and augite which are in graphic intergrowth, so that, though the mica appears as brown scales in a matrix of augite, the position of extinction is the same for the greater number of these scattered individuals. The augite is of the pale- coloured variety usual in the phenocrysts, and has a thin border of brown augite like that of the groundmass. The interior of the crystal is darkened by the presence of great numbers of small grains of iron oxides, and the rounded shape of the crystal indicates that it has suffered much corrosion. Similar intergrowths consisting of augite and hornblende are very common in some of the camptonite dykes of the Orkneys. J * " The Igneous Rocks of Arkansas," Ann. Report Geol. Survey of Arkansas for 1890, 1891, vol. ii. p. 392. f A. Lacroix, " Matdriaux pour la Mineralogie de Madagascar," Nouv. Archives du Mtisenm (V Bisioire Naturelle (de. Paris), 1902, vol. iv. p. 141. ■| John S. Flett, " The Trap-dykes of the Orkneys," Trans. Boy. 8oc. Edin., 1900, vol. xxxix. p. 880. North-West Dykes. 45 These large porphjo'itic crystals belong to a very early stage in tlie crystallisation of the rock. They were followed by the formation of a second group of phenocrysts of augite which are only of small size, purplish-brown colour and have marked hour-glass structure and slight dichroism ; their central parts are usually paler than the margins. The augite, which is very abundant in the groundmass, is in much smaller prisms, with good crystalline outlines, but often long and narrow. No hornblende occurs with them, but there is a considerable amount of red-brown biotite which often sur- rounds the little augites, but has at times rather good hexagonal outhnes. In polarised light the mica sometimes shows twinning. It has a strong dichroism and is in good preservation. Small octahedra of magnetite and needles of apatite are common also in the groundmass. Surrounding these minerals there is a colourless substance some- times clear and transparent, but for the most part granular and cloudy. It contains decomposition products such as chlorite and carbonates. There is no felspar in the groundmass, but analcite occurs, sometimes isotropic, but often doubly refracting. With it there is another colourless mineral which has a weak double refraction, a straight extinction and the optical sign and refractive index of nepheline ; it is weathering to a fibrous zeoHte, presumably natrolite. In many specimens this mineral is almost completely decomposed ; in others, it is still present in fair quantity. It does not show good crystalline form, but fills up the spaces between the augite and biotite of the groundmass. Whether there was originally a glassy base in addition to nepheline cannot now be determined. Many small steam cavities in the rock are filled with analcite and carbonates, and some of these cavities have their walls lined with a zone of brush-Uke felspars and small blades of biotite converging towards the centre, which is occupied by analcite and carbonates. These are the ocelli which are visible in the hand specimens as white or reddish spots, and it is only in them that felspar occurs in the rock. In one of the largest ocelli the biotite has the usual radiate arrangement in the peripheral part of the ocellus, but has a tangential disposition in the part which is next the central cavity. Mr. Radley has made a chemical analysis of the nephehne ouachitite of Kilchattan in the Survey's laboratory. It is given on p. 46 (No. IV.), and with it are a number of analyses of north-west dykes of allied character from Skye and Argyllshire. No. II. is an ohvine dolerite from Skye which contains a small amount of interstitial zeolites. No. I. is a crinanite or analcite ohvine dolerite from Jura, No. III. is a camptonite from Ardmucknish, and No. V. is the remark- able monchiquite from the same district which carries large corroded phenocrysts of anorthoclase. The close similarity of these dykes is obvious from a glance at the analyses. In silica, alumina and (total) iron oxides they differ very little. The most interesting feature of the series is the progressive decrease of lime and magnesia and increase of alkalies as we pass from the dolerite through crinanite and camptonite to monchiquite. The apparently high amount of lime in the camptonite is due to abundance of calcite, as is evident on comparing the amount of carbonic acid in this and the other rocks. 46 Minor Intrusions of Colonsay. Slice No. 14174 7862 11817 13676 13191a I. n. ni. IV. V. Si02 43-94 44-01 42-22 39-54 43-55 TiOa . . 2-45 2-03 2-49 2-72 2-52 Al^O., . . 1403 12-32 10-62 13-74 14-60 Ve,0, . . 1-95 3-62 4-74 405 5-34 Cr^Oa . . trace trace 0-10 — V2O3 . . undet. — 005 FeO 11-65 8-75 6-18 , 7-8S 7-46 MnO .. 0-32 0-21 0-50 i 0-40 0-55 (CoNijO . . nt. fd. trace 005 0-05 0-05 BaO nt. fd. — 0-04 019 0-27 SrO nt. fd. — . nt. fd. 0-14 nt. fd. CaO 8-99 10-57 14-80 1 9-64 7-94 MgO . . 10-46 12-86 8-68 8-66 1 4-77 K2O 0-33 0-49 1-41 3-70 1 1-40 Na^O . . 2-68 1-68 2-46 2-24 501 LisO nt. fd. — ? trace trace trace H20atl05°C. . 0-85 0-89 0-50 0-61 0-63 H2O above 105° C 2-31 2-73 116 2-77 ' 4-58 P2O5 . . 0-20 0-17 0-73 1-01 1 0-88 FeSa . . 004 — 0-27 1 0-31 FejSs • ■ 0-06 — — S — 0-11 0-12 1 __ CO2 0-16 trace 3-57 i 2-48 0-56 T( )tal . • 100-42 100-44 100-42 10006 1 100-42 I. Crinanite, Sloo nan Sgarbh, 1 mile north of Inver Lodge, Jura (Sheet 27) (analysed by E. G. Eadley). II. Olivine dolerite, Ciche na Beinne Deirge, Skye (analysed by Dr. W. Pollard) ; see " The Tertiary Igneous Rocks of Skye," Mem. Oeol. Survey, 1904, p. 325 (the titanium dioxide and alumina have been corrected according to Dr. Pollard's latest tests). III. Camptouite, Sailean Shgenach, Ardmuckuish, Argyllshire (analysed by E. G. Radley) ; see " The Geology of the Country near Oban and Dalmally," Mem. Oeol. Survey, 1908, p. 126. IV. Nephehne ouachitite, Kilohattan, Colonsay (analysed by E. G. Radley). V. Monohiquite, west side of Camas an Fhais, Argyllshire (analysed by E. G. Radley); see ''^The Geology of the Country near Oban and Dalmally," Mem. Qeol. Survey, 1908, p. 128. CHAPTER VII. Tectonics of Colonsay and Oeonsay.* General Account of the Structure. — The tectonic arrangement of the Torridonian sediments, of which the islands of Colonsay and Oronsay are largely composed, possesses in its major features a mach greater simplicity than one would be led to expect by the observation of isolated exposures. In many places, it is true, intense puckering and contortion, coupled with isoclinal folding, can be traced ; and cleavage is everyTfhere apparent in the more peUtic beds. None of these folds, however, attain an amplitude sufficient to disturb to any extent the general disposition of the- strata, which preserve, when viewed broadly, a uniform dip, or locally form themselves into comparatively simple anticlines and syncUnes. The Colonsay Limestone, which, with the beds above and below it, constitutes an easily recognisable horizon, occurs on the eastern coast of the island, dipping out to sea at a low angle. If a traverse be made in a westerly or southerly direction across the islands from this outcrop, successively lower beds are passed over, all dipping east or north-east at gentle angles ; and on finally reaching the extreme outlying parts of Oronsay and Ardskenish, there is still no indication of any base to this enormously thick series of sediments. Throughout this whole traverse one has passed continuously from higher to lower beds, and there has been no repetition of any importance. In the northern part of Colonsay the relations are slightly more com- plex. The outcrop of the limestone mentioned above skirts the east coast of the island for some distance, and then passes out to sea north of Scalasaig. A traverse made thence to Kiloran Bay passes at first over successively lower beds dipping south-eastwards, and then the dip is reversed and' the same series is repeated in ascending order until the limestone is once more reached. The anticline thus crossed has a north-easterly trend and brings to the surface along its axis the rocks of the Kiloran and Milbuie Groups which imderhe the limestone. From the manner in which the hmestone circles round Kiloran Bay, it is clear that the latter occupies the centre of a synchnal basin. Finally the north end of the island has an anticlinal structure and a mass of gneiss, presumably of Lewisian age, occupies the centre of the fold which has a north-east trend. It has proved impossible to determine whether this^ is part of the old irregular floor on which the Torridon sediments have been laid down, or whether its relation to these sedimients is due to thrusting. It will thus be seen that the structure is, broadly speaking, not at all complex. Over the greater part of the islands we find a single conformable sequence of sediments dipping east and' north-east * See also W. B. Wright : " The Two Earth-Movements of Colonsay," Quart. Journ. Qeol. Soc, 1908, vol. Ixiv. p. 297. 48 Tectonics of Colonsay. at gentle angles, and in the extreme north the upper portion of the same series forming a marked synclinal basin flanked by two anticlines. Complexity of Minor Folding. — The simplicity, however, does not extend to details. The dip is frequently steep, and cases of inversion or even of overfolding can be cited. Thus on Beinn Bhreac to the JS^ Fig. 8. — Section through reonmbent fold of grit with mudstone core, Carn Spiris. south of Balnahard, the flags and grits are locally overturned, and dip away from the syncline of Kiloran Bay at an angle of 80° instead of towards it. A remarkable case of overfolding can be seen on the south side of Carn Spiris, a pro- minent white hill south of the blown sand of the golf links and north of Carn Chaointe. One of the white grit bands belonging to the Machrins grit and mudstone Group is here bent into a recum- bent fold of 40 yds. amplitude (Fig. 8). The same group affords some pretty examples of folding on the shore south of Beinn Eibhinn, while in the north end of the island, on the shore near Cnoc Corr, Balnahard, some grit bands in the phyllites of the Kilchattan Group exhibit the remarkable contortions shown in FiQ. 9. — Diagrammatic sketch of complex Pig. 9. folding of grit bands in phyllite on Faulting and Overthrmtinq.— shore at Cnoc Corr, m the north end of mi i li.- x .t_ r>i i "' i Colonsay. The nearly vertical cleavage ihetaultmg Ot the Lolonsay rocks is that of the second movement. The is of a tjrpe .similar to that usually section is perpendicular to the axes of j^e^ with in the SOUth-wesb High- '"^' lands. One case of overthrust has been observed. This occurs on Sliabh Eiabhach, and the fault can be traced downhill in a W.S.W. direction along a hollow feature till it crosses the road. Its incUnation is steep, and the amount of dis- placement by no means large. Cleavage. 49 Cleavage of the Sediments. — ^Perhaps one of the most noticeable features of the Torridon sediments of these islands is the presence of a well marked cleavage. It is of course best developed in the more pelitic beds, being a true close cleavage of the slaty type, produced by parallel arrangement of the mineral particles of which the rock is composed. Owing to its obvious similarity to the slaty cleavage of other districts, there can be no doubt that it was produced, during a period of movement, in the planes normal to the greatest pressure. In the northern part of Colonsay it preserves a remarkable parallelism to the bedding, but may be seen here and th.ere to cross it on the crest of an isoclinal fold, having been produced at the same time as this fold and along its axial plane. In the south of the islands we find what appears to be the same cleavage cross- ing the bedding over large areas without the accompaniment of any such folding. In its pro- duction, the arrangement of the pre-existing clastic micas in planes perpendicular to the greatest stress has probably gone hand in hand with the development of secondary sericitic mica along these planes. This growth of secondary minerals in a definite series of planes, however close these planes may be, is an incipient type of foliation. We shall, however, in future refer to it as the first cleavage. On the western side of the islands the first cleavage presents plane or only slightly wavy faces, but, eastwards, it becomes progressively more folded and crinkled, and on the east coast has been subjected to intense puckering. At the same time a new cleavage is developed, which, slight at first, becomes towards the east coast fairly well marked, and in the extreme north-east entirely dominates the minor features in the structure of the country. This relation is expressed in. Fig. 10, in which zones of increasing intensity of both foldiag and cleavage areshown bydifferent degreesof shading. Inzone 1 theeSect is trivial, amounting onlytoa slight waviness of thelaminaeof the first cleav- age in the more pelitic beds. In zone 2 a pronounced puckering of these laminae is observable, but little or no cleavage yet accompanies it. Inzone 3 the puckering becomes acute, and the accompanying cleavage well marked, and as zone 4 is approached its planes become so close as almost completely to destroy any facility of splitting along the first cleavage. 4 Fig. 10. — Diagram illustrating by shading the general increase in the intensity of the second movement from west to east, the curved lines being lines of approximately equal movement. Zone 1 is unshaded, zones 2, 3 and 4 are shown with increasing degrees of shading. 50 Tectonics of Colonsay, The relation of the second cleavage to a corresponding system of folds is, as might be expected, far more obvious than in the case of the first cleavage. It preserves, at least in the softer beds, a strict general parallelism with the axial planes of these folds, and un- doubtedly owes its origin to the same compressive forces Q"j!a^3(Sa=^^'*""'"' which produced them. Fig. 11.— Section about 2 ft. long on the coast north W® have, in the two of Port na Ciiiloe, Colonsay, showing a small cleavages of the islands lamprophyre sill (L) and quartz reins (Q) in phyllite. ^nd in the folding ac- fThe quartz veins lie along the planes of the first • j-i, j-u cleavage, and the lamprophyre sill has also been COmpanymg them, tne intruded along the same set of planes. Both the record of twO periods quartz veins and the lamprophyre have been folded Qf g'Jjregg and consequent and cleaved by the second movement.] . j. i • i_ •' movement, of which one is clearly later than the other. How much later remains to be determined. The analogy with the sequence of events described by Mr. C. T. Clough * in the Cowal district of Argyll is remarkable. In both instances there is a primary " foliation " of the rocks, and in both the planes of this foliation are affected by folding and cleavage due to secondary movement. Mr. Clough held that these two movements were distinct, because the planes of secondary cleavage cut not only those of the first cleavage, but also the quartz veins formed along the latter. Where actual cleavage has not been developed, the effect of the later movement can still be recognised in the folding and crinkling of both the early cleavage and the accompanying quartz veins. An exactly similar relation occurs in Colonsay. Quartz veins clearly formed along the first cleavage have been folded and even cleaved during the second movement (see Fig. 11). Still good as is this evi- dence, it is, perhaps, not quite convincing. We have had the good fortune in Colonsay to come upon better. Relation of the Cleavages to the Lamprophyres. — A large number of the lamprophyre dykes so abundant in Colonsay are obviously intruded along the first cleavage, and as obviously folded and / ' _',''/■ ^''-^■. ■*.' ,^ '^ ~ ■* cleaved by the secondary movement. These run in a general way parallel to the strike of the firsb cleav- age, dip with it, and occa- sionally cut across it in such a manner as to leave no doubt that they are subsequent to it. More- over, marginal relations such as that shown in Fig. 12 are not uncommon, and are quite inconsistent with the idea that the primary * " The Geology of Cowal," Mem. Oeol. Survey. 1897. Fig. 12. — Margin of lamprophyre in phylKte, x>n the coast north of Kiloran Bay, showing its relation to the first cleavage, which stops short at the edge of the sill, and does not pene- trate it. Relations of the Cleavages. 51 cleavage could be subsequent to the dykes. The relations of the secondary cleavage are quite as clear. Again and again this cleavage can be traced from the adjoining schists right into the body of the lamprophyre, and repeatedly we find the lamprophyre affected by the folds which contort the planes of primary cleavage along which it liesi Moreover, the relation of distribution of cleaved lamprophyre dykes and sheets throughout the islands is one of remarkable signifi- cance. Their number increases rapidly as we pass from the area of slight to that of intense secondary movement. Thus in zone 1 (Fig. 10) ■we scarcely ever see a lamprophyre showing even incipient cleavage. In zone 2 they are to be found, but are neither strikingly cleaved nor abundant. In zone 3 even the larger and more massive dykes have suffered, and the smaller ones are often remarkably folded and cleaved. In zone 4 they are re- duced to a schistose state, and do not sbrike one at first glance as markedly different from the adjoin- ing phyllites. It will, perhaps, be conceded that we have in the intrusion of this series of dykes a fair in- dication of a very con- siderable lapse of time between the two move- ments. It has been found possible, however, to go even further than this. Relation of the Cleav- uges to the Phdonio p„, i3._Section on the shore south-west of Eilean Dubh, and north of Port ua Cuilce, Colonsay, showing a lamprophyre sill (A) intruded along the primary cleavage in the phyllite and folded and cleaved by the secondary movement. A small dyke (B) of later age than the secondary movement is seen at the bottom of the section. Scale : 1 inch = approximately 2 ft. Masses. — A careful ex- amination of the syenite to the north of Kiloran Bay has thrown consider- able light on the ques- tion. The remarkable breccia occurring along the margins of the plutonic mass (see Fig. 4), and referred to above, contains abundant fragments of the adjoining Torridon sediments, exhibiting clearly the primary cleavage and lying at all angles to one another. They plainly had this cleavage impressed upon them before their inclusion in the breccia, which must therefore have been formed at a period subsequent to the first movement. The syenite, which contains caught-up fragments of the breccia, cannot be earlier than it ; and it is, therefore, quite safe to conclude that both syenite and breccia are distinctly later than the first movement. Evidence of the relation of the syenite and breccia to the secondary movement is also forthcoming. They are traversed by several dykes which have been affected by this movement. The most remarkable of these is to be seen cutting the syenite on Eilean Easdail. It is about 52 Tectonics of Colonsay. 