Composition of Garnet from the Xianghualing Skarn Sn Deposit, South China: Its Petrogenetic Signiﬁcance and Exploration Potential

: The Xianghualing skarn Sn deposit in the southwestern part of the southern Hunan Metallogenic Belt is a large Sn deposit in the Nanling area. In this paper, the garnet has been analyzed by laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) to obtain the concentrations of the major and trace elements. The results reveal that the garnets from the Xianghualing deposit mainly belong to andradite-grossular (grandite) solid solution and are typically richer in Al than in Fe. They show enrichment in heavy rare earth elements (HREEs) and notably lower light rare earth elements (LREEs), and commonly negative Eu anomalies, indicative of a relatively reduced formation environment. The garnets have high Sn concentrations between 2313 ppm and 5766 ppm. It is also evident that there is a positive correlation between Sn and Fe, suggesting that Sn 4 + substitutes into the garnets through substituting for Fe 3 + in the octahedral position. Combined with previous studies, it can be recognized that the Sn concentrations of garnet in skarn Sn deposits are generally high, whereas the W concentrations are relatively low. This is just the opposite in garnets from skarn W deposits that typically have high W, but low Sn concentrations. In polymetallic skarn deposits with both economic Sn and W, the concentrations of both metals in garnets are relatively high, although varying greatly. Therefore, the Sn and W concentrations in garnets can be used to evaluate a skarn deposit’s potential to produce Sn and (or) W mineralization, which is helpful in exploration.


Introduction
The Xianghualing skarn Sn deposit is an important skarn deposit in the center of the Nanling Range, which is situated in the northern part of Linwu County, southern Hunan Province. It contains 0.17 Mt Sn with a grade ranging from 0.93% to 1.39%, and ranks the 10th largest among the Sn skarns in China [1]. The deposit has been widely investigated since the 1980s, which provides systematic understanding of the regional metallogenic background [2], geological characteristics, ore genesis [3][4][5], geochronology, geochemistry of rocks and ores in the mining district [6,7], and crystal chemistry of Sn-bearing minerals [8]. Previously, many attempts have been made to reveal the Sn mineralization

Geology of the Deposit
The Xianghualing deposit is located in the northern part of the southern Hunan Province, which is near the center of the Nanling Range [29]. It is one of the largest Sn (-W) deposits in the southwestern part of the South Hunan Metallogenic Belt, and tectonically it lies in the intersection area between the NE-trending, deep-seated Chenzhou-Linwu fault zone and the S-trending fault zone. The study area appears as a NE-trending tectono-magmatic dome, which is due to Jurassic magmatic diapirism under a regional extensional setting [6]. The Cambrian slate and meta-sandstone are the main components of the core of the Xianghualing deposit (Figure 2), and the flanks are mainly composed of the Devonian clastic rocks, including the Middle Devonian Tiaomajian Formation conglomerates, sandstones, and shales, the Middle Devonian Qiziqiao Formation limestone and dolomites, and the Upper Devonian Shetianqiao Formation dolomitic limestones and sandstones. In addition, Carboniferous carbonates and Jurassic clastic rocks are also present locally [35]. The Cambrian and Middle Devonian strata are considered to be the major host rocks in the Xianghualing Sn deposit [29,36]. The Laiziling and Jianfengling stocks are the two largest granitic stocks in this area, which intrude into Cambrian Minerals 2020, 10, 456 4 of 17 meta-sandstone and Devonian carbonate of the Tongtianmiao dome, respectively. The Laiziling intrusion can be divided into four vertical zones from the base to the top: albite-microcline granite zone, albite granite zone, greisen zone, and pegmatite zone [30,32]. The U-Pb zircon age determined from the Laiziling Granite is 155 ± 2 Ma [37].
