JP WeiRong

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/318402769
Back-reaction of Peritectic Garnet as an Explanation for the Origin of Mafic

Enclaves in S-type Granite from the Jiuling Batholith in South China
Article in Journal of Petrology · May 2017

DOI: 10.1093/petrology/egx029
CITATIONS READS
0 226
3 authors:
Wei Rong Shao-Bing Zhang

University of Science and Technology of China University of Science and Technology of China
3 PUBLICATIONS 11 CITATIONS 32 PUBLICATIONS 3,108 CITATIONS
SEE PROFILE SEE PROFILE
Yong-Fei Zheng
University of Science and Technology of China
305 PUBLICATIONS 19,177 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Crust-mantle interactions at subduction zones View project
zircon and evolution of continental crust View project
All content following this page was uploaded by Wei Rong on 01 November 2017.
The user has requested enhancement of the downloaded file.

JOURNAL OF Journal of Petrology, 2017, Vol. 0, No. 0, 1–29
doi: 10.1093/petrology/egx029
PETROLOGY Original Article
Back-reaction of Peritectic Garnet as an

Explanation for the Origin of Mafic Enclaves in
S-type Granite from the Jiuling Batholith in
South China
Wei Rong, Shao-Bing Zhang* and Yong-Fei Zheng
CAS Key Laboratory of Crust–Mantle Materials and Environments, School of Earth and Space Sciences, University
of Science and Technology of China, Hefei 230026, China
*Corresponding author. E-mail: [email protected]
Received August 3, 2016; Accepted May 1, 2017
ABSTRACT
The origin of mafic enclaves in granites can provide significant information on granite petrogen-
esis. A combined petrology and geochemistry study was carried out on biotite-rich enclaves and
their host S-type granites from the Jiuling batholith in South China. The results demonstrate that
the biotite-rich enclaves were formed by back-reaction of peritectic garnet with the granitic melt at
a late stage of magma crystallization. The enclaves are fine-grained and contain more biotite and
less K-feldspar than the host granites. Garnet debris occurs in some biotite aggregates.
Petrographic observations show the occurrence of biotite rims on garnet fragments and the em-
bayment of biotite and quartz into garnet, suggesting that some biotites are formed by consump-
tion of garnet. These metasomatic biotites are characterized by higher Mg# [¼ Mg/(Mg þ Fe),
molar], lower K2O and TiO2 contents, lower (La/Yb)N and (Gd/Yb)N ratios and distinct negative Eu
anomalies. Some biotites in the enclaves exhibit similar geochemical characteristics to these meta-
somatic biotites. The biotites in the enclaves contain no K-feldspar inclusions and more plagioclase
inclusions than those in the host granites. Furthermore, these enclaves have indistinguishable
whole-rock Sr–Nd–Hf–O isotope and consistent zircon U–Pb ages and similar zircon Hf–O isotope
compositions to the host granites. Hence, we propose that the biotite-rich enclaves were formed by
back-reaction of peritectic garnet aggregates with the host granitic melt at an increased water fuga-
city in the granitic magma. Thermodynamic modeling indicates that the quantity of directly crystal-
lized biotite is low (<5%). Therefore, it is inferred that most of the biotites in the S-type granites
were formed by back-reaction of peritectic garnet entrained from the source.
Key words: mafic enclave; peritectic garnet; back-reaction; S-type granite; biotite; the Jiuling
batholith
INTRODUCTION Pascual et al., 2008; Farner et al., 2014). Mafic enclaves

Mafic enclaves are common in S-type granites, espe- in peraluminous granites are usually fine-grained and
cially in relatively more mafic ones (e.g. Chappell & more mafic than the host granite, although they usually
Wyborn, 2012). They provide valuable petrogenetic in- have similar mineral assemblages (biotite, plagioclase
formation about the nature of the source of the granites and quartz, with or without K-feldspar; minor garnet,
and magmatic processes, which are not readily appar- orthopyroxene and cordierite). Mafic enclaves usually
ent from the granites themselves (e.g. Didier, 1973, have identical emplacement ages to the host granites.
1987; Barbarin & Didier, 1991; Chappell & White, 1991; However, they show a large variation not only in min-
Maas et al., 1997; Barbarin, 2005; Donaire et al., 2005; eral assemblage and modal proportions but also in
C The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: [email protected]
V 1
2 Journal of Petrology, 2017, Vol. 0, No. 0
geochemistry. This leads to different petrogenetic inter- (e.g. Zhao & Cawood, 2012; Wang et al., 2013; Zhang &
pretations of field-based observations. In summary, Zheng, 2013). The metasedimentary sequences are
there are three schools of viewpoints on their structurally divided into a folded basement and an
petrogenesis. overlying cover with an unconformity between them.
The first school proposes that mafic enclaves repre- The magmatic rocks include a small number of mafic–
sent the unmelted remnants of the partially melted ultramafic plutons, tholeiitic basalts and voluminous
source rocks (e.g. Chappell et al., 1987; Chen et al., peraluminous granitoid plutons (e.g. Wang & Li, 2003;
1989; Chappell & White, 1991; White et al., 1999; Zhang & Zheng, 2013). The peraluminous granitoid plu-
Chappell & Wyborn, 2012; Wyborn, 2013). This residue tons are mainly Neoproterozoic intrusions such as
model gains support from the complementary relation- Bendong, Sanfang and Yuanbaoshan in the southwest-
ship in composition between the enclaves and their ern part and Jiuling, Xucun, Shexian and Xiuning plu-
host granites, the metamorphic texture of the enclaves tons in the northeastern part (Li et al., 2003; Wu et al.,
and the presence of residual minerals such as calcic 2006; Zheng et al., 2007; Zhao et al., 2013). Among
plagioclase, plagioclase cores with very high An con- them, the Jiuling batholith is the largest with an outcrop
tents, magnesium cordierite, and orthopyroxene area of about 2500 km2 (Fig. 1b).
enclosed in biotite (Chappell et al., 1987; White et al., The Jiuling batholith is geographically located in
1999). The second school advocates that the mafic en- northern Jiangxi Province. It intruded into the
claves represent batches of a mafic magma that have Shuangqiaoshan Group metasediments, which experi-
been mixed with a felsic magma (e.g. Reid et al., 1983; enced significant deformation and greenschist-facies
Elburg & Nicholls, 1995; Maas et al., 1997; Collins et al., metamorphism. The Shuangqiaoshan Group is com-
2000; Barbarin, 2005; Yang et al., 2007; Chen et al., posed of sandstone, siltstone, pelitic siltstone, slate and
2009; Vernon, 2014). The mixing model can explain the less amounts of volcanic rocks. The Jiuling batholith is
igneous texture (e.g. pseudo-doleritic texture), linear unconformably covered by the Neoproterozoic
bulk composition relationship and the different isotope Dengshan Group, which is mainly composed of sand-
compositions between the enclaves and the host gran- stone, slate, conglomerate, pelite and lesser carbonate,
ites. The third school suggests that the mafic enclaves spilite and volcanoclastic rocks (e.g. Wang et al., 2007).
may be cumulate crystals formed at an earlier stage in This study focuses on mafic enclaves and their host
the host granitic magma system (e.g. Tindle & Pearce, granites at Aozicun in the southeastern part of the
1983; Clemens & Wall, 1984; Dodge & Kistler, 1990; Jiuling batholith (Fig. 1c). Our samples were collected
Donaire et al., 2005; Collins et al., 2006; Huang et al., from a small quarry (28651 N, 115330 W) where there
2014; Chen et al., 2015; Moita et al., 2015). The main are abundant fresh enclaves hosted by medium to
lines of evidence for the cumulate model are similar coarse biotite granite. Aggregates of quartz (1–2 cm)
mineral compositions, identical formation ages, and and biotite (2 mm) can be observed in the field and in
similar isotope compositions. hand specimen. Most of the mafic enclaves occur as el-
All three schools of thought have succeeded in some lipsoids (Fig. 2a, b and d) and some of them are irregu-
case studies, but none of them is universally applicable lar polygons (Fig. 2b and c). All the enclaves are rich in
to the petrogenesis of mafic enclaves in granites. It biotite. They are fine-grained, dark gray in color, and
seems that the origin of mafic enclaves in granites is range in size from less than 1 cm to tens of centimeters
variable, just like the famous statement ‘there are gran- in the longest dimension. Most enclaves are several
ites and granites’ by H. Read. To provide insights into centimeters in size, whereas those smaller than 1 cm
the petrogenesis of mafic enclaves, we performed a and larger than 20 cm are rare. In the transitional zone
combined petrological and geochemical study on between some of the large enclaves and the host gran-
biotite-rich enclaves and their host S-type granites from ites there are biotites coarser than those in the interior
the Jiuling batholith in South China. The results demon- of the enclaves (Fig. 2a). Five host granites and three
strate that the mafic enclaves can be formed by back- mafic enclaves were analyzed in this study.
reaction of peritectic garnet with the host granitic melt
during magma emplacement. This increases our under-
standing of the origin of mafic enclaves in granites as ANALYTICAL METHODS
well as the geochemical diversity of S-type granites. Before handling the enclaves were isolated by remov-
ing any surrounding granite. However, the host granite
cannot be completely prevented from enclave contam-
GEOLOGICAL SETTING ination because some enclaves are too small to be sep-
The Jiuling batholith in South China consists mainly of arated. The rocks were crushed and powdered at the
S-type granitoids of Neoproterozoic age (Fig. 1). It is Institute of Geophysical and Geochemical Exploration
part of the Jiangnan orogen between the Yangtze in the Chinese Academy of Geological Science,
Craton in the NW and the Cathaysia Block in the SE (Li Langfang. After crushing, the samples were processed
et al., 2009; Zhang & Zheng, 2013; Zheng et al., 2013). by heavy-liquid separation, magnetic separation and
The Jiangnan orogen is mainly composed of binocular microscopic selection to separate various
Neoproterozoic metasedimentary and magmatic rocks minerals.
Journal of Petrology, 2017, Vol. 0, No. 0 3
Fig. 1. (a) Simplified tectonic map showing the major cratonic blocks and orogenic belts in China (modified after Zheng et al., 2013);
(b) geological map of the Jiangnan orogen in South China (modified after Yao et al., 2014); (c) simplified geological map showing
the distribution of S-type granites and mafic enclaves at Aozicun in the Jiuling batholith.
Mineral inclusions in garnet and biotite were Mn), spinel (Cr) and apatite (F, Cl). Routine detection
observed under an optical microscope and identified limits were 002 wt % for F and Cl, and 001 wt % for all
by laser Raman spectroscopy (ThermoFisher DXR) at other elements.
the CAS Key Laboratory of Crust–Mantle Materials and The trace element compositions of biotite were meas-
Environments, University of Science and Technology ured by laser ablation inductively coupled plasma mass
of China (USTC). Some large mineral inclusions spectrometry (LA-ICP-MS) at the CAS Key Laboratory of
(>5 lm) on the polished garnet and biotite surfaces Crust–Mantle Materials and Environments, USTC. The ana-
were analyzed using an EDAX GENESIS APEX Apollo lytical procedures have been described in detail by
System (EDS) on a TESCAN MIRA 3 LMH FE scanning Scheibner et al. (2007). The beam size was 44 lm and the
electron microscope at the CAS Key Laboratory of frequency was 10 Hz. Element contents were calibrated
Crust–Mantle Materials and Environments, USTC, against multiple reference materials (NIST610, BCR-2G,
Hefei. Mineral abbreviations used through the text, BIR-1G and BHVO-2G), and the SiO2 contents measured
including the figures and tables, are after Whitney & by electron microprobe (EMP) were used as internal stand-
Evans (2010). ards. Off-line selection and integration of background and
sample signals, and time-drift correction and quantitative
calibration were performed by ICPMSDataCal (Liu et al.,
Biotite major and trace elements 2008, 2010). The analytical uncertainty of our measure-
The major element compositions of biotite in polished ments is less than 610% (2r).
thin sections were analyzed using a JEOL JXA-8230 at
Hefei University of Technology, China. The analyses
were obtained using an accelerating voltage of 15 kV Whole-rock geochemistry
and beam current of 20 nA, with a beam size of 5 lm. Whole-rock major and trace elements were analyzed at the
Reference minerals used in the analyses are biotite State Key Laboratory of Isotope Geochronology and
(Mg, K, Fe, Ti), tremolite (Si), kyanite (Al), garnet (Ca, Geochemistry in the Guangzhou Institute of Geochemistry,
Fig. 2. Representative field and hand-specimen photographs for samples in this study. (a) Fine-grained and ellipsoidal mafic en-
claves in the medium-grained host granite. The gradational rims are coarser-grained and lighter in color than the interior of the en-
claves. (b) Small enclaves scattered in the host S-type biotite granite. They are darker and finer-grained than the host granite. (c)
Small enclaves with different sizes and shapes in hand specimen. (d) One representative small enclave with a panoramic photo-
micrograph and enlarged photomicrograph to show the details.
Chinese Academy of Sciences, Guangzhou. Major elem- Single-stage Hf model ages (TDM1) were calculated rela-
ents were analyzed by inductively coupled plasma atomic tive to the depleted mantle with a present-day
176
emission spectrometry (ICP-AES) using a Varian Vista Pro Hf/177Hf ratio of 028325 and 176Lu/177Hf ratio of
system; the analytical precision is better than 62–5%. 00384 (Griffin et al., 2000), and two-stage Hf model
Trace elements were determined by ICP-MS using a ages (TDM2) were calculated relative to the average con-
Perkin-Elmer Sciex ELAN 6000 system with an analytical tinental crust with a 176Lu/177Hf ratio of 0015 (Griffin
precision better than 65% for most elements. Two rock et al., 2002).
standards (granite GSR-1, basanite GSR-3) and one ran-
domly selected duplicate sample (12JL39B) were chosen
to monitor the data quality and reproducibility. Zircon U–Pb ages and O–Hf isotopes
Whole-rock Rb–Sr, Sm–Nd and Lu–Hf isotope were Zircon grains were selected by hand picking under a
analyzed by multicollector (MC)-ICP-MS on a Neptune binocular microscope. Representative zircon grains
system at the State Key Laboratory of Lithospheric were mounted in epoxy resin and polished to expose
Evolution in the Institute of Geology and Geophysics, the grain center for analysis. Cathodoluminescence (CL)
Chinese Academy of Sciences (CAS), Beijing. Detailed imaging of the zircons was performed using a JXA-
analytical methods were described by Yang et al. 8800R electron microprobe at the Institute of Mineral
(2010). Measured 176Hf/177Hf, 87Sr/86Sr, and 143Nd/144Nd Resources in CAS, Beijing. Both optical photomicro-
ratios were normalized to 179Hf/177Hf ¼ 07325, graphs and CL images were taken as a guide to selec-
86
Sr/88Sr ¼ 01194 and 146Nd/144Nd ¼ 07219, respect- tion of analytical spots.
ively, using the exponential law. During our analyses, Zircon in situ O isotopes were analyzed by sensitive
the 176Hf/177Hf, 87Sr/86Sr and 143Nd/144Nd ratios of the high-resolution ion microprobe (SHRIMP) on the
BCR-2 and GSP-2 standards were identical within error SHRIMP II at the Beijing SHRIMP Center in CAS, Beijing.
to the recommended values (Weis et al., 2006, 2007). The detailed instrument parameters and operating
principles have been described by Ickert et al. (2008). It PETROGRAPHY

