The GEOLOGY of the SHIANT ISLES (HEBRIDES)
 The Crinanite-Picrite Sill of Garbh Eilean and Eilean an Tighe. [Continued]
The percentage mineralogical composition by weight, or mode, of many of the types described above was calculated by the aid of the Shand recording micrometer, the results being tabulated below:-
* Includes mesostasis.
A. Picrite from the base of the cliffs in the extreme south of Garbh Eilean.
B. Basic olivine-dolerite, 30 feet higher in the sill than A.
C. Crinanite, 70 feet higher in the sill than A.
D. Crinanite, 125 feet higher in the sill than A.
E. Crinanite, 250 feet higher in the sill than A.
F. Crinanite 400 feet higher in the sill than A.
G. Pegmatite-vein in crinanite, in the bay south of Sgeirean Mol na h-Athadh.
The large size of the ophitic minerals in most of the types enumerated above renders micrometric analysis somewhat unreliable, especially in the case of the crinanites, where a crystal of augite may extend over an entire section. For this reason the results have been given only to the nearest integer, and apatite, which does not make up 1 per cent. of any of the rocks, has been omitted.
* Forsterite, 44,6, Fayalite, 16,9.
I. Crinanite; southern face of Garbh Eilean, 125 feet above the shingle-beach. Anal. E. G. Radley.
A. Dolerite cone-sheet ; Cuillins (Skye). Anal. W. Pollard. Quoted from the Mull Geol. Surv. Memoir, 1924, p. 15.
II. Olivine-dolerite; southern face of Garbh Eilean, 30 feet above the shingle-beach. Anal. E,. G. Radley.
B. Olivine-dolerite apophysis from a peridotite-dyke; Sgurr na Banachdich (Skye). Anal. M. G. Keyes. Quoted from N. L. Bowen, ' Evolution of Igneous Rocks' 1928, p. 154.
III. Picrite; southern face of Garbh Eilean at the shingle-beach. Anal. F. G. Radley.
C. Peridotite-dyke; Coir' a' Ghreadaidh (Skye). Anal. M. G. Keyes. Quoted from N. L. Bowen, loc. supra cit.
D. Average plateau-basalt magma-type. Quoted from the Arran Geol. Surv. Memoir, 2nd ed. (1928) p. 121.
Of the foregoing analyses, I, II, & III were made for me by Mr. E. G. Radley through the good offices of Sir John Flett & Dr. H. H. Thomas, who very kindly interested themselves in the work. A fourth analysis of a type from Eilean Mhuire was also made by Mr. Radley, and it is scarcely necessary to point out that these four excellent analyses have proved a valuable addition to the petrology of the group.
The crinanite of Garbh Eilean is distinctly more acid than any example hitherto analysed, and finds its nearest analogue in Analysis A of the dolerite cone-sheet from the Cuillins. Despite its slightly higher magnesia content, the Skye rock does not appear to contain olivine.
Analysis II, of the olivine-dolerite, can be matched only by a slightly more basic type, also from the Cuillins, stated, however, by Bowen, to contain only 16 per cent. of modal and normative olivine.
Analysis III, of the picrite, can be matched alone, fairly closely, by a peridotite - again from the Cuillins - which. contains 55 per cent. of normative olivine.
Tbe average composition of the 'plateau-basalt magma-type' has been calculated by Dr. G. W. Tyrrell, and is given under D. On comparing this average with Analyses I, II, & III, it will he seen that the crinanite is more acid, while the olivine-dolerite and picrite are more basic, than the plateau-basalt magma-type. It is further apparent that the sinking by gravity of iron-ore and olivine-crystals of normal composition from a magma corresponding in composition to the more acid examples of the plateau-basalt magma-typecould produce the basic types of Analyses II & III, while the upper part of the magma-basin might well correspond in composition to Analysis I, except in the case of the alkalies which would be lower in I. This deficiency in alkalies in the case of the Shiant rock might be caused by the expulsion of the last alkaline residuum, of the magnia through the picrite in the manner described above.
The case for gravitational differentiation in the great Shiant sill will be Considered in detail below.
All the above measurements were made with exceedingly fresh material, and are therefore trustworthy. The heights given, in the case of 45 feet and above, were obtained by aneroid, and are accurate to the nearest 10 feet. Those below 45 feet were interpolated by visual estimation with a probable error of ±5 feet.
I believe that the picrite-crinanite sill of Garbh Eilean and Eilean an Tighe provides a singularly clear demonstration of crystallization -differentiation unique in the British Tertiary province at least.
The gravitational settling of early-formed olivine-crystals is thought to be the first and most important phase, and will now be considered.
