Isotopic and Geochemical Data.
The elvan dykes.
The Cornubian Batholith of South West England underlies the counties of Cornwall and Devon, running down the axis of the peninsula for in excess of 200 km. It is exposed onshore in five major plutons (east to west: Dartmoor, Bodmin Moor, St Austell, Carnmenellis and Land's End) and eleven satellite intrusions (east to west: Hemerdon Ball, Hingston Down, Kit Hill, Belowda Beacon, Castle an Dinas, Cligga Head, St Agnes, Carn Marth, Carn Brea, Tregonning-Godolphin and St Michael's Mount) of Lower Permian age (Chen et al., 1993). It outcrops further west at the Scilly Isles (28 km west of Land's End) and beyond (gravity surveys in 1963-65 by the Bedford Institute of Oceanography indicated a 100 mile seaward extension of the batholith). The Haig Fras granite bosses, 95 km WNW of the Scilly Isles, out in the Western Approaches are also of Variscan age (Evans, 1990) though appear to represent a separate plutonic body. The batholith intrudes a succession of deformed, low-grade, regionally metamorphosed sediments and igneous rocks (Edmonds et al., 1975) of Devonian and Lower Carboniferous age (Figure 1).
Figure 1. A simplified geological map of Cornwall, showing the major and minor granite outcrops.
The rocks of the batholith are granitic (Exley and Stone, 1964; 1982) in nature and their origin is related to the later stages of the Variscan Orogeny (late Carboniferous) that had previously deformed (Alexander, 1997) and metamorphosed the sedimentary pile. The batholith is also highly mineralised (Dines, 1956) and this mineralisation has been exploited by deep mining continuously for the last 400 years within local records, and for some 2000 years prior to that by shallow surface mining and working placer deposits. This area has been used as a model for vein mineralisation and contributed significantly to the understanding of ore forming processes.
The batholith has been extensively studied since the Geological Survey first entered the region in 1836 (the publication of the first memoir, by H.T. De La Beche, was in 1839). The excellent coastal exposures, coupled with those afforded by mining activity, have allowed the rocks to be studied at outcrop at a number of levels, from surface to depths of ~1000 metres. Today the petrological characteristics, isotope and whole rock geochemistry, mineralogy and form of the batholith are well known. Significant recent advances in the field of geochronology have allowed us to fairly well constrain the time frame of the intrusion of the batholith and the subsequent mineralisation. However; despite being the subject of such intense study, important questions relating to the origin of the granite, its mode of emplacement and the origin of the economic mineralisation remain unanswered. Current research is being directed into the fields of magma generation and mixing and emplacement tectonics within the Cornubian Batholith.
Back to top of Page.
The Space Form of the Batholith.
1. Gravity Modelling.
The first substantial geophysical work to be undertaken across Cornwall was a pendulum gravity survey by Bullard & Jolly (1936). This survey established the pattern of bouger anomalies (with pronounced negative anomalies over the granite outcrops) across the peninsula, but made no interpretation of the results. This work was followed up by a major survey (acquiring gravity and magnetic data) in the late 1950's covering Cornwall, Devon and Somerset. The results of this survey were published in Bott et al. (1958) and this seminal work has formed the basis of the gravity interpretation of the batholith to the present day.
Bott and his co-workers established a chain of gravity stations across the southwest, with a series of regional centres to aid in local corrections of data. Utilising a set of over 400 density measurements they established the density contrast between the granites and metasediments (granites ranged from 2.58 - 2.64 g/cm3; metasediments averaged 2.75 g/cm3, though none of these samples was from Cornwall) to be ~0.10 g/cm3. This density contrast is of great importance in the development of any model of gravity data as it is used to calculate the overall thickness of the granite batholith and also to infer the angle of the granite contacts on the margins of the individual plutons (as it affects the contouring of the anomalies). In the processing of the raw data a model, based on geological observation at surface, is created and the area under study broken down into a set of polygonal units. The slope and 'break points' on the processed curves give the direction of contact dip and the density contrast, used in the processing calculations, defines the angle of dip and the thickness of the polygonal block, which is usually modelled as having a flat base. Changing the density contrast to a higher value results in a steeper contact angle and a thinner modelled slab (Kearey and Brooks, 1991), without any compromise of the original raw data; therefore a number of possible models can be developed from the same data depending on the density contrasts used.
The model produced by Bott et al. (1958) placed the base of the granite batholith between 8 and 20 km below surface (dependent on the density contrast) and divided it into three major units; Dartmoor, Bodmin Moor-St Austell (or Hensbarrow) and Carnmenellis-Land's End. They found that the pronounced negative bouger anomalies were coincident with the major plutons and described a belt running down the axis of the Cornubian Peninsula (see Figure 2). They found that the granite contacts dip outwards at varying angles and that most of the plutons are connected to each other (something which had been postulated as early as 1837 by Sedgwick and Murchison) by narrow ridges that lie at no great depth below the surface (often in the region of 1 - 1.5 km); the exception being between the St Austell and Carnmenellis granites, where a deep trough separates the two plutons. These two bodies were also found to have a large areal cover of roof rocks, the actual outcrops making up only the southern part of each pluton. The St Austell granite also encompasses the outlying Belowda Beacon and Castle-an-Dinas granites, and the Carnmenellis granite is connected to the Cligga Head and St Agnes granites by a broad shelving subcrop that reaches the north coast.
Figure 2. The pattern of Bouger anomalies across southwest Cornwall. Note the northerly extension of the Carnmenellis granite to the north coast and the submarine extension of the Land's End Granite westwards. After the IGS 1:250000 Bouger gravity anomaly map sheet 50°N-06°W, 1975.
One other important factor of this model (Bott et al., 1958; Bott and Scott, 1964) is the southerly disposition (with respect to an individual pluton) of some of the anomaly minima (which equate to the areas of maximum 'slab' thickness). In the cases of Bodmin Moor and Dartmoor this is particularly pronounced and was accompanied by data that suggested, with a uniform density contrast, that both plutons had bases that thinned (by between 33 and 50%) northwards. In the case of Dartmoor (Hawkes, 1982) evidence for this southerly deeper 'cylindrical' feeder zone was substantiated by Brammall (1926) who, studying feldspar phenocryst orientations, postulated that the granite rose vertically in the south before flowing horizontally to the north. In both of these cases, to create a flat base to the modelled pluton it would be necessary to reduce the density of the granite across the northern area (see Figure 3), a proposition which would infer a composite make-up for the plutons in question.
Figure 3. A section through the Bodmin Moor Granite with the corresponding gravity data. In this example a change of density over the north of the granite has been used to create a flat base for the pluton at 12 km. Without this correction (i.e., using a uniform density contrast) the pluton would thin ('step up') to the north. Note that the underlying gravity profile would support either option. After Bott and Scott, 1964.
