The GEOLOGY of the SHIANT ISLES (HEBRIDES)
 The Lower Sill of Garbh Eilean.
The Jurassic beds of Garbh Eilean are divided and underlaid by another crinanite-sill. It is probable that only a small portion of the top of the sill is exposed, and that the lower part is under water. If its course be traced above water from east to west, it is first seen forming the conspicuous promontory of Bidean a'Roimh at the extreme east of Garbh Eilean, and has here a maximum visible thickness of about 100 feet. It shows columnar structure at this point, but not to the same extent as the sill above. At the west of the promontory the sea has eroded out a very fine natural arch, known as Toll a'Roimh, along a major joint, while immediately west of the arch the Jurassic strata make their appearance. From this point the top of the sill descends in a westerly direction until it forms the skerries of Sgeirean a'Bhaigh, where, still overlaid by Jurassic beds, it may be followed off-shore for a quarter of a mile to the west. The upper contact of the sill is seen in the skerries, and shows very little chilling. It is, however, sharply defined. At the east of Sgeirean a'Bhaigh the igneous rock divides the Jurassic strata into two portions, but is otherwise non-transgressive, conforming its dip to the sedimentary deposits.
The rock of the lower sill is fine-grained and dark grey (with a tinge of green) - the colourless olivine being more serpentinized than in the crinanite of the upper sill, and 'ophimottling' of the augite less conspicuous. Microscopic examination confirms these differences, which are, however, of degree only.
 The Upper Sill of Eilean Mhuire.
Above the Jurassic strata of Eilean Mhuire there is a sill of crinanite, now almost entirely 'removed by denudation, but of which a fragment forms the highest part of the island. Exposures are poor, and are limited to a few large blocks projecting from the grassy slopes at, and immediately below, the summit.
In the hand-specimen, the rock has a somewhat troctolitic appearance, the ferromagnesian minerals standing out as black spots in a matrix of greasy-looking felspathic minerals. Under the microscope, the rock is seen to be a crinanite not unlike that of the lower sill of Garbh Eilean, but containing only 4 per cent. of partly serpentinized olivine and as much as 5 per cent. of analcite and natrolite.
The lower sill of Eilean Mhuire, which occurs underneath the Jurassic strata of that island, is the most complex of the whole group. It is, moreover, excellently exposed, although many of the sections are quite inaccessible, being on vertical cliffs. The visible thickness of the intrusion is about 200 feet, but a great part is doubtless below water.
Along the shore-line the sill is generally quite easy of access, by boat at least; but, except at the eastern extremity of the island, sheer cliffs tower above the rocky shore-platform. Immediately east of Bid na Faing, however, grassy slopes are intermingled with the cliffs, and provide relatively accessible exposures. At the eastern part of the island, which is joined to the main mass by a narrow and lofty neck of rock, the cliffs are much less steep, andl may be scaled with comparative case, and it is a stroke of good fortune that the most interesting rocks appear to be concentrated here. Columnar jointing is not conspicuous, but the joint-planes are usually quite well marked. The most striking feature of the sill is, perhaps, its stratiform appearance at the eastern end due to the injection (after solidification) of horizontal basaltic sheets. The sill itself, indeed, seldom departs from a horizontal position. By far the greater part of the intrusion consists of crinanite, and it is only east of Bagh Chlann Neil that other types are seen. At several points on the north-eastern and north-western cliffs the upper contact appears to be exposed, but it is invariably quite inaccessible. Elsewhere the contact is grassed over.
While the crinanite is a medium-grained dark rock, both on the weathered surfaces and on freshly fractured specimens showing typical '’ophimottling', some of the types of the east of Eilean Mhuire are of very different appearance. A coarse subophitic analcite-dolerite or gabbro, allied to teschenite and comparatively rich in analcite, forms the lowest visible part of the sill from the neck eastwards. In this rock the augite- and felspar-crystals give to it a mottled black and white appearance. This modification becomes even richer in analcite in an upward direction, passing into true teschenite, and, some 20 feet above sea-level, the teschenite gives place to a coarse-grained crumbling rock which is more leucocratic and obviously rich in alkali-felspar and zeolites, being,, in fact, a syenite. The syenite persists upwards until a height of about 90 feet above sea-level is reached, and is then followed by a coarse decomposed dolerite which caps the eastern eminence of the island.
All these modifications contain schlieren, especially the svenite, which is also drusy and riddled with felsic segregations of irregular shape and disposition. These latter may attain quite a considerable size, and as a rule are distinctly fresher than the crumbling, syenite. It was probably either in the syenite, or in its segregations, that Heddle detected nepheline. The schlieren have comparatively sharp junctions, which are, however, unchilled, and in the teschenitic types a rude parallelism of the minerals may be observed. The junctions between the various types are also unchilled and inconspicuous.