Although the theory of gravitational differentiation of a magma through the settling of early-formed crystals was initiated by Charles Darwin, it gained little support until the present century, notwithstanding its favourable consideration by such eminent petrologists as Sir Jethro Teall and Prof. F. J. Loewinson-Lessing. During the last twenty-five years, however, numerous examples of the settling in igneous masses of early-formed minerals, such as olivine and augite, have been described from all parts of the world; while in 1915 the theory received important support from the brilliant experimental work of Dr. N. L. Bowen, who was able to demonstrate the settling of olivine and pyroxene in artificial melts in a platinum crucible. All examination of the literature of natural occurrences indicates that even the best known type-examples possess some features which prevent them from being entirely convincing. Other processes of differentiation, or subsequent intrusions, may, for instance, conspire to confuse the issue.
The references given, which may be taken as representative of the literature of gravitational differentiation, display some features in common. All the intrusions or flows, except the Lugar sill, the magma of which was abnormally aqueous and mobile, are exceedingly thick ; and slow cooling, ensuring a long period of fluidity, appears to be a sine qua nonof this type of differentiation. It will also be noted that, all six examples show discontinuous variation, usually by the occurrence of schlieren or banding. This may be due, as in the Lugar sill, to differentiation prior to intrusion, or more probably, as Bowen has suggested, to intrusion during crystallization of the more liquid portions of the magma into rifts in the crystal-mesh. Dr. G. W. Tyrrell has recently developed the latter hypothesis in connexion with the crinaniteanalcite-syenite sills of Ayrshire, and it is applicable with equal pertinence to a similar sill on Eilean Mhuire, as will be seen below. There can be no doubt, at any rate, that the processes mentioned tend to obscure the variation caused by the gravitational settling of early-formed crystals, and the ideal case of this type of differentiation is one where the magma has been intruded (or extruded) in a completely liquid condition, and where no deformation of the mass takes place during cooling. Such a case is, of course, impossible of realization, for in a lava-flow there will always be movement during crystallization, while in all intrusion there will always be deformation due to the pressure exerted by the overlying cover. It is hoped, however, to show thatthe crinanite-picrite sill of Garbh Eilean and Eilean ail Tighe makes a close approximation to the ideal.
A continuous variation should be shown in all cases of gravitational differentiation (the chilled contacts being excluded) unless the sinking constituents ceased to crystallize at all early stage while themagma was still quite fluid. In this latter case a layer consisting chiefly of the sinking minerals would be formed at the bottom of the mass, and would have a relatively sharp contact with the upper part, as in parts of the Palisade ‘diabase'. If, however, the heavy minerals continued to crystallize until the magma became gradually too viscous to permit of further sinking, then a continuous increase of specific gravity, together with a corresponding rise in the proportion of the sinking minerals would be seen from top to bottom of the mass. This is beautifully illustrated in a microphotograph of a vertical section of the melt, in which Bowen demonstrated gravitational sinking of crystals in the laboratory. No contacts are seen in the Shiant sill, but the most complete section of the intrusion, the lowest exposures of which must closely approach the lower contact, shows a wonderfully continuous increase in the proportion of olivine (the only mineral tosink appreciably)from top to bottom, while the corresponding rise in specific gravity is, if anything, still more striking (p. 373). These two features are brought out clearly in two graphs (below) where height above the visible base of the sill is measured along the abscissae and specific gravity or percentage of olivine along the ordinates.
The smoothness of the specific gravity curve is particularly noteworthy, any slight divergence of the points from it being more than accounted for by the probable error in estimating the heights. (The crinanite of Eilean an Tighe also shows a slight downward increase in specific gravity.) The normative olivine of the three analyses shows a similar downward increase.
It has already been mentioned that there are no schlieren or banding. in the sill, the only discontinuous variation being the occurrence of the thin pegmatite-veins by which residual magma was expelled at several stages during crystallization.
That olivine was the only mineral to sink appreciably is shown by the comparative constancy of the ratio of the proportions of the other two major constituents - augite and plagioclase. The divergences from the mean ratio show no regular tendency (see table below), and may he fully accounted for by the errors of the
RATIO OF PLAGIOCLASE TO AUGITE
micrometric method of estimating the mode. It is obviously a very difficult matter to obtain accurate figures when the whole of a large rock-section may be occupied by a single augite-felspar cluster, even if several slices are measured. If there had been any sinking of pyroxene-crystals, a concentration of that mineral would have occurred in and above the pierite zone, with a corresponding fall in the augite-plagioclase ratio. The texture of the rocks, too, indicates that the pyroxene commenced to crystallize at a distinctly later period than the olivine.
The history of the cooling of the sill may now be considered. The first stage is the intrusion of a magma of normal crinanitic composition in an almost completely fluid condition, as the extreme rarity of phenocrysts testifies. As the temperature fell, olivine began to crystallize, and, by virtue of its light specific gravity,, was able to sink in the fluid magma. An accumulation of olivine-crystals was thus formed near the bottom of the sill. There was probably a slight fall of magnetite-crystals at the same time, but their small size militated against sinking. The variation in the size of the olivine-crystals in the more basic types is considered to be due to growth during settling. A similar variety of size is seen in sinking forsterite-crystals in one of Bowen's artificial melts. The crystals of olivine in the picrite are distinctly smaller than the majority in the overlying olivine-dolerite, this being due to the shorter distance travelled before coming to rest, and consequent limitation of the opportunities of growth.