3i ft. wide and lias a general trend about N. 10° E., but takes si decided double bend about the middle of its course. A band about a foot wide on the east side of the dyke {"s rather fine-grained, but the rest is coarser. There are, besides, very fine-grained chilled margin* in contact with the syenite. The secondary cleavage in the dyke is very conspicuous in the coarse-grained portion, and is also well shown by the elongation of the black spots in the chilled margin. It iff throughout vertical or dips at very high angles. Its distinctness from flow-structure is very obvious, since it may be seen crossing the banding of the dyke at a very considerable angle. At the southern end it strikes, in the fine-grained portion N. 20° W., and in the coarse-grained W. 20° N. ; a little north of the bend it is W. 1-5° N., farther north N. 30° W., N. 40° W- and W. 25° N. to N.W. In the adjoining syenite of Eilean Easdail there are marked lines of crushing and. shearing, accompanied bv a rude mineral arrangement with a trend W. 40° N. The great scar of breccia and syenite which forms the landward continuation of Eilean Easdail is crossed by five lamprophyre dykes: Three of these, which are of small breadth, from 6 to 8 in., show the secondary cleavage in a very marked way. Another, which is very massive and about 4 ft. broad, shows no cleavage where it cuts the breccia and syenite, but its obvious continuation can be seen to the south, in the schists, and there it shows a distinct, though not powerful, cleavage. As the syenite is thus cleaved, and is traversed by cleaved lamprophyres, and as the direction of this cleavage corre- sponds fairly well with that produced in the adjoining schists by the secondary movement, it is pretty clear that it must be prior in age to the secondary movement. The breccia contains, as far as could be seen, no fragments obviously affected by the second movement before their inclusion in it. A few of the fragments certainly exhibit crink- ling, but on account of their hornfelsed condition it is difficult to determine which movement this is to be referred to. In most cases the fragments occur under such relations that it is probable the crink- ling was impressed on them while in the breccia. It is important to note that the lamprophyre dykes which traverse the syenite are chilled against it. The syenite must clearly have been cool before their intrusion. We may, therefore, draw the conclusion that, between the first and second movements in the schists, there elapsed sufficient time for — (1) the intrusion of a fair-sized plutonic mass ; (2) the cooling of the same ; (3) the intrusion of an extensive series of lamprophyre dykes. The other plutonic masses of the island are also later than the first cleavage movement with its concomitant introduction of quartz veins, for the latter are often conspicuously altered from white to blue in the baked sediments adjoining the igneous rock. The fragments, too, in the marginal breccia of the Scalasaig diorite were cleaved before they found their way into the breccia, just as in the analogous case of Kiloran Bay. It is extremely interesting there- fore to find a strong cleavage in some of the marginal portions of this diorite continuing out into the neighbouring sediments (Fig. 5). Clearly this is a second cleavage which may with much confidence be Nature of the Second Cleavage. 53 correlated with tliat which has affected the lamprophyre dykes crossing the Kiloran. Bay syenite. The field evidence has failed to throw any hght on the relations of the Balnahard kentallenite to the second movement, though it is, hke the otlass masses, clearly later than the first. No cleavage has been observe' in the body of the rock, and nO dykes have been proved to cut it. The exposures, however, are so limited that negative evidence •of this sort is of little value. An attempt to determine this relation naicroscopically was equally unsuccessful. Beyond a sUght faulting of the garnets developed in the baked sediments, no evidence of move- ment later than the hornfelsing was afforded by the slides. The question cannot therefore be regarded as settled one way or the other, and though the absence of evidence to the contrary might be taken as indicating that the kentallenite was, unlike the other plutonics, later than the second movement, its petrographical affinities to these masses points on the other hand to a common period of intrusion. Finally, it is worthy of note that some of the minor pipes and plugs of syenite lying to the north of Balnahard can be shown to have been, like the larger mass at KUoran, intruded between the two periods of movement. In one case the margin is sharply folded, and a pronounced cleavage parallel to the axial plane of folding can be traced from the sediments right into the igneous mass. The Nature of the Second Cleavage. — The secondary cleavage ia Colonsay is always, at least in the softer beds, strictly parallel to the axial plane in the corresponding folds, and has therefore been un- doubtedly produced in the plane normal to the direction of greatest compression. Emphasis is laid on this very obvious fact, because the ■cleavage is not a slaty cleavage, but a true strain-slip cleavage. The investigations of many eminent geologists, from Sharpe and Haughton down to Heim, Harker and Van Hise, have proved beyond a doubt that slaty cleavage has in all cases been produced in the normal plane. No such confident general assertion can be made with regard to strain- slip cleavage, which, on account of its analogy to thrusting, has been supposed by some to have probably been produced parallel to the planes of easiest shearing. Indeed, it is not at all improbable that structures arising in different ways may be at present grouped under the term " strain-slip," which is, however, generally understood to be ■equivalent to the " ausweichungschvage " of Heim.* In the present case we are dealing with what may be called " alter- nating strain-sUp," by which it is meant to imply that the shearing itnotion along the various planes of gliding has not been everywhere in the same direction. On two adjacent ghding-planes the motions may be equal and opposite, and thus compensate one another ; or on several successive planes the shear may be continuously in the same direction, and then continuously back again on the next succeeding ones. It is exactly analogous to cases figured by Heimf and Harker. J * " Untersuohungen tiber den Meohanismus der Gebirgsbildung," Basel, 1878, vol. ii. p. 54. t Op. cit. Atlas, plate xv. t " On Slaty Cleavage and Allied Rook-Struotures, with special reference to- the Mechanical Theories of Uieir Origin," R-p. Brit. Assoc. (Aberdeen), 188.5, pp. 838, 840. 54 Tectonics of Colonsay. Owing to the compensating action of tlie alternation, there may be on a large scale no differential movement in the direction of the strain- slip cleavage planes. This is one step towards the explanation of the production of the latter in the normal plane. There is a difficulty, pointed out by Harker,* in supposing any gUding or faulting motion whatsoever strictly perpendicular to the direction of greatest pressiire, since there is no component of shear in this direction. Such motion would certainly seem to be theoretically impossible in. a rock made up of lam in as of absolutely uniform strength. The phyUites in which this structure is produced in Colonsay are, however, far from uniform. Strong quartzose bands traverse the phyllites parallel to the first or slaty cleavage, and are buckled iato folds of which the axial planes are parallel to the second cleavage. We may presume that each lamina gave rise under compression to folds whose amplitudes were a direct function of its strength. As the movement progressed, the consider- able displacement of the crest and trough of the folds in the stronger laminae drew and thrust to one side the adjacent phyllite, and adjust- ment took place most easUy by shearing along the limbs of the much smaller folds produced in. the mica laminae of the phyllite. The folding and subsequent strain-slip would seem therefore to be the result of stresses acting on a decidedly heterogeneous rock, in which the weaker laminae folded on a smaller scale, and then adjusted themselves by fold-faults to the form of the stronger laminae. Heim long ago showed how this structure arose ia pelitic bands intercalated between stronger beds of dolomite or grit, the successive strain-slip bands being moved over one another along planes parallel to the axial planes of folding in such a manner as to adjust themselves to the form of the stronger beds — " Die Mikrofaltenverwerfungen sind jedoch nicht an der umgebogenen Stelle am starksten ausgebildet, sie erlangen ihre Hohepunkt an beiden Schenkeln zu der Seite der Umbiegungsstelle, wahrend an der letzteren rnehr nur wellenformiger Sohichtenverlauf vorkommt. Zu beiden Seiten gegen die Sobenkel Tiin belierrschen die Verwerfungslinien das Bild. Die Reste der Schichtung zwischen denselben aind nur bei starker Vergrosserung noch ganz sicher in ihrem S-formig gekriiromten Verlauf zu erkennen, nnd liegen symmetrisch. auf beidem Seiten der Mittellinie der ganzen Umbiegung. Die Verwerfungslinien zeigen naoh beiden Schenkeln bin die Tendenz, dieser Mittellinie parallel sich zu stellen, und schneiden deshalb die Biegungsschenkel unter schiefem Winkel, sie lassen sich mit scharfer Loupe auf frischem Bruche als Rutsehspiegel erkennen. Diese Faltenverwer- fungslinien sind Plachen starker Diflerentialbewegungen, welchem entlang durch Verschiebungen das plastiche Material von den Mittelschenkeln, wo es gequetscht wurde, gegen die Umbiegungsstellen auswlch, und dort hinstromend sich anhaufte." + It is worth remarking that similar straia-shp can be observed in connection with the slaty cleavage of the first movement, the glid- ing planes being coincident with the cleavage. Mr. Clough % has figured a case of this associated with the analogous slaty cleavage of Cowal. An interesting result was obtained by mapping the planes of secondary cleavage, or, where these were not well enough developed * " On Slaty Cleavage and Allied Rock-Structures, with special reference to the Mechanical Theories of their Origin," Re-p. Brit. Assoc. (Aberdeen), 1885, pp. 838, 840. t " Untersuchungen iiber den Mechanismus der Gebirgsbildimg," 1878, vol. ii. pp. 52-53. X " The Geology of Cowal," Mem. Geol. Survey, 1897, p. 12, Fig. X Trsnd of the Second Cleavage. 55 to be readily observable, the axial planes of secondary folding. The result is shown in Fig. 14. The curved lines indicate the strike of the planes, and their dip is shown by the arrows. The constancy of the strike and dip in some areas contrasts with their irregularity in others, where the condition of affairs cannot be adequately expressed on a COIiONS.«Y- Strike of the most remarkable physio- graphical features of Colonsay and Oronsay is an old plain of marine erosion, which can be traced throughout the greater part of the islands. It has suffered much from subaerial and glacial denudation, and large areas of it have no doubt been removed, at various stages, by the sea. Nevertheless it is possible in many places to trace its inner margin, which is frequently marked by a line of lofty chffs. The denudation to which the rocks forming the platform have been subjected must have been largely effected in pre-glacial times. There is considerable difficulty, however, in discriminating between irregularities which pre-existed the platform and those due to subse- quent erosion, so that no certain conclusion can be drawn as to the lapse of time between the elevation of the platform beyond the reach of the waves and the advent of the land-ice. The post-glacial atmospheric waste has obviously been slight. Everywhere one finds the surface of the plain ice-moulded, and in many places it is striated and covered by boulder clay ; the old line of cliff also fre- quently shows marks of the passage of the ice, and one cannot but be struck by the sHght modification which these much exposed glaciated surfaces have undergone. This remarkable rock shelf and its accompanying cliffs are best preserved in Uragaig to the west of Kiloran Bay (see Fig. 17). The platform is here nearly haK a mile wide, and probably onc« extended much farther seaward. Its inner angle beneath the magnificent cliffs of Tormach Mor is about 135 ft. above high-water mark of ordinary spring tides. It can be traced from here round the west coast for about a mile and a half to Port Ban, forming along this coast the grassy platforms between the upper and lower cliffs blown as Aonan Ceann a' Gharraidh, Aonan nam Bo, Aonan Mhic Mhuirtean and Aonan nam Muc* South of the last-named locality no sure trace of it can be found along the cliffs as far as the south side of Binnein Riabhach. Here a great strip of the platform can be seen stretching away inland towards Upper Kilchattan, although it is much denuded and its inner margin is ill defined. South of Port Mor, however, the * Aonan means a grassy pasture open to tlie sea, and surrounded by rooky cliffs. It does not strictly imply a step between the higher and lower cliffs," as stated by Mr. Symington Grieve, although it often is so. A small aonan, the Aonan nam Clach Mora, north of the Cailleaoh Uragaig, is a grassy flat of beach gravel surrounded by cliffs and lying only a httle above high-water mark. A former proprietor of the imi at Scalasaig, Mr. Donald M'Neill, being worried by his guests as to the meaning of Aonan nam Muc, translated it for their benefit as The Pig's Paradise, and by this name it has been commonly known since. To Mr. Grieve belongs the credit of having first recognised that some of the higher aonans along the west coast owe their origin to marine erosion ; he does not, however, appear to have noticed that they are of pre-glacial age. See Proc. 8oc. Ant. Scot., 1882-83, vol. v. p. 351. 62 Pre-Cflacial Shore Platform. 63 old coast-line is marked by a fine clifE forming the western faces of Beinn nan Caorach and Sliabh Riabhach. The pass to the south of the last-mentioned hill was a strait at the time of the formation of /~\",-A Fig. 17. — Uragaig from the north side of Kiloran Bay, showing the cliff and rock- platform of the 135-foot pre-glacial raised beach. the old shore-hne, and the platform can be traced right through it to Scalasaig along the north side of the road. The higher parts of the southern half of Colonsay, with Beinn Oronsay, formed at this period outstanding islands with a series of narrow straits between (see Fig. 19), The platform of erosion among these islands is in a few places still well preserved, as on the plateau known as L6n Mor to the south-west of Loch Staosnaig. A distinct shelf can also be traced round Beinn Eibhinn, and it is again particu- larly well marked on the south side of Beirm Oronsay. Some outlying hills at the north of Colonsay were also isolated, while the big inland cliffs south of Babiahard marked, at this stage, the coast of the main island. On the east side of the islands the effect of full exposure to the impact of the ice is noticeable in the planing down of the platform a.nd the rounding off of the old cliff feature. This is, of course, especially marked in the softer, more pelitic beds, but where the rock Fis. 18. — The shores of Colonsay looking north from Port Olmsa (a mile and a half N.E. of Scalasaig), showing the pre-glacial cliff and rock-platform and the ice-planed slopes of the east coast. ehelf is cut in the harder grits, as in the case of some of the headlands north of Scalasaig, both cliff and platform are sometimes well pre- served (see Fig. 18). 64 Raised Beaches of Colonsay. In addition to this well marked high-level platform, there exista another, a little above sea-level, which may, perhaps, also be taken as indicating a pre-glacial plain of marine denudation. This lower platform includes the ice-worn skerries, many of them covered at high tide, which render the western coast so dangerous to navigation, and which, along the shores of Ardskenish and Oronsay, are the favourite haunts of seal and cormorant. The margin is nowhere COLONSAY *BALNAHA/iD URABAIGj AONAN CEANNA'GHARFtAIDH AONANNAMMUC {PORTOLUSA PORT MOR^ RUDH'AIROALANAi DUU 6ALLAIN ARDSHENIl ORONSAY UNL OF PRE.GLACIAL SEA-CUFF Oo' PftEGLACIAL PLATFORM O OF MARINE EROSION Fig. ,19. — Sketch-map showing the development of the 13,5-foot pre- glacial cliff and shore-platform in Colonsay and Oronsay, very definite, so that doubt may well exist as to the interpretation of the origin of the platform here suggested. A fairly good margin may, however, be noted at Dun Gallain and also behind Rudh' Ard Alanais to the north and Rudh' an Dunain farther south, as well as occasion- ally in Oronsay. It is not impossible that the fine seacliffs along the western shores of Colonsay to the north of Kilchattan may, in the main, have been fashioned \ij the pre-glacial sea, and only modified in detail by subsequent denudation. Post-Glacial Raised Beaches. 