Petrogenesis of Heavy Rare Earth Element Enriched Rhyolite: Source and Magmatic Evolution of the Round Top Laccolith, Trans-Pecos, Texas

: The Round Top rhyolite located in Trans-Pecos Texas is enriched in Be, F, Li, Nb, Rb, Sn, Th, U, Y, Zr, and rare earth elements (REEs). REE-bearing minerals are mainly ubiquitous nano-scale accessory phases throughout the groundmass, incorporated in synchysite-group minerals, xenotime-(Y), Y- and Ce-rich ﬂuorite, and zircon. The rhyolite is peraluminous, high-silica, alkaline (not peralkaline), with elevated heavy rare earth element concentrations and anonymously negative Eu values. Pervasive spongy groundmass and recrystallization textures are consistent with the elevated and remobilized Zr, Th, and Y + HREE (heavy rare earth element) concentrations and a high ﬁeld strength element (HFSE) soluble, sub-alkalic, F-rich, magmatic system. REE-bearing minerals are present as late-magmatic, interstitial phases and attributed with closed-system, post-magmatic, hydrothermal alteration. Petrogenetic modeling provides scenarios that explain the geochemical evolution and REE complexing behavior in evolved rhyolite magmas, and determines possible source compositions and evolution. Trace element models suggest a system typical of having extensive magmatic differentiation. The resulting rhyolite magma is indicative of a silica-rich magmatic system enriched in H 2 O, Li, and/or F that could be considered transitional between pure silicate melt and hydrothermal ﬂuid, where ﬂuorine-ligand complexing was prevalent through late magmatic cooling and crystallization processes. Thorough differentiation and high ﬂuorine activity contributed to the late stage crystallization of REE-bearing minerals in the Round Top rhyolite.


Introduction
The Trans-Pecos region of Texas hosts more than 100 Paleocene to Miocene age intrusive igneous and extrusive bodies related to subduction and later rifting along the southwest margin of North America. Many of the localities, including Round Top and the other laccoliths of the Sierra Blanca Complex (SBC) are hypabyssal, sub-volcanic, and rhyolitic in composition. Despite the numerous magmatic rocks in the Trans-Pecos area and contiguous regions, only a few have been analyzed for rare earth element (REE) geochemistry and recognized to have REE mineral occurrences.
Several sills and dikes outcrop around and within the SBC laccoliths. These intrusions precede the main laccolith magmatism and are composed of hypabyssal hornblende andesite porphyry, porphyritic diorite, and latite porphyry dikes [17][18][19][20]. Diorite to andesitic dikes and sills from around Sierra Blanca and Round Top were dated at 40.6 ± 0.1 Ma (K/Ar age; [19]). A sill outcropping along the northern margin of Round Top is hornblende diorite porphyry and sits unconformably beneath the laccolith. Parts of the sill and surrounding limestone are partially brecciated and fluoritized. Whether the dikes and sills have any relationship to the Round Top mineralization is unclear, but it is probable that the same magmatic conduits that emplaced the sills were used during emplacement of SBC laccoliths.

Materials and Methods
More than 350 samples were collected from Round Top, Little Round Top, and Sierra Blanca. Samples were collected from outcrops, road cuts and drill cuttings obtained during exploratory drilling by Texas Mineral Resources Corporation. Petrographic work on more than 40 thin sections was conducted using petrographic microscopes. Thin sections were made to the standard thickness  [16]). Inset map of Texas shows the location of the SBC (black square) and the extent of the Trans-Pecos Magmatic Province (red shapes; adapted from [16]) in west Texas.
Magmatism in the region occurred between 100 and 40 Ma ago, related with the subduction of the Farallon plate under the North American continent. Bimodal volcanism extended east from the west coast subduction zone [3,12]. Trans-Pecos magmatism occurred  Ma and is separated into early and later phases [3,6,7,13]. The early phase  is characterized by voluminous silicic to mafic igneous intrusions, lava flows, and caldera eruptions [3,6]. The late phase  was compositionally variable, ranging from alkali-calcic to alkalic, and represents the most active magmatism in the region [3,5,6,10]. The Sierra Blanca laccolith, adjacent to and associated with Round Top, was emplaced during the main phase of bimodal TPMP activity at 36.2 ± 0.6 Ma (K/Ar in biotite; [4]). Volcanism during this period is characterized by silicic to mafic intrusions, and large-volume caldera eruptions associated with voluminous ash-flow tuffs [3,6,8]. Basin and Range extension was initiated by a change in the region's stress around 31 Ma, and extensive normal faulting initiated around 24 Ma [3]. Mafic volcanism in the early stages of rifting has been dated at 17 Ma [3].