The syenite. and the teschenite are both cut by several horizontal sheets of basalt which may reach a thickness of 20 feet, and can be traced for 200 or 300 yards along the cliffs. This basalt is a fine-grained, black, non-porphyritic rock, that shows marked chilling against both teschenite and syenite, being obviously of considerably later date. Neither basalt nor syenite is seen much to the west of the rocky neck, and both probably die out near Bagh Chlann Neil. The sea-cave mentioned by Heddle in his record of nepheline is doubtless one of the many rifts in the rock caused by wave-erosion along major joints.
Crinanite.-The crinanite of the lower sill presents few striking features, being very similar to that of the lower sill of Garbh Eilean, except for the granularity of the olivine and the conspicuous part played by natrolite.
Teschenite or analcite-dolerite. - The teschenite or analcite-dolerite of the eastern end of the island is different in texture from any of the crinanites. If we take the less alkaline type first, the olivine is seen under the microscope to be much less abundant than in the crinanites, comprising only 4 per cent. of the rock. It occurs in well-shaped crystals averaging 0.6 min. across, and, although generally altered to green or brown serpentine, is colourless when fresh. Occasionally it is penetrated by the ends of plagioclase-laths.
The pyroxene is abundant, and its optical properties are almost identical with those of the crinanite of the picrite-crinanite sill of Garbh Eilean. It is, however, subophitic in habit, forming stout prisms up to 3 mm. across, which are usually penetrated in every direction by plagioclase -laths, although they sometimes exhibit crystal boundaries. The 'ophimottling' of the crinanites is conspicuous by its absence, but zoning of the hour-glass type is common.
The felspar is almost entirely zonal plagioclase, varying from medium labradorite (Ab2An3) at the centre of the crystals to oligoclase at the margin. It occurs in well-formed elongated laths up to 1.5 min. in length, which penetrate both augite and olivine, and make up nearly 60 per cent. of the rock. These laths are sometimes, to a slight extent, replaced by analcite along cracks, and even in small patches. Analcite, besides occurring as a replacement product of plagioclase, is, as one might expect, quite abundant interstitially, the usual proportion being about 8 per cent., and it is generally unaltered. Natrolite, however, was not recorded in this type.
The iron-ore is mainly titaniferous magnetite or ilmenite, but pyrites occurs to a limited extent, the two minerals comprising 7 per cent. of the rock. Small needles of apatite form, as usual, an ubiquitous accessory. The specific gravity of the rock is 2.97, and the mode is given below.
A gradual upward transition is seen from analcite-dolerite into teschenite. This is due to a considerable increase in the proportions of analcite and acid plagioclase and a dwindling almost to vanishing point in the proportion of olivine, although a good deal of serpentine may he observed in the rock. The augite in the teschenite assumes an idiomorphic columnar habit, the crystals measuring up to 1 cm. in length and 2 mm. in breadth, some of them showing pronounced elongation in the direction of the c-axis. Although they usually show good crystal boundaries, they may, on rare occasions, be penetrated by the ends of plagioclase-laths. The pleochroism of this pyroxene is slightly different from that of the analcite-dolerite, agreeing well with that given by Dr. A. Harker for the augite of a gabbro from Eilean Mhuire that is: X lemon-yellow, Y purplish-brown, Z lemon-yellow. Zoning with stronger tints in the interior is very frequent in the augite, which may assume a faint greenish tinge at the margin, through the presence of a small proportion of the aegirine molecule.
The felspar of the teschenite, like that of the analcite-dolerite, is almost entirely plagioclase ; but, although the centres of the crystals are still medium labradorite (Ab2An3), the proportion of oligoclase at the margin is decidedly greater. Much of the plagioclase is replaced in a patchy fashion by analcite, or altered to turbid brown decomposition-products. The well-formed laths are larger than in the analcite-dolerite, but do not show the same elongation. The analcite occupies large areas between the felspars, amounting to about 12 percent. of the rock. It is perfectly clear, and is sometimes penetrated by needles of apatite or flakes of biotite, a circumstance which seems to indicate a primary origin for the zeolite.
The iron-ore is mainly ilmenite in skeletal crystals, occasionally altered to leucoxene, and frequently fringed by a reaction-border of biotite; but pyrites also has been observed. Apart from the olivine and analcite, the proportion in which the various minerals occur is the same as in the analcite-dolerite.
In some varieties of the teschenite there is distinct parallelism of the minerals, and it seems probable that the gabbros described by Dr. Harker are simply modifications of teschenite poor in analcite.
