A Report on the geology of the Northern Marginal Zone of the Central Layered Intrusion: Isle of Rum, Scotland
Table of Contents
The aim of this study is to establish the nature and sequence of events that formed the Central Layered Intrusion on the Isle of Rum in the Inner Hebrides of Scotland. More specifically, the focus of this project is on the North Marginal Zone (NMZ), which comprises the area close to the boundary between the Main Ring Fault (MRF) of Rum’s volcanic centre and the Torridian sandstone country rock to the north.
The primary method that was chosen for use in the field was outcrop mapping, this was due to the abundance of easily accessible outcrops in the mapping area. This technique consisted of adding any observed and classified outcrop to the basemap as accurately as possible (in terms of location and size/shape). Transect mapping was also used occasionally to try and establish lithological changes over specific distances or areas. During the reconnaissance of the mapping area it was clear that the majority of lithologies didn’t have a clear boundary to them, with most being gradational and/or diffused, so boundary mapping was not used as a method. As a result, most geological boundaries on the map are inferred (or conjectured: where no outcrops were present).
These techniques were used in conjunction with standard field observations, such as interpreting the lithology of outcrops, construing structural features from the geomorphology, using sketches to simplify data and enable further interpretation and hypothesis formation etc. Using a compass clinometer, structural data was collected from a variety of different features (bedding/igneous layering/dykes) to enable further analysis out of the field.
These method strategies did have their limitations though, with the rapidly fluctuating weather hindering several of the mapping techniques on occasion. Also, I was critically ill for a short portion of the mapping trip so that limited the extent of data that was collected, and the map area that was covered overall.
The RS is found mostly in the north of the mapping area, with very little outcropping in the south. The majority of the RS outcrops that were mapped were south of the main path that runs east-west from Kinloch to Harris and north of the Main Ring Fault.
In general, most of the exposures had a yellow brown/pink brown colour to the weathered surfaces, with fresh surfaces exhibiting a slightly lighter brown colour. The only noticeable change in colour was when the sandstone was in close proximity to igneous intrusions, in which case it was greyer in colour, likely due to contact metamorphism.
The RS is mostly well bedded and has a varying grain size, though it normally ranges between 1-2 mm, which is medium to coarse. Overall the sphericity tended to range from sub-rounded to well rounded. In some good examples, sedimentary structures such as cross bedding were seen, though there was no evidence of any fossil material. The dip/dip direction trend of the RS was averaged to be ~30° to the north-west (excluding anomalous data).
In areas where there were igneous intrusions identified, the contact with the sandstone exhibited some contact metamorphic alteration, with baked/chilled margins observed on occasion. Sometimes this made the fresh surfaces look like a quartz rich igneous rock, due to the interlocking textures of the grains. The RS also displayed frequent fractures and jointing, with rare weak cleavage. However most of these features were rather indistinct and hard to establish a general trend from.
One distinct change in the RS unit was that the outcrops dipped more steeply (differing from the general unit trend) the closer to the Northern Marginal Zone/MRF contact they got, with up to ~50-60° dip to the north.
The RP is located predominantly in the centre section of the mapping area, with most of the mapped areas trending broadly north-south, like the exposures both east and west of Long Loch and the large ridge between the Priomh Lochs and Loch Bealach Mhic Neill. In addition, most of the exposures at the summit of Barkeval were composed of peridotite. Occasional plugs of peridotite were also found in the RS unit in the north.
The RP outcrops have a distinct rusty brown colour on the weathered surfaces, with fresh surfaces showing a dark grey/green colour. The colour of these outcrops may indicate a high proportion of ferrous (iron) rich minerals.
The RP was phaneritic in general, with rare pegmatitic outcrops found. Most fresh surfaces showed an equigranular medium grained texture with an average crystal size of ~1-2mm. In some localities layering and pseudo-sedimentary structures were observed in the peridotite, though often this was close to the gradational boundary with feldspar-poor gabbro and thought to be a mixing/settling feature (Fig. 1). The peridotite outcrops had almost no dykes cutting through them, which is different from all the other lithologies that often had a high frequency of them.
The RP had a high percentage of both dark grey/black coloured minerals (likely pyroxene) and slightly translucent bottle green/brown coloured ones (likely olivine). There was one outcrop with a few lumps of a black crystal cluster/assemblage that had a metallic/vitreous lustre. This was inferred to be a nodule of chromite or similar spinel group mineral (monomineralic pod). Also at another outcrop there were some well-formed deep red/brown coloured phenocrysts observed that were atypical, though these were not fully classified.
The overall composition of the RP was ultramafic, though occasionally rare feldspathic minerals were seen in fresh surfaces of the peridotite. This unit was often difficult to classify to a high degree of certainty as it was compositionally and texturally very similar to the FpG, in addition there is no distinct boundary between these units (most were inferred) and that they tend to be gradational in transition to one another.
