Origin of S-, A- and I-Type Granites: Petrogenetic Evidence from Whole Rock Th/U Ratio Variations

: The origin and evolution of granites remain a matter of debate and several approaches have been made to distinguish between different granite types. Overall, granite classiﬁcation schemes based on element concentrations and ratios, tectonic settings or the source rocks (I-, A-, S-type) are widely used, but so far, no systematic large-scale study on Th/U ratio variations in granites based on their source or tectonic setting has been carried out, even though these elements show very similar behavior during melting and subsequent processes. We therefore present a compiled study, demonstrating an easy approach to differentiate between S-, A- and I-type granites using Th and U concentrations and ratios measured with a portable gamma ray spectrometer. Th and U concentrations from 472 measurements in S- and I-type granites from the Variscan West-Bohemian Massif, Germany, and 78 measurements from Neoproterozoic A-type Malani granites, India, are evaluated. Our compendium shows signiﬁcant differences in the average Th/U ratios of A-, I-and S-type granites and thus gives information about the source rock and can be used as an easy classiﬁcation scheme. Considering all data from the studied A-, I- and S-type granites, Th/U ratios increase with rising Th concentrations. A-type granites have the highest Th/U ratios and high Th concentrations, followed by I-type granites. Th/U ratios in S- to I-type granites are lower than in A-type and I-type granites, but higher than in S-type granites. The variation of Th/U ratios in all three types of granite cannot be explained by fractional crystallization of monazite, zircon and other Th and U bearing minerals alone, but are mainly due to source heterogeneities and uranium mobilization processes.

The ongoing discussion shows that it is difficult to decipher the petrogenesis of granite formations. Obviously, there are several variables that control the magma composition, such as pressure, temperature, volatiles, source rocks and potentially tectonic setting. To make things even more complicated, the magma composition is subsequently modified by assimilation, fractional crystallization and mixing processes. A reliable tracer that is not altered by subsequent processes after the melting process would provide information about the source rock of the granite and the petrogenetic processes leading to magma generation. A very common method found in the literature is the use of tectonic discrimination diagrams in which trace element concentrations and their ratios and Rb/Sr and Sm/Nd soluble hexavalent U species [56]. Bea et al. [57] showed, in a study about metapelites in the Ivra-Verbano Zone, that U was compared to Th enhanced into the fluid phase that was released at the end of crystallization of leucosomes. On the other hand, studies from Variscan I-and S-type granites from central Spain show that Th/U ratios can also be affected by fractionation [47,48], as the Th concentrations are influenced by this process.
So far, a systematic global study on Th and U concentrations and ratios in order to distinguish between different types of granites has not been carried out. This is therefore the first systematic Th and U study of S-, A-, and I-type granites on a larger scale, including 550 measurements. Here we present data from the western Bohemian Massif comprising I-type to S-type granites and NW India with vast exposure of A-type granites. We here show that average Th/U ratios vary significantly between S-, I-, and A-type granites and that Th/U ratios and concentrations can be used to distinguish between petrogenetic processes leading to the formation of granitic rocks. This implies that the variation of Th/U ratios in all three types of granite cannot be explained by fractional crystallization of monazite, zircon and other Th-and U-bearing minerals alone, but is mainly due to source heterogeneities and additionally, in some S-type granites due to uranium mobilization processes.