Upper dolerite. - The upper dolerite, which caps the eastern extremity of Eilean 11huire, is somewhat different in texture and mineralogical composition from the analcitic types in the lower part of the sill. It is a coarse, crumbly, decomposed-looking rock in the hand-specimen, but is seen under the microscope to be in fair preservation. Purplish-brown zonal augite, similar in its optical properties to the pyroxene of the teschenite, but showing inclusions of magnetite-dust, is by far the most abundant constituent of the rock, amounting to 50 per cent. It occurs in large ophitic plates tip to 1 cm. across, which do not, however, exhibit ‘ophimottling', and have often distinct crystal boundaries. A very few shapeless serpentinous pseudomorphs probably indicate the former presence of olivine, but the proportion of this mineral cannot have exceeded 2 per cent.
Elongated laths of zonal plagioclase, mainly (Ab2An3), make up nearly 40 per cent. of the rock, and are often enclosed in augite. They are usually about 1 mm. long, and in many cases are stained by migratory decomposition-products. The interstices between the felspars are frequently occupied by a decomposed brownish mesostasis (4 per cent.) which contains clear analcite areas, and sometimes quartz, the latter mineral being probably of secondary origin. Large skeletal ilmenite (6 per cent.) and small apatite- needles make up the remainder of the rock.
Dr. H. H. Thomas, having seen my sections, commented on the resemblance of this dolerite to the Ben Hiant type of quartz-dolerite so well represented in Ardnamurchan.
The junctions of both dolerite and teschenite with the central syenitic portion of the sill were not detected, either in the field or in thin section, and it is probable that they are both gradual, although the transition must be complete within a few feet in each case.
Syenite - A layer some 60 feet thick in the centre of the sill is formed by the syenite, which, like the upper dolerite, is a coarse, brownish, crumbly rock apparently much altered. This decomposition is seen also in some thin sections, whereas others are surprisingly fresh.
The chief constituent of the rock is alkali-felspar, which is frequently unaltered. It consists partly of oligoclase (Ab4An1) and partly of soda-orthoclase. The plagioclase either forms well-shaped elongated laths up to 1 cm. in length showing albite, Carlsbad, and pericline twinning, or occurs as the core of a crystal surrounded by soda-orthoclase. It is usually zonal, the most basic variety being andesine, while the more acid parts approach albite in composition. Alteration to brown decomposition -products is more prevalent than in the soda-orthoclase. The latter felspar, besides forming broad fringes round crystals of oligoclase, occurs as untwinned, or simply twinned, tables equal in size to the plagioclase-laths towards which it is allotriomorphic. The mean refractive index was found by oils to be 1.529±.003, and the mineral is optically negative. It has an abnormally low optic axial angle, which, from the marked curvature of the isogyre, was estimated to be about 30°. The two felspars together make up 50 per cent. of the rock.
Pyroxene comes next in order of abundance, the commonest variety being a pale purple titanaugite, which shows hour-glass zoning to a slight extent. This mineral is usually surrounded by a fringe of grass-green aegirine-augite, or by deep-green aegirine, and sometimes contains cracks bordered by these alkali-pyroxenes. It forms stout prisms up to 6 mm. in length, which allows a certain amount of idioimorphism, but are sometimes penetrated by the extremities of plagioclase-laths. The aegirine-augite and aegirine may also form separate prisms, idiomorphic towards all minerals except plagioclase, and seldom exceeding 1 mm. in length. Some of the prisms of aegirine are associated with small crystals of blue arfvedsonite, which also occurs very sparingly in small allotriomorphic crystals. These metasilicates comprise about 20 per cent. of the rock.
Zeolites, of which the chief is analcite, are also very common, occupying large interstitial areas. In these patches the analcite is unaltered, and was the first zeolite to crystallize, for it invariably presents idiomorphic boundaries towards the others. The nature of these other zeolites is a little uncertain, but refractive-index tests with oils show that a variety with low double refraction and =1.498 ± .003 occurring in minute divergent fibres is probably scolecite : stilbite also occurs in spheroidal masses of radiating fibres. The zeolitic content of the rock approaches 20 per cent.
The syenite also contains about 3 per cent. of dark-brown pseudomorphs measuring tip to 2 mm. across, and presenting frequently a rectangular or hexagonal outline, the shape being strongly suggestive of the former presence of nepheline. Skeletal ilmenite in large crystals is abundant, and often shows reaction borders of biotite. Apatite is conspicuous, too, in needles measuring up to 1.5 mm. in length, usually coloured bluish grey by inclusions.
A large amount of serpentinous matter derived from the decay of ferromagnesian minerals is irregularly distributed in the syenite, and it is possible that olivine was originally present. This is particularly the case in an essexitic modification of the syenite which occurs at the top of the neck, and contains a fair amount of plagioclase as basic as AblAnl.
The mode of the rock is given below, but it was not considered fresh enough for analysis.
Greyish-white segregations are common in the syenite, and are of irregular shape and size, the maximum breadth observed being about 2 feet, while the distribution is apparently haphazard. The rock is slightly finer in grain than the normal syenite, and distinctly richer in felsic constituents. It is also visibly porous and crumbly, containing numerous druses lined with analcite and spheroidal aggregates of radiating fibrous zeolites.