Without these density corrections the plutons would assume the form of a stepped laccolith, with a southerly disposed vertical feeder pipe (see Figure 4). Intrusions with very similar forms have been modelled in the laboratory (Roman-Berdiel, 1999) in analogue experiments of emplacement into brittle, shallow, crust under extensional conditions. In these experiments the 'plutons' emplaced into extending crust with a top sense of shear to the south were extended northwards with feeder zones located towards the south in exactly the same form as seen in Figure 4.
Figure 4. A section through the Bodmin Moor Granite with the corresponding gravity data. In this example a uniform density for the granite is assumed. The resulting laccolithic shape of the pluton has been suggested as the true form of both the Bodmin Moor and Dartmoor granites and closely resembles analogue models of shallow plutons emplaced in extensional regimes. Modified after Bott and Scott, 1964.
The similarity of the modelled scenario with the tectonic history of Cornubia, field observations and equivocal gravity data suggest that this is a valid counter-interpretation to the flat-bottomed batholith (for Dartmoor and Bodmin Moor at least) which is currently in vogue. The application of AMS and 3D seismic inversion techniques at some time in the future will be necessary to deduce the deeper form of the batholith and its component plutons.
Further density data collection was undertaken in the 1970's (McCann, 1973); together with a further onshore regional gravity survey (completed by the IGS in 1973) which added significantly to the earlier 1950's dataset. This information was extended by the use of localised detailed surveys and by data collected from boreholes (Beer et al., 1975); the results were published in Tombs (1977).
Tombs' model used a density contrast of 0.10 g/cm3 as a starting point (with minor alterations in selected areas) and was constructed from the data in a series of polygons. A total of 40 runs of the computer program were undertaken to achieve the final model, which viewed the batholith as a wedge-shaped body with its base some 20 km deep below Dartmoor, thinning to around 15 km below Land's End and around 10 km at the Scilly Isles. Changes in density to the east or changes in the background anomaly would modify this model significantly, but both were rejected in favour of the 'wedge' model.
There are no major differences in the form of the near-surface (<2 km) batholith between the models of Bott et al. (1958) and Tombs (1977), but Tombs' model benefits from greater sample density, particularly around the St Agnes, St Austell and Hingston Down areas and between the Tregonning-Godolphin and Carnmenellis granites (Beer et al., 1975), allowing greater definition of the connecting 'ridges' between the major plutons and the shallow subcrop extensions of the St Austell and Carnmenellis granites.
Willis-Richards (1990) reworked Tombs' data and developed a computer program to create a new model, building in assumptions established by seismic studies (Brooks et al., 1984; Doody, 1985) about the position of a major crustal reflector (R2) beneath Cornwall that had been correlated with the floor of the batholith. Using a more refined density dataset (augmented by data from the Hot Dry Rocks Geothermal Energy Project based at Rosemanowes [SW736346] on Carnmenellis) he ran a series of models with density contrasts between 0.07 and 0.13 g/cm3 (taking account of density changes with depth). The final version used in his thesis (model 17), used a contrast of 0.118 g/cm3 and gave depths of 14-17 km to the batholith floor, which provided the best fit with the R2 reflector. Using this program he was able to model the batholith in three dimensions (Willis-Richards and Jackson, 1989; Willis-Richards, 1990) and establish a total volume for the batholith of 68,000 km3.
Willis-Richards' model (see Figure 5) gave the batholith a floor that sloped down to the south (coincident with the R2 reflector) and averaged 13.5 km deep below Land's End, Carnmenellis, St Austell and parts of Bodmin Moor. The base under the northern parts of Bodmin was shallower (~10 km); while under southern Dartmoor it reached ~17 km. In this model the batholith can be divided into two sections - the eastern section (Dartmoor, Bodmin Moor, St Austell) and the western section (Carnmenellis, Land's End, Scilly), which are divided by a 3 km deep fault-controlled trough that runs beneath the Truro district. Except between St Austell and Bodmin Moor, the cupolas of each section are connected by saddles no more than 2 km below the surface (with pronounced, steep, south-facing embayments). In some instances sections of these saddles are exposed as minor granite outcrops (e.g. Kit Hill and Hingston Down form a saddle ridge between Bodmin Moor and Dartmoor). Elsewhere the minor granite outcrops represent the apices of narrow tubular bodies connected to the main granite at depth (St Michael's Mount, Cligga Head and Hemerdon Ball). The overall appearance of the batholith is a tabular body, which is 40-60 km wide at its base, with steep sides, a sloping base (possibly tilted 2-3° further south during post-emplacement movements) and an irregular upper surface.
Figure 5. Isometric diagram of the space form of the batholith. From Willis-Richards and Jackson, 1989.
Post-emplacement movements along NW-SE trending wrench faults are postulated to have broken the batholith up into a series of blocks (Dearman, 1963; Willis-Richards and Jackson, 1989) with Dartmoor translated 34 km (Dearman, 1963) to the southeast - restoration of Dartmoor to its original position would see its southern margin on the same latitude as the southern margin of Bodmin Moor. Isostatic movements along these faults, in response to mass compensation (Bott et al., 1958), have also been used to support the model of the westward-thinning batholith (Tombs, 1977).
Back to top of Page.
2. Seismic Data.
Seismic reflection and refraction studies across Cornubia and the Western Approaches from the 1970's onwards have identified a number of major crustal reflectors both on and offshore (Bott et al., 1970; Brooks et al., 1984; Doody, 1985; Camborne School of Mines, 1988; Evans, 1990). Beneath the Cornubian Peninsula three major reflectors have been identified by seismic refraction methods (Brooks et al., 1984; Doody, 1985): R1 marks the top of a low velocity zone and lies at depths of 7-8 km; it is restricted to the interior of the major plutons and has been interpreted as the top of a zone of xenolith accumulation (see Figure 4.6) within the batholith (Bromley and Holl, 1986), but this is as yet unresolved. A second, lower, southerly dipping R2 reflector lies between 10-15 km depth and extends beyond the northern margin of the batholith (Brooks et al., 1984).
Figure 6. A section through a typical Cornubian pluton, showing the position of the R1 and R2 reflectors bounding the 'zone of xenolith accumulation'. Areas of textural zonation are also shown. After Bromley and Holl, 1986.
The R2 reflector has been widely equated with the base of the batholith; it is interpreted as a thrust plane of major regional significance (the abrupt velocity change across the reflector has been stated (Willis-Richards and Jackson, 1989) as typical of a granite to schist boundary) and has been cited in some models (Shackleton et al., 1982) as playing a major role in the emplacement of the Cornubian granites by acting as a tectonic pathway for the south-to-north movement of the batholith as a whole into its present position.