The Xianghualing deposit is a stratabound skarn Sn deposit [5], and the orebodies are mainly stratiform, pipe-and vein-like in shape [38][39][40]. Field investigations have revealed three major stages of skarn alteration and metal mineralization in the Xianghualing deposit: (1) A prograde pre-mineralization skarn stage (Figure 3a), followed by (2) a dominant Sn-mineralization stage with significant retrograde alteration, and finally (3) a subordinate Pb-Zn sulfide stage. Typical minerals of the prograde stage consist of garnet, pyroxene and olivine. The Sn mineralization stage is mainly composed of cassiterite, with retrograde minerals including actinolite, tremolite, chlorite, vesuvianite and quartz. Retrograde alteration often overprinted prograde garnet-bearing assemblages. The Pb-Zn sulfide stage is composed of Pb-Zn sulfides, with minor pyrite, pyrrhotite and arsenopyrite (Figure 3b), as well as coexistent phlogopite, chlorite, fluorite, quartz and calcite ( Figure 3). The ore mineralization is generally zoned in the altered country rocks surrounding the plutons: Sn mineralization occurs in skarns proximal to the granites and Pb-Zn sulfides in more distal areas away from the granites [30]. The cassiterite U-Pb dating and muscovite 40 Ar- 39 Ar dating for the Xianghualing deposit gave two ages of 157 ± 6 Ma and 154.0 ± 1.1 Ma [35,38], respectively; these results are consistent with the zircon U-Pb dating results of the Laiziling intrusion, implying that the magmatism and mineralization could be genetically related and then quickly cooled below the muscovite Ar retention blocking temperature (<350 °C). The Xianghualing deposit is a stratabound skarn Sn deposit [5], and the orebodies are mainly stratiform, pipe-and vein-like in shape [38][39][40]. Field investigations have revealed three major stages of skarn alteration and metal mineralization in the Xianghualing deposit: (1) A prograde pre-mineralization skarn stage (Figure 3a), followed by (2) a dominant Sn-mineralization stage with significant retrograde alteration, and finally (3) a subordinate Pb-Zn sulfide stage. Typical minerals of the prograde stage consist of garnet, pyroxene and olivine. The Sn mineralization stage is mainly composed of cassiterite, with retrograde minerals including actinolite, tremolite, chlorite, vesuvianite and quartz. Retrograde alteration often overprinted prograde garnet-bearing assemblages. The Pb-Zn sulfide stage is composed of Pb-Zn sulfides, with minor pyrite, pyrrhotite and arsenopyrite (Figure 3b), as well as coexistent phlogopite, chlorite, fluorite, quartz and calcite ( Figure 3). The ore mineralization is generally zoned in the altered country rocks surrounding the plutons: Sn mineralization occurs in skarns proximal to the granites and Pb-Zn sulfides in more distal areas away from the granites [30]. The cassiterite U-Pb dating and muscovite 40 Ar- 39 Ar dating for the Xianghualing deposit gave two ages of 157 ± 6 Ma and 154.0 ± 1.1 Ma [35,38], respectively; these results are consistent with the zircon U-Pb dating results of the Laiziling intrusion, implying that the magmatism and mineralization could be genetically related and then quickly cooled below the muscovite Ar retention blocking temperature (<350 • C).

Sampling and Analytical Methods
The samples analyzed in this study were all collected from the Xianghualing mining area ( Figure 2). Generally, the garnet could be divided into two types. The first type garnet is euhedral and larger in size (>3 mm) (Figure 3c-e); the second type is anhedral and much smaller (<100 μm) (Figure 3f), and is clearly overprinted by later retrograde alteration. Based on microscopic observation, neither garnet type has obvious growth zoning. In this study, three garnet samples (XHL-1 and XHL-2 of the first type, and XHL-3 of the second type) have been selected for laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) element analysis.
The analysis was performed at the Ore Deposit and Exploration Centre (ODEC), Hefei University of Technology, Hefei, China, using an Agilent 7900 ICP-MS coupled with a 193 nm ArF excimer Resonetics RESOlution laser-ablation system, named PhotonMachines Analyte HE. Helium (0.90 L/min) and argon (0.87 L/min) were used as the carrier gas and make-up gas, respectively. The two gases were mixed through a T-joint before entering ICP. The laser was operated at 8 Hz, with a laser fluency of 2 J/cm 2 , and a spot size of 30 μm. Each analysis included 20 s of background acquisition (gas blank), 40 s of data acquisition, and again 20 s of background signal, and therefore

Sampling and Analytical Methods
The samples analyzed in this study were all collected from the Xianghualing mining area ( Figure 2). Generally, the garnet could be divided into two types. The first type garnet is euhedral and larger in size (>3 mm) (Figure 3c-e); the second type is anhedral and much smaller (<100 µm) (Figure 3f), and is clearly overprinted by later retrograde alteration. Based on microscopic observation, neither garnet type has obvious growth zoning. In this study, three garnet samples (XHL-1 and XHL-2 of the first type, and XHL-3 of the second type) have been selected for laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) element analysis.