Under the microscope the rock is seen to consist of the same minerals with the same habit as the normal syenite, but in different proportions. The felspars, which amount to over 60 per cent. Of the rock, have the same habit as in the normal syenite, but the plagioclase is nearly all albite with positive sign and very minute multiple twinning. The mean, refractive index by oils was found to be 1.534 ± .003 Both the albite and the soda-orthoclase are usually quite fresh, and their relative abundance is exceedingly variable.
Pyroxenes form only 10 per cent of the rock, aegirine and aegirine-augite together being as abundant as titanaugite, while the optical properties of all three varieties are the same as in the normal syenite. Alkaline amphiboles comprise about 1 per cent, .and. besides arfvedsonite with deep-blue to lavender pleochroism, occasional crystals of riebeckite may be detected which vary in tint from dark blue to greenish-yellow, while barkevikite with brown-pale yellow pleochroism is still more rare. All three minerals occur in intimate association, and frequently form crystal aggregates with aegirine. The habit of these alkali-pyroxenes and amphiboles is, indeed, very similar to those described by Dr. H. H. Thomas in the alkali-syenite of Gamhnach Mor in Mull.
The zeolites of the segregation veins do not differ notably in occurrence, constitution, or abundance from those of the normal syenite. Nepheline, however, is distinctly more common, and occasional patches of the unaltered mineral may be detected in the dark-brown pseudomorphs. The mineral is seen to have a slightly higher refractive index than the albite, straight extinction, negative optical character, and one good cleavage accentuated by decomposition-products. It alters locally to a mineral with slightly higher refractive index, positive sign, and double refraction slightly above that of quartz, which resembles closely a mineral associated with nepheline, described by Dr. G. W. Tyrrell and Prof. E. B. Bailey. The iron-ore is distinctly less abundant than in the normal syenite, and is probably titaniferous magnetite, as the skeletal habit is rare; but apatite is again conspicuous.
Both the normal rock and its segregations considered by themselves are analcite-syenites, but the occurrence of nepheline renders them unique in British petrographic literature. Nepheline-syenites with analcite from various localities have been described, principally by Prof. A. Lacroix, but these rocks tend to be richer in the felspathoid than in the zeolite, which is not usually primary. The habit of the zeolites in the Shiant rocks renders it extremely probable that they were the last products of solidification of a sodic magma, and they are therefore considered to be primary, like the zeolites of the analcite-syenite of Howford Bridge. The segregation-rock, being reasonably fresh, was chosen for analysis, and the results are tabulated below, together with the mode and a discussion of the affinities of both syenitic types.
Olivine-basalt - The sheets of this rock which cut the eastern end of the sill have been described microscopically by Dr. A. Harker, whose account is confirmed by my own observations in almost every particular. The unchilled rock is a virtually non-porphyritic olivine-basalt, containing as rare corroded phenocrysts only olivine and labradorite up to 4 mm. in length. Elongated laths of labradorite (Ab2An3), measuring up to 0.5 mm. in length, comprise the bulk of the rock: greenish-brown patches of subophitic augite, rounded grains of unaltered olivine averaging 0.2 mm. in diameter, a small quantity of mesostasis or analcite, and minute cubes of magnetite making up the remainder.
Towards the margin of the sheets the rock becomes much finer in grain, and contains a good deal of devitrified glass. The specific gravity, the mode, and the analysis of the normal basalt are given below, while a microphotograph appears here.
A. Analcite-dolerite, sea-level, extreme east of Eilean Mhuire. (Photo below.)
B. Analcite-syenite, 60 feet above sea-level, on the cliff at the eastern end of Eilean Mhuire.
C. Acid segregation in syenite, 25 feet above sea-level, on the cliff at the eastern end of Eilean Mhuire.
D. Basalt-sheet, 20 feet above sea-level, on the cliff at the south-eastern end of Eilean Mhuire (Photo)
The modes; of A, B, & C represent the average type at each locality, and it is possible in all three cases to find schlieren of distinctly more felsic or more mafic composition.
*The total given has been subjected to a deduction of 0.04 per cent. 0 for sulphur, and 0.81 per cent. for moisture calculated before analysis.
I. Acid segregation in syenite, 25 feet above sea-level, on the cliff at the eastern end of Eilean Mhuire; analyst, E. G. Radley.
a. Syenite, Gamhnach Mor, Carsaig Bay (Mull) ; analyst, E. G. Radley. Quoted from the Mull Geol. Surv Memoir, 1924, p. 27.
b. Analcite-syenite, Howford Bridge, Mauchline; analyst, M. Dittrich. Quoted front G. W. Tyrrell, Q. J. G. S. vol. lxxxiv (1928) p. 559.