Sierra Blanca Complex
The Sierra Blanca Complex comprises five separate rhyolite and rhyolite porphyry laccoliths. SBC laccoliths are aphanitic to slightly porphyritic, and are located in a 90 km 2 area about 18 km southeast of the Finley Mountains (47 Ma) and 5 km north of the Quitman Mountains (35 Ma; [6]). The largest laccolith, Sierra Blanca, was emplaced 36.2 ± 0.6 Ma [6]. Emplacement ages of the other SBC laccoliths have not been measured, but a relative emplacement sequence can be interpreted through major-and trace-element differentiation. The increasing enrichment of REEs, especially the increasing ratio of HREEs to LREEs (light rare earth elements) between laccoliths, suggests that Triple Hill was emplaced first, followed by Sierra Blanca, Round Top, Little Sierra Blanca, and Little Round Top [3,[14][15][16]. The slight differences in compositions and rapid cooling textures suggest the timing between emplacement and duration of emplacement was relatively short.
Sierra Blanca is the largest of the SBC laccoliths. The laccolith is pink to gray in color, chemically and texturally homogenous and composed of quartz rhyolite porphyry. Alteration of the rhyolite is not pervasive compared to Round Top, but hematitic and manganese oxide staining is evident throughout jointed and faulted portions of the laccolith. Round Top is the most westerly of the SBC ( Figure 1) and comprise four variations of rhyolite that differ slightly by color, mineralogy, and texture. The rhyolite is fragmented, heavily jointed, blocky, and locally cemented by multiple generations of fluorite, calcite, and quartz.
Several sills and dikes outcrop around and within the SBC laccoliths. These intrusions precede the main laccolith magmatism and are composed of hypabyssal hornblende andesite porphyry, porphyritic diorite, and latite porphyry dikes [17][18][19][20]. Diorite to andesitic dikes and sills from around Sierra Blanca and Round Top were dated at 40.6 ± 0.1 Ma (K/Ar age; [19]). A sill outcropping along the northern margin of Round Top is hornblende diorite porphyry and sits unconformably beneath the laccolith. Parts of the sill and surrounding limestone are partially brecciated and fluoritized. Whether the dikes and sills have any relationship to the Round Top mineralization is unclear, but it is probable that the same magmatic conduits that emplaced the sills were used during emplacement of SBC laccoliths.

Materials and Methods
More than 350 samples were collected from Round Top, Little Round Top, and Sierra Blanca. Samples were collected from outcrops, road cuts and drill cuttings obtained during exploratory drilling by Texas Mineral Resources Corporation. Petrographic work on more than 40 thin sections was conducted using petrographic microscopes. Thin sections were made to the standard thickness of 30-microns. Microbeam analyses were conducted in the University of Austin at Texas, Department of Geological Sciences (Austin, TX, USA), Electron Microbeam Laboratories on a JEOL JXA-8200 Electron Microprobe.
Samples from Round Top were analyzed at a commercial facility (ACMELABS, Vancouver, BC, Canada; see Appendix B for blind duplicate analyses). Samples of~10 kg from each locality were broken into smaller unweathered pieces and combined into a single sample of~5 kg. Each sample was jaw-crushed, and splits pulverized in a carbon-steel ball-mill. Major elements were prepared with LiBO 2 fusion and analyzed by X-ray fluorescence (XRF). Detection limits for major elements were 0.01 wt %, and 0.1% loss on ignition (LOI). For trace and rare earth elements, a multi-acid digestion was conducted where 0.25 g was heated in HNO 3 -HClO 4 -HF and dried. The residue was dissolved in HCl and analyzed by inductively couple plasma mass spectroscopy (ICP-MS). Detection limits for SiO 2 , Al 2 O 3 , MgO, CaO, Na 2 O, K 2 O, TiO 2 , P 2 O 5 , Li, FeO and MnO were 0.01 wt %; Fe 2 O 3 was 0.04 wt %; Cr 2 O 3 was 0.02 wt %. Detection limits for Sc, Ba, Be, Sn, and Zn were 1 ppm; Co and Th were 0.2 ppm; Cs, Hf, Nb, Rb, Ta, U, Zr, Y, La, Ce, Mo, Cu, Pb, Ni, Cd, Sb, Bi, Ag and Tl were 0.1 ppm; Ga, Sr, W, As, Au and Se were 0.5 ppm; V was 8 ppm; Pr, Eu and Ho were 0.02 ppm; Nd was 0.3 ppm; Sm, Gd, Dy and Yb were 0.05 ppm; Tb, Tm, Lu and Hg were 0.01 ppm; Er was 0.03 ppm. F was analyzed separately by specific ion electrode, with a detection limit of 0.01 vol %.