The analysis was performed at the Ore Deposit and Exploration Centre (ODEC), Hefei University of Technology, Hefei, China, using an Agilent 7900 ICP-MS coupled with a 193 nm ArF excimer Resonetics RESOlution laser-ablation system, named PhotonMachines Analyte HE. Helium (0.90 L/min) and argon (0.87 L/min) were used as the carrier gas and make-up gas, respectively. The two gases were mixed through a T-joint before entering ICP. The laser was operated at 8 Hz, with a laser fluency of 2 J/cm 2 , and a spot size of 30 µm. Each analysis included 20 s of background acquisition (gas blank), 40 s of data acquisition, and again 20 s of background signal, and therefore the total time for every spot analysis was 80 s [41]. The external standards were BCR-2G, NIST610, and NIST612, analyzed once every 10 analyses in order to correct the time-dependent drift of sensitivity and mass discrimination. The laser for LA-ICP-MS mapping was operated at 10 Hz, with a sample movement speed of 40 µm/s. The mapping was conducted in line scanning style, with each line being parallel and the size consistent with the laser ablation spot [41]. Standards were analyzed in duplicate at the start and end of each mapping run, which included reference glasses NIST610, NIST612 and garnet standard MON-GT. The total analysis time of mapping was approximately 2 h for each sample [42]. Elements analyzed in this study include Mg, Al, Si, Ca, Ti, Mn, Fe, Cu, Zn, Y, Mo, Sn, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, W, and Pb. The off-line data processing was performed using the ICPMSDataCal program [43]. Trace element compositions were calibrated against multiple-references materials without applying internal standardization. The sum of all element concentrations, expressed as oxides according to their oxidation states, is normalized to be 100 wt % [43].

Major Element Geochemistry
The major element compositions of the garnets from the Xianghualing skarn Sn deposit are summarized in Table 1, and a full list of the data are given in Table S1 in Supplementary Materials. The three analyzed garnet samples share similar compositions ( The data show that the garnets from the Xianghualing deposit generally form from a grossular-andradite solid solution. They are Al-rich and dominated by grossular (60.59-70.75 mol %) and less andradite (18.91-29.76 mol %), with minor amounts of spessartine, pyrope, and almandine accounting for less than 5.5 mol % of the total (Table 1; Figure 4). In general, the contents of grossular and andradite are negatively correlated, but there is no obvious compositional zoning from the core to the rim of an individual garnet grain, although in some grains there are narrow rims surrounding the cores based on the mapping results ( Figure 5). The Xianghualing garnets are similar to other skarn Sn deposits worldwide in composition [44] as shown in Figure 4.
Minerals 2020, 10, x FOR PEER REVIEW 6 of 17 the total time for every spot analysis was 80 s [41]. The external standards were BCR-2G, NIST610, and NIST612, analyzed once every 10 analyses in order to correct the time-dependent drift of sensitivity and mass discrimination. The laser for LA-ICP-MS mapping was operated at 10 Hz, with a sample movement speed of 40 μm/s. The mapping was conducted in line scanning style, with each line being parallel and the size consistent with the laser ablation spot [41]. Standards were analyzed in duplicate at the start and end of each mapping run, which included reference glasses NIST610, NIST612 and garnet standard MON-GT. The total analysis time of mapping was approximately 2 h for each sample [42]. Elements analyzed in this study include Mg, Al, Si, Ca, Ti, Mn, Fe, Cu, Zn, Y, Mo, Sn, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, W, and Pb. The off-line data processing was performed using the ICPMSDataCal program [43]. Trace element compositions were calibrated against multiple-references materials without applying internal standardization. The sum of all element concentrations, expressed as oxides according to their oxidation states, is normalized to be 100 wt % [43].