II. Olivine-basalt Sheet, in the cliff at the extreme south-east of Eilean Mhuire, 20 feet above sea-level ; analyst, N, Sahlbom (photo).
e. Average plateau-basalt magma-type. G. W. Tyrrell, loc. supra cit.
The foregoing analysis of the syenitic segregation-material of Eilean Mhuire brings out clearly its very sodic nature. Reasonable agreement is seen between the Shiant analysis and one of the Gamhnach Mor alkali-syenite in Mull, quoted alongside; but it will he observed that the Shiant rock is the more alkaline, probably owing to its high analcite content, the Mull rock containing neither nepheline nor analcite. Both these Tertiary syenites are distinctly more acid than the typical analcite-syenite of Howford Bridge, which is probably of Permian age, and the analysis of which is also quoted. An analcite-syenite from Emery (Utah) is, however, still more basic.
The analysis of the olivine-basalt which was made for me by Dr. N. Sahlbom shows very close agreement with the average composition of the Tertiary plateau magma-type as calculated by Dr. Tyrrell. Moreover, no agreement is shown with any analysis of a British Tertiary rock that has not been referred to this magma-type: for instance, any type of the tholeiites.
The crumbly and porous nature of the syenites renders any determination of specific gravity untrustworthy, but 2.50 was the mean of several widely differing estimations.
The mode of differentiation of the lower sill of Eilean Mhuire must have been quite unlike that of the crinanite-picrite sill of Garbh Eilean, the history of the cooling of which is so clear. In the western, or main, portion of Eilean Mhuire there does not appear to be any trace of gravitational differentiation, the crinanite of the lower sill retaining practically the same mineralogical composition from top to bottom of the cliffs; while in the eastern portion there is actually a definite concentration of augite in the coarse, slowly-cooled top of the sill. (In the following discussion the later olivine-basalt sheets will be neglected.) Thus, sinking by gravity of early-formed crystals was not a cause of differentiation.
There remain three hypotheses open to account for this differentiation:-
(1) The intrusion of an already heterogeneous magma.
(2) The subsequent intrusion of the syenite into the partly cooled analcite-dolerite and teschenite.
(3) The injection of residual magma during crystallization into rifts in the crystal-mesh.
The first hypothesis has been applied by Dr. A. Harker to account for the banded gabbros and peridotites of Skye, and by Campbell & Stenhouse to account for the various modifications in the picrite-teschenite sill of Inchcolm. The recent work of Greig on immiscibility (or better, limited miscibility) in silicate-melts shows that the occurrence of a natural unhomogeneous silicate liquid is most unlikely. There is, however, in the case of the Eilean Mhuire sill, the remote possibility of the simultaneous intrusion of two viscous magmas of widely differing composition, which, on account of their increasing viscosity through cooling, had not time to mix perfectly. But crystallization in such viscous magmas would almost undoubtedly have been far advanced before intrusion, resulting in a much greater development of flow-structure and hybridism than is actually seen in the sill. The first hypothesis is, then, improbable.
Turning to the second, we find that, while the junctions between syenite, dolerite, and teschenite do not altogether preclude this hypothesis, there is no veining of the basic rocks by the more acid, or vice versa, and this absence of a feature which might well be looked for, if the sill was intruded in two stages, is significant.
The third hypothesis was suggested by Bowen to account for differentiation of banded sills and laccolites, and has recently been developed by Prof. Bailey and by Dr. Tyrrell. Bowen considers that the 'auto-intrusion', as he very appropriately terms it, is due to deformation of the masses by lateral thrust during crystallization ; but Tyrell, applying the hypothesis to the crinanite-analcite--syenite sill of Howford Bridge and other unhomogeneous sills in Central Scotland, attributes the auto-intrusion to the pressure exerted on the residual magma by the superincumbent rock-column, while Bailey considers it to be due to the expansive force of volatile constituents prior to the final consolidation of the rock.
The resemblance, in the mode of occurrence of the various types, of the lower sill of Eilean Mhuire to the much thinner one of Howford Bridge is noteworthy; but Tyrrell's development of Bowen's hypothesis, although exceedingly simple and ingenious, presents difficulties of application in certain cases. Why, for instance, is there no sign of segregation-veins in many of the dolerite-sheets of Skye which are thick, of similar composition to the Shiant sills, and must have been subjected to very considerable pressure by the superincumbent rock in the case, at least, of the lower members? It is also obvious that some other reason must be found for the occurrence of segregation -veins in the plateau-lavas of Mull,the pressure from the superincumbent rock being in this case extremely small.
I am inclined to attribute the main differentiation of the Eilean Mhuire sill into alkaline dolerite and syenite to auto-intrusion, due either to deformation during cooling, or to the pressure exerted by the superincumbent rock-column - the evidence for and against either process being inconclusive. In the case of the later segregations of syenitic composition, however, the drusy nature of the syenite seems to give good support to Prof. Bailey's suggestion, and the felsic segregations are then thought to be due to the development of cracks in the crystal framework in the manner outlined by him, and to the infilling of these by residual magma. This hypothesis may also account for the pegmatite-veins in the crinanite of Eilean an Tighe.
