Petrogenesis of the Nanling Mountains granites from South China: Constraints from systematic apatite geochemistry and whole-rock geochemical and Sr–Nd isotope compositions

Petrogenesis of the Nanling Mountains granites from South

China: Constraints from systematic apatite geochemistry and

whole-rock geochemical and Sr–Nd isotope compositions

Pei-Shan Hsieh

, Cheng-Hong Chen

, Huai-Jen Yang

, Chi-Yu Lee

a

Department of Geosciences, National Taiwan University, No. 1, Roosevelt Road Section 4, Taipei 106, Taiwan b

Department of Earth Sciences, National Cheng Kung University, Tainan, Taiwan Received 1 August 2007; received in revised form 12 February 2008; accepted 18 February 2008

Abstract

The widespread Mesozoic granitoids in South China (135,300 km2_{) were emplaced in three main periods: Triassic (16% of the total} surface area of Mesozoic granitoids), Jurassic (47%), and Cretaceous (37%). Though much study has been conducted on the most abun-dant Jurassic Nanling Mountains (NLM) granites, their rock affinities relative to the Triassic Darongshan (DRS) and Cretaceous Fuz-hou–Zhangzhou Complex (FZC) granites which are typical S- and I-type, respectively, and the issue of their petrogenetic evolution is still the subject of much debate. In this study, we discuss the petrogenesis of NLM granites using apatite geochemistry combined with whole-rock geochemical and Sr–Nd isotope compositions. Sixteen apatite samples from six granite batholiths, one gabbro, and three syenite bodies in the NLM area were analyzed for their major and trace element abundances and compared with those collected from DRS (n = 7) and FZC (n = 6) granites. The apatite geochemistry reveals that Na, Si, S, Mn, Sr, U, Th concentrations and REE distribution patterns for apatites from DRS and FZC granites basically are similar to the S and I granite types of the Lachlan Fold Belt (Australia), whereas those from NLM granites have intermediate properties and cannot be correlated directly with these granite types. According to some indications set by the apatite geochemistry (e.g., lower U and higher Eu abundances), NLM apatites appear to have formed under oxidizing conditions. In addition, we further found that their REE distribution patterns are closely related to aluminum saturation index (ASI) and Nd isotope composition, rather than SiO2content or degree of differentiation, of the host rock. The majority of apatites from NLM granites (ASI = 0.97–1.08 and eNd(T) =8.8 to 11.6) display slightly right-inclined apatite REE patterns distinguishable from the typical S- and I-type. However, those from few granites with ASI > 1.1 and eNd(T) <11.6 have REE distribution patterns (near-flat) similar to DRS apatites whereas those from granites with ASI < 1.0 and eNd(T) >6.6 and gabbro and syenite are similar to FZC apatites (strongly right-inclined). In light of Sr and Nd isotope compositions, magmas of NLM intrusives, except gabbro and syenite, and few granites with eNd(T) >8, generally do not involve a mantle component. Instead, they fit with a melt derived largely from in situ melting or anatexis of the pre-Mesozoic (mainly Caledonian) granitic crust with subordinate pre-Yanshanian (mainly Indosinian) gra-nitic crust. We suggest that an application, using combined whole-rock ASI and eNd(T) values, is as useful as the apatite geochemistry for recognizing possible sources for the NLM granites.

Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Apatite geochemistry; ASI (aluminum saturation index); Mesozoic granitoids; Nanling Mountains; South China

1. Introduction

South China is composed of Yangtze Block in the north-west and Cathaysia Block in the southeast (Fig. 1a), and the

latter is characterized by record of repeated granitic magma-tism since the Neoproterozoic. Mesozoic igneous rocks are the most widely exposed in the Cathaysia Block (approxi-mately 50% of the total surface area) and many are enriched in Sn, W, Bi, Mo, Pb, Zn, and Cu ore deposits (GRGNP, 1989). Three main episodes of the Mesozoic magmatic events have occurred; Triassic, Jurassic, and Cretaceous, 1367-9120/$ - see front matterÓ 2008 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jseaes.2008.02.002 *

Corresponding author. Tel.: +886 2 33665872; fax: +886 2 23636095. E-mail address:chench@ntu.edu.tw(C.-H. Chen).

www.elsevier.com/locate/jaes

Available online at www.sciencedirect.com

Fig. 1. (a) Simplified geological map of the Mesozoic granitoids in S. China, with sample localities of (b) Indosinian (Triassic) Darongshan (DRS) granitic suites (modified afterDeng et al., 2004), (c) Early Yanshanian (Jurassic) Nanling Mountains (NLM) batholiths and plutons, and (d) Late Yanshanian (Cretaceous) Fuzhou–Zhangzhou Complex (FZC) granitic plutons (Fuzhou, Dayang, and Liangjiang plutons of the Fuzhou Complex and Zudi, Yanqian, and Changtai plutons of the Zhangzhou Complex) (modified afterChen et al., 2000). Major faults in S. China include r Jiangshan-Shaoxing, Dongxiang-Pingxiang, and Xupu-Sanjiang deep fault zone; s Shi-Hang zone; t Zhenghe-Dapu deep fault zone; u Changle-Nanao deep fault zone; v Bobai-Cenxi deep fault zone; and w Lingshan-Tangxian deep fault zone.

and they are conventionally referred as the Indosinian, Early Yanshanian (EY), and Late Yanshanian (LY) stages of orogeny (Zhou and Li, 2000; Li et al., 2004). Lithologi-cally, granites and rhyolites (>95%) predominate over basic and intermediate rocks, and volumetrically, Indosinian granitoids crop out in a total area of20,900 km2(16% of the total surface area of Mesozoic granitoids), EY grani-toids64,100 km2(47%), and LY granitoids50,300 km2 (37%) (Sun, 2006). For more than half a century, the origin and petrogenesis of extensive EY granitoids and the process of related mineralization have stimulated the interest of geologists. Moreover, whether the EY magmatism caused the extensive crustal growth or whether the crustal evolution is related to the geodynamic environment of the Mesozoic South China continent is still hotly debated (Gilder et al., 1996; Chen and Jahn, 1998; Zhou and Li, 2000; Pirajno and Bagas, 2002; Chen and Grapes, 2003; Li et al., 2003; Xu et al., 2005; Li and Li, 2007). To better constrain mantle input to the crust, an important step is to understand the petrogenesis of EY granitoids.

SinceChappell and White (1974)first proposed the con-cept of two contrasting granite types (S and I) based on the distinction of chemical and mineralogical compositions, granites have commonly been related to derivation by par-tial melting of two different types of source material-sedi-mentary and igneous, and the derived granites can inherit the geochemical and isotopic characteristics from their source rocks. S-type granites are commonly thought to rep-resent melting of ‘‘reworked” continental crust in contrast to I-type granites whose sources basically have not been sub-ject to significant chemical weathering (Chappell and White, 2001). The term ‘‘granite” they used and subsequently in this study, in a broad sense, includes all plutonic rocks domi-nated by quartz and feldspar, and is analogous to ‘‘granit-oid” (Streckeisen, 1976). However, to identify such two types of rock using the criteria proposed byChappell and White (1974)is not always possible, particularly if mixed-characteristics of source regions or multiple sources with magma mingling have occurred (Clemens, 2003). Therefore, different petrogenetic discrimination schemes, such as the iron oxide series (Ishihara, 1977), relative abundances of whole-rock trace elements (Pearce et al., 1984), combining field observations and petrographical, chemical, and isoto-pic criteria (Barbarin, 1999), and chemical discrimination of apatite from granites (Zhang et al., 1985; Sha and Chap-pell, 1999; Belousova et al., 2002), have been proposed.

It has long been suggested that the Mesozoic S-type granites in S. China are mainly distributed in the interior of the Cathaysia Block, and I-type granites only crop out near the coast (Jahn et al., 1976, 1990; Gilder et al., 1996; Chen and Jahn, 1998; Pirajno and Bagas, 2002). However, the origin of the voluminous EY granitoids in the center of this Block is still under debate because of the complexity of these rocks. For example, the largest Fogang batholith (6000 km2

) is thought to include the S-, I-, and A-type granites based on mineralogical, whole-rock geochemical and Sr–Nd isotopic criteria (Li et al.,

2007 and references therein). Any petrogenetic model developed to account for magma generation of such a large and complicated batholith needs to address the regional control for the distribution of magmatism, rather than rely-ing on a srely-ingle batholith alone. Furthermore, the use of effective geochemical and isotopic discrimination parame-ters is critical to understanding of the overall picture of the Mesozoic granitoids in S. China. As apatites in the meta- and peraluminous rocks appear to contain many ele-ments that are sensitive to different physical conditions (Bea, 1996), apatite geochemistry is considered as an addi-tional distinguishing criterion.

Apatite, with an ideal formula of Ca10(PO4)6(F, OH, Cl)2, occurs as an accessory mineral in almost all kinds of rock. In igneous rocks, its abundance varies directly with the phosphorous content and inversely with the increasing silica content of the host rock (Frietsch and Perdahl, 1995). Apatite is also an excellent host of some trace elements such as REE, Sr, U, and Th in the natural system which are useful not only for understanding ore genesis (Treloar and Colley, 1996), but also in the field of mineral explora-tion (Belousova et al., 2002; Mordberg et al., 2006). With regard to petrogenetic applications, Sha and Chappell (1999) found that apatites can concentrate many minor and trace elements whose abundances and ratios are sensi-tive to factors controlling the fundamental differences between I- and S-type granites and suggested that the results have important implications for identifying different types of granites and potential significance for determining the provenance of sedimentary rocks.

The aims of this study are threefold. Since apatite geo-chemistry has been shown to be an effective parameter that correlates well with S- and I-type granites in the Lachlan Folded Belt (LFB), Australia (Sha and Chappell, 1999), it is used here for the first time to differentiate the well-established Mesozoic S- and I-type granites in S. China. On the basis of apatite geochemistry, many other granites in S. China, particularly those in the Nanling Mountains (NLM) area that show more complicated geochemical and isotopic characteristics than the simple S–I division, are examined. They are compared in order to reveal the conditions of magma crystallization, such as the redox state and the effect of coexisting accessory minerals. This may help to better constrain the potential for ore mineral exploration related to these granites (Belousova et al., 2002). Finally, correlations between the apatite geochemis-try and the whole-rock parameters (major and trace ele-ment abundances and Sr and Nd isotope compositions) are investigated with regard to the NLM granites. These results help to provide better constraints on the petroge-netic model and tectonic setting for these granites. 2. Geological background and sampling

The Yangtze and Cathaysia Blocks are separated by a fault system including the Jiangshan-Shaoxing fault in Zhe-jiang province, the Dongxiang-Pingxiang fault in Jiangxi 430 P.-S. Hsieh et al. / Journal of Asian Earth Sciences 33 (2008) 428–451

province, and the Xupu-Sanjiang in Guangxi province (Fig. 1a). Basement of the Cathaysia Block was Mesopro-terozoic to Ordovician flysch sequences, possibly with a small Archean nucleus, before being subject to greenschist to amphibolite facies metamorphism (Li et al., 1991; Wang and Mo, 1995; Xu et al., 2005). It was unconformably cov-ered by the Devonian to Triassic sediments and intruded by the wide spreading Mesozoic igneous rocks (Gilder et al., 1996; Chen and Jahn, 1998). Temporally, the Triassic rocks expose dispersively in the Cathaysia interior, whereas the Jurassic and Cretaceous rocks are more concentrated in the central and a narrow zone close to the SE coastal areas, respectively, of the Cathaysia Block (Fig. 1a).