took c. 7 min for each 18O/16O analysis (5 min waiting for Host granite
the isotopic composition to stabilize and 2 min of O iso- The granite of the Jiuling batholith mainly consists of
tope measurement). The spot diameters were about quartz, plagioclase, K-feldspar, biotite and muscovite.
20 lm. The TEMORA 2 zircon (d18O ¼ 820‰; Black et al., Quartz (30–40%) and plagioclase (20–30%) occur as
2004) was analyzed at the start of each session and after euhedral and subhedral crystals. K-feldspars (10–15%)
every three samples for calibration of instrumental are present as subhedral grains. The abundance of bio-
mass fractionation. tite in the granite ranges from 15% to 18%. Most of
After the O isotope analysis, the mounted zircons these biotites occur as clots of subhedral and anhedral
were re-polished and vacuum-coated with gold for U– crystals; minor biotites are euhedral and coarse
Pb dating by the SHRIMP ion microprobe at the same grained. Their sizes range from 350 lm to 2 mm in the
domain. The detailed analytical procedure was longest dimension (Fig. 3a). Muscovite (3–5%) is
described by Compston et al. (1992). The U–Pb isotope scarce and occurs as subhedral crystals. The accessory
data were collected in sets of five scans and every four minerals are zircon, garnet, apatite and pyrite.
analyses was followed by a reference zircon TEM
(417 Ma). Common Pb was corrected using the meas-
ured 204Pb, and the model crustal Pb isotope compos- Mafic enclaves
ition of Stacey & Kramers (1975) was used for the Mafic enclaves are widespread in the host granite. Their
common Pb correction. The data were treated following distribution is irregular, neither close to the edge of the
Compston et al. (1992) with the ISOPLOT program of pluton, nor forming an alignment. The enclaves are rich
Ludwig (2003). in biotite (30–45%), and contain quartz (25–35%) and
Zircon Lu–Hf isotope analysis was conducted by LA- plagioclase (30–40%). Generally, the mineral assem-
MC-ICP-MS at the State Key Laboratory of Geological blages of these biotite-rich enclaves are similar to those
Processes and Mineral Resources in the China of the host granite, but with significantly higher abun-
University of Geosciences, Wuhan. Detailed instrumen- dances of biotite and very scarce K-feldspar. The bio-
tal conditions and data acquisition were described by tites have elongated to equigranular shapes and are
Liu et al. (2010). Analyses were performed using a subhedral to anhedral (Fig. 3b–d). Their sizes range
Neptune MC-ICP-MS system equipped with a GeoLas from 200 to 600 lm in the longest dimension. Some
193 nm excimer LA system, with a beam diameter of quartz grains are small and granular, whereas others
44 lm and a repetition rate of 8 Hz. The spots chosen for are large and anhedral with smaller biotite and plagio-
Lu–Hf isotope analyses were near or at the sites of the clase inclusions (Fig. 3b and c). Plagioclase occurs as
U–Pb isotope analyses located using CL images. During laths and subhedral crystals. The distribution of plagio-
our analysis, the standard zircons gave 176Hf/177Hf ratios clase in the enclaves is generally homogeneous
of 0282308 6 0000025 (2r, n ¼ 22) for 91500 and throughout all the thin sections. Some small stumpy
0282025 6 0000022 (2r, n ¼ 4) for GJ-1, within the ana- plagioclase crystals grow together with equigranular
lytical errors (Woodhead & Hergt, 2005; Morel et al., anhedral biotite (Fig. 3d). As described above, some
2008). The decay constant for 176Lu of 1865 1011 a–1 large enclaves have light-coloured and coarse-grained
(Scherer et al., 2001), and the present-day chondritic transitional zones at the contact with the host granite.
ratios of 176Hf/177Hf ¼ 0282772 and 176Lu/177Hf ¼ 00332 These rims contain slightly more quartz and plagioclase
(Blichert-Toft & Albarède, 1997) were adopted for calcu- than the interior of the enclaves.
lating eHf(t) values. In addition to the typical large enclaves with sizes of
5–20 cm, we have also studied some small enclaves
with sizes less than 1 cm (Figs 2d and 4). The mineral
Whole-rock and mineral O isotopes abundances in these small enclaves are biotite 30–45%,
Whole-rock and mineral oxygen isotope analyses were quartz 25–35% and plagioclase 30–40%, similar to those
performed using the laser fluorination technique and a in the large enclaves. The most prominent feature of
Finnigan Delta XP mass spectrometer equipped with a the small enclaves is the occurrence of garnet as broken
25 W MIR-10 CO2 laser at the CAS Key Laboratory of crystals (Figs 4 and 5c), fragments (Fig. 5a) or small
Crust–Mantle Materials and Environments in USTC, debris (Fig. 5b). These garnets are generally anhedral
Hefei. Oxygen isotope data are reported as parts per and their sizes range from 100 to 300 lm. They are rela-
thousand difference (‰) from the reference standard tively scarce in the mafic enclaves and are typically
VSMOW in the d18O notation. Errors for repeat meas- associated with biotite (Fig. 5c–f). The growth of biotite
urements of each standard on a given day were is usually accompanied by consumption of garnet (Fig.
about 6 01‰ (1r) for d18O. Reference minerals used in 5f). Some small garnets are rimmed by biotite (Fig. 5e).
the laboratory are as follows: garnet UWG-2 with d18O Other garnets are surrounded by biotite or even
value of 58‰ (Valley et al., 1995), zircon 91500 with enclosed in biotite (Fig. 5b and d). K-feldspar is absent
d18O value of 100‰ (Zheng et al., 2004) and home in the vicinity of the garnet debris.
standard garnet 04BXL07 with d18O value of 37‰ To study the relationship between garnet and biotite,
(Gong et al., 2007). we separated and mounted some of the garnet grains.
Fig. 3. Representative photomicrographs of the host granites and the enclaves at Aozicun in the Jiuling batholith, South China. (a)
Large euhedral to subhedral biotite in the host granite; (b) subhedral to anhedral elongated biotite with a high aspect ratio and
coarse anhedral quartz crystal in an enclave; (c) subhedral biotite intergrown with euhedral plagioclase embedded in a coarse
quartz crystal; (d) abundant stumpy plagioclase intergrown with equigranular anhedral biotite. All under crossed polars.
Some biotites overgrown as rims on the garnet are anhe- biotites in the host granite that are euhedral and coarse
dral, with sizes ranging from 20 to 50 lm (Fig. 6a). They grained were studied for their mineral inclusions. The
generally coexist with quartz. The inward curvature of mineral inclusions in the biotites from three enclaves
these biotites on the garnet edges suggests the con- (12JL39B, 12JL40B and 12JL41) are quartz, zircon,
sumption of garnet during the overgrowth of the biotite. apatite, plagioclase, ilmenite and titanite. The major differ-
Some biotites occur in the interior of the garnet (Fig. 6b), ence in the populations of mineral inclusions in the bio-
and represent primary inclusions. Furthermore, some tites between the host granites and enclaves lies in the
small leptosomatic biotites (<10 lm) coexist with xeno- abundance of feldspars. High abundances of K-feldspar
morphic quartz on the rims of the garnet (Fig. 6c and d). and low or zero abundances of plagioclase were observed
These coexisting biotite and quartz grains penetrate in the biotites from the host granites, whereas plagioclase
sharply into the rim of the garnet, whereas the garnet is common but K-feldspar is absent in the biotites from
pseudomorph still remains. The morphological charac- the enclaves. The biotites in the enclaves contain fewer in-
teristics of these small leptosomatic biotites are very clusions than the biotites in the host granites.
similar to those of the neoblastic biotites reported in the
melting experiment of Himalayan metapelitic rocks
~ o-Douce & Harris, 1998), which occur predomin-
(Patin
GEOCHEMISTRY
antly as small subhedral flakes (<10 lm). In addition,
these small leptosomatic biotites resemble those lepto- Biotite major and trace elements
somatic biotites in our enclaves (Fig. 3b–d) having a high Based on petrographic observations, different kinds of
aspect ratio. biotite in the host granite and the enclaves were identi-
fied. They include euhedral and coarse-grained biotites
in the granite, those in contact with garnet, those on
Mineral inclusions in biotite garnet rims and biotite inclusions inside garnet. To
Mineral inclusions were identified in the biotites from evaluate the relationship between these biotites and
both the host granites and the enclaves (Table 1). All link them to the garnet debris described above, we ana-
the mineral inclusions listed in Table 1 are primary ra- lyzed the major and trace element compositions of
ther than secondary. The mineral inclusions in biotites these biotites. Generally, all the biotites are homoge-
from two host granites (12JL39A and 12JL40A) are neous in major element composition. Therefore, an
mainly titanite, ilmenite, zircon, apatite, K-feldspar and average composition is calculated for each grain.
quartz. Plagioclase is scarce in the biotite of granite As shown in Fig. 7 and Table 2, biotites in the host gran-
12JL40A and absent in granite 12JL39A. Only magmatic ite have low SiO2 contents of 3360–3616 wt %, MgO
Journal of Petrology, 2017, Vol. 0, No. 0
Fig. 4. Three biotite aggregates (small enclaves) in one slice. Garnet fragments are observed only in (a). The modal proportion of biotite in (a) is obviously lower than that in (b) and
(c). The biotites in (a) are anhedral to subhedral whereas those in (b) and (c) are subhedral to euhedral.
7
Fig. 5. Backscattered electron (BSE) images of representative enclaves showing the texture between garnet and biotite. (a, b) Small
anhedral garnet debris randomly occurring in small enclaves; (c) fractured garnet crystal with high-Mg# biotite in cracks; (d) garnet
fragment surrounded by radiating biotite; (e) biotite rim mantling garnet fragment; (f) biotite embayed into garnet fragment.
contents of 659–770 wt % with Mg# [¼ Mg/(Mg þ Fe) positive Eu anomalies with Eu/Eu* from 230 to 489
molar] of 34–38, A/CNK [¼ Al2O3/(CaO þ Na2O þ K2O) (Fig. 8; Table 3). Their middle REE (MREE)–heavy REE
molar] values of 135–167, high FeOT contents of 2040– (HREE) patterns are flat with (Gd/Yb)N of 031–137. Only
2271 wt %, TiO2 contents of 237–471 wt % and K2O con- one biotite grain in an enclave rim was analyzed for its
tents of 981–1026 wt %. In contrast, biotites in contact trace elements. It shows a similar pattern to the biotites
with garnet, inlaid on the garnet rims and included in the in the host granite, with a weak light REE (LREE)/HREE
garnet interiors have relatively higher SiO2 contents of fractionation (La/YbN ¼ 044), a strong positive Eu anom-
3432–4058 wt %, higher MgO contents of 764–1381 wt aly (Eu/Eu* ¼ 508) and a flat MREE–HREE pattern (Gd/
%, higher Mg# of 40–64, higher A/CNK ratios of 161–217 YbN ¼ 087). However, the biotites in enclaves and in
but relatively lower FeOT contents of 1358–2187 wt %, contact with garnet have distinctive REE patterns with
lower K2O contents of 836–983 wt % and abnormally low (La/Yb)N of 001–010, Eu/Eu* of 027–062 and (Gd/Yb)N
TiO2 contents of 004–403 wt %. Some of the biotites in the of 008–033. The significant negative Eu anomalies indi-
large enclaves have similar compositions to the biotites in cate the co-growth of plagioclase in a relatively closed
the host granite, whereas others resemble the biotites environment.
associated with garnet (Fig. 7). The biotites in enclave rims
have compositions intermediate between these two types. Whole-rock major and trace elements
All the biotites in the host granite display similar Major and trace element data for the host granites and en-
chondrite-normalized rare earth element (REE) patterns claves are presented in Table 4. Nearly all of these sam-
with (La/Yb)N varying from 016 to 077 and strong ples are peraluminous with A/CNK values of 102–147.
Fig. 6. BSE images of representative garnets and associated biotites. (a) Biotites on garnet rims; (b) biotite inclusions inside garnet
crystals; (c, d) micro-leptosomatic biotites (<10 lm) and xenomorphic quartz embayed into garnet grains. These biotites are needle
shaped, without a uniform growth orientation. The coexisting biotites and quartz pseudomorph the garnet.
On the total alkalis–silica (TAS) diagram (Fig. 9a), d18O values of 105–118‰ (Fig. 9b). In Harker variation
all the host granites plot in the granodiorite domain diagrams (Fig. 10) the host granites have consistent
and all the biotite-rich enclaves plot in the diorite field. Na2O contents of 242–275 wt % and K2O contents of
The host granites and enclaves have consistently high 321–425 wt % with K2O/Na2O ratios of 133–169 and
Table 1: Species and frequencies of mineral inclusions in biotites from the Aozicun granites and mafic enclaves in the Jiuling batho-
lith, South China
Granites Mafic enclaves
Sample: 12JL39A 12JL40A 12JL39B 12JL40B 12JL41
Total inclusion numbers/ 72/7 39/3 47/10 25/12 28/8