The reaction series to be observed in the lower Eilean Mhuire sill are practically the same as those made out in the picrite-crinanite sill of Garbh Eilean, but the continuous reaction series afforded by the plagioclase-felspars is carried a stage farther by the imperceptible passage of albite into soda-orthoclase.
The basalt-sheets are not in any way connected with the main differentiation of the sill, being clearly intruded after complete solidification had taken place.
The line of rocks known as the Galtachean appears to be a remnant of a single great sill of crinanite, perhaps to be correlated with the crinanite-picrite sill of Garbh Eilean and Eilean an Tighe. The same well-marked columnar structure is to be seen in all eleven islets; but, while the disposition of the columns indicates a constant southerly direction of dip, the angle is very variable, changing from 20° to almost 90° in places. These changes of dip are rapid, and are accompanied in the eastern islets by marked curvature of the columns. All the islets present steep rocky faces towards the north, but the southern, or dip-slopes, of the larger islets are more accessible, particularly in the case of Galta Mor. No trace of sedimentary strata is to be seen ,on any of them.
The group appears to be made up almost entirely of fresh crinanite, which in the hand-specimen bears a close resemblance to types from the upper parts of Garbh Eilean and Eilean an Tighe. In Galta Mor, however, there is a considerable development of pegmatite segregation-material - not merely in veins, but in large bands reaching a thickness of at least 10 feet. These bands are most conspicuous near sea-level in the south-west of that island, and again near the summit. They consist of a coarse grey rock, in which prisms of augite stands out clearly from a felspathic matrix.
Crinanite. - Under the microscope, the crinanite of the Galtachean is seen to be identical with that of Garbh Eilean and Eilean an Tighe. Specimens from all heights on Galta Mor show the markedly ophitic olivine typical of the upper parts of the crinanite-picrite sill; but the rock is fresher, and has unaltered olivine which is absolutely colourless, thus confirming the hypothesis that the occasional brown colour of the mineral in Garbh Eilean is due to decomposition.
Pegmatite. - Three different types of pegmatite were foundon Galta Mor. The first is an essexitic or teschenitic type which forms the summit of the island. Felspar is the most abundant constituent, consisting partly of analcitized zonal plagioclase-laths measuring up to 3 mm. in length, varying from acid labradorite to oligoclase, and partly of untwinned soda-orthoclase fringing the plagioclase or occurring as broad separate tables. The two felspars together make up about 60 per cent. of the rock, and the plagioclase occurs in ophitic relationship with large plates of zonal titanaugite, sometimes fringed with green aegirine-augite and comprising about 20 per cent. of the rock. Small crystals of aegirine occur separately, and the remainder of the rock consists of clear interstitial analcite (10 per cent.), iron-ore (6 per cent.), and serpentinized olivine (3 percent.).
At a lower point, about 100 feet above sea-level, another variety is encountered, forming small veins reminiscent of those of Eilean an Tighe. The resemblance is confirmed by microscopic examination, the two rocks being very similar indeed. There is, however, little or no brown mesostasis in the Galta Mor example.
The third variety of pegmatite forms thick bands in the lower part of the island, and is syenitic in composition, having a strong resemblance to the syenitic segregations of Eilean Mhuire and also to the syenite of Gamhnach Mor in Mull. Analcite is scarcer than in the Eilean Mhuire rock, and felspar - particularly soda-orthoclase - moree abundant. The soda-orthoclase has a distinctly larger axial angle than that from the Eilean Mhuire syenite. Nepheline is much scarcer and soda-amphiboles entirely absent in the Galta Mor rock, which is, on the whole, the freshest. syenitic type in the group. The mode is given below, and a microphotograph is reproduced here.
Mode of syenitic pegmatite from the south-west of Galta Mor.
The occurrence of the pegmatite-bands on Galta Mor is considered to be due to auto-intrusion of residual magma into the crinanite in three distinct stages during crystallization: (1) the thin teschenitic veins, (2) the essexitic bands of the summit, and finally the syenite-bands of the lower south-western part. The cause of this auto-intrusion must, as in the lower sill of Eilean Mhuire, remain doubtful, the same alternatives being offered.
There can be little question that all the igneous rocks of the Shiant isles are of Tertiary age, for in some cases they are seen to invade strata of Jurassic age, while there is an exceedingly strong genetic connexion between all types, except the olivine-basalt sheets, which are to be regarded as the latest manifestation of igneous activity. This genetic connexion is well shown by plotting the major oxides against silica on a variation-diagram (fig. 3).