Darongshan (DRS) granitic suites, including Darong-shan, Pubai, Taima, Jiuzhou, and Nadong batholiths, are the most important intrusive rocks related to the Indosini-an orogeny (Triassic) in the Cathaysia Block. They are distributed in an elongate fashion for >400 km in the southeastern part of Guangxi Province, with a total area of exposure for granites over 10,000 km2(Fig. 1b). During Late Permian to Triassic, this area was strongly folded, and intrusion of DRS granitic suites was controlled by the Bobai-Cenxi and Lingshan-Tangxian deep faults and their bifurcated faults (Deng et al., 2004). DRS granitic suites are mainly composed of cordierite–biotite granite, gar-net–cordierite granite and hypersthene granite porphyry. Although there are few systematic geochemical and isoto-pic studies, the presence of abundant Al-rich minerals, such as cordierite, garnet, hypersthene, almandine, sillimanite, and andalusite, in these rocks suggests that they can be classified as S-type granites ofChappell and White (1974)

in the mineralogical sense.

EY magmatism (180–140 Ma) (Zhou and Li, 2000) has resulted in a large number of granitic intrusives forming the Nanling Mountains (NLM). This area, situated in Longitude 110°E–117°E and Latitude 22°400_N–26°200_N, is the watershed of the Yangtze and Pearl Rivers (Fig. 1c). For decades, related mineralization in the NLM granites has been the center of attention because of the existence of some world-class W, Si, and Bi ore deposits (Chen and Jahn, 1998; Zhou and Li, 2000; Pirajno and Bagas, 2002; Li et al., 2003, 2004; Xu et al., 2005). Large NLM batholiths are composed predominantly of biotite granite, granodiorite and A-type granite, and sporadically distributed small bodies are mainly gabbro and syenite. They are accompanied by some Late Jurassic bimodal bas-alts and rhyolites in southern Jiangxi (Li et al., 2003). Based on recent geochronological data, ages of the NLM granites appear to be largely concentrated in the range of ca. 165–155 Ma (Li, 2000; Xu et al., 2005).

Products of LY magmatism (140–80 Ma) are mainly restricted in the SE coast of S. China forming a NE–SW magmatic belt about 900 km long and 150 km wide (Fig. 1a). Rock types are mainly granodiorite, monzogra-nite, syenogranite and alkali feldspar granite as a rock complex (110–100 Ma), A-type granites (100–90 Ma) and basalt–rhyolite bimodal volcanics (90–80 Ma) (Chen

et al., 2004). Judging from the general appearance of horn-blende in the intermediate rocks, as well as the geochemical and Sr–Nd isotopic characteristics, LY granitoids are typ-ical I-type granites (Chen et al., 2000). Granites of the Fuz-hou and ZhangzFuz-hou Complexes (FZC) in Fujian are the most extensively studied representatives of this belt (Fig. 1d).

In this paper, 25 granitic samples and their apatite sep-arates from four batholiths (Pubai, Jiuzhou, Taima, and Nadong) of the DRS granitic suites (Fig. 1b), six batholiths (Jiufeng, Qitianling, Dadongshan, Guidong, Fogang, and Lianyang) of the NLM region (Fig. 1c), and six FZC plu-tons (Fuzhou, Danyang, and Lianjiang from the Fuzhou Complex and Zudi, Yanqian, and Changtai from the Zhangzhou Complex) of the Fujian Province (Fig. 1d) are studied. Rocks from these places are commonly referred to in the discussion of Mesozoic granitic magma-tism of the Cathaysia Block (e.g., Chen and Jahn, 1998; Zhou and Li, 2000; Xu et al., 2005). In addition, samples from one gabbro body (Chebu) and three syenite bodies (Quannan, Paitan, and Ejinao) are also included especially for revealing the apatite geochemistry in magmas with lar-ger involvement of the mantle component (Li et al., 2003), in which Paitan and Ejinao bodies are closely associated with the Fogang batholith. Batholith, pluton and igneous body are arbitrarily set by the surface exposure of rocks >500, 500–100, and <100 km2.

3. Analytical methods

Major element contents of rock samples were deter-mined with an X-ray fluorescence using a RigakuÒ RIX 2000 spectrometer on fused glass disk at Department of Geosciences, National Taiwan University (Lee et al., 1997). Trace element abundances were measured by induc-tively coupled plasma-mass spectrometry (ICP-MS) using a Perkin ElmerÒ Elan-6000 spectrometer at Guangzhou Institute of Geochemistry, the Chinese Academy of Sci-ences (Liu et al., 1996). The analytical precision and accu-racy are generally better than 5% for most elements. Sr and Nd isotope compositions were measured using a FinniganÒ MAT 262 mass spectrometer at Department of Earth Sci-ences, National Cheng Kung University. The isotopic ratios were corrected for mass fractionation by normalizing to 86Sr/88Sr = 0.1194 and 146Nd/144Nd = 0.7219. Long-term laboratory measurements for SRM 987 Sr and La Jolla Nd standards yield 0.710239 ± 0.000012 (n = 30) and 0.511847 ± 0.000012 (n = 30), respectively.

Apatite separates were mounted in epoxy and well-pol-ished on the exposed surfaces suitable for both the electron microprobe (EMP) and laser ablation inductively coupled plasma-mass spectrometer (LA-ICP-MS) analyses. Major and minor elements (Ca, P, F, Cl, Si, Na, and S) were analyzed using a Shimadzu-ARL EMX-SM7 electron microprobe equipped with four channels of wavelength dis-persive spectrometer at National Taiwan University. The analyzing conditions were 15 kV acceleration potential, P.-S. Hsieh et al. / Journal of Asian Earth Sciences 33 (2008) 428–451 431

6–8 nA sample current, and a beam spot10 lm in diam-eter. Data reduction is based on the ZAF correction proce-dure (Chen and Tung, 1984). Concentrations of trace elements (Mn, Sr, Ba, Th, U, Pb, and Y) including rare earth elements (REE) were determined on single apatites by an excimer LA-ICP-MS at the Key Laboratory of Con-tinental Dynamics, Northwest University, Xi’an, China. The ICP-MS used is an Elan 6100 DRC (Dynamic Reac-tion Cell) from Perkin Elmer/SCIEX (Canada) coupled to the GeoLas 200M laser-ablation system (MicroLas, Go¨ttingen, Germany) equipped with a 193 nm ArF-exci-mer laser and a homogenizing, imaging optical system. Detailed analytical procedures and instrumental operating conditions have been described by Gao et al. (2002). The spot size has been adjusted to 40–60 lm in this study. Helium was used as a carrier gas for argon to maintain sta-ble and optimum excitation conditions. Coexisting acces-sory uraninites, if any, were subjected to EMP analysis for U, Th, Pb, and Y.

Calcium content of apatite obtained from EMP analysis was used as an internal standard to correct for differences in the ablation yield between sample and reference materi-als, matrix effects and signal drift in the ICP-MS. The external standard used for the apatite in situ analysis was NIST SRM 610 (Pearce et al., 1997). Each spot analysis consisted of approximately 30 s background (gas blank) followed by 60 s data acquisition from the sample. Trace element concentrations were calculated using GLITTER 4.0 software (Macquarie University). Accuracy and preci-sion of the present LA-ICP-MS setting have been reported for known international standards, including NIST glass standards (SRM 610, 612, and 614) and USGS rock stan-dards (BCR-2G, BHVO-2G and BIR-1G) (Gao et al., 2002). Generally, detection limits for a spot size of 60 lm are well below the ppm level for all elements analyzed, and the precision of analyses on the NIST SRM 610 are 1–2.6% for trace elements in the studied samples.

4. Whole-rock geochemistry and Sr–Nd isotope compositions For a large number of Mesozoic intrusive rocks in S. China analyzed, whole-rock geochemical data reported here (Table 1) only include those which have accompanied by apatite data (n = 29). Additional major element compo-sitions of the NLM granites (n = 37) referred in this study are listed inAppendix. All the Sr and Nd isotopic compo-sitions (n = 38) are given in Table 2.

DRS granites are evolved rocks (SiO2= 68–75 wt %), with relatively low Na2O contents (1.6–2.2 wt %). The alu-minum saturation index (ASI, or A/CNK = 1.15–1.47) indicates that these rocks are strongly peraluminous (ASI > 1.1, Frost et al., 2001). In contrast, FZC granites have a large variation of silica (SiO2= 58–77 wt %), rela-tively high Na2O contents (3.0–4.5 wt %), and are mostly metaluminous (ASI = 0.89–1.04). Voluminous NLM gran-ites also have a wide range of silica (SiO2= 63–78 wt %) or aluminum saturation (ASI = 0.91–1.21) but moderate

Na2O contents (2.4–4.0 wt %), as compared with wider fields of DRS and FZC granites illustrated (Fig. 2a and b). Chebu gabbro (SiO2= 50 wt %; Na2O = 2.6 wt %) and Quannan syenite (SiO2= 54 wt %; Na2O = 4.1 wt %) in S. Jiangxi as well as Paitan and Ejino syenites (SiO2= 59–62 wt %; Na2O = 3.6–7.5 wt %) in Guangdong appear to be Na-rich rocks as a whole. They are grouped together because of a crystal fractionation relationship between gabbro and syenite (Li et al., 2003).

The total FeO and CaO contents of the studied granites are positively correlated and the DRS, NLM and FZC granites are indistinguishable at lower CaO–total FeO con-tent. At higher CaO and total FeO levels, FZC granites have higher CaO contents than DRS granites at a given total FeO content, and NLM granites fall between the FZC and DRS suites (Fig. 2c). Using the Na2O–K2O divide for differentiating S- and I-type granites (Chappell and White, 2001), FZC and DRS granites can be assigned to S- and I-types, respectively, whereas NLM granites astride the dividing line (Fig. 2a). Also the ASI (at 1.1) is a very effective parameter to separate FZC from DRS granites, again, NLM granites spread in the middle (Fig. 2b).

Chondrite-normalized REE distribution patterns of all the Mesozoic intrusive rocks in S. China, except gabbro and few syenitic rocks, invariably show gentle LREE enrichment and moderate negative Eu anomalies (Fig. 3). Here we draw attention to the general similarity in the shape and relative abundance of REE distribution patterns for all granitic rocks, probably indicating insignificant feld-spar crystal fractionation between the mafic and felsic granites. Only in the gabbro–syenite series, feldspar crystal fractionation is important (Fig. 3c).