I-and S-Type Variscan Granites along the Western Bohemian Massif
The granites studied here are representative for Variscan intrusions [58,59] and cover the Fichtelgebirge (Saxothuringian Zone), the Oberpfalz area (transition zone between Saxothuringian and Moldanubian Zone) and the Bavarian Forest (Moldanubian Zone) (Figure 1). For a wider context of the European Variscan chains, see e.g., [57,[60][61][62][63]. The granites intruded in late-Variscan time expanding a period of about 40 m.y. [59,64,65] whereby the geochemical characteristics allow a subdivision into two granite suites, known as older intrusive complex (OIC) and younger intrusive complex (YIC), respectively. The Fichtelgebirge and Oberpfalz granites are peraluminous S-to I-type granites that formed through anatexis of metasediments in middle or deep levels of the crust.
Quartz-monzodiorites (locally named redwitzites) are the precursors of the granites in the Fichtelgebirge and Oberpfalz area. They are predominately metaluminous and interpreted as I-type granites, which represent melting of crustal rocks triggered and mixed with mantle-derived melts [59,64,66]. These diorites represent the initial phase of the late orogenic intrusions at about 323 Ma [64].
Zircon geochronology in NE Bavaria indicates simultaneous intrusion of the diorites and the spatially associated older granites in regions 1 and 2 [64]. Co-existence of the diorites and the older granites as contemporaneous but different magmas may suggest that there was also a link between melt extraction from the mantle, heat input into the crust, and the generation of diorites and larger amounts of well-mixed homogeneous older granites melts [64]. Formation of the younger granites, at 328-320 Ma, 315-312 Ma and 298 Ma, marks the time when pure, intracrustal, acidic S-type magmas were produced [65]. The granites are classified as monzogranites and their geochemical characteristics reach from weakly peraluminous biotite-granites to moderate and strongly peraluminous ± biotite muscovite granites, e.g., [65].
The Bavarian Forest (region 3) is part of the Moldanubian Zone and high-grade regional metamorphism is recorded in this internal part of the Variscan orogeny (~340 Ma, [67]). In contrast to the above described regions 1 and 2, the Bavarian Forest exposes a deeply eroded middle to lower continental crust [68]. Depth of granite intrusion has been constrained at 16 to 18 km for the Hauzenberg Pluton [69]. Regional high temperature metamorphism associated with migmatization and anatexis of Moldanubian rock associations in the Bavarian Forest is constrained at 326-322 Ma [68,70] and formation of major granite complexes are related to these processes. This includes the Hauzenberg and Fürstenstein intrusive complexes. Granodiorites are similar to the diorites of the Fichtelgebirge and are I-type granitoids, while granites from these complexes are classified as S-type granites [69,71].
The granites of this area have been the subject of various geochemical studies (e.g., [72,73]). Compared with the granites from the worldwide granite database GRAN-ITE2017 [23], the studied granites are typical for S-, I-and S-to I-type granites. Data on radioelement concentration have been obtained from in situ gamma radiation measurements, reported in previous studies [74,75].
The Malani granites were formed by the anatexis of an older Neoproterozoic granitic crust (Erinpura granites, see [82]). The granites are peraluminous to metaluminous and classified as A-type (e.g., [83]). Compared with the granites from the worldwide granite database GRANITE2017 [23], the Malani granites are typical A-type.
Felsic rocks of the Malani Igneous Suite are well known for their high radiogenic heat production [84]. The Malani granites are characterized by high but varying Th and U concentration [85]. Some intrusions such as granites forming the Mt. Abu batholith reach Th content of >100 ppm and U of up to 20 ppm [79,86].

Methods
We use Th and U concentrations measured with a portable gamma ray spectrometer RS 230 manufactured by Radiation Solutions, Ontario, Canada. It uses a bismuth germinate (BGO) scintillation detector. Individual data (550 samples) and coordinates of locations are all recorded in our previous published papers [74,75], but can also be found in Table S1. We provide statistical evaluation in Table S2 and statistical analysis using the Mann-Whitney pairwise U-test with Bonferroni correction (Past 4.01) in Table S3 (see Supplementary Materials). Measurements were taken in the "assay-mode" over a 120 s interval. The device integrates an area of about 1 m in diameter with a depth of about 15-25 cm when in direct contact with the outcrop. The manufacturer used concrete pads of 1 m × 1 m × 30 cm size containing known amounts of potassium, uranium and thorium and a fourth pad with none of these three elements to measure the background for calibration of the instrument, as described in [87]. Possible cross interferences can occur, as error components for a single element will also influence the remaining elements (Radiation Solutions, Ontario, Canada) [87]. For a detailed description and comparison with ICP-MS analyses, see [74].

Malani Granites
The concentrations of the Malani granites vary between 20 and 100 ppm and U concentrations vary between 1 and 20 ppm (Figure 3a,b). Average Th and U concentrations of the Malani granites are 47.7 ppm ± 2.1 for Th and 7.6 ppm ± 0.4 for U. Mt. Abu Malani A-type granites show higher average Th and U concentrations with 68.5 ppm ± 5.1 for Th and 8.5 ppm ± 0.6 for U, respectively.