Mineralogy
Phenocryst and groundmass typically comprise equal proportions of the rhyolite (50% phenocrysts, 50% groundmass; Figure 2). Plagioclase feldspar comprises less than 15% of the rhyolite as subhedral to euhedral, rectangular (smaller than 0.5 × 1.2 cm), albite lathes, commonly with K-feldspar overgrowths. These lathes occur in clusters, as penetrating twins, or as individual phenocrysts, and occasionally K-feldspar mantles have their own albite overgrowths. The albite before the potassium feldspar overgrowth is typically equant without reaction textures, indicating a lack of dissolution before the formation of the K-feldspar. Potassium feldspar constitutes about 45-55% of the rhyolite and is found as 0.75-2 cm phenocrysts with hourglass sector zonation, or as mantles around early formed sodic plagioclase ( Figure 3). Quartz comprises about 25-35% of the rhyolite, occurring as zoned, anhedral to subhedral 0.2-1.5 cm phenocrysts. The core is typically equant, euhedral to subhedral, early beta quartz, pseudomorphed to trigonal, low temperature, alpha quartz. Quartz typically exhibits pseudo-concentric rings (snowball texture), where repeated corrosion and crystallization of later quartz overgrowths occurred. Biotite is annitic and comprises less than 5% of the rhyolite. The annite occurs as interstitial subhedral to anhedral phenocrysts less than 1 mm in size, displaying moderate pleochroism, alternating from a greenish-yellow to dark brown-green, where the edges of are commonly irregular, ragged, and undulatory, filling the space between earlier phenocrysts.
Minor phases comprise less than 10% of the rhyolite, including magnetite, hematite, and zircon. Magnetite and hematite comprise 2-3% of the rhyolite, where hematite is directly related to oxidation. Magnetite occur as euhedral to subhedral 0.1-1.5 cm phenocrysts, and as micro-crysts in the groundmass. Magnetite is typically found as anhedral inclusions in other phenocrysts, ranging from 1-5 mm in size, in close association with other accessary phases, including cassiterite, changbaiite (PbNb 2 O 6 ), columbite-(Fe) group minerals, zircon and Y ± REE fluorides. Hematite occurs as replacement around the edges of magnetite phenocrysts, and less than 1-2 µm inclusions within the K-feldspar groundmass.
Trace minerals include zircon, Y-REE-fluorides, cassiterite, cerianite-(Ce), changbaiite, columbite-(Fe), cryolite and thorite. Zircon is the most common trace mineral, approximately 0.5-1% of the rhyolite at Round Top. Two populations of zircon are distinguished: one has abundant thorite inclusions and in the other these inclusions are rare. Most grains are 0.1-0.4 mm anhedral to subhedral, occasionally euhedral, phenocrysts. Rubin et al. [21] hypothesized the two zircon populations represent an early magmatic (thorite inclusion-rich), and a later population introduced by migrating hydrothermal fluids (thorite inclusion-poor). Ref. [21] noted the ubiquitous presence of late-stage hydrothermal zircons within fluorite replacement in surrounding limestone, but not within Round Top rhyolite. Both zircon populations (thorite inclusion-rich and inclusion-poor) were observed in Round Top rhyolite, commonly in close proximity to each other.

Geochemistry
The Round Top laccolith comprises relatively homogenous, high silica, non-topaz bearing rhyolite with low MgO and CaO, high Na 2 O and K 2 O ( Figure 5; Table 1), and relatively high Zr (960-1100 ppm), Rb (1660-1930 ppm), Th (160-190 ppm), and U (28-58 ppm; Table 2). Felsic rocks from the Round Top and Sierra Blanca suite of rocks show predominantly Within Plate (WPG) tectonomagmatic affinity ( Figure 6), they are F-rich (up to 2.4 wt %), peraluminous with ASI ratios (Aluminum Saturation Index) around 1.2, but not peralkaline as is typical of many REE enriched systems ( Figure 7). Initial Sr 87 /Sr 86 ratios for the mafic constituents of the magmatic suite are around 0.702 [19], where values of felsic members are around 0.728 [15]. These contrasting ratios imply a more crustal source for the felsic rocks. The Rb/Sr system appears to be affected by late-to post-magmatic processes increasing Rb relative to Sr ( Figure 8). The post magmatic process makes it difficult to rely on Sr in determining crustal contribution to the source of these magmas.
Each of the rhyolite laccoliths of the SBC and other regional magmatic rocks have distinct geochemical characteristics that help to discriminate them from each other [14,15,19]. The Finlay Mountains represent nearby mafic to intermediate compositions (40-65 wt % SiO 2 ), conforming temporally and compositionally to early gabbro/diorite dikes in the immediate SBC area. The Texan Mountain rhyolite represents the lowest silica content from felsic rocks in the magmatic series (64-70 wt % SiO 2 ), followed by Triple Hill and Sierra Blanca with Round Top, Little Blanca and Little Round Top representing extremely evolved rhyolitic compositions (70-76 wt % SiO 2 ). These evolved compositions show an increase in Na 2 O relative to K 2 O, decreasing Al 2 O 3 and very low MgO and P 2 O 5 ( Figure 5).