Sr and Nd isotope compositions of the Mesozoic grani-toids in S. China are more variable as compared with I-and S-type granites in LFB (Fig. 4). DRS granites generally exhibit the most enriched character with a small range of eNd(T) (12 to 14) but a wide range of Isr (0.722– 0.730). Such isotopic characteristics, even more enriched than the LFB S-type granites (Fig. 4), have been inter-preted as indicating the involvement of sedimentary mate-rials (Shen and Lin, 2002). These rocks also have significantly more enriched Sr and Nd isotope composi-tions than the pre-Mesozoic (predominantly Caledonian) granites that overwhelmed other rocks in the Cathaysia Block before overprinted by the Mesozoic magmatism (see later section). FZC granites, on the other hand, have more depleted Sr and Nd isotope compositions with nar-rower ranges of both eNd(T) (3 to 6) and Isr (0.705– 0.708). They conform to I-type granites in the SE China coastal area (Chen et al., 2000), and overlap with the more enriched field of I-type granites in LFB (Collins, 1996).

In the NLM region, gabbro, and syenites have Sr and Nd isotope compositions (eNd(T) = 3.0 to 2.6 and Isr = 0.704–0.708) consistent with similar rocks reported byLi et al. (2003). They overlap with more depleted LFB I-type granites, but can be clearly distinguished from 432 P.-S. Hsieh et al. / Journal of Asian Earth Sciences 33 (2008) 428–451

Table 1

Chemical compositions of representative Darongshan (DRS), Fuzhou–Zhangzhou Complex (FZC), Nanling Mountains (NLM) granites, gabbro and syenites of S. China

DRS granites FZC granites

Locality: Jiuzhou Taima Pubai Pubai Nadong Nadong Nadong Zudi Changtai Danyang Fuzhou Yanqian Lianjiang

Sample no.: 22GX12 22GX01 22GX13 22GX16 22GX07 22GX03 22GX04 93ZUD01 93CHT01 92BD01 92KS01 93YAN01 92GT01

Latitude (°N): 22°080 _22°080 _22°200 _22°240 _21°420 _21°430 _21°410 _24°400 _24°380 _26°280 _26°050 _24°440 _26°110 Longitude (°E): 108°350 _108°260 _109°180 _109°340 _107°470 _108°030 _107°580 _117°490 _117°460 _119°290 _119°240 _117°470 _119°330 wt % SiO2 68.05 70.68 70.40 71.24 71.25 73.67 74.47 58.61 63.26 67.54 74.94 75.57 77.04 TiO2 0.73 0.42 0.50 0.42 0.52 0.30 0.32 0.90 0.69 0.52 0.21 0.16 0.13 Al2O3 14.34 14.72 14.61 14.27 13.98 13.22 12.96 17.40 17.27 15.89 13.96 13.22 13.07 Fe2O3 4.79 2.50 3.43 3.32 3.41 1.77 1.90 7.45 5.08 3.03 1.17 1.04 0.78 MnO 0.05 0.03 0.04 0.03 0.04 0.02 0.03 0.12 0.08 0.06 0.07 0.05 0.07 MgO 1.57 0.49 1.01 1.02 0.67 0.29 0.30 3.49 1.77 0.95 0.16 0.09 0.00 CaO 2.02 2.15 1.47 1.26 1.73 1.34 1.23 6.49 4.75 2.81 1.08 0.96 0.50 Na2O 1.61 1.89 1.73 1.77 2.05 2.16 2.06 2.99 3.39 3.69 4.18 3.57 4.46 K2O 3.88 4.91 4.26 4.14 4.80 5.05 4.96 2.65 2.84 4.68 4.24 4.92 4.38 P2O5 0.15 0.14 0.20 0.15 0.13 0.08 0.08 0.25 0.23 0.15 0.07 0.06 0.04 Total 97.19 97.93 97.65 97.62 98.58 97.90 98.31 100.35 99.36 99.32 100.08 99.64 100.47 ASI 1.36 1.19 1.44 1.47 1.19 1.15 1.18 0.89 1.00 0.98 1.04 1.02 1.01 A/NK 2.09 1.75 1.96 1.93 1.63 1.47 1.48 2.23 2.00 1.43 1.22 1.18 1.08 ppm Ni 21.2 6.14 11.3 12.3 8.10 4.58 4.82 16.2 2.74 3.65 n.d. n.d. n.d. Ga 17.9 18.0 17.5 17.0 17.5 17.1 16.8 18.0 19.0 16.9 13.5 14.3 16.8 Rb 195 242 236 241 266 322 337 134 134 185 145 266 157 Sr 109 95.4 78.0 70.8 78.8 45.6 38.9 512 599 408 159 75.5 33.4 Y 35.8 42.7 43.2 37.7 44.6 44.2 103 26.5 19.0 26.9 18.1 15.2 28.8 Zr 274 220 219 194 279 171 186 222 175 251 117 108 119 Nb 15.7 12.0 14.9 13.3 14.2 10.7 11.2 10.3 9.28 13.3 11.3 14.5 19.6 Cs 10.0 10.2 19.6 14.7 9.23 21.1 18.5 9.08 7.17 8.39 1.51 6.09 1.20 Ba 631 758 498 396 661 375 291 782 806 1155 1409 167 339 La 46.7 46.7 43.0 30.2 50.8 42.0 43.3 28.3 34.0 45.9 30.4 30.4 26.9 Ce 93.4 92.0 88.1 59.2 101 83.2 84.1 60.0 67.0 88.5 57.4 47.2 52.5 Pr 10.6 11.1 10.2 7.13 11.3 9.87 10.5 7.59 8.08 10.3 6.33 4.71 5.85 Nd 39.6 41.3 38.2 26.3 42.0 36.0 39.0 28.9 29.6 35.8 20.7 14.1 18.6 Sm 7.80 8.53 8.01 5.85 8.50 7.48 9.43 5.89 5.52 6.39 3.57 2.23 3.41 Eu 1.26 1.31 1.00 0.865 1.16 0.794 1.05 1.27 1.18 1.16 0.800 0.377 0.430 Gd 7.50 8.14 7.71 5.92 8.38 7.40 11.7 5.19 4.53 5.27 3.05 2.18 3.15 Tb 1.12 1.30 1.27 1.04 1.32 1.23 2.40 0.804 0.642 0.784 0.460 0.312 0.550 Dy 6.11 7.05 7.13 6.12 7.34 6.93 14.9 4.45 3.28 4.31 2.63 1.84 3.51 Ho 1.18 1.49 1.51 1.28 1.50 1.51 3.62 0.880 0.611 0.864 0.548 0.41 0.796 Er 3.25 4.09 4.14 3.46 4.09 4.11 9.67 2.39 1.66 2.45 1.64 1.32 2.51 Tm 0.509 0.601 0.635 0.518 0.619 0.640 1.46 0.352 0.235 0.366 0.257 0.232 0.406 Yb 3.40 3.88 4.03 3.21 3.97 4.02 8.37 2.20 1.47 2.43 1.81 1.72 2.80 Lu 0.579 0.604 0.644 0.504 0.646 0.633 1.29 0.338 0.222 0.384 0.290 0.300 0.451 Hf 6.80 6.17 5.77 5.38 6.83 4.94 5.28 5.72 4.86 6.45 3.49 3.83 4.42 Ta 1.22 1.13 1.28 1.21 1.17 1.10 1.14 0.841 0.817 1.01 0.914 1.01 1.32 Pb 29.1 36.4 28.8 29.4 33.1 36.7 34.9 19.3 17.5 28.1 21.4 39.4 28.3 Th 22.1 25.5 21.2 16.0 26.4 28.3 29.8 23.1 26.8 21.4 17.2 38.0 13.3 U 5.54 6.35 4.40 4.25 6.38 6.76 7.53 5.82 6.46 3.98 4.22 11.8 3.64

(continued on next page)

P.-S. Hsieh et al. /Journal of Asian Earth Sciences 33 (2008) 428–451 433

Table 1 (continued)

NLM granites

Locality: Guidong Guidong Guidong Fogang Fogang Fogang Fogang Qitianling Lianyang Jiufeng Jiufeng Dadongshan

Sample no.: 99GD16 99GD17 99GD18b 99GD36a 99GD02a 99GD03b 99GD19 21HUN01 21GD01a 97GD61 97GD54 21GD07