Variscan Granites
The uranium concentrations of the Fichtelgebirge (7 ppm to 34 ppm) and the Bavarian Forest (4 ppm to 20 ppm) are quite similar. Compared to the two other regions, the Oberpfalz intrusives have a much greater variation to high values (8-61 ppm U). YIC granites tend to have higher U concentrations than the OIC granites. The Th content in all granites is <40 ppm (Figures 4-6).

Comparison between the Different Types of Granite
In general, the A-type granites tend to have higher Th concentrations compared to the other granite types, followed by I-type granites ( Figure 7). A-, I-and S-type granites all lie on an L-shaped Th versus U concentration line. The Malani A-type granites have the highest Th concentrations and Th/U ratios. Th concentrations are lower in S-to I-type granites compared to A-type and I-type granites, but they have higher Th values than the S-type granites. S-type granites have the highest U and lowest Th concentrations These observations are supported by statistical evaluation (Table S2 in the Supplementary Materials) and statistical analysis, using the Mann-Whitney U-test with Bonferroni correction (Past 4.01) ( Table S3 in the Supplementary Materials). The statistical analysis show that the Th concentrations differ significantly for all granite types at the 95% confidence interval. The U concentrations differ significantly for all granite types except S-I-type and A-type granites at the 95% confidence interval.
S-type granites with highest U and lowest Th concentrations have been shown from the Iberian Variscan granites, too [48].
Calculated averages of Th and U concentrations of A-and I-and I-to S-type granites all fall within the field for Th/U > 1, but the S-type granites lie partly in the Th/U > 1 and partly in the Th/U < 1 field ( Figure 8). Overall, the average Th concentrations vary between 4.5 ppm and 68.5 ppm, the average U concentrations between 2.9-27.8 ppm. The I-type granites plot nearest to the average continental crust [88].   [2,23]. All A-and I-and I-to S-type granites fall within the field for Th/U > 1, but the S-type granites lie partly in the Th/U >1 and partly in the Th/U < 1 field. Data and statistical analysis are given in Tables S1-S3 in the Supplementary Materials.
The Th/U ratios display a large variation from 0.08 to 15 ( Figure 9) and considering all data, Th/U ratios rise with increasing Th concentrations. The medians of the Th/U ratios differ significantly for the different granite types at the 95% confidence level, with the exception of the S-I-type and I-type medians, which only differ at the 90% confidence interval (see Table S3 in the Supplementary Materials). The S-type granites show the lowest Th/U ratios, followed closely by the S-to I-type granites. The I-type granites have higher Th/U ratios and the A-type granites have the highest Th/U ratios. In the Th/U versus U diagram (Figure 10), the samples show a negative line. The S-type granites have higher and quite variable U concentrations with different Th/U ratios. The average continental crust has Th/U ratio of 4 [50] just as the I-type granites. Considering all data, increasing Th concentrations correlate with increasing Th/U ratios. The S-type granites show the lowest average Th/U ratios, followed closely by the S-to I-type granites. The I-type granites have, on average, higher Th/U ratios and the A-type granites, on average, the highest Th/U ratios.