The weathered surface of most of the RP outcrops was very varied, but the majority showed an uneven/rough weathered surface (honeycomb like appearance), with rare white feldspar crystals seen on these surfaces, likely due to it being more resistant and stable than olivine at the earth’s surface.
Figure 1 – Outcrops of peridotite on the east side of Long Loch, close to the boundary with the FpG unit. This outcrops texture is very uneven and rough, with occasional small feldspathic xenoliths/veins (A) observed in increasing proportions closer to the gradational boundaries. (Field notebook for scale).
The RGn outcrops in the central section of the mapping area, with the majority on the north, west and east sides of the Priomh Lochs (Mor and Beag); only a small section outcrops to the south. There is also a north-south trending zone of gneiss on the east side of Long Loch.
The RGn has a distinct colour scheme to its less weathered/fresh surfaces, with light grey/white bands alternating with darker grey/brown bands. Though its weathered surfaces tend to be less distinct and rather grey in colour overall.
In almost all outcrops the gneiss showed a very clear and well-developed foliation/banding. This gneissic banding is very distinct and has been deformed extensively in several ways. As a result, this has formed hummocky/lensoidal shapes throughout the gneiss. In addition, certain outcrops exhibit folding, faulting and boudinage within the bands of the unit (Fig. 2 & 3). Some of these bands are rather thick (up to 20 cm), but most tend to be thinner with a rough average of ~1-3 cm. The thinner bands have been altered and deformed more intensely than the larger ones.
This RGn unit is feldspathic in composition overall and is possibly in the granulite/amphibolite facies. The light-coloured bands (leucosome) are those rich in felsic type minerals, or their polymorph/metamorphic equivalents. The darker bands (melanosome) have a more mafic mineral composition and are less frequently observed compared to the leucosome bands.
Multiple weathered dykes cross cut through the RGn (Fig. 2), which were mostly basaltic in composition (from fresh surfaces observed), though they have been significantly deformed. The concentration of dykes within the gneiss unit was very high compared to other lithologies, there was also evidence observed for 2 different sets of dyke swarms. Many of these dykes show slight chilled margins (only a few mm wide), which indicate they formed more recently than the gneiss in terms of relative age.
The RGn was the only regionally metamorphosed rock observed in the mapping area and as a result it is likely the oldest of the lithologies. It is also probable that it is the basement rock, though the RS unit could also be interpreted as that.
Figure 2 – Outcrop of gneiss seen west of the Priomh Lochs that has been cut by a small basaltic dyke (A). Small scale faulting has also occurred in both the light bands (C) and dark bands (B) along strike of the dyke. In this outcrop the darker bands (B) were much thicker than that of the lighter bands, which was uncommon for this unit. (Hammer for scale).
Figure 3 – Gneiss outcrop observed east of the Priomh Lochs showing clear boudinage structures (A) within the darker bands. Foliation is also clearly observed in the lighter bands above and below the boudin necks. (Hammer for scale).
Textural and compositional changes were frequent between gabbro outcrops, some examples include: gabbros with low feldspar content; gabbros with high feldspar content; gabbroic/peridotite breccia; and layered gabbro with peridotite xenoliths. All boundaries between these mafic and ultramafic units were difficult to establish due to the majority being gradational in nature. They transitioned gradually which made the mapping more difficult to keep to a high level of accuracy. The differing sub-types of gabbro are detailed in the next sections.
The FpG unit outcropped across the majority of the mapping area, with small plugs within the RS unit in the north, large north-west/south-east trending ridges on the west side of Long Loch, and several plugs intermixed with other units in the centre of the mapping area. There were also several outcrop layers of FpG on the flanks of Barkeval in the south-east.
The weathered surfaces often have a light brown/grey colour, with fresh surfaces showing a darker grey colour. The mineral ratios for the FpG varied, but an average outcrop showed ~5% or less olivine, a range of ~5-25% feldspathic minerals, and the remaining percentage being made up of pyroxenes/amphiboles. The main distinction between this unit and the peridotite is that the feldspar-poor gabbro has a low percentage of olivine crystals. This gabbro unit had a varying grain size, which ranged from less than 1mm, up to a maximum observed of ~10mm.
The FpG outcrops were mostly massive in structure, though there were a few exceptions. One notable exception was that of a small cliff section (located south-east of Priomh Loch Beag) that showed clear fabric/small-scale layering, the surface of which was weathered and displayed the honeycomb appearance (as seen in other localities). The layers on this exposure ranged from less than 1cm up to 30cm in thickness and alternated between the feldspar-poor gabbro layers (thicker) and the highly felsic/pegmatitic layers (thinner).
The unit is distinct for its peridotite xenolith honeycomb/dendritic texture, where there were peridotite xenoliths that have weathered away and left only the feldspar crystals and other mafic minerals to remain (Fig. 4).
