U-Pb age of a late Cenozoic ultra-high temperature metamorphic event under Central Mexico, as inferred from granulite xenoliths from Cerro El Toro, Mexico

ABSTRACT Cerro El Toro, a Pliocene scoria cone that carried abundant lower crust xenoliths to the surface, is located in the Mesa Central (MC), Mexico, where mafic alkalic intra-plate volcanism has occurred since the early Miocene (end of the Sierra Madre Occidental volcanism). Early studies of El Toro xenoliths showed that the feldspathic granulite equilibrated at P = 0.9–1.4 GPa and T = 900−1100°C under anhydrous conditions (ternary feldspars calculations). A young (Oligocene-Quaternary?) pulse of ultra-high temperature (UHT) metamorphism was thus proposed for the region, without a strict age constrain. Zircon crystals recovered from a set of nine xenoliths (7 Grt-Sil bearing metapelites, a Px-bearing meta-quartz diorite, and a Grt-Opx bearing orthogneiss) were selected for LA-(MC)-ICPMS U-Pb geochronology. Palaeozoic to Neoproterozoic zircon crystals are scant, whereas Mesozoic to Cenozoic ages are more abundant. Based on their chondrite-normalized REEs all Neoproterozoic to Mesozoic zircons are interpreted as igneous, whereas those of late Oligocene to late Miocene age (ca. 27–6 Ma) are mostly metamorphic and grew during a protracted pulse of UHT metamorphism in the lower crust. The presence of Cenozoic metasediments in the lower crust under Cerro El Toro is indicative of the action of the subducted Farallon plate, coupled with tectonic erosion of continentally derived sediments, either from a forearc basin and/or an accretionary prism that were relaminated to the lower crust by sediment diapirism. Similarities among the xenolith zircon ages with those from modern sediments belonging to Central Mexico Pacific coast point towards a NW-SE 100 km long coast stretch across Zihuatanejo as a possible sediment source. The paucity of Grenville-age detrital zircon grains in the recovered xenoliths suggests that the El Toro area is not underlain by a Proterozoic basement, thus implying a substantial reduction of Oaxaquia extension beneath Central Mexico and presence of the younger Guerrero terrane.


Introduction
Feldspathic granulite-facies xenoliths are fundamental records of deep-seated processes, which occur in the continental crust where no direct observation is possible. In Mexico, they are commonly brought to the surface, together with mantle fragments, by mafic alkalic intraplate-type magmas (e.g. Ruiz et al. 1988;Luhr et al. 1989;Rudnick and Cameron 1991;Schaaf et al. 1994). Their study can, among other things, help to identify different compositional and age-wise basement types that are directly unreachable, as crystalline basement outcrops are notoriously absent in vast portions of central and northern Mexico. The southern part of the Mesa Central of Mexico ( Figure 1) is characterized by a thick cover (an average of 1000 m) of dominantly felsic volcanic rocks, mostly Tertiary in age, which buries the underlying Mesozoic marine sediments (Nieto-Samaniego et al. 1999;Orozco-Esquivel et al. 2002), laying atop the crystalline basement. While the Mexican crystalline basement is exposed in isolated localities in northern, southern, and some eastern areas of Mexico (e.g. Ortega-Gutiérrez et al. 1995Solari et al. 2014Solari et al. , 2018Weber and Schulze 2014;Weber et al. 2010Weber et al. , 2018Weber et al. , 2020Maldonado et al. 2020) it is almost completely hidden in Central Mexico, and its nature and age has only been locally inferred from the study of scant xenoliths found in volcanic rocks (e.g. Hayob et al. 1989;Aguirre-Díaz et al. 2002;Ortega-Gutiérrez et al. 2012. Whereas geochemical and isotopic compositions of these rocks are sometimes hard to interpret, as it is difficult to establish links between different xenoliths sampled in the same locality, because they are random samples from a complex column with unknown thickness, geochronology (especially if obtained by U-Pb) is a proven method that yields sensible information, which may be used to infer the broad nature of xenolith protoliths, which are lithologies that formed in the lower crust prior to the time of the eruption that brought the xenoliths to the surface (e.g. Rudnick and Cameron 1991). Likewise, zircon data may help to establish significant differences (igneous vs metamorphic origin) among crystals of this phase and formulate hypotheses on the tectonomagmatic evolution of the source area.