Systematic A-, I-, and S-Type Th/U Ratios and Concentrations in Granites
Our aim is to investigate how granite types of different tectonic settings correlate with Th and U concentrations and Th/U ratios. Artemieva et al. [23] argued that U and Th concentrations as well as Th/U ratios are useful tools for the characterization of granitic rocks. They also showed that the geodynamic environment for the generation of granites might have changed through the early Earth history and speculated that after the Middle Proterozoic, the production of granites occurred in more diverse tectonic settings, which led to an increase in the diversity of Th and U concentrations [23]. Within the continental crust, Th and U concentrations show a large variation [2,51,[53][54][55] and even within rock types, or at pluton scale, the Th and U concentrations vary considerably [41,89]. Adams et al. [90] proposed that differences within rock types are commonly greater than the average differences between the various igneous rock types. The granite database GRANITE2017 [23] reports mean Th concentrations of 14.8 ± 13.2 ppm and mean U concentrations of 3.93 ± 3.27 ppm in granites. However, such Th and U concentrations are significantly lower compared to most granites of this study and we therefore suggest that the Th and U variation is even larger than previously thought (Figure 7). The database GRANITE2017 reports Th/U ratios that are in the range of the predicted global mean ratio of Th/U = (3.7-4.0) for the bulk Earth and primitive mantle [49] and is in agreement with the average Th/U ratio of the continental crust of about four [2,40,[50][51][52]91], with variations reported from individual plutons [2,51,[53][54][55]. Thus, in contrast to the Th and U concentrations, the Th/U ratios in granitoids are believed to be rather constant [41,92], but vary from 3.8 [51] to 5.2-5.3 [53,55]. However, these results rely on a relatively small number of systematic regional Th and U studies, e.g., for some A-type (e.g., [93] and I-and S-type granites (e.g., [30,31], but similar global studies are absent for different granite types and for granitic rocks in general. Exceptions are reported for some I- [46,47] and S-type granites [48], and references therein. This study covers systematic regional Th/U ratio analysis of A-, I-and S-type granites of Malani, Fichtelgebirge, Oberpfalz and the Bavarian Forest. The Th/U ratios of all four regions show a large range between 0.07-12.4 (Figure 3a, Figure 4a, Figure 5a, and Figure 6a). Artemieva et al. [23] proposed that only granites of the Archean and Early Proterozoic have Th/U ratios >10. However, the here presented Malani A-type granites show that granites younger than the Early Proterozoic can also have Th/U ratios >10. A-type granites have higher Th/U ratios and Th concentrations compared to the other granite types. This implies a Th-enriched source, which might have developed during earlier small-scale partial melting processes in the mantle over a prolonged time.
On the other hand, many S-type granites show high U concentrations and low Th/U ratios. In addition, nearly all samples have Th/U ratios >1, except most S-type granites from the Oberpfalz region that lie along a suture zone in which the Th/U ratios <1 are due to U mobilization ( Figure 8). This is associated with carbon-rich sedimentary units and mobilization during hydrothermal processes [94][95][96]. The Flossenbürg intrusion has very high U concentrations. The long axis of the oblong shape of this intrusion parallels late-Variscan faults and such setting might favor the mobilization of U. Fluid mobilization can change the U concentrations, but does not affect the Th concentrations (e.g., [97][98][99]. This is because U is far more mobile because of its increased solubility at higher oxidation states, whereas Th has only one valence state and remains inside the lattice sites of the minerals. Although in S-type granites of the Oberpfalz region, U mobilization is more likely, as hydrothermal processes did take place in wide areas of these intrusions, the involvement of fractional crystallization in some P-rich S-type granites has also been invoked for explaining their low Th/U ratios. The chemical variation of magmatic phosphates of those Variscan granites (monazite, xenotime, see [48], apatite, see [100]) evolves within each granite unit, generating in the most fractionated one the (Th-poorest) U-richest accessories ever described in granites: up to 23 wt% UO 2 in monazite, up to 13 wt% UO 2 in xenotime [48] and up to 460 ppm U in apatite [100].
Thus, large Th/U ratio variation can be explained by different source compositions/tectonic settings, degree of melting, fractional crystallization and late-stage U mobilization. The positive Th/U versus Th trend line together with the U and Th concentrations show that the variation in Th/U ratios is mostly dependent of Th, except in the cases of U mobilization.