Latitude (°N): 24°320 _24°320 _24°420 _24°010 _23°160 _23°560 _24°080 _25°330 _24°190 _25°160 _25°200 _24°450 Longitude (°E): 114°110 _114°110 _114°040 _114°040 _113°450 _114°480 _113°550 _112°590 _112°070 _113°290 _113°180 _112°510 wt % SiO2 63.34 64.77 73.53 66.11 70.89 71.02 75.56 66.18 66.67 69.03 71.66 77.43 TiO2 0.71 0.58 0.15 0.59 0.43 0.29 0.15 0.58 0.36 0.52 0.31 0.11 Al2O3 16.63 16.83 14.64 15.75 14.75 14.60 12.57 14.43 15.82 15.16 14.63 11.18 Fe2O3 4.66 3.94 1.19 4.14 2.63 2.37 1.53 4.20 2.55 3.50 2.19 0.78 MnO 0.08 0.07 0.05 0.06 0.05 0.05 0.05 0.06 0.05 0.08 0.06 0.02 MgO 1.79 1.43 0.02 0.75 0.33 0.22 b.d.l. 0.58 1.07 1.06 0.54 0.09 CaO 3.68 3.26 0.93 2.47 1.54 1.89 1.03 2.22 2.34 2.76 1.81 0.71 Na2O 3.09 3.03 2.97 2.59 2.42 2.92 2.82 3.57 4.00 2.86 3.00 2.95 K2O 3.82 4.46 5.10 5.64 5.53 5.02 4.71 5.43 5.05 4.37 5.34 4.88 P2O5 0.28 0.24 0.15 0.17 0.23 0.09 0.05 0.20 0.13 0.17 0.13 0.05 Total 98.08 98.61 98.73 98.27 98.80 98.47 98.47 97.45 98.04 99.51 99.67 98.20 ASI 1.04 1.07 1.21 1.06 1.15 1.07 1.08 0.91 0.97 1.05 1.04 0.98 A/NK 1.87 1.72 1.47 1.52 1.48 1.43 1.29 1.23 1.31 1.61 1.37 1.10 ppm Ni 9.67 7.28 0.329 0.218 0.394 0.767 0.532 3.30 12.8 6.62 7.46 1.65 Ga 22.2 22.2 19.0 22.1 8.59 27.1 17.9 21.1 15.0 17.9 20.9 12.5 Rb 216 218 396 257 139 429 279 175 310 267 448 356 Sr 324 345 52.4 152 58.9 182 65.5 48.9 136 221 148 26.8 Y 26.7 22.2 27.2 50.0 14.2 51.1 43.8 279 17.2 30.5 30.4 42.7 Zr 250 216 86.7 247 96.3 234 156 30.7 214 175 163 96.6 Nb 16.2 13.1 23.4 32.1 8.53 28.5 25.4 b.d.l. 13.8 20.5 26.8 16.7 Cs 12.8 11.3 32.3 8.69 2.65 17.6 3.39 28.0 24.1 18.7 56.9 18.4 Ba 943 1219 266 833 261 609 183 792 688 686 669 72.0 La 66.3 75.9 23.9 73.0 22.1 60.1 47.6 62.0 20.2 31.1 47.7 34.4 Ce 127 142 51.7 142 49.3 119 100 124 39.0 64.4 89.4 52.80 Pr 14.7 15.9 6.27 16.4 6.06 13.5 12.0 15.0 4.70 8.03 9.98 8.87 Nd 54.6 55.3 20.8 57.6 21.4 47.5 41.8 55.8 17.1 29.0 32.9 32.2 Sm 9.61 8.77 4.88 11.8 4.23 10.0 9.21 11.0 3.31 6.08 6.25 7.76 Eu 1.59 1.61 0.437 1.01 0.434 1.09 0.448 1.75 0.860 1.10 0.900 0.370 Gd 6.56 5.82 5.09 10.1 3.85 8.33 8.21 10.0 3.11 5.20 5.60 6.79 Tb 0.971 0.836 0.897 1.67 0.577 1.44 1.41 1.47 0.460 0.850 0.847 1.17 Dy 5.20 4.52 4.95 9.79 2.92 8.46 8.02 8.00 2.62 5.19 5.02 6.80 Ho 1.00 0.838 0.950 1.78 0.542 1.73 1.57 1.63 0.560 1.03 0.972 1.42 Er 2.71 2.22 2.65 4.87 1.37 5.06 4.67 4.53 1.66 2.87 2.81 4.09 Tm 0.383 0.315 0.403 0.675 0.207 0.814 0.704 0.680 0.270 0.477 0.472 0.670 Yb 2.35 1.99 2.50 3.88 1.21 5.33 4.36 4.43 1.97 2.96 2.96 4.37 Lu 0.361 0.296 0.360 0.569 0.179 0.840 0.668 0.730 0.350 0.437 0.466 0.710 Hf 7.25 6.10 3.06 7.85 2.72 7.97 5.76 7.61 6.30 5.01 5.20 4.06 Ta 1.44 1.14 5.34 3.27 0.846 3.74 2.68 2.64 1.65 3.03 4.95 2.82 Pb 39.2 46.4 39.2 24.0 15.5 44.0 28.7 37.9 66.3 n.d. n.d. 44.3 Th 24.7 27.5 26.5 38.1 17.0 77.1 66.2 30.0 32.9 26.2 45.4 45.1 U 4.60 3.86 27.5 14.9 3.43 21.3 18.7 7.35 9.04 10.0 21.8 10.7 434 P.-S. Hsieh et al. /Journal of Asian Earth Sciences 33 (2008) 428–451

NLM gabbro and syenites

Locality: Chebu Quannan Paitan Ejinao

Sample no.: 99CB02 99QN06 97GD77 97GD84 Latitude (°N): 24°500 _24°460 _23°300 _23°430 Longitude (°E): 115°020 _114°290 _113°470 _113°380 wt % SiO2 49.87 54.25 62.14 59.18 TiO2 1.91 2.06 0.71 0.24 Al2O3 15.79 16.88 15.92 18.94 Fe2O3 11.69 9.52 5.61 5.14 MnO 0.16 0.19 0.12 0.23 MgO 6.54 2.57 0.73 0.36 CaO 9.44 6.31 2.90 1.70 Na2O 2.55 4.11 3.58 7.45 K2O 1.11 2.83 5.75 5.62 P2O5 0.28 0.65 0.25 0.10 Total 100.29 99.77 98.30 100.05 ASI 0.70 0.79 0.92 0.88 A/NK 2.93 1.72 1.31 1.03 ppm Ni n.d. 0.882 37.2 3.67 Ga n.d. 22.8 22.5 23.2 Rb 58.8 94.9 135 264 Sr 307 615 236 66.2 Y 32.2 35.5 32.7 26.2 Zr 104 245 578 339 Nb 19.7 59.3 37.0 69.3 Cs 1.69 6.78 3.17 3.08 Ba 202 790 997 132 La 18.9 51.6 63.2 90.2 Ce 40.3 101 124 156 Pr 5.29 12.4 14.8 16.2 Nd 22.2 49.1 54.8 50.0 Sm 5.09 10.0 10.3 7.01 Eu 1.48 2.94 2.49 0.533 Gd 5.29 8.76 7.89 6.02 Tb 0.850 1.33 1.23 0.835 Dy 4.99 7.39 6.52 4.63 Ho 1.05 1.43 1.25 0.914 Er 2.99 3.87 3.51 2.77 Tm 0.413 0.544 0.510 0.456 Yb 2.54 3.40 3.24 3.08 Lu 0.378 0.484 0.497 0.518 Hf 2.60 6.97 13.2 8.12 Ta 1.17 4.09 2.07 2.75 Pb 4.91 23.0 15.8 n.d. Th 5.51 11.2 15.6 17.1 U 1.27 3.50 3.25 3.17

ASI={molar [Al2O3/(CaO + Na2O + K2O)]} and A/NK={molar [Al2O3/(Na2O + K2O)]}.

Major element results of FZC granites are taken from the work ofChen et al., 2000. n.d., not determined.

b.d.l., below detection limit.

P.-S. Hsieh et al. /Journal of Asian Earth Sciences 33 (2008) 428–451 435

Table 2

Sr and Nd isotope compositions of Darongshan (DRS), Fuzhou-Zhangzhou Complex (FZC), Nanling Mountains (NLM) granites, gabbro and syenites of S. China

Locality Sample SiO2(wt %) Rb (ppm) Sr (ppm) Sm (ppm) Nd (ppm) 87Rb/86Sr 87Sr/86Sr 2r 147Sm/144Nd 143Nd/144Nd 2r Isr eNd(T)

DRS granites (T = 230 Ma) Jiuzhou 22GX12 68.05 195 109 7.80 39.6 5.19 0.742478 10 0.1191 0.511852 12 0.725500 13.1 Taima 22GX01 70.68 242 95.4 8.53 41.3 7.35 0.747977 10 0.1249 0.511862 12 0.723948 13.1 Pubai 22GX13 70.40 236 78.0 8.01 38.2 8.80 0.757267 14 0.1267 0.511812 12 0.728479 14.1 Nadong 22GX07 71.25 266 78.8 8.50 42.0 9.81 0.754211 8 0.1224 0.511880 12 0.722119 12.7 FZC granites (T = 110 Ma) Zudi 93ZUD01 58.61 134 512 5.89 28.9 0.758 0.707164 12 0.1232 0.512423 12 0.705979 3.2 Changtai 93CHT01 63.26 134 599 5.52 29.6 0.648 0.706770 11 0.1127 0.512410 12 0.705757 3.3 Danyang 92BD01 67.54 185 408 6.39 35.8 1.31 0.709549 11 0.1079 0.512281 12 0.707497 5.7 Fuzhou 92KS01 74.94 145 159 3.57 20.7 2.64 0.711478 10 0.1043 0.512269 12 0.707350 5.9 Yanqian 93YAN01 75.57 266 75.5 2.23 14.1 10.2 0.720960 12 0.0956 0.512340 11 0.705012 4.4 Lianjiang 92GT01 77.04 157 33.4 3.41 18.6 13.6 N.A. 0.1108 0.512358 11 4.3 NLM granites (T = 160 Ma) Guidong 99GD14 63.46 197 301 9.00 48.9 1.90 0.726641 11 0.1114 0.511950 12 0.722330 11.7 99GD16 63.34 216 324 9.61 54.6 1.93 0.726703 13 0.1064 0.511958 10 0.722303 11.5 99GD17 64.77 218 345 8.77 55.3 1.83 0.726501 12 0.0959 0.511936 10 0.722339 11.7 99GD18b 73.53 396 52.4 4.88 20.8 17.5 N.A. 0.1418 0.511947 13 12.4 Fogang 99GD36a 66.11 257 152 11.8 57.6 4.90 0.726432 11 0.1239 0.512005 11 0.715287 10.9 97GD93 69.81 304 89.4 10.5 60.6 9.84 0.736744 14 0.1043 0.512045 12 0.714358 9.7 99GD04a 70.10 259 731 8.92 49.3 1.03 0.718974 12 0.1093 0.511963 12 0.716640 11.4 97GD51 70.80 202 171 6.57 40.0 3.42 0.719924 13 0.0993 0.512040 13 0.712144 9.7 99GD02a 70.89 139 58.9 4.23 21.4 6.86 0.731616 16 0.1196 0.511965 10 0.716023 11.6 99GD03b 71.02 429 182 10.0 47.5 6.82 0.729667 14 0.1276 0.512105 10 0.714147 9.0 97GD79 74.92 226 72.1 4.72 21.2 9.09 0.741271 14 0.1346 0.511987 11 0.720589 11.5 99GD22 75.14 277 74.4 7.74 38.2 10.8 0.739791 14 0.1224 0.512033 10 0.715288 10.3 99GD19 75.56 279 65.5 9.21 41.8 12.3 0.743255 16 0.1331 0.512124 11 0.715216 8.8 97GD82 77.57 256 21.1 11.2 44.3 35.1 N.A. 0.1526 0.512128 11 9.1

Qitianling 21HUN01 66.18 175 48.9 11.0 55.8 10.4 N.A. 0.1192 0.512220 12 6.6

21HUN02 68.01 189 41.1 8.28 37.9 13.3 N.A. 0.1321 0.512182 12 7.6

Lianyang 21GD01a 66.67 310 136 3.31 17.1 6.60 N.A. 0.1170 0.512060 11 9.7

21GD02b 74.83 276 34.4 3.28 12.8 23.2 N.A. 0.1549 0.512104 12 9.6 Jiufeng 97GD53 68.55 288 177 6.54 34.3 4.72 0.723628 14 0.1152 0.512034 9 0.712903 10.2 97GD61 69.03 267 221 6.08 29.0 3.49 0.723439 12 0.1269 0.512037 10 0.715507 10.3 97GD54 71.66 448 148 6.25 32.9 8.77 0.732235 11 0.1149 0.512020 12 0.712298 10.4 Dadongshan 21GD08 72.00 364 54.6 10.7 61.1 19.3 N.A. 0.1059 0.512003 12 10.6 21GD11 75.10 539 8.78 7.28 20.3 178 N.A. 0.2168 0.512136 11 10.2 21GD07 77.43 356 26.8 7.76 32.2 38.5 N.A. 0.1457 0.512065 12 10.2

NLM gabbro and syenites (T = 173–136 Ma)

Chebu 99CB02 49.87 58.8 307 5.09 22.2 0.554 0.707825 13 0.1384 0.512569 13 0.706462 0.1a

Quannan 99QN06 54.25 94.9 615 10.0 49.1 0.447 0.705864 22 0.1233 0.512715 11 0.704816 3.0b

Paitan 97GD77 62.14 135 236 10.3 54.8 1.65 0.710877 13 0.1142 0.512431 10 0.707492 2.6c

Ejinao 99GD84 59.18 264 66.2 7.01 50.0 11.5 0.729378 12 0.0848 0.512472 9 0.707059 1.3d

eNd(T) = [(143_Nd/144_Nd)

sample(T)/(143Nd/144Nd)CHUR(T)l]  l04, (143Nd/144Nd)sample(T) = (143Nd/144Nd)sample (147Sm/144Nd)sample(expkTl), (143Nd/144Nd)CHUR(T) = 0.512638–0.1967

 (expkT  l), k = 0.00654 Ga1_.