Some FpG outcrops exhibited feldspathic veins, most with preferential orientation and some outcrops had small clusters of very coarse grained pegmatitic gabbro. As with most of the other units, there were an abundance of dykes that cross cut these outcrops, some of which showed slight chilled margins.
Figure 4 – Feldspar-poor gabbro outcrop west of Long Loch at a gradational contact with peridotite. A fresh surface (A) reveals the dark grey aphanitic texture, with a low proportion of lighter feldspathic minerals. Close to this there is a dendritic/radiating texture (B) that commonly occurs near the gradational boundary of this unit. Further to the right (C) the FpG grades into to peridotite, indicated by a colour change and the massive texture.
The FrG unit outcropped in several locations across the mapping area, with several small plugs in the RS unit to the north and some small plugs on the west side of the Long Loch fault. The majority of the larger FrG outcrops were in the east close to Loch Bealach Mhic Neill and Loch Gainmhich, there were also abundant feldspar-rich gabbro outcrops layers on the flanks of Barkeval in the south-east.
Weathered surfaces often had a light grey colour, with fresh surfaces showing a slightly darker grey colour. This unit is similar to the FpG, with the major difference being that this unit has a higher percentage of feldspathic minerals (~30% and above).
The FrG unit was often coarser in nature than most other mafic units. In a few outcrops there has been slight alteration to the feldspathic minerals, as some had slightly different coloration, ranging from milky brown to green colours.
In addition, there were some outcrops than had more elongate and interlocking feldspathic minerals compared to the more frequently observed equigranular texture. Another texture observed was that of a dendritic nature in the feldspar crystals/layers, which is more common in felsite type rocks, however there was still enough dark mafic minerals in these outcrops for them to be classified as a gabbroic rock. Some of these felsite type layers had a crystal arrangement that looked spherulitic/radiating in appearance.
In most FrG outcrops, it was very difficult to distinguish a clear boundary between the feldspar-rich and feldspar poor gabbro units, as the contact was never sharp and almost always gradational (Fig. 5). As a result, most of the boundaries were marked on the map with a specific line symbol to reflect this.
Figure 5 – Outcrop of Feldspar-rich gabbro north of Long Loch, showing 3 distinct sections. A and B show typical massive feldspar-rich gabbro, whereas the central band (C) displays feldspar-poor gabbro that contains pegmatitic layers. This was a common feature observed in the gradational contacts that most of the igneous units exhibited. (Field notebook for scale).
The LGX unit outcrops in a few localities in the west and north-west areas around Long Loch. This unit is only clearly exhibited in two outcrop areas in the mapping area and they both progressively thinned towards the south, eventually ending in between the peridotite and feldspar-poor gabbro units.
This unit, sometimes referred to in the notebook as layered xenolithic gabbro, is mapped on as a separate unit regardless of their being compositional similarities to other units. This is due to its structural features being vastly different from the other gabbro units. The composition of this sub-unit is typical for a gabbro, however there is a distinct layering texture and frequent finer grained gabbroic xenoliths throughout its outcrops (Fig. 6).
The contact of this unit and the peridotite is seen in certain outcrops, where the contact was varied and exhibited a fair amount of intermingling. Parts of the layered gabbro units were observed to have veins, that ran along what appeared to be fractures into the peridotite unit.
The darker (thinner) layers of the LGX unit are slightly coarser than the lighter layers and have a higher proportion of mafic minerals (likely olivine and pyroxene). Some of these thin dark layers are very coarse with regards to crystal size, and almost pegmatitic in nature. The lighter (thicker) layers in the LGX unit are slightly finer grained than the darker layers and have less mafic minerals (more felsic).
Measurements of the layering within the LGX showed that it shallowed slightly towards the south (average dip of ~40°), though the westerly direction of the layering remained almost the same.
Figure 6 – Outcrop of Layered Gabbro with Xenoliths north-west of Long Loch, showing distinct and well developed igneous layering (A). Also xenoliths of both finer grained FpG and FrG were frequently observed within the unit, with rare peridotite xenoliths seen close to the units boundaries. (Field notebook for scale).
The RPB outcrops are only seen in a few places, though mostly they are found on the western edge of our mapping area, roughly west/south-west of Long Loch.
This unit is seen in some other locations within other units, though few of these were large scale enough to be mappable. For example, some of the RPB is observed in the transitional/gradational boundary between the RP unit and the LGX unit. However these sections were unlikely to be discrete RPB outcrops, and probably the result of a different process.
The weathered and fresh surfaces have the same basic characteristics as the RP and FpG units, the only difference is that they are not homogenous across outcrops, having a clear separation of distinct large blocks of peridotite and a matrix of varied feldspar-poor/feldspar-rich gabbro (Fig. 7). The clasts in this breccia had a size that ranged from smaller than 1cm, up to several 10’s of metres in size and greater.