A third reflector (R3) is found at depths of 27-30 km and is taken to be the MOHO, which maintains this depth under Cornubia, rising to 13 km in the Western Approaches. The average seismic velocity within the granite (see Figure 6) above the R1 reflector is 6.0 km/sec. Below R1 it falls abruptly to 5.6 km/sec, rising to 6.5 km/sec below R2 and 8.0 km/sec below R3 (Brooks et al., 1984; Bromley 1989).
A fourth reflector was detected a little above R2, at ~12 km, below Carn Brea, but was discontinuous (Brooks et al., 1984) and has largely been disregarded. It may represent a splay from the main R2 thrust or part of R2 offset by later faulting.
Seismic reflection studies (Camborne School of Mines, 1988), undertaken as part of the Geothermal Energy Project, failed to pick up the R1 and R2 reflectors or any other reflectors in the granite. The granite/killas boundary was also transparent due to poor impedance contrasts between the two rock types. A number of smaller reflectors were picked up in the killas and differentiation between the granite and metasediments was largely based on the absence of these structures in the plutons themselves.
Back to top of Page.
3. Granite Classification.
The dominant rock type exposed (accounting for over 90% of the total outcrop) in the Cornubian Batholith is a medium to coarse grained two mica granite (see Figure 7) which is commonly megacrystic, though the size, proportion and alignment of these megacrysts is highly variable (they have a mean size range of 1 - 4 cm, but may reach up to 17 cm in length). These are set in a groundmass that has a mean grain size of 3 mm (Stone and Exley, 1985).
Exley and Stone (1982) classified the granites into six petrographic categories (see Table 1). Type A identifies rafts and inclusions of rocks ranging from granodiorite to quartz diorite. These may be of early igneous origin, but some authors suggest they may represent partially assimilated masses of stoped country rocks. The dominant rock type in the major plutons (Type B) is the megacrystic biotite-muscovite monzogranite described above. Fine-grained porphyritic and non-porphyritic granites (Type C) usually make up less than 10% of the major plutons at outcrop. Lithium enriched granites (Type D) only occur to any extent in the St Austell and Tregonning-Godolphin plutons, while topaz and fluorite granites (Types E and F) are restricted to the western lobe of the St Austell Granite.
Figure 7. The distribution of granite varieties across the Cornubian Batholith. After Hawkes et al., 1987.
Table 1. The granite classification system of Exley and Stone (1982).
A second classification scheme was put forward by Dangerfield and Hawkes (1981). This scheme (see Table 2) is based on the size and abundance of the K-feldspar megacrysts and on the average grain size of the groundmass.
Table 2. The granite classification system of Dangerfield and Hawkes (1981).
It also takes into account the specialised Lithium granites (Hawkes et al., 1987). This scheme has particular significance for the mapping of various megacrystic and equigranular granites within the major plutons. It has been found that coarse-grained megacrystic granites (megacrysts > 15 mm length, types 1A and 1B) dominate the Dartmoor, St Austell and (the Northern and Southern lobes of) Land's End granites. Coarse grained megacrystic granites with small megacrysts (< 15 mm length, type C) are characteristic of the Bodmin Moor (Exley, 1996), Carnmenellis and Scilly Isles plutons (Stone, 1995).
Detailed physical and geochemical mapping of the various plutons has shown that the relationships between the granite facies are a result of a combination of fractionation and intrusive events (Knox and Jackson, 1990; Stone, 1995; Powell et al., 1999a, 1999b; Stone, 2000). Some supporting evidence that the type 1A and 1B granites are roof facies of the major plutons comes from the vertical distribution of granite types. On Dartmoor the coarse megacrystic variants are more common at higher topographic levels, with a reduction in size and number of megacrysts in the deeply eroded river valleys (Exley and Stone, 1985, 1986).
In the Rosemanowes HDR borehole, in the Carnmenellis pluton [SW736346] there was an observed reduction in megacryst size, distribution and groundmass size with depth, from coarse megacrystic granite (type 1C) at surface, through to equigranular granite below 2000 m depth (Willis-Richards and Jackson, 1989: Bromley, 1989).
Based on the limited outcrop exposure, it has been suggested that the coarse, megacrystic granite (type 1A) forms a relatively thin carapace at the top of the Dartmoor, (sections of) St Austell and Lands End plutons. It grades downwards, within a few hundred metres, to type 1B granites and thence into type C granites at depth. Types 1A and 1B granite have been eroded away in the Bodmin Moor (Selwood et al., 1998; Exley, 1996), Carnmenellis and Scilly Isles masses or they may never have developed (Exley and Stone, 1985, 1986).
The Type 2A granites include Exley and Stone's Type D, E and F granites. The topaz granites (Type E), seen in the St Austell and Tregonning-Godolphin masses are considered to be products of extreme differentiation, while the type D (lithionite) and type F (fluorite) are presumed to be of metasomatic origin; derived, respectively, from biotite and topaz granites by hydrothermal alteration and ionic exchange (Manning and Exley, 1984). Type 3, fine-grained megacrystic and equigranular granites form restricted outcrops in all the major plutons. Crosscutting relationships indicate that they represent a later intrusive phase, coeval with the elvan dykes of the region (Exley and Stone, 1985, 1986).
Back to top of Page.
The mineralogy and petrology of the granite types has been described in detail by Stone and Exley (1985, 1986), and Exley and Stone (1964, 1982).
Approximately 90% of the total outcrop of the Cornubian Batholith consists of medium to coarse-grained biotite granite with a mean grain size of 3 mm (Stone and Exley, 1985). This contains varying proportions of K-feldspar megacrysts that typically range from 10 to 40 mm in length, but can reach 200 mm in exceptional cases.
The K-feldspars in the groundmass may be either monoclinic or triclinic and both are perthitic. The megacrysts are monoclinic. They typically enclose other minerals, especially plagioclase and quartz and make up, on average, 32% of the mode. The origin of the megacrysts has been the subject of some debate. Ghosh (1934) described them as magmatic phenocrysts. Exley and Stone (1964) suggested they were late subsolidus features produced by potassium metasomatism. Hawkes (1967) proposed that they replaced earlier plagioclase crystals (a view supported by Edmonds et al. (1975), for the megacrysts of the Dartmoor granite). Leveridge et al. (1990) working on the Carnmenellis Granite proposed a primary igneous origin for the megacrysts. They found no evidence of replacement of plagioclase by K-feldspar (though they did find the reverse) or metasomatic effects within the feldspars, though they did note that some of the quartz shows replacement textures.