N.A., not available due to low Sr content (<50 ppm) or high Rb/Sr ratio (>5.0).

a and b, age = 173 and 165 Ma (Li et al., 2003); c and d: age = 145 and 136 Ma (Hsieh et al., 2005).

436 P.-S. Hsieh et al. /Journal of Asian Earth Sciences 33 (2008) 428–451

FZC granites – the I-type representatives of S. China (Fig. 4). However, granites differ markedly from gabbro

and syenite in having more enriched Sr and Nd isotope compositions for the majority of samples (eNd(T) =8.8 to 11.6 and Isr = 0.712–0.717). Few exceptions include (1) the Qitianling granodiorite (sample 21HUN01: Fig. 2. (a) Na2O vs. K2O diagrams for the Mesozoic intrusive rocks from S.

China (b) ANK (molar Al2O3/(Na2O+K2O)) vs. ACNK, or the aluminum saturation index (ASI) (molar Al2O3/(CaO+Na2O+K2O)), and (c) CaO vs. total FeO (Table 1andAppendix;Chen et al., 2000). DRS granites are strongly peraluminous (ASI > 1.1) and have low CaO and Na2O contents, whereas FZC granites are metaluminous to mildly peraluminous, and high in CaO and Na2O, generally corresponding to the S- and I-type granites of the Lachlan Fold Belt (LFB), respectively. NLM granites have character-istics between DRS and FZC granites, while NLM gabbro and syenites have low aluminosity and various CaO and Na2O contents.

Fig. 3. Chondrite-normalized REE distribution patterns for (a) DRS granites, (b) NLM granites, (c) NLM gabbro and syenites, and (d) FZC granites of S. China. Normalizing values are those recommended bySun and McDonough (1989). M, mafic granites (SiO2= 57–70 wt %); F, felsic granites (SiO2> 70 wt %). The M–F division is that proposed bySha and

Chappell (1999).

eNd(T) =6.6 and Isr is not available due to high Rb/Sr ratio) that is close to FZC granites and (2) some Guidong granodiorites and granites (eNd(T) =11.5 to 12.4 and Isr = 0.722) that are close to DRS granites. Our data for the Fogang granites (eNd(T) =8.8 to 11.6 and Isr = 0.712–0.721; n = 10, Table 2) show significantly higher Isr values than those (n = 17) reported by Li et al. (2007), indicating the complexity of magma sources for this huge batholith. Other NLM granites, although wide spreading in the Cathaysia Block, distribute in a small field somewhat deviated from that of the Fogang granites, but in large part coincide with the160 Ma Hong Kong gran-ites to the south (Darbyshire and Sewell, 1997; n = 6) (Fig. 4). Overall, the NLM granites, except Qitianling granodiorite, can be regarded as possessing intermediary Sr and Nd isotopic compositions between FZC and DRS granites.

5. Apatite geochemistry

Representative EMP and LA-ICP-MS analyses of apa-tites are shown in Tables 3 and 4, respectively. All these apatites are fluorapatites (averaged as F = 3.2–5.4 wt % and Cl < 0.3 wt % except sample 93ZUD01 that has F = 3.1 wt % and Cl = 1.0 wt %), typical of igneous origin (Nash, 1984). Consistency of F and Cl contents among these apatites minimizes the possibility of being signifi-cantly affected by F metasomatism. Following the group-ing of apatite geochemistry on LFB granites (Sha and Chappell, 1999), apatites are dealt separately with respect to the mafic (SiO2= 57–70 wt %) and felsic host rocks (SiO2> 70 wt %), except the DRS granites that are treated as a whole because of the small range of SiO2(68–74 wt %).

5.1. Sodium, silicon, and sulfur

In contrast to the clear difference in the bulk Na2O con-tents (Fig. 2a), there is a large overlap on Na2O contents for apatites from DRS (Na2O = 0.15–0.47 wt %) and FZC (Na2O = 0.04–0.38 wt %) granites. However, the DRS apatites are distinct from FZC apatites for low SiO2 and SO3contents (Fig. 5a and b). The FZC apatites display a positive Na2O–SO3correlation whereas the DRS apatites are characterized by low and relatively constant SO3 con-tents over a fourfold range in Na2O contents. These fea-tures resemble those shown by the LFB S- (SO3< 0.05 wt % and Na2O > 0.1 wt %) and I-type granites (SO3= 0.03–0.75 wt % and Na2O = 0.02–0.15 wt %) (Sha

and Chappell, 1999). Overall, the NLM apatites have Na2O and SiO2contents similar to FZC apatites, but have low SO3contents that are the characteristics of DRS apa-tites. Therefore, based on the Na2O–SO3-SiO2relationship of the NLM apatites, their host granites cannot be explic-itly assigned to S- or I-type.

5.2. Thorium and uranium

Thorium contents of the DRS apatites are the lowest (5–40 ppm) although the U concentrations are the highest (20–130 ppm) among all granitic suites. In contrast, FZC apatites have low U concentrations (<35 ppm) but high Th contents (10–90 ppm), and both are positively corre-lated. One mafic sample (93ZUD01) is exceptional for containing higher U concentrations (Fig. 5c). Covariation between U and Th of the NLM apatites mimics that of the FZC apatites, in which those from the felsic granites have low Th and U contents (<35 and <20 ppm, respec-Fig. 4. eNd(T) vs. Isr plots for DRS, FZC, and NLM granites, and NLM gabbro and syenite in S. China (Table 2). They are compared with LFB S-type and I-type granites, gabbros, and syenites in S. Jiangxi, granites from Fogang batholith, and Hong Kong. Noted that many NLM mafic and felsic rocks are indistinguishable for Sr and Nd isotope compositions.

tively) in spite that very few have slightly higher U con-tents. Low Th contents in the DRS apatites relative to oth-ers probably reflect the role of monazite, a Th-rich mineral that is rather abundant in the DRS granites (Chen et al., 2006). The possible mechanism for high U concentration in DRS apatites will be discussed later.

5.3. Strontium and manganese

DRS apatites generally contain less Sr and more Mn (Sr < 85 ppm; Mn = 2100–8200 ppm) than FZC apatites (Sr = 40–360 ppm; Mn = 550–4100 ppm), although there is an overlap between the fields for DRS and FZC apatites (Fig. 5d). Based on Mn content alone, FZC apatites can be distinguished between those forming in the maifc (Mn = 550–1200 ppm) and felsic (Mn = 1900–4100 ppm) rocks. Apatites from NLM granites apparently can be grouped for forming in the mafic (Mn < 800 ppm) and fel-sic (Mn > 2100 ppm) rocks as well, and those from gabbro and syenites (Mn = 230–1230 ppm) are close to the mafic rocks. In the case of LFB granites, low-Mn apatites (<1000 ppm) only appear in mafic I-type granites that also contain more Sr (110–400 ppm) than apatites from S-type and felsic I-type granites (30–210 ppm) (Sha and Chappell, 1999). However, Sr content of apatites from NLM granites,

varied from 55–200 ppm in mafic to 30–160 ppm in felsic rocks, is not as obvious as Mn content for discriminating these two groups of rock.

5.4. REE (plus Y) abundances and ratios

Although total REE abundances of DRS and FZC apa-tites are generally the same (5000–14,000 ppm), their LREE (La–Eu) and HREE (Gd–Lu) behave differently. Hence using the LREE/HREE ratio, it is possible to differ-entiate between DRS (<3.5) and FZC (6.0–30) apatites (Fig. 6a). Similarly, it is possible to distinguish the

varia-tion of Eu and the Sm/Nd ratio between DRS

(Eu < 7.0 ppm and Sm/Nd > 0.30) and FZC (Eu > 11 ppm and Sm/Nd < 0.23) apatites (Fig. 6b). In addition, the dominance of Y (in place of Ho) in DRS (Y = 3000– 6000 ppm; Ce = 1000–3400 ppm) whereas Ce in FZC (Y = 700–2700 ppm; Ce = 2500–6000 ppm) apatites is noted (Fig. 6c). Behavior of REE in apatites from FZC mafic and felsic granites are generally the same, except one felsic sample 92KS01 that shows a tendency towards DRS apatites (Fig. 6a–c).

Total REE concentrations are rather variable in apatites from NLM granites (1800–14,500 ppm) and gabbro plus syenite (3800–47,000 ppm). Except sample 99GD36a (a Table 3

Electron microprobe analyses in representative apatites from Darongshan (DRS), Fuzhou–Zhangzhou Complex (FZC), Nanling Mountains (NLM) granites, gabbro, and syenites of S. China

Locality Sample No. (number) CaO (wt %) P2O5(wt %) SiO2(wt %) Na2O (wt %) SO3(wt %) F (wt %) Cl (wt %) Total (wt %) DRS granites Jiuzhou 22GX12-5 54.15 41.43 0.19 0.21 0.01 3.94 0.17 100.10 Taima 22GX01-2 52.19 42.26 0.25 0.28 0.03 4.52 0.29 99.82 Pubai 22GX13-5 54.55 41.39 0.17 0.19 0.02 4.52 0.03 100.87 22GX16-2 55.19 40.21 0.15 0.17 0.02 4.38 0.03 100.15 Nadong 22GX07-2 53.35 42.55 0.27 0.25 0.03 4.11 0.08 100.64 22GX03-3 52.45 41.46 0.28 0.23 <0.01 5.33 0.06 99.81 22GX04-2 51.16 43.32 0.25 0.41 <0.01 3.60 0.09 98.83 FZC granites Zudi 93ZUD01-6 52.99 40.44 0.36 0.13 0.03 3.00 1.11 98.06 Changtai 93CHT01-8 54.33 40.69 0.42 0.11 0.08 4.55 0.25 100.43 Danyang 92BD01-3 54.37 39.85 0.42 0.16 0.25 4.56 0.16 99.77 Fuzhou 92KS01-7 52.55 40.86 0.56 0.20 0.16 4.73 0.14 99.20 Yanqian 93YAN01-6 54.22 42.14 0.43 0.09 0.22 4.13 0.01 101.24 Lianjiang 92GT01-3 53.12 40.45 0.48 0.26 0.42 4.87 0.02 99.62 NLM granites Guidong 99GD16-4 53.95 40.95 0.24 0.01 0.02 5.13 0.01 100.31 99GD18b-1 53.99 40.33 0.08 0.16 <0.01 4.00 <0.01 98.56 Fogang 99GD36a-9 52.89 42.28 0.19 0.04 0.02 4.79 0.12 100.33 99GD02a-1 52.49 41.95 0.20 0.16 0.04 4.79 <0.01 99.63 99GD19-3 53.03 40.60 0.52 0.16 <0.01 5.41 0.01 99.73 Qitianling 21HUN01-6 54.41 39.98 0.48 0.08 0.05 5.16 0.02 100.18 Lianyang 21GD01a-1 54.15 39.77 0.48 0.09 0.03 5.10 <0.01 99.62 Jiufeng 97GD54-3 51.31 42.62 0.29 0.12 <0.01 5.42 0.01 99.77 Dadongshan 21GD07-5 52.27 40.45 0.66 0.09 <0.01 5.03 0.01 98.51