Major Element Geochemistry
The major element compositions of the garnets from the Xianghualing skarn Sn deposit are summarized in Table 1, and a full list of the data are given in Table S1 in Supplementary Materials. The three analyzed garnet samples share similar compositions ( The data show that the garnets from the Xianghualing deposit generally form from a grossular-andradite solid solution. They are Al-rich and dominated by grossular (60.59-70.75 mol %) and less andradite (18.91-29.76 mol %), with minor amounts of spessartine, pyrope, and almandine accounting for less than 5.5 mol % of the total (Table 1; Figure 4). In general, the contents of grossular and andradite are negatively correlated, but there is no obvious compositional zoning from the core to the rim of an individual garnet grain, although in some grains there are narrow rims surrounding the cores based on the mapping results ( Figure 5). The Xianghualing garnets are similar to other skarn Sn deposits worldwide in composition [44] as shown in Figure 4.

Trace Element Geochemistry
The concentrations of trace elements and rare earth elements of the garnets from the Xianghualing skarn Sn deposit are summarized in Table 2 and listed in Table S1. The three garnet samples show little compositional variations. All garnets contain very low concentrations of Cu (<0.31 ppm), Mo (<0.46 ppm), and Pb (<5.50 ppm), but have higher and variable Sn concentrations from 2314 to 5766 ppm (4295 ppm on average). The concentrations of W vary from 0.22 to 30.0 ppm, Zn from 7.95 to 12.6 ppm, and Pb from 0 to 0.55 ppm.
The rare earth elements (REEs) concentrations of the Xianghualing garnets are all very low. Among them, the average Nd, Sm, Gd, Dy, Er, and Yb contents are generally higher than 1 ppm, whereas the concentrations of all the remaining REE elements (including La, Ce, Pr, Eu, Tb, Ho, Tm, and Lu) are lower than 1 ppm. The REE abundances in garnets range from 6.97 to 30.2 ppm, with an average of 14.7 ppm. Both LREE/HREE and (La/Yb) N ratios are generally lower than 1, ranging from 0.01 to 0.75 and from 0 to 0.04, respectively ( Table 2). The garnets display obvious negative Eu anomalies (Eu/Eu* of 0.15-0.79), except for one spot that has a Eu/Eu* of 2.37. Similarly, Ce also shows negative anomalies (Ce/Ce* = 0-0.71).

REE Patterns in Garnet and Their Significance
The REE enrichments in skarn minerals like garnet are controlled by several factors, including the REE concentrations in the source magmas and/or wallrocks (typically carbonate), the REE concentrations in the hydrothermal fluids forming the skarns, and the mobility of REEs (i.e., their ability to be released during hydrothermal alteration) [45,46]. In addition, several physicochemical parameters of the hydrothermal fluids responsible for the skarn formation, including salinity, pH, f O2 , X CO2 , pressure, and temperature, have also been suggested to be critical in the REE incorporation into garnets [47][48][49]. As shown in Figure 6, the chondrite-normalized REE patterns [50] of the garnets from the Xianghualing skarn Sn deposit indicate that the garnets are generally enriched in heavy REE (HREE) relative to light REE (LREE), with pronounced negative Eu anomalies. This is also similar to the garnets from many other Sn-bearing skarn deposits in China, such as the Shizhuyuan skarn W-Sn-Mo-Bi deposit [51,52] and the Jinchuantang skarn Sn-Bi deposit [53] in Hunan Province, the Gejiu skarn Sn deposit in Yunnan Province [54], and the Huanggangliang skarn Sn-Fe deposit in Inner Mongolia [55]. As documented in previous studies, such REE patterns indicate that the incorporation of the REEs into garnets was controlled by crystal chemistry through isomorphic substitution [19,53], and that the HREEs with smaller ionic radii are more compatible to be accommodated in the octahedral coordination site of garnet lattice than the LREEs with larger ionic radii [56,57]. The REEs in garnets of the Xianghualing deposit display similar patterns and therefore are believed to be controlled by the crystal chemical structure.