biotite grains
Major inclusions in Ttn(27), Zrn(13), Ilm(14), Ap(9), Ttn(21),Ap(14), Zrn(7), Pl(6), Ilm(17), Pl(6)
biotites Kfs(13), Ap(13), Zrn(6), Qz(6), Qz(7), Pl(4) Ap(5), Qz(4),
Ilm(4) Kfs(3) Ilm(3)
Minor inclusions in Qz(2) Pl(1) Zrn(1) Qz(2), Zrn(2),
biotites Ap(1)
CaO of 190–214 wt %. They show high Al2O3 contents 493 ppm) with high Th/U ratios of 01–10, which are typ-
of 1508–1689 wt %, high MgO þ FeOT values of 645– ical of zircon of magmatic origin (Hoskin & Schaltegger,
779 wt % and high TiO2 contents of 065–084 wt % 2003). Zircon grains from the enclave (12JL39B) are
with Al2O3/TiO2 ratios of 1967–2448. The Mg# of the subhedral to anhedral, opaque to translucent. They are
host granites ranges from 40 to 41. Compared with the smaller than those from the host granite, with sizes of
host granites, the mafic enclaves show similar Na2O 50–100 mm and length to width ratios of 1:1 to 2:1.
contents of 173–297 wt %, CaO contents of They have higher U contents of 326–3514 ppm and Th
195–464 wt % and similar Al2O3 contents of 153– contents of 77–1539 ppm, but with similar Th/U ratios of
1684 wt %, lower K2O contents of 281–346 wt %, and 016–062.
K2O/Na2O ratios of 100–200, higher FeOT þ MgO val- Fourteen of the 15 analyses on 11 zircon grains of the
ues of 1024–1080 wt % and higher TiO2 contents host granite 12JL39A are concordant. Seven analyses on
of 072–101 wt %, with Al2O3/TiO2 ratios of 1634– relict cores gave apparent 206Pb/238U ages ranging from
2125. The Mg# of the biotite-rich enclaves is slightly 1431 6 14 Ma to 832 6 5 Ma; only one is discordant. The
higher than that of the host granites, ranging from 40 remaining seven analyses yield a weighted mean
206
to 57. Pb/238U age of 822 6 4 Ma (1r, n ¼ 7, MSWD ¼ 116).
The host granites and the enclaves have indistin- Zircon U–Pb dating on 14 zircon grains for enclave
guishable REE and trace element patterns 12JL39B gave a similar result, with a weighted mean
206
(Supplementary Data Fig. S1; supplementary data are Pb/238U age of 823 6 3 Ma (1r, n ¼ 7, MSWD ¼ 094) as
available for downloading at http://www.petrology. crystallization age, and seven ages on relict zircons rang-
oxfordjournals.org). All samples display enrichment in ing from 838 6 5 Ma to 863 6 4 Ma. This is consistent
LREE relative to HREE, with (La/Yb)N from 670 to 1605, with previous dating results from the Jiuling batholith,
negative Eu anomalies with Eu/Eu* from 027 to 079 including 819 6 9 Ma (Li et al., 2003), 828 6 8 Ma (Zhong
and total REE contents from 15399 to 20247 ppm. They et al., 2005) and 815 6 7 Ma (Zhao et al., 2013) for granit-
are characterized by strong enrichment in Rb (1505– oid samples and 813 6 5 Ma for mafic enclaves (Zhao
255 ppm), Th (1035–1835 ppm), U (21–305 ppm), K et al., 2013). It appears that the host granite and the en-
(23317–33689 ppm) and Pb (14–31 ppm), but pro- clave were emplaced at the same time.
nounced negative anomalies in Ba (194–482 ppm), Sr
(628–1465 ppm), Nb (91–162 ppm), Ta (09–18) and Ti
(3896–6053). The calculated Zr saturation temperatures Whole-rock Sr–Nd–Hf–O isotopes
(TZr) for the host granites and enclaves range from 797 Two host granites and three enclaves were analyzed for
to 825 C (av. 806 C) and from 724 to 826 C (av. 780 C), their Rb–Sr, Sm–Nd and Lu–Hf isotope compositions
respectively. Relict zircons occur in both the host gran- (Table 6). Initial (87Sr/86Sr)i ratios, eNd(t) values and eHf(t) are
ite and enclaves, indicating the saturation of Zr in the calculated at t ¼ 820 Ma based on the zircon U–Pb dates.
melts. Therefore, the calculated Zr saturation tempera- The results show that the two host granites and three en-
tures represent the upper limit of temperature for zircon claves have similar initial Sr–Nd–Hf isotope compositions
crystallization (Boehnke et al., 2013). (Fig. 12). The host granites have high (87Sr/86Sr)i ratios of
0708–0712, negative eNd(t) values of –334 and positive
eHf(t) values of 410–749. Their single-stage Nd and Hf
Zircon U–Pb ages model ages (TDM1) are 1856–1937 and 1194–1288 Ma, re-
One host granite and enclave pair (12JL39A and spectively. The enclaves yield high (87Sr/86Sr)i ratios of
12JL39B) was selected for SHRIMP U–Pb dating (Table 0710–0714, eNd(t) values of –274 to –075 and positive
5). Zircon grains from the host granite (12JL39A) are eu- eHf(t) values of 426–469. The moderately negative eNd(t)
hedral to subhedral, colorless and transparent. The values fall in the range of the country-rocks; that is, the
grain sizes range from 100 to 350 mm with aspect ratios Shuangqiaoshan Group and its equivalent strata in the
of 15:1 to 4:1. Most of the zircons show oscillatory or vicinity (Li, 1996; Zhang et al., 2000; Wang et al., 2010).
banded zoning in CL images (Fig. 11). The zircons have The corresponding single-stage Nd and Hf model ages
moderate contents of U (109–699 ppm) and Th (41– (TDM1) are 1626–1870 and 1249–1319 Ma, respectively.
Fig. 7. Major element compositions of biotite from the Aozicun granites and enclaves in the Jiuling batholith, South China.
Whole-rock O isotope analyses for five granites and same domains as for the U–Pb dating. The results are
three enclaves yield consistently high d18O values of presented in Fig. 13 and Table 7. The eHf(t) values and
105–118‰ (Fig. 9b and Table 4). Hf model ages were calculated at t ¼ 820 Ma for syn-
magmatic zircons, whereas the relict zircon cores were
Zircon Hf–O isotopes calculated at their apparent U–Pb ages.
In situ zircon Hf–O isotopic analyses of the host granite For host granite 12JL39A, the synmagmatic zircon
and enclave were conducted on the same spot or the domains have essentially constant eHf(t) values of
Table 2: Average major element compositions of representative biotites in the Aozicun granites and mafic enclaves from the
Jiuling batholith in South China
Bt in the host granite
bt1 bt2 bt3 bt4 bt5 bt6 bt7 bt8 bt9 bt10 bt11 bt12 bt13 bt14 bt15 bt16 bt17 bt18
SiO2 (wt %) 3616 3581 3606 3560 3574 3543 3569 3561 3578 3572 3526 3556 3522 3537 3541 3530 3360 3410
TiO2 308 271 237 279 261 246 287 275 282 265 423 401 384 468 279 334 453 421
Al2O3 1749 1799 1820 1704 1693 1762 1739 1716 1697 1724 1767 1775 1805 1753 1765 1791 1762 1757
Cr2O3 004 005 007 002 001 002 002 002 002 001 001 001 002 000 010 010 001 002
FeO 2234 2264 2211 2255 2193 2189 2221 2271 2245 2207 2157 2155 2123 2040 2237 2239 2111 2089
MnO 036 037 036 015 016 015 015 015 016 015 009 011 011 011 001 001 012 010
MgO 770 760 745 715 729 740 703 715 723 747 689 671 695 668 696 659 700 717
CaO 000 000 000 000 000 005 002 000 000 000 000 000 006 000 000 000 000 000
Na2O 013 013 015 011 008 023 016 007 008 007 008 008 009 009 010 007 117 065
K2O 986 988 981 1014 1009 990 1005 1017 1014 1026 1016 1016 1003 998 993 1003 1026 1029
NiO 000 000 000 002 002 001 003 002 002 002 001 002 002 003 002 002 003 002
F 000 000 000 008 009 013 006 013 012 011 012 011 013 010 016 008 012 012
Cl 000 000 000 002 002 002 002 002 002 002 002 003 003 003 003 003 003 003
Total 9717 9719 9658 9564 9494 9525 9568 9590 9575 9573 9606 9604 9570 9494 9544 9583 9554 9510
Mg# 3805 3742 3754 3612 3721 3758 3607 3595 3645 3762 3629 3571 3685 3686 3567 3442 3715 3795
A/CNK 160 165 167 152 153 157 155 154 153 153 158 159 162 160 161 163 135 144
Bt in the host granite Bt in the enclave
SiO2 (wt %) 3585 3529 3517 3712 3710 3616 3515 3582 3622 3617 3661 3626 3626 3661 3593 3857 3606 3626
TiO2 450 471 376 127 125 396 211 284 232 251 202 232 232 142 365 137 430 232
Al2O3 1743 1733 1750 1784 1792 1869 1899 1813 1756 1744 1771 1710 1710 1659 1866 1558 1790 1710
Cr2O3 003 001 001 004 005 004 000 001 002 002 004 001 001 011 004 010 003 001
FeO 2125 2098 2139 1979 1938 2119 2082 2090 2168 2131 2106 2111 2111 003 2130 001 2152 2111
MnO 011 010 009 025 026 024 010 010 020 020 018 011 011 1844 023 1525 023 011
MgO 704 673 695 964 1010 710 783 757 854 862 922 847 847 1081 738 1310 713 847
CaO 000 001 000 000 000 000 000 000 001 003 001 000 000 000 000 000 000 000
Na2O 007 012 014 012 009 017 024 021 027 015 012 008 008 006 030 005 023 008
K2O 1013 994 998 983 978 975 977 976 937 932 954 1022 1022 1022 936 1031 943 1022
NiO 002 001 001 000 000 000 000 000 000 000 000 003 003 000 000 000 000 003
F 011 011 008 000 000 000 017 017 034 030 036 016 016 017 000 033 000 016
Cl 003 002 003 000 000 000 002 003 010 007 003 003 003 001 000 001 000 003
Total 9651 9531 9508 9590 9594 9731 9516 9547 9646 9598 9673 9583 9583 9441 9685 9457 9685 9583
Mg# 3712 3637 3670 4648 4816 3740 4014 3924 4125 4223 4384 4170 4170 5109 3817 6049 3714 4170
A/CNK 157 158 158 164 166 172 173 166 165 168 168 152 152 148 175 138 169 152
Bt in the enclave
SiO2 (wt %) 3624 3591 3694 3556 3569 3539 3547 3526 3572 3534 3917 3565 3570 3568 3583 3622 3633 3586
TiO2 417 422 160 294 377 300 266 356 320 351 131 381 394 403 400 230 245 416
Al2O3 1832 1783 1776 1819 1738 1817 1815 1784 1814 1770 1650 1693 1729 1754 1828 1795 1784 1823
Cr2O3 006 006 006 001 001 000 002 000 001 001 010 001 003 000 005 002 002 004
FeO 2149 2238 1958 2059 2160 2104 2040 2094 2024 2129 1591 2244 2168 2167 2142 2160 2125 2205
MnO 026 023 027 011 012 009 010 011 011 010 027 011 010 011 023 018 018 021
MgO 737 746 989 803 746 724 790 711 754 727 1291 741 713 704 752 870 884 714
CaO 004 000 000 000 000 001 000 000 001 001 000 000 000 000 000 002 000 000
Na2O 027 023 011 020 017 020 019 015 019 018 007 026 024 022 025 016 008 021
K2O 929 951 986 974 981 979 971 984 980 976 991 955 982 974 954 950 951 967
NiO 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000
F 000 000 000 013 018 017 019 018 020 018 000 019 018 012 000 030 023 000
Cl 000 000 000 003 003 004 003 003 004 003 000 003 003 003 000 006 002 000
Total 9751 9782 9605 9548 9616 9509 9476 9497 9511 9532 9616 9633 9607 9615 9712 9685 9665 9756
Mg# 3795 3727 4739 4100 3809 3804 4084 3771 3989 3783 5912 3707 3697 3667 3851 4180 4257 3660
A/CNK 173 167 163 167 159 166 167 163 165 162 152 157 156 161 170 169 171 168
(continued)
27–59, with single-stage Hf model ages of For enclave 12JL39B, the synmagmatic zircons have
1146–1279 Ma and high d18O values of 84–106‰. eHf(t) values of 38–71, with single-stage Hf model ages
The relict zircon cores exhibit highly variable eHf(t) of 1103–1229 Ma and d18O values of 89–101‰. The rel-
values from 04 to 134, with single-stage Hf ict zircon cores yield relatively higher eHf(t) values of
model ages of 1163–1377 Ma and d18O values of 28– 41–113, with single-stage Hf model ages of 998–
99‰. 1246 Ma and d18O values of 80–92‰.
Table 2: Continued
Bt in the enclave Bt in the enclave rim
SiO2 (wt %) 3604 3609 3533 3529 3901 3564 3552 3561 3532 3880 3676 3528 3524 3517 3538 3639 3596 3583
TiO2 428 395 386 355 139 349 401 357 395 128 150 406 410 390 379 167 227 248
Al2O3 1806 1859 1744 1765 1657 1742 1759 1795 1757 1563 1662 1755 1718 1750 1690 1768 1756 1722
Cr2O3 004 006 000 002 007 001 002 002 001 011 012 002 002 001 003 000 000 000
FeO 2110 2082 2200 2192 1618 2180 2132 2179 2203 002 002 2204 2175 2105 2184 2010 2013 2036
MnO 017 018 009 009 026 009 010 009 009 1530 1814 009 010 010 009 016 020 022
MgO 722 736 713 716 1267 758 725 700 689 1302 1078 723 693 721 716 901 917 867
CaO 000 001 000 000 000 000 001 000 000 000 000 000 000 000 003 000 000 001
Na2O 017 021 025 017 008 020 026 019 019 006 006 020 024 024 028 021 019 020
K2O 972 971 965 969 981 956 956 984 983 1032 1038 983 959 957 956 919 908 913
NiO 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000
F 000 000 017 015 000 016 017 015 018 035 019 015 015 017 016 013 017 012
Cl 000 000 004 004 000 004 003 003 004 001 002 003 003 004 005 004 004 004
Total 9679 9698 9589 9568 9604 9593 9580 9619 9604 9478 9451 9643 9528 9489 9521 9452 9470 9420
Mg# 3788 3864 3662 3677 5827 3828 3774 3641 3581 6027 5145 3688 3624 3793 3689 4440 4481 4315
A/CNK 167 171 160 164 154 163 163 164 160 138 146 160 159 163 155 172 173 168
Bt in the enclave rim Bt in contact with garnet
SiO2 (wt %) 3596 3592 3601 3605 3559 3599 3609 3596 3610 3585 3614 3688 3906 3763 3623 3583 3829 3854
TiO2 255 270 273 245 073 184 403 282 064 328 286 000 034 061 060 055 032 050
Al2O3 1700 1717 1705 1874 1936 1962 1780 1875 1824 1781 1799 2044 1788 1949 1824 1793 1738 1691
Cr2O3 000 000 000 017 018 026 024 013 014 009 018 003 024 001 006 011 018 004
FeO 2090 2020 2073 1988 2059 2023 2023 1939 2095 1986 1914 1568 2055 1927 2152 2161 1951 1943
MnO 020 022 019 010 005 012 041 041 030 005 005 009 029 025 035 032 032 032
MgO 817 879 825 857 846 825 770 831 864 892 917 1194 815 764 855 848 892 897
CaO 026 004 022 001 000 000 000 000 000 000 011 001 000 001 000 000 000 000
Na2O 014 015 012 014 033 038 009 010 003 022 032 022 003 004 003 004 015 002
K2O 901 924 893 951 903 892 968 983 975 926 885 960 860 974 970 924 954 964
NiO 000 000 000 003 000 001 003 001 002 003 003 000 001 000 004 002 000 000
F 011 016 010 021 031 014 008 023 027 027 027 044 012 026 027 027 047 031
Cl 003 004 003 003 003 002 003 003 002 002 002 001 002 002 002 002 004 003
Total 9428 9455 9431 9580 9451 9573 9637 9586 9498 9554 9500 9514 9522 9484 9547 9430 9482 9457
Mg# 4104 4368 4152 4344 4228 4209 4041 4332 4210 4447 4607 5758 4143 4140 4146 4118 4489 4514
A/CNK 163 166 166 177 187 190 167 173 172 171 174 190 191 183 173 178 164 161
Bt on garnet rim Bt inclusions in garnet
SiO2 (wt %) 3904 3696 4016 3707 3743 3534 3891 3568 3578 3432 3614 3627 3504 3530 3752 4058 3753 3664
TiO2 005 288 034 112 049 052 007 048 104 133 006 032 029 004 026 017 008 009
Al2O3 1895 1841 1958 1946 2131 1874 2183 1878 1884 1867 1849 2291 1901 1914 1913 1825 2043 2073
Cr2O3 023 057 011 019 007 014 008 016 024 016 002 001 005 004 002 008 003 010
FeO 1357 1671 1834 1725 1693 1718 2031 1906 1858 2183 2135 1757 2187 2176 1999 1930 1662 1694
MnO 011 011 012 010 008 006 121 018 026 033 009 008 038 042 017 017 008 026
MgO 1381 1163 953 1058 1106 1233 825 1043 1028 846 1051 912 836 837 924 943 1225 1081
CaO 002 000 000 000 001 000 005 000 000 000 004 004 000 000 000 001 003 000
Na2O 023 019 019 015 031 029 019 016 024 017 032 049 005 003 011 009 031 026
K2O 865 933 836 964 948 931 624 900 916 937 849 893 971 976 928 870 927 914
NiO 000 001 000 000 001 006 002 002 000 000 001 000 001 000 000 000 002 001
F 056 027 030 037 053 033 028 023 045 014 037 017 048 018 005 016 061 017
Cl 002 003 002 002 004 002 002 004 002 003 004 003 002 003 004 003 005 003
Total 9500 9697 9691 9578 9749 9416 9731 9409 9469 9473 9576 9586 9507 9500 9578 9687 9703 9510
Mg# 6446 5538 4809 5224 5380 5613 4199 4938 4967 4085 4673 4806 4051 4066 4518 4657 5677 5321
A/CNK 194 176 209 182 197 177 305 187 182 179 189 217 179 180 187 190 193 200
DISCUSSION cordierite 6 sillimanite. These coronas are interpreted as