NLM gabbro and syenites

Chebu 99CB02-2 52.94 42.61 0.08 0.12 <0.01 3.40 0.11 99.26

Quannan 99QN06-4 53.12 38.57 0.63 0.25 0.20 5.13 0.14 98.04

Paitan 97GD77-1 54.61 40.95 0.42 0.10 0.05 4.38 0.06 100.57

Ejinao 97GD84-3 54.77 40.30 0.76 0.11 0.01 3.96 0.01 99.92

Table 4

LA-ICP-MS analyses of trace element abundances (ppm) in representative apatites from Darongshan (DRS), Fuzhou–Zhangzhou Complexes (FZC), Nanling Mountains (NLM) granites, gabbro, and syenites of S. China

DRS granites FZC granites

Locality: Jiuzhou Taima Pubai Pubai Nadong Nadong Nadong Zudi Changtai Danyang Fuzhou Yanqian Lianjiang

Sample no.: 22GX12-5 22GX01-2 22GX13-3 22GX16-1 22GX07-1 22GX03-1 22GX04-2 93ZUD01-1 93CHT01-1 92BD01-1 92KS01-1 93YAN01-1 92GT01-1

Mn 3074 3313 2448 2403 2773 8116 5451 669 810 1114 2908 2027 3428 Sr 65.3 48.7 49.9 60.7 47.8 27.9 18.6 359 249 97.0 81.3 260 46.3 Ba 0.035 0.284 0.084 0.876 0.714 1.00 0.083 0.680 0.398 0.369 0.509 0.550 0.524 Th 13.2 15.8 13.1 12.5 30.4 30.5 51.7 24.4 76.4 26.3 15.0 91.9 10.6 U 36.8 104 59.2 127 24.7 23.6 82.2 55.1 18.1 9.36 4.59 14.9 1.75 Pb 7.11 6.66 5.85 6.52 7.01 9.35 7.46 5.11 6.01 4.53 4.03 6.26 7.11 La 420 749 391 369 597 606 548 1141 1623 1489 567 3312 1460 Ce 1426 2338 1328 1179 1990 1881 2075 3012 3581 2912 2038 5887 3462 Pr 229 323 210 195 312 301 307 387 435 304 313 514 432 Nd 1123 1395 987 995 1582 1554 1361 1645 1906 1156 1519 1634 1850 Sm 472 529 420 454 622 675 592 316 357 224 510 196 375 Eu 2.93 2.01 3.85 6.23 2.48 30.5 2.57 12.8 16.3 9.29 17.6 29.4 30.9 Gd 605 647 503 583 827 876 692 303 309 225 554 254 360 Tb 121 137 108 131 167 184 151 36.0 35.1 26.1 100 19.5 49.5 Dy 732 893 694 792 1030 1121 911 189 179 139 621 107 286 Ho 137 175 139 142 202 217 166 35.5 31.4 26.2 119 23.7 55.3 Er 323 438 365 336 500 524 404 85.9 73.8 64.5 306 67.3 142 Tm 41.1 57.7 50.6 43.0 58.6 63.6 54.7 10.9 8.36 7.51 41.8 10.0 17.0 Yb 244 347 324 266 326 365 355 62.6 50.0 45.8 281 68.1 99.0 Lu 29.7 40.7 40.9 31.0 35.9 41.2 37.5 8.18 6.37 6.22 35.1 11.6 11.9 Y 3698 4391 3372 3722 5384 5974 4243 979 841 707 3432 726 1634 P LREE 3673 5335 3340 3199 5105 5020 4885 6513 7918 6093 4965 11573 7610 P HREE 2233 2735 2224 2323 3148 3391 2751 731 693 540 2058 562 1021 LREE/HREE 1.65 1.95 1.50 1.38 1.62 1.48 1.78 8.91 11.4 11.3 2.41 20.6 7.45 Sm/Nd 0.420 0.379 0.426 0.457 0.393 0.435 0.435 0.192 0.188 0.194 0.336 0.120 0.203 (La/Sm)N 0.560 0.891 0.586 0.512 0.604 0.565 0.583 2.27 2.86 4.18 0.699 10.6 2.45 (Ho/Lu)N 2.06 1.92 1.53 2.04 2.52 2.47 1.98 1.94 2.21 1.88 1.51 0.97 2.08 (La/Yb)N 1.16 1.46 0.817 0.938 1.24 1.12 1.10 12.3 22.0 22.0 1.36 32.8 9.95 Eu/Eu* _{0.017 0.010 0.026 0.037 0.011 0.012 0.012 0.125 0.147 0.125 0.100 0.402 0.253} NLM granites

Locality: Guidong Guidong Guidong Fogang Fogang Fogang Fogang Qitianling Lianyang Jiufeng Jiufeng Dadongshan

Sample no.: 99GD16-1 99GD17-5 99GD18b-1 99GD36a-1 99GD02a-1 99GD03b-1 99GD19-1 21HUN01-1 21GD01a-1 97GD61-5 97GD54-1 21GD07-4

Mn 645 509 6507 799 5851 993 2290 708 695 656 2180 4087 Sr 195 170 30.7 71.7 69.2 52.2 29.6 117 57.1 166 105 54.1 Ba 0.510 0.500 6.77 5.84 2.05 1.58 0.042 0.241 0.120 0.498 0.071 0.114 Th 75.0 50.0 26.6 82.9 18.6 26.0 16.9 39.9 67.9 25.5 97.5 10.4 U 33.1 25.9 80.6 52.4 16.2 11.1 5.03 16.5 18.1 17.2 41.7 3.86 Pb 9.68 8.68 24.6 5.70 5.13 18.0 8.02 6.20 5.66 6.97 6.62 11.0 La 732 488 332 111 336 232 797 3429 1555 461 940 651 Ce 2124 1387 1235 381 1300 877 2325 6442 4243 1399 2437 1891 Pr 304 209 226 69.4 213 165 352 668 598 219 362 284 440 P.-S. Hsieh et al. /Journal of Asian Earth Sciences 33 (2008) 428–451

Nd 1470 965 984 438 1045 900 1765 2472 2788 998 1778 1330 Sm 327 209 426 290 471 337 550 408 696 245 483 383 Eu 37.9 24.4 6.87 5.66 14.6 4.95 6.63 7.16 6.62 14.8 26.1 5.30 Gd 297 185 493 533 545 393 583 387 681 226 474 381 Tb 40.7 23.1 87.3 123 106 61.0 96.7 45.1 98.1 31.9 83.3 62.5 Dy 223 132 446 849 644 370 561 244 522 188 505 360 Ho 43.3 25.7 72.6 180 116 72.8 103 47.5 95.8 37.0 104 68.9 Er 107 66.7 187 489 288 190 266 121 225 96.5 275 179 Tm 13.2 8.44 22.5 66.9 36.0 23.9 34.5 15.7 26.4 12.7 40.3 25.0 Yb 84.6 52.1 131 417 221 146 214 96.8 152 78.4 293 182 Lu 11.9 8.01 16.6 52.2 25.0 20.6 26.6 12.8 19.3 10.9 39.3 23.9 Y 1183 768 2619 5239 3163 2335 3157 1405 2524 1170 3497 2017 P LREE 4994 3281 3209 1296 3379 2515 5795 13426 9886 3336 6026 4544 P HREE 821 501 1456 2712 1982 1279 1884 970 1819 682 1814 1282 LREE/HREE 6.09 6.55 2.20 0.478 1.71 1.97 3.08 13.8 5.43 4.89 3.32 3.54 Sm/Nd 0.222 0.216 0.432 0.662 0.451 0.374 0.312 0.165 0.249 0.245 0.272 0.288 (La/Sm)N 1.41 1.47 0.504 0.242 0.448 0.433 0.912 5.29 1.41 1.19 1.22 1.07 (Ho/Lu)N 1.62 1.44 1.97 1.55 2.06 1.58 1.73 1.67 2.22 1.52 1.18 1.29 (La/Yb)N 5.85 6.32 1.82 0.18 1.02 1.07 2.52 23.9 6.92 3.97 2.17 2.42 Eu/Eu* _{0.365 0.372 0.046 0.043 0.088 0.042 0.036 0.054 0.029 0.189 0.165 0.042}

NLM gabbro and syenites

Locality: Chebu Quannan Paitan Ejinao

Sample no.: 99CB02-2 99QN06-5 97GD77-5 97GD84-4 Mn 361 915 680 724 Sr 239 35.8 518 268 Ba 0.380 31.4 68.9 2.32 Th 34.2 31.6 30.5 82.6 U 11.0 6.87 7.30 9.83 Pb 3.68 2.49 2.24 3.52 La 1439 2989 1014 4006 Ce 2883 6141 2174 7692 Pr 301 763 303 814 Nd 1127 3323 1516 2972 Sm 188 613 307 416 Eu 22.6 40.4 94.6 27.2 Gd 86.0 557 280 384 Tb 20.6 65.6 33.5 39.9 Dy 111 332 166 208 Ho 21.6 59.3 29.6 40.0 Er 54.9 129 67.2 96.3 Tm 6.89 14.1 7.15 11.4 Yb 44.1 74.4 41.0 66.5 Lu 5.93 8.51 5.52 8.29

(continued on next page)

P.-S. Hsieh et al. /Journal of Asian Earth Sciences 33 (2008) 428–451 441

mafic pegmatitic granite), apatites from NLM mafic gran-ites (LREE/HREE = 4.0–14; Eu = 4.7–45 ppm; Sm/ Nd = 0.16–0.28) and gabbro plus syenite (LREE/ HREE = 8.5–21; Eu = 18–95 ppm; Sm/Nd = 0.13–0.21) have chemical characteristics similar to FZC apatites. In contrast, apatites from NLM felsic granites have interme-diate chemical signatures between DRS and FZC apatites, i.e., slight enrichment in LREE (1600–6500 ppm) com-pared with HREE (900–2000 ppm), lower LREE/HREE (1.4–7.0), higher Sm/Nd ratios (0.23–0.49) and enrichment in Y (Fig. 6a–c).