65 Post-Glacial Raised Beaches. — The post-glacial raised beaches of Colonsay and Oronsay are on the whole not at all strikingly developed. With the exception of some huge shingle banks and a little irregular erosion we find throughout the islands few traces of the old shore- lines. Even the most recent of these, the Neolithic or 25-foot beach, which forms such a well marked rock shelf on the islands in the Firth of Lome, has here only a slight development. The smooth ice-planed slopes which rise from the sea on the east side of Colonsay have little to show for this prolonged period of higher sea-level but a slight corrosion and breaking-up of their surfaces. On the west, where the Atlantic waves have exerted their full force, we find at this level only a little sand and gravel among irregular knobs of rock, with here and there a well marked storm-beach. The great banks of shingle at Lower Kilchattan are among the finest on the island. We can trace in them the shore-lines of the well known 100-foot, 50-foot and 25-foot beaches of Islay, Jura and the mainland. The finer stratified gravel of the 100-foot beach, which came within reach of the ordinary scour of the waves, can be observed up to a height of 95 ft. above high-water mark of spring tides, while Fig. 20.- -Section of the Kilchattan storm-beaches. Horizontal scale : 6 inches = 1 mile. Vertical scale twice horizontal. the storm-beaches of this period reach a level of 109 ft. The inner angle of the ordinary shore-terrace of the 50-foot beach lies at 60 ft. above the same datum, while its storm-beaches attain a height of 87 ft., and the shingle in part probably lies on the outer edge of the 100-foot shore. Owing to the different configuration of the adjoining coast when the sea stood at the 25-foot level, no storm-beaches were formed at this period. There is, however, a marked shelf of shelly sand and fine gravel with its inner angle at a height of 26 ft. above high-water mark of spring tides. Finally, a comparatively recent accumulation of shelly sand reaches a heighb of 10 or 12 ft.* The relations of the various beaches are shown in Fig. 20. After the abandonment of the country by the land-ice a deep valley stretched along the site of Loch Fada to Port Mor. This either immediately or at a later stage became occupied by the sea. The upper reaches of this valley were then separated off, for shingle spits grew out from the sides until they formed a continuous barrier right * The lowest point of the road where it passes the head of Port Mor is about 5 ft. above high-water mark of spring tides. In storms the sea occasionally washes over it. In the great storm of 1880 the waves swept over the road right into the margin of the 12-foot shelf, depositing a huge rampart of seaweed and shingle across the fields. The low-lying cottages by the roadside at this point were flooded, and the inhabitants had to escape to the rooks above. 5 66 Raised Beaches of Colonsay. across. The supply of shingle at this stage must have been consider- able, as the sea continually attacked fresh ground, more or less covered with drift. A strand of fine stratified gravel was formed near the 100- foot level, and the highest parts of the storm-beach thrown up above high-water mark. During the elevation of the land a pause evidently occurred when the sea stood at the 60-foot level, and another well defined strand with its accompanying storm-beach was developed. The 25-foot shore-line does not mark a further stage in the elevation, but is the record of a much later period of depression. No direct proof of this was forthcoming in Colonsay, but on the adjoining coasts peat deposits have been found beneath the gravels of this beach, showing that its formation was preceded by a period of higher level. The 100-foot raised beach is the oldest post-glacial beach, being presumably of late glacial age. Except at Kilchattan (see description above) it is not remarkably developed anywhere on the islands. Traces of it can be detected in the slopes of waterworn gravel banked against the rock slopes to the south of Kiloran House, the whole of the wide meadow in front of the house being at that time under the sea. Again, the narrow valley which runs across the island from Scalasaig to Machrins contains gravel up to more than 90 ft. above high-water mark of spring tides. At Turnicil there is gravel at about 100 ft. above the same datum, and on Lon Mor to the south of Balarumindubh there is a bar of gravel overgrown by heather enclosing a small marsh and lying at much the same level. Gravel and sand occiir among rocks up to 90 ft. at Garvard. The 50-foot raised beach is much more remarkably developed, and is, in fact, the most conspicuous beach in the islands. This is largely due to the quantities of shingle associated with it, the amount being generally in excess of that found along the later 25-foot beach. In addition to Kilchattan it is found developed in the following localities : — Balnahard. — The great gravel spit which connects Meall na Suiridhe with the high ground south of the Bahiahard valley was probably formed at this period. Its summit is 60 ft. above high-water mark of spring tides, but there is a re- markable shelf on both the northern and southern sides of it which has its inner angle at a height of 47 or 48 ft. above the same datum. XJragaig. — The gravel shelf occupying the hollow which stretches from Port nam Fhuchan to Port an Tighe Mhoir is probably composed of storm shingle of the 50-foot beach. Its upper surface is about 77 ft. above high-water mark of spring tides. As the modern storm shingle in Port nam Fhuchan reaches a height of 20 ft., and that of the 25 foot beach a height of 60 ft., the storm-shingle of 50-foot age may weU have reached a height of 77 ft. A little further east in Uragaig, close to the southern end of Traigh BhJin and immediately west of Uamh Ur (the New Cave), there is a well marked grassy shelf with its inner angle about 60 ft. above high-water mark of spring tides. The geos rotmd about this shelf show smoothing by wave action up to 73 ft. or so. On the modem shore wave-pohshing reaches 13 ft., and, comparing, we get 60 ft. elevation, speaking roughly, for the beach, which agrees fairly well with the height of the inner angle above spring tide. KilchaUan.—llhe inner angle of the 50-foot shore-line on the Kilchattan shingle banks is 60 ft. above high-water mark of spring tides (see description and figure above). This shore-hne can be traced for some distance both to the north and south of Lower Kilchattan, but to the north it is difficult to follow, the shingle having been washed up into irregular geos to a height of 70 ft. To the south of Lower Kilchattan the road runs along the gravelly shelf of the 50-foot beach, which hes in a hollow in the pre-glaoial platform. The gravel shelf, which abuts mland against a steep chff, lies at 55 ft. The modem shore Post-Glacial Raised Beaches. 67 here is rooky, with reef -sheltered bays having sandy bottoms 10 or 15 ft. below high-water mark. As we pass south along the road to its bend west of Beinn nan Caoraoh the inner edge of the gravel in the hollow drops to a level a little over 50 ft. It is possible that the whole of the sand and gravel in this slack may have been formed well below high-water mark, like the sand in the bottom of the modem bays. Just at this point there are some magnificent modern shingle banks. Machrins. — The blown sand of the golf links obscures the greater part of the raised beach beyond this point, but to the north of the road the ground is clear, and the shingle of the 50-foot beach extends among the rock knobs to eleva- tions of 60 or 70 ft. Machrins itself is beyond the limits of the blown sand, and the gravel may be traced in horseshoe form round the swampy hollow occupying the heart of the valley. When the com is cut, the sweeping curve of these yellow stubble-covered fields recalls irresistibly the contour of a modem beach laid bare by the retreating tide. Oarvard. — At the farm of Garvard there is a fine gravel shelf and spit stretching south to some rock knobs which must have been islands at the time of the 50-foot beach. The surface of this gravel lies from 65 to 70 ft. above high-water mark of spring tides. Loch Staosnaig. — At the head of Loch Staosnaig the gravel of the 50-foot beach, which hes at 55 to 60 ft. above high-water mark of spring tides, has dammed up a Uttle stream, and the latter has since out a gorge for itself partly through the rock. Scalasaig. — About 150 yds. S.W. of the Manse there is a fairly good shelf of deposition with its inner angle just at the 50-foot level. The best development, however, occurs east of the Church and to the south of the Inn, in both of which localities the gravel shelf is just 60 ft. above high- water mark of spring tides. East Coast North of Scalasaig. — Traces of the raised beaches along this part of the coast are exceedingly scanty. At Glas Ard gravel occurs in the geos up to 68 or 70 ft., and in Sloe nan Sgarbh immediately south of the Rudha Carraig nan Darrach there is a small ledge of shingle overlying tumbled blocks such as occur on the present shore. The inner edge of the shingle ledge is about 60 ft. ; but the measurement was difficult and is only an approximation. A glance over the above details will show how hopeless is any attempt to estimate accurately the difference of level of the " 50-foot " and present seas. We never find, either in the case of the raised beach or the modern shore, anything of the nature of a regular rock shelf which might serve as a basis for comparison of levels. We have to rely wholly upon sand, gravel and shingle deposits, and these may be formed well below high-water mark or well above it. They are thus very unreliable, and great care is needed in drawing conclusions from them. The evidence on the whole seems, however, to point to a fairly prolonged pause of the sea at a level about 60 ft. higher than the present. No confidence, however, can be felt that there are not gravels on Colonsay which were formed at a time when the sea stood at levels intermediate between the 100-foot and 60-foot levels, and it is not at all impossible that there may have been a halt at a slightly lower level of about 45 to 50 ft. This difficulty of correlation is well exemplified in the case of the terraces to the south-east of Cuirn Mhora on the northern side of the Strand. A series of shelves, apparently of gravel and sand, can be seen here, one above the other, at heights of 4, 12, 23, 36 and 47 ft. re- spectively, and, close by, in another embayment they lie at 5, 15, 30-33, 50 and 68 ft. The lower terraces (at 4 ft. and 5 ft.) are modern accumula- tions being added to at the present day; the present sand-fiat over which the tide rises and falls is 2 or 3 ft. below high-water mark of spring tide. The 12-foot and 15-foot terraces probably correspond. The 23-foot terrace in the first case is that of the "25-foot" beach, The corresponding sheH is hard to locate in the other case ; the 30-33- foot terrace is very ill defined. The 47- and 50-foot terraces may have 68 Raised Beaches of Colonsay. been formed at a period wlieii, as suggested above, the sea stood at a level 45 to 50 ft. higher than at present ; they may, in fact, correspond to the'marked shelf in the flanks of the Balnahard shiagle spit. In the first- mentioned of the Cuirn Mhora series of terraces there is a grassy spit at 65 ft. which may correspond to the 68-foot terrace in the second series. The above-cited cases wiU give some conception of the difficulty of forming any clear idea from the Colonsay evidence alone regarding the oscillations of sea-level, at any rate during the formation of the ear Her and higher beaches. The poor development and extreme ruggedness and unevenness of the rock-platform of the ^5-foot raised beach round Colonsay is a direct result of the lowness of the ground into which the sea was in most places advancing during the period of its formation. With little or no clifE to undercut, the advance of the shore-line was in places very rapid. On a coast of this kind when the waves have once left behind an outlying stack or island, they lose, from that on, their power of destroying it subsequently, for all the eroding material in the form of sand and shingle is at the extreme limit of advance, in the heads of the geos that are being out. Thus a rapidly advancing sea yields an immature plane of erosion, while the necessity of under- cutting a lofty clifi guarantees the removal of all irregularities. The pre-glacial platform of Uragaig exhibits, for example, a degree of evenness of its surface which has not been approached by any later plane of erosion now visible. It must not be forgotten, however, that the time required for its erosion was considerably greater than that during which the sea stood at the 25-foot or any other level which it occupied during post-glacial times. The estimation of the height of the 25-foot shore in Colonsay is, therefore, a matter of some difficulty. As it has not been found possible to state with confidence any very exact figure, it will be necessary to consider briefly the various localities in which it has been possible to make measurements at all. Balnahard. — The sand of the 25-foot beaoh occupies a considerable area at Balnahard. The relations of its storm-banks to the shingle spit and terraces of the 50-foot beach are shown diagrammatically in plan and section in Kg. 21, where the levels of the various ledges are given. On the western or Port Sgiobinish side, ledges of ungrassed shingle occur up to 27 ft. above high- water mark of spring tides, and grassed shingle to 32 ft., while on the eastern side adjoining Port an Obain there is a grassed shelf at 35 ft., which, however, is very obscure. At this side also we appear to have the upper limit of the waves during ordinary tides indicated by a notch at 25 ft. above high-water mark of spring tides. On the west side it is covered by later storm-beaches. One of these, that at 27 ft., may represent a pause during the retreat of the sea from the 25-foot level. It would correspond to a sea-level about 16 ft. above the present. Kiloran Bay. — Shingle spits of 25-foot age can be seen crossmg Kiloran Meadow south of the sandhills. The meadow was evidently formed as a marsh behind them. Uragaig.— Immedi&tely east of Dun Taltaig there are four ledges of storm- shingle one behind the other at levels of 3, 10, 17, and 35 ft. above high-water mark of spring tides. The stones of the three first are very clean and fresh, and obviously relatively recent. Those of the 35-foot ledge are weathered and lichen- covered, partly overgrown with grass and strewn with scree-blocks from the clifiEs. Port nan Fliuchan.— At the head of Port nan FUuchan the recent ordinaiy tidal shelf of gravel only extends about half a foot above high-water mark of spring tides. Above this are recent shingle banks reaching levels of 16 and 20 ft. The inner angle of the 25-foot sheK, on to the edge of which the recent Post-Glacial Raised Beaches. 69 shingle has been thrown, varies from 24 to 25 ft. The storm -shingle of the 25-foot beach forms a terrace still further inland, attaining at its outer margin a height of 45 ft. and at its inner 50 to 52 ft. On the whole the deposits here seem to indicate a difference of about 23 to 25 ft. between the levels of the sea at present and during 25-foot beach times. Port 2ffjr, Kilchattan. — The 25-foot shelf which is partly cut in the older shingle beaches at Port Mor is shown in Kg. 20. Its inner angle is about 26 ft. above high-water mark of spring tides, and probably did not he far above the spring-tide level at the time of its formation. The 25-foot margin is traceable for a short distance along the northern shore of Port Mor. Ardskenish. — Ungrassed shingle banks of 25-foot beach age are well developed on Ardskenish Point. They occur at heights of 38 to 42 ft. above high-water mark, but numerous later ledges and banks occur at all levels down to 12 ft. The recent sea is now throwing up shelly sand and gravel to a height of 10 ft. The following liae of levels taken over a series of these shingle beaches may be regarded as fairly representative : — High-water mark (approximately only) . . . . feet. 1st patch of ungrassed shingle .. .. .. 11-12 2nd ledge „ „ „ . . . . . . 14 3rd „ „ „ „ 20 Top of 1st big shingle beach, ungrassed . . . . 28 „ 2nd „ „ „ „ . . . . 30-33 „ 3rd or main „ „ .. .. 40 42 Northern Shore of the Strand. — The 25-foot beach seems to be represented in the series of terraces to the south-east of Cuim Mhora (see above) by a shelf at a height of 23 ft. above high-water mark of spring tides, and at Garvard the inner angle of a well - marked shelf of deposition Ues at just about the same height. In both these locaUties there is a terrace of later age, having at Ouim Mhora a height of 12 to 15 ft. and at Garvard of about 10 or 12 ft. Oronsay. — A large part of the surface of Oronsay is composed of the gravel and sand of the 25-foot beach. On Rudha Caol, the extreme southerly point of the island, there are some vmgrassed storm-beaches of this age forming graceful cuspidate curves some distance in from the coast. Shingle banks are also conspicuous on Eilean Ghurdmail and Eilean nan Ron. Sccdasaig. — Gravel deposits of the 25-foot beach occur fringing the boggy flat which lies north of the road between Scalasaig and the pier. To the south of the road they attain a height of 35 ft. above high-water mark of spring tides. The gravelly slopes at Glas Ard to the north of Scalasaig show a very indefinite sort of shelf at 25 ft., and there is the usual lower shelf at 13 to 15 ft. above high-water mark of spring tides. SOFOOT UNGflASSEO BEACH. SANDtrSHIHeLe. Fig. 21.