Biotites formed by back-reaction of peritectic formed by the back-reaction of spinel with a silica-
garnet undersaturated melt during a heating–cooling cycle.
Back-reaction of peritectic minerals with anatectic melts Beard et al. (2004, 2005) interpreted amphibole and bio-
was identified by Kriegsman & Hensen (1998) in their tite rims on partially reacted and resorbed pyroxenes
study on partial melting of metapelites. They found that and Fe–Ti oxides as the principal petrographic evidence
peritectic spinel in the metapelites is separated from the for hydration crystallization, which is the reverse of
felsic leucosome by coronas of garnet 6 sillimanite or dehydration-melting reaction. Similar back-reaction
phenomena are recognized in other studies (e.g. Waters, needles and flakes represent the product of melt–garnet
2001; Erdmann et al., 2009; Lavaure & Sawyer, 2011). We reaction. Some garnet crystals with larger biotites on
propose that the biotites associated with garnet in the their rims (Figs 5e and 6a) suggest higher degree of re-
host granite and the biotite aggregates (small enclaves) action. Other textures, such as embayment of biotite
are also produced by the back-reaction of peritectic gar- and quartz into garnet crystals (Fig. 5f), the surrounding
net with a granitic melt via the following reaction: of small garnet grains by biotite and quartz (Fig. 5d), the
fracture of garnet by biotite (Fig. 5c) and small garnet
Garnet þ K-feldspar þ melt þ H2 O grains in biotite aggregates (Figs 4 and 5a, b), all repre-
¼ Biotite þ Plagioclase þ Quartz þ melt: (I)
sent replacement of garnet by biotite to different de-
The petrological and geochemical evidence for the grees. (2) These metasomatic biotites show some
back-reaction includes the following. (1) Metasomatic chemical characteristics inherited from garnet. Firstly,
textures at various scales: the needle-shaped biotites they have high MgO, low FeOT contents and high Mg#
(<10 lm) growing with xenomorphic quartz penetrate (Fig. 7). This is consistent with the reported metasom-
into the garnet crystals (Fig. 6c and d). The morphology atic biotites by Lavaure & Sawyer (2011). Secondly,
of these small leptosomatic biotites is very similar to they have low K2O and TiO2 concentrations, consistent
those neoblastic biotites produced by incongruent with the consumption of K2O and TiO2 by the formation
breakdown of muscovite in melting experiments of metasomatic biotite via reaction (I). The enrichment
~ o-Douce & Harris, 1998). These subhedral biotite
(Patin in HREE and fractionated LREE/HREE patterns in biotite
associated with garnet (Fig. 8) are also consistent with
the garnet inheritance. Its negative Eu anomaly is con-
sistent with plagioclase growth via reaction (I). It is
noted that the primary biotite inclusions in the garnet
have similar compositions to the metasomatic biotites.
These inclusions are very likely to be the residue of bio-
tites from the granite source. This implies that reaction
(I) is the reverse of biotite dehydration melting.
Mechanism for the back-reaction of peritectic

garnet
Reaction (I) can be rewritten as
ðFe; MgÞ3 Al2 ½SiO4 3 þ 0 5 K2 OðmÞ þ H2 OðmÞ
(II)
¼ KðFe; MgÞ3 ½AlSi3 O10 ðOHÞ2 þ 0 5 Al2 O3 ðmÞ
Fig. 8. REE patterns for different kinds of biotite from the where oxides represent components in the granitic
Aozicun granites and enclaves in the Jiuling batholith, South melt. Because quartz and plagioclase are abundant in
China. the system, the SiO2 activity is buffered and the Na2O
Table 3: Representative REE concentrations (ppm) of biotites in the Aozicun granites and mafic enclave from the Jiuling batholith in
South China
Bt in host granite Bt in Bt in Bt in contact
enclave rim enclave with garnet

bt1 bt2 bt3 bt4 bt bt bt
La 0054 0417 0057 0045 0117 0011 0028

Ce 0326 2086 0247 0108 0267 0150 0120
Pr 0079 0305 0033 0017 0044 0005 0010
Nd 0398 1322 0358 0171 0293 0193 0113
Sm 0186 0687 0149 0136 0090 0130 0069
Eu 0136 0500 0166 0161 0223 0012 0015
Gd 0142 0644 0122 0074 0200 0132 0079
Tb 0033 0079 0021 0030 0058 0058 0006
Dy 0234 0802 0076 0127 0241 0366 0231
Ho 0046 0159 0022 0045 0048 0144 0017
Er 0132 0391 0074 0132 0146 0520 0102
Tm 0043 0071 0027 0026 0028 0099 0016
Yb 0140 0389 0094 0197 0191 1288 0196
Lu 0040 0073 0013 0035 0042 0134 0026
La/YbN 028 077 044 016 044 001 010
Gd/YbN 084 137 107 031 087 008 033
Eu/Eu* 257 230 378 489 508 027 062
Subscript N indicates REE normalized to chondrites from Sun & McDonough (1989).
Table 4: Whole-rock compositions and oxygen isotope ratios of the Aozicun granites and mafic enclaves from the Jiuling batholith,
South China
Sample no.: 12JL35 12JL36 12JL38 12JL39A 12JL40A 12JL39B 12JL40B 12JL41
Rock type: biotite biotite biotite biotite biotite biotite-rich biotite-rich biotite-rich
granite granite granite granite granite enclave enclave enclave
Major elements (wt %)

SiO2 6294 6791 6338 6515 6342 5912 5967 6181
Al2O3 1652 1508 1689 1646 1664 153 1684 1651
Fe2O3 631 528 574 524 548 686 854 827
CaO 214 205 204 19 213 464 231 195
MgO 211 176 200 173 184 463 29 28
Na2O 252 242 251 275 261 173 297 281
K2O 347 321 425 366 406 346 296 281
Cr2O3 001 001 001 001 001 002 001 001
TiO2 084 065 069 068 072 072 101 101
MnO 01 009 009 008 009 01 013 015
P2O5 0139 0142 0144 0158 0167 0155 0159 0122
LOI 155 129 154 157 165 181 192 167
Total 9868 9994 9934 9942 9886 9858 9943 9994
A/CNK 140 135 136 138 132 102 137 147
MgO þ FeOT 779 651 717 645 677 108 1059 1024
Trace elements (ppm)
Rb 200 1505 1790 195 203 1885 255 220
Ba 365 330 482 393 481 224 194 261
Sr 1125 997 1130 1165 1225 1465 754 628
Th 1745 1425 1515 136 154 1035 135 1835
U 293 257 210 261 305 212 215 231
Nb 14 103 115 107 117 91 162 157
Ta 15 10 10 11 12 09 16 18
Pb 24 22 31 25 31 16 15 14
Zr 266 225 220 208 225 195 223 260
Hf 74 66 62 57 63 51 62 79
Y 446 237 394 212 38 252 277 177
Ga 253 196 225 249 241 204 265 268
V 121 95 98 107 116 126 182 175
La 381 342 369 327 335 289 311 358
Ce 829 730 788 71 729 638 682 786
Pr 973 859 904 832 873 764 807 939
Nd 355 331 359 309 327 284 302 349
Sm 774 702 772 647 719 619 66 717
Eu 12 122 137 134 146 154 077 059
Gd 76 580 681 597 736 581 608 632
Tb 118 086 114 08 117 082 092 082
Dy 723 468 683 393 673 453 506 39
Ho 152 089 146 075 138 09 099 069
Er 441 264 415 196 387 251 272 175
Tm 064 036 063 028 056 036 038 025
Yb 408 225 385 171 351 223 24 16
Lu
P 064 035 058 027 054 036 04 026
REE 20247 17496 19518 1664 1816 15399 16389 18204
(La/Yb)N 670 1090 687 1372 685 930 930 1605
Eu/Eu* 048 058 058 066 061 079 037 027
d18OWR (‰) 1070 1071 1095 1181 1087 1090 1050 1074
TZr ( C) 825 810 798 798 797 724 792 826
Zircon saturation temperature were calculated after Boehnke et al. (2013); subscript N indicates REE normalized to chondrites from
Sun & McDonough (1989).
and CaO activities are roughly buffered, thus they are schematically in Fig. 14a. As can be seen from equation
not considered in this reaction. For simplicity, the end- (III) and Fig. 14a, increases in K2O and H2O activities or
member chemical formulae for garnet and biotite are a decrease in Al2O3 activity in the melt will lower the
theoretical. At a given temperature and pressure, the free energy and drive the conversion of garnet to
Gibbs free energy for reaction (II) can be calculated: biotite.
It is also clear that an increase of H2O activity in the
–DGr =RT ¼ 0 5 ln aAl2O3ðmÞ – 0 5 ln aK2OðmÞ – ln aH2OðmÞ
melt must be the primary driver of converting garnet to
(III) biotite. However, granitic melts produced by anatexis of
where R is the gas constant, T is temperature, and –DGr crustal rocks are usually unsaturated in water at the ini-
represents the P–T specific Gibbs free energy of pure tial stage (Clemens, 1984; Scaillet et al., 2000; Clemens
substances at a standard state. This is illustrated & Stevens, 2012). Thus, this initially low aH2O granitic
calculated by Fe2þ–Fe3þ–Mg2þin our biotites

(Supplementary Data Fig. S2). The conditions at the
FMQ (fayalite–magnetite–quartz) buffer were also mod-
eled, showing little influence on the results. A closed-
system equilibrium is assumed because of insufficient
crystal–melt segregation in granitic magma. All main
mineral phases in our host granites were reproduced
by the modeling results (Fig. 14b and c). The water con-
tent in the residual melts increase gradually with a de-
crease in melt fraction. For a magma system with an
initial water content of 4 wt % and 2 wt %, water will be-
come saturated at a crystallinity of 65% and 85%, re-
spectively. Hence, along with the crystallization of the
granitic magma, the water contents in the magma will
become high enough to convert most of the garnets to
biotites.
The garnets that reacted to form biotites are con-
sidered peritectic in origin. They are generally anhedral
and most of them occur as fragments or broken grains
in biotite aggregates (small enclaves), consistent with
typical peritectic garnets in S-type granites (Dorais &
Tubrett, 2012). The occurrence of biotite inclusions in
the garnets also implies that they are of peritectic ori-
gin, because magmatic garnet usually crystallizes ear-
lier than biotite (Fig. 14b) and thus is unlikely to enclose
biotite. As discussed above, these biotite inclusions
seem to be the residue from the source because they
have higher Mg# and lower Ti than the magmatic biotite
in the host granite (Fig. 7). As such, only the peritectic
garnet has the opportunity to capture the biotite resi-
due. Hence, the observed garnets in our samples
should be peritectic ones that were produced by the in-
Fig. 9. Whole-rock compositions of granites and enclaves at
Aozicun in the Jiuling batholith, South China. (a) Total alkalis congruent dehydration melting of biotite in the metase-
vs SiO2 (TAS) diagram (after Middlemost, 1994); (b) d18O vs A/ dimentary rock and were subsequently entrained into
CNK diagram. The ranges of d18O values and A/CNK ratios for the granitic magma.
I- and S-type granites are from O’Neil & Chappell (1977) and
Chappell & White (2001). At the level of emplacement, extensive back-reaction
of peritectic garnet would occur for the following rea-
sons: (1) peritectic garnet from the magma source falls
melt cannot directly react with the garnet. Because out of its original stability owing to the change of pres-
water behaves like an incompatible element, it has been sure, temperature and oxygen fugacity in the magmatic
shown that any magma that contains even a trace system during the migration and ascent of the granitic
amount of water will eventually become water- magma; (2) the change of magma composition owing
saturated after sufficient differentiation (Eggler, 1972). to fractional crystallization (i.e. elevated H2O and K2O
Therefore, only in a residual melt where the water con- content in the magma) would drive the conversion
tent is high enough will the reaction of garnet to biotite of garnet into biotite; (3) peritectic minerals are frag-
occur. mented or broken, which favors the infiltration of mag-
We have modelled the crystallization of minerals and matic melts into the peritectic garnet for extensive
the H2O content variation of a residual melt during evo- back-reaction.
lution of the host magma using Rhyolite-MELTS
(Gualda et al., 2012; Gualda & Ghiorso, 2015). One host
granite sample (12JL39A) is selected as the initial The origin of mafic enclaves in S-type granites
magma with addition of different H2O contents (2–4 wt Several models have been proposed to explain the ori-
%). The adoptive pressure is 45 kbar, corresponding to gin of mafic enclaves in S-type granites, including (1)
a crustal depth for the emplacement of common S-type restite from anatectic sources, (2) hybridization between
granitic magmas (e.g. Anderson et al., 2008). This is mafic and felsic magmas, and (3) early formed cumu-
also consistent with the greenschist-facies metamorph- lates from the host granitoid magma.
ism recorded by the country-rocks of the Jiuling batho- In the restite model (Chappell et al., 1987; Chen et al.,
lith (Wang et al., 2007). The selected oxygen fugacity 1989; Chappell & White, 1991; White et al., 1999;
was below the NNO (Ni–NiO) buffer according to that Chappell & Wyborn, 2012; Wyborn, 2013), mafic
Journal of Petrology, 2017, Vol. 0, No. 0
Fig. 10. Whole-rock variations of major elements and some trace elements for the Aozicun granites and enclaves in the Jiuling batholith, South China.
17
Table 5: SHRIMP zircon U–Pb data for the Aozicun granites and mafic enclaves from the Jiuling batholith, South China
Spot Element (ppm) Th/U Isotope ratio Age (Ma)