Apatites of sample 99GD36a are extremely low in total REE abundance (<4000 ppm), especially LREE (<1300 ppm) and Eu (<6 ppm), and LREE/HREE (0.48– 1.3), but high in HREE abundance, and Sm/Nd ratio (0.46–0.66), with Y (>1300 ppm) more strongly enriched over Ce (<400 ppm) (Fig. 6a–c). Such features are in accord with those described for apatites from granite peg-matites (Belousova et al., 2002).

5.5. Chondrite-normalized REE distribution patterns Apatites from DRS granites generally show very gentle convex or near-flat REE distribution patterns ((La/ Sm)N= 0.47–1.17 and (La/Yb)N= 0.66–1.75) with strong negative Eu anomalies (Eu/Eu*_{< 0.04). Like the case of} LFB S-type granites (Sha and Chappell, 1999), a slight Nd depletion is observed (Fig. 7a). On the other hand, most apatites from FZC granites are characterized by straight and strongly right-inclined REE distribution pat-terns ((La/Sm)N= 1.50–2.97 and (La/Yb)N= 5.80–22.1) with weak to moderate Eu anomalies (Eu/Eu*_{= 0.10–} 0.30). Again these patterns resemble those of LFB I-type granites (Fig. 7b). However, unusual behavior is shown by apatites from fractionated I-type granites. One example is sample 93YNC01 which shows concave HREE patterns with low (Ho/Lu)Nratios in apatites (0.97–1.30), whereas the generalized straight and right-inclined LREE charac-teristics of apatites in this group of rocks are maintained (Fig. 7c). The host rock of these apatites shows a low (Ho/Lu)Nratio (0.55) as well (Fig. 3d), implying a linkage to the whole-rock geochemistry for such a concave HREE pattern of apatites. REE patterns of apatites from another fractionated sample 92KS01, which possesses the highest ASI value (1.04) in the FZC granites (Table 1), differ from the FZC I-type granite proper by having slightly right-inclined REE distribution patterns ((La/Sm)N= 0.70–1.05 and (La/Yb)N= 1.36–2.95) with moderate Eu anomalies (Eu/Eu*_{= 0.10–0.14) (Fig. 7d). In fact, such REE patterns} resemble apatites from the majority of NLM granites (see later sections).

Apatites from NLM granites show various kinds of REE patterns. Those from rocks with high ASI values (P1.1) are similar to DRS apatites, e.g., sample 99GD02a of the Fogang batholith and sample 99GD18b of the Guidong batholith (Fig. 7e), and those from rocks with low ASI values (60.91) are similar to apatites from

Ta ble 4 (continue d ) NLM gabbro and syenites Loca lity: C hebu Quannan Paitan Ejinao Samp le no.: 99CB 02-2 99QN06 -5 97GD7 7-5 97GD8 4-4 Y 549 1438 810 1051 P LREE 5960 13870 5408 15927 P HR EE 351 1240 631 855 LR EE/HRE E 17.0 11.2 8.57 18.6 Sm /Nd 0.167 0.185 0.203 0.140 (La/ Sm) N 4.82 3.07 2.08 6.07 (Ho /Lu) N 1.63 3.12 2.40 2.16 (La/ Yb) N 22.1 23.4 16.7 40.7 Eu/E u * 0.421 0.208 0.969 0.204

the FZC granite proper, e.g., sample 21HUN01 (Qitianling batholith; Fig. 7f). In fact, the majority of apatites from NLM granites are metaluminous to mildly peraluminous (ASI = 0.97–1.08) and characterized by slightly right-inclined REE distribution patterns ((La/Sm)N= 0.40–1.68 and (La/Yb)N= 0.76–10.9) with various Eu negative anomalies (Eu/Eu*_{= 0.02–0.39) (Fig. 7g–j). They are} dis-tinguished clearly from NLM gabbro and syenites which are close to the FZC unfractionated I-type granite (Fig. 7k–n). A special type is shown by apatites from sam-ple 99GD36a, the pegmatitic granite, that has strong desam-ple- deple-tion in LREE ((La/Sm)N= 0.22–0.27) and variable HREE content ((La/Yb)N= 0.18–0.67) plus moderate Eu anoma-lies (Eu/Eu*_{< 0.06), leading to the unique convex apatite} REE patterns comparable with those from Norwegian granite pegmatites (Fig. 7o).

6. Discussion

Trace element abundances of apatite often vary with some parameters of the host rock, such as the oxidation state, SiO2content, total alkali and ASI values (Belousova

et al., 2002). Here we examine the validity of these param-eters (except total alkali) in relation to the geochemistry of

apatites from different kinds of the Mesozoic granites in S. China. The effect of total alkali is not a major concern because it can be important only to the syenitic rocks which have very small population in the studied rocks that gener-ally have calc-alkaline affinities.

6.1. Preferential substitution of elements in apatite and oxidation state of magma

In apatite, substitution of Ca2+by trivalent lanthanides (REE3+) and Y mainly involves charge compensated for smaller (Si4+) and larger (Na+) atoms as expressed in the following reactions (Fleet and Pan, 1994; Sha and Chap-pell, 1999):

LREE3þþ Si4þ ¼ Ca2þþ P5þ ð1Þ

HREE3þþ Naþ _{¼ 2Ca}2þ _ð2Þ

The Na2O–SiO2relationship (Fig. 5a) and total LREE rel-ative to HREE (Fig. 6a) for apatites from DRS strongly peraluminous granites, FZC and NLM metaluminous to slightly peraluminous granites, and NLM syenites and gab-bro provide evidence to support the preference of reaction

(2) substitution in the former and reaction(1)in the latter 0 0.2 0.4 0.6 0.8 1.0 0 0.1 0.2 0.3 0.4 0.5 DRS granites NLM mafic granites NLM felsic granites FZC granites NLM Gb andSy 0 0.2 0.4 0.6 0.8 Na₂O(wt%) 0 0.1 0.2 0.3 0.4 0.5 Na₂O(wt%) DRS granites NLM mafic granites NLM felsic granites FZC mafic granites FZC felsic granites 0 100 200 300 400 0 2000 4000 6000 8000 10000 500 600 (756,915) 0 40 60 80 100 120 20 80 U (ppm) 0 20 40 60 100 120 14 0 (24. 0, 25 1) (21. 9, 21 3) (35. 5, 40 7) DR Sg ra ni te s NL Mm af ic gran it es NL Mf el si cg ra ni te s FZ Cg ra nite s Sa mp le 93Z UD 01 NL MG ba nd Sy NLM Gb and Sy (11187, 40.2) (15466, 45.7) Mn (ppm)

Fig. 5. (a) Si and Na concentrations, (b) S and Na concentrations in apatites from DRS, FZC, NLM granites, gabbro, and syenites of S. China. (c) Th and U concentrations, and (d) Sr and Mn concentrations in apatites from DRS, FZC, NLM granites, gabbro, and syenites of S. China.

two kinds of rock. A few NLM felsic rocks that have ASI > 1.1 also follow reaction(2).

The ionic state of S is another controlling factor for Si or Na enrichments in apatites: high oxygen fugacity condi-tions favor the presence of the S6+state that can couple with Na+ or Si4+ to substitute for P5+ or Ca2+, whereas low oxygen fugacity environments enhance conversion of S6+ to S2 which is highly incompatible in apatites (Sha

and Chappell, 1999). Apparently, substitution of S6+ is insignificant for DRS apatites (Fig. 5b), indicating the dominance of S2under the lower oxygen fugacity environ-ment for strongly peraluminous magmas. On the contrary, the positive correlation between SO3and Na2O in apatites from FZC granites supports the substitution of S and Na for Ca and P in more oxidizing metaluminous magmas. The Na and S relationship in apatites from NLM syenites and gabbro generally follows the trend of FZC apatites, but SO3and Na2O contents in apatites from NLM granites are low. Only those from felsic rocks have higher Na2O contents up to the level of DRS granites.

Abundances of Mn, U, and Eu in apatites have also been proposed to be controlled by the oxygen fugacity of the melt (Belousova et al., 2001). Extensive Mn and U enrichment in DRS apatites (Figs. 5c–d) strongly indicates that their host magmas have been subject to low oxygen fugacity condi-tions. The presence of accessory uraninite (UO2= 86.7– 93.3 wt %, ThO2= 3.1–6.9 wt %, PbO = 2.5–2.8 wt %, and Y2O3= 0.3–0.8 wt %) in DRS granites is also diagnostic. Conversely, mafic granites of FZC and NLM granites (gab-bro and syenites as well) are representative of more oxidized rocks. Again, magmas of a few NLM felsic granites are more likely to be equivalents at lower oxygen fugacity (Fig. 5c). It is worth noting that high U concentrations in apatite from sample 93ZUD01 in the FZC granites (55–85 ppm) is also accompanied by exceptionally high Cl (1.1 wt %) and low F (3.0 wt %) contents (Table 3). Because the host rock is not particularly rich in U (Table 1), this enrichment could be due to earlier crystallization of apatite, rather than the presence of another U-retaining mineral (most probably zir-con) in the melt.

Owing to large amounts of Eu2+entering into the feldspar structure, reduced melts, usually high in Eu2+/Eu3+ratios (Sha and Chappell, 1999), would hold a strong Eu negative anomaly in the enclosing apatites. This can explain the case of the DRS apatites. On the contrary, a high amount of Eu3+ (or lower Eu2+/Eu3+ratio) in oxidized melts can assist Eu occupancy in the apatite structure because the ionic radius of Eu3+ is similar to Ca2+. Such a mechanism could be responsible for the slight to medium Eu negative anomalies in FZC and NLM apatites. On this ground, apatites from Paitan syenite (sample 99GD77) that have higher Eu concen-trations (20–95 ppm) leading to very slight or even obscure Eu negative anomalies (Eu/Eu*_{= 1.0–0.24) in the REE} pat-terns (Fig. 7l), could have crystallized from melts with higher oxygen fugacities as well. In summary, DRS apatites appear to have crystallized from relatively reduced magmas, whereas FZC and the majority of NLM apatites have crys-tallized from more oxidized magmas, except those in a few felsic granites that may have formed from magmas with lower oxygen fugacity similar to the DRS apatites.