Why Do the Th/U Ratios Vary within One Region?
Assuming that granites are not derived from a simple source, each intrusion should show a unique Th/U ratio depending on the source composition. It is therefore inevitable to have a closer look at some individual plutons.
U and Th concentrations of all studied granites-except the U concentrations of Steinwald-are higher than the estimated values for the bulk continental crust (1.3 ppm U and 5.6 ppm Th) [2]. In general, high U and Th concentrations can be explained by the presence of monazite and zircon in granites. Fractionation of monazite lowers both the Th/U ratio and the Th concentration as monazite has high Th concentrations. In contrast, zircon fractionation does not affect the Th concentration much, but lowers the Th/U ratio, as zircon incorporates more U than Th. Monazite hosts more than 80% of Th in all studied rocks [101]. The importance of zircon for Th and U in the rock budget in granites is lower, ranging from 4-26% for U and less than 5% for Th. The study of [101] showed that for the peraluminous granite plutons of the Bohemian Massif granites of similar mineral composition, they contain zircon and monazite of different chemistry and no simple proportional relationship exists between the chemical composition of magma and the chemistry of accessory mineral phases crystallized from the melt. The concentration of minor elements in monazite and zircon are strongly influenced by the order of their crystallization [101]. It is therefore doubtful to only use the chemical composition of accessory minerals for a geochemical interpretation of parental granites.
Malani Granites A-type Malani granites have Th and U concentrations well above the average given by [23] in GRANITE2017 for granites and Th concentrations increase with rising U concentrations (Figure 3a,b). Th/U ratios are mainly controlled by the U contents, as Th/U versus U diagram shows a weak negative data correlation whereas Th/U versus Th only shows a data cluster (Figure 3a,b). These patterns define a typical trend of fractional crystallization in felsic magmas with slightly more zircon fractionation compared to monazite. The correlation of Th and U concentrations indicate that secondary fluid mobilization has not played a major role in the Malani granites.
Mt. Abu Malani granites show higher average Th/U ratios and higher average U and Th concentrations compared to the Malani granites. In the Th/U versus Th diagram, the samples show a slight positive correlation; in the Th/U versus U diagram, the samples show a cluster. This implies that, in this case, the zircon fractionation was not dominant compared to the monazite fractionation. Together with previous research, this suggests that the Mt. Abu Malani granites have the same source as the Malani granites, but a lower degree of melting or a higher degree of fractionation [79]. Both processes would support the model of emplacement of several magma batches to form this large intrusion of approximately 125 km 2 [79,86].