Occasionally pegmatitic layers (mafic composition) with crystals up to roughly 5cm large were found within the breccia unit. There was a distinct lack of dykes observed in the RPB outcrops, which was similar to that of the other ultramafic lithologies.
Figure 7 – Outcrop of Peridotite Breccia located on the ridges west of Long Loch. Clear brown peridotite clasts (A) are observed within a gabbroic matrix (B) with varying feldspar content. (Field notebooks for scale).
The RGr unit only outcrops in a few areas north-east of Long Loch. These sections of granite are bounded by the Main Ring Fault boundary to the north and the gneiss and gabbro contacts to the south.
This unit displayed a blocky, uneven weathered surface with abundant joints and fractures throughout. The weathered outcrops were light grey in colour, with the fresh surfaces being very similar. No large phenocrysts were observed within the granite and its overall texture was phaneritic with an average grain size of ~2-3mm.
Fresh surfaces showed a highly felsic composition with lighter coloured minerals dominating the rock (>65%). With feldspathic minerals being abundant and platy micas and dusty translucent crystals (inferred to be quartz) having been identified via hand lens, it was clear it was a granite. Using a penknife, these crystals were scratched to test for hardness and they could not be scratched, indicating the hardness level is greater than 7.
In a few outcrops assimilation between the RGr unit and the FpG unit were seen. The granite here was observed to be slightly coarser than previously seen (though still rich in felsic minerals), with some increased degree of alteration observed. Along the outcrop there were inclusions of granite within a section of the feldspar-poor gabbro. Using the principle of inclusion, this is evidence that the granite is the older of the two units.
Most of the RGr outcrops displayed a distinct texture with the lighter coloured felsic minerals. Rather than having a more homogenous equigranular texture (typical of a granite), these lighter minerals were often elongate and had a radial interlocking texture to them, with no obvious preferred orientation. This is a granophyric texture, though for simplicity the name of this unit was kept as a granite.
The RHSM unit predominantly outcrops in the central and eastern sections of the mapping area. With the major sections being south-west of Loch Duncan/north-east of the Priomh Lochs, the north and west flanks of Loch Bealach Mhic Neill and the north and south sides of Loch Gainmhich.
The RHSM (changed from ‘SSLB unit’ in notebook due to inaccurate unit name) is a heterolithic sedimentary unit, that has fresh surfaces showing a coarse-grained sandstone that alternates with layers of a well bedded fine-grained silt/mudstone (both of which can have localised brecciation). The sandstone beds typically have a course/medium grain size and bed thickness of roughly 10 to 30cm in thickness. The silt/mudstone beds typically have a finer grain size overall with thinner beds of roughly 1 to 5cm thick. The weathered surface of the RHSM outcrops is typically a red brown/grey colour, with the fresh surfaces being a darker brown/grey colour.
Contact with the RHSM and the RGn unit is rather gradational in places and makes identification difficult for some outcrops. The contact with the igneous lithologies also exhibited a similar gradual change, often with brecciation of the RHSM unit closest to the boundary.
In some outcrops the laminations/bedding of this unit has been fractured and fragmented, which makes certain sections appear like a chaotic poorly sorted breccia. There is also evidence for slight contact metamorphism near the contact with igneous units and occasional micro-folds were observed, but again this was localised to small areas/outcrops (Fig. 8).
Figure 8 – Small outcrop of the HSM unit located south of Loch Duncan, displaying thinly laminated mudstone beds with a microfold structure (A), and coarser sandstone beds above (B). (Pencil for scale).
The RVBt unit forms the edge of the boundary of the Northern Marginal Zone, with it being located on the edge of the Main Ring Fault. Its main boundary is with the RS unit to the north. It runs in an east-west trend along the Main Ring Fault.
Fresh surfaces of the RVBt are of a light grey colour, showing a fine grained and almost aphanitic/crystalline texture, with the clasts being fairly distinguishable and having differing grain size. The weathered surface of these outcrops shows a messy amalgamation of clasts and matrix material (Fig. 9), with it being difficult to determine and identify individual clasts in some outcrops as they had been deformed (likely metamorphosed). As a result, this breccia is chaotic overall with relatively poor sorting. In places the matrix has some feldspathic minerals within it, which do appear to show a slight degree of preferential orientation.
In a few outcrops (notably at Locality 41, north of the Priomh Lochs) there were what appeared to be small welded tuff layers/inclusions. These could be up to several m’s in size and displayed a very blocky and jointed surface, that is smoother and less weathered than the rough, uneven breccia. These outcrops had a darker grey aphanitic texture, with fresh surfaces showing some faint grain grading, likely from ash that is now devitrified. There was also an abundance of small white feldspathic crystals (phenocrysts) in the darker grey matrix.