The points given by the analysis of the olivine-basalt are then seen to diverge greatly from the smooth curves or straight lines yielded by the other four analyses, which are plainly all on a single line of descent from it common stock.
Although the Tertiary age of the Shiant sills does not admit of much argument, there is considerable doubt as to the stage of Tertiary igneous activity to which they belong. Dr. Harker has suggested that the Shiants are close to the focus of activity which gave rise to the innumerable dolerite-sills of Northern Skye. There is certainly, so far as our information goes, a resemblance in composition between the Shiant crinanites and the Skye sills, and it may indeed be the case that the Shiant group belongs to a deeper-seated phase of this activity, where slower cooling permitted differentiation of the sills to take -place to a marked degree; but, while the probability of this is admitted, a definite opinion must he withheld until our knowledge of the Trotternish sills is less scanty . It may, however, be mentioned that the description and microphotograph given by Prof. T. J. Jehu & Mr. R. M. Craig of the olivine-dolerite of the Maddy More sill, which lies off the eastern coast of North Uist, agree well with the crinanite of Garbh Eilean and Eilean an Tighe. An investigation of the islands of Trodday and Fladdachuan, which I propose shortly to undertake, may throw more light on this point.
There is also the possibility that the Shiant sills may belong to an earlier phase of activity than the Northern Skye sills, and that the latter were contemporaneous with the later olivine-basalt sheets of Eilean Mhuire, which apparently belong to the early and parent plateau magma-type.
It seems clear, however, that, whatever the exact age of the Shiant sills may be, they belong to a fairly early stage of Tertiary igneous activity, and that they throw considerable light on the mode of origin of many of the ultrabasic and alkaline types of the Tertiary Province of the Inner Hebrides. It has been shown that both picrite (sometimes approaching dunite) and syenite like that of Gamhnach Mor may be produced in thick sills from crinanite by crystallization-differentiation alone, without any assimilation of the country-rock; the former through the sinking by gravity of early-formed olivine-crystals, and the latter by squeezing-out or auto-intrusion of alkaline residual magma."
In his account of the glaciation of the Shiant Isles, Heddle considers that the glaciated appearance of the western shores of Garbh Eilean and Eilean an Tighe and the occurrence of erratics along the shores indicates a flow of the ice, which undoubtedly overrode the group during the Glacial Period, from west to east. He states also that the erratics are recognizable 'Long Island' rocks, and records the presence of hornblende-gneiss and Cambrian (Torridonian ?) grit among them. No striae were found on any part of the bed-rock.
1 am able to confirm the above observation., but I give them a very different interpretation, based largely on recent work carried out by Prof. T. J. Jehu & Mr. R. M. Craig on the 'Long Island’. I consider that the unglaciated appearance of the eastern shores of Garbh Eilean and Eilean an Tighe is due to the westward dip of the sill alone, that all traces of glaciation there must have been obscured long ago by the constant fall of columns from the scarp-cliffs, and that the vast scree-slopes of these shores have buried numerous erratics. I disagree, too, with Heddle’s statement that the blocks of gneiss and red Torridonian arkose found on the dip-slopes of the western islands and on the grassy top of Eilean Mhuire must necessarily come for the ‘Long Island’, for identical types are to be found at many points on the mainland. The glaciated appearance of the dip-slope of Garbh Eilean seems to me to indicate a flow of ice from south to north rather than from east to west, and the formation of the very distinct fault hollow is considered to be due mainly to the superior erosive action of the ice working along this soft band. It is thought, in fact. that the mer-de-glace of the Minch flowed northwards from the mainland and from Skye, and completely covered the Shiant group. In confirmation of this hypothesis we have the record by Prof. Jehu & Mr. Craig of north-and-south striae off the island of Harris west of the Shiants.
Raised beaches -There are no definite traces of raised beaches on any of the islands, but the comparatively level ground north-west of Eilean an Tighe, on which stood the shepherd's house and potato patch, may perhaps represent the 25-foot beach-platform, for the bluff east of it has the appearance of a sea-cliff.
Peat .- A great part of the dip-slopes Of Garbh Eilean and Eilean an Tighe is covered by a thin coat of peat ; but in three places indicated on the map this covering thickens out considerably, and fills up hollows with a swampy variety of peat.
The Gaelic place-names on the geological map accompanying this paper have been revised in accordance with information kindly supplied by Mr. Malcolm Macsween, of Tarbert.
The investigation of the islands was greatly facilitated by the cordial co-operation of a great many gentlemen. Permission to survey the group was very kindly given by Mr. Compton Mackenzie, who also rendered assistance in many other ways, and the visits to the group were made easy by the never-failing help of Mr. Malcolm Macsween.