Variscan Granites
The Variscan granites show increasing Th/U ratios with rising Th concentrations and a negative line on the plot of Th/U versus U (Figures 4-6). The positive trend in Th/U versus Th suggests that the sources of the individual plutons in the region vary in Th and U. In addition, the variation in Th/U ratios within individual plutons are then subsequently caused by fractional crystallization. The Th/U versus U plot shows that the U concentration stays over most of the Th/U ratio range relatively low, but at low Th/U ratios, the U concentrations are high. This implies that the Th/U ratios of the Variscan granites are mostly dependent on monazite fractionation, except at low Th/U ratios. Zircon, however, is a major host for U, and thus, fractionation of zircon would also affect the high Th/U ratios. However, the extremely low Th/U ratios in the Oberpfalz granites are mainly due to high U and low Th concentrations. As the Oberpfalz granites have been intruded along fault intersections with the Moldanubian-Saxothuringian suture zone, it is quite possible that the high U concentrations are due to U mobilization. The high U concentrations due to U mobilization have a much higher impact on the depleted samples than a few ppm addition or loss of U would influence samples with higher U and Th concentrations.
The diorites (redwitzites) are less abundant among the igneous rocks in NE Bavaria. They belong to a group of peculiar late-Variscan igneous rocks [102,103]. These diorites have relatively high Th/U ratios and Th concentrations, but quite low U values.
Fichtelgebirge (Region 1) Within the Fichtelgebirge granites (G1 to G4, e.g., [58], Th/U versus Th ratios show a positive correlation, whereas Th/U versus U shows a trend to higher U concentrations with low Th/U ratios (Figure 4a,b). This is in agreement with the interpretation of [104] who suggested that the G1 granites (OIC) show a progressive differentiation trend, which could be explained by enhanced fractionation of monazite in relation to zircon.
Siebel et al. [105] suggested that G1 granites have different source compositions. One model for the origin of G1 is a mixing between mantle-derived magmas and Saxothuringian or Moldanubian paragneisses [105]. The G1 magma could have been formed in a subsequent stage with less or no involvement of mantle-derived material. This interpretation is supported by this study, as G1 granites have higher Th/U ratios that were internally fractionated (Figure 4a,b).
The G2 and G3 (YIC 1) granites have the same pattern as the OIC granites, except that the G3 and Flossenbürg granites partly show low Th/U ratios and high U concentrations that can be explained by U mobilization or/and fractional crystallization. G2 granites show higher Th/U ratios suggesting that they have fractionated less monazite than G3 granites. Hecht [106,107] suggests that G2 and G3 were formed from the same magma that crystallized in two textural varieties in the course of differentiation. It could be that crystallization at the sides leads to a more differentiated magma inward, from which a new batch can be issued. Alternatively, multiple injections emplace successively from a source differentiated at depth.
G4 granites (YIC-2) have high U concentrations with Th/U ratios <1 due to U mobilization and/or fractional crystallization, but lie on a fractionation line together with G2 and G3 granites (Figure 4a,b). Most authors also suggest that the continuous evolution in mineralogy and rock texture, as well as in whole-rock and mineral chemistry from G2&G3 towards G2*&G4 is an indicator for the rocks being related to each other (e.g., [106][107][108][109]. However, Siebel et al. [105] suggested, based on εNd(t) values, that due to a larger scatter of G4 compared to G2&G3, a more heterogeneous source or wall-rock contamination is responsible for the heterogeneity. Hecht et al. [107] and Irber et al. [110] state that within the YIC (G2, G3 and G4), autometasomatism increases with increasing differentiation, culminating in the most differentiated samples of G4 [107,110]. The high U concentrations in the samples definitely support a late stage metasomatism.
Oberpfalz (Region 2) Based on chemical and isotopic composition the granites of Leuchtenberg and Zainhammer, are transitional between S-and I-type, whereas those of Falkenberg, Friedenfels, Steinwald and Flossenbürg are S-type granites [64]. Most intrusions display chemical characteristics, which can be explained by fractional crystallization processes [105,108,111,112]. Leuchtenberg and Zainhammer granites show, on average, higher Th/U ratios than the S-type granites, but mostly lower U concentrations (Figure 5a,b). This suggests that the Zainhammer and Leuchtenberg granites have a source with a considerable higher mantle component and were not or only little affected by U mobilization compared to the S-type granites. Our observation here is supported by Nd isotopic characteristics of the older granitoids as reported by [59,113,114]. These authors suggest a higher degree of contribution from the mantle in the older granites than in the younger granites. Siebel et al. [64] suggested, based on zircon analysis, that magma-generating processes were different for the older granitoids (diorites and older granites) and the two groups of younger granites. The S-type granites of Falkenberg, Friedenfels, Steinwald and Flossenbürg lie on a positive trend plotting Th/U ratios versus Th. In the Th versus U diagram, the samples of each location lie in a separate field. In contrast, Th/U ratios versus U concentrations show that U values stay rather constant, except for high U concentrations (Figure 5a,b). These results show that the different granites show different degrees of fractional fractionation of monazite and in a lesser amount, xenotime and apatite (e.g., [48,100]). In addition, many of the Steinwald and Friedenfels granites were affected by U mobilization. However, high U concentrations and a decrease in Th contents could also be due to differentiation of the magma (e.g., [48]) and then reflected in the Th/U ratios that are <1, on average, in these granites. This is also true for some of the granites from Flossenbürg and Falkenberg. Bea et al. [115,116] shows that the strongly peraluminous and P-rich granites from the Pedrobernado pluton, central Spain, are extremely depleted in Th and U due to monazite and apatite fractionation. We suggest that both fractionation of the magmas and source composition are the dominant signatures in the S-type granites of the Variscan Bohemian granites.
Bavarian Forest (Region 3) According to [117], the Hauzenberg I and II granites belong to the S-type granites which originated by re-melting of the continental crust. The Hauzenberg granodiorite contains monazite, but the Hauzenberg I and II S-type granites do not contain significant monazite. Some of the Hauzenberg diorites contain relatively low Th and U concentrations compared to the other granites. This suggests that monazite is not evenly distributed in the granites.
The Fürstenstein diorites have Sr and Nd isotope ratios, which suggest mantle contribution to melts or melting of young mafic lower crust [105]. The diorites are isotopically similar to the dioritic rocks, from northern Oberpfalz in NE Bavaria [105]. Zircon data indicate that the diorites are contemporaneous with the granites of the Fürstenstein granite massif and the "older granites" in northern Oberpfalz (Leuchtenberg, Zainhammer granites) [59,64].