The obviously negative Eu anomalies of the garnets from the Xianghualing deposit imply a relatively low oxygen fugacity of the mineralizing fluid. Under reduced conditions, unlike other REEs that only have +3 valence (and +4 valence for Ce), Eu can also occur as Eu 2+ . The reduced Eu 2+ (1.25 Å) can less preferentially substitute for Ca 2+ (1.12 Å) than the oxidized cation Eu 3+ (1.07 Å) in the octahedral position [58,59]. As a consequence, negative Eu anomalies would be generated in a reduced environment. On the contrary, many Cu-rich skarn deposits formed in oxidized conditions [1,60], and as a result, they show typical positive Eu anomalies (Figure 7). To sum up, the negative Eu anomalies of the Xianghualing garnets indicate a reduced condition of the magmatic systems and their exsolved hydrothermal fluids. This is consistent with many other hydrothermal Sn deposits in China and worldwide, which are commonly related to highly reduced granitic system [1,60,61].
concentrations in the hydrothermal fluids forming the skarns, and the mobility of REEs (i.e., their ability to be released during hydrothermal alteration) [45,46]. In addition, several physicochemical parameters of the hydrothermal fluids responsible for the skarn formation, including salinity, pH, fO2, XCO 2 , pressure, and temperature, have also been suggested to be critical in the REE incorporation into garnets [47][48][49]. As shown in Figure 6, the chondrite-normalized REE patterns [50] of the garnets from the Xianghualing skarn Sn deposit indicate that the garnets are generally enriched in heavy REE (HREE) relative to light REE (LREE), with pronounced negative Eu anomalies. This is also similar to the garnets from many other Sn-bearing skarn deposits in China, such as the Shizhuyuan skarn W-Sn-Mo-Bi deposit [51,52] and the Jinchuantang skarn Sn-Bi deposit [53] in Hunan Province, the Gejiu skarn Sn deposit in Yunnan Province [54], and the Huanggangliang skarn Sn-Fe deposit in Inner Mongolia [55]. As documented in previous studies, such REE patterns indicate that the incorporation of the REEs into garnets was controlled by crystal chemistry through isomorphic substitution [19,53], and that the HREEs with smaller ionic radii are more compatible to be accommodated in the octahedral coordination site of garnet lattice than the LREEs with larger ionic radii [56,57]. The REEs in garnets of the Xianghualing deposit display similar patterns and therefore are believed to be controlled by the crystal chemical structure.
The obviously negative Eu anomalies of the garnets from the Xianghualing deposit imply a relatively low oxygen fugacity of the mineralizing fluid. Under reduced conditions, unlike other REEs that only have +3 valence (and +4 valence for Ce), Eu can also occur as Eu 2+ . The reduced Eu 2+ (1.25 Å) can less preferentially substitute for Ca 2+ (1.12 Å) than the oxidized cation Eu 3+ (1.07 Å) in the octahedral position [58,59]. As a consequence, negative Eu anomalies would be generated in a reduced environment. On the contrary, many Cu-rich skarn deposits formed in oxidized conditions [1,60], and as a result, they show typical positive Eu anomalies (Figure 7). To sum up, the negative Eu anomalies of the Xianghualing garnets indicate a reduced condition of the magmatic systems and their exsolved hydrothermal fluids. This is consistent with many other hydrothermal Sn deposits in China and worldwide, which are commonly related to highly reduced granitic system [1,60,61].  [53], Shizhuyuan [51] and Huanggangliang [55], which have been shown here for comparison. All data were normalized to the C1 chondrite values [50].  [53], Shizhuyuan [51] and Huanggangliang [55], which have been shown here for comparison. All data were normalized to the C1 chondrite values [50]. Figure 7. Garnet Eu/Eu* ratios of Xianghualing and several other skarn deposits. Note: The compiled skarn Sn-bearing deposits include Jinchuantang skarn Sn-Bi deposit [53] and Shizhuyuan skarn W-Sn-Mo-Bi deposit [51]. The Cu skarns include Hongniu-Hongshan [19] and Yongping [62] skarn Cu deposits.