— Diagrammatic plan and section of the raised beaches of Balnahard. Horizontal scale : 3 inches = 1 mile. Vertical scale three times horizontal. The shelf marked 35 ft. on the north- east side is ill-defined. The above details embrace practically all the evidence available for the determination of the sea-level during the formation of the 25-foot beach in Colonsay and Oronsay. It is, to say the least of it, unsatisfactory. Even an approximate estimate from such evidence has elements of doubt about it, and it is hard to state Hmits for the possibility of error. AU the observed facts are, however, explainable 70 Post-Glacial Deposits of Colonsay. on the supposition of a difference in level of tlie 25-foot and present seas of 23 ft., and the figure is probably not more than 25 ft. nor lesa than 2] ft. Any difference in level from place to place due to a gradient such as can be observed on the adjoining coasts of Scotland would not amount to more than a foot within the limits of the islands, and this is quite inappreciable with the measurements at our disposal. The ledges of gravel and sand mentioned above as occurring at several locahties at a height of 12 or 15 ft. above high-water mark of spring tides, and the corresponding storm-beaches at a somewhat higher level, may represent a pause in the late Neohthic uphft, but the Colonsay cases are hardly sufficient to stand by themselves. If, however, the occurrence of such a " recessional beach " is estabhshed. on the mainland of Scotland and on the adjoining islands, then these must have the same interpretation. Neolithic Remains, Caves and Cave Deposits. — In the sandhills north- east of Balnahard, there is a well-marked Neolithic floor at a level of 22 to 24 ft. above high-water mark of spring tides. It has yielded abundant flint implements, including scrapers, small knife-hke flakes struck from a core and a well formed arrow-head. A few of the flakes show obvious secondary working along the edges, and may have been intended for heads of arrows or fish spears. The accumulation is obviously Neohthic, but to what stage of the Neolithic culture it belongs it is hard to say. It is noteworthy that these implements do not occur in the sandhills at a lower level than 22 ft. above high-water mark of spring tides, a relation which suggests their association with the 25-foot shore. It is not clear, however, whether they are contemporaneous with the maximum submergence or a little later than it. Recovery, at any rate, had probably not proceeded very far at the time of their manufacture. A remarkable feature of the 25-foot or Neolithic shore-Hne in Colonsay is the number and size of the caves that were formed along it. The distance to which these penetrate, their spaciousness and the smoothing and polishing of their walls testify to the long-continued action of the sea. Their remarkable development is in striking contrast to the otherwiss trifling signs of erosion at this level throughout the island. Even taken alone, however, they show that if there was a prolonged pause in post-glacial times at any shore-hne now above sea-level it was at the 25-foot. Few such caves have been developed along the present sea margin, and none are found in association with any of the higher beaches.* The deposits of the New Cave (Uamh Ur) on the south side of * The following list of the more remarkable caves may be of use to visitors, who are, however, requested to refrain from tampering with the floor deposits. Alljthe finer examples occur on the north-west coast. The Port Easdail Caves. — These caves, four in number, are situated at the north end of Kiloran strand. Three of them are at a very low level, the modern shingle being washed into them by storms. The fourth, the Widow's Cave, has its floot about 17 ft. above the modern shingle. Two of the low-lying caves are very extensive. The New Cave, or Vamh Pr.— This is Mr. Symington Grieve's " Crystal Spring Cavern." It is an extensive and intricate cave with numerous chambers, and is situated on the coast about 200 yds. west of the mouth of Kiloran Burn. Two other Caves to the South-west of Kiloran Bay. — These lie on the same coast within a quarter of a mile of the New Cave. The Bonny Care, and another on the coast north of Dun Taltaig. Caves. 71 Kiloran Ba.y have been studied by Mr. Symington Grieve.* In the fossiliferous strata which locally cover the shingle on the floor of the cave he found bones of ox (Bos longifrons), sheep (a small variety), horse, pig and red deer, as well as a broken bone needle and other signs of human occupation. A more abundant fauna was obtained by Mr. Grieve and the late WilHam Galloway f from the shell mound of Caisteal nan Gillean in Oronsay. The relation of this deposit to the raised beach is, however, not so obvious, but a certain interest attaches to a comparison of the remains with those of the New Cave. The mound yielded bones of the red deer, marten, otter, sheep, rat, rabbit, pig, common seal and grey seal, as well as of the great auk {Alca impennis, L.). Associated with these were bone harpoon-heads, bone awls, elongated stones and bones rubbed at one end so as to give a chisel edge, and other rude implements. Mr. Grieve draws the conclusion that the occupation of the cave was more recent than that of the mound. Thus the red deer, which is plentiful in the lower strata of the mound and persists throughout it, is represented only among the early deposits of the cave, and even there by a single fragment of an antler. The pig appears only in the lower deposits of the cave. The sheep, represented by one doubtful bone near the surface of the mound, occurs aU through the cave deposits, while the ox &st shows evidence of its presence about the middle of the latter. The horse is only found in the very uppermost deposits. The chief interest attaching to these archaeological researches is the proof they afford us that the earUest inhabitants, those of Caisteal nan Gillean, were a hunting and fishing folk, and that later, at the period of occupation of the cave, they had either been replaced by or had developed into a pastoral race possessing at first only sheep but at a later period cattle and horses as well. We learn also from these investigations that the mammalian fauna of the islands was at one time not nearly so restricted as it is at present, and we cannot but realise that any deduction from its meagreness regarding the absence of posb-glacial connection with the mainland is entirely unworthy of confidence. The caves formed by the sea during NeoUthic times in Colonsay are interesting, and in some cases command attention by their spaciousness and intricate development. It is, however, merely to court disappointment to compare them, as Mr. Grieve has done, with those of Stafia. Their archaeological interest also, though undeniable, is restricted, for it must be remembered that they were occupied by the sea during the earher phases of Neohthic culture, and are thus very unlikely to contain any traces of those phases. Much more promising as subjects for excavation are a couple of smaller caves at a higher level, namely the Widow's Cave north of The Caves of Gailleach TJragaig. — The finest of these is in alittle bay west of Aonau nam Claoh Mora. The Meals Cave, below Pigs' Paradise, north of Port Ban. The Piter's Cave, south of Port Ban. According to tradition the devil used to live here, and the usual piper and dog story is told about the cave. * " Preliminary Note on the Discovery of a Bone Cave at the Island of Colonsay," Proc. Soc. Ant. Scot, 1879-80, p. 318, and " On the Crystal Spring Cavern, Colonsay (Second Communication)," ibid. 1882-3, p. 351. •(• " Notice of the Discovery of Remains of the Great Auk or Garefowl [Alca im- pennis, L.) on the Island of Oronsay, Argyllshire," Journ. of the Linnean Society, Zoology, 1883, vol. xvi. p. 479. 72 Post-Glacial Deposits of Golonsay. Port Easdail, Kiloran Bay, and especially the small cave known as the Goat's Cave on the southern face of Beinn Bibhinn. The latter is little more than a rock-shelter in the cliffs, but it has the appearance of having a considerable thickness of floor deposits. Submerged Forest of Loch Fada. — Several stone axe-heads, proving beyond a doubt the presence of Neolithic man, have been found in Colonsay, and are preserved in the National Museum of Antiquities of Scotland. One of these, labelled AF. 173, was found on the surface of a remarkable bed of peat which occurs in many places round Loch Fada. Huge stumps of trees, many in the position of growth, are imbedded in this peat to a depth of at least 4 ft. below the surface of the lake. The cause of the submergence of these trees is a question of some interest. The stumps are obviously in place, and some large tree trunks are lying about, one being about 18 in. in diameter. No trees of this size exist in the valley at the present day, with the exception of the artificiaUy protected plantations and woods at Kiloran.* It is fairly certain that the cUmatic conditions at the time of their growth were more favourable to the existence of forests than those of the present day, and it is probable that they are to be referred to one of the " Forestian " periods of Professor Geikie. Although covered by its waters it is unlikely that they antedate the actual production of the loch, which appears to have come into exist- ence as a result of the damming action of the great shingle banks of the 100-foot beach at Kilchattan. This latter being .almost certainly of late glacial age, it is unhkely, though certainly not impossible, that trees of this size could have occupied the country previous to its formation. It is much more probable that the submergence has been merely caused by a later transgression of the lake due to the raising of its outlet, and since the outlet is through a peat moss we find in the growth of the peat the easiest explanation of such a rise. * Samples of the wood of some of these submerged stumps were sent by Mr. Symington Grieve to Professor J. M. Macfarlane, who identified it as that of the Goat Willow {Saiix caprea, L. ). It is probable, however, that other trees are also represented. CHAPTER X. Economic Resoubces of Colonsay and Oeonsay. Economically considered, the geology of Colonsay presents few features of any interest. The rocks show no promise of mineral wealth, and the richness of the island is entirely agricultural. It bears, however, among the Western Isles a considerable reputation for its fertility, which is the more astonishing iu view of its rugged and forbidding aspect. Its barren chfis and crags serve, however, as shelter to the fertile valleys between, and afford protection to the sheep and cattle that are reared among them. As Pennant says, Colonsay is "full of rocky hiUs, running transversely, with variety of pretty meandering vales full of grass, and most excellent for pasturage." * The island was at one time famous for its breed of Highland cattle, but at the present day a considerable area is devoted to the raising of Ayrshires and Ayrshire crosses. The richness of the hUl pasture is to a large extent due to the widespread distribution by the wind of the shell sand that is continually being thrown up on the western shores of the islands. This is especially noticeable in the north end of Colonsay, where the hiUs are clothed with a grassy turf, and there is an almost complete absence of the heather which aboimds ia locaUtiea which are not reached by the sand. The main distributing reservoirs of this invaluable natural fertihser are the sandhills of Balnahard, Kiloran, Dun Gallain, Ardskenish and Oronsay. A sample of Kiloran sand digested in hydrochloric acid proved to contain about 55 to 60 per cent., while one from Ardskenish contained as much as 87 per cent, by weight of shell debris. The blown sands on the south-east coast of Colonsay present a remarkable contrast to these, in that they are largely, If not entirely, made up of quartz grains and other inorganic material. The debris from the phylUte crags north of Kilchattan produces a soil which also supports a grassy sward. These rocks weather to an impalpable powder having a greasy feel, as a result of the quantity of sericitic mica it contains. This material forms between the rock- scars talus-like slopes, which soon become clothed with a green turf, and the combination of this pasturage with the abundant natural shelter afiorded by the crags enables sheep to be raised on the hills without much attention. The sand and gravel deposits of the raised beaches, which cover considerable areas in some parts of the island, form an excellent light soil easy to till and requiring little drainage in comparison with the wet heavy boulder clay slopes of the interior valleys. Corn win ripen on these hght soils in years when it wiU remain green right up to shearing-time on the heavier clay land. The meadow in front of Kiloran House has a soil mainly composed of peaty alluvium, * Thomas Pennant, "A Tour in Scotland and Voyage to the Hebrides ; 1772," London, 1790, part i., p. 274. 73 74 Economic Resources of Colonsay. deposited behind the gravel spits of the 25-foot beach. It had at one time the reputation of being the finest meadow in Argyll. Very little quarrying is now carried on in the island. Small open- ings in the flags and grits are worked here and there for road metal, but the boulders of the raised beaches are very often used, as near Kilchattan, for this purpose. A small quarry in the felspathic grit was opened in the grounds of Kiloran House during the building of the recent additions. The beautifully banded semi-pehtic rocks of Kiloran and Uragaig split in many places easily along the bedding, and would make ex- cellent flags, were it not that they possess an inconvenient obUque cross-jointing which would render squaring a matter of difficulty. The demand for natural flags is, moreover, not so great as it was formerly. The dry-stone dykes of Colonsay, built largely of the flat slabs of this rock, have quite a reputation for excellence. In the south of Colonsay and in Oronsay the mudstones are used instead of the flags for wall-building. The kentallenite of Balnahard closely resembles the well-known " black granite " of Ballaohulish, but it is imlikely that it could ever compete successfully with a quarry so suitably situated- as that of Kentallen. The diorite of Scalasaig was employed for the construction of the harbour at that place, but, as it proved rather soft, granite from the Ross of Mull was used as facing for the pier. The syenite of Kiloran Bay is a rather handsome stone, but owing to its situation is quite unsuited for exploitation. The quartzose breccia adjoining the last-mentioned rock contains large quantities of pyrites, and the presence of gold being possible, specimens were submitted for assaying, but without any promising result. Concerning these, Dr. Pollard says : " I have examined the above specimens, and beg to report traces of gold probably not ex- ceeding 10 grains per long ton." PART II.— SOUTH PART OF THE ROSS OF MULL, AND THE ADJACENT ISLANDS. By E. H. Ctjkningham Cbaio. CHAPTEE XL Introduction. geographical position and character. The portion of the long south-western promontory of Mull, called the Ross * of Mull, which comes into the northern part of one-inch map 35, is in most places less than a mile broad, and is probably less than 9 square miles in extent. The coast-hne is, however, of considerable length, being repeatedly indented in the eastern portion by narrow bays, running N.N.E. parallel to the strike of the adjacent rocks. Among the neighbouring islands we may specially mention Soa, EUean a' Chalmain and the Torran Rocks. The latter are all bare of vegetation and wave-swept at every storm. They extend from the Ross in a south-westerly direction for about 5 miles. About 10 miles farther in the same direction is the rock- island of Dubh Artach with its lonely Lighthouse. The population is extremely scanty, being confined to the farm of Scoor and the seaside hamlets of Ardalanish and Uisken. But during the summer, particularly in the Glasgow holidays, a few visitors come to enjoy the sea breezes and the fine sandy bays which are a special feature of this coast. To a considerable number of tourists the sea view of this part of Mull is no doubt well known, for comfortable passenger steamers have, for many summers past, made daily trips round the island ; the ship's course hes near the south coast of the Ross, and the geological characters of some of the clifis are recognisable. The district is rough and craggy, but no part much exceeds 400 ft. in height. A striking feature of the whole of the Ross when seen from the sea at a httle distance, whether from the south or the north, is the uniform level attained by the hills. Most of them rise to about 400 ft. but no higher, so that a distant view gives the im- pression of an even plain. This plain stretches westward over lona,']' but in an easterly or north-easterly direction it ends rather abruptly a little north of the north-east corner of one-inch map 35, and gives place to terraced hills which rise to considerable heights and form a great contrast in geological structure to the ground to the west. The cHmate is often stormy, especially during the winter, but the temperature does not sink low at that season, and snow does not * The Gaelic word " Ros " is said to signify a projecting point or promontory. t ■'^ corresponding plain is also recognisable in Colonsay, see Chapter T. 76 76 Introduction : Boss of Mull. lie long. The average rainfall is less than in the more hilly parts of Mull, but is greater in the eastern part of the Eoss than in the western.