There are subtle differences between this unit and the RIB unit. The clasts within this unit are composed of sandstone, gneiss and gabbro. In addition, this unit has a slightly higher ratio of matrix to clasts (~60%:40%) than the RIB and there are inclusions/outcrops of tuff observed within this breccia, which is another clear difference. It is not clear from field observations as to whether the matrix material is of igneous origin or not, however it had an aphanitic texture on fresh surfaces.
Also, some sections of this unit were cut by large basaltic dykes (up to 1.5m wide) as with most of the other lithologies. With the lack of depositional/graded bed features (typical of extrusive volcanic units) observed, the interpretation of this unit was changed to be that of a volcanic breccia (with tuff), though it’s process of formation is unclear.
Figure 9 – Outcrop of Volcanic Breccia with tuff unit observed NW of Loch Duncan, showing a variety of poorly sorted sub angular clasts in a chaotic matrix. (Pencil for scale).
The RIB unit outcrops close to the east margin of the mapping area in the centre section, on the south-east flank of Loch Gainmhich.
Weathered surfaces of the RIB are light grey in colour and have a rough and chaotic appearance due to the variety of clasts. Fresh surfaces show a dark grey colour with a fine-grained texture; the clasts within having lost some of their internal texture and have been amalgamated into the matrix, which is likely due to metamorphism at time of formation/emplacement.
This breccia exhibits relatively angular clasts of sandstone, gneiss, gabbro and a undetermined fine grained feldspar-rich igneous clast. This was all contained within a supportive igneous matrix, with the clasts being poorly sorted and showing no preferential orientation. The dominant clasts are sandstone, with some clasts even displaying relict bedding. The matrix material of this breccia is aphanitic/crystalline in general and likely igneous in origin.
This unit is similar to the RVBt unit but there are some subtle differences, this unit has a slightly lower proportion of matrix and has the presence of the undetermined fine-grained feldspar-rich igneous clasts. It also lacks any tuff material or inclusions within it like the RVBt unit. However, in places near the contact with the RWT unit, there is interbedding between the units, but there is no tuff within any other outcrops.
Several basaltic dykes cut the RIB unit, though the lack of baked margin within the breccia indicates that it was probably at a high/similar temperature relative to the dykes when they intruded.
The RWT unit is located in the central section of the mapping area, on the far east side, and on the eastern flank of Loch Gainmhich. This was the only location where it was found to outcrop for the mapping area.
Th fresh surfaces of the outcrops were a light grey colour and had a blocky and fractured weathered surface. Fresh surfaces showed a very fine grained dark grey matrix with abundant white phenocrysts (likely feldspathic minerals), though these showed little to no preferred orientation. The matrix material is igneous and has a crystalline texture, it is likely made up of devitrified ash and other volcanic fragments.
In some outcrops, dark stretched fragments were observed, these were later determined to be fiamme structures (welded pumice clasts); they also appeared to have a slight preferential orientation. These clasts were darker in colour than the matrix and had a diffused outline to them. There were no observed way-up structures in this unit, though the RIB and RHSM both graded up to and were interbedded with the welded tuff in certain outcrops, which gave the relative age of the tuff as younger. Dykes were also found in some outcrops of the welded tuff, sometimes cross cutting through it and other lithologies, meaning that they are likely the youngest relative unit of all those seen.
Dykes were found across the entire mapping area, intruding into every unit identified, though there were very few observed in the peridotite outcrops (especially in the western sections). However, most of the dykes observed were found in the central portion of the mapping area and within the gneiss and gabbro units.
Most of the dykes were dark/light grey in colour and more fine-grained than the rock they had intruded. Often there were multiple dykes crosscutting one another and relative ages were established, which was prevalent in the RGn unit.
All dykes found on the island were either basaltic or doleritic (dependent on the crystal size) and mafic in composition, though the majority were basaltic with an aphanitic texture. The fresh surfaces showed a very dark grey/black groundmass, which were likely composed of pyroxenes with some olivine. The occasional phenocrysts were lighter in colour and likely feldspathic, with sizes ranging from <1mm to ~4mm. Some dykes observed exhibited flow banding near their margins and an accumulation of larger crystals (phenocrysts) in the centre/dyke core (Fig. 11). A few dykes observed showed chilled margins (Fig. 10), though almost none showed any evidence of baked margins in the host rocks.
A few rare dykes were of very large scale, some getting up to 2m in width and over 30m in length across outcrops. This meant that these were large enough to be put onto the base map with more than just a symbol. These larger dykes had a more phaneritic texture and crystals were observable with a hand lens/naked eye. This meant they were doleritic due to the coarser crystal size.
Most of the dykes had a steep dip of ~75° up to vertical (90°), though a few uncommon examples had more shallow dip angles.
Figure 10 – Outcrop of gneiss with a basaltic dyke cutting through it, observed south-east of Priomh Loch Beag. This dyke exhibits a clear chilled margin (A) that is roughly 1cm wide and of a much darker colour, this was frequently observed with other dykes that were intruded into the gneiss unit. (Field notebook for scale).