Substitution of Sn into Garnet
Garnet has a general chemical formula X3Y2Z3O12, where X represents divalent cations in eightfold coordination, Y represents trivalent cations in octahedral coordination, and Z is dominantly Si in tetrahedral coordination [63,64]. According to the results listed in Table 1, the Xianghualing garnet belongs to the series of grossular-andradite solid solution. It can be inferred that the X position of the eightfold coordination is mainly Ca 2+ , containing a small amount of Fe 2+ , Mn 2+ , Mg 2+ , and other divalent cations (e.g., REEs), and the Y position is mainly occupied by Fe 3+ and Al 3+ .
The incorporation of trace elements into garnet crystals, like the REEs proposed above, is essentially controlled by crystal chemistry [65]. According to the Goldschmidt's first rule, the ionic radius and charge mismatch between the host cation and its substituent, at a given site, exert the principal control on substitution mechanisms [19,60]. The ionic radius of Sn 4+ is 0.69 Å, which is close in size to Fe 3+ (0.67 Å) in the octahedral site of garnet [60], making it possible for Sn 4+ to substitute Fe 3+ . In this study, there is a generally positive correlation between the content of Fe and Sn in Xianghualing garnets based on the mapping results ( Figure 5). This suggests that Sn 4+ is likely to occupy the octahedral sites of garnet through replacement of Fe 3+ . The possible mechanism for such a substitution (Equation (1)) could be: Here □ is a divalent cations-site vacancy in the octahedral site [66].
It should be noted that the above substitution of Sn 4+ for Fe 3+ requires not only a modification of the crystal structure, but also charge balance. The charge difference between quadrivalent Sn and trivalent Fe makes it necessary to get a compensatory mechanism to maintain charge balance. As a result, two other cosubstitution mechanisms (Equations (2) and (3)) have also been proposed [52,[67][68][69]: (2) Figure 7. Garnet Eu/Eu* ratios of Xianghualing and several other skarn deposits. Note: The compiled skarn Sn-bearing deposits include Jinchuantang skarn Sn-Bi deposit [53] and Shizhuyuan skarn W-Sn-Mo-Bi deposit [51]. The Cu skarns include Hongniu-Hongshan [19] and Yongping [62] skarn Cu deposits.

Substitution of Sn into Garnet
Garnet has a general chemical formula X 3 Y 2 Z 3 O 12 , where X represents divalent cations in eightfold coordination, Y represents trivalent cations in octahedral coordination, and Z is dominantly Si in tetrahedral coordination [63,64]. According to the results listed in Table 1, the Xianghualing garnet belongs to the series of grossular-andradite solid solution. It can be inferred that the X position of the eightfold coordination is mainly Ca 2+ , containing a small amount of Fe 2+ , Mn 2+ , Mg 2+ , and other divalent cations (e.g., REEs), and the Y position is mainly occupied by Fe 3+ and Al 3+ .
The incorporation of trace elements into garnet crystals, like the REEs proposed above, is essentially controlled by crystal chemistry [65]. According to the Goldschmidt's first rule, the ionic radius and charge mismatch between the host cation and its substituent, at a given site, exert the principal control on substitution mechanisms [19,60]. The ionic radius of Sn 4+ is 0.69 Å, which is close in size to Fe 3+ (0.67 Å) in the octahedral site of garnet [60], making it possible for Sn 4+ to substitute Fe 3+ . In this study, there is a generally positive correlation between the content of Fe and Sn in Xianghualing garnets based on the mapping results ( Figure 5). This suggests that Sn 4+ is likely to occupy the octahedral sites of garnet through replacement of Fe 3+ . The possible mechanism for such a substitution (Equation (1)) could be: Here is a divalent cations-site vacancy in the octahedral site [66].