207 207 206 207 207 206
U Th Pb/ 1r Pb/ 1r Pb/ 1r Pb/ 1r Pb/ 1r Pb/ 1r
206 235 238 206 235 238
Pb U U Pb U U
12JL39A
12JL39A-1 331 41 012 00661 00009 12429 00194 013637 000096 810 29 820 9 824 5 M
12JL39A-2 115 44 038 00646 00022 12106 00433 013596 000148 761 72 805 20 822 8 M
12JL39A-3 320 160 050 00632 00012 11877 00300 013629 000231 715 40 795 14 824 13 D
12JL39A-4 324 55 017 00646 00011 11949 00220 013405 000095 763 36 798 10 811 5 M
12JL39A-5 580 58 010 00655 00007 12429 00152 013767 000081 790 22 820 7 832 5 I
12JL39A-6 492 493 100 00671 00007 12600 00148 013629 000083 839 21 828 7 824 5 M
12JL39A-7 374 46 012 00671 00007 12542 00160 013558 000091 841 23 825 7 820 5 M
12JL39A-8 109 59 054 00911 00012 31229 00529 024855 000271 1449 25 1438 13 1431 14 I
12JL39A-9 110 55 050 00665 00020 12978 00405 014148 000151 823 61 845 18 853 9 I
12JL39A-10 483 310 064 00655 00007 12557 00162 013900 000085 791 24 826 7 839 5 I
12JL39A-11 256 158 062 00666 00012 12501 00248 013608 000113 826 38 823 11 822 6 M
12JL39A-12 386 388 101 00655 00008 12724 00173 014082 000093 791 25 833 8 849 5 I
12JL39A-13 427 118 028 00665 00008 12583 00170 013731 000087 821 25 827 8 829 5 M
12JL39A-14 699 142 020 00665 00006 13035 00144 014211 000100 823 18 847 6 857 6 I
12JL39A-15 198 79 040 00671 00013 12755 00265 013793 000119 840 39 835 12 833 7 I
12JL39B
12JL39B-1 713 112 016 00655 00007 12232 00149 013540 000074 791 23 811 7 819 4 M
12JL39B-2 339 82 024 00648 00011 12113 00219 013558 000093 768 35 806 10 820 5 M
12JL39B-3 665 399 060 00662 00005 12552 00153 013742 000130 814 16 826 7 830 7 M
12JL39B-4 2118 798 038 00659 00004 12445 00099 013700 000061 803 14 821 4 828 3 M
12JL39B-5 467 92 020 00669 00007 12811 00152 013884 000089 835 21 837 7 838 5 I
12JL39B-6 1135 373 033 00658 00005 12712 00108 014011 000071 800 14 833 5 845 4 I
12JL39B-7 2730 649 024 00661 00003 12985 00086 014243 000065 810 10 845 4 858 4 I
12JL39B-8 1462 254 017 00660 00004 12425 00096 013656 000066 806 13 820 4 825 4 M
12JL39B-9 326 77 024 00669 00009 12570 00192 013617 000099 836 28 827 9 823 6 M
12JL39B-10 3514 1539 044 00659 00003 13012 00076 014320 000063 803 8 846 3 863 4 I
12JL39B-11 511 155 030 00651 00011 12594 00236 014029 000088 778 37 828 11 846 5 I
12JL39B-12 1410 738 052 00656 00004 12576 00098 013898 000067 794 13 827 4 839 4 I
12JL39B-13 1593 666 042 00658 00004 12285 00105 013542 000074 800 14 814 5 819 4 M
12JL39B-14 2979 752 025 00659 00003 12648 00086 013917 000063 804 11 830 4 840 4 I
M, magmatic zircon; I, inherited core; D, discordant core.
enclaves are composed of both residual and peritectic source, they should contain more inherited zircons than
minerals of a refractory nature and the residual min- the host granite, which is not observed in our study.
erals are thus inherited from the source rock. These re- Hence, the restite model is invalid at least for the
fractory minerals are successively unmixed from the biotite-rich enclaves in the Jiuling batholith.
host granitic melt during the rise of a crystal mush from The mixing model (Reid et al., 1983; Vernon, 1984,
the anatectic source. A diagnostic feature of restite 2007, 2014; Elburg & Nicholls, 1995; Maas et al., 1997;
unmixing is complementary compositions between the Collins et al., 2000; Barbarin, 2005; Yang et al., 2007,
enclaves and their host granite (e.g. Chappell et al., 2015; Chen et al., 2009) is the most popular explanation
1987). However, the enclaves and host granites of this for the formation of mafic enclaves in felsic granitoids.
study show similar compositions in most major and It is envisaged that the mafic enclaves in a granitoid are
trace elements (Fig. 10 and Supplementary Data Fig. globules of mantle-derived mafic magma that mingles
S2). Further evidence for the restite model is the pres- or mixes with the host felsic magma. If the enclaves
ence of residual minerals or residual cores, such as cal- represent the mantle-derived magma, they would be
cic plagioclase crystals, plagioclase cores with very gabbroic in lithochemical composition. However, our
high An contents, magnesium cordierites, orthopyrox- biotite-rich enclaves are dioritic rather than gabbroic
ene cores enclosed in biotite and inherited zircons (e.g. (Fig. 9a). Importantly, the mineral assemblage of our en-
Chappell et al., 1987; Chen et al., 1989; Chappell & claves is biotite, quartz and plagioclase, with a lack of
Wyborn, 2012). The magnesium cordierites and ortho- mafic minerals such as clinopyroxene and amphibole.
pyroxene cores can be interpreted as peritectic minerals In addition, a linear lithochemical variation between en-
formed by the incongruent dehydration melting of bio- claves and host granites is usually regarded as key
tite (e.g. Clemens, 1984; Ellis & Obata, 1992; Watkins proof for hybridization between the mantle-derived en-
et al., 2007; Ward et al., 2008). More importantly, the clave-forming mafic magma and the host felsic magma
plagioclases in the enclaves and host granites in the (e.g. Collins et al., 2000; Barbarin, 2005). This is not
present study have identical compositions (not shown). observed in our samples (Fig. 10). Isotope compositions
Furthermore, if the enclaves were restites from the may offer a means to testify the hybridization between
Fig. 11. Zircon CL images and U–Pb concordia diagrams for the Aozicun granite and enclave from the Jiuling batholith, South
China. Black ellipses represent concordant zircon U–Pb crystallization ages, whereas red ellipses are ages that are disconcordant or
were obtained on inherited grains.
the mantle-derived enclave-forming mafic magma and enclaves are considered to be cumulates formed at an
the host felsic magma (e.g. Maas et al., 1997; Chen earlier stage of the same magma system as the host
et al., 2009). The Sr–Nd–Hf–O isotope compositions of granite. However, minerals in the mafic enclaves of our
our enclaves and host granites overlap with each other study (Figs 2 and 3) are finer grained and less automor-
and are thus indistinguishable (Figs 9b and 12), incon- phic, inconsistent with the cumulate crystal texture
sistent with the mixing model. Although the similar iso- described by Irvine (1982). In addition, the notably high
tope compositions between enclaves and host granite viscosity of granitic melts is still a potential obstacle for
can be argued to be a result of geochemical equilibra- separation of crystals from them (Glazner, 2014) when
tion between the different compositions of magmas referring to cumulates in granites. The consistent zircon
(e.g. Barbarin, 2005; Alves et al., 2015; Wang et al., U–Pb ages and Sr–Nd–Hf–O isotope compositions be-
2015), zircon Hf–O isotopes in the enclaves and host- tween the enclaves and the host granites can be pro-
rocks cannot be equilibrated readily because of the duced as long as they come from the same source at
sluggish rate of zircon Hf isotope exchange, which the same time. More importantly, our enclaves are en-
stands as a powerful tool to test the mixing model (e.g. riched in biotite, which has different geochemical char-
Yang et al., 2007). The zircon Hf–O isotopes in our en- acteristics from that in the host biotite granite (Figs 7
claves and host granites are similar to each other (Fig. and 8). This is inconsistent with the cumulate model be-
13). All the syn-magmatic zircons exhibit high d18O val- cause biotite does not crystallize until the middle to late
ues (>8‰) and low eHf(t) values (Fig. 13). Therefore, our stage of granitic magma evolution (Fig. 14c and d).
biotite-rich enclaves cannot be explained by a mixing Therefore, the cumulate model is also not applicable to
model. our biotite-rich enclaves.
A cumulate model has been proposed to account for All the models discussed above may be applicable to
mafic enclaves that have similar mineral assemblages, some mafic enclaves in granites. But none of them can
mineral compositions, crystallization ages and isotope adequately account for our biotite-rich enclaves.
compositions to their host granites (Tindle & Pearce, Instead, our field observations, petrographic and geo-
1983; Clemens & Wall, 1984; Dodge & Kistler, 1990; chemical data indicate that these enclaves are produced
Donaire et al., 2005; Collins et al., 2006; Huang et al., by the back-reaction of garnet aggregates with hydrous
2014; Chen et al., 2015; Moita et al., 2015). Mafic granitic melt at a later stage of the host granitic magma
Table 6: Whole-rock Rb–Sr, Sm–Nd and Lu–Hf isotope compositions of the Aozicun granites and mafic enclaves from the Jiuling
batholith, South China
87
Sample no. Rb Sr Rb/86Sr 87
Sr/86Sr 2r (87Sr/86Sr)i
(ppm) (ppm)
12JL39A 195 1165 486 0768618 0000091 0711741

12JL40A 203 1225 481 0764687 0000036 0708398
12JL39B 179 143 363 0756915 0000083 0714429
12JL40B 255 754 987 0825572 0000049 0710017
12JL41 220 628 1023 0832347 0000016 0712572
147
Sample no. Sm Nd Sm/144Nd 143
Nd/144Nd 2r eNd(t) TDM1 TDM2
(ppm) (ppm) (Ga) (Ga)
12JL39A 647 309 01266 0512086 0000009 –334 186 176

12JL40A 719 327 01329 0512120 0000008 –334 194 176
12JL39B 603 288 01266 0512219 0000010 –075 163 155
12JL40B 66 302 01321 0512146 0000010 –274 187 172
12JL41 717 349 01242 0512105 0000007 –271 177 171
176
Sample no. Lu Hf Lu/177Hf 177
Hf/176Hf 2r eHf(t) TDM1 TDM2
(ppm) (ppm) (Ga) (Ga)
12JL39A 027 57 0006726 0282480 0000012 410 129 171

12JL40A 054 63 0012171 0282659 0000013 749 119 133
12JL39B 036 50 0010224 0282550 0000006 469 132 156
12JL40B 04 62 0009161 0282522 0000011 426 132 162
12JL41 026 79 0004673 0282455 0000014 435 125 176
The initial isotopic ratios are calculated at the time of magma crystallization (820 Ma). The Chondrite Uniform Reservoir (CHUR) val-
ues are as follows: 147Sm/144Nd ¼ 01967, 143Nd/144Nd ¼ 0512638; 176Lu/177Hf ¼ 00332, 176Hf/177Hf ¼ 0282772. eHf(t) and TDM are cal-
culated using the formulae eHf(t) ¼ [(176Hf/177Hf)S – (176Lu/177Hf)S (ekt – 1)]/[(176Hf/177Hf)CHUR – (176Lu/177Hf)CHUR (ekt – 1)] –
1) 10000, TDM1 ¼ 1/k (1 þ [(176Hf/177Hf)S – (176Hf/177Hf)DM]/[(176Lu/177Hf)S – (176Lu/177Hf)DM)], TDM2 ¼ TDM1 – (TDM1 – t)[(fCC – fS)/(fCC –
fDM)], fLu/Hf ¼ (176Lu/177Hf)S/(176Lu/177Hf)CHUR – 1, where (176Lu/177Hf)CHUR ¼ 00332, (176Hf/177Hf)CHUR ¼ 0282772; (176Lu/177Hf)DM ¼ 0
0384, (176Hf/177Hf)DM ¼ 028325, fs ¼ (176Lu/177Hf)S/(176Lu/177Hf)CHUR1; (176Lu/177Hf)CC ¼ 0015, fCC ¼ –0.55; k ¼ 1865 10–11 a–1. S,
sample; DM, depleted mantle; CC, average continental crust.
evolution. As discussed above, the biotites in some explained by the fact that metasomatic biotites can
small enclaves (<1 cm in biotite aggregates) are inevit- be readily re-equilibrated with the host melts
ably produced by the back-reaction of garnet. Although (Erdmann et al., 2009; Lavaure & Sawyer, 2011). The
garnets were not observed in the three enclaves biotites in the enclave rims display an intermediate
(12JL39B, 12JL40B and 12JL41), there is plenty of evi- composition between the metasomatic biotites and
dence demonstrating that the biotite in the enclaves is the biotites in the host granite, confirming the chem-
the product of garnet back-reaction. The main lines of ical exchange between the enclaves and host granite.
evidence are as follows. 5. The biotite in the enclaves has a similar REE pattern to
the metasomatic biotite with a progressive enrichment
1. K-feldspar is absent or scarce in the mafic enclaves. in HREE, a slight depletion in LREE [low (La/Yb)N ratio]
A similar phenomenon is observed in the biotite ag- and a significant negative Eu anomaly. The enrichment
gregates (Figs 4 and 5). This is because the back- in HREE and depletion in LREE are inherited from the
reaction of garnet to biotite has consumed K2O. garnet, a common feature in metamorphic, peritectic
2. The biotites in the mafic enclaves are fine-grained and magmatic garnets. The negative Eu anomaly is
and have a higher aspect ratio than those in the host caused by co-growth of plagioclase.
granite. This is similar not only to those nascent bio- 6. Plagioclase is a major inclusion whereas K-feldspar
tites formed on the garnet rims (Fig. 6c and d), but is absent in the biotites from the enclaves. This is
also to the neoblastic biotites in the experiments of consistent with K-feldspar consumption and plagio-
Patin~ o-Douce & Harris (1998). clase formation during the back-reaction of garnet.
3. Mineral abundances of these enclaves are consistent 7. The biotite-rich enclaves have indistinguishable zircon
with the theoretical proportion from the product of U–Pb ages and Hf–O isotope compositions and whole-
reaction (I), namely biotite (35%), quartz (30%) and rock Sr–Nd–Hf–O isotopes from those of the host bio-
plagioclase (35%). tite granites. This suggests that they are from the same
4. Some biotites in the biotite-rich enclaves exhibit source and formed from the same magma batch.
major element characteristics consistent with those 8. The relatively higher U contents in zircons from the
of the biotites associated with garnet (Fig. 7). This enclaves suggest that the zircons in the enclaves
demonstrates that they are also produced by con- would have crystallized at a relatively later stage be-
sumption of garnet. Although the other biotites in cause the U concentration increases significantly
the enclaves show similar major element features to only with the significant crystallization of U-poor
the biotites in the host granites, this can be minerals from a granitic magma.
Fig. 12. Whole-rock initial Sr–Nd isotope compositions for the Aozicun granites and enclaves from the Jiuling batholith, South
China. The initial isotopic ratios are calculated at the zircon crystallization age (t ¼ 820 Ma). Data for the Nd isotopic compositions of
metasedimentary rocks in the Shuangqiaoshan Group and its equivalents are after Li (1996), Zhang et al. (2000) and Wang et al.
(2010).
and temperature. As the granitic magma cools and crys-