* EARLIER GEOLOGICAL INVESTIGATIONS. Very little has been written about the geology of the area to be described. In 1819 MaccuUoch t gave a general account, and in his geological map showed granite in the western portion, and in the eastern a belt of " primary " strata, consisting of " mica slate " and " quartz rock." He describes the general character of the granite, and illustrates its relation to the primary rocks by a plate showing a junction of the two near Loch Laigh.f He states that the posteri- ority of the granite is quite clear, and that at the places of contact " there may often be observed fragments of the different stratified rocks, detached from the main body and embedded in the granite they are in some instances so distinct that the parts whence they were separated may be traced, and the fragments re-united by the imagination." Professor Judd § in 1874 expressed the view that the granite was of later Palaeozoic age, and stated that no veins were seen proceeding from it into the Secondary and Tertiary strata. In 1892 Mr. J. G. Goodchild || called attention to the granite junctions at various locahties in the Ross of Mull, including some within one-inch map 3-5, and he aroused a good deal of opposition by advancing the view that the " granite veins have evidently eaten their way into the surrounding rock, along joints or other planes of weakness, melting the adjoining rock as they advanced, and without wedging any of it apart." LIST OF FORMATIONS. IT The following tabular hst includes the various formations in the portion of the one-inch map now being described. The arrangement of the divisions is not intended in all cases to indicate relative age. Recent and Pleistocene — Blown Sand. Peat. 25-foot Raised Beach. 50-foot „ „ 100-foot „ Glacial Drift. Tertiary Intrusive Sheets — Dolerite. Granophyre. * In the neighbouring village of Bunessan, Jlr. Duncan Cameron states that for the seven years prior to 1909 the average rainfall was 87 '73 in. April. May and June were, on the average, much the driest months. On the other hand, the rainfall at the Erraid Observatory, belonging to the Northern Light Commissioners, was on the average only 37-42 in. a year for the last seven years. This observatory is about 6 miles from Bunessan, in a direction slightly south of west. ■j- " A Description of the Western Islands of Scotland," vol. i. pp. 541, 553-556, and vol. iii. map facing p. 73, and plate xxvi. I One-inch map 43. § " The Secondary Rocks of Scotland. Second Pape/. On the Ancient Volcanoes of the Highlands and the Relations of their Products to the Mesozoio Strata," Quart. Journ. Qeol. Soc, vol. xxx., 1874, pp. 244 and 290. II " Note on a Granite Junction in the Ross of Mull," Oeol. Mag., 1892, p. 447. IT See also Note at end of Chapter XIII. General Account of the Geology. 77 North-West Dykes — Dolerites and Basalts. Camptonites. Monchiquites and Nepheline Monchiquites. Newer Palaeozoic — Intrusive Sheets and Dykes of Vogesite and Porphyrite. Granite. Diorite or Tonalite, Diorite Porphyrite or Tonalite Porphyrite. Metamorphic — Igneous — Intrusive Sills of Hornblende Schist. Moine Series — PeHtic Schists or Gneisses, with thin homblendic calcareous seams. Psammitio Schists or Gneisses, including Quartzo-felspathic Granulites and Quartzites. Lewisian Gneiss — Acid and Homblendic Gneisses, with some Pegmatites. It is believed, from evidence in various parts of the North- West Highlands,* that a great hiatus exists between the Lewisian Gneiss and the Moine Series. There is no doubt also that great periods of time are represented by the intervals between the intrusive sills of hornblende schist and the so-called Newer Palaeozoic intrusions, and also between the latter and the Tertiary intrusions. In this portion of the map the Lewisian Gneiss is only certainly knovm in the island of Soa and some closely adjacent smaller islands. Soa is hardly 2 miles S.S.W. of lona, which has long been known to be chiefly composed of Lewisian Gneiss. The Gneiss of Soa is separated from the chief tract of the Moine Series by the Sound of lona and a considerable breadth of granite. In lona the Lewisian Gneiss is closely associated with breccias and arkoses,t which are now considered to be Torridonian. These are in a much less altered condition than any of the Moine rocks in the district, and it is hence suggested that the lona arkoses and the Moine schists on the east may be separated by the great Moine thrust, of post-Cambrian age, which so often in the North- West Highlands has brought the Moine schists on the east against much less altered Cambrian and Torri- donian rocks on the west. A wide space of sea separates the Moine rocks of the Ross of MuU from the Torridonian rocks of Colonsay, described in Part I. of this Explanation, and it seems probable that the great Caledonian Glen Fault passes under this sea. On the shores of Loch Linnhe it is observed that the Moine rocks on the north-west side of the fault are generally in a more highly altered state than most of the rocks on the south-east side, and similar or even greater differences exist between the Moine rocks of Mull and the Torridonian of Colonsay. Various facts have recently been brought to light which lend strong support to Professor Judd's view that the Ross of Mull granite is of later Palaeozoic age, and may thus be classed as one of the Newer Granites of the Highlands. It has nowhere been observed in a dis- * See, for instance. Dr. Peach, in the Annual Beport of the Geological Survey for 1892, p. 262. t Sir A. Geikie, " Recent Researches into the Origin and Age of the Highlands of Scotland and the West of Ireland." Lecture at the Royal Institution, June 7, 1889 ; Nature, vol. xl. pp. 299 and 320. 78 Introduction : Ross of Mull. tinctly sheared condition, but, on the other hand, in one-inch map 43, it is cut by various dykes of mica trap and lamprophyre,* like those which in other parts of the Highlands traverse the Newer Granites. If we take into account the Torran Rocks, which are nearly all formed of granite, we see that the extent of the plutonic mass is greatest in a S.S.W. direction, being more than 9 miles, while the breadth is rather less than 6. The granite is closely associated, particularly at the south-vrest end of the Ross, with masses of dioritic rock which seem to have consolidated slightly before the granite, and are often cut by veins of it or are included within it. This, it need hardly be said, is in agree- ment with the rule usually found in other plutonic masses. f The North-West dykes are not generally abundant, but are fairly numerous near the eastern end of the area, from which the plateau- basalts of Tertiary age are only about a quarter of a mile distant. J It is probable that many of these dykes are Tertiary, and possibly all are, but some belong to petrographical types so different from the Tertiary dykes of Skye that at present we cannot feel confident about their age. c. t. c. * T. 0. Bosworth, Summary of Progress of the Geological Survey for the year 1908, p. 55. t See, for instance, J. R. Dakyns and J. J. H. Teall, " On the Plutonic Rocks of Garabal Hill and Meali Breac," Quart. Journ. Geol. Soc., 1892, vol. xlviii. p. 107. J A large fault probably intervenes. CHAPTER XII. The Crystalline Gneisses and Schists of the Ross of Mull. 1. LBWISIAN gneiss. The Lewisian Gneiss of Soa is chiefly composed of tolerably weU- defined bands varying in breadth from less than an inch to several yards. The usual strike is north-west, and the dip norbh-east, from 40° to 60°. The commonest varieties are of pale grey or pale reddish colours, and are chiefly composed of felspar with small admixtures of quartz and biobite. In other bands quartz is abundant, in cloudy or milky grains which somebimes attain the breadth of a quarter of an inch or more. Dark gneisses with abundant distinct hornblende are also tolerably frequent, and are sometimes found in large knots 10 or 20 yards in breadth. The whole assemblage has a much closer re- semblance to bhe Lewisian Gneiss of West Sutherland and Wester Ross than has the Lewisian Gneiss of Colonsay.* The Soa gneisses have a more massive character and a coarser grain than those of Colonsay, and have probably been much less crushed and granulitised. Nevertheless, yellowish green epidotic crush lines are very common in Soa, and form a striking contrasb in colour with the redder gneisses between them. They generally strike north-east or north-north-east across the banding, and incline steeply south-east. Possibly they have been produced in connection with a great line of thrust — the Moine bhrusL — which, as already stated, may separate the Lewisian Gneiss of Soa from the Moine rocks farther east. The gneiss is crossed by some coarse red pegmatites, which may attain a breadth of 6 or 7 yds., and are best developed in the eastern part of the island. They occasionally show cleavage faces of red alkali felspar, as much as 3 or 4 in. in breadth, together with shorb strings of biotite or chlorite, and veins, perhaps a foot broad in places, which mainly consist of magnetibe. No such pegmatites have been noticed in the supposed Torridonian rocks of lona. They are also much broader and redder than any pegmatites seen near the granite of the Ross of Mull, and it seems probable that they are of pre-Torridonian age. Among certain specimens collected for microscopic examination, Dr. Flett has detected a pyroxene-hornblende gneiss, with a pale green augite surrounded by darker green amphibole. He states, also, that the biotites of the biotite gneisses contain, in their middle portions, needles of rutile arranged in three directions crossing one anobher,f and that some of the gneisses contain exceptionally large crystals of zircon. c. t. c, e. h. o. c. * See pp. 10-12. ■\ Similar inclusions are also found in biotite in the Lewisian Gneiss of Colonsay. Seep. 12. 79 80 Gneisses and Schists of Ross of Mull. The middle and the eastern rocks of the West Reef, 3 miles south of Soa, may possibly belong to the Lewisian Gneiss. They consist of biotite gneiss striking north-west and dipping north-east. Perhaps they represent patches entangled with the granite. A small patch, not shown on the map, entangled with the granite on the west side of Ard Mor may also perhaps belong to the Lewisian Gneiss. e. h. c. c. 2. THE MOINE SERIES. The Moine Series of metamorphic schists or gneisses forms most of the ground on the east side of the granite for a distance of about 6 miles. Farther west it also composes many masses, both large and small, which are entangled with the granite. Some of the masses near Cnoc Breac are as much as half a mile in length. It is important to notice that in most of the larger masses the dominant strike is much the same as in the Moine rocks outside the granite, its general direction being N.N.E. It is, therefore, supposed that these masses are not " floating " in granite, but have solid " roots," connecting them with the Moine rocks elsewhere. The Moine rocks are well exposed on the coast between Rudh' Ardalanish and the north-east corner of the map. Their main con- stituents — quartz, felspar, biotite and muscovite — are generally in rather large grains or flakes, so that perhaps in most cases the rocks deserve the name of gneiss rather than schist. They are repeatedly folded along axial planes, striking north-east, which are vertical or steeply inclined, often towards the south-east. Isoclinal folds with both limbs inclining south-east are also general. C. t. c, e. h. c. c. The pitch of the folds is often seen to be southward at rather gentle angles. By means of it the order of succession, as it now exists in the area, can be made out. It is indicated briefly in the following list, in descending order : — 3. Quartzo-felspathic Granulites. 2. Pelitic Gneiss. 1. Lowest Quartzose Group. The lowest beds (1) are seen in one or two small exposures on the south-west side of Ardalanish Bay. They consist of bands of quartzose schist, almost quartzite, separated by thinner partings of pelitic gneiss. In the map this quartzose series has not been separated from the overlying pelitic gneiss. It is overlain by a band or bands of hornblende schist, of intrusive origin but keeping almost to one stratigi'aphical horizon. Higher still comes a peUtic or semi-pelitic gneiss (2), of various types, which makes the common rock of the country all the way from Arda- lanish Bay to Traigh Bhan an Sgoir, and, still farther east, comes again into the map, near Rudha nam Braithrean. It also forms most of the patches of gneiss which are entangled with the granite west of Ardalanish Bay. One common type of the pelitic gneiss consists chiefly of flakes of biotite and muscovite closely mixed together and lying more or less evenly along parallel foliation planes, together with grains of quart,z Maine Series. 81 and felspar interspersed between the flakes or combining to form thin lenticles : smaU garnets are tolerably abundant. Specimen 13421, from Traigh Bhan an Sgoir, is called by Dr. Flett a typical muscovite biotite gneiss, contaiaing biotite, quartz, albite etc. and a few small garnets. e. h. c. c. In certain localities the qiiartzo-felspathic lenticles or streaks are closely crowded together and form coUectively a considerable pro- portion of the rock, though individually they are very short, and perhaps never more than half an inch broad. Perhaps they represent little pegmatites which have been developed in areas of special metamorphism. They are repeatedly disposed in sharp folds, together with the micaceous constituents of the rock, and, as they are very well developed on the coast south of Scoor, at a distance of three miles and a half from the granite crop, it hardly seems likely that they are due to the action of the granite. An additional reason for this opinion is the common occurrence of similar streaks in the gneisses of other parts of the Highlands, even in areas where no granite is known for many miles. c. T. c. The above described typical muscovite biotite gneiss is mixed with various other types of gneiss, biit none of these are differentiated on the map. In one type, perhaps more general than that just described, the quartzo-felspathic constituents and the micas, which are often rather small, are closely mixed together, and form speckled grey rocks of semi-pelitic character, which have sometimes, from their appear- ance, been called "pepper and salt " rocks. Such rocks are well seen between Port Uisken and the east side of A' Chleit. They are often arranged in bands differing slightly in character and thickness, which no doubt represent beds of original deposition. The bedding surfaces of some of the bands are spangled over with large white micas of secondary origin, with their cleavage planes lying across the bedding at various angles to one another. Such spangled bands are particu- larly well seen about 300 yds. W.S.W. of Dun a' Ghaird, where some of them are hardly a quarter of an inch thick. In various places near the sill or sills of hornblende schist, already alluded to, we find, interbedded with the pehtic gneiss, thin lenticular bands of quartzose rock, rarely more than a foot thick, which are rich in garnets, large flakes of black biotite and dark green hornblendes, generally in feathery blades. Microscopic examination (13425) shows also that felspar is occasionally abxmdant. These bands are well seen on both sides of Ardalanish Bay and also in some of the patches of gneiss entangled with the granite. They have a considerable resemblance to the zoisite hornblende garnet rocks of the Moine series of the Northern Highlands,* though no zoisite has yet been identified with certainty in any of the Mull bands. They also often contain a distinct proportion of carbonate which causes certain laminae to weather with a carious surface. They are believed to represent bands which originally contained some carbonate of hme, but which in the course of regional metamorphism have lost much of their carbonic acid and given up most of their lime to help to form garnet and hornblende. Besides the thin blades of hornblende, stouter hornblendes, or groups of such, are occasionally seen, which are poikihtio with numberless * First described by W. Gunn and J. J. H. Teall, in the Summary of Progress of ilie Geological Survey for 1897, p. 41. 