Alkalinity Index versus Aluminum Saturation Index show that all of the igneous rocks in the SBC and the Finlay mountains are alkalic, and all but the more primitive Finlay mountains samples are peraluminous. Tectonomagmatic affinity diagrams including Ta vs. Yb suggest a within-plate tectonic setting for the rhyolites from the SBC (see Figure 6), including the samples from Round Top. Rb/Sr ratio versus weight percent oxide of silicon shows a significant enrichment in rubidium for late stage, highly evolved rhyolites in the Sierra Blanca magmatic complex, as Rb fractionates preferentially into residual melts ( Figure 8).
Rare earth element plots show LREE enrichment for felsic, intermediate and low SiO 2 rocks in the SBC. Mafic/Intermediate patterns show little to no Eu anomaly, but felsic members are enriched in HREEs and show a moderate to strongly negative Eu anomaly (Round Top, Sierra Blanca, Little Round Top and Little Blanca rhyolites; Figure 9).
The petrogenetic model evaluates starting melts of a composition similar to the andesitic basalts of the Finlay Mountains, differentiated by closed-system Rayleigh fractional crystallization (FC) to produce early Sierra Blanca Complex melts-compositions similar to Texan Mountain and Triple Hill, and crustal melts represented by xenoliths from the Trans-Pecos region (Figure 10a). Solid lines represent the differentiation sequence of a range of regional basaltic compositions, dotted line represents batch melting (BM) of a range of regional crustal xenolith sources. Both products of regional melted crust and differentiated mafic magma compositions can produce appropriate melts similar to the mafic diorite sills that precede the Round Top rhyolite in the Sierra Blanca Complex, though major element least sum-residual-squared regression [25] of Tran-Pecos crustal xenolith (Σrisidual 2 = 1.003 at 33.70% melt) and Finlay Mountain andesitic-basalt (Σr 2 = 10.265 at 28.12% fractionation) mineralogies suggests a crustal melt is more appropriate for the rhyolite source.

Figure 5.
Major element (wt %) and trace element (ppm) vs. SiO 2 (wt %) concentrations in rhyolite from Round Top and other laccoliths in the Sierra Blanca Complex, compared with other regional igneous rocks from the Quitman mountains and Finlay mountains [14,15,19].  [14,15,19]. Symbology is the same as Figure 5.  [14,15,19]. Symbology is the same as Figure 5.  [14,15,19]. Symbology is the same as Figure 5.    [14,15,19]. The Round Top rhyolite shows extreme differentiation with very high Rb/Sr ratios, over a couple of orders of magnitude compared to most the other igneous rocks. Symbology the same as Figure 5.   [14,15,19]. The Round Top rhyolite shows extreme differentiation with very high Rb/Sr ratios, over a couple of orders of magnitude compared to most the other igneous rocks. Symbology the same as Figure 5.  [14,15,19]. The Round Top rhyolite shows extreme differentiation with very high Rb/Sr ratios, over a couple of orders of magnitude compared to most the other igneous rocks. Symbology the same as Figure 5.

Geochemical Modeling
The rhyolite of the Round Top laccolith does not have any direct source material indicators (autoliths, xenoliths, etc.) and identification of protolith compositions is determined from indirect geochemical and isotopic constraints, and petrogenetic modeling of probable regional source rocks. The Finlay Mountain andesitic to basaltic compositions represent earlier more primitive magmas that could be differentiated to form rhyolite similar to Round Top. The variation in Sr isotope ratios between mafic sills in the SBC (Sr 87 /Sr 86 = 0.702) [19] and Round Top (Sr 87 /Sr 86 = 0.728) [15], may be the result of late-to post-magmatic and hydrothermal processes that effected the Rb/Sr system. It is possible that crustal melts contributed, at least partially, to the initial melts that produced the Round Top rhyolite. Granulitic crustal rocks in the Trans-Pecos region provide possible source material at depth. Mafic, intermediate, and felsic xenolith compositions have been identified in the Trans-Pecos region and represent crustal lithologies [24].