This work presents a set of U-Pb geochronologic data obtained from 9 lower crustal, granulite-facies, feldspathic xenoliths collected at the Cerro El Toro locality in San Luis Potosi state, Central Mexico ( Figure 2). U-Pb data obtained studying their zircon crystals are used to constrain the ages of igneous and metamorphic events registered in each one of the xenoliths. It is assumed that no single xenolith in the studied suite contains a complete record of all the metamorphic and igneous processes that occurred during the evolution of the lower crust underneath the region where Cerro El Toro is located. A combined set of inferred processes, derived from individual xenoliths in the suite, may give a better picture of the evolution of the basal complex. Results are also used to set important geographic and temporal constraints on at least three basement domains, which are considered important tectonic boundaries in Central Mexico, namely the Guerrero, Sierra Madre, and Oaxaquia terranes (e.g. Campa and Coney 1983;Sedlock et al. 1993;Keppie 2004) and reflect a complex tectonomagmatic evolution of the southern end of North America. Uncertainties related to the boundary locations between different basement domains (terranes) stem from the existence of a thick cover of Mesozoic marine sediments, and/or mid-Tertiary volcanic rocks, and/or basin fill Quaternary sediments ( Figure 2) almost everywhere in Central Mexico.

Regional geology
Cerro El Toro is located in the Mesa Central, which is a high (1800−2300 masl) plateau surrounded by mountain chains (Figure 2). South of the Mesa Central is the Trans-Mexican Volcanic Belt (TMVB), an arc related with active subduction along the Middle America Trench. Activity at the TMVB began in the late Miocene (~10 Ma: Ferrari et al. 1999), with the magmatic front near the Mesa Central. Currently, the magmatic front of the TMVB is located closer to the trench, ~230 km south of the Mesa Central-TMVB boundary. West of the Mesa Central is the Sierra Madre Occidental, a silicic large igneous province that was active during the Eocene-Miocene, and it was contemporaneous at least in part with subduction of the Farallon Plate under North America, and it was later coeval with the arrival of the East Pacific Rise to the trench and with a major plate reorganization along the Pacific Coast of North America (e.g. Stock and Hodges 1989), which culminated with Baja California Peninsula rifting and capture by the Pacific plate and the formation of oceanic crust at the mouth of the Gulf of California in the Pliocene (~3.7 Ma: Castillo et al. 2002).
The northern and eastern borders of the Mesa Central are formed by the Sierra Madre Oriental, a fold and thrust belt active during the Late Cretaceous-Eocene and related with the Mexican Orogen. Syn-orogenic magmatism accompanied folding and thrusting from the Cenomanian to the Eocene in the Mexican Orogen. Cárdenas is in the Sierra Madre Oriental. Based on their geographical distribution, petrology of host magmas, and/or xenolith type(s) the intraplate-type vents have been divided into four different groups. On the basis of the abundance, size, and xenolith variety, the most important localities in the Ventura-Espíritu Santo group are Cerro El Toro (CT), Joyuela (J), Joya Honda (JH) and Laguna de los Palau (LP), whereas in the Santo Domingo group are Joya Prieta (JP), Joya Los Contreras (JC), and the Santo Domingo (SD) maar. The general trend and approximate influence area of San Luis-Tepehuanes fault system within the state is shown in green. Cities: San Luis Potosí (SLP) and Matehuala (M). VRG = Villa de Reyes Graben.
Magmatism during the Mesozoic was prevalent in the hinterland region of the orogen where it occurred in two phases: the first one was late Triassic-Late Jurassic and the second phase was Late Jurassic-Early Cretaceous, producing submarine lava flows, dike swarms, and plutons, which are exposed in the Sierra de Guanajuato, at the southern end of the Mesa Central. Voluminous calcalkaline plutonism and silicic volcanism occurred close to the Pacific margin of Mexico during the Late Cretaceous, and scattered plutons were emplaced in central and eastern Mexico (Fitz-Díaz et al. 2018). Magmatic activity either related with subduction or intra-plate type continued in the region during most of the Cenozoic.