tallizes, the water activity of the residual melt increases
gradually. This would induce reaction (I) to operate.
Small biotites begin to grow at the rim of the peritectic
garnets (Fig. 15b). This stage is recorded in Fig. 6. As
the crystallization of the granitic magma continues and
more water is available for the metasomatic reaction,
most of the peritectic garnets are converted into biotite
(Fig. 15c and d). This stage is recorded in Figs 4 and 5,
in which only a small quantity of garnet debris is left.
When the garnet aggregate is completely converted
into biotite (Fig. 15e), a biotite aggregate or a mafic en-
clave is produced according to the size of the original
garnet aggregate. Then some of the biotite aggregates
or the enclaves become distributed in the granite (Fig.
15f). Actually, disaggregation of the garnet cluster or
Fig. 13. Zircon d18O–eHf(t) diagram for the granite and biotite-
the biotite aggregate is inevitable during this process
rich enclave at Aozicun in the Jiuling batholith, South China. and most of these micaceous enclaves become scat-
tered in a pluton and only a few survive. After the scat-
The above arguments demonstrate that the enclaves tering of the biotites, they would grow continuously and
and the biotite aggregates represent the product of the equilibrate with the granitic melt.
back-reaction between garnet and granitic melt at differ- The above process is generally similar to that re-
ent stages (Fig. 15). If a granitic magma is derived from ported by Lavaure & Sawyer (2011) in the Wuluma
the incongruent melting of micas and other minerals, it Pluton, in which wispy schlieren are considered to be
is a mixture of the anatectic melt and peritectic and re- derived from reaction of porphyroblasts of garnet, cor-
sidual minerals. During the ascent of the magma, peri- dierite and orthoproxene with a granitic melt. However,
tectic garnets tend to accumulate because they are the garnet porphyroblasts of their study are incorpo-
denser than the melt. The difference in rheology be- rated into the granite from the country-rocks. Lavaure &
tween the peritectic garnets and the melt causes these Sawyer (2011) observed that screens of country-rocks
solids to cluster in local regions during melt migration. gradually disappeared from the margin to the center of
When the granitic magma is emplaced at a shallower the intrusion. In general, granitic melts leave their
depth, the entrained peritectic garnet aggregates be- source rocks and migrate upwards into shallow levels.
come metastable owing to the decrease of pressure In this regard, the Wuluma Pluton represents a special
Table 7: LA-MC-ICP-MS Lu–Hf and SHRIMP O isotopes in zircon from the Aozicun granites and mafic enclaves in the Jiuling batho-
lith, South China
176
Spot Yb/177Hf 176
Lu/177Hf 176
Hf/177Hf 2r Age (Ma) (176Hf/177Hf)i eHf(t) 2r TDM1 TDM2 fLu/Hf d18O 2r
12JL39A
1 0046978 0001654 0282363 0000020 824 6 5 0282337 27 04 1280 1210 –095 999 028 M
2 0036693 0001322 0282449 0000019 822 6 8 0282428 59 03 1147 1220 –096 843 014 M
3 0039696 0001375 0282413 0000020 824 6 13 0282392 47 04 1198 1218 –096 1059 014 D
4 0055887 0001961 0282302 0000019 832 6 5 0282271 06 04 1378 1220 –094 968 021 I
5 0080795 0002588 0282436 0000027 824 6 5 0282396 48 05 1205 1183 –092 866 030 M
6 0040848 0001466 0282404 0000025 820 6 5 0282381 43 05 1214 1216 –096 908 020 M
7 0051906 0001838 0282302 0000025 1431 6 14 0282252 134 05 1373 2155 –094 645 013 I
8 0029567 0001026 0282381 0000019 853 6 9 0282365 44 03 1232 1281 –097 277 018 I
9 0036441 0001275 0282359 0000021 839 6 5 0282339 32 04 1272 1251 –096 730 014 I
10 0057880 0001950 0282391 0000022 822 6 6 0282361 36 04 1249 1201 –094 823 021 M
11 0039831 0001383 0282459 0000033 849 6 5 0282437 69 06 1134 1264 –096 633 011 I
12 0031178 0001078 0282396 0000023 829 6 5 0282379 42 04 1213 1227 –097 926 018 M
13 0065110 0002148 0282455 0000026 857 6 6 0282420 65 05 1164 1253 –094 991 015 I
14 0044377 0001625 0282419 0000021 833 6 7 0282394 50 04 1198 1231 –095 944 011 I
12JL39B
1 0037951 0001391 0282411 0000031 819 6 4 0282390 46 06 1202 1216 –096 987 021 M
2 0043625 0001591 0282409 0000030 820 6 5 0282385 44 05 1211 1212 –095 954 021 M
3 0032326 0001217 0282387 0000032 830 6 7 0282369 38 06 1230 1223 –096 1009 023 M
4 0094991 0003364 0282531 0000033 828 6 3 0282479 78 06 1089 1160 –090 803 031 M
5 0048119 0001782 0282503 0000171 838 6 5 0282475 80 31 1082 1234 –095 918 017 I
6 0089835 0003383 0282590 0000039 845 6 4 0282536 103 07 1001 1197 –090 — — I
7 0105671 0003757 0282620 0000037 858 6 4 0282560 115 07 966 1204 –089 882 019 I
8 0052769 0001976 0282473 0000030 823 6 6 0282442 64 05 1132 1201 –094 966 028 M
9 0150439 0005147 0282625 0000035 863 6 4 0282542 109 06 998 1168 –084 — — I
10 0038783 0001391 0282380 0000026 846 6 5 0282358 40 05 1246 1259 –096 898 016 I
11 0040933 0001507 0282483 0000041 839 6 4 0282459 75 07 1103 1244 –095 885 024 I
M, magmatic zircon; I, inherited core; D, discordant core; fLu/Hf ¼ (176Lu/177Hf)s/(176Lu/177Hf)CHUR1.
case. In our case, the biotite-enriched enclaves are simi- could obscure the original composition of the biotite.
lar in petrography to typical mafic enclaves in most Physical exchange between the granitic magma and the
granites. This implies that our conclusion is applicable enclave or the aggregate is also possible. Therefore,
to other plutons. In other words, our study suggests most of the metasomatic biotites in the aggregates or
that the back-reaction of peritectic minerals is an effi- the enclaves have indistinguishable compositions from
cient way to form mafic enclaves. Another difference in those in the host granite. Only small quantities of them
our study is that different stages of back-reaction are all preserve the original compositions determined by the
recognized in our samples. back-reaction of peritectic garnets.
It should be noted that the compositional exchange
between the mafic enclaves and the host granite during
back-reaction is selective. The major elements that are
considerably changed are those with a significant differ-
IMPLICATIONS FOR THE ORIGIN OF S-TYPE
ence between the mafic enclaves and the host granite. GRANITES
Because quartz and plagioclase are abundant in the The results of this study show that the granites hosting
magmatic system, the SiO2 activity is buffered and the the mafic enclaves at Aozicun in the Jiuling batholith
Na2O, CaO and Al2O3 activities are roughly constant in are peraluminous with A/CNK values of 102–147 and
the bulk system. For mafic components such as MgO, d18O values of 105–118‰ (Fig. 9b and Table 4). The
FeOT, and TiO2, in contrast, these would partly transport syn-magmatic zircons from the host granite and the en-
from the enclaves to the host granite. The exchange clave also show similarly high d18O values of 84–106‰
rates of trace elements during the back-reaction are (Fig. 13 and Table 7). Such whole-rock lithochemical
even faster. This is why only part of the biotite in the en- and zircon O isotopic characteristics indicate that their
claves has inherited relatively higher HREE contents source rocks underwent chemical weathering on the
from the garnets. Owing to the high magma tempera- surface. Therefore, the host granites and their mafic en-
ture, the equilibrium isotope exchange is readily claves originated from a metasedimentary crustal
reached. Therefore, we are not able to observe any dif- source and thus belong to the S-type category. The
ference in stable and radiogenic isotope compositions negative eNd(t) values and high (87Sr/86Sr)i values are
between the enclaves and their host granite. prominent for both the host granite and enclaves (Fig.
Furthermore, as soon as the metasomatic biotites are 12 and Table 6), indicating their derivation from rework-
produced by the back-reaction, chemical exchange ing of the relatively ancient continental crust.
would occur between them and the host magma. This Nevertheless, there are positive eHf(t) values for whole-
Fig. 14. (a) Schematic activity diagram for reaction (I); (b, c) thermodynamic modeling results of mineral crystallization and water concentration in the residual melt of the host granitic
rocks and zircons from these rocks (Tables 6 and 7),
suggesting an additional contribution from relatively ju-
venile crust (Wu et al., 2006; Zheng et al., 2008).
Moreover, relict zircon cores in the host granites exhibit
highly variable eHf(t) values from 04 to 134 and d18O
values of 28–99‰ (Table 7), suggesting mixed contri-
magma at Aozicun in the Jiuling batholith using Rhyolite-MELTS model (Gualda et al., 2012; Gualda & Ghiorso, 2015) assuming initial water contents of 2 and 4 wt %.
butions from both relatively ancient and juvenile crustal
rocks experiencing the supracrustal processes at both
low and high temperature (Zheng et al., 2007).
The calculated Zr saturation temperatures (TZr) for the
host granites and enclaves range from 797 to 825 C and
from 724 to 826 C (Table 4), respectively. Relict zircons
occur in both the host granites and enclaves, indicating
the saturation of Zr in the melts. Therefore, the calcu-
lated Zr saturation temperatures represent the upper
limit of temperatures for zircon crystallization (Boehnke
et al., 2013). The zircon U–Pb dating yields concordant
ages of 822 6 4 Ma for the host granite and 823 6 3 Ma
for the mafic enclave (Fig. 11 and Table 5). It appears
that the host granite and the enclaves crystallized at the
same time. Such ages are consistent with known zircon
U–Pb ages of 813 6 5 to 828 6 8 Ma for S-type granites
from the Jiuling batholith (Li et al., 2003; Zhong et al.,
2005; Zhao et al., 2013). Taken together, the metasedi-
mentary rocks underwent partial melting to produce
granitic magmatism at 810–830 Ma in the Jiangnan
orogen.
Like most S-type granites (e.g. Scheepers &
Armstrong, 2002; Scheepers & Poujol, 2002; Chappell &
Wyborn, 2012), the granites at Aozicun in the Jiuling
batholith have relatively higher MgO þ FeOT contents
(645–779 wt%) than natural melts (Acosta-Vigil et al.,
2007; Cesare et al., 2007, 2009) or the experimental
melting products of various sediments (Holtz &
Johannes, 1991; Patin ~ o-Douce & Johnston, 1991;
Vielzeuf & Montel, 1994; Patin ~ o-Douce & Beard, 1996;
Montel & Vielzeuf, 1997; Patin ~ o-Douce & Harris, 1998;
Castro et al., 1999; Garcıa-Arias et al., 2012; Fig. 16).
Dehydration and hydration melting of crustal rocks are
the two basic mechanisms for granitic magmatism at
elevated temperatures (Weinberg & Hasalova , 2015;
Gao et al., 2016). However, S-type granites are not the
crystallized products of pure anatectic melts. Instead
they are composed of magmatic, peritectic and residual
minerals in different proportions.
We cannot interpret mafic enclave-bearing S-type
granites as mixtures between felsic and mafic magmas.
As argued by Stevens et al. (2007), addition of a basaltic
magma to a granitic melt in a proportion sufficient
(35%) to produce the observed MgO þ FeOT values
would result in A/CNK values as low as 10 (Fig. 16b).
This is inconsistent with the observation that the mafic
enclaves in the Jiuling batholoth have similar A/CNK
values to the host granites. The mafic mineral in the
Aozicun granite is biotite (about 15–18%). Hence, the
high biotite abundances must be the dominant reason
for the whole-rock high MgO þ FeOT contents (Fig. 16a).
However, our modeling for the crystallization of min-
erals from a granitic magma discussed above suggests
Fig. 15. Schematic illustrations showing the formation of biotite-rich enclaves by back-reaction of peritectic garnet with granitic
melt. (a) The peritectic garnet aggregates entrained in the granitic melt; (b) the peritectic garnets become metastable owing to a de-
crease of pressure and an increase of water activity; (c) some metasomatic biotites formed at the rim of peritectic garnets and
some quartz and plagioclase crystallized between the peritectic garnet grains; (d) with more water available, the garnets are nearly
completely replaced by biotite. Only a small quantity of garnet debris is left in the biotite aggregates; (e) the complete replacement
of garnet by biotite; (f) biotite aggregates dispersed in S-type granite.
that the quantity of directly crystallized biotite from the 10% peritectic garnet must have been entrained into
granitic magma is less than 5% (Fig. 14b and c). our granitic magma from the anatectic source. Most of
Therefore, it is possible that most of the biotites in our these garnets would have reacted with the granitic melt
biotite-granites are converted from peritectic garnets to form the metasomatic biotite at a later stage of gran-
entrained from the anatectic source. Because most of itic magma evolution. This results in the high Mg and
the metasomatic biotites would be scattered in the Fe contents in the Aozicun S-type granites.
magma, the resultant S-type granites acquire their high The results of this study are also of great significance
MgO þ FeOT contents in this way. As such, the back- for understanding the fate of peritectic garnet in S-type
reaction of peritectic garnet to biotite identified in this granites. Stevens et al. (2007) noted the high
study provides the most likely mechanism for the in- MgO þ FeOT contents in S-type granites from the Cape
crease of Mg and Fe in the S-type granites. Granite Suite in South Africa. They considered that the
It has been shown that the MgO þ FeOT contents of compositional variation in the S-type granites is primar-
natural anatectic melts are about 19 wt % (Fig. 16). ily controlled by the proportion of peritectic garnet and
Based on this, the amount of garnet or biotite entrained ilmenite entrained in the granitic melt in view of the
by a granitic magma can be calculated following the composition of experimental melts produced by partial
mass-balance principle. About 18% biotite is required to melting of metasedimentary rocks. This can account for
mix with a natural S-type granitic melt to reach a whole- the variation of many elements (Mg, Fe, A/CNK, Mg#,
rock MgO þ FeOT content of 69 wt % for the Aozicun S- Si, K, Ca, Ti and Dy þ Ho þ Er) in these S-type granites.
type granites (Fig. 16). If all these metasomatic biotites However, the absence of peritectic garnet in the S-type
are converted from peritectic garnet, the entrained peri- granites is a vital imperfection in their model. The result
tectic garnets must be about 15% because of the higher of our study demonstrates that peritectic garnets cannot
MgO þ FeOT contents in the garnets. Of course, we can- be the sole component that controls the MgO þ FeOT
not ignore magmatic biotites, which are about 5% ac- contents of S-type granites owing to their limited abun-
cording to the thermodynamic modeling (Fig. 14b and dance. During the ascent of granitic melts to shallow
c). If we accept 5% biotite as the magmatic mineral, we crustal levels, most of the peritectic minerals should
still need 10% peritectic garnet. In this regard, at least ‘vanish’ from the melt via the mechanism of dissolution
and reprecipitation or be replaced by other minerals consideration seems consistent with the mantle input to
(e.g. biotite in our study). It is the metasomatic biotite, granite magmas as inferred from the radiogenic isotope
either in mafic enclaves or scattered in the host granite, signature of many granites (e.g. Keay et al., 1997; Kemp
that directly contributes to the high MgO and FeOT con- et al., 2007; Yang et al., 2007). However, such basaltic
tents of S-type granites. Therefore, the result of the pre- magmas should have a juvenile composition, because
sent study is an important supplement to the peritectic only decompressional melting of the asthenospheric
mineral entrainment model in granite petrogenesis.The mantle has the capacity to deliver the high heat flow for
S-type granites and mafic enclaves in the Jiuling batho- lithospheric melting, especially for batholith-sized gran-
lith were emplaced at 810–830 Ma, consistent with itic magmatism. The subcontinental lithospheric mantle
zircon U–Pb ages of 819 6 3 to 827 6 3 Ma for emplace- (SCLM) is cold and requires external heat flow for its
ment of gabbros in the ophiolite suite at Fuchuan in the partial melting (Zheng & Chen, 2016). It is also known
Jiangnan orogen (Zhang et al., 2012, 2013). The con- from experimental petrology that partial melting of
temporaneous occurrence of felsic and mafic magmatic ultramafic rocks in the mantle is not able to produce fel-
rocks is generally associated with lithospheric exten- sic melts (Wyllie, 1984; Johannes & Holtz, 1996). In this
sion in response to thinning of the thickened orogen regard, it is inappropriate to interpret the relatively
(e.g. Collins & Richards, 2008). The heat source for bi- depleted radiogenic isotope compositions in granites as
modal magmatism is often considered to be provided a material contribution to granitic magmatism (Zheng
by basaltic magmas derived from mantle melting (e.g. et al., 2015). Therefore, the heat for granitic magmatism
Huppert & Sparks, 1988; Voshage et al., 1990). This may be from high internal concentrations of heat-
producing elements such as U, Th and K in thickened
crustal zones (e.g. Clark et al., 2011). Alternatively, the
accretionary wedge and its adjacent SCLM at conver-
gent plate margins may be directly heated by the lat-
erally convective asthenospheric mantle on their base
as soon as the subducting oceanic slab rolls back
(Zheng et al., 2016). This process has the capacity to
provide the high heat flow for partial melting of metase-
dimentary rocks and metasomatic mantle domains in
accretionary orogens (Collins & Richards, 2008; Zheng
& Chen, 2016).
CONCLUSIONS
Some biotites in the S-type granite of the Jiuling batho-
lith are produced by back-reaction of peritectic garnet
with granitic melts at a late stage of their evolution. The
growth of biotite and quartz microcrystals on the garnet
rims, the garnet debris near neoblastic biotite, and bio-
tite aggregates with or without garnet are all snapshots
of the back-reaction at different stages. The biotites
formed by the consumption of garnet have high Mg#,
low K2O and TiO2, low (La/Yb)N ratio, negative Eu
anomalies and low (Gd/Yb)N ratios.
Some of the biotites in the mafic enclaves have simi-
lar major and trace element compositions to the biotites
formed by the back-reaction of garnet. The enclaves
have more biotite and less K-feldspar than the host
granite. The enclaves and their host granite have indis-
tinguishable whole-rock Sr–Nd–Hf–O isotope compos-
itions, identical zircon U–Pb ages and similar zircon Hf–
O isotope compositions. Mineral inclusions in the en-
clave biotites differ from those in the host granites; the
Fig. 16. Variation of MgO þ FeOT vs SiO2 (a) and MgO þ FeOT
vs A/CNK (b) showing the composition of the Aozicun biotite former have more plagioclase but scarce K-feldspar. All
granites, experimental quenched glasses and naturally these lines of evidence demonstrate that these biotite-
quenched anatectic melts. The arrowed line denotes the mix- rich enclaves were produced by the back-reaction of
ing of different proportions of biotite or basalt with the average
peritectic garnet aggregates with the host granitic melt.
natural melts. Also shown for comparison are data for the
Cape Town S-type granites (Stevens et al., 2007, and refer- Thermodynamic modeling suggests that the quantities
ences therein). of biotite directly crystallized from the magma are low
(<5%). Therefore, most of the biotites in the host granite Black, L. P., Kamo, S. L., Allen, C. M., Davis, D. W., Aleinikoff, J.
are also the products of garnet back-reaction. N., Valley, J. W., Mundil, R., Campbell, I. H., Korsch, R. J.,
Williams, I. S. & Foudoulis, C. (2004). Improved 206Pb/238U
microprobe geochronology by the monitoring of a trace-
element-related matrix effect; SHRIMP, ID-TIMS, ELA-ICP-
ACKNOWLEDGEMENTS
MS and oxygen isotope documentation for a series of zircon
Thanks are due to Qiang He for his assistance with field- standards. Chemical Geology 205, 115–140.
work, to Juan Wang and Yonghong Shi for their assist- Blichert-Toft, J. & Albarède, F. (1997). The Lu–Hf isotope geo-
ance with EMP analysis of biotite, to Mei Xia for her chemistry of chondrites and the evolution of the mantle–crust
assistance with mineral inclusion analysis, to Yueheng system. Earth and Planetary Science Letters 148, 243–258.
Boehnke, P., Watson, E. B., Trail, D., Harrison, T. M. & Schmitt,
Yang and Long Chen for their assistance with Sr–Nd–Hf
A. K. (2013). Zircon saturation re-revisited. Chemical
isotope analyses, to Dunyi Liu, Ning Li and Wancai Li Geology 351, 324–334.
for their assistance with SIMS zircon O isotope and U– Castro, A., Patin ~ o-Douce, A. E., Corretgé, L. G., Jesu s, D., El-
Pb dating, and to Zhaochu Hu for his assistance with zir- Biad, M. & El-Hmidi, H. (1999). Origin of peraluminous gran-
con Lu–Hf isotope analyses. We are grateful to G. ites and granodiorites, Iberian massif, Spain: an experimen-
Bergantz, A. Patin~ o-Douce and A. Acosta-Vigil for their tal test of granite petrogenesis. Contributions to Mineralogy
comments, and to J. Beard for his editorial handling. and Petrology 135, 255–276.
Cesare, B., Maineri, C., Toaldo, A. B., Pedron, D. & Vigil, A. A.
(2007). Immiscibility between carbonic fluids and granitic
melts during crustal anatexis: a fluid and melt inclusion
FUNDING
study in the enclaves of the Neogene Volcanic Province of
This study was supported by funds from the National SE Spain. Chemical Geology 237, 433–449.
Natural Science Foundation of China (41222018, Cesare, B., Ferrero, S., Salvioli-Mariani, E., Pedron, D. & Cavallo,
41572168). A. (2009). ‘Nanogranite’ and glassy inclusions: The anatectic
melt in migmatites and granulites. Geology 37, 627–630.
Chappell, B. & White, A. (1991). Restite enclaves and the restite
SUPPLEMENTARY DATA model. In: Didier, J. & Barbarin, B. (eds) Enclaves and
Granite Petrology. Developments in Petrology, Vol. 13.
Supplementary data for this paper are available at Elsevier, Amsterdam, pp. 375.
Journal of Petrology online. Chappell, B. W. & White, A. J. R. (2001). Two contrasting granite
types: 25 years later. Australian Journal of Earth Sciences
48, 489–499.
REFERENCES Chappell, B. W. & Wyborn, D. (2012). Origin of enclaves in S-
type granites of the Lachlan Fold Belt. Lithos 154, 235–247.
Acosta-Vigil, A., Cesare, B., London, D. & Morgan, G. B. (2007).
Microstructures and composition of melt inclusions in a Chappell, B., White, A. & Wyborn, D. (1987). The importance of
crustal anatectic environment, represented by metapelitic residual source material (restite) in granite petrogenesis.
enclaves within El Hoyazo dacites, SE Spain. Chemical Journal of Petrology 28, 1111–1138.
Geology 237, 450–465. Chen, B., Chen, Z. C. & Jahn, B. M. (2009). Origin of mafic en-
Alves, A., Pereira, G. d. S., Janasi, V. d. A., Higgins, M., Polo, L. claves from the Taihang Mesozoic orogen, north China cra-