While the olivine still continued to crystallize, basic plagioclase and augite made their appearance, but their mode of crystallization was very different. In the picrite zone they certainly both formed large plates, which enclosed poecilitically the closely packed mass of olivine-crystals at a comparatively early stage, but above, where chemical (and probably physical) conditions were different, the plagioclase had a distinct tendency to crystallize out in small well formed laths, while the pyroxene obstinately refused to show crystal boundaries, forming large clumps which enclosed numerous plagioclase-crystals. These pyroxene-felspar clusters make up the bulk of the crinanite and much of the olivine-dolerite. Olivine, meanwhile, continued to crystallize, but its power to sink vanished as the magma became more viscous and full of crystals. It was then partly enclosed by the pyroxene-felspar clusters, or, if it persisted in its crystallization to a late stage, as it did in the crinanite, ophitic olivine-felspar clusters were. formed in the same manner as the pyroxene-felspar examples. The later olivine of the crinanite rivalled, indeed, the pyroxene in its obstinate refusal to form crystal boundaries, and in the upper portions of the crinanite the olivine-felspar clusters reach a diameter exceeding half an inch.
There does not appear to have been any perceptible change during crystallization in the composition of the olivine; but, while the first pyroxene was practically colourless, the outer zones, in places at least, are distinctly purple and probably rich in iron. The plagioclase, also, shows distinct zonary banding - especially outside the pyroxene-felspar clusters-varying from basic labradorite to oligoelase. In the olivine-dolerite and picrite the iron-ore is magnetite of early crystallization occurring in small grains, but in the crinanite it is itinenite, and assumes a characteristic skeletal habit. It is frequently enclosed in pyroxene-felspar clusters, however, and is probably, therefore, another mineral to crystallize out at an early stage.
The latest stages of crystallization are seen between the ferromagnesian-felspar clusters in the shape of acid plagioclase round the margins of basic laths, and finally, in interstitial analcitc and natrolite which represent the thoroughly aqueous and sodic residuum of the inagma. During these last stages the augite ceased to crystallize as norinal titanaugite, and the occurrence of a few scraps if xgirine-augite, due to reaction with the sodic liquid, shows the end of its story. Crystallization of olivine appears to have ceased a little earlier.
If there had been no pressure upon the crystal-mesh during solidification, the crinanite and olivine-dolerite would have been distinctly more alkaline. As it was, a considerable amount of alkaline magma was expelled through veins in the picrite (and probably through the chilled upper margin of the sill) at various periods during crystallization. The first fraction seems to have been expelled soon after the complete solidification of the picrite, and successive fractions of increasing alkalinity were removed until the sill was rendered free from liquid. The later fractions were of zeolitic composition, and had a marked corrosive effect upon the earlier.
In Garbh Eilean. no examples of 'auto-intrusion,' that is, intrusion of magmatic liquid into rifts in the crystal-mesh, are seen in the crinanite or olivine-dolerite. The Jurassic strata of Eilean an Tighe have, however, brought about this phenomenon in the crinanite. These Jurassic beds end abruptly against the igneous rock at Sgeirean Mol na h-Athadh, and were almost undoubtedly floated up from below. They reacted with the magma, and gave off a stream of gas-bubbles which ultimately became filled with zeolites ; but they -also gave rise to lines of weakness and cracks in the crystal-mesh, into which pegmatite -veins of teschenitic composition were intruded.
The cause of the pressure that gave rise to the intrusion of all the pegmatite-veins is doubtful, but may, as Dr. G. W. Tyrre11 suggests for certain sills of Central Scotland, be due simply to the weight of the cover. The matter will, however, be more fully discussed in connexion with the lower sill of Eilean Mhuire.
Two well-marked continuous reaction-series may be traced in the various types of the sill-
(1) Lime felspar -> lime-soda felspar --> soda-lime felspar --> soda-felspar.
(2) Titanaugite --> aegirine-augite --> Aegirine.
Any member of the first series may take part in the discontinuous reaction series - plagioclase --> analcite --> natrolite, while the following reactions have also been noted:-
Ilmenite --> biotite, olivine --> biotite, olivine --> barkevikite, ilmenite --> arfvedsonite, aegirine -> arfvedsonite.
No process other than gravitational settling of olivine seems able to account for the differentiation of the sill. Assimilation may be ruled out at once, for there is no indication of appreciable reaction between the crinanite and the Jurassic beds of Eilean an Tighe, which show a fairly well-defined junction, neither were any traces of xenoliths found. Density stratification while still liquid is the only other alternative, and this, too, may be dismissed with equal brevity, in view of Bowen's experimental work on the diffusion of silicate melts.
Continued on next page