It should be noted that the above substitution of Sn 4+ for Fe 3+ requires not only a modification of the crystal structure, but also charge balance. The charge difference between quadrivalent Sn and trivalent Fe makes it necessary to get a compensatory mechanism to maintain charge balance. As a result, two other cosubstitution mechanisms (Equations (2) and (3)) have also been proposed [52,[67][68][69]: As shown in Figure 8a, however, the atoms of divalent iron and the content of Sn do not show any positive correlations, indicating that the cosubstitution (Equation (2)) might be weak, although it cannot be fully ruled out. As for the Mg-Sn cosubstitution (Equation (3)), because of the relatively low Mg concentrations of the garnets in Xianghualing (Table 1) and the poor correlations between the Mg 2+ and Sn content (Figure 8b), it is also unlikely to be the major diadochic substitution mechanism.
Minerals 2020, 10, x FOR PEER REVIEW 11 of 17 Sn 4+ + Mg 2+ = 2Fe 3+ As shown in Figure 8a, however, the atoms of divalent iron and the content of Sn do not show any positive correlations, indicating that the cosubstitution (Equation (2)) might be weak, although it cannot be fully ruled out. As for the Mg-Sn cosubstitution (Equation (3)), because of the relatively low Mg concentrations of the garnets in Xianghualing (Table 1) and the poor correlations between the Mg 2+ and Sn content (Figure 8b), it is also unlikely to be the major diadochic substitution mechanism.  Based on the above observations, it is concluded that the incorporation of Sn into prograde skarn garnets in the Xianghualing deposit are mainly through the substitution 3Sn 4+ + = 4Fe 3+ .
Interestingly, as shown in the mapping results ( Figure 5), the concentrations of Sn and W in the garnets are also positively correlated. This has also been confirmed by the LA-ICP-MS analysis results as shown in Figure 8c. This implies that the mechanism for W substitution into the garnet may be similar to that of Sn.
It should be noted that the ability of Sn substitution into garnet is controlled by the Sn contents in hydrothermal fluids and the partitioning coefficient between fluids and garnet. The partitioning coefficient is further related to the Sn availability in the hydrothermal system; the mineral composition into which Sn substitutes [16,70]; as well as the fluid temperature, pH, salinity, pressure, and the chlorine and possibly fluorine contents [70][71][72]. Based on the microthermometric data for primary fluid inclusions from the different Sn-bearing deposits [73][74][75][76], including Xi'ao skarn Cu-Sn deposit, Huanggang skarn Fe-Sn deposit, Shizhuyuan skarn W-Sn deposit and Xianghualing skarn Sn deposit, it can be concluded that they have similar formation temperatures of garnet (mainly 400-500 • C), salinities (mainly around 10 wt % NaCl equivalent) and pressures (around 500 bars). Hence, it is reasonable to conclude that these parameters of the Sn-bearing hydrothermal fluids have similar effects on the partitioning coefficient of Sn between fluids and garnets. Therefore, the ability of Sn substitution into garnet largely depends on the concentrations of Sn in the fluids.

Exploration Implications
The garnets of the Xianghualing deposit contain relatively high Sn concentrations (>2000 ppm), especially for the sample XHL-2-2, which reaches up to >5700 ppm (Table 2). Similarly high Sn concentrations in garnets have also been observed in other skarn Sn deposits from previous studies. The garnet from a skarn Sn polymetallic deposit in Finland contains up to 1.44 wt % SnO 2 [77]. Also, the content of SnO 2 in garnets at Doradilla, New South Wales, Australia is as high as 0.5-3.5 wt % [68]. The studies on several Sn-bearing skarn deposits in South China (e.g., Yaogangxian, Dachang, Xinlu, Debao, Gejiu, and Shizhuyuan) also show that the contents of SnO 2 in garnets are generally in the range of 0.02-2 wt %, particularly in Gejiu, where the highest value can be up to 5.14 wt % [51,52]. For garnets from skarn deposits containing both Sn and W, including the Baiganhu skarn W-Sn deposit and the Huangshaping skarn W-Sn deposit, the LA-ICP-MS analysis results show that both are rich in Sn (Figure 9), with a maximum of 4990 ppm and 8364 ppm, respectively [23,65]. Combined with the results of this study, it can be generally concluded that the garnets in Sn-bearing skarn deposits are commonly enriched in Sn. A reasonable explanation given by Zhou et al [23] is that the substitution of Sn 4+ for Fe 3+ can more readily take place when the hydrothermal fluids responsible for the Sn mineralization contain elevated concentrations of Sn, regardless of other factors, such as redox state, pH, and temperature conditions [23]. As mentioned above, the abundance of Sn in garnet largely depends on the Sn concentrations in the hydrothermal fluids. Accordingly, the Sn content of garnet can be used to evaluate the potential of Sn mineralization in a skarn system.