The petrogenetic model evaluates starting melts of a composition similar to the andesitic basalts of the Finlay Mountains, differentiated by closed-system Rayleigh fractional crystallization (FC) to produce early Sierra Blanca Complex melts-compositions similar to Texan Mountain and Triple Hill, and crustal melts represented by xenoliths from the Trans-Pecos region (Figure 10a). Solid lines represent the differentiation sequence of a range of regional basaltic compositions, dotted line represents batch melting (BM) of a range of regional crustal xenolith sources. Both products of regional melted crust and differentiated mafic magma compositions can produce appropriate melts similar to the mafic diorite sills that precede the Round Top rhyolite in the Sierra Blanca Complex, though major element least sum-residual-squared regression [25] of Tran-Pecos crustal xenolith (Σrisidual 2 = 1.003 at 33.70% melt) and Finlay Mountain andesitic-basalt (Σr 2 = 10.265 at 28.12% fractionation) mineralogies suggests a crustal melt is more appropriate for the rhyolite source.  The sum residual squared from differentiated andesitic basalt, or melted crust, shows good correlation when compared with rare earth patterns for Round Top diorite compositions (i.e., The diorite compositions correlate well with melted xenolith (r 2 = 0.9551) and fractionated basalt compositions (r 2 = 0.9753); Figure 11). The FC of the observed early phenocryst assemblage from melt products in the melting model produce compositions similar to Texan Mountain and Triple Hill at 60% FC (Figures 10b and 12). Melt compositions similar to Texan Mountain fractionating observed mineral phases become the highly evolved compositions of Round Top, Sierra Blanca, Little Round Top and Little Blanca between 50% and 90% FC (Figures 10b and 13). The REE model correlates well with early FC modeling, but HREE enrichment starts to effect correlation (r 2 = 0.9353) when modeling compositions from Triple Hill through the most evolved compositions, even though LREEs fit very well. The sum residual squared from differentiated andesitic basalt, or melted crust, shows good correlation when compared with rare earth patterns for Round Top diorite compositions (i.e., The diorite compositions correlate well with melted xenolith (r 2 = 0.9551) and fractionated basalt compositions (r 2 = 0.9753); Figure 11). The FC of the observed early phenocryst assemblage from melt products in the melting model produce compositions similar to Texan Mountain and Triple Hill at 60% FC (Figures 10b and 12). Melt compositions similar to Texan Mountain fractionating observed mineral phases become the highly evolved compositions of Round Top, Sierra Blanca, Little Round Top and Little Blanca between 50% and 90% FC (Figures 10b and 13). The REE model correlates well with early FC modeling, but HREE enrichment starts to effect correlation (r 2 = 0.9353) when modeling compositions from Triple Hill through the most evolved compositions, even though LREEs fit very well. Figure 11. Chondrite normalized [27] rare earth element plots of diorite dikes in the Sierra Blanca Complex compared to the best possible melt compositions of modeled crustal xenolith source rocks (dashed grey line) and differentiation by closed system fractional crystallization of andesitic-basalt compositions (solid black line). Modeling parameters are described in Appendix A.
The significant HREE enrichment in Round Top and Sierra Blanca are difficult to account for with the observed late stage crystallizing phases. A model that incorporates just 0.5% of REE bearing accessory minerals known to be present, but in the matrix as pervasive micro to nano-crystalline crystals, accounts for the observed enrichment. The HREE enrichment is occurring at the transition between late stage magmatic crystallization (with an exsolving vapor phase) and alteration/precipitation related to high-temperature hydrothermal fluids, but with high enough fluorine to retain a large amount of silicate melt component in the fluid. The Round Top rare earth geochemistry is modeled and compared to similar highly evolved systems (Kymi complex in southern Finland [28], Erzgebirge and adjacent Fichtelgebirge and Oberpfalz rocks in Germany [29,30]) to understand the timing and processes that form economically viable rare earth element deposits in highly evolved magmatic rocks (Figures 14-16).    [27] rare earth element plots of modeled andesitic basalt dike/sill compositions from the Sierra Blanca Complex (Red circles), Texan Mountain (grey circles) and Triple Hill laccolith (light blue circles) compared to modeled 30% (dashed grey line) and 60% (dashed light blue line) FC of mineral phases from early melt (modeling parameters are described in Appendix A). Figure 13. Chondrite normalized [27] rare earth element plots of Texan Mountain rhyolite composition (grey circles) and Sierra Blanca (purple circles) and Round Top (pink circles) compositions compared with 30% late fractionating mineral phases (at 30%; red dotted line). Modeled LREEs correlate well, but HREEs are grossly underestimated (modeling parameters are described in Appendix A).

Figure 13.
Chondrite normalized [27] rare earth element plots of Texan Mountain rhyolite composition (grey circles) and Sierra Blanca (purple circles) and Round Top (pink circles) compositions compared with 30% late fractionating mineral phases (at 30%; red dotted line). Modeled LREEs correlate well, but HREEs are grossly underestimated (modeling parameters are described in Appendix A).