A., Juriaans, O. S. & Ribeiro, B. V. (2015). The origin of felsic ton. Lithos 110, 343–358.
microgranitoid enclaves: Insights from plagioclase crystal Chen, S., Niu, Y., Sun, W., Zhang, Y., Li, J., Guo, P. & Sun, P.
size distributions and thermodynamic models. Lithos 239, (2015). On the origin of mafic magmatic enclaves (MMEs) in
33–44. syn-collisional granitoids: evidence from the Baojishan plu-
Anderson, J. L., Barth, A. P., Wooden, J. L. & Mazdab, F. (2008). ton in the North Qilian Orogen, China. Mineralogy and
Thermometers and thermobarometers in granitic systems. Petrology 109, 577–596.
In: Putirka, K. D. & Tepley, F. J., III (eds) Minerals, Inclusions Chen, Y., Price, R. & White, A. (1989). Inclusions in three S-type
and Volcanic Processes. Mineralogical Society of America granites from southeastern Australia. Journal of Petrology
and Geochemical Society, Reviews in Mineralogy and 30, 1181–1218.
Geochemistry 69, 121–142. Clark, C., Fitzsimons, I. C. W., Healy, D. & Harley, S. L. (2011).
Barbarin, B. (2005). Mafic magmatic enclaves and mafic rocks How does the continental crust get really hot?. Elements 7,
associated with some granitoids of the central Sierra 235–240.
Nevada batholith, California: nature, origin, and relations Clemens, J. D. (1984). Water contents of intermediate to silicic
with the hosts. Lithos 80, 155–177. magmas. Lithos 17, 273–287.
Barbarin, B. & Didier, J. (1991). Enclaves of the Mesozoic calc- Clemens, J. D. & Stevens, G. (2012). What controls chemical
alkaline granitoids of the Sierra Nevada Batholith, variation in granitic magmas?. Lithos 134, 317–329.
California. In: Didier, J. & Barbarin, B. (eds) Enclaves and Clemens, J. D. & Wall, V. J. (1984). Origin and evolution of a per-
Granite Petrology. Developments in Petrology, 13. Elsevier, aluminous silicic ignimbrite suite: the Violet Town Volcanics.
pp. 135–153. Contributions to Mineralogy and Petrology 88, 354–371.
Beard, J. S., Ragland, P. C. & Rushmer, T. (2004). Hydration Collins, W. J. & Richards, S. W. (2008). Geodynamic signifi-
crystallization reactions between anhydrous minerals and cance of S-type granites in circum-Pacific orogens. Geology
hydrous melt to yield amphibole and biotite in igneous 36, 559–562.
rocks: description and implications. Journal of Geology 112, Collins, W., Richards, S., Healy, B. & Ellison, P. (2000). Origin of
617–621. heterogeneous mafic enclaves by two-stage hybridisation in
Beard, J. S., Ragland, P. C. & Crawford, M. L. (2005). Reactive magma conduits (dykes) below and in granitic magma
bulk assimilation: A model for crust–mantle mixing in silicic chambers. Transaction of the Royal Society of Edinburgh:
magmas. Geology 33, 681–684. Earth Sciences 91, 27–45.
Collins, W. J., Wiebe, R. A., Healy, B. & Richards, S. W. (2006). teaching of magma properties and evolution.
Replenishment, crystal accumulation and floor aggradation Geochemistry, Geophysics, Geosystems 16, 315–324.
in the megacrystic Kameruka Suite, Australia. Journal of Gualda, G. A. R., Ghiorso, M. S., Lemons, R. V. & Carley, T. L.
Petrology 47, 2073–2104. (2012). Rhyolite-MELTS: a modified calibration of MELTS
Compston, W., Williams, I. S., Kirschvink, J. L., Zhang, Z. & optimized for silica-rich, fluid-bearing magmatic systems.
Guogan, M. A. (1992). Zircon U–Pb ages for the Early Journal of Petrology 53, 875–890.
Cambrian time-scale. Journal of the Geological Society, Holtz, F. & Johannes, W. (1991). Genesis of peraluminous gran-
London 149, 171–184. ites I. Experimental investigation of melt compositions at 3
Didier, J. (1973). Granites and their Enclaves: the Bearing of and 5 kbar and various H2O activities. Journal of Petrology
Enclaves on the Origin of Granites. Developments in 32, 935–958.
Petrology, 3. Elsevier. Hoskin, P. W. O. & Schaltegger, U. (2003). The composition of
Didier, J. (1987). Contribution of enclave studies to the under- zircon and igneous and metamorphic petrogenesis. In:
standing of origin and evolution of granitic magmas. Hanchar, J. M. & Hoskin, P. W. O. (eds) Zircon. Mineralogical
Geologische Rundschau 76, 41–50. Society of America and Geochemical Society, Reviews in
Dodge, F. C. W. & Kistler, R. W. (1990). Some additional obser- Mineralogy and Geochemistry 53, 27–62.
vations on inclusions in the granitic rocks of the Sierra Huang, H., Niu, Y., Nowell, G., Zhao, Z., Yu, X., Zhu, D. C., Mo,
Nevada. Journal of Geophysical Research 95, 17841. X. & Ding, S. (2014). Geochemical constraints on the petro-
Donaire, T., Pascual, E., Pin, C. & Duthou, J.-L. (2005). genesis of granitoids in the East Kunlun Orogenic belt,
Microgranular enclaves as evidence of rapid cooling in gran- northern Tibetan Plateau: Implications for continental crust
itoid rocks: the case of the Los Pedroches granodiorite, growth through syn-collisional felsic magmatism. Chemical
Iberian Massif, Spain. Contributions to Mineralogy and Geology 370, 1–18.
Petrology 149, 247–265. Huppert, H. E. & Sparks, S. J. (1988). The generation of granitic
Dorais, M.J., & Tubrett, M. (2012). Detecting peritectic garnet in magmas by intrusion of basalt into continental crust.
the peraluminous Cardigan pluton, New Hampshire. Journal of Petrology 29, 599–624.
Journal of Petrology 53, 299–324. Ickert, R., Hiess, J., Williams, I., Holden, P., Ireland, T., Lanc, P.,
Eggler, D. H. (1972). Amphibole stability in H2O-undersaturated Schram, N., Foster, J. & Clement, S. (2008). Determining
calc-alkaline melts. Earth and Planetary Science Letters 15, high precision, in situ, oxygen isotope ratios with a SHRIMP
28–34. II: Analyses of MPI-DING silicate-glass reference materials
Elburg, M. A. & Nicholls, I. A. (1995). Origin of microgranitoid and zircon from contrasting granites. Chemical Geology
enclaves in the S-type Wilson’s Promontory Batholith, 257, 114–128.
Victoria: Evidence for magma mingling. Australian Journal Irvine, T. (1982). Terminology for layered intrusions. Journal of
of Earth Sciences 42, 423–435. Petrology 23, 127–162.
Ellis, D. J. & Obata, M. (1992). Migmatite and melt segregation Johannes, W. & Holtz, F. (1996). Petrogenesis and Experimental
at Cooma, New South Wales. Transactions of the Royal Petrology of Granitic Rocks. Springer, 335 pp.
Society of Edinburgh: Earth Sciences 83, 95–106. Keay, S., Collins, W. J. & McCulloch, M. T. (1997). A three-
Erdmann, S., Jamieson, R. A. & MacDonald, M. A. (2009). component Sr–Nd isotopic mixing model for granitoid gen-
Evaluating the origin of garnet, cordierite, and biotite in esis, Lachlan fold belt, eastern Australia. Geology 25,
granitic rocks: a case study from the South Mountain 307–310.
Batholith, Nova Scotia. Journal of Petrology 50, 1477–1503. Kemp, A. I., Hawkesworth, C. J., Foster, G. L., Paterson, B. A.,
Farner, M. J., Lee, C. T. A. & Putirka, K. D. (2014). Mafic–felsic Woodhead, J. D., Hergt, J. M., Gray, C. M. & Whitehouse, M.
magma mixing limited by reactive processes: A case study J. (2007). Magmatic and crustal differentiation history of
of biotite-rich rinds on mafic enclaves. Earth and Planetary granitic rocks from Hf–O isotopes in zircon. Science 315,
Science Letters 393, 49–59. 980–983.
Gao, P., Zheng, Y.-F. & Zhao, Z.-F. (2016). Experimental melts Kriegsman, L. M. & Hensen, B. J. (1998). Back reaction between
from crustal rocks: A lithochemical constraint on granite restite and melt: implications for geothermobarometry and
petrogenesis. Lithos 266–267, 133–157. pressure–temperature paths. Geology 26, 1111–1114.
Garcıa-Arias, M., Corretgé, L. G. & Castro, A. (2012). Trace elem- Lavaure, S. & Sawyer, E. W. (2011). Source of biotite in the
ent behavior during partial melting of Iberian orthogneisses: Wuluma Pluton: Replacement of ferromagnesian phases
An experimental study. Chemical Geology 292, 1–17. and disaggregation of enclaves and schlieren. Lithos 125,
Glazner, A. F. (2014). Magmatic life at low Reynolds number. 757–780.
Geology 42, 935–938. Li, X. H. (1996). Sm–Nd isotopic systematics of Sibao Group
Gong, B., Zheng, Y. F. & Chen, R. X. (2007). TC/EA-MS online de- from the southern margin of Yangtze block: implications for
termination of hydrogen isotope composition and water the crustal evolution. Scientia Geologica Sinica 31, 218–228.
concentration in eclogitic garnet. Physics and Chemistry of (in Chinese with English abstract).
Minerals 34, 687–698. Li, X. H., Li, Z. X., Ge, W. C., Zhou, H. W., Li, W. X., Liu, Y. &
Griffin, W. L., Pearson, N. J., Belousova, E., Jackson, S. E., Wingate, M. T. D. (2003). Neoproterozoic granitoids in South
Achterbergh, E. V., O’Reilly, S. Y. & Shee, S. R. (2000). The China: crustal melting above a mantle plume at ca. 825 Ma?
Hf isotope composition of cratonic mantle: LAM-MC-ICPMS Precambrian Research 122, 45–83.
analysis of zircon megacrysts in kimberlites. Geochimica et Li, X. H., Li, W. X., Li, Z. X., Lo, C. H., Wang, J., Ye, M. F. & Yang,
Cosmochimica Acta 64, 133–147. Y. H. (2009). Amalgamation between the Yangtze and
Griffin, W. L., Wang, X., Jackson, S. E., Pearson, N. J., O’Reilly, Cathaysia Blocks in South China: constraints from SHRIMP
S. Y., Xu, X. & Zhou, X. (2002). Zircon chemistry and magma U–Pb zircon ages, geochemistry and Nd–Hf isotopes of the
mixing, SE China: In-situ analysis of Hf isotopes, Tonglu and Shuangxiwu volcanic rocks. Precambrian Research 174,
Pingtan igneous complexes. Lithos 61, 237–269. 117–128.
Gualda, G. A. R. & Ghiorso, M. S. (2015). MELTS_Excel: A Liu, Y., Hu, Z., Gao, S., Günther, D., Xu, J., Gao, C. & Chen, H.
Microsoft Excel-based MELTS interface for research and (2008). In situ analysis of major and trace elements of
anhydrous minerals by LA-ICP-MS without applying an in- Scherer, E., Münker, C. & Mezger, K. (2001). Calibration of the
ternal standard. Chemical Geology 257, 34–43. lutetium–hafnium clock. Science 293, 683–687.
Liu, Y., Gao, S., Hu, Z., Gao, C., Zong, K. & Wang, D. (2010). Stacey, J. S. & Kramers, J. D. (1975). Approximation of terres-
Continental and oceanic crust recycling-induced melt– trial lead isotope evolution by a two-stage model. Earth and
peridotite interactions in the Trans-North China Orogen: U– Planetary Science Letters 26, 207–221.
Pb dating, Hf isotopes and trace elements in zircons from Stevens, G., Villaros, A. & Moyen, J. F. (2007). Selective peritec-
mantle xenoliths. Journal of Petrology 51, 537–571. tic garnet entrainment as the origin of geochemical diversity
Ludwig, K. R. (2003). User’s Manual for Isoplot 3.0: A in S-type granites. Geology 35, 9–12.
Geochronological Toolkit for Microsoft Excel. Berkeley Sun, S. S. & McDonough, W. F. (1989). Chemical and isotopic
Geochronology Center, Special Publication 4. systematics of oceanic basalts: implications for mantle com-
Maas, R., Nicholls, I. A. & Legg, C. (1997). Igneous and meta- position and processes. In: Saunders, A. D. & Norry, M. J.
morphic enclaves in the S-type Deddick granodiorite, (eds) Magmatism in the Ocean Basins. Geological Society,
Lachlan Fold Belt, SE Australia: Petrographic, geochemical London, Special Publications 42, 313–345.
and Nd–Sr isotopic evidence for crustal melting and magma Tindle, A. G. & Pearce, J. A. (1983). Assimilation and partial
mixing. Journal of Petrology 38, 815–841. melting of continental crust: evidence from the mineralogy
Middlemost, E. A. (1994). Naming materials in the magma/igne- and geochemistry of autoliths and xenoliths. Lithos 16,
ous rock system. Earth-Science Reviews 37, 215–224. 185–202.
Moita, P., Santos, J. F., Pereira, M. F., Costa, M. M. & Corfu, F. Valley, J. W., Kitchen, N., Kohn, M. J., Niendorf, C. R. &
(2015). The quartz-dioritic Hospitais intrusion (SW Iberian Spicuzza, M. J. (1995). UWG-2, a garnet standard for oxygen
Massif) and its mafic microgranular enclaves—Evidence for isotope ratios: Strategies for high precision accuracy with
mineral clustering. Lithos 224, 78–100. laser heating. Geochimica et Cosmochimica Acta 59,
Montel, J. M. & Vielzeuf, D. (1997). Partial melting of metagrey- 5223–5231.
wackes, Part II. Compositions of minerals and melts. Vernon, R. (1984). Microgranitoid enclaves in granites—
Contributions to Mineralogy and Petrology 128, 176–196. globules of hybrid magma quenched in a plutonic environ-
Morel, M. L. A., Nebel, O., Nebel-Jacobsen, Y. J., Miller, J. S. & ment. Nature 309, 438–439.
Vroon, P. Z. (2008). Hafnium isotope characterization of the Vernon, R. H. (2007). Problems in identifying restite in S-type
GJ-1 zircon reference material by solution and laser- granites of southeastern Australia, with speculations on
ablation MC-ICPMS. Chemical Geology 255, 231–235. sources of magma and enclaves. Canadian Mineralogist 45,
O’Neil, J. R. & Chappell, B. W. (1977). Oxygen and hydrogen 147–178.
isotope relations in the Berridale batholith. Journal of the Vernon, R. H. (2014). Microstructures of microgranitoid en-
Geological Society, London 133, 559–571. claves and the origin of S-type granitoids. Australian
Pascual, E., Donaire, T. & Pin, C. (2008). The significance of Journal of Earth Sciences 61, 227–239.
microgranular enclaves in assessing the magmatic evolu- Vielzeuf, D. & Montel, J. M. (1994). Partial melting of metagrey-
tion of a high-level composite batholith: A case on the Los wackes. Part I. Fluid-absent experiments and phase relation-
Pedroches Batholith, Iberian Massif, Spain. Geochemical ships. Contributions to Mineralogy and Petrology 117,
Journal 42, 177–198. 375–393.
~ o-Douce, A. E. & Beard, J. S. (1996). Effects of P, f(O2) and
Patin Voshage, H., Hofmann, A. W., Mazzucchelli, M., Rivalenti, G.,
Mg/Fe ratio on dehydration melting of model metagrey- Sinigoi, S., Raczek, I. & Demarchi, G. (1990). Isotopic evi-
wackes. Journal of Petrology 37, 999–1024. dence from the Ivrea Zone for a hybrid lower crust formed
~ o-Douce, A. E. & Harris, N. (1998). Experimental con-
Patin by magmatic underplating. Nature 347, 731–736.
straints on Himalayan anatexis. Journal of Petrology 39, Wang, H. Z., Chen, P. R., Sun, L. Q., Ling, H. F., Zhao, Y. D. &
689–710. Lan, H. F. (2015). Magma mixing and crust–mantle inter-
~ o-Douce, A. E. & Johnston, A. D. (1991). Phase equilibria and
Patin action in Southeast China during the Early Cretaceous:
melt productivity in the pelitic system: implications for the ori- Evidence from the Furongshan granite porphyry and mafic
gin of peraluminous granitoids and aluminous granulites. microgranular enclaves. Journal of Asian Earth Sciences
Contributions to Mineralogy and Petrology 107, 202–218. 111, 72–87.
Reid, J. B., Evans, O. C. & Fates, D. G. (1983). Magma mixing in Wang, J. & Li, Z. (2003). History of Neoproterozoic rift basins in
granitic rocks of the central Sierra Nevada, California. Earth South China: implications for Rodinia break-up.
and Planetary Science Letters 66, 243–261. Precambrian Research 122, 141–158.
Scaillet, B., Whittington, A., Martel, C., Pichavant, M. & Holtz, F. Wang, W., Wang, F., Chen, F., Zhu, X., Xiao, P. & Siebel, W.
(2000). Phase equilibrium constraints on the viscosity of sili- (2010). Detrital zircon ages and Hf–Nd isotopic composition
cic magmas II: implications for mafic–silicic mixing proc- of Neoproterozoic sedimentary rocks in the Yangtze Block:
esses. Transactions of the Royal Society of Edinburgh: Earth constraints on the deposition age and provenance. Journal
Sciences 91, 61–72. of Geology 118, 79–94.
Scheepers, R. & Armstrong, R. (2002). New U–Pb SHRIMP zir- Wang, X. L., Zhou, J. C., Griffin, W. L., Wang, R. C., Qiu, J. S.,
con ages of the Cape Granite Suite: implications for the O’Reilly, S. Y., Xu, X., Liu, X. M. & Zhang, G. L. (2007).
magmatic evolution of the Saldania Belt. South African Detrital zircon geochronology of Precambrian basement se-
Journal of Geology 105, 241–256. quences in the Jiangnan orogen: Dating the assembly of the
Scheepers, R. & Poujol, M. (2002). U–Pb zircon age of Cape Yangtze and Cathaysia Blocks. Precambrian Research 159,
Granite Suite ignimbrites: characteristics of the last phases 117–131.
of the Saldanian magmatism. South African Journal of Wang, X. L., Zhou, J. C., Wan, Y. S., Kitajima, K., Wang, D.,
Geology 105, 163–178. Bonamici, C., Qiu, J. S. & Sun, T. (2013). Magmatic
Scheibner, B., Wörner, G., Civetta, L., Stosch, H. G., Simon, K. & evolution and crustal recycling for Neoproterozoic
Kronz, A. (2007). Rare earth element fractionation in mag- strongly peraluminous granitoids from southern China: Hf
matic Ca-rich garnets. Contributions to Mineralogy and and O isotopes in zircon. Earth and Planetary Science
Petrology 154, 55–74. Letters 366, 71–82.
Ward, R., Stevens, G. & Kisters, A. (2008). Fluid and deform- Yao, J., Shu, L., Santosh, M. & Zhao, G. (2014). Neoproterozoic
ation induced partial melting and melt volumes in low- arc-related mafic–ultramafic rocks and syn-collision granite
temperature granulite-facies metasediments, Damara Belt, from the western segment of the Jiangnan Orogen, South
Namibia. Lithos 105, 253–271. China: constraints on the Neoproterozoic assembly of the
Waters, D. J. (2001). The significance of prograde and retro- Yangtze and Cathaysia Blocks. Precambrian Research 243,
grade quartz-bearing intergrowth microstructures in par- 39–62.
tially melted granulite-facies rocks. Lithos 56, 97–110. Zhang, C. L., Santosh, M., Zou, H. B., Li, H. K. & Huang, W. C.
Watkins, J. M., Clemens, J. D. & Treloar, P. J. (2007). Archaean (2013). The Fuchuan ophiolite in Jiangnan Orogen: geo-
TTGs as sources of younger granitic magmas: melting of chemistry, zircon U–Pb geochronology, Hf isotope and im-
sodic metatonalites at 06–12 GPa. Contributions to plications for the Neoproterozoic assembly of South China.
Mineralogy and Petrology 154, 91–110. Lithos 179, 263–274.
Weinberg, R. F. & Hasalova , P. (2015). Water-fluxed melting of Zhang, H. Y., Sun, D. Z., Zhu, B. Q. & Tu, X. L. (2000). Pb, Nd iso-
the continental crust: a review. Lithos 212–215, 158–188. topic study of Proterozoic metamorphic sediments in North
Weis, D., Kieffer, B., Maerschalk, C., Barling, J., de Jong, J., Jiangxi and its tectonic significance. Regional Geology of
Williams, G. A., Hanano, D., Pretorius, W., Mattielli, N., China 19, 66–71. (in Chinese with English abstract).
Scoates, J. S., Goolaerts, A., Friedman, R. M. & Mahoney, J. Zhang, S.-B. & Zheng, Y.-F. (2013). Formation and evolution of
B. (2006). High-precision isotopic characterization of USGS Precambrian continental lithosphere in South China.
reference materials by TIMS and MC-ICP-MS. Gondwana Research 23, 1241–1260.
Geochemistry, Geophysics, Geosystems 7, Q08006, doi: Zhang, S.-B., Wu, R.-X. & Zheng, Y.-F. (2012). Neoproterozoic
10.1029/2006GC001283. continental accretion in South China: geochemical evidence
Weis, D., Kieffer, B., Hanano, D., Nobre Silva, I., Barling, J., from the Fuchuan ophiolite in the Jiangnan orogen.
Pretorius, W., Maerschalk, C. & Mattielli, N. (2007). Hf iso- Precambrian Research 220–221, 45–64.
tope compositions of U.S. Geological Survey reference ma- Zhao, G. & Cawood, P. A. (2012). Precambrian geology of
terials. Geochemistry, Geophysics, Geosystems 8, Q06006, China. Precambrian Research 222, 13–54.
doi:10.1029/2006GC001473. Zhao, J. H., Zhou, M. F. & Zheng, J. P. (2013). Constraints from
White, A. J., Chappell, B. W. & Wyborn, D. (1999). Application of zircon U–Pb ages, O and Hf isotopic compositions on the ori-
the restite model to the Deddick granodiorite and its gin of Neoproterozoic peraluminous granitoids from the
enclaves—a reinterpretation of the observations and data of Jiangnan Fold Belt, South China. Contributions to
Maas et al. (1997). Journal of Petrology 40, 413–421. Mineralogy and Petrology 166, 1505–1519.
Whitney, D. L. & Evans, B. W. (2010). Abbreviations for names Zheng, Y.-F. & Chen, Y.-X. (2016). Continental versus oceanic
of rock-forming minerals. American Mineralogist 95, subduction zones. National Science Review 3, doi:10.1093/
185–187. nsr/nww049.
Woodhead, J. D. & Hergt, J. M. (2005). A preliminary appraisal Zheng, Y.-F., Wu, Y.-B., Chen, F. K., Gong, B., Li, L. & Zhao, Z.-F.
of seven natural zircon reference materials for in situ Hf iso- (2004). Zircon U–Pb and oxygen isotope evidence for a
tope determination. Geostandards and Geoanalytical large-scale 18O depletion event in igneous rocks during the
Research 29, 183–195. Neoproterozoic. Geochimica et Cosmochimica Acta 68,
Wu, R.-X., Zheng, Y.-F., Wu, Y.-B., Zhao, Z.-F., Zhang, S.-B., Liu, 4145–4165.
X. & Wu, F.-Y. (2006). Reworking of juvenile crust: Element Zheng, Y.-F., Zhang, S.-B., Zhao, Z.-F., Wu, Y.-B., Li, X. H., Li, Z.
and isotope evidence from Neoproterozoic granodiorite in X. & Wu, F.-Y. (2007). Contrasting zircon Hf and O isotopes
South China. Precambrian Research 146, 179–212. in the two episodes of Neoproterozoic granitoids in South
Wyborn, D. (2013). Reply—Origin of enclaves in S-type granites China: Implications for growth and reworking of continental
of the Lachlan Fold Belt. Lithos 175–176, 353–354. crust. Lithos 96, 127–150.
Wyllie, P. J. (1984). Constraints imposed by experimental pet- Zheng, Y.-F., Wu, R.-X., Wu, Y.-B., Zhang, S.-B., Yuan, H. L. &
rology on possible and impossible magma sources and Wu, F.-Y. (2008). Rift melting of juvenile arc-derived crust:
products. Philosophic Transactions of the Royal Society of geochemical evidence from Neoproterozoic volcanic and
London, Series A 310, 439–456. granitic rocks in the Jiangnan Orogen, South China.
Yang, H., Ge, W. C., Zhao, G. C., Dong, Y., Xu, W. L., Ji, Z. & Yu, Precambrian Research 163, 351–383.
J. J. (2015). Late Triassic intrusive complex in the Jidong re- Zheng, Y.-F., Xiao, W. J. & Zhao, G. C. (2013).
gion, Jiamusi–Khanka Block, NE China: Geochemistry, zir- Introduction to tectonics of China. Gondwana Research 23,
con U–Pb ages, Lu–Hf isotopes, and implications for magma 1189–1206.
mingling and mixing. Lithos 224–225, 143–159. Zheng, Y. F., Chen, Y. X., Dai, L. Q. & Zhao, Z. F. (2015).
Yang, J. H., Wu, F. Y., Wilde, S. A., Xie, L. W., Yang, Y. H. & Liu, Developing plate tectonics theory from oceanic subduction
X. M. (2007). Tracing magma mixing in granite genesis: in zones to collisional orogens. Science China: Earth Sciences
situ U–Pb dating and Hf-isotope analysis of zircons. 58, 1045–1069.
Contributions to Mineralogy and Petrology 153, 177–190. Zheng, Y.-F., Chen, R.-X., Xu, Z. & Zhang, S.-B. (2016). The
Yang, Y. H., Zhang, H. F., Chu, Z. Y., Xie, L. W. & Wu, F. Y. transport of water in subduction zones. Science China: Earth
(2010). Combined chemical separation of Lu, Hf, Rb, Sr, Sm Sciences 59, 651–682.
and Nd from a single rock digest and precise and accurate Zhong, Y. F., Ma, C. Q., She, Z. B., Lin, G. C., Xu, H., Wang, R. J.,
isotope determinations of Lu–Hf, Rb–Sr and Sm–Nd isotope Yang, K. G. & Liu, Q. (2005). SHRIMP U–Pb zircon geochronology
systems using multi-collector ICP-MS and TIMS. of the Jiuling granitic complex batholith in Jiangxi Province.
International Journal of Mass Spectrometry 290, 120–126. Earth Science 30, 685–691. (in Chinese with English abstract).
View publication stats