The Mesa Central was considered as the southernmost part of the Basin and Range Province Aranda-Gómez 1991 andAranda-Gómez, 1992), as there is abundant evidence of Neogene extension in it. Isolated mountain chains within the Mesa Central are often cored by folded Mesozoic sedimentary successions, shortening structures that were formed in the hinterland of the Mexican Orogen. The southern part of the Mesa Central is overlain by mid-Cenozoic volcanic rocks that have been affected by three sets of normal faults with NS, NW, and NE trends (Nieto-Samaniego et al. 2007). Thus, mountain ranges in that portion of the Mesa Central are fault-bounded blocks in which the folded marine successions are partially buried under a thick cover of felsic volcanic rocks of the Sierra Madre Occidental (Figure 2). The NW-SE trending, ~700 kmlong, San Luis -Tepehuanes normal fault system cuts the Mesa Central and divides it into two segments. Mountain ranges in the northern segment are cored by Mesozoic rocks and in a few places exist remnants of the Palaeogene volcanic rocks and small outcrops of intrusive bodies of the same age. The southern segment of the Mesa Central, located between the San Luis-Tepehuanes ( Figure 2) and El Bajío normal fault systems is mostly covered by Palaeogene volcanic rocks and outcrops of Mesozoic sedimentary and plutonic rocks are subordinated (Nieto-Samaniego et al. 2007). The Sierra de Guanajuato is of special relevance as it exposes submarine volcanic rocks and Mesozoic plutons that have been referred to the Guerrero Terrane, a controversial concept that has either been interpreted as allochthonous (i.e. an exotic Pacific arc developed over oceanic crust and accreted to southern end of North America, e.g. Lapierre et al. 1992;Tardy et al. 1994) or as a para-autochthonous block (a volcanic arc developed on continental crust, rifted from nuclear Mexico by back-arc extension, and then thrusted above nuclear Mexico during the early stages of the Mexican Orogen, e.g. Martini et al. 2011;Ortega-Flores et al. 2020).
A small outcrop of Mesozoic pillow-lavas exposed 60 km S8W of Cerro El Toro has been attributed to the Guerrero terrane (Dávalos-Elizondo 2011). Similar submarine lavas have been documented at the Zacatecas and Fresnillo mining districts (De Cserna 1976;Escalona-Alcázar et al. 2009), 110 km S84W and 140 km N46W, respectively, from Cerro El Toro.
Local vents where felsic magmas arrived at the surface during the Eocene to early Miocene (45.5-20.5 Ma, Aguillón-Robles et al. 2009) are abundant in the southern portion of the Mesa Central. Two volcanic styles appear to be prevalent in that region: emplacement of large, voluminous dacitic to rhyolitic lava dome complexes and fissure-related eruptions that formed felsic ignimbritic successions. Early Miocene activity was bimodal (Aguillón-Robles et al. 2009).
Extensively scattered intraplate-type mafic alkalic rocks occur in the Mesa Central forming small volcanic clusters. Based on their geographic distribution, isotopic age, geochemistry of the volcanic rocks and type of inclusions (xenoliths and/or megacrysts), the intraplate volcanoes of the Mesa Central have been divided into three volcanic fields. A large set of middle Miocene (10.6 −13.6 Ma, K-Ar on groundmass, Luhr et al. 1995) volcanic necks and lava capped mesas, mostly made of xenocryst-and megacryst-bearing hawaiites were referred as Los Encinos Volcanic Field. These volcanoes define two distinct lineaments, more than 60 km long each, which trend N45W and N45E, respectively ( Figure 2). Most hawaiites from Los Encinos contain complex assemblages of resorbed and reacted xenocrysts and megacrysts, which in some samples are accompanied by partly disaggregated lower-crustal orthogneisses and paragneisses. Isotopic data and trace element geochemistry of Los Encinos volcanic rocks are consistent with AFC-style contamination with lower crustal garnetbearing granulitic paragneisses (Luhr et al. 1995).