The significant HREE enrichment in Round Top and Sierra Blanca are difficult to account for with the observed late stage crystallizing phases. A model that incorporates just 0.5% of REE bearing accessory minerals known to be present, but in the matrix as pervasive micro to nano-crystalline crystals, accounts for the observed enrichment. The HREE enrichment is occurring at the transition between late stage magmatic crystallization (with an exsolving vapor phase) and alteration/precipitation related to high-temperature hydrothermal fluids, but with high enough fluorine to retain a large amount of silicate melt component in the fluid. The Round Top rare earth geochemistry is modeled and compared to similar highly evolved systems (Kymi complex in southern Finland [28], Erzgebirge and adjacent Fichtelgebirge and Oberpfalz rocks in Germany [29,30]) to understand the timing and processes that form economically viable rare earth element deposits in highly evolved magmatic rocks (Figures 14-16).
Irber [30] discusses quantification of the tetrad effect and ligand dependent differences in the ionic radii of REEs [32]. The variation in ionic radii are based on variable expansion of the electron cloud (nephelauxetic-or tetrad-effect), type of complexing and internal atomic structure. The inference is that the tetrad effect can be generated during REE partitioning at the transition between silicate melt and high-temperature hydrothermal systems, or between coexisting melt, aqueous high-temperature late stage fluid and crystallizing minerals, in both cases with strong complexing of fluorine and REE.
The observed decoupled behavior, where Sr is elevated relative to Eu, is a result of strong Eu depletion with early feldspar crystallization (Figure 14b). Zr/Hf ratios shift toward lower ratios with more evolved compositions of silicate melts. Round Top ratios represent increased Hf in late stage zircons transitioning to very low Zr/Hf ratios of pegmatitic systems (e.g., [34]). Y/Ho ratios have near chondritic values, but only coincidently due to an original path of differentiation and non-CHARAC behavior followed by fluorine ligand complexing in late silicate melt before aqueous fluid transition. Y/Ho ratios combined with Zr/Hf ratios show that the Round Top samples do not have the elevated Y and Hf relative to Ho and Zr, respectively, that would be typical of hydrothermal alteration processes in these types of rocks (cf. [35]; Figures 15 and 16). Y and HREEs were retained in the melt due to the F-rich nature of the magmas (cf. [36][37][38]).  [30]) and Round Top rhyolite compared with samples from Erzgebirge [29,30], Kymi [28] and chondritic crust. The tetrad effect shows irregularity from expected REE behavior by comparing the sum of Ce + Pr to La and Nd, and Tb + Dy to Gd and Ho [30].  [30]) and Round Top rhyolite compared with samples from Erzgebirge [29,30], Kymi [28] and chondritic crust. The tetrad effect shows irregularity from expected REE behavior by comparing the sum of Ce + Pr to La and Nd, and Tb + Dy to Gd and Ho [30].   [29,30], Kymi [28] and chondritic crust. Hydrothermal alteration and CHARAC field(s) adapted from [31]. Symbology the same as Figure 14.
Irber [30] discusses quantification of the tetrad effect and ligand dependent differences in the ionic radii of REEs [32]. The variation in ionic radii are based on variable expansion of the electron cloud (nephelauxetic-or tetrad-effect), type of complexing and internal atomic structure. The inference is that the tetrad effect can be generated during REE partitioning at the transition between silicate melt and high-temperature hydrothermal systems, or between coexisting melt, aqueous hightemperature late stage fluid and crystallizing minerals, in both cases with strong complexing of fluorine and REE.
The observed mineral textures trend toward late melt phase REE mineralization in a rapid nucleation environment with a moderate to high degree of undercooling. Geochemical modeling suggests that the Round Top rhyolite HREE enrichment is a result of thorough and long-lived differentiation processes in a fluorine-rich magmatic system, where fluorine complexing with HREE and similar trace elements was pervasive and transitional between low temperature silicate melt and magmatic hydrothermal fluids.

Magma Evolution and Emplacement
The conditions and geochemistry necessary to form a rare earth element deposit in highly evolved sub-alkaline rhyolite has not been well documented. The peraluminous composition of Round Top rhyolite differs from peralkaline and metaluminous rhyolites within the TPMP, suggesting a difference in the source material, formation process and/or emplacement timing [3,22,39].
TPMP magmatism and timing of tectonic events were favorable for the formation of continental sourced, highly enriched silicic magma evolving at depth and being emplaced in a shallow environment. The complex tectonic history of the TPMP has had a key role in the development of magmatic source material, providing heat and hosting early mafic to intermediate magmatism.
The periodic compressive events punctuated by abrupt stress orientation changes and subsequent extension, overprinted with multiple faulting events and volcanism, created a geologic setting favorable for prolonged isothermal fractionation and differentiation.