JP WeiRong

Uploaded by

Copyright:

Available Formats

JP WeiRong

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

JP WeiRong

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author proﬁles for this publication at: https://www.researchgate.

Back-reaction of Peritectic Garnet as an Explanation for the Origin of Maﬁc

Article in Journal of Petrology · May 2017

Wei Rong Shao-Bing Zhang

SEE PROFILE SEE PROFILE

Crust-mantle interactions at subduction zones View project

zircon and evolution of continental crust View project

The user has requested enhancement of the downloaded ﬁle.

Back-reaction of Peritectic Garnet as an

INTRODUCTION Pascual et al., 2008; Farner et al., 2014). Mafic enclaves

principles have been described by Ickert et al. (2008). It PETROGRAPHY

Granites Mafic enclaves

Sample: 12JL39A 12JL40A 12JL39B 12JL40B 12JL41

Total inclusion numbers/ 72/7 39/3 47/10 25/12 28/8

DISCUSSION cordierite 6 sillimanite. These coronas are interpreted as

Mechanism for the back-reaction of peritectic

Bt in host granite Bt in Bt in Bt in contact

enclave rim enclave with garnet

La 0054 0417 0057 0045 0117 0011 0028

Major elements (wt %)

calculated by Fe2þ–Fe3þ–Mg2þin our biotites

Spot Element (ppm) Th/U Isotope ratio Age (Ma)

M, magmatic zircon; I, inherited core; D, discordant core.

12JL39A 195 1165 486 0768618 0000091 0711741

12JL39A 647 309 01266 0512086 0000009 –334 186 176

12JL39A 027 57 0006726 0282480 0000012 410 129 171

and temperature. As the granitic magma cools and crys-

M, magmatic zircon; I, inherited core; D, discordant core; fLu/Hf ¼ (176Lu/177Hf)s/(176Lu/177Hf)CHUR1.

View publication stats

You might also like

M, magmatic zircon; I, inherited core; D, discordant core; fLu/Hf ¼ (176Lu/177Hf)s/(176Lu/177Hf)CHUR1.