Figure 11 – Close up of one of the larger dolerite dykes, located north of Priomh Loch Mor. This dyke showed clear flow differentiation, with the central portion having a much higher concentration of small feldspathic minerals compared to the edges. (Field notebook for scale).
Figure 12 – Thin-section GC-4-2013: Sandstone (Quartz Arenite) viewed in XPL with quartz crystals showing undulose extinction, relic bedding (A) and some grain boundary migration/intergrowth (B).
When viewed in PPL, the slide GC-4-2013 (Fig. 12) had a dusty brown appearance and consisted mostly of quartz crystals (average width of ~250μm), making up >90% of the thin-section. Rare small feldspars crystals and some opaque minerals made up the remaining percentage of the rock. As a result, this sandstone is a quartz arenite.
When viewed in XPL the quartz showed a variety of different features; most of the quartz grains had undulose extinction when the stage was rotated. Some grains exhibited grain boundary migration and reduction, though this was less common. In addition, some relic bedding features were observed (Fig. 12) across the section which all had the same orientation, though due to the metamorphic alteration of this lithology they lacked lateral continuity across the thin-section.
The features observed are likely to have formed due to the low grade regional and contact metamorphism. The undulose extinction of the quartz grains alludes to it being deformed plastically by dislocation processes, whereas grain boundary migration/reduction is likely a result of contact metamorphism which caused recrystallisation of the quartz grains.
Figure 13 – Thin-section PD1-2103: Peridotite viewed in XPL with an abundance of olivine crystals (A) forming the bulk of the section. Some plagioclase feldspar is observed as interstitial material (B) or oikocrysts, but this is rare.
The thin-section PD1-2103 (Fig. 13) shows a relatively equigranular olivine rich peridotite, with a low proportion of feldspathic minerals (<10%). Both ortho/clinopyroxene were seen (~15%) and some small opaque minerals were also observed.
The bulk of this rock is made up of euhedral/subhedral crystals that exhibit abundant fractures and no cleavage, with interference colours ranging from the 2nd up to 3rd order blue/green. These crystals are olivine (Fig. 13) and make up ~70-80% of this section, making this peridotite lherzolitic in composition overall.
Some of the material within the olivine fractures and between crystals showed some of the properties of serpentine (Fig. 14); light green pleochroism, mesh like texture, 1st order interference colours. This could have been formed by contact metamorphism after the initial crystallisation of this rock.
Some small crystals of feldspar were seen with this section, most being plagioclase due to the diagnostic lamellar/polysynthetic twinning. In one part of the section the plagioclase was observed as an oikocryst, containing some smaller olivine’s as chadacrysts (Fig. 13), but this was not observed again across this section.
There were also occasional opaque inclusions observed (Fig. 14) in some of the minerals, this was likely chromite (or similar spinel group mineral) due to the ultramafic composition and geological context of Rum.
This peridotite likely originated from a primitive magma source and had undergone only a small amount of fractional crystallisation, much less than all the other igneous units observed on Rum. However, due to the low percentage of plagioclase feldspar in this section it is likely close to the transition between the peridotite and feldspar-poor gabbro lithologies.
Figure 14 – Thin-section JM-A1: Peridotite viewed in PPL to show the clear bands/layers (A) of opaque minerals (Chromite) that were uncommonly observed in this lithology. (B) shows the light brown/green serpentine that was observed in some of the fractures within the olivine crystals.
The thin-section TW.16-G80 (Fig. 15) was identified as a feldspar-poor gabbro, due to its increased proportion of feldspathic minerals (~30%) compared to the peridotite (<10%). The mineral assemblage identified in this section was similar to thin-section PD1-2103, with the main difference being the mineral ratios relative to one another.
The crystal size was coarser grained than the peridotite, with the overall texture of the section varying between equigranular and poikilitic in places. In (Fig. 15) we see this poikilitic texture, displaying larger plagioclase feldspar oikocrysts containing smaller olivine and pyroxene chadacrysts.
Figure 15 – Thin-section TW.16-G80: Feldspar-poor gabbro viewed in XPL showing a variety of different minerals; with large well-formed pyroxenes (A), plagioclase feldspar oikocrysts exhibiting clear lamellar/polysynthetic twinning (B), and small olivine assemblages (C).
Due to the poikilitic texture, the order of crystallisation can be established with a higher degree of certainty, as crystals enclosed within another must have been the first to crystallise. This means the olivine and pyroxene will have been the first to crystallise out of the melt, and as the magma evolved with increasing fractional crystallisation, the plagioclase feldspar would have then formed (later stage).
It was also determined that the plagioclase feldspar crystals were bytownite. This was done by using the Michel-Levy method to establish the extinction angles relative to the lamellar twinning, which gave an average An% of ~70%.