Plagioclase is generally euhedral to subhedral (anhedral in some of the Carnmenellis granites) and is commonly twinned on some combination of the Carlsbad, Albite and Pericline laws and is typically zoned from An25-30 in the cores to An8-4 in the rims (Stone and Exley (1985), report values of An20-32 in the cores and An6 close to the rims of plagioclase crystals from Carnmenellis). The average plagioclase content of these rocks is 22%.
Quartz tends to be equigranular, though irregularly shaped and is commonly interstitial. The grains are frequently strained and contain tiny inclusions, often in trains. Occasionally, each quartz 'grain' is found to consist of an aggregate of rather irregularly outlined, but separate, interlocked crystals. These agglomerated crystals (some of which may be secondary) may reach up to 40 mm in size. Quartz makes up about 34% of the rock in an average mode. Biotite amounts to some 6% of the mode and is subhedral, often in clusters of flakes and sometimes intergrown with muscovite. It has a 2V less than 15 degrees and a typical lithium content of 1000 - 2000 ppm and can reach up to 5000 ppm on occasion (Exley and Stone, 1986).
Muscovite is also subhedral, but is usually subordinate to biotite, giving an average modal value of 4%. The total mica content of the granites is always close to 10%. In Dangerfield and Hawke's classification (1981) the biotite content of Type 1A granite is typically above 5%. In Type 1C it is about 5%. In Type 2B, 2C and Type 3 it is always less than 5%. The biotite/muscovite ratio in Type 1A is greater than 2.In Type 1C it is 1 and in 2B, 2C and Type 3 (medium and fine grained granites) it is commonly less than 0.5.
Biotite occurs in two main morphologies. It may form single euhedral crystals (presumed to be of magmatic origin), up to 5 mm in length, in the megacrystic and equigranular granites (it also forms euhedral inclusions in the K-feldspar megacrysts, plagioclase and quartz) or it may form aggregates within the megacrystic granites. These aggregates are spatially associated with andalusite or cordierite, which Charoy (1986) considered to be of restite origin. The compositional differences between the two forms are negligible, and this has been attributed to magmatic re-equilibration during crystallisation and subsolidus hydrothermal activity (Exley et al., 1983).
In the Carnmenellis pluton, Type 1C granite some 50 - 60% of biotite occurs in aggregates with partially muscovitised andalusite, apatite and ilmenite (Leveridge et al., 1990). Less commonly it is associated with altered cordierite and heavily sericitised, anhedral plagioclase. These aggregates are purely interstitial and show all gradations between individual aggregates of two to three crystals and large pelitic xenoliths with relict sedimentary structures and slatey cleavage.
These xenoliths can be directly matched with the hornfelsed slates in the contact aureoles (Lister, 1984), however, some work suggests that at least some of the xenoliths from the Carnmenellis granite were entrained at depth and are not the product of high-level stoping (Jeffries, 1985). Xenoliths are common in the Type 1A granites, rare in Types 1B and 1C and absent from Types 2B, 2C and 3. In the Dartmoor granite around 70% of biotite occurs as aggregates in the Type 1A granite. In the Rosemanowes borehole the Type 2B granites carry only 20% of their biotite as aggregates and the equigranular Type 3 granite carries biotite as single crystals only (Bromley, 1989).
This pattern of distribution has been used to suggest that the biotite aggregates are the product of assimilation and disaggregation of pelitic xenoliths rather than being of restite origin. Bromley (1989) suggests that perhaps only 2-3% modal biotite in the granites may be of direct magmatic origin, the rest being inherited from metamorphosed supracrustal rocks.
Brown tourmaline (schorl to dravite in composition), often with blue patches makes up a further 1% of the megacrystic granites (Exley and Stone, 1985, 1986) in which it is a common accessory. This is primary tourmaline in contrast to the later pneumatolytic tourmalines that are the products of boron metasomatism and first stage mineralisation processes. Other minerals, also amounting to a total of about 1%, include cordierite, pink andalusite, apatite, zircon, monazite, thorite, xenotime, uraninite (Jeffries, 1984) and ore minerals including copper, zinc and iron sulphides. Rare garnets also occur. These have been described from Dartmoor by Stone (1988) where their composition (almandine, <10% spessartine) was used to infer that these garnets could potentially represent a restite phase. A recent discovery of another garnet species (a high Mn spessartine) from the Carnmenellis Granite (LeBoutillier et al., 1999; Wheeler et al., 2000) is indicative of a primary, late stage, magmatic origin.
Cordierite is a common accessory mineral, particularly in the coarse grained biotite granites, where it is seen in crystals up to 1 cm (almost always altered to pinite). It is usually interpreted as a restite phase, being inherited from pelitic xenoliths entrained at depth, but Clarke (1995) has shown that magmatic cordierite may form in granites at high crustal levels under the right chemical and temperature/pressure conditions; indeed he argues against a restite origin for the origin of cordierite in granitic rocks emplaced at high crustal levels due to the minerals instability at anatectic depths and conditions.
The fine grained granites (Type C/Type 3) and the topaz, Li mica, and fluorite granites (Types D, E and F/2A) make up the remaining 10% of the exposed plutons. Type C (Type 3) granite occupies most of the 10% of non-Type B (Types 1A - 2C) areas. It is medium to fine grained, the groundmass grain size averaging about lmm, but may also be somewhat megacrystic. K-feldspar is subhedral to anhedral, sometimes microperthitic, and averages 30%. Plagioclase is euhedral to subhedral, normally zoned from An10-15 to about An5 and amounts to an average 26%. Biotite resembles that in Type B, but is less plentiful at about 3%, while muscovite, also similar to that in Type B, is more abundant at about 7%. The proportion of all other minerals is similar to those of Type B, apart from size, although zircon, apatite and monazite are less abundant, and fluorite, which is rare in Type B, is more common (Exley and Stone, 1986).
Type D is a megacrystic, medium to coarse-grained rock in which the K-feldspar averages 27%. Megacrysts of this mineral have a mean length of about 2 cm, but may be as long as 8.5 cm. The groundmass K-feldspar is euhedral to subhedral and microperthitic, constituting up to 26% of the rock. Plagioclase is euhedral to subhedral and is mostly unzoned with a composition of An7. Quartz makes up some 36% of the rock (sometimes as aggregates of crystals) and primary tourmaline makes up 4%. In this granite the biotite and muscovite micas have been replaced by a lithium bearing variety. This was formerly referred to as 'lithionite', but is now known to be zinnwaldite. It is bronze-brown, subhedral, has a small 2V and makes up 6% of the mode. Other minerals, totalling about 0.5% on average, include apatite and topaz as important phases, with small amounts of ore and occasional fluorite. This type only occurs within the St Austell pluton (Exley and Stone, 1986; Hill and Manning, 1987). The topaz granites contain a smaller proportion of accessory minerals, but a greater variety including Mn-apatite, amblygonite, zircon, Mn-ilmenite, columbite-tantalite, Nb-Ta rutile, arsenopyrite, sphalerite, pyrite, cassiterite and monazite (Scott et al., 1998).