6 82 Gneisses and Schists of Ross of Mull. inclusions of quartz or felspar. Occasionally, wken the band is folded, the blades of hornblende and the flakes of biotite are seen to be arranged across the bedding, and parallel to the axial plane of fold. Various thin leaden-grey seams, about a quarter or half an inch thick, have been found in the neighbourhood of the garnet hornblende quartzose bands on the west side of Ardalanish Bay. Dr. Flett reports that these (13426, 13854) " consist very largely of small shining flakes of muscovite. The microscope shows the presence of another mineral in very small hexagonal prisms. It has the optical properties and chemical reactions of apatite, and its abundance proves that these seams must have been originally phosphatic lumps or nodules, which were presumably phosphatic concretions in the muddy sediments from which these schists were derived. The amorphous phosphate of the concretions has been changed by metamorphism into the small crystalline prisms of apatite. Other ingredients are biotite in small quantity and opaque dust, which includes both iron oxides and graphite." On the coast between Port Uisken and A' Chleit various thin garnetiferous calcareous siliceous seams have been noticed which cross the bedding of the neighbouring rocks distinctly. It is probable that these seams represent calcareous veins which were formed before the advent of the regional metamorphism. At Traigh Bhan an Sgoir the quartzo-felspathic granulites (3) are separated from the pelitic gneiss on the west by thick bands of unusually white siliceous granuHtes, or quartzites, which have not been differ- entiated in the map from the other granuUtes. The outcrop of the more siliceous series is perhaps 100 yds. in breadth, but the margins are not sharply defined. Many of the granulites show indications of false-bedding and of the granulitised remains of pebbles, and it is evident that they represent a series of coarse felspathic grits or arkoses. Remains of pebbles are particularly abundant and large in a false-bedded band, about 80 yds. broad, which strikes N.N.W. across Slochd a' Mhuilt, and is approximately vertical. Remains of pebbles, both of quartz and felspar, are here very abundant, but always in a granulitic condition, and often, particularly in the case of quartz, greatly elongated. When the observer faces the bedding planes, he sees the direction of elonga- tion dipping S.S.E. at an angle of 60° or 70°. Specimen 13427, from the locality just mentioned, is stated by Dr. Flett to contain both biotite and muscovite, to be fairly rich in miorochne, and to show bands of heavy minerals indicating bedding. Towards the top of the granulites, fine-grained schists with more biotite, and allied to the " pepper and salt rocks " already alluded to, begin to appear ; but they have not been mapped separately. The steep bedding and foliation of the Moine rocks are frequently crossed by aknost horizontal or gently inclined joints, some of which have provided paths for intrusive dolerite sheets of Tertiary age {see Chap. XV.). Along one of the nearly horizontal joints in Slochd a' Mhuilt, movement has taken place to the extent of at least 6 or ^ in- c. T. c, E. H. c. c. Hornblende Schist. 83 3. INTRUSIVE SILLS OF HORNBLENDE SCHIST. None of the crops of hornblende schist or epidiorite are of large extent, but they are very numerous near Ardalanish Bay and between ■there and Dun a' Ghaird. With one possible exception, to be mentioned shortly, they represent a thin intrusive sill, often only a "few feet thick, or various thin sills, on nearly the same stratigraphical position, and folded together with the schists and gneisses of sedi- mentary origin. The folding is so close and repeated that there can be little doubt that many of them represent folded portions of one sill, and to some observers it may seem not impossible that all of them do so. The horizon of the siU or sUls is a little above the lowest quartzose group of the Moiae series of the district. Garnets, often a quarter of an inch in diameter, and sometimes as much as an inch and a half, usually occur in these sUls in great numbers, but they are considerably cracked, rounded in outline and ■encircled with wisps of biotite. They often contain many inclusions of felspar and hornblende, and in some localities the place of the :garnet seems entirely taken by aggregates of these minerals, as if they were pseudomorphous after it. c. t. c, e. h. c. c. A peculiar hornblendic rock, on the coast about 280 yds. west of Dun a' Ghaird, makes a crop about 60 yds. long in an N,N.E. direction and 8 yds. broad at the maximum. It contains some .seams of ferriferous carbonate 5 or 6 in. broad, and parts which are mainly composed of white mica, biotite or chlorite, together with garnets and blades of actiaohte crossing the foliation. The garnets are sometimes 2 or 3 in. long. Prisms of black "tourmaline, nearly an inch long, and others of pale blue Icyanite, sometimes more than an inch and a half long, are common in some positions, and on microscopical examination (13849, 13850) Dr. Flett has also detected staurolite, andalusite and silUmanite. The occurrence of the last three minerals suggests contact alteration by the Ross of Mull granite, but the locahty is more than two miles from the nearest exposure of granite, and no tourmaUne or kyanite have been noticed near the granite. In some respects the band •calls to mind various altered sediments in the Gairloch and Loch Maree district (one-inch maps 91 and 92) which belong to the Lewisian Gneiss series, but it is difficult to see how it can belong to this series. In the adjoining beds of the Moine series no kyanite or tourmaline or sillimanite could be detected macroscopically, nor is it known that staurolite or andalusite occur. Hornblende schists of normal tjrpe are found close to the rock described, and differ notably from it. c. t. c. CHAPTER XIII. The Eoss or Mull Igneous Complex and Associated Intrusions: The Ross of Mull Igneous Complex* forms the western portion of the Ross, and extends at least to the Torran Rocks, far out to sea in a south-west direction. 1. THE DIOEITIC ROCKS. The dioritic rocks form a very subordinate part of the complex and are confined to its western side, where they compose Eilean. a' Chalmain, considerable masses on Torr Mor and Ard Mor, and innumerable smaller lumps. The mass which forms EUean a' Chalmain is more than two-thirds of a mile long and a third broad. These dioritic rocks represent the earlier and more basic stages in the differentiation of the plutonic magma, and have been intruded earlier than the granite, which burst through them and includes numberless fragments of them. Most of the masses which are sufficiently large to be shown on. the map appear to be quite surrounded by granite, but others occur at junctions of granite with schist masses, and though they are entangled with the granite, have probably, as already suggested, soHd " roots." None of the varieties are very basic. Specimen 13412, from. Eilean a' Chalmain, may, Dr. Flett states, be described as tonalite. It contains much green hornblende, biotite and sphene : plagioclase (andesine and oligoclase) is the common felspar, but orthoclase occurs, inters titially in some abundance, and there is also quartz and micro- pegmatite. Another specimen (13415) from the Sound of Erraid,. contains glomero-porphyritic groups of quartz and oligoclase, and may be classed as a dioritic porphyrite or a tonalite porphyry. 2. THE GRANITE. Most of the granite is a reddish, rather coarse-grained muscovite biotite granite, without conspicuous felspar phenocrysts. It shows- no chilling at the margins, and few or no indications of flow structure, A specimen (13417) from Ardalanish Point is stated by Dr. Flett to be of the same type as many of the Newer Granites of the Highlands ; f it contains a good deal of microcline and microcline microperthite, as well as some oligoclase and perthitic orthoclase : albite occurs as individual crystals, and as veins and patches in potash felspars : the quartz is sometimes slightly crushed and shows, the beginning of mosaic formation. * An important paper by T. 0. Bosworth on " Metamorphism around the Ross of Mull Granite," Quart Journ. Geol. Soc, 1910, vol. Ixvi. p. 376, was published while this memoir was in the press. t Many of these are beUeved to belong to some part of the Lower Old Bed Sand- stone period. 84 Granite. 85 Thin fine-grained aplite veins are tolerably common in the granite, and muscovite pegmatites have also been detected in several places, but are rarely more than a few inches thick. On the south-west side of Ardalanish Bay some of these pegmatites show distinct "graphic" structure. Thin pegmatites are also not uncommon in the Moine rocks east of the granite, and are perhaps genetically connected with it. The country formed by the granite, though never of great elevation, is full of small hills of a rough rocky character, and is crossed by various long hoUows which may perhaps represent lines of crush. Joint planes are sometimes conspicuous, inclined at moderate angles and giving a bedded appearance in the distance. Further facts confirming Professor Judd's view of the later Palaeozoic age of the granite have now been disclosed, especially the discovery, in one-inch map 43, of various dykes of mica trap and hornblendic lamprophyre,* and sheets of porphjnrite which cut and are chilled against the granite. In one-inch map 35, also, two intrusions of porphyrite, with chilled edges, cut the veins of granite on the west side of Ardalanish Bay. The eastern margin of the granite, from Eudh' Ardalanish north- wards is very irregular, sheets and irregular veins extending in all directions from the main mass into the adjoining Moine gneiss. The number of pieces of gneiss entangled with the granite is countless. As the larger entangled masses possess the same strike as the main mass of Moine gneiss east of the granite, they can hardly have been floating in the granite magma. It seems probable, rather, that they lepresent parts of a very irregular floor below the granite. Metamorphic Effects of the Granite. As regards the sizes of the quartz and felspar grains, and of the mica flakes, there is no striking difference between the Moine rocks •close to and entangled with the granite and those three or four miles away from the granite. It seems probable, therefore, that these constituents of the rock were already so coarsely crystalHsed before tflfe intrusion of the granite that they could not be appreciably ■enlarged in size afterwards. e. h. c. c. Certain minerals, however, which in other districts are usually confined to the neighbourhood of large intrusions,t have been noted in abimdance in the gneiss close to and entangled with the granite. In one-inch map 35, small clusters of thin white needles of sillimanite, large enough to be readily discerned macroscopicaUy (Plate VI. i'ig. 6), are, for instance, very abimdant in the pehtic gneiss entangled with the granite 470 yds. E.S.E. of the Ordnance Station 279 on Torr Fada, by the northern walls of Breac-achadh, and 400 yds. E.S.E. of these walls. In a specimen (13847) from Breac-achadh, Dr. Flett has also detected cordierite, andalusite and spinel (Plate VI. Fig. 6) : and in a specimen from the locality near Torr Fada he noted silli- manite, andalusite and spinel. The gneisses next the granite on the west side of Ardalanish Bay frequently display yellowish tints, which are probably due to the * T. 0. Bosworth, Summary of Progress of the Oeological Survey for 1908, p. 55. t J. J. H. Teall, "The Geology of the Country near Oban and Dalmally," Mem. Qeol. Survey, 1908, pp. 143, 145. J. S. Flett, " The Geology of Lower Strathspey," Mem. Qeol. Survey, 1902, p. 53. 86 Ross of Mull Igneous Complex etc. weathering of finely disseminated iron pyrites specially introduced near the granite. c. t. o. 3. THE POEPHYRITE AND VOGESITE INTRUSIONS. It is probable that the porphyrites and most of the unfoliated lamprophyres in the Eoss of MuU were intruded shortly after the con- sohdation of the granite, though their margins are chiUed wherever they have been observed in contact with this rock. On the other hand, these intrusions are themselves cut by dykes, striking north-west, some of which are probably of Tertiary age. Only two intrusions of porphyrite have been mapped. The larger band forms a sheet, about 15 ft. thick, at and to the south of Carraig Mhor. It strikes N.N.E., incKnes gently north-west, and cuts many irregular granitic intrusions, but is itself crossed by two faults and accompanying dykes ; one of the faults throws the sheet down to the west for perhaps 40 ft. Judging from hand specimens- this sheet is composed of two rather different varieties of rock, one of which contains conspicuous quartz blebs, while the other does not. These varieties are often rather sharply separated, but no distinct chilling was observed at the planes of contact. Another thinner more steeply inclined intrusion, usually about a foot and a half thick, is well seen about 100 yards north and north- west of Carraig Mhor. The strike is in places only a httle north of east, but in other places nearly north-east. This intrusion also cuts many granitic bands, but is itself cut by a nephehne monchiquite dyke striking north-west. Dr. Flett has examined microscopic sections of the two porphyrite bands from Carraig Mor, and finds that both of them contain porphyr- itic felspar and hornblende, but the southern mass (14456) is of more acid type, as it contains quartz both as small phenocrysts and in the groundmass, while the other rock (14455) is practically free from quartz, and approaches the lamprophyres in many points of its microscopic structure. j- Two or three intrusive sheets of dark grey vogesite strike N.NJE. across the Lewisian Gneiss of Soa. They incUne north-west and are only 4 or 5 ft. thick. The margins are distinctly chiUed and often show closely clustered spheruHtic forms, about a quarter of an inch in diameter, which are conspicuous on the weathered face. These sheets are later than the pegmatites in the island, also than the epidotic crush ILaes and a nearly horizontal plane of movement which displaces the bands of gneiss. On the other hand, they are cut by various basaltic N.N.W. dykes. In various localities between Port Uisken and Dun a' Ghaird we see a thin dyke, apparently a lamprophyre, striking nearly east and west and weathering with a reddish brown surface. Perhaps all the exposures belong to one dyke. c. t. c, e. h. c. c. NOTE. Since this Explanation was sent to the printer we have noticed at Port Mor, on the west side of Ardalanish Bay, an intrusive sheet, inclining west, which is clearly older than the granite, though much Pre-granite Lamprophyres . 87 less foliated than, and evidently later than the common hornblende schists. This sheet is cut by many veins of granite. In microscopic section it proves to be a vogesite of peculiar type, and has suffered considerable contact metamorphism. The foliation only affects the top and bottom parts, merely for breadths of a few inches near the surfaces : it is approximately parallel to these surfaces, and is thus in quite a different direction from the foliation noticed in the adjoining Moine gneisses. Just on the east side of Ardalanish Bay another lamprophyre sheet, inclining west, shows a peculiar varied character, but it does not seem to be foliated in any part. On the west side of Loch na Lathaich, in the sheet to the north of this, we have also observed various lamprophyric sheets that resemble very closely those mentioned above. They are occasionally slightly foliated and are sometimes cut and contact altered by the granite. They are quite distinct in microscopic character from the lamprophyre sheets that are later than and cut the granite. We have accordingly two sets of lamprophyres in the Koss of Mull, one older than the granite, and one which cuts the granite. Of these the former may possibly correspond to the fohated or cleaved lamprophyres of Colonsay, that have been described in an earlier part of this memoir. c. t. c, J. s. f. CHAPTER XIV. The Noeth-West Dykes of the Ross of Mull. geneeal desceiption. As already intimated many of these dykes are of Tertiary age, and perhaps all of them are. But the camptonites and monchiquites differ greatly in petrological character from any of the Skye Tertiary dykes described by Mr. Harker, and it is possible that in the Ross of Mull we have two sets of north-west dykes, one Tertiary and another much older. The north-west dykes are not generally abundant, but they are tolerably numerous, as already stated, at the eastern end of the area, in the vicinity of the Tertiary plateau-basalts. Thus, on the west side of Traigh Bhan eleven dykes are seen, all striking north-west or N.N.W. within a distance of 700 yds. measured at right angles to their general direction. This apparent increase in abundance in an easterly direction, near the plateau-basalts, may in itself suggest the Tertiary age of the dykes, but it should be stated (1) that in the neighbouring plateau- basalts dykes do not seem abundant, (2) that two of the Traigh Bhan dykes are cut by a dolerite sheet supposed to be of Tertiary age, and (3) that one of these two dykes is crossed by a number of small faults. The dyke just referred to (14446) is well exposed about 333 yds. south-west of the head of Traigh Bhan, though generally less than IJ ft. in breadth. It dies out altogether here and there, but appears to be cut by an adjacent dolerite sheet. It contains many conspicuous crystals of black hornblende, sometimes as much as an inch in length. Dr. Flett names it fourchite and describes its petrographical characters in the second part of this chapter. The parallel dyke, about 20 yds. farther west, is a dolerite (14445) and is very clearly cut by the almost horizontal top of the dolerite sheet, which is chilled against it. The breadth is about 6 ft., but a considerable strip of granulite is partly included near the north-east side. A third parallel N.N.W. dyke (14447), about 4 ft. thick, which is well seen on the shore about 180 yds. south-west of the head of Traigh Bhan, is too fine-grained to be readily distinguished in the hand specimen from the dyke last mentioned ; but after microscopic examination it has been determined as a camptonite. At the head of Traigh Bhan two thin north-west dykes are seen, as well as a dolerite sheet. The thinner dyke, only 4 in. broad, cuts the sheet, shows tachylite selvages, and may be confidently classed as Tertiary. The other is a camptonite and of more uncertain age. It is about 2 ft. thick and" contains numbers of small red ocelli, but is a good deal decomposed and veined with thin strings of ferri- ferous carbonate. Petrography. 