The thin-section JM.16-J1 (Fig. 16) was identified as a feldspar-rich gabbro (troctolite) due to the very high proportion of plagioclase feldspar observed (~70-80%). Again, the overall mineral assemblage was similar to that of the FpG and RP lithologies, but the proportions of pyroxene and olivine were much lower in this section.
Olivine and pyroxene crystals were observed in this section but in much lower quantities than seen in the feldspar-poor gabbro (Fig. 16). The olivine crystals made up ~10% of the composition and were mostly euhedral, additionally they displayed the alteration to serpentine within the internal fractures as seen before.
Figure 16 – Thin-section JM.16-J1: Feldspar-rich gabbro viewed in XPL showing a high proportion of euhedral elongate feldspathic minerals (A), some small olivine crystals (B), and a large oikocryst of pyroxene (C).
The pyroxene crystals made up ~5% of the section and were observed only as a few oikocrysts (Fig. 16), with most being clinopyroxene due to the higher order interference colours and inclined extinction angles. Also, the crystals were mostly sub/anhedral and contained multiple smaller plagioclase chadacrysts within them.
The feldspar crystals were mostly elongate in shape and euhedral, with evidence of overgrowth rims and grain boundary migration observed, likely due to a slower cooling rate compared to the other igneous units. The cleavage was poor in most of the crystals and the twinning ranged from simple to lamellar in form, though the majority had simple twinning only. A cumulate texture was clearly observed in this section and is likely the result of crystal settling within a magma chamber. In places the plagioclase crystals had some preferential orientation and appeared to form layering, though in other parts they were grouped radially (Fig. 16).
The overall texture of this section most closely resembles an adcumulate due to the compaction of the cumulate pile (in the parts where the plagioclase crystals were layered), the overgrowth rims commonly observed, and that the interstitial space/minerals had mostly been replaced by plagioclase. The occasional oikocrysts of pyroxene does infer the possibility of this being a heteradcumulate though.
Figure 17 – Thin-section GN1-2103: Gneiss viewed in XPL showing elongate and altered mica minerals (A), recrystallised/migrated quartz grains (B), and feldspars with both lamellar and cross-hatched twinning (C).
The thin-section GN1-2103 (Fig. 17) was identified as a high-grade metamorphic rock in the field, more specifically it is a feldspathic gneiss from the granulite facies. The mineral assemblage includes: quartz, plagioclase and orthoclase feldspar, microcline, and biotite mica (some retrograde).
The overall crystal size was quite coarse throughout the section, with the majority being sub/anhedral, likely due to metamorphism that this rock has been subject to. There was a distinct lack of foliation or tectonic cleavage in the gneiss (Fig. 17), which alluded to high grade regional metamorphism being the dominant process that formed this rock.
There was a small amount of quartz within the gneiss, which made up <10% of the section overall. It exhibited gradual undulatory extinction and grain boundary migration/intergrowth, all likely because of the high-grade metamorphism.
Microcline, plagioclase and orthoclase feldspar were observed with varying degrees of alteration, and made up ~70% of the section. The distinct cross-hatched twinning seen (Fig. 17) is diagnostic of microcline, and it would have only formed due to high grade regional metamorphism.
The biotite mica crystals in this section appeared to have undergone retrograde alteration, as the typical features of the mineral were not always present. In addition, some of the biotite crystals did form weak foliation in parts of the section, but this was uncommon. The retrogradation in the mica could have been due to later stage contact metamorphism from the other intrusive igneous units on Rum.
The thin-section GC-2-2013 (Fig. 18) displays a chaotic breccia with varying different types of clasts and textures. Some of these clasts took up large portions of the thin-section (eg: Clast B is ~6.5mm wide) and were of sedimentary, igneous and metamorphic origin.
Figure 18 – Thin-section GC-2-2013: Volcanic Intrusive Breccia (Am Mam Breccia) viewed in XPL, displaying a variety of different clasts in a chaotic matrix. (A) shows a large amalgamation of clasts with varied texture and composition, with elongate randomly orientated feldspars mixed with larger oikocrysts of pyroxene and hornblende. (B) is a ~6500μm width clast with large feldspar phenocrysts (C) within a fine groundmass.
The portion marked with (A) shows a chaotic amalgamation of minerals and clasts. Subhedral elongate crystals with 1st order interference colours and simple/lamellar twinning are plagioclase, which showed no discernible preferred orientation. In PPL some light green coloured pleochroic minerals were visible, which had 2nd order blue/green interference colours and 2 planes of cleavage, they also had a subhedral hexagonal shape; this is hornblende. Pyroxenes, quartz and rare olivine was also observed across the section.
Clast B exhibited plagioclase feldspar, quartz, biotite and hornblende phenocrysts within a fine aphanitic matrix (Fig. 18) The matrix appeared to be made up of volcanic glass and small quartz crystals, making this clast much more felsic than the others in this section. It is likely a rhyodacite clast, due to its composition and extrusive texture.