Xenolith-bearing late Miocene (Tristán-González 2008) to Pleistocene-latest Miocene (ca. 6-0.4 Ma: Aranda-Gómez and Luhr 1996;Aguillón-Robles et al. 2009;Saucedo et al. 2017) mafic alkalic rocks have been divided into the Santo Domingo and the Ventura-Espíritu Santo volcanic fields ( Figure 2). Maars, cinder cones and associated lava flows made up of xenolithbearing (mantle and lower crustal, e.g. Aranda-Gomez and Ortega-Gutiérrez 1987) olivine-nephelinite and basanite form the Ventura-Espiritu Santo volcanic field (Figure 2), which is located in or near the influence area of the San Luis-Tepehuanes fault system (Aranda-Gómez and Dávila-Harris 2014). Mantle xenoliths in these volcanoes generally have a coarse granular texture, according to Mercier and Nicholas (1975) classification scheme, whereas megacrysts and hydrated phases in the mantle peridotite are absent or extremely rare (Heinrich and Besch 1992;Aranda-Gómez and Dávila-Harris 2014). The Santo Domingo volcanic field is located ~60 km to the N35E of the Ventura volcanic cluster (i.e. Joya Honda, Joyuela and neighbouring cinder cones and lava flows). Most mantle peridotites in the Santo Domingo group (Figure 2) are foliated tectonites with distinctive porphyroclastic (Mercier and Nicholas 1975) texture. Kaersutite and pargasite megacrysts and pegmatitic hornblendites are always present and they are abundant in one of the Santo Domingo maars (Dávalos-Elizondo 2018). Rare composite xenoliths, made of foliated peridotites intruded by thin kaersutite-rich dikes are also present. Furthermore, the Santo Domingo volcanic field lies on the N45Wtrending San Tiburcio lineament, which is interpreted as a regional-scale basement fault (Mitre-Salazar 1989;Aranda-Gómezet al. 2007a). Volcanic rocks in Santo Domingo are hawaiites and in some of these rocks crustal contamination may be significant . Age of some of the volcanoes in Santo Domingo (0.3-0.5 Ma, K-Ar, groundmass) is like the ~0.4 Ma age of Joya Honda (Ar-Ar, whole rock, Saucedo et al. 2017) at the Ventura cluster.
Cerro El Toro is a cinder cone located in the westernmost sector of the Ventura-Espíritu Santo Volcanic Field (Aranda-Gómez 1982;Luhr et al. 1989;Hayob et al. 1989;(Figure 3). Morphology of the breached cone suggests that it is in a mature stage of degradation, as its crater is broad and shallow, its outer slopes are ~22°, and volcanic agglutinate beds are clearly exposed on the edifice. Thus, based on comparison with isotopically dated, mature cinder cones in similar climatic conditions (e.g. Camargo Volcanic Field: Aranda-Gómez et al. 2003), it is here hypothesized that the volcano is Plio-Pleistocene in age. Cerro El Toro volcanic complex appears to be formed by two vents, being the northern one more conspicuous and better preserved. The topography, as seen in vertical air-photo stereopairs, suggests that there are three different lava flows issued from the volcano. Mantle and lower crustal xenoliths are present in all of them, but feldspathic granulite are particularly abundant in the southern lava flow, in which they may be as large as 40 cm in length and 15-20 cm in width. All the xenoliths studied in this investigation were collected in the southern lava flow.
Temperature equilibration estimates, based on ternary feldspar geothermometry yielded metamorphic temperatures of 950-1125°C for metapelites with the Kfs + Pl + Grt ± Sil + Qz+ Rt + Gr paragenesis (Hayob et al. 1989; phase's abbreviations according to Whitney and Evans 2010). The Grt + Sil, + Qz + Pl geobarometer gave pressure estimates of 0.9−1.4 GPa, or 0.9-1.3 GPa with the Grt + Rt+ Sil + Ilm + Qz barometer for the same samples. The P-T equilibration conditions inferred for these xenoliths, which were interpreted by Hayob et al. (1989) as products of regional-scale, ultra-high temperature (>900°C) granulite facies metamorphic episode caused by basaltic underplating in the past, but not prior than 30 Ma. The UHT condition has probably been preserved in the lower crust beneath the Mesa Central, as mafic alkalic volcanism in the region occurred in several pulses and the youngest known volcano is very young (~311 ka, Ar-Ar, whole rock, Saucedo et al. 2017).
Geochronologic studies of crustal xenoliths from Mexico are scant. Xenoliths have been collected from a few volcanoes widely dispersed in a vast region, which range from N to S in La Olivina in southeastern Chihuahua, where Grenville-age (~1.3-1.0 Ga), granulitefacies xenoliths are found in Pleistocene basalts (1.8 Ma according to Rudnick and Cameron 1991;Aranda-Gómez et al. 2003), to El Toro, Ventura, and S. Domingo maars in San Luis Potosí (Aranda-Gómez 1982; Ruiz et al. 1988;Hayob et al. 1989). Further south, scant xenoliths were described in La Goleta Cenozoic volcanics, as highgrade metapelite (Elías-Herrera and Ortega-Gutiérrez 1997) and in the Amealco caldera as granulite-facies meta-igneous sample (Aguirre-Díaz et al. 2002). Xenoliths from the Miocene Chalcatzingo trondhjemite were interpreted as high-grade metapelite and igneous rocks with concordant igneous ages spanning from the Devonian to the Oligocene (Gómez-Tuena et al. 2008;Ortega-Gutiérrez et al. 2012). In the Rincón de Parangueo maar, a dated, high-grade felsic xenolith yielded a Late Cretaceous age . A further age was presented by Schaaf et al. (1994) with a Sm-Nd isochron age of 1248 ± 69 Ma, calculated on four wholerock orthogneiss granulites and two whole-rock pyroxenite and websterite from the Santo Domingo and Ventura maars in San Luis Potosí.