Type E granite is an aphyric, reasonably equigranular rock with a similar suite of minerals to that in the Type D granites. The proportions are different however, plagioclase at 32% now exceeding K-feldspar (24%), quartz having fallen to 30%. Tourmaline (enriched in the Li bearing elbaite component, Manning, 1998) is reduced back to 1%, topaz having risen to 3%, fluorite and apatite together to 2%. There are no ore minerals recorded. Plagioclase now has a composition close to albite, typically An0-4. Zinnwaldite (9%) is similar to that in the Type D granite but is poorer in iron and richer in Li and F. These topaz granites form only 1% of the surface outcrop of the batholith (St Austell and Tregonning) but may be more common than has been realised. Topaz granite has been recorded from the dumps of Botallack Mine, near St Just, but never recorded in situ (Manning, 1998).
Type F granite is texturally very similar to Type E, but is distinguished from it by its mode (K-feldspar 27%; plagioclase 34%; quartz 30%; muscovite 6%; fluorite 2%; topaz 1%; apatite less than 1%) which includes no iron-bearing minerals, tourmaline being absent and the place of zinnwaldite being taken by muscovite. There is a marked increase in fluorite, much of which is sited in the cleavages of the mica. This type is of very limited extent and only occurs in the St Austell pluton (Exley and Stone, 1986).
A number of other granite types are present as marginal facies or dykes. Medium to fine grained biotite and/or tourmaline-bearing microgranite and fine-grained aplite dykes cut the biotite-granites and often follow joint directions in the latter. The microgranites have many of the features of the Type C granites, but the typical aplites are more leucocratic and fine grained (grain diameters less than l mm). In the Carnmenellis granite, the former commonly have Na2O/K2O (wt.%) less than 1.0, whereas the latter tend to be richer in plagioclase (and albite), having Na2O/K2O more than 1.0. The aplites are probably the products of crystallisation of residual magma occupying very early-formed joints in the biotite granites, whereas the microgranites may be related to some of the Type C granites, and could have the same diverse origins (Exley and Stone, 1985,1986).
Another group of granites is characterised by Li-bearing micas, albite and topaz. Lepidolite-albite-topaz leucogranites differ from the associated Type E granites in containing lepidolite in place of zinnwaldite, a paler-coloured tourmaline (poorer in iron) and a higher albite/potash feldspar ratio. These rocks together with lepidolite and zinnwaldite bearing aplite and pegmatite form the roof rocks of the Tregonning granite and the granitic sheets of the Megilligar Rocks (Hawkes et al., 1987) and the roof zone of the Type E granite intrusion at Gunheath (St Austell pluton).
Tourmaline is a ubiquitous accessory in most of the Cornubian granites. In some rocks it attains major proportions and gives rise to a suite of tourmaline granites. These can be divided into primary granites and those of metasomatic origin. The primary granites occur mainly as microgranite dykes in all the major plutons, consisting of a quartz-feldspar (K-feldspar and plagioclase)-muscovite-tourmaline assemblage. Primary tourmaline granite also occurs in the St Austell pluton (Hill and Manning, 1987; Hawkes et al., 1987).
Metasomatic tourmaline granites show every gradation from the above, with progressive elimination of micas and feldspars, to a quartz-tourmaline tourmalinite (some of which appear to be intrusive - products of extreme fractionation, such as the dyke at Roche Rock [SW992594] and the tourmalinite at Porth Ledden [SW354320] near Cape Cornwall). Examples include rocks such as Luxullianite and Trevalganite. Boron (and hence, tourmaline) has played a major part in the transport of metals during the high temperature phases of mineralisation (Charoy, 1982) and is a common constituent of main stage Sn lodes (Farmer, 1991) at depth (e.g., South Crofty Mine).
Back to top of Page.
5. Isotopic and Geochemical Data.
The granites of the Cornubian Batholith have a distinctive geochemical signature, typical of S-type granites (Chappell and White, 1974; Willis-Richards and Jackson, 1989). They are silica-rich (70-75% SiO2) and rich in K2O with high K2O/Na2O ratios. They have moderate to low CaO and low MgO, total FeO and Fe2O3/FeO ratios close to 0.5 and high normative corundum (1-4%). They have high concentrations (relative to most granitic rocks) of As, B, Cl, Cs, F, Ga, Li, P, Pb, Rb, Sn, Ta, U and Zn. They have less than average Ba, Sr, Ti, Y, Th and Zr. They have high Rb/Sr, Rb/Ba and Ba/Sr ratios, while K/Rb, Nb/Ta and Th/U ratios are generally low (Exley et al., 1983; Charoy, 1986; Floyd et al., 1993).
The high initial 87Sr/86Sr ratio (0.713), enrichment in 18O (13%, SMOW), relatively low sodium content, low oxygen fugacity, low ferric/ferrous ratio and presence of ilmenite, cordierite, monazite, garnet and muscovite are all diagnostic of S-type granites, as is the association with tin/tungsten mineralisation (Beckinsale, 1979).
The distribution of trace elements and various oxides within the granites can be used to observe patterns of partitioning and differentiation within the evolving magmas (Bromley, 1989). Plots of total Fe against wt% TiO2 show a negative slope developed from Type 1A and 1B granites through to topaz granites, which show the lowest concentrations of both components. A similar pattern emerges for wt% TiO2 against zirconium. Here there is a continuous negative slope from Type 1A and 1B granites, through Types 2B and 2C through granite porphyries, tourmaline granites to topaz granites, which again show the lowest concentrations. Thorium against zirconium shows the same pattern (with the Type 1A, B and C types oversaturated with respect to zircon, indicating that much of this material is xenocrystic and inherited from xenoliths or is restite), while niobium against zircon shows a strong positive correlation with the topaz granites effectively partitioning niobium while containing very low levels of zirconium. The topaz granites show the highest concentrations of rubidium and the lowest concentrations of barium and strontium (though the Type 1A and B granites show higher levels than Type 1C).
These variations can be partly explained by the movement of elements such as barium into the plagioclase lattice; more differentiated granites contain less of the calcium (anorthite) component that barium (and strontium) can substitute for, hence the concentrations are lower. Lister (1984) showed Ba and Sr can also be admitted to the magma by xenolith assimilation in the same way that most of the zircon is assumed to have been brought into the granite. Xenolith concentrations are greatest in the 1A and 1B granites and this is reflected in the geochemical profiles.
Yttrium is preferentially concentrated in the Type 1A and 1B granites and is present in apatite and xenotime (Jeffries, 1984; Stone, 1987); it is also found extensively in fluorite associated with later mineralisation. The LREE and HREE elements provide different chemical profiles between the various plutons with steeper profiles for the Bodmin Moor and Carnmenellis granites. This again could be a function of granite type exposed at outcrop; The HREE elements are selectively partitioned into zircon, which is most prevalent in Type 1A and 1B granites. The Bodmin Moor and Carnmenellis granites, showing deeper structural profiles are lacking in these granite types and hence have lower HREE concentrations (Stone 1995, 2000). LREE elements are concentrated by differentiation and so are most prevalent in the more evolved Type 2 granites and more specialised derivatives.
Fluorine is present in the Cornubian granites in varying amounts. It is concentrated preferentially in the more specialised granites and is eventually incorporated into topaz, tourmaline, apatite or fluorite, though it is also able to be accepted into the lattices of micas. The fluorine content varies from 0.20%, in the biotite granites, to 1.39% in the lithium mica granites. Late stage aplites such as those at Megilligar Rocks (1.15%) and Meldon (1.67%) also carry elevated values. Manning (1982) found that at this level of concentration the presence of fluorine could depress the granite minima, at 1 kbar, to between 690°C (1% F) and 670°C (2% F), allowing late stage crystallisation to continue through to these low temperatures (4% F would depress the minima to 630°C) which has important implications for the crystallisation and emplacement of the more specialised Type 2 granites.
The differentiation sequence biotite granite - lithium mica granite - lithium mica leucogranite - aplite exhibits increases in Na, Li, Mn, P, F, Rb, Cs, Ni and Sn; with corresponding decreases in Ti, Fe, Mg, K, Zr, Zn, Pb and As (Stone, 1982). This series of trends reflects the ability of the various elements to substitute in the lattices of the main silicate phases; the volatile group being accommodated within micas, particularly the lithium-rich varieties.
There has been considerable two-way movement of elements and compounds between the granite and the host rocks from the time of initial emplacement and throughout the period of mineralisation. Hall and Alderton (1994) found that ammonium, present in the country rocks (in amounts up to 1000 ppm ammoniacal nitrogen), was also present in the granites in varying amounts. High levels of ammonium are also present in areas of hydrothermal mineralisation and alteration (143-340 ppm in greisens) and in elvan dykes (Praa Sands elvan, 407 ppm; Swanpool elvan, 1212 ppm). In the case of the elvans and mineralised areas the ammonium is brought in by hydrothermal convection from the country rocks and fixed in secondary white micas. Elvans, often acting as pathways for hydrothermal fluids, are particularly likely to receive high levels of ammonium. The source of ammonium in the granites is either from xenolithic material, or may have been derived from the source region of original granite melts. The ammonium substitutes for potassium (up to 2%) during the growth of K-feldspar. The Bodmin Moor (97 ppm) and Carnmenellis (87 ppm) granites are the most ammonium rich. The Land's End (32 ppm), St Austell (15 ppm) and Dartmoor (11 ppm) granites all contain lesser amounts. This variation is the reverse of what would be expected as the Bodmin Moor and Carnmenellis granites contain a lower proportion of xenoliths (if they are the source). It may be due to the different structural levels seen in the plutons, with the deeper levels reflecting zones of greater xenolith assimilation away from the carapace.
Other elements such as Sn, W, F, Rb, Cs and Li have moved out from the granite, into the killas, by volatile streaming and metasomatism (Ball et al., 1998); halogens and mobile elements such as Rb are able to substitute for Al, etc, within the lattices of phyllosilicates within the slates. Beer and Ball (1986) found elevated levels of Sn and W up to 0.8 km beyond the granite contacts of a number of the plutons, indicating a primary phase of dispersion, prior to the onset of the first stage of lode mineralisation. The area immediately adjacent to the pluton (20-50 m) often carries medium values and this is succeeded by a belt of high metal values to a distance of 500-800 m. Background values for the pelites of Cornwall are 3.3 ppm for Sn and 4.6 ppm for W. Background values for the granites are 32 ppm for Sn and 19 ppm for W. Sn levels in the granites vary from around 30 ppm to 70 ppm (Carnmenellis), reaching up to 1000 ppm (St Agnes) and values for W may reach 25 - 100 ppm. Elevated values in the contact aureole may reach 120 ppm (Sn) and 30 ppm (W), with W typically lower than Sn and declining in concentration more rapidly. At the contact of the Land's End Granite favourable geological conditions lead to the development of a Sn-rich skarn deposit (Van Marcke de Lummen, 1985).
Stone (1982) noted elevated concentrations of alkali metals and base metals within the contact aureoles of the granites; he also discovered preferential concentration along horizontal contacts (upward vapour streaming) which were up to three times higher than those for vertical contacts. Values within the aureole are twice the background values for Rb (414 ppm), Li (324 ppm), Zn (241 ppm) and Pb (232 ppm), and three to eight times higher for F (0.20 ppm) and Cs (65 ppm) respectively.
Ball et al. (1998) noted alkali metasomatism within the contact aureoles of the granites, but also found dispersion haloes above concealed granite bodies. The dispersion patterns they found highlight the varying mobility of the various elements. Na concentrations were found above background, but lower than in the granite, with a low concentration close to the granite. K behaves in a similar way, with highest values some 1km from the contact. This high grade zone marks the metasomatic 'wave front' for these elements. Rb shows elevated values (500 ppm) up to 500 m from the contact and close to the contact these are twice those found in the granite itself. Cs behaves similarly; values close to the contact exceed the values in the granite by 10 times (up to 400 ppm). Li follows he same pattern (up to 600 ppm). These elements are typically hosted in secondary white micas in the surrounding country rocks. They reflect initial metasomatic dispersal of volatile elements that could not be incorporated within the granite. A similar study looking at REE (Ball et al., 2000) was inconclusive, but this was partially a reflection of the sample base (minor plutons, e.g. Cligga Head, only) and the fact that the country rocks have a higher REE content than the granites to begin with. Although no REE's with a 'granitic signature' were found it is likely that there is some transport of these elements from the granites into the aureole rocks. From the above we can see that this dispersion also involved base metals, including Sn. We can assume that the majority of the Sn and W was unavailable for this metasomatic phase as it was locked into a series of chloro-fluoro-complexes and therefore too large for molecular transport by diffusion (Horsnail, 1979; Taylor, 1979).
Back to top of Page.
The geochronology of the Cornubian Batholith has been constrained in a number of important works over the last 15 years. Using a variety of techniques, the timing of the emplacement of individual plutons and the onset of mineralisation has been deduced. This chronology shows that the emplacement of the batholith is diachronous, and spans a 20 million year period, beginning in the Lower Permian, during the period of orogenic extension (D3; Alexander, 1997). The data by Darbyshire and Shepherd (1985; 1987); Clark et al. (1993); Chen et al. (1993) and Chesley et al. (1993) is summarised in Figure 8.
The earliest of the granites is the coarse-grained megacrystic granite of the Carnmenellis pluton, which was intruded at around 293 Ma, followed by the main Bodmin Moor granite at 291.4 Ma (Chesley et al., 1993). These two granites have several geochemical characteristics in common that set them apart slightly from the reset of the pluton (Stone, 1987), perhaps reflecting a slightly different source rock or greater degree of partial melting. The Carnmenellis granite was already undergoing significant differentiation and volatile concentration on its margins by 289 Ma, which is the date for the onset of W-As mineralisation in the Roskear Complex of South Crofty Mine (Chen et al., 1993); this is roughly contemporaneous with the intrusion of the outer granite (288.4 ± 1.2 - 290 ± 1.1 Ma). This was followed at 288.5 ±1.5 Ma by the emplacement of the inner granite (Gd of Ghosh, 1934) in the centre of the outcrop (Chen et al., 1996).
Figure 8. A compilation of geochronological data (with 2s error bars) for the Cornubian batholith. (after Chesley et al., 1993).
The emplacement of part of the composite Scilly Isles Granite (Mullis et al., 2001) took place around 287 Ma; followed by Carn Marth at 284 Ma, Kit Hill and Hingston Down at 284 and 283 Ma respectively and St Michael's Mount and the De Lank granite on Bodmin moor at 281 Ma (Chen et al., 1993).
By 280 Ma the emplacement of the Dartmoor Granite (Edmonds et al., 1968; Stone, 1995) and the St Austell biotite granite had begun (by this time according to the data of Clark et al. (1993) the development of main stage Sn lode mineralisation in the Carnmenellis pluton was nearing its end, with the 'first pass' of mineralisation already in place in several of the lodes at South Crofty mine). The biotite granites were then intruded by a later set of tourmaline-bearing dykes and small bodies of granite (accounting for some 5% of the St Austell pluton, where their distribution has a close spatial connection with later sites of kaolinisation; Manning et al., 1996) and these were succeeded by topaz granites between 280-274 Ma. To the west, at this time (280 Ma), the Tregonning (topaz) Granite was also being emplaced. In the Carnmenellis Granite, by 283 Ma, chlorite-dominated Sn-Cu mineralisation was being developed and was probably fully established at South Crofty Mine. Greisen related Sn-W mineralisation was initiated at St Michael's Mount and Cligga Head at 281.8 ± 1.4 Ma and 279.9 ± 0.8 Ma respectively (Clark et al., 1993). Isotopic data (Clark et al., 1993) shows that the oxide-dominated Sn-Fe lode mineralisation of Central Dartmoor was emplaced at 278 ± 0.6 Ma, giving a similar 2-3 million year hiatus to that seen in Carnmenellis.
Data from Evans (1990) for the submarine Haig Fras Granite gives an age (K-Ar) of 277 Ma, and for the granodiorite outcropping at Granite Cliff on the continental slope (approximately 400 km WSW of Land's End) a K-Ar date of 275 Ma (whole rock) and a biotite age of 290 Ma; assigning both to the Lower Permian intrusive episode.
At this time the results of significant differentiation and elemental partitioning is leading to the formation of highly specialised tourmaline and topaz-bearing granite bodies (particularly in the St Austell Granite) that have intrusive contacts with the pre-existing batholith rocks. These show a variety of contact relationships from gradational to sharp to the development of stockschieder pegmatites. These granites are associated with extensive tourmaline veining and pervasive tourmalinisation, and later with the development of breccia pipes and veins by hydraulic fracturing of the host granites (Manning and Exley, 1984).
The earliest granites of the Land's End pluton, those of the Castle-an-Dinas Granite and northern, Zennor Lobe (coarse grained megacrystic granite; Salmon and Powell, 1998; Powell and Salmon, 1999a, 1999b), were emplaced, in rapid succession, around 275 Ma (Chesley et al., 1993). This was followed within 1-2 million years by the emplacement of the southern, St Buryan Lobe (Chen et al., 1993) and then (occupying the central area of the pluton) the poorly-megacrystic coarse-grained granite of the St Just Wedge (Powell et al., 1998, 1999b). The final intrusives are the variably sized, irregular masses and sheets of fine-grained biotite-muscovite granite that occur in all of the three major granites.
The latest granites were intruded in the Carnmenellis pluton at Boswyn at 276.4 ± 1.1 Ma and Sn mineralisation began in the St Austell pluton around 271.4 ± 4.5 Ma. Greisen related Sn-W mineralisation was initiated at Bostraze and Balleswidden in the Zennor Lobe of the land's End Granite at 271.2 ± 0.8 Ma and the main Sn-Cu phase of mineralisation in the St Just area was initiated between 265 and 262 Ma (Halliday, 1980; Clark et al., 1993).
Dartmoor was unroofed in the Lower Permian (Edmonds et al., 1975) and granite clasts can be seen in red bed sediments in the Crediton Trough. Other granites may have been unroofed in the early Triassic as eastwards draining rivers deposited granite clasts in the conglomerates of Budleigh Salterton in Devon.
The isotopic evidence for the various plutons gives information on their cooling rates. Using U/Pb data for monazite (trapping temperature 750°C) and xenotime (trapping temperature 650°C) and Ar/Ar data for muscovite (trapping temperature 325°C) cooling rates can be calculated. Cooling times for individual plutons, from emplacement to closure in muscovite crystals, vary from 1.9 to 6.7 Ma (mean is 3.9 Ma). Initial cooling would be rapid, followed by a long period of protracted cooling towards an ambient value (Chesley et al., 1993). This data shows, that although there is variation between the ages and cooling rates of the various plutons, there is a decrease in the cooling rates from Land's End (210°C/myr-1) in the West to Dartmoor (60°C/myr-1) in the east. This goes directly against the assumption of Willis-Richards and Jackson (1989) that the Dartmoor granite cooled rapidly. Chen et al. (1993) did not observe this variation in their data and assert that all the major plutons cooled at between 80-100°C/myr-1 and that the present surface of the batholith had cooled below 320°C by 270 Ma.
Further work by Chen et al. (1996) using apatite fission track data (closure temperature 125°C) on rocks from Carnmenellis and the Rosemanowes borehole shows variations in cooling rates within this single pluton, relative to depth within the granite. Rocks at the surface cooled to <125°C at 155 ± 18 Ma (cooling rate of 2.8°C/myr-1), while those at 2481 m depth cooled to the closure point at 58 ± 7 ma (cooling rate of 0.45°C/myr-1) indicating the length of time taken to reach this point and the fact that the interior of the pluton remained at over 100°C for nearly 240 million years after emplacement. This is a function of the radiogenic heat flow produced by the pluton. Argon dates for muscovites from the borehole give dates of 289 ± 0.7 Ma for rocks at surface and 286 ± 1 Ma for a greisen at 2481 m. An age of 284.4 ± 0.9 Ma for a fine-grained granite at 2299 m depth indicates the presence of a younger intrusive phase at depth.
Chen et al. (1996) also discovered that, on the basis of U-Pb, Ar-Ar and AFT data, the Carnmenellis pluton had, after crystallisation, undergone protracted cooling (due to the effects of radiogenic heat production) at approximately 1.5°C/myr-1 up to 155 Ma. At this point the cooling rate accelerated to 2.8°C/myr-1 and continued at this rate to 137 Ma. This late Jurassic - early Cretaceous date matches an uplift event recorded in the sedimentary basins of the English Channel and Bristol Channel (Evans, 1990). At this time the Cornubian Ridge was uplifted (initiating the 'fast cooling event') and the basins immediately to the north and south also experienced strong uplift. It is likely that at this time all the granites would have been exposed and that the region would have been significantly denuded. Although it may have remained as land throughout the Jurassic it seems likely that Cornubia would have been submerged by the Cenomanian Chalk Sea. Later mild uplift and tilting in the Tertiary was accompanied by reactivation of older structures and substantial wrench faulting (Dearman, 1963) and block faulting throughout the massif, in response to Alpine movements, with the development of onshore basins along the Sticklepath Fault (Edmonds et al., 1975; Selwood et al., 1984; Smithurst, 1990) in Devon. This lead to the breaking up and translation of the pluton into a series of blocks; later peneplanation during the Late Tertiary lead to the current outcrop patterns being established with the differing structural levels observed between the plutons.
Back to top of Page.
7. The elvan dykes.
The elvans are a suite of quartz porphyry dykes (sometimes described as rhyolite porphyry) that occur in close association with the granite batholith (Floyd et al., 1993) and its associated mineralisation. The elvans occupy the same (high-angle extensional faults) fracture set as the main-stage lodes, often trending parallel with them in each major mineralised district (E-W around St Austell, NNE-WSW around Camborne and Redruth), and show a variety of cross-cutting relationships with them (Hosking, 1988; Rayment et al., 1971; Kettaneh and Badham, 1978) that display a contemporaneous or overlapping origin. They cut coarse-grained granites throughout southwest England, but not fine-grained granites (Selwood et al., 1998) and have been dated at between 282 ± 6 Ma and 269 ± 8 Ma (Hawkes et al., 1975; Darbyshire and Shepherd, 1985).
The Model developed in Thorne and Edwards (1985), Bromley and Holl (1986), Bromley (1989) and Jackson et al. (1989) sought to explain the processes of elvan magmatism and mineralisation in terms of radiogenic reheating and protracted cooling of a (assumed) single original magma pulse. In this model radiogenic reheating in the interior of the pluton generates a 'second magmatic event' in which differentiated specialised granites, elvan dykes and mineral lodes are emplaced. Thermal modelling of the Carnmenellis Granite by Willis-Richards and Jackson (1989) showed that, using known heat flow and conductivity values, and by restoring a 4 km thick sequence of cover rocks, that a significant volume of the granite (10%) could remain above the solidus for in excess of 20 million years. This time period would be long enough for significant differentiation to occur and to generate successive intrusive events (as is seen in Carnmenellis) and to allow full development of later elvan magmatism and mineralisation. Where the 'thermal blanket' of cover rocks is thinner or removed quicker, there is likely to be less development of these features as complete crystallisation takes place much more rapidly. Willis-Richards and Jackson (1989) suggests this as a reason why there is a relative scarcity of elvan dykes and mineralisation over Northern Dartmoor, which was unroofed during the Lower Permian.
This model no longer appears valid in the light of new chronological data on the punctuated nature of emplacement throughout the batholith (and Carnmenellis itself) and in terms of the work by Cruden (1998) and others, on the composite (by necessity) nature of larger intrusions; it also does not remain valid if the R1 reflector is identified as the base of the batholith. The origins of the elvans now appears far more complex; they are still poorly understood, but the resolution of granite chronology, and to some extent that of mineralisation, shows that elvan emplacement is also diachronous and, like the mineralisation events, tied to the development of individual plutons.
Elvans vary from grey to light green or pink in colour and can contain concentrations of phenocrysts that may reach up to 25% (usually euhedral to subhedral microperthitic and ophitic orthoclase up to 30 mm long, glomerocrysts of quartz up to 5 mm across and/or zoned subhedral plagioclase up to 4 mm long), set in a groundmass that may vary in grain size from cryptocrystalline to ~0.5 mm (usually 0.1 - 0.2 mm) and display textures from granophyric to micrographic (Selwood et al., 1998; Manning, 1998). Orthoclase and quartz are the most abundant minerals. Elvans are often subject to secondary alteration; tourmalinisation is fairly common, as is chloritisation; but kaolinisation is the most widespread and may often be so locally pervasive that many elvans formerly quarried for building stone also had working claypits along strike in close proximity.
Elvans are chemically and texturally heterogeneous as a group, reflecting a complex origin, and more potassic than their associated granites (Hawkes et al., 1975; Stone, 1968), a factor that may reflect more specialised parent magmas or in-situ ion exchange during emplacement (Stone, 1968).
Elvan dykes vary from a few cm in thickness to >40 m (Selwood et al., 1998) and may have strike lengths that reach several km. Their margins may show a variety of textures from chilling to flow banding. Several dykes show evidence of fluidised gas entrainment of clasts along the margins (Stone, 1968; Rayment et al., 1971; Kettaneh and Badham, 1978) indicating the presence of a gas/vapour phase which may have aided emplacement (by gas-coring); some dykes are also associated with hydrothermal breccias (Manning, 1998).
Just as some of the fractures utilised by the main-stage lodes communicated with lower pressure zones or the surface (Halls, 1994), it is likely that similar conditions applied to the elvan dykes which would have resulted in surface rhyolitic volcanism. Today only ~1 km2 of rhyolite lavas is preserved at Kingsand [SX435506] and Withnoe [SX404517] in southeast Cornwall (Manning, 1998). These lavas are reddened by Permo-Triassic oxidative desert weathering and, as evidence from the quantity of rhyolites pebbles in Permian sediments suggests, formed part of a formerly much more extensive suprabatholithic volcanic region that may have extended to several parts of Cornubia.
Back to top of Page.
Forward to Next Page
Back to Introduction
This page last updated 28/02/2013