As for the two skarn deposits containing both economic Sn and W (Baiganhu and Huangshaping), the W concentrations in garnets are also very high, although variable: the range is between <1 to 2092 ppm [23,65]. For the Weondong skarn W deposit (without Sn) from South Korea [78], the garnet W concentrations (5-760 ppm) are as high as in these Sn-W skarn deposits, and significantly higher than the Xianghualing skarn Sn deposit ( Figure 9). However, the Sn concentrations (32-228 ppm) in Weondong garnets are lower than these Sn-bearing skarn deposits (Figure 8). These observations indicate that skarn Sn deposits commonly contain Sn-rich, but W-poor garnets, whereas skarn W deposits commonly contain W-rich, but Sn-poor garnets; for skarn deposits containing both Sn and W, the garnets may be enriched in both the two ore-forming elements. This might be applied in skarn occurrences where exploration is ongoing to reveal their Sn and/or W mineralization potential. Figure 9. Binary plots of Sn vs. W concentrations of garnets from different deposits. Note: skarn W-Sn deposits include Baiganhu [23] and Huangshaping [65]. The data of the Weondong skarn W deposit are from [78].

Conclusions
(1) Garnets in the Xianghualing skarn Sn deposit in South China belong to typical grossular-andradite solid solution. They are characterized by HREE enrichment, LREE depletion and negative Eu anomalies, indicative of a reduced magmatic-hydrothermal environment. The enrichment of Sn in garnet is mainly controlled by the substitution into the Fe 3+ octahedral site.
(2) Garnets from skarn Sn deposits like Xianghualing are generally Sn-rich and W-poor, whereas those from skarn W deposits (e.g., Weondong) are W-rich and Sn-poor. Interestingly, the Sn and W concentrations in garnets from skarn Sn-W deposits can both be very high. Such observations are potentially useful in guiding exploration for these skarn occurrences without defined Sn and/or W resources to evaluate their potentials to form economic Sn and (or) W mineralization.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Table S1: Summary of the major element compositions (in wt %) and trace and rare earth elements compositions (ppm) of the Xianghualing garnets.
Author Contributions: Q.S. designed the project; F.Y., L.L., Q.Z., and W.Y. collected the samples; F.Y., X.N., and K.X. performed the LA-ICP-MS analyses and reduced the data; F.Y. and Q.S. wrote and organized the paper, with a careful discussion and revision by D.R.L. All authors discussed the results and evaluated the data. All authors have read and agreed to the published version of the manuscript.   [23] and Huangshaping [65]. The data of the Weondong skarn W deposit are from [78].

Conclusions
(1) Garnets in the Xianghualing skarn Sn deposit in South China belong to typical grossular-andradite solid solution. They are characterized by HREE enrichment, LREE depletion and negative Eu anomalies, indicative of a reduced magmatic-hydrothermal environment. The enrichment of Sn in garnet is mainly controlled by the substitution into the Fe 3+ octahedral site.
(2) Garnets from skarn Sn deposits like Xianghualing are generally Sn-rich and W-poor, whereas those from skarn W deposits (e.g., Weondong) are W-rich and Sn-poor. Interestingly, the Sn and W concentrations in garnets from skarn Sn-W deposits can both be very high. Such observations are potentially useful in guiding exploration for these skarn occurrences without defined Sn and/or W resources to evaluate their potentials to form economic Sn and (or) W mineralization.
Author Contributions: Q.S. designed the project; F.Y., L.L., Q.Z., and W.Y. collected the samples; F.Y., X.N., and K.X. performed the LA-ICP-MS analyses and reduced the data; F.Y. and Q.S. wrote and organized the paper, with a careful discussion and revision by D.R.L. All authors discussed the results and evaluated the data. All authors have read and agreed to the published version of the manuscript.