6.2. Effects of coexisting REE-rich accessory minerals Based on small variations of REE concentration and similar whole-rock REE distribution patterns for the LREE/HREE DRS granites NLM mafic granites NLM felsic granites FZC granites Sample 92KS01 Sample 99GD36a NLM Gb and Sy 0 10000 20000 30000 40000 5 0 10 15 20 25 30 99GD36a 92KS01 80 0 20 40 60 10 0 0. 1 0. 2 0. 3 0. 4 0. 5 0. 6 0. 7 Eu (ppm) 99GD36a 92KS01 La 0 10 20 30 40 50 60 70 80 90 100 Y 0 10 20 30 40 50 60 70 80 90 100 Ce 0 10 20 30 40 50 60 70 80 90 100 92KS01 99GD36a

Fig. 6. (a) Total LREE concentration and LREE/HREE ratio plots, (b) Sm/Nd ratio and Eu concentration plots and (c) Y–Ce–La triangular plots for apatites from DRS, FZC, NLM granites, gabbro, and syenites of S. China.

Mesozoic granitoids in S. China (Fig. 3), large differences of REE abundances and distribution patterns present in their enclosing apatites are unlikely to be controlled by the whole-rock REE concentration. Therefore, internal redistribution of REE among different REE-rich minerals is considered more likely. Major factors are the mineral assemblage, crystallization sequence of REE-rich minerals, and partition coefficients of REE between the mineral and the melt. Among the common REE-retaining accessory minerals, apatite is high in all REE, zircon and xenotime are enriched in HREE, and monazite and allanite are char-acteristically enriched in LREE. Progressive replacement of allanite and sphene by monazite and xenotime commonly occur in rocks with increasing ASI and decreasing CaO (Bea, 1996).

Depletion of LREE in apatites from LFB S-type and fel-sic I-type granites has been ascribed to the crystallization of monazites which usually become saturated earlier than apatites in the low-Ca, reduced and strongly peraluminous magmas (Sha and Chappell, 1999). Generally, DRS granites and two NLM granitic samples (99GD02a and 99GD18b) possess the criteria of such magmas and indeed contain more abundant monazite (Montel et al., 1996; Chen et al., 2006) and allanite (Liu et al., 2005). The slight Nd negative anomaly for apatite from these rocks (Fig. 7a and e) matches early crystallization of monazite due to

pre-dominance of Nd, or the largest partitioning of Nd among REE, in this mineral (Charoy, 1986; Yurimoto et al., 1990). In the high-Ca, more oxidized and metaluminous magma that usually forms I-type granites, monazite is absent. Chemical variations between mafic (granodiorite and monzogranite) and felsic (syenogranite and alkali feld-spar granite porphyry) suites in the FZC granies mainly result from magmatic processes of crystal fractionation (Martin et al., 1994; Chen et al., 2000). Both apatite and zircon are considered to be early magmatic phases in the FZC granites and hence the REE composition of apatite may reflect a counterpart from zircon (Bea, 1996). That is why apatites from mafic and most felsic suites have the same right-inclined distribution patterns (Fig. 7b). In some cases (e.g., felsic rocks) the effect of predominant zircon fractional crystallization over apatite would result in concave-upward HREE distribution patterns (i.e., (Ho/ Lu)N< 1). This can appear not only in the host rocks (Fig. 3c), but also in the containing apatites (Fig. 7c).

Ward et al. (1992)reported that magma evolution con-trolled by crystal fractionation of plagioclase, biotite and accessory apatite, monazite, zircon and xenotime can lead to strong reduction of REE concentrations with SiO2 varying from 71 to 74 wt % in the Dartmoor granitic plu-ton (SW England). Basically, crystallization dominated either by LREE-rich monazite or by HREE-rich zircon Fig. 7. Chondrite-normalized REE distribution patterns of apatites from (a) DRS and LFB S-type granites, (b–d) FZC granites (including LFB I-type granites in b), (e–j) NLM granites, (k–n) NLM gabbro and syenites, and (o) NLM pegmatitic granite. Normalizing values are afterSun and McDonough (1989).

fractionation at two different stages is responsible for such distinctions. Judging from the similarity of whole-rock REE concentrations and patterns within a wide range of silica contents (SiO2= 63–78 wt %) as well as the scarcity of monazite, xenotime, and allanite in the NLM granites, involvement of these accessory minerals during the course of magma evolution is very small, not even to affect the REE shape of coexisting apatites. In few cases, zircon crys-tallization indeed can cause the whole-rock HREE patterns to become concave upward (Fig. 3b), but not their enclos-ing apatites (Fig. 7g–j).

The particular type of REE distribution patterns for apatites from NLM pegmatitic granites (sample 99GD36a; Fig. 7o) is considered to be the only case that has been significantly affected by crystallization of REE-rich accessory minerals. Petrographically, there are many large (1.3–1.5 mm) euhedral crystals of allanite and sphene in this rock, and apatites occur mainly as inclusions in the large biotite (Fig. 8). Early crystallization of allanite (and sphene to a lesser degree;Sawka et al., 1990) can account for LREE depletion in apatites, because partition coeffi-cients of LREE between allanite and high-silica melt (Mahood and Hildreth, 1983) are very similar to those between monazite and peraluminous melt (Montel, 1986, 1993). In other words, allanite alone has the ability to cause a strong LREE depletion in later apatites for sample 99GD36a.

6.3. Apatite REE distribution patterns and whole-rock silica contents, ASI and eNd(T) values

The geochemistry of apatite seems to provide a tool for distinguishing mafic from felsic granites in the Nanling Mountains area. For example, Na, Mn, and REE contents of apatites from NLM felsic rocks are more akin to those from DRS granites (Figs. 5a and d and 6), that are felsic or near-felsic (SiO2= 68–75 wt %). Also, the apatite REE distribution pattern becomes more flattened, relative to gabbro and syenites (Fig. 7k–n), as the whole-rock SiO2 content increases (Fig. 7g–j). Therefore, NLM granites seem to follow the general tendency that, when rocks become felsic in composition, their geochemical properties would be limited by the low-temperature melts (Chappell and White, 2001). Such apatite geochemical features may thus be related to the low crystallization temperature of magmas.

Furthermore, apatite REE geochemistry, especially the REE distribution patterns, varies concordantly with the ASI value of the host granites, even including gabbro and syenites. Apatites from NLM high ASI (P1.1) granites (e.g., samples 99GD02a and 99GD18b) are more akin to the DRS pattern (host rock ASI P1.15); those from gab-bro and syenite bodies and low ASI (60.91) batholiths (e.g., Qitianling) are same as the FZC pattern (host rock ASI 61.04); others with ASI values in between (0.97– 1.08), such as Guidong (1.04–1.07), Lianyang (0.97), Jiuf-eng (1.04–1.05), Fogang (1.07–1.08), and Dadongshan

(0.98), are intermediary but closer to the FZC pattern (Fig. 7g–j). In fact, this group of rocks are basically mildly peraluminous (ASI = 1.0–1.1) varying in a wide range of Fig. 8. (a) Photomicrograph showing the large euhedral allanite crystal in the NLM pegmatitic granite (sample 99GD36a). Small apatites crystal-lized in the boundary between allanite and biotite (open nicol) are noted. (b) Photomicrograph showing the euhedral sphene crystal (open nicol). (c) Photomicrograph showing apatites occur mainly as inclusions in the large biotite (open nicol).

silica (SiO2= 63–78 wt %). Moreover, the pegmatite sam-ple (99GD36a), although having an ASI value (1.06) simi-lar to this group, follows the unique apatite REE patterns of well-known pegmatitic granites elsewhere (Fig. 7o).

Isotopic composition is a useful parameter for indicating the source material of granitic magmas. Here we examine the REE distribution patterns of apatite from the studied rocks with their corresponding eNd(T) and ASI values. A striking feature is that samples with the characteristic apa-tite groups shown inFig. 7 are clustered at the particular place on the eNd(T) vs. ASI diagram (Fig. 9). If this dia-gram is arbitrarily separated by two dividing lines on ASI = 1.1 and eNd(T) =8, DRS granites and few NLM plutonic rocks (samples 99GD02a and 99GD18b) fall in the 4th quadrant (Group I); most NLM granodior-ites and grangranodior-ites fall in the 3rd quadrant (Group IIa) with only Qitianling granodiorite (21HUN01) in the 2nd quad-rant; FZC granites (Group IIIa) and NLM gabbro and sye-nites (Group IIIb), although all plotted in the 2nd quadrant, are classified into two groups according to rock type. On this basis, Qitianling granodiorite is thus included into Group IIIa. For the particular sample of the fraction-ated FZC granite (92KS01) that shows the same REE pat-tern of apatite as the majority of NLM rocks, it is further categorized as Group IIb by possessing high eNd(T) value. Therefore, the majority of NLM granites, including both the mafic and felsic types (SiO2= 63–78 wt %), belong to Group IIa, suggesting that REE distribution patterns of their enclosing apatites are unlikely affected by the whole-rock SiO2contents, hence, the process of crystal fraction-ation. It is worthy to note that, although the Guidong granodiorites display Group IIa apatite patterns whereas

the associated granite displays Group I patterns, they are plotted separately in the eNd(T)-ASI space more akin to the NLM granite proper and DRS granites. Granitic rocks with higher ASI values favor crystallization of monazite – the LREE phosphate in the magma (Chappell, 1999), coex-isting apatites are thus depleted in LREE. Those which have ASI < 1.1 need to be further examined for the eNd(T) value. Generally, the higher the eNd(T) value, the more enriched LEEE in the REE distribution pattern of apatites (Fig. 9). Conclusively, the case of NLM granites demon-strates that the host rock ASI and eNd(T) values, rather than SiO2contents (or the mafic–felsic relationship) as sug-gested by Belousova et al. (2002), are the most sensitive parameters to correlate the shape of REE distribution in the enclosing apatites.

6.4. Petrogenetic model of NLM granites

Chappell, (1999) gave a thorough petrogenetic discus-sion on the I- and S-type granites in LFB based on the ASI values of these rocks. Major conclusions were (1) All the granites in the LFB are resulted from partial melting of the crust, with I-type rocks derived from low ASI (<1.0) and S-type rocks from high ASI (>1.0) sources; only their mafic suites can reflect the ASI of the sources. (2) Most mafic and felsic suites represent temperature-depen-dent partial melts, and felsic suites in I and S granite types may be indistinguishable due to similar compositions approaching the temperature minimal of granites. (3) If the mafic–felsic relationship is achieved through fractional crystallization, it shall lead to distinctly high ASI (>1.1) in S-type granites, but close to 1.0 in I-type granites.

How-Fig. 9. eNd(T) vs. ASI plots for DRS, FZC, NLM granites, gabbro, and syenites of S. China. Groupings are based on the REE distribution patterns of enclosing apatites. Group I: near-flat patterns including strongly peraluminous and Nd isotope depleted DRS granites and NLM high ASI granites (99GD02a and 99GD18b). Group II: slightly right-inclined patterns, (a) including metaluminous to mildly peraluminous and Nd isotope depleted NLM granites; and (b) including FZC high ASI granite (92KS01). Group III: strongly right-inclined patterns, (a) including metaluminous and Nd isotope enriched FZC granites and NLM high eNd(T) granites (21HUN01); and (b) including Nd isotope enriched NLM gabbro and syenites. Noted that the pegmatitic granite (99GD36a) shows unique patterns among all samples.