The matrix of this breccia was of igneous composition and phaneritic overall, with little of it being aphanitic or glassy. This suggests that this breccia is likely of intrusive origin; whereas the similar RVBt unit had features that pointed towards an extrusive volcanic or collapse origin. This is the major difference between those units.
Figure 19 – Thin-section GC-7-2013: Dolerite viewed in XPL, showing a fine grained crystalline texture with several minerals (olivine, pyroxenes and feldspars) large enough to be identified. Little to no groundmass observed.
The thin-section GC-7-2013 was identified as Dolerite (or diabase) and was taken from a fine grained phaneritic dyke. It displays an ophitic texture; with eu/subhedral randomly orientated feldspars amongst interstitial sub/anhedral pyroxene and olivine. The average crystal size (long axis) for this section was ~150μm, making it much finer grained than the majority of the other intrusive igneous units observed.
Doleritic dykes like this one were less common than basaltic dykes on Rum, however there was not a basalt section available. It would have had a different appearance in thin-section to this, likely being aphanitic overall and displaying occasional phenocrysts within a fine groundmass, though the mineral assemblage would have been very similar.
The Isle of Rum has an abundance of igneous intrusions, the most notable being dykes. These vary in size from small scale dykelets (<10cm) to large scale dykes that can be several metres in size (largest observed was ~30m long). Most of the dykes found on Rum are mafic in composition, with both aphanitic (basalt) and phaneritic (dolerite) types seen, though basaltic dykes were more common. The main proportion of the dykes were found within the Central Layered Intrusion, though all lithologies had dykes intruded within them (except the peridotite unit).
There were also cross-cutting relationships found in several locations, which evidenced that there were differing dyke swarm phases that occurred over time. Establishing the relative ages from the cross-cutting dykes was relatively simple in individual cases, however in certain outcrops the smaller basaltic dykes would be younger, and in others the larger doleritic dykes were younger. This suggested that there were multiple dyke swarm phases and not just the major two groups that were originally identified.
The basaltic dyke group (Fig. 20) were more widely distributed in terms of dip and strike than the dolerite dykes, but the averaged trend was broadly N-S. These are likely to be radial dykes associated with the Rum Central Complex, but there is not enough evidence from the field data to definitively prove this, as they could just be several smaller dyke swarm phases.
The dolerite dyke group (Fig. 21) shows a clear NW-SE trend with few anomalous results, this correlates directly with the regional trending dykes across the Hebridean complexes (Speight et al., 1982). These dykes are therefore unlikely to be related to the Rum Central Complex.
Most of the igneous units on Rum exhibited layering in some form, with the most prevalent being in the feldspar-poor gabbro and LGX outcrops. Layering was also observed in the peridotite and feldspar-rich gabbro lithologies, but it was less common.
The layering observed across the Northern Marginal Zone (NMZ) was infrequent and as a result the dataset for Fig. 22 is rather small and is probably not representative of the overall layering trend in the Central Layered Intrusion. The data does show that the igneous layering dips moderately in a NE-SW trend, which is slightly different from the N-S trend that would be expected for this area of the NMZ. The LGX unit, located on the west side of Long Loch, had layering with a shallow to moderate dip in an easterly direction. This is anomalous when compared to the rest of the data; but it could be explained by displacement from the inferred Long Loch fault.
The Rum Sandstone unit is located in the northern portion of the mapping area covered. The dataset for the sandstone is small due to the difficulty of establishing bedding planes in the field, as most of the outcrops observed had been subject to moderate levels of alteration/metamorphism. Overall the unit dipped in a west/north-west direction (Fig. 23), with shallow dip angles in the north, that increased with movement south towards the Main Ring Fault (MRF).
Fig. 23 shows a few red coloured planes, which were taken from outcrops in close proximity to the MRF. In general, these had a steeper angle of dip which was orientated in a more northerly direction. This is likely to have been caused by resurgence or doming of the Rum caldera and the associated uplift that occurred, which would have displaced the surrounding country rocks (RS, HSM and RGn units).
Allmendinger, R., Cardozo, N. and Fisher, D. (2013). Structural Geology Algorithms. Cambridge: Cambridge University Press, p.289.
Cardozo, N. and Allmendinger, R. (2013). Spherical projections with OSXStereonet. Computers & Geosciences, 51, pp.193-205.
Emeleus, C. and Troll, V. (2008). Geological excursion guide to Rum. Edinburgh: NMSE – Pub.
Emeleus, C., Bell, B. and Stephenson, D. (2005). The Palaeogene volcanic districts of Scotland. 4th ed. Nottingham: British Geological Survey.
Speight, J., Skelhorn, R., Sloan, T. and Knaap, R. (1982). The dyke swarms of Scotland. In: D. Sutherland, ed., Igneous rocks of the British Isles. Chichester: John Wylie & Sons.