Average Th/U Ratios of the A-, I-, S-Type Granites
Our data show that the average Th/U ratios for the S-, A-and I-type granites are significantly different, implying that the Th/U ratios are a petrogenetic fingerprint of the sources (Figure 8, Table S3 in the Supplementary Materials). The Th/U ratios of each field overlap with each other and therefore cannot be taken by itself as a definite characteristic. The A-type Malani granites have the highest average Th/U ratios of 6.2 and 8.0, respectively. They are followed by the I-type diorites of the Fichtelgebirge (average Th/U ratio of 4.8) and the Oberpfalz (average Th/U ratio of 6.0) and the diorites from the Bavarian Forest (Hauzenberg average Th/U ratio of 4.6 and Fürstenstein average Th/U ratio of 3.7). Within the Fürstenstein Massif, the high Th concentrations can be explained by a high amount of monazite in this late-stage granite formation [71].
The S-to I-type granites from the Oberpfalz have average Th/U ratios of 3.5 (Zainhammer 2.9 and Leuchtenberg 2.8). The S-type granites from the Fichtelgebirge have average lower Th/U ratios of 2.1 (G2 2.3 and G3 2.0) and the ones from the Oberpfalz of 1.0 (Falkenberg 2.8, Friedenfels 0.7, Steinwald 0.2 and Flossenbürg 0.8); the Hauzenberg granites I and II of 3.2 and 1.8. The tin granite has an average Th/U ratio of 0.5. The Falkenberg granites have high Th/U ratios due to their high Th concentrations. Most granite intrusions studied here are characterized by Th/U ratios >1 that can be explained by enriched sources and additionally, fractionation processes. These are all A-and I-type granites. In contrast, the S-type granites show average Th/U ratios <1 due to high U concentrations. Our results show that it is possible in a first approach to use the Th/U ratios in order to distinguish between different granite types. Fractionation of the magmas does change the Th/U ratios, but the source composition seems to be the dominant control.

Conclusions
This is the first systematic Th and U study of S-, A-and I-type granites on a larger scale (550 measurements) with samples from typical settings for I-and S-type (Variscan granites) and A-type Malani (India) granites. The main results of this work can be summarized as follows: • Average Th/U ratios of A-, I-and S-type granites differ significantly from each other due to the involvement of different sources.

•
The Th/U ratios of all four studied regions show a large range between 0.07-12.4 and therefore, a larger range than assumed so far. • Middle Proterozoic granites are assumed to have Th/U ratios <10. However, this study shows that there are middle Proterozoic granites with Th/U ratios >10 (A-type granites).

•
The A-type Malani granites and the I-type, I-to S-type and S-type Variscan granites from the Fichtelgebirge, Oberpfalz and Bavarian Forest show a large variation in Th and U concentrations, with the A-type granites showing a higher variation in Th than in U, whereas S-type granites vary the most in U.

•
In addition, Th/U ratios within individual regions vary significantly, too. • A-type Malani granites have the highest average Th/U ratios and high Th concentrations, followed by I-type diorites.
• S-to I-type granites have lower average Th/U ratios compared to A-type and I-type granites, but higher values than the Bohemian S-type granites. • A-, I-and S-type granites all lie on a positive Th/U versus Th trend line, showing that the variation in Th/U ratios is mostly dependent of Th, except in the cases of U mobilization.

•
The variation of Th/U ratios in all three types of granite cannot be exclusively explained by fractional crystallization of monazite, zircon and other Th-and U-bearing minerals alone, but in the studied plutons, it is probably mostly due to source heterogeneities and U mobilization processes.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/min11070672/s1, Table S1: Th and U concentrations of the A-type, S-type, I-type and S-I-type granites from the Fichtelgebirge, Oberpfalz, Bavarian Forest and India (550 samples). All measurements and coordinates of locations are recorded in our previous published papers [74,75], Table S2: Statistical evaluation of the data. Shown here the mean, standard deviation (1 sigma) and the mean square error, Table S3: Statistical analysis using the Mann-Whitney U pairwise test with Bonferroni correction (Past 4.01). Medians of the Th/U ratios differ significantly for all granite types on the 95% confidence interval, with the exception of the S-I-type and I-type medians, which only differ on the 90% confidence interval. U concentrations differ significantly for all granite types except the S-I-type and A-type granites. Th concentrations differ significantly for all granite types on the 95% confidence interval. Funding: This research received no external funding.

Data Availability Statement:
The data of this publication has already been published in [74,75,85].