The Round Top magma was emplaced in upper Cretaceous fine-grained siliciclastic and carbonate strata, relatively contemporaneous with the formation of the SBC laccoliths, at 36 Ma [6]. Magmatic material that formed the laccolith structure stalled during ascension, extended laterally, and was prevented from erupting at the surface. Rapidly cooled fine-grained groundmass suggests loss of pressure at a higher level in the system, allowing the stalled laccolith magma to quench. The migration of magma through the surrounding Cretaceous country rock was controlled by buoyancy and density contrasts. The nature of the contacts between the laccolith structure, mafic sills and Cretaceous strata suggest the magma was injected along bedding planes at several depths. The rhyolite magma most likely used preexisting magmatic conduits of the diorite sill during emplacement, although enclaves of the preexisting mafic yet to be documented within the rhyolite.
The elevated fluorine content and volatile-rich nature of the magma contributed to the episodic migration of igneous material by altering the viscosity of the magmatic fluids. Elevated fluorine content has been identified as a mechanism for decreasing viscosity of volatile-rich silicic magmas [40,41]. Crystallization of the magma began at depth, and formed phenocrysts of K-feldspar, quartz and plagioclase. Subsequent FC processes continued to differentiate the magma, where isothermal decompression events caused dissolution of K-feldspar and recrystallized as mantle overgrowths on albite phenocrysts. The zoned quartz phenocrysts also provide evidence for prolonged phenocryst residence, where quartz corrosion and re-precipitation (snowball texture) occurred, probably from loss of pressure during multiple minor degassing events. Undercooling textures suggest the groundmass nucleated and crystallized quickly, probably during an eruption event. Interstitial spongy texture in the groundmass and late-stage vapor-phase mineralization in those pore spaces suggests a lateto post-magmatic hydrothermal overprint. The surrounding Cretaceous rock provided an ideal environment to trap residual magmatic fluids, facilitating rapid quenching and crystallization of the groundmass. The rapidly quenching vapor phase formed late-stage trace phases and REE-fluorides. The rhyolite is uniformly, and remarkably, aphanitic and homogenous, suggesting that crystallization of the laccolith as a whole was relatively rapid.
Post-magmatic fluid migrated along fractures and joint surfaces within the rhyolite, oxidizing parts of the laccolith.
Fluid-mineral interaction caused magnetite oxidization, forming pervasive hematite alteration. Any color variation in rhyolite is the product of increasing proportions of hematite and oxidation processes, primarily within the micropores of K-feldspar in the groundmass. Post-magmatic mineralization in the brecciated fractures and joint surfaces is ubiquitous, producing multiple generations of fluorite, calcite and quartz. The pervasive post-magmatic mineralization suggests a prolonged period of remobilization and reprecipitation, facilitated by the circulation of oxidizing, low-temperature, meteoric fluids. The lack of elevated REE concentrations in the surrounding fluoritized limestone and fractures within the rhyolite suggests the post-magmatic conditions were insufficient to remobilize REEs and be incorporated in secondary fluorite.
Disequilibrium textures and the anhedral morphology of groundmass mineralogy throughout the laccolith suggest rapid crystallization. The hourglass sector zonation of the K-feldspar phenocrysts is an unusual disequilibrium texture that occurs as a result of preferential site allocation of sodium and potassium during formation, indicating rapid crystallization and growth during disequilibrium (see Figure 3; [42,43]). Additionally, the pristine outer edge of many of the interior plagioclase laths and interior quartz cores indicate disequilibrium or corrosion of the phenocrysts did not occur until after the formation of overgrowths. The three-part quartz phenocrysts indicate transition from euhedral beta quartz to rapidly crystallizing low-temperature quartz, entraining magmatic melt inclusions. The uniform occurrence of these disequilibrium textures throughout the rhyolite suggests that the change in pressure-temperature conditions affected the entirety of the laccolith, and it is likely that the Round Top laccolith was emplaced as a single magmatic event.

Timing and Crystallization of REE-Enriched Minerals
The paragenetic sequence of the Round Top mineralogy includes early, middle, and late crystallization phases. These phases were subjected to late-stage alteration from a volatile-rich vapor phase and post-magmatic partial oxidation from circulating fluids. The observed mineralogical relationships between the major, accessory, and trace phases include the formation of early phenocrysts within the magma chamber and continued through emplacement. Albite lathes and the core portions of high-temperature beta quartz phenocrysts formed early. Magma emplacement was relatively rapid, inhibiting significant crystallization during transport as evident by the aphanitic and homogenous nature of the groundmass. As temperature and pressure conditions decreased during emplacement, the subhedral to euhedral, hourglass-type K-feldspar phenocrysts, the K-feldspar overgrowths, the smaller accessory phases, and the larger constituents of the groundmass (K-feldspar, albite, quartz, and annite) began to crystallize. Interaction of these minerals with the circulating high temperature volatile-rich vapor phase produced the dissolution textures observed with the feldspar and quartz populations. During the final stages of crystallization, interaction with the volatile-rich vapor phase produced the alteration textures now observed within the feldspars and quartz overgrowths. Late-stage phases occur interstitial to and within the K-feldspar pores, and include cassiterite, cerianite-(Ce), changbaiite, tantalite, Y-and Ce-rich fluorite, and other nanoscale Y+REE-bearing phases. The intimate association of the REE-fluoride minerals, magnetite, zircon and other trace phases suggests the timing of crystallization was relatively close. Rubin [7,21] suggested that the early and late zircon populations of the Round Top rhyolite (thorite-rich and thorite-poor) indicate dissolution and recrystallization during subsequent hydrothermal alteration. Selective remobilization of only zircon would require a fluid that could dissolve minerals with relatively immobile geochemical constituents and leave the less robust minerals, such as K-feldspar, plagioclase and quartz, intact. The zircon populations in Round Top rhyolite more likely represent one population with varying degrees of thorite exsolution.