Outcropping high-grade basement in eastern and southern Mexico constitutes what is known as the Oaxaquia microcontinent, a Grenvillian, granulite-facies block firstly proposed by Ortega-Gutiérrez et al. (1995) and interpreted as the backbone of Mexico, around which the Mesozoic and Cenozoic terranes were accreted. Oaxaquia's most extensive and continuous outcrop forms the Oaxacan Complex in southern Mexico Solari et al. 2003Solari et al. , 2014Weber et al. 2010;Weber and Schulze 2014). Other isolated portions of Oaxaquia are the Huiznopala Gneiss in east-Central Mexico (Lawlor et al. 1999;Weber and Schulze 2014), the Novillo Gneiss in eastern Mexico (Cameron et al. 2004), which was affected by dikeswarm magmatism during the Ediacaran (~635-541 Ma) break-up of Rodinia (Weber et al. 2019), and the Guichicovi Complex in southeastern Mexico (Weber and Kohler 1999), which is tentatively associated with its lower metamorphic grade counterpart further to the southeast, in the El Triunfo Complex (de León A et al. 2017;Weber et al. 2018).

Sample description
The studied xenoliths (a summary of their petrographic features is available in Table 1) are generally 10-14 cm in length, with poorly defined segregation banding in hand sample, and characterized by a coarse-grained granoblastic texture (Suppl. Figure 1). Based on their composition, xenoliths can be grouped into quartzofeldspathic, mafic, and aluminous classes. Quartzofeldspathic and mafic xenoliths display variable mineral assemblages (Examples in Figure 4), mostly made up by plagioclase, orthopyroxene and clinopyroxene, quartz, K-feldspar, garnet, and kaersutitic hornblende, whereas accessory minerals are rutile, titanite, zircon, and opaques. The limited amount of modal quartz (2-15%) in these xenoliths, abundant plagioclase (25-30%, mostly oligoclase to andesine) and K-feldspar (12-25%), and relatively large amounts of garnet (up to 25%) and pyroxene (up to 40%) suggest that they had igneous protoliths of intermediate to mafic composition. This composition is similar to that previously described by Schaaf et al. (1994). Mineral assemblages in aluminium-rich xenoliths are mainly composed of plagioclase, K-feldspar, garnet, quartz, sillimanite, and minor amounts of orthopyroxene, cordierite, rutile, zircon, and opaques. The abundance of Al-rich phases coupled with the age distribution obtained by zircon dating (next sections) strongly support the interpretation that most of the aluminous xenoliths are in fact metasedimentary rocks.
Thin glassy films are often recognized at the intergranular contacts between phases in some xenoliths. This glass, as well as the symplectitic reaction rims around garnet porphyroblasts are interpreted as the products of decompression melting, which occurred during the host magma transit from the lower crust to the surface. It is here noted that glass selvage thickness increases when the contact is among different mineralogical phases. Locally, small triangular glass-filled pockets occur at triple-points. The conspicuous granoblastic texture, widespread occurrence of well-equilibrated 120° triple joints, and stability of garnet in the presence of orthopyroxene prior to the decompression melting are all features consistent with granulite facies metamorphism.

Methods
A total of nine xenoliths were chosen for U-Pb geochronology. They were crushed, sieved, and zircon crystals were separated, carefully handpicked and mounted. One xenolith (XES-1013, see below) was devoid of zircon, but it had abundant titanite that was thus chosen instead. Mounting was performed under binocular microscope, in 1" round mounts that were glued in epoxy resin, polished and then imaged by cathodoluminescence (Suppl. Figure 2). U-Pb analyses were conducted by laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) at the Laboratorio de Estudios Isotópicos (LEI) of the Centro de Geociencