Introduction

Mafic microgranular enclaves (MME) are widespread in intermediate and felsic granitoids. The characterisation, genesis and evolution of MMEs have been widely discussed by several authors (e.g., Didier 1973; Hibbard 1981; Vernon 1983; Didier and Barbarin 1991). The three main theories for the petrogenesis of microgranular enclaves are as follows: (1) they form by settling of early crystals from the host magma or by fragmentation of early solidified wall-rock facies closely related to the host magma cognate fragments of cumulate minerals from a mafic magma (e.g., Noyes et al. 1983; Chappell et al. 1987; Dahlquist 2002; Ilbeyli and Pearce 2005); (2) they are fragments of recrystallised, refractory metamorphic rocks and fragments of melt residues from the granite source (e.g., Chappell et al. 1987; Chappel and White 1992; White et al. 1999); or (3) they are globules of more mafic magma, generally hybrid magma, comingled with a more felsic host magma (e.g., Vernon 1984; Dorais et al. 1990; Blundy and Sparks 1992; Wiebe et al. 1997; Kadioglu and Gulec 1999; Perugini et al. 2003; Barbarin 2005; Kocak 2006; Feeley et al. 2008; Kaygusuz and Aydıncakır 2009). The hybridism model, in particular has been well documented by many geologists (e.g., Frost and Mahood 1987; Didier and Barbarin 1991; Michael 1991; Blundy and Sparks 1992; Silva et al. 2000; Barbarin 2005) and has been applied in studies of granites worldwide.

Previous studies of the Horoz Granodiorite have reported that magma mingling had occurred (e.g., Çevikbaş et al. 1995; Kocak 2000; Kadioglu and Dilek 2009), but the mineral chemistry of the enclaves has not been investigated. The aim of this study is to present the mineral chemistry of the Horoz granitoid and to investigate the chemical and physical interaction between felsic and mafic magmas.

Geological setting

Turkey is situated on an E–W-trending fragment of the Alpine–Himalayan orogenic belt. This fragment contains a few continental blocks divided by suture zones developed by the closure of the different branches of the Neo-Tethyan Ocean during the late Cretaceous–Eocene (Sengor and Yilmaz 1981) (Fig. 1). Two subduction zones with opposite vergences are proposed to have created a double granitic arc in the north (Central Anatolian Crystalline Complex, CACC, Kirsehir Block of Gorur et al. 1984) and in the south (Taurides).

Fig. 1
figure 1

Generalised geological sketch map of the main lithologic units of the Central Anatolian Crystalline Complex (CACC, after Bingol 1974)

Full size image

The granitic intrusions along the northern edge of the Taurides may have been developed in relation to the evolution of the Neo-Tethyan “Inner Tauride Ocean” (Gorur et al. 1984; Dilek et al. 1999), which began to form as early as the Jurassic between CACC to the east and the Taurides to the west. The ocean was consumed by an intra-oceanic subduction northwards (Gorur et al. 1998) along the Inner-Tauride Suture Zone (ITSZ), which is defined by discontinuous exposures of the Cenomanian–Turonian suprasubduction zone ophiolites (i.e., Alihoca, Aladag) and melanges that occurred by the latest Cretaceous (Clark and Robertson 2002). The measured ages of the different parts of the ophiolites along the Tauride belt show that the subduction-related magmatic activity was initiated in the Turonian and continued until the end of Maastrichtian prior to their emplacement onto the Tauride platform (Parlak et al. 2010).

The study area is situated at the eastern part of the Bolkar Mountains (Nigde, Central Southern Turkey, Fig. 1), adjacent to ITSZ, between two continental blocks, the Taurides and the CACC. The Bolkar Mountain units, the oldest in the area, make up the heights of the Bolkar Mountains due to an anticline they form. The units include variably metamorphosed Upper Permian-Upper Triassic platform carbonates with siliciclastic intercalations, which were intruded by an ENE–WSW-trending, sill-like pluton (the Horoz granitoid, Fig. 2), along with the development of some economic skarn mineralisation. The Horoz granitoid is of shallow-seated intrusions and is cropped out in the so-called Horoz Valley. The pluton was intruded by later tonalitic/dioritic dikes (Fig. 3a) with a sharp contact that has a thickness ranging from 10 cm to 10 m and may continue locally for 100–250 m along-strike (Çalapkulu 1980). Aplite veins may cut the enclave mingled with plastic granitoid magma. The granitoid was overlaid by Upper Cretaceous Alihoca ophiolites, which are composed of the following: (a) medium-grained basic and ultrabasic rocks; (b) volcanosedimentary rocks, observed as klippe at the high hill of the Bolkar Mountains; (c) diabases, spilite and volcanic rocks; and d) albite-epidote-actinolite-glaucophane bearing schists. The ophiolite is dismembered as thrust sheets and blocks within an extensive ophiolitic mélange and is overlain by a U. Maestrichtian-L. Paleocene Kalkankaya formation, including conglomerates, limestone with rudists and argillecous limestone. The granitoid was covered with nonconformity by Oligocene terrace deposits at the southern part of the valley. Accordingly, the granitoid was suggested to be emplaced between the Lower Paleocene and Lower Eocene (Çalapkulu 1980). The age of the Horoz pluton was recently determined to be 56.1 Ma (earliest Eocene, Ypresian) using U–Pb zircon dating of its granitic end member (Y. Dilek, unpublished data; Kadioglu and Dilek 2009).

Fig. 2
figure 2

Geological map of the study area (modified from Çevikbaş et al. 1995, and Kadioglu and Dilek 2009)

Full size image
Fig. 3
figure 3

Field photographs displaying a mafic dyke (dk); b MMEs displaying sharp contact with their host rock; c back veining in MME; d double enclave (e) in the dioritic dike

Full size image

The granitoid is typically light-coloured and medium-fine grained with a slightly developed chilled margin at the contacts of the country rocks. This granitoid contains MMEs, which is suggestive of mingling in the felsic magma. Rare pelitic xenoliths with a diameter mostly less than 5 cm were also observed at the outermost margins of the pluton. Based on its mineralogical composition, the pluton can be classified into two distinct groups: granodiorite and granite (sensu stricto) with gradational contact. In comparison with the granodiorite, the felsic granite is relatively finer grained, more fractured and altered, and has less enclaves. The weathering can locally become intensive so as to produce sandy regoliths or grus.

The MMEs are rounded and tabular in shape, and their contacts with the host rocks are mostly sharp. In general, MMEs have dramatically different sizes (from several cm up to meters) and shapes (ellipse/round-cornered, Fig. 3b). Smaller MMEs tend to be finer grained and more isotropic than larger ones. The enclaves are less altered and are kept in better shape compared with the host granitoid. The MMEs are distributed randomly throughout the pluton. In no outcrops do the enclaves appear to be in any preferred alignment. The MMEs usually exhibit a fine-grained margin against the host rock, with sharp to crenulate contact. The enclaves sometimes show a core-tail structure in which the tails exhibit an S-shaped bend, implying that they were at least partly plastic when introduced into the felsic magma. Small blisters (1–10 cm in diameter) of the host rocks are incorporated into the MMEs (Fig. 3c). The feldspars are situated in the MMEs and near the host-enclave boundary. The MMEs in the dioritic dikes are texturally highly heterogeneous and may contain a “double enclave” structure in which they partially or fully enclose smaller, finer-grained, and more mafic enclaves (Fig. 3d).

In this study, we employ the terminology of Sparks and Marshall (1986), where ‘magma mixing’ produces homogeneous hybrid rocks and ‘mingling’ leads to inhomogeneous hybrid rocks.

Petrography and mineral chemistry

Methods

Sixty petrographic thin sections were studied under microscope to determine their composition and texture. Polished sections (25 · 46 mm) of representative rock samples were made at the thin-section Laboratory of the Geological Engineering Department, Selcuk University (Konya). Polished slides were coated with carbon and then analysed at the Electron Microprobe Laboratory of McGill University, Montreal, Quebec, Canada. The mineral analyses were obtained using a JEOL JSM35 Electron Microprobe running Link QX2000 energy dispersive analytical software. The electron beam condition was 15 keV and 15 nA. The Fe3+ content in the amphiboles was determined according to Droop (1987).

General petrography

The granodiorite is composed of plagioclase (55–70 vol%), quartz (15–25%), biotite (5–15%), orthoclase (10–15%), amphibole (0–10%), minor chlorite, and apatite along with accessory apatite, tourmaline and zircon. The granite consists of zoned plagioclase (35–55 vol%), orthoclase (30–50%), quartz (15–30%), biotite (10–15%), amphibole (0–5%), and accessory apatite and zircon. In both granitoids, the hypidiomorphic granular texture is dominant, exhibiting various levels of sericitisation, chloritisation and kaolinitisation. The host rocks are sometimes highly fractured and altered, particularly in the granite, and the crystals are flattened by stretching or shearing, which produced the post-crystallisation features. Accordingly, the quartz crystals display strong undulose extinction and usually form numerous subgrains. The plagioclase may be slightly epidotised, particularly in its core. The quartz forms anhedral to the subhedral grains clustered between feldspar crystals, and it locally occurs as micrographic, granophyric and worm-like myrmekitic intergrowths. The biotite is usually subhedral, pleochroic from β-dark brown to α-pale yellow and contains inclusions of zircon. Amphibole is rare and is found only next to the contacts between the enclave and host rocks.

The MMEs have an almost identical mineral assemblage as their host. The main constituents are plagioclase (75–85%), amphibole (5–15%), biotite (5–10%), and orthoclase (0–5%), and the minor constituents are quartz, chlorite, titanite, clinozoisite, opaque iron ore and apatite (Fig. 4a–f). The MMEs are generally equigranular, fine-grained and sometimes porphyritic. Albite and albite-carlsbad twinnings are characteristic in the plagioclase, which may have sieve and boxy cellular textures (Fig. 4a–b). A poikilitic texture is characteristic in some samples that contain large euhedral-subhedral plagioclase and/or alkali feldspar (Fig. 4c) phenocrysts up to 2.0 mm in size, with the inclusion of subhedral biotite and amphibole. The K feldspar megacrysts are identical to those in the host, and they exhibit a perthitic texture and resorbtion. The quartz is interstitial and sometimes occurs as ocelli quartz, which is developed in the samples where the quartz crystals were engulfed by subhedral biotite, amphibole and/or feldspars (Fig. 4d). Quartz ovoids also exist in the samples. The amphibole frequently demonstrates conversion to brown biotite and chlorite (Fig. 4e) and is strongly altered into the chlorites along its cleavage, particularly in the granites. The titanite in the MMEs is predominantly anhedral to subhedral with an interstitial habit, but it sometimes forms typical wedge-shaped large crystal, up to 2.7 mm in maximum length, with plagioclase inclusion. Small subhedral titanite crystals are often associated with chlorite and opaque iron ore. The apatite has a typical acicular shape (Fig. 4f), suggesting quenching of the mafic melt globules trapped in silicic magma (Wyllie et al. 1962), but stubby prismatic shapes also exist.

Fig. 4
figure 4

Thin sections photographs of the MMEs (a-f) showing disequilibrium crystallization features and textures. a Plagioclase displaying sieve texture; b Boxy cellular plagioclase; c Orthoclase phenocryst with poikilitic texture and d Ocelli quartz; all photos under crossed polars. (e, f) Resorbed hornblende and acicular apatite (f) with quenched morphologies respectively (plane-polarized light). Quartz: qtz; orthoclase: or; plagioclase: pl; hornblende: hb; biotite: bt; ap apatite; sv sieve texture

Full size image

The plagioclase in all samples is usually complexly zoned with a variety of compositional patterns; it shows a patchy zoned or resorbed slightly zoned core mantled by oscillatory zones, mostly planar and euhedral or gently curved (Fig. 5a–d). Electron microprobe traverses of the plagioclase show that many grains exhibit a distinct increase in An content (an An “spike”, e.g., Wiebe 1968; Hibbard 1981), which interrupts and/or transects the normally zoned plagioclase. Two types of chemical zoning are determined in the zoning profiles: (a) in the first, the An content appears to increase from ~29 to ~41 and 38 to 49 between the core and rim respectively. (Fig. 5a–c), and there is a pronounced shift of ±<5 and ~15% of the An content in the outer part of the mantle at the granodiorite and enclave plagioclase, respectively (Fig. 5a, b and c); (b) in the second, the An content decreases regularly towards the rim (Fig. 5d). The oscillations within the rim are typically on the order of +25 and +22 mol% An in the MMEs and granodiorites, respectively. A sieve-textured zone that is 0.1–0.3 mm wide usually forms mantles on planar, locally fritted surfaces bounding oscillatory euhedral zones, sometimes with curvilinear boundaries.

Fig. 5
figure 5

The crossed polars image of the (a, b) Oscillatory zoning in plagioclase from granodiorites and enclave (c). d backscattered electron image of enclave plagioclase with oscillatory zoning

Full size image

Plagioclase

In the granitoids, the plagioclase occurs both as large euhedral crystals (up to 2.8*1.1 mm) and as small subhedral crystals (0.8*0.3 mm), usually with resorption features. The plagioclase composition of the granodiorites varies from An21 to An55 in the cores to An07–An33 in the rims (Figs. 5 and 6, Table 1). The plagioclase composition of the granite is distinct from that of granodiorites, which varies from An17 to An38 in the core and from An18 to An24 in the rim (Table 1). The plagioclase crystal may exhibit albite twinning and/or normal zoning.

Fig. 6
figure 6

Nomenclature of the feldspars from granodiorite (a), leucogranite (b) and MMEs (c)

Full size image
Table 1 Representative analyses of plagioclase from host rocks
Full size table

The MME plagioclase crystals (An17–64, Fig. 5) are relatively smaller than those in their hosts, though some large crystals, up to 1.8 mm in size, form a poikilitic texture with numerous subhedral magnesiohornblende and biotite inclusions. Their composition ranges from An18–An64 in the cores to An17–An29 in the rims (Figs. 5 and 6, Table 2). The compositional similarity between the plagioclase phenocrysts from the MMEs and their host rocks suggests that the MMEs were predominantly equilibrated with their hosts. The plagioclase inclusions in the K-feldspar megacrysts have a composition ranging from An55 to An27 in the cores and An22 to An19 in the rims.

Table 2 Representative analyses of plagioclase from MMEs
Full size table

Amphibole

The amphibole is usually subhedral and ranges in shape from blade to blocky in the granitoid and enclave samples. The enclave amphibole occurs in two grain sizes: (a) coarse-grained (2 × 0.7 mm) and (b) fine-grained (0.3 × 0.12 mm). In the IMA-approved nomenclature (Leake et al. 1997), both amphibole groups are classified as magnesio-hornblende and edenite (Fig. 7a, b), which have high alkali (Na + K) contents and plot closer to the pargasitic substitution line (Fig. 7c) with an aluminium number (i.e., Al# = [6]Al/AlT) ≤ 0.21. The enclaves and their host also include unusual amphiboles, occasionally with Si > 7.3, which is possibly not actually magmatic but rather the product of subsolidus reactions in the presence of an exsolved fluid (Chivas 1981; Hendry et al. 1985). The secondary amphiboles occur both as subhedral to anhedral homogeneous crystals or as patchy domains in the magnesiohornblende crystals (Fig. 7d), which show intensive resorption and some alteration to biotite. They are all magnesiohornblende in composition and rarely contain plagioclase inclusions. These Mg-rich amphiboles were possibly formed under conditions of high oxygen fugacity and were preferentially depleted in elements that partition into “late-magmatic” hydrothermal solutions.

Fig. 7
figure 7

a, b Nomenclatures of the amphiboles (Leake et al. 1997). c Na + K vs. Aliv. Edenitic (Ed), pargasitic (Prg), and tschermakitic (Ts) substitutions are indicated along with an idealised hornblende (Hbl) substitution (Hietanen 1974). d Patchy domains in the hornblende refer to secondary hornblende. e Inverse chemical zoning from enclaves amphiboles. Filled and open diamonds represent rim and core of the enclave hornblendes, respectively. X: granodiorite ◊: enclave

Full size image

In comparison with the amphibole of the granodiorites, the MME amphiboles have generally higher Al, Ti, Fe2+, Ca, Na and K and lower Si and Fe3+, Mg contents (Tables 3 and 4), though they nearly overlap in the classification diagram of Leake et al. (1997) (Fig. 7a), which may suggest that the amphibole has not been equilibrated. The amphibole phenocrysts from both groups demonstrate reverse zoning in terms of decreasing Mg/(Mg + Fe2+), indicating the disequilibrium of crystallisation with magma (Fig. 7e). [Fe3+/(Fe3+ + Mg)] = 0.32–0.52 for the Hbl in the MMEs, and [Fe3+/(Fe3+ + Mg)] = 0.42–0.53 for the Hbl in the granodiorite. These ratios are much higher than 0.2 (Anderson and Smith 1995), suggesting crystallisation at high fO2 for the granitoids and MMEs.

Table 3 Representative analyses of the hornblende from granodiorite (Sample 40b)
Full size table
Table 4 Representative analyses of amphiboles from MMEs
Full size table

Biotite

Enclave biotite forms as phenocrysts (up to 1.02 × 0.7 mm) and as inclusions (0.2 × 0.13 mm) within the amphibole and plagioclase. This biotite varies in morphology from anhedral patches to subhedral blade-shaped crystals. The biotite is frequently bent and altered into green chlorite along its cleavage.

Representative analyses and calculated formulae of mica from the granodiorites and their MMEs are presented in Tables 5 and 6. All biotites show compositions with FeO = 18.24–20.26 wt.%, Al2O3 = 10.05–15.17 wt.%, and K2O = 8.88–9.83 wt.% and can be classified as Fe-biotites and eastonite based on the classification diagram of Tischendorf et al. (1997) (Fig. 8a–d). The biotites are rich in TiO2, with contents of 3.16–3.87 mass% in the enclaves and 3.29–3.92 mass% in the host rocks. The mole fractions of octahedrally coordinated Mg2+, Fe2+, and Fe3+ in the biotite of the MMEs and their host suggest that their nature is comparable to the primary biotite, which has been unaffected by hydrothermal alteration (Beane 1974). The FeOt/MgO ratio of biotites in the granodiorite and MMEs ranges between 1.30 and 1.55, with an average of 1.41, and between 1.27 and 1.46, with an average of 1.37, respectively. These values are close to those of the Mg–biotites (FeOt/MgO = 1.76) typically associated with calcic pyroxene and/or calcic amphibole commonly found in calc–alkaline (mostly orogenic and subduction-related) I-type granitoid suites (Abdel-Rahman 1994). The Fe2+/Fe2+ + Mg ratio and the iron-enrichment index of biotite in the host granodiorite are identical to the biotite in the MMEs, suggesting that the biotites could be entirely equilibrated with the MMEs. The SiO2 content of the samples also supports this suggestion, as it ranges from 36.26 to 37.08 mass% and 35.12 to 36.97 mass% in the enclaves and granodiorites, respectively.

Table 5 Representative analyses of biotite from the granodiorites
Full size table
Table 6 Representative analyses of the biotite from the MMEs. For explanations see Table 5
Full size table
Fig. 8
figure 8

Nomenclature and classification of the biotites from granodiorite (a, c) and their enclaves (b, d) (Foster 1960; Tischendorf et al. 1997)

Full size image

The MICA+ (Yavuz 2003) computer programs were utilised for the mineralogical calculations of the mica analyses. The micas analysed from the enclave and their hosts are all trioctahedral micas (2.74 ≤ M site <2.95 atoms per formula unit) and, for the most part, eastonite. Figure 8c and d show the octahedrally coordinated cations in the terms of Foster (1960). Biotites are found on the line along which the Mg:Fe ratio is 1:1.

Structural formulae calculated based on 22 positive charges show that the Si and Al cation p.f.u. usually fills the tetrahedral sites. In contrast, the sum of octahedral sites show somewhat more variability, particularly in enclaves, between 2.89 and 3.09 cations p.f.u. The occupation of 12-fold coordination sites ranges between 0.867 to 1.397 and 0.917 to 0.973 in the enclaves and their host, respectively. All these values suggest that all micas are near the ideal stoichiometric values.

The Fe3+ contents and Fetot/(Fetot + Mg) ratios give at least relative information about the oxygen fugacity during crystallisation (Wones and Eugster 1965; Dodge et al. 1969; Albuquerque 1973; Barriere and Cotten 1979; Neiva 1981). Both groups of biotites plot above the Ni-NiO (NNO) oxygen fugacity buffer (Fig. 9a, b) in the Fe2+-Fe3+-Mg diagram of Wones and Eugster (1965), indicating moderately oxidising conditions.

Fig. 9
figure 9

Relationship of the Fe3+-Fe2+-Mg content of the biotites from granodiorite (a) and their enclaves (b) compared with buffered biotites from Wones and Eugster (1965)

Full size image
Fig. 10
figure 10

Geotectonic setting of the biotites from granodiorite (a) and their enclaves (b) according to Abdel-Rahman (1994)

Full size image

Abdel-Rahman (1994) introduces discrimination diagrams on the basis of the major-element composition of biotite in igneous rocks crystallised from three distinct magma types. The biotites in granodiorites and enclaves in terms of MgO FeOt Al2O3 components show their crystallisation in a subduction-related calc-alkaline magma (Figs. 9a, b and 10).

Discussion

The mineralogical compositions of the MMEs are fairly comparable to those of the host granites. The ubiquitous finer grain size of the enclaves and the lack of cumulate texture are inconsistent with their being cumulates (Barbarin and Didier 1992). The enclaves do not include sillimanite, andalusite, cordierite, garnet or residual minerals produced from mica dehydration melting. Consequently, they do not represent restite. Correspondingly, a combination of characteristics such as an elliptical to rounded shape, contact morphologies (such as phenocrysts of host granites partly enclosed in MMEs and host granite magma), and igneous microtextures with acicular apatite strongly suggest that the MMEs may represent remnants of a mafic component added to intermediate or felsic magma chambers (Eichelberger 1980; Vernon 1984; Didier and Barbarin 1991; Bonin 2004; Barbarin 2005; Yang et al. 2006, 2007; Qin et al. 2009).

The magma mixing has been suggested from a number of lines of evidence:

  1. (1)

    The coexistence of contrasting magmas may be suggested by the chilling of the basic mass and disrupted late syn-plutonic dykes or an abundance of mafic microgranular enclaves (Didier 1973; Vernon 1984; Didier and Barbarin 1991; Wiebe 1996). In a slowly cooled and intrusive environment, the quenched texture could be in part obscured by recrystallisation effects (e.g., Wall et al. 1987). Alternatively, the breakdown of larger enclaves with chilled margins could create many smaller ones that mainly lack the chilled margins. Enclave magma quenching against the cooler felsic host is indicated by the shape of the enclaves, which are rounded to ovoid, subangular to angular and have lobes that are convex towards the host granitoids. Vernon et al. (1988) have stressed that rounded- to ellipsoidal-shaped enclaves are a strong sign of magma mingling and flow.

  2. (2)

    The preliminary quenching and multistage crystallisation of the assumed hybrids can be shown by the acicular apatite and blade-like biotite, together with quartz and K-feldspar megacrysts with many inclusions of biotite and amphibole (Vernon 1990, 1991; Hibbard 1991, 1995). Wyllie et al. (1962) experimentally showed that long prismatic-acicular crystals could result from strong undercooling during growth, and this is mostly taken as indirect evidence of magma mixing, showing the quenching of the basic melt by cooler acidic magma (Didier 1987; Hibbard 1991, 1995). The poikilitic texture in the enclaves is also possibly the result of the late stage crystallisation of felsic melt following the crystallisation of numerous quench-generated amphibole and biotite in relation to magma mixing (Hibbard 1991).

  3. (3)

    An exchange of crystals (xenocrysts) is displayed by plagioclases and amphibole with resorption surfaces, quartz ocelli rimmed by biotite and amphibole, quartz ovids, and a sieve texture. The quartz ocelli are developed by the partial dissolution of quartz xenocrysts in the mafic magma that extracts latent heat from the adjacent liquid. The undercooled surface of the quartz grain was transformed into a preferential nucleation substrate for minerals of the mafic magma (Vernon 1990). The sieve texture may be developed in at least two ways. If a plagioclase crystal is put into a magma in which it is not in equilibrium (by magma mixing, Wiebe 1968; Pringle et al. 1974; Tsuchiyama 1985; Nelson and Montana 1992), then it will become corroded, and melt will move into the crystal structure. New plagioclase of a distinct composition will crystallise from the magma and possibly form a rim around the corroded core. The sieve texture may also be formed by decompression when magma rises towards the surface (Stomer 1972). The exchange of crystals is also displayed by overgrowths and sharp zoning discontinuities (compositional spikes) in the plagioclase (e.g., King 1964; Wiebe 1968; Hibbard 1981; Barbarin 1990; Baxter and Feely 2002; Kocak 2006). Many plagioclase grains in the MMEs show evidence of resorption or dissolution and the subsequent re-growth of zones of differing An composition. These processes caused the formation of boxy cellular plagioclase (Figs. 4b), which may develop when plagioclase, originating from the mafic material, comes in contact with the cooler, felsic magma (Hibbard 1995). The plagioclase is then converted into more sodic compositions as equilibration occurs.

The homogenisation of the interacting melt phases results in the conversion of any pre-existing crystals to minerals stable in hybrid melt (enclave environment) or their rimming by stable minerals (e.g., Reid et al. 1983; Vernon 1983, 1984). Thus, the compositionally similar biotite and the similar sodic rims on the zoned plagioclase in the host rocks and their enclaves were developed. The high-level emplacement of the Horoz magma led to its saturation with H2O and the exsolution of fluids at an early stage of crystallisation. According to Blundy and Sparks (1992) and Waight et al. (2001), the diffusion of alkalis and water is much faster than that of other components, which promotes the crystallisation of biotite at the expense of amphibole. Finally, the dissolution of xenocrysts takes place in a new hybrid (enclave) magma environment, giving way to a reduction in grain size and the formation of ocellar mineral grains.

Some chemical transfer could occur in place in the felsic host before the enclaves solidify (e.g., Frost and Mahood 1987; Vernon 1990; Wiebe 1994, 1996; Wiebe and Collins 1998), but the local ratio of mafic to felsic magma needs to be high enough to preserve the more mafic magma, which is sufficiently fluid for mixing to take place. In general, MMEs are too small and insulated for this process to take place, apart from possibly in the very margins of some enclaves. Consequently, the mixing typically takes place outside the present host magma chamber, most probably in deeper magma bodies that are disrupted upon encountering the main magma body, thus forming the MMEs that were scattered in the host magma.

As a geobarometer, the Al contents of calcic amphiboles can be used in plutonic rocks (Hammarstrom and Zen 1986; Hollister et al. 1987; Johnson and Rutherford 1989; Thomas and Ernst 1990; Schmidt 1992; Anderson and Smith 1995). The total pressure, fH2O, and oxygen fugacity, in addition to the coexisting phases, influence the Al-content of hornblende. The tschermak substitution (Si + R2+ = AlIV + AlVI) is pressure sensitive (Hollister et al. 1987), while the edenite substitution (Si + vacA = AlIV + (K + Na)A and the reactions involving Ti (e.g., Ti + R2 = 2AlVI and Ti + AlIV = AlVI + Si) are temperature sensitive (Anderson and Smith 1995). The oxygen fugacity also plays a substantial role in the Altot content of the amphiboles. A high oxygen fugacity causes the insertion of Fe3+ into the lattice, which preferably substitutes Al, resulting in low Al-contents in the hornblende. Accordingly, it has been recommended that only amphiboles with a Fe3+/(Fe2+ + Fe3+)-ratio ≥0:25 be used for barometric analyses (Anderson and Smith 1995). The amphiboles from the MMEs and granodiorites have a sufficiently appropriate chemical composition and can be used as good pressure indicators in igneous rocks. The method of Al-in-hornblende barometry of Anderson and Smith (1995) was utilised to infer the pressure at which the magma crystallised as it responded to the effect of temperature. The amphiboles from the enclaves and granodiorites yield a P of 0.7–4.1 ± 0.6 kbar and 2.1–2.3 ± 0.6 kbar at 730°C, respectively. Thus, magma mixing between mafic and felsic magmas took place possibly at some depth below the present exposure level during ascent from at least ~8 km (2.3 kbar). The discrepancy in pressure data for the enclaves may either reflect problems inherent to the Al-in-hornblende barometer or suggest that amphibole-melt equilibration may have occurred at levels deeper than that of the emplacement. The chemical composition of the secondary magnesiohornblendes (Si > 7.3) were not used in the geobarometric calculations due to their low Altot contents, which induce too small (<0.5 kb) or too negative (e.g., −0.51 kb) pressure values.