Heavy Rare Earth Element Enrichment
The Round Top laccolith underwent a pervasive alteration event during late-stage crystallization prior to being oxidized. The alteration event is characterized by partial dissolution and the formation of the "spongy" texture pores spaces within the K-feldspar phenocrysts, overgrowths, and groundmass [44], in addition to forming the undulatory edges in late quartz overgrowths [42].
The strong correlation of REE-minerals to the K-feldspar portion of the rhyolite suggests the availability of open pore space was a significant factor in the retention of the REE-fluoride mineralization. The ubiquitous nature of the quartz embayment and feldspar dissolution suggests the vapor phase alteration was pervasive throughout the entirety of the laccolith, and helps to constrain the timing of REE-bearing mineral precipitation.
The evolution of the Round Top rhyolite and elevated heavy rare earth element concentration requires a suitable chemistry of the protolith and evolving magma that keeps HREEs in the liquid phase until late precipitation at the end of the SBC magmatic evolution. The increased Na, K, and Al in the magma causes increased solubility of rare earths [43]. The fluorine content of the melt is critical to increasing the solubility of REE bearing constituents and keeping the REE in the melt through the evolution of the magma [45,46]. Studies have shown that these compositions can have intermediate properties between silicate melt and hydrothermal fluid allowing for significant amounts of REE-and HFSE-incorporating minerals to crystallize [38,[47][48][49][50]. Further transport in late-magmatic hydrothermal conditions depends on the availability of ligands in solution that can form stable REE complexes. Although there are environmental considerations and exceptions [51,52], fluorine complexes have been shown to dominate hydrothermal transport of REEs [53,54]. The lack of REE-enrichment in post magmatic fluorite from the surrounding fluoritized limestone suggests that REE mobilization conditions were restricted to late-magmatic to transitional timing.

Conclusions
The Round Top rhyolite is enriched in Be, F, Li, Nb, Rb, Sn, Th, U, Y, Zr, and REEs. REE-bearing minerals are mainly ubiquitous nano-scale accessory phases throughout the groundmass, incorporated in synchysite-group minerals, xenotime-(Y), Y-, Ce-, and REE-bearing fluorides, and zircon. The rhyolite is slightly peraluminous, silica-rich, alkaline, with elevated HREE concentrations and a strongly negative Eu-anomaly. Pervasive spongy groundmass and recrystallization textures are consistent with the elevated and remobilized Zr, Th, and Y + HREE concentrations and a HFSE soluble, alkali-and F-rich, magmatic system. REE mineralization is secondary and interstitial, attributed with late-magmatic to transitional vapor-phase mineralization, prior to post-magmatic hydrothermal alteration.
Geochemical modeling provides a better understanding of the geochemical evolution and REE complexing behavior in these rocks, and determines possible source compositions and petrogenetic evolution. The Round Top model suggests an intermediate to mafic crustal protolith or fractionating intermediate to mafic magma could be viable source material. Trace and rare earth element modeling suggests a system typical of having extensive magmatic differentiation, with pulses of emplaced magma represented by the laccoliths of the SBC. The resulting Round Top rhyolite magma is indicative of a high-silica magmatic system rich in H 2 O, Li, and/or F transitional between pure silicate melt and hydrothermal fluid, where fluorine-ligand complexing was prevalent through late magmatic cooling and crystallization processes. High fluorine activity and thorough differentiation contributed to the late stage precipitation of REE-bearing minerals in the Round Top rhyolite. Table A4. Partition coefficients (K d values) and bulk distribution coefficients (BDC) used in fractional crystallization modeling of late felsic melt (dotted line in Figure 10b) fractionating 5% biotite (Bt), 75% alkali feldspar (Kspar) and 20% quartz (Qtz). Ba and Sr values are in ppm. Appendix B Table A7. Geochemistry quality control was conducted by blind duplicate analyses, with a known standard(s) and conducted by the commercial analytical facility ACMELABS in Vancouver, B.C., Canada. Major elements and F in weight percent (%), trace and rare earth elements in ppm.