The working-up of the results was made pleasurable by the constant encouragement and advice of the Director and officers of the Geological Survey, and for this I am particularly indebted to Sir John Flett, Mr. Murray Macgregor, Dr. H. H. Thomas, and Prof. E. B. Bailey, while Mr. Radley's four excellent analyses have proved of great value. Similar good offices were most kindly rendered by Prof. T. J. Jehu and Dr. R. Campbell, of the University of Edinburgh, and the work would not have been completed by this date if it had not been for the great facilities for research on the group granted to me by Mr. D. E. Innes, of the University of St. Andrews, during the summer term of the academic year 1927-28. In addition, I was exceedingly fortunate in being able to draw upon the great knowledge of Jurassic ammonites of the late Mr. S. S. Buckman and Mr. J. W. Tutcher, a knowledge which was placed most generously at my disposal.
I also desire to express my extreme indebtedness to Dr. A. Harker for his great kindness in permitting me to undertake the work which he himself was prevented through illness from carrying out, and for his valuable advice and encouragement.
Finally, grateful acknowledgement of the generous financial assistance rendered by the Carnegie Trust for the Universities of Scotland must be expressed.
The PRESIDENT (Prof. J. W. Gregory) congratulated the Author on having filled a gap in the geology of Scotland by his description of these islands, and on having shown their bearing on general geological problems. The Author had supplied an especially clear case of gravitational differentiation, which had been advanced by Darwin and supported by Dr. G. W. Tyrrell's interpretation of the Lugar sill. The speaker remarked that the iron-ores had not been affected by the process, as they were twice as abundant in the upper as in the lower layers. The paper is also of physiographic interest, as the fissures illustrated by the Author are frequent on the western coast of Scotland, and vary in size from narrow tension-clefts to the troughs occupied by some of the sea-lochs.
Dr. L. Hawkes remarked that the calculations of Ingersoll & Nobel (‘Mathematical Theory of Heat Conduction' 1913) indicated that a sill 500 feet thick would take so long to crystallize that the conditions would be eminently favourable for crystal- settling. In the earlier stages of cooling, the olivines from the upper border of the sill would, if the magma were superheated, be resorbed on sinking towards the middle, but the even increase in olivine content front the top to the bottom, as demonstrated by the Author, showed that this had not occurred. It was significant that, normally, magmas on intrusion were either crystallizing or on the point of crystallizing. It was regrettable that the contacts of the sill with the country-rocks were not exposed, as the selvages would provide samples representing the original homogeneous magma.
Under conditions of equal downward pressure, it was to be expected that the residual liquor in the crystal-mesh would be squeezed into vertical contraction-fissures thus accounting for the vertical veins described by the Author. With unequal downward pressure, the interstitial liquor would migrate towards those areas that were under less pressure, and a general enrichment by interstitial material would result over such areas; but not a segregation into horizontal layers. H asked whether the syenite-band in the interior of one of the sills might represent the liquor left there on the hardening of rock from above and below. Such a liquid layer would readily migrate with variations in overhead pressure, being expelled from some regions and collecting in others, thus giving rise to the horizontal schlieren and bands described.
Prof. S. J. Shand stated that perhaps no more convincing illustration of the sinking of early-formed olivine-crystals had ever been described, for, although many thick sills show a concentration of olivine towards the base, very few of them have been subjected to close chemical and micrometric study. The production of nepheline under conditions which seemed to exclude the co-operation of limestone was also a matter of great theoretical interest, which the speaker hoped to discuss further with tile Author of the paper.
Dr. C. E. TILLEY wished to express his appreciation of the work of the Author, who had, he thought, convincingly demonstrated that the picritic basal layer of this great sill owed its origin to gravitational differentiation. He (the speaker) was interested particularly in the Author's conclusion that the olivine of the picrite had essentially the same composition as the olivine of the upper crinanite-layers, despite the differences in microscopical features that distinguished them. He was loth to believe - unless the Author could provide more extended data than had been put forward - that the distinctive yellow colour of the olivine of the upper part of the sill was to be attributed to an incipient and selective decomposition. It was a matter of common observation. amply attested by quantitative data, that, as differentiation proceeded, there was a concentration of iron relative to the magnesium in the residual liquid, a feature denoted in the iron-magnesium ratio of the later-formed ferromagnesian minerals. It was very obvious here in the case of the pyroxenes, and might be expected to hold in the case of the olivines. One might expect these ophitic olivines, yellow in thin slice, to be considerably enriched in the fayalite molecule.
The Author, in reply to the President's question as to the non-sinking of the iron-ores, stated that the small size of the crystals militated against gravitative settling.
In reply to Dr. Hawkes's questions, he stated that the alkaline veins penetrating the picrite ran more or less vertically, and that Dr. Hawkes's suggestion as to the intrusion of the syenitic band on Eilean Mhuire had been considered as a possibility.
Finally, in reply to Dr. Tilley, he stated that the refractive index measurements by immersion in oils had confirmed the estimates of double refraction, by means of which the composition of the coloured olivine was deduced.