Temperature data were calculated for the individual hornblende-plagioclase pairs of the different samples. A crystallisation temperature between 730 and 940 ± 40°C for a pressure of 5 kbar was obtained from the amphibole–plagioclase geothermometer (Holland and Blundy 1994) in the MMEs. The temperature range estimated from thermometry is in accordance with experimental data on the liquidus temperature of the hornblende in granodiorites (<950 8C; Gilbert et al. 1982). In the geobarometer calculation, a temperature of 730°C was preferred, as it is lower than the interpreted peak magmatic temperatures within the hornblende stability field and because it eliminates both scatter and negative pressure estimates.

For the Horoz granitoids, the biotite chemistry of the samples suggests a calcalkaline orogenic setting. However, Kadioglu and Dilek (2009) suggest that Horoz granitoids have adakitic geochemical characteristics developed in the convergent margin and collisional tectonic settings. The samples could have been crystallised under moderately oxidising conditions, as evidenced by high Fe2+/Fe3+ and [Fe3+/(Fe3+ + Mg)] ratios in the biotite and amphibole, respectively. The extensive occurrence of amphibole and biotite as the main hydrous phases in the enclaves and their host supports the idea that these rocks, particularly mafic magma, have crystallised from a relatively water-rich magma. The existence of amphibole in a rock shows that the melt must have had a water content above some threshold value (the threshold level is controversial). Estimates of the amount of water needed to saturate a magma with amphibole range from a minimum of 6 wt.% (Merzbacher and Eggler 1984) to approximately 5% (Eggler 1972; Helz 1973; Naney 1983), and then to 2–3% (Luhr 1992). Therefore, the water content of the Horoz magma is possibly greater than 3%.

Two types of igneous enclaves have been determined in the calc-alkaline plutons of the CACC (Kadioglu and Gulec 1999; and Ilbeyli and Pearce 2005): (1) fine-grained to rarely porphyritic (Type-I); and (2) medium-grained to porphyritic with feldspar megacrysts (Type-II). Most common are the Type-I enclaves with quartz dioritic to quartz monzodioritic compositions and fine-grained margins. Less common are the Type-II enclaves, which have quartz dioritic to gabbro compositions and cumulate to porphyritic textures. The texture and mineralogy of the Horoz MMEs usually shows a similarity to that of Type-1 with quartz dioritic to quartz monzodioritic compositions, indicating an interaction between silicic magmas and mantle-derived melts.

Conclusions

The combination of field, textural and mineral chemical evidence leads to the conclusion that magma mixing and mingling of mafic basic to intermediate and felsic (granodiorite and granite) magmas played an important role in the petrogenesis of the adakitic Horoz intrusion. These plutons have textures that show significant magma mixing and mingling processes. Poikilitic, sieve and boxy cellular textures and inverse and oscillatory-zoned plagioclase with compositional spikes probably record the mixing of coexisting mafic and felsic magmas. The existence of acicular apatite and blade-like biotite also supports this suggestion. The textural and mineralogical evidence from the MMEs and host rocks demonstrates that magma mixing between the mafic and felsic magmas with high f O2 took place at some depth below the present exposure level during its ascent from at least ~9 km and at more than 730°C. The identical mineral chemistry of biotite and plagioclase in the MMEs and their host suggest that these two minerals were possibly equilibrated. In comparison with the granodiorites, however, the higher content of Al, Ti, Fe2+, Ca, Na and K and the lower content Si, Fe3+ and Mg in MMEs suggest that the MME hornblende is possibly of xenocrysts origin and was transferred from the MMEs into the felsic melt.