89 Near Carraig Mhor, on the south-west side of Ardalanish Bay, three north-west dykes have been detected, all rather fine-grained and of much the same aspect in the hand specimen. The most northerly example (14452) is well seen about 100 yds. shghtly west of north of Carraig Mhor. It is about 3 ft. thick and cuts through a, porphyrite intrusion running north-east or east, already described in Chapter XIII. A specimen has been examined under the microscope by Dr. Flett, and is classed as a nepheline monchiquite. Another ■dyke of nephehne monchiquite may also be mentioned here, as it is well seen a httle within one-inch map 43, to the north of the map being described. This dyke (13846) is exposed in a Uttle quarry by the roadside, 690 yds. east of Ardachy. It strikes W.N.W., is about 3 ft. thick, and cuts across an iutrusive sheet parbly com- posed of granite porphyry. Most of the dyke is composed of a very ■dark compact rock, but it contains also a number of small reddish streaks and ocelU, many of which show centres of white calcite— probably infiUings of vesicles. The other dykes in this neighbourhood are best seen in libtle creeks on the shore about 170 and 233 yds. S.S.W. of Carraig Mhor. They both cut an intrusive sheet of porphyrite, and are considerably crushed, being accompanied by faults which displace the sheet. The throw in the first locahty is about 6 ft. down towards the north-east, and that in the second is about 40 ft. down towards the south-west. Dr. Flett states that the dyke in the first locality (14453) is a dolerite, in a considerably decomposed state, while that in the second (14454) is a camptonite. In the island of Soa three or four north-west dykes are excellently exposed, and cut the vogesitic intrusive sheets already referred to in Chapter XIII. Eight or nine other north-west dykes have been traced short dis- tances in the six-inch maps, but it has not been possible to show them all on the one-inch scale. None of these have been sliced. During the field survey they were all supposed to be dolerites or basalts, ■but, in view of the characters disclosed in the shced specimens, it is probable that the field determinations are often wrong. Out of ten sliced specimens of north-west dykes five turned out to be camptonites, two monchiquites, and three dolerites or basalts. Only three of these specimens showed characters in the field which induced suspicion that they were not dolerites or basalts. c. t. c. PBTEOGEAPHICAL CHARACTERS. The north-west dykes of the south part of the Ross of Mull may be dividedinto three series : — fi (a) oUvine dolerites. (b) camptonites. (c) monchiquites, including fourchites and nephehne monchiquites. A general resemblance may be recognised between them and those of Colonsay ; monchiquites, for example, and nepheline monchi- quites occur in both districts, but the camptonites are represented only in Mull, while the crinanites occur only in Colonsay. 90 North-West Dykes of Ross of Mull. Dolerites. — Only two dolerite dykes of southern Mull have been, examined microscopically, and both of them proved to contain olivine (14445, 1050 ft. south-west of low water at the head of Traigh Bhan, and 14453, 500 ft. south-west of Carraig Mhor). Both are rather decomposed, their olivine being replaced by serpentine and carbonates. The first of these rocks has ophitic structure, and its steam cavities are filled with radiate fibrous zeolites. The second is peculiar in containing a small olivine nodule replaced by carbonates. Camptonites. — Pour dykes of camptonite have been examined from this quarter. Three of these occur in Traigh Bhan (14448, near the head at low water ; 14447, 550 ft. south-west of the head ; and 14451, 1600 ft. south-west of the head). The other (14454) was found 700 ft. south-south-west of Carraig Mhor. These camptonites are distinguished from the olivine dolerites by several features. They contain a good deal of brown hornblende in addition to augite, olivine and plagioclase felspar, and most of the slides show abundance of analcite. They are never ophitic in structure, but their augite forms well-crystallised prisms which sometimes are aggregated into radiate groups. Most of them also contain more or less perfect ocelli. The rocks are not conspicuously porphyritic, though the olivines are sometimes larger than the other constituents. This mineral is usually altered into serpentine, but some of the slides contain it in a fresh condition. The dark brown hornblende surrounds the augite in parallel growth and has the outUnes characteristic of amphibole ; small crystals also occur which contain no augite at their centres. The augite is purplish, with zonal and hour-glass structures and a strong dispersion. The plagioclase felspar has fairly good crystalline form in most cases, so that the rocks tend to be " panidiomorphic." Many of the felspars have the long rectangular outhnes so frequent in the basalts ; they are zonal, with centres of labradorite and andesiae and margins of oHgoclase. A little untwinned alkali felspar with refractive indices lower than balsam may now and then be detected between the larger crystals, and is devoid of crystalline form. In some slides much of the felspar occurs in large plates which have imperfect out- lines and are poikilitic with inclusions of the other minerals. There is much analcite in these camptonites, mixed usually with carbonates. It is never eumorphic, but fills up spaces which were left after the crystallisation of the felspar, and also occurs in steam cavities and at the very centres of the ocelli. The latter structures are not well developed in most of these dykes, but in 14451 they are sometimes nearly half an inch in diameter (Plate VI. Fig. 3.), rounded or oval, and are easily seen as reddish spots in the hand specimens. They consist of radiate felspars, long prismatic brown hornblendes and crystals of augite in smaller numbers. Monchiquites. — The monchiquite dykes include one fourchite and two nepheline monchiquites, the fourchite being a variety of monchi- quite which contains no olivine and is rich in brown hornblende in addition to augite. The dyke of fourchite (14446) is one of the group which occurs in Traigh Bhan ; it is found at a point 1000 ft. south-west of low water at the head of the bay. In the hand specimens it shows many large black phenocrysts of hornblende, up to half an inch in length ; in the Petrography. 91 sections these are deep brown with few inclusions and perfect crystalline forms. Smaller prisms of hornblende, equally perfect, occur in great numbers in the groundmass. The augite of the rock is of the purple zonal variety found in crinanites and camptonites ; it is always idiomorphic and is less abundant than the amphibole, but it Kkewise occurs in two generations, viz., as small phenocrysts and as an element of the groundmass. There is no felspar and no nepheline in the rock, but an abundant turbid yellowish base which is isotropic between crossed nicols. Small steam cavities are filled with analcite and scalenohedral crystals of calcite. The isotropic groundmass is not analcite, but a somewhat decomposed glass, as it has a refractive index above that of xylol (1"495), while that of analcite is lower. The two dykes of nepheline monchiquite occur 690 yds. east of Ardachy (13846) and 100 yds. sHghtly west of north of Carraig Mhor (14452). The former is a httle within one-inch map 43, to the north of the map being described. They are both fine-grained dark rocks, not obviously porphyritic, and contain small olivines which are decomposed to serpentine and carbonates. The augite, which is far the most abundant mineral, is of the purpHsh-brown, zonal variety, and occa- sionally forms small phenocrysts. The groundmass is crowded with minute idiomorphic crystals of pyroxene. Biotite occurs also, as in the nepheline ouachitite of Colonsay, and being less idiomorphic is of slightly later crystallisation, but often its crystals show traces of hexagonal outUnes. There is no felspar in these rocks except occasion- ally in the small ocelli, but the matrix consists of analcite, carbonates, chlorite and weathered nepheline. The last-named mineral is not in good preservation, but the cleavage, straight extinction, optical sign and refractive index serve to distinguish it from analcite and alkali felspar. It seems to form irregular plates without crystalline outlines, and it is uncertain whether the rock contained originally a small amount of glassy base ; if so, the glass has been entirely replaced by carbonates and analcite. ♦ j. s. f. CHAPTER XV. Teetiaey Intrusive Sheets. In this chapter we shall describe (1) various sheets of dolerite and (2) a few very thin squirts of granophyre. THE DOLERITE SHEETS. The dolerite sheets make conspicuous .features in the eastern portion of the Ross, not far from the plateau-basalts. They are markedly columnar, vary in thickness from 4 to 15 ft., and are generally nearly horizontal, or but gently inclined. They have evidently been intruded along a set of gently inclined joint planes, crossing the steep isoclines of the Moine gneisses at a high angle. One of the sheets has been traced for more than a mile. A normal specimen of a sheet at Traigh Bhan an Sgoir is stated by Dr. Flett to be a dolerite of the olivine-free type, and to consist of labradorite felspar and greenish-brown augite, with some apatite and iron oxides ; the structure is somewhat ophitic, the felspar crystals being well formed, and the augite occurring as irregular masses between them, but not in large optically continuous areas. The bottom portion of this sheet contains a good many xenoliths, some of them a foot or more in breadth, of coarse grain, which may be referred to the " cognate group " of Mr. Harker.* Dr. Flett states that these inclusions " belong to two classes ; one of these consists of bytownite and pale green augite, the other of bytownite and a nearly colourless enstatite or bronzite. In both cases there is very weU- marked ophitic structure, for the felspars are embraced by irregular areas of pyroxene, but the enstatite forms masses which have more tendency to crystalline outlines than the augite. The bytownite contains many rounded glass enclosures sometimes arranged in zones ; in the augite and enstatite there are branching and very irregular fluid cavities. The bytownite crystals are short and broad, but have rectangular outlines in most sections. In one of these nodules there are soft dark red grains which appear to be serpentine after olivine ; in another, secondary analcite occurs filled with vermi- form chlorite." e. h. c. c. Some parts of the above sheet contain also many xenoliths of the " accidental " group. For instance, in the upper part of the sheet, .in a position about 100 yds. north-east of the north-east corner of the bay, pieces of highly siliceous banded schist, sometimes more than a foot long, are exceedingly numerous. These pieces are quite unlike the pelitio gneiss which forms most of the floor and roof of the sheet at this locality, and have probably been carried from the siliceous schists which occur in situ 10 or 20 yds. further east. This seems * " The Tertiai-y Igneous Rooks of Skye," Mem. Oeol. Survey, 1904, p. 351. Alteration of Adjoining Sediments. 93 the more likely, because, just above the sheet, the banding of the pelitic gneiss is twisted sharply from the vertical, and made to assume a gentle westerly dip. Further west, in the upper part of the sheet somewhat west of the centre of the Isay, similar inclusions of highly siliceous schist occur again, and have perhaps been carried westward a considerably greater distance. In this last locahty there are also various small inclusions of black highly graphitic rock, of a shivery or shale-like character. They are less massive than the graphitic rocks which have been as yet observed in association with the plateau- basalts, and, if they have been derived from former outliers of such rocks, they must have been carried downward for a considerable distance, probably 300 ft. at least. It seems not impossible that they have been brought up from below, from some graphitic schist belong- ing to the Lewisian Gneiss, which may here underUe the Moine rocks. This dolerite sheet is well seen again some distance further east, on the west side of Traigh Bhan, where the nearly flat top cuts dis- tinctly through a vertical north-north-west dolerite dyke, about 6 ft. broad, already described ; it appears also to cut through a thinner north-north-west camptonite dyke about 20 yds. further east, but the cliff at this place is not readily accessible. On the foreshore near the head of Traigh Bhan, a dolerite sheet, probably a continuation of the one being described, is cut by a narrow vertical string about 4 in. broad, which strikes north-west, and has distinct tachylite selvages. At Traigh Bhan an Sgoir the landscape colour of the pelitic gneiss which is there cut by the dolerite sheet is changed from pale grey to a much darker grey for a distance of about 5 ft. below the sheet. On nearer examination the contact alteration found in this gneiss near the dolerite both at Scoor and Traigh Bhan an Sgoir is found to be very great, the foliation planes for a foot or two off the sheet being welded together, and their micaceous glimmer quite lost. Just below the base of the sheet we find in places a rock, a few inches thick, which macroscopically has some resemblance to an altered sandstone, but is really a partly fused gneiss, closely allied to buchite. A similar altered rock is also found just below the dolerite sheet at Scoor, which is probably continuous with the Traigh Bhan an Sgoir sheet. Dr. Mett thus describes the altered rocks below the Scoor sheet : "In their original state these rocks consist of quartz, felspar, muscovite, biotite and garnet, with a small amount of apatite and iron oxides. These minerals are not sensibly altered except the micas, and these retain their shapes and to some extent also their optical properties, though they have been partly fused and changed into new mineral aggregates. The muscovite has become very cloudy and gives its normal high polarisation colours only in a few places. Sections examined with high powers of the microscope prove to contain small isotropic grains of spinel and minute needles, which may perhaps be sillimanite, in a clear substance, with weak double refraction, which is possibly felspar. The biotite is altered also, though less completely than the muscovite, as portions of its crystals still show the intense pleochroism and dark brown colour which are characteristic of this mineral ; others have a ragged or eroded appear- ance, and are filled with minute dark grains which, with high magni- fication, are found to be magnetite and green spinel in small octahedral S4 Tertiary Intrusive Sheets. crystals. A narrow zone of clear, pale-coloured glass surrounds the spinels. Small crystals, which are probably sillimanite and cordierite, may be seen in some of the fused biotites. The felspar of these rocks is often turbid, but this may be a consequence of weathering." Dr. Mett also goes on to remark that the process involved in this alteration " is one of simple fusion, for the chemical composition of the altered rocks is not modified ; moreover, similar changes in the biotite of Highland gneisses are found in the fused rocks which form the walls of the old vitrified forts." c. t. c, e. h. c. c. A still more completely fused rock, a true buchite, which in parts has, in hand specimens, a close superficial resemblance to tachylite, forms most of a mass 4 or 5 ft. thick, just above the sheet, in a position 85 yds. shghtly north of east of the north-east end of the bay.* This buchite varies in character a good deal and shows a banded structure which in places is nearly horizontal and in other places vertical or sharply twisted. Oval specks, partly composed of opal or chalcedony and with a close resemblance to the amygdules of true igneous rocks, are not uncommon and sometimes as large as peas. In one section some nearly horizontal bands are cut by a nearly vertical string composed of a slightly different variety of buchite, and the buchite which forms most of specimen 13841 is also cut by a thin vein of slightly different character, so that it appears as if parts of the buchite had, on a small scale, acted as intrusive racks. Various specimens of the buchite have been examined under the microscope by Dr. Mett. Slice 13841 " consists mainly of small com- posite crystals of cordierite in a brown glassy base. The cordierite is transparent and colourless ; longitudinal sections are sharply rectangular, but cross sections have more or less perfect hexagonal outlines, and between crossed nicols the largest of them break up into radiating sectors, most clearly visible when a selenite plate is used. These crystals contain few enclosures, and most of them are not more than a hundredth of a millimetre in diameter. The brown glassy base closely resembles that of many pyroxene andesites in colour. Through most of the rock it is scanty, but there are patches which consist mainly of this glass ; in these there are long straight crystals, sometimes forked at their ends. From their optical properties they may belong to sillimanite, and the cross sections are bounded by faces which correspond to the usual prism in that mineral, while others are nearly square. The latter are evidently sections of another prism which is frequent in sillimanite, and this places the identity of the mineral beyond doubt ; combinations of the two prisms are also met with. Iron oxides are present as small octahedra (magnetite), as rather stout black rods, and as minute hair-like crystals, often bent or curved, forming typical trichites. Occasional very minute crystals of deep green colour may be seen among the small cordierites embedded in the brown glass, and from analogy with other buchites, like those of Ardmucknish',t may be assigned to green spinel." Another specimen (14839), taken from the same locality, is a fine- grained violet-grey rock, with a dull rather than a vitreous lustre. * Some rocks referred to granophyre also occur in this locality (see p. 9fi). t J. S. Flett, in " The Geology of the Country near Oban and Dalmallv " Mem 9, 57-61. of Ross of Mull, 97. Glas Ard, 67, 69. Glen Orchy, Comparison with kentaOenite of, 31, 32, 33. Gneiss, Lewisian, 7, 10-12, 77, 79-80. of Moiue Series, 77, 80-82, 93-95. Metamorphism of, to huchite, 93-95. Gold in Colonsay, 74. GOODCHILD, J. G., 76. Granite, 77-78, 84-85, 99. Erratics of, 6, 60, 97. Metamorphic effects of, on gneiss, 85-86. Granophyre, 76, 95-96. Quartz xenoliths in, 96. Granulite, 77, 80, 82, 104. Graphitic xenoliths in dolerite, 93. Gravel of possible glacial origin, 59-60. Great Auk, Remains of, in Oronsay, 6, 71. Grieve, Symington, 6-7, 62. 70, 71, 72. Grits of Torridonian formation, 13. Gtojst, W., 35, 81. Haekee, a., 30, 41, 53, 64, 88, 92. Hauohton, S., 53, 99. Hebrides, Outer, Comparison with glaciation of, 57, 58. Heim, a., 6, 63, 54. Hill, J. B., 32, 33, 36, 37. Hornblende schist, Intrusive sills of, 77, 80, 83. Hornbleudite, 8, 29-30. Hundred-foot raised beach of Colonsay, 7, 65, 66, 67, 72. of Ross of Mull, 76, 97-98. " Hymba," 1. Inoorporation of sedimentary material by solution, 8, 29, 30, 96. Intersections of dykes, 41. Intrusive rocks of Colonsay and Oronsay, 6, 6, 8, 28^6. of Ross of Mull, 76-78, 84^96. Inversions of strata, 19, 20, 23, 48. Iron pyrites, 74, 86. Islay, Comparison with, 8, 16, 17, 18. JuDD, Peoe. J. W., 76, 77, 85. Kkntallenite, 8, 31-33, 34, 36-37, 53, 74. 104. ICilchattan, 13, 15, 20, 21, 45-46, 66, 73, 74. Group, 18, 23. Lower, 43, 44, 65. Kilchiar.an Group of Islay, Comparison with, 16-17. ICiloran, 1, 15, 35, 36, 5.3, 59, 73, 74. Bay, 2, 3, 6, 7, 8, 15, 18, 20, 26, 28, 40, 47, 51, 52, 53, 68, 70, 74. Flag Group, 18, 25-26. House, 66, 73, 74. Kitchen-midden or shell - mound IQ Oronsay, 71. " Kniokungsebene," 56. Kyanite, 83, 100. Kyn ASTON, H., 32, 33, 36 37. Index. 107 Laceoix, a., 44, 95. Laimaolean, 28, 31. Lamprophyre dykes and silk of Colonsay, 8, 9, 36, 38-41, 50-51. Lamprophyres of Ross of Mull, 77, 86, 88-90. Lamprophyric phase of diorite, 34. syenite, 30. Leao Bhuidhe, 10, 11, 14, 18, 40. Lewisian Gneiss of Colonsay, 7, 10-12. Soa and Ross of Mull, 77, 79-80. Liath Sgeir, 18. Limestone, 5, 15, 20, 26, 47. Alteration of, 16, 26, 35. breccia, as erratics, 6, 60. erratics, 6, 60. in breccias, 35. Possible correlation of, 17-18, 26. Literature : Colonsay and Oronsay, 5-6, 101. Ross of Mull, 76, 101-102. Loch ChoUa, 23, 24, 25, 59. Fada, 3, 15, 18, 23, 59, 65, 72. Laigh, 76. na Lathaich, 87. Staosnaig, 15, 19, 20, 33, 59, 67. L6n M6r, 63, 66. M'Neill, Donald, 60, 62. Macculloch, J. A., 5, 76, 99. Macfabt.ani!, J. M., 72. Maohrins, 3, 15, 23, 67. Group, 18, 22-23. ■Magnetite in pegmatite, 79. Maol Buidhe, 4. Marine erosion, Pre-glacial, 3, 02. Marten, Remains of, in Oronsay, 71. Maufe, H. B., 35. Meall na Suiridhe, 66. Metamorphism of Lewisian Gneiss, 11- 12. of pelitic gneiss, 93-95. of Torridonian sediments, 6, 8, 15-16, 49-56. Thermal, by augite diorite, 33. by dolerite sheet, 93-95. — . — . by kentallenite, 3 1 , 5.3. by syenite, 28, 35. Mica porphyry, as erratics, 60. Millbuie, 13, 14, 16, 18, 24. Group, 18, 24^25. Minor intrusions of Colonsay and Oron- say, 7, 38-46. of Ross of Mull, 76, 77, 86-96. Moiue Series, 77, 80-82, 93-95, 100. Apatite in, 82. Calcareous siliceous seams in, 82. Granulites of, 77, 80, 82. Lowest Quartzose Group of, 80. Metamorphism in, 93-95. Politic gneiss of, 77, 80-82, 9.S-95, 100. Moine thrust, 77. Monohiquite dykes of Colonsav, 5, 8, 41, 43-46. J. of Ross of Mull, 77, 89-91. Mudstones of Torridonian, 15. Muscovite biotite gneiss, 80, 81. Names, Origin of; Colonsay, Oronsay, and Ross of Mull, 1-2, 75. Neolithic floor, 9, 70. man, 9, 70-72. raised beach, 65, 68, 70, 71. Nepheline monchiquite, 77, 89, 90, 91. ouachitite dyke in Colonsay, 44r-46. NewCave, 6-7, 70-71. " Newer Granites," Relation of local rocks to, 36-37, 77-78, 85, 87. North-west dykes of Argyllshire, Age of, 41. of Colonsay, 38, 41-46. of Ross of Mull, 77, 78, 88-91. Olivine dolerite dykes in Colonsay, 41, 42. of Ross of Mull, 89, 90. Oronsay Group, 18, 22. Ouachitite dyke in Colonsay, 44-46. Overfolding, 19, 20, 23, 48. Overthrust fault, 23, 48. Peach, Db. B. N., 7, 16, 17, 35, 77. Peat, 7, 72, 76, 98, 100. Pebbles in Moine rocks, 82. in Torridonian, 11, 14, 18, 22. Pebbly quartzite in Torridonian, 11, 14, 18. Pegmatites in granite, 79, 85. in Lewisian Gneiss, 10, 11, 12, 77. 79, 86. in Torridonian grits, 16. Pennant, Thomas, 5, 73. " Pepper and salt rocks," 81, 82. Petrography of buchite, 94^95. of dolerite sheet of Ross of Mull, 92, 93. — — of granophyre of Ross of Mull. 95, 96. of hornblende schist, 83. of lamproph3Te dykes, 40. of Lewisian Gneiss, 11-12, 79. of Moine schists, 81-82. of north-west dykes, 41-46, 89-91. — — of plutonic rocks, 29-34, 84-8.J. of porphyrite intrusions of Ross of Mull, 86. of Torridonian epidotic grits, 13-14. . of TOgesites of Ross of Mull, 86. Phosphatic seams in Moine Series, 82. Phyllites in Torridonian, 15. ; Decomposition of, 73. Picrite, 29, 31, 32-33, 104. Pipe, Possible volcanic, 35. Plain, Coincidence of hiUtops in, 2, 75. of marine denudation, Pre-?lacia!, 3, 62-64, 97. Playeair, J., 6. PleochroTc halos, 30. Plutonic rocks of Colonsav, 5, 6. S, 28-37, 51-53. of Ross of Mull, 77-78, 84-86. Pollakd, De. W., 46, 74. Porphyrite, 77, 85, 86. as erratics, 6. Porphvritic type of kentallenite, 32. Port an Obain, 11, 34, 41, 68. an Tighe V.hbir, 3, 66. 108 Index. Port Ban, 71. Easdail, 30, 70. Lotha, 3. M6r, Colonsay, 2, 3, 20, 22, 23, 40, 43, 65, 69. , Boss of Mull. 86. na Cuilce, 39. nam Fliuchan, 3, 66, 68. Olmsa, 25, 59. Sgiobinish, 3. Uisken, 8], 82, 86, 97, 98. Post-glacial deposits of Colonsay and Oronsay, 65-72. raised beaches of Colonsay and Oronsay, 65-70. Pre-glacial plain of marine denudation, 3, 62-64, 68. Pyroxene-hornblende gneiss, 79. Qtjaekying in Colonsay, 74. in Eoss of Mull, 99. Quartz in buohite, 95, 104. xenoliths in granophyre, 96. in lamproph3Tes, 39. Quartzite, 80. boulders in breccia, 28, 35. in hornblendite, 29-30. in lamprophyres, 39. of Islav, Possible evidence of age of, 18. pebbles in Torridonian, 11, 14, 18, 22. Quartzose bands in pelitic gneiss, 81. Radley, E. G.,_45. Eainfall of Colonsay, 3. of Ross of Mull, 75-76. Raised beaches, 6, 7, 9, 62, 70, 73. 76, 97-98. later than 25-f t. beach, 70. Reasagbuie, 9, 25, 43. Recrystallization, Absence of, in Torri- donian sediments, 16. Recumbent fold, 23, 48. Red deer. Remains of, in Colonsay and Oronsay, 71. Reversed fault, 23, 48. Rhinns of Islay, Comparison with, 7, 16- 17. Roches moutonn^es, 57, 59. Rock basin, Submerged, 6, 57-59. groups. Local variations within, 22- 27. Rudh' an Dunain, 22. Ard Alanais, 64. Ardalanish, 80, 85, 97, 98. Rudha Caol, 69. nam Braithrean, 80. Rutile in biotite, 12, 79. Saint Columba, 1. Oban, 1. Sandstones of Torridonian formation, 14. Old Red, as erratics, 6, 60, 61. Soalasaig, 3, 6, 8, 13, 15, 16, 20. 26, 27, 33, 34, 47, 67, 69, 74. and Kilchattan Road, 13. Scoor, 75, 81, 93, 95, 97, 98, 100. Shabpe, D., 53. Shatter-belts, 2-3. Shearing at margin of Lewisian Gneiss, 11. Sheets, Intrusive, of Colonsay and Oronsay, 8, 38-41, 50. of Ross of MuU, 76, 77, 86-87, 88, 89, 92-96. Shell sand, 1, 73, 98. mound, 6. Siliceous Torridonian mudstones, 15. SilUmanite in buchite, 93, 94, 95, 104. gneiss, 85, 104. SiUs of hornblende schist, 77, 80, 83. of lamprophyre. See " Sheets." Skye, Comparison vrith Torridonian rocks of, 8, 16-17. Sliabh Rhiabhach, 23, 48, 60, 63. Sloe nan Sgarbh , 67. Slochd a' Mhuilt, 82. Soa, 75, 77, 86, 89. Solution, Incorporation of sedimentary material by, 8, 29-30. Sound of Erraid, 84, 97. Spinel in metamorphosed gneiss, 85. Staosnaig PhylUte Group, 18, 27. StauroUte, 83. Stevenson, Alan, 99. Storm beaches of Colonsay and Oronsay, 65-70. Strain-slip cleavage, 53-36. Stratigraphy of Moine schists, 80-82. — : — of Torridonian sediments, 13-27. Striae in Colonsay and Oronsay, 57—59. in Ross of Mull, 97. Structure of Colonsay, General account of, 47, 48. Submerged forest in Loch Fada, Colonsay, 72. rook basin, 6, 57-59. Succession of Torridonian sediments, 18-19. Syenite, 8, 28-31, 36. 51-53, 74. Syncline of Kiloran Bay, 3, 20, 47. Tachyutic selvages of dyke, 88, 93. Teall, Db. J. J. H., 32, 78, 81, 85. Tectonics of Colonsay and Oronsay, 47, 56. Temperature of Colonsay, 3. of Ross of Mull, 75. Tertiary intrusions of Colonsay, 7, 38, 41-46. intrusive sheets of Ross of Mull, 76, 88, 92-96. Tesohenites, CoBiparison with crinanites, 43. Thermal metamorphism, 16, 85-86, 87, 93-95, 96. Thickness of Torridonian sediments of Colonsay, 18. TonaUte, 77, 84. Topography of Colonsay, 1-3. of Ross of MuU, 75. Torr Fada, 85-97. Mor, 84, 98. Torran Rocks, 75, 78. Torridonian of Colonsay, 7, 8, 13-27. Correlation of, 8, 16-18, 77. Lateral variability in, 8, 13, 19, 22-27. Metamorphism in, 6, 8, 15-16, 49-56. Pebbles in. 8, 10-11, 14, 18, 24. Possible floor to, 10. Index. 109 Torrid onian of Colonsay, Kock types in, 8, 13-15. of lona, 77. Torrnaoh Mor, 62. Tourmaline, 83, 100. Triiigh BhJin, Colonsay, 30, 66. Mull, 83, 95, 104. an Sgoir, 80, 81, 82, 92, 93, 95, 96, 98, 104. nam Barca, 3. Trees, Distribution of, 3-4, 100., Submerged stumps of, 72. Troueh on sea bottom east of Colonsay, 57—59. Turnicil, 13, 22, 60, 66, 104. Twenty - five foot raised beach of Colonsay and Oronsay, 5, 9, 66, 68- 72. of Ross of Mull, 97-98. Uamh Ur, 66, 70. Uisken, 75, 100. Uragalg, 9, 15, 38, 62, 66, 68, 74. Van Hise, C. R., 53. Variability of Torridonian rock groups, 19, 22-27. Vogesite dykes and sills, 7, 8, 39-41, 77, 86, 87. associated with kentallenite, 33. Volcanic pipe or fissure, 35. West Reep, 80. Williams, J. F., 44. Wind in Colonsay, 3. Weight, W. B., 9, 41, 47, 98. Xenoliths in dolerite sheet, 92, 93. in granophyre, 96. in igneous rocks of Colonsay, 8, 29- 30, 39, 40-41. ZlECON, 79. ,is of the Geological Survey of Scotland — continued. ,mess-shire, Elginshire, Banffshire, Aberdeenshire (parts of). 19s. Zd. .Aberdeenshire, Kincardineshire (parts of). 13s. Zd. ' ,i. Aberdeenshire, S.E. part; Kincardineshire, N.E. corner. 8s. ' '81. Eoss-shire, S.W. ; Islands of Raasay and Bona, partof Skye. 14s. 9d.- 85. Elginshire, Banffshire, Aberdeenshire (parts of). 20s. Sd. . 87. North-East Aberdeenshire and BanflFshire (detached portions). 13s. 3d. 91. Ross-shire (part of). 15s. 6d. 94. Rossrshire, Cromartyshire, Sutherlandshire (parts of). 8s. M. 95. Elginshire, Banffshire (parts of). 5s. Sd. 96. Aberdeenshire, Banffshire (parts of). 7s. Sd. 97. Aberdeenshire and Eastern Banffshire (parts of). 4s. 6d. 100. N.W. Eoss-shire (part of). 2s. 6d. 101. Eoss-shire, Cromartyshire, and Sutherlandshire (parts of). 14s. 103. Sutherlandshire, E. 8s. 9d. 107. Sutherlandshire (part of). 25s. 3d. 113. Sutherlandshire, K.W. (part of). 5s. Sd. 114. Sutherlandshire (part of N.). 13s. 3d. 115. Sutherlandshire, Caithness-shire (parts ot). 19s. Sd. Maps on Six-inch Scale, illustrating the Coal-fields. Edinburghshire. Haddingtonshire. Fifeshire. Ayrshire. Renfrewshire. Lanarkshire, Sheets 1, S.W. (and iB, S.E.), S.E. ; 3, K.W., N.E., S.W., S.E. ; 4, N.W., N.E., S.W., S.E. ; 4a., S.W.; 5, N.E., S.E. ; 7, N'.E., S.E. ; 8, N.W., N.E., S.W., S.E. ; 9 (and Haddington 14), N.W., S.W., S.E. ; 11, N.W., S.W. ; 12, S.E. ; 13, N.W., N.E., S.W„ S.E. ; 14, N.W.; N.E., S.W., S.E. ; 15, N.W. (and Haddington 18, N.W.) ; 18, N.E. ; 19, N.W., N.E. (revised editions). Is. 6d. (unooloured). Sheets 4, S.W., S.E. ; 9, N.W., N.E., S.W., S.E. ; 14, N.W., N.E., S.W., S.E. ; 18, N.W. (revised editions). Is. 6d. (uncoloured). Sheets 24, 6s. Sd. ; 25, 4s. Sd. ; 30, 31, 8s. Sd. ; 32, 37, 5s. ; 33, 2s. ; 35, 7s. ; 36, Us. Sheets 7, 12, 13, 17, 19, 22, 23, 24, 28, 30, 36, 42, 46, 52, 8s. Sd. ; 8, 40, 8s. ; 9, 4s. 3d. ; 11, 18, 29, 34, 41, 9s. 9d. ; 16, 6s. 3d. ; 26, 27, 31, 5s. 9d. ; 33, 7s. ; 35, lis. ; 47, 7s. 9d. ; 50, 10s. 3d. ~ - - " 8 and 9, 13, 5«. ; 12, 8s. 3d. ; 14, 17, 3s. 9d. (old editions). 13, S.W. (revised editions). Is. 6d. Sheets 7, 11, 9s. 9d. 3s. ; 15, 16, 7s. Sheets 12, N.E.,S.E. (uncoloured). Sheets 1, 3s. 3d. ; 24, 7s. ; 11, 17 25, 32, 7s. 9d. 3, 3s. 9d. ■ 4, 2s. 9d. ; 5, 2s. Sd. ; , 20, 31, 38, 41, 42, 8s. 3d. ; 8, 23, 9, 19, 6s. Sd. ; 12, 13, 16, 5s. 9d. ; 10, 5s. ; 18, 37, 9s. ; 49, 4s. 3d. (old editions). Sheets 1, S.E. ; 2, S.W., S.E. ; 6, N.W., N.E., S.W., S.E. ; 7, N,W„ N,E,, S,W„ S.E. ; 8, N.W. ; 10, N.W. (revised editions). Is. 6d. (uncoloured). i 1, 3s. ; 5, 7, 5s. 9d. ; 6, 8s. Sd. Sheets 19A, 20, 24, 2s. Sd. ; 23, 7s. 9d. ; 25, 6s. Sd. ; 26, 3s. 3d. ; 28, including 29, 3s. Sheets 17, 28, 8s. Sd. ; 18, 31, 4s. 3d. ; 23, 27, 9s. ; 24, 5s. 9d. ; 25, 2s. 9d. ; 29, 7s. ; 30, 32, 5s. ;,33, 36, 2s. 3d. ; 35, including 34, 6s. Sd. (old editions). Sheets 27, S.E. ; 28, S.W. ; 32, N.E. ; 33, N.W. (revised editions). Is. 6d. (uncoloured). Sheet 8, 5s. 9d. (old edition). Sheets 1, N.E., S.E. ; 2, S.W. and N.W., S.E. ; 3, S.W. ; 6, N.W., N.E., S.W., S.E. ; 7, N.W. and S.W. ; 9, N.E., S.E. ; 10, N.W., S.W. ; 12, N.E. (revised editions). Is. 6d. (uncoloured). Sheets 133, 9s. 9d. ; 134, 140, 8s. Sd. ; 135, 5s. ; 139, 142, 3s. ; 141, 2s. 9d. ; 143, 23. Maps on Six-inch Scale, illustrating Structure of N.W. Highlands. Sutherlandshire. Sheets 5, lis. ; 71, 20s. 3d. Skye Sheets 38, 20s. ; 39, 16s. ; 44, 22s. 9d. ; 45, 24s. Maps on the Scale of Four Miles to One Inch, Sheet 12. Perthshire, Forfarshire, Kincardineshire, &c. ,, 16. Galloway and part of Ayrsliire, J^e. „ 17. Eoxbur^shire, Dumfriesshire, &c. Printed in Colours. ^2s. 6d. each. Horizontal Sections. Sheet 1. Edinburghshire and Haddingtonshire. 3s. 9d. ,, 2. Edinburghshire, Haddingtonshire, Berwickshire. 4s. 3d. ,, 3. Peeblesshire, Edinburghdiire, Linlithgowshire. 4s. 3d. ,, 4. Ayrshire Coal-fields (west side). 4s. 8d. , 6. Ayrshire Coal-fields (east side). 5s. 9d. V Dumfriesshire. Dumbartonshire. Stirlingshire. Linlithgowshire. Perthshire. List of Publications of the Geological Survey of ^», ?,,, Sheet 6. Renfrewshire, Dumbartonshire, Ayrshire. 5s. 9a. ■ ,, 7. Cheviot and Lanimermoor Hills. 5s. > , , 8. Clyde Coal-field and Campsie Hills. 5s. ,, 9. Ayrshire Coal-fields (Miiirkirk and New Cumnock). 5s. 9d. Vertical Sections. Is. 8d. p^r Sheet. Sheet 1. Edinburgh and Haddington Coal-fields. ,, lA. Midlothian Coal Basin (middle portion). ,, iB. „ ,, ,, (western side). „ Id. ,„ ;, „ (eastern „ ). ,, iB. East Lothian Coal Basin. „ 2. Fife Coal-fields. „ 2a. Kfe Coal-fields. ,, 3. Kilmamook Coal-field. ,, 4. Clyde Basin Coal-field. ,, 5. Stirling and Clackmannan Coal-fields. ,, 6. Muirkirk, Lesmahagow, and Douglas Coal-fields. ,, 7. La.narkshire Coal-fields (Rutherglen and Carluke). Memoirs of the Geological Survey of Scotland. (1.) Genbkal Memoik : — Silurian Rocks of Britain. Vol. I. Scotland. 15s. (2.) Economic Memoir: — The Oil-Shales of the Lothians. 4s. (3.) Museum Guide ; — Guide to the Collections of the Geological Survey. Id. (4.) DisTKiCT Memoiks : — Cowal, Argyllshire. 6s. The Geology of the Neighbourhood of Edinburgh. (Second edition.) 7s. 6d. East Lothian. (Second edition. ) 4s. 6d. Central and "Western Fife and Kinross. 5s. 6d. Eastern Fife. 8s. Skye, the Tertiary Igneous Eocks of. 9s. Islay, &c. 2s. Sd. The Geological Structure of the North-West Highlands of Scotland. 10s. 6d. Palaeontology. The Higher Crustacea of the Carboniferous Rooks of Scotland. 4s. (6.) Sheet Memoirs : — Sheet 1. 'Wigtownshire, South-Western Districts. 3a!. ,, 2. Wigtownshire, South-Eastern Districts. Sd. „ 3. Wigtownshire, Western Districts. 3d. ,, 4. Wigtownshire, E. part; Kirkcudbrightshire, portion of S.W. Division. 9d. ,, 5. Kirkcudbrightshire, Southern Districts. Is. &d. ,, 7. Ayrshire, South- Western District. Zd. ,, 9. Kirkcudbrightshire, N.E. ; Dumfriesshire, S.W. Is. ,, 13. Ayrshire, Turnberry Point. Zd. ,, 14. Ayrshire, Southern District. Zd. ,, 15. Dumfriesshire, N.W. ; Ayrshire, S.E. ; and Lanarkshire, S. Zd. ,, 21. Buteshire (Arran, Central, and N. and S. Bute), Argyllshire, Ayrshire (parts of). 4s. ,, 22. Ayrshire, Northern District, and parts of Renfrewshire and Lanark- shire. 3i^. ,, 23. Lanarkshire, Central Districts. Is. ,, 24. Peeblesshire, Lanarkshire, Edinburghshire, Selkirkshire (parts of). Zd. ,, 31. Lanarkshire, N. ; Stirlingshire, S. ; Linlithgowshire, W. 2s. 3d. ,, 34. Eastern Berwickshire. 2s. ,, 35. Colonsay and Oronsay, with part of the Ross of Mull. 2s. 3d. „ 36. Seaboard of Mid Argyll. 2s. Zd. ,, 37. Mid Argyll. 3s. -„ 45. Argyllshire, The Country near Oban and Dalmally. 2s. 6d. „ 55. Perthshire, The Country round Blair- AthoU, Pitlochry, and Aberfeldy. 3s. „ 60. The Small Isles of Inverness-shire, is. 6d. „ 70. Inveruess-shire. West-Central Skye, with Soay. Is. ,, 71. Inverness-shire. Glenelg, Loohalsh and the South -East part of Skye. Zs..Sd. ,, 75. Inverness-shire, Elginshire, Banffshire, Aberdeenshire (parts of). Is. 6d. ,, 76. Aberdeenshire, Central. Is. ,, 85. Elginshire, Banffshire, Aberdeenshire (parts of). Is. 6d. ,, 87. Aberdeenshire and Banffshire (parts of). 9d. „ 97. Northern Aberdeenshire. Eastern Banffshire, id. (6.) Catalogue of Geological Photographs. Series B and C. 6d. (7.) Description of Arthur's Seat Volcano. 6d. A detailed Catalogue may be had on application to Messrs. W. & A. K. Johnston. LId. 2 St. Andrew i'-yr .',-7:i"""M;;.~ ' ^% *^^^ r'' 'V*f^V '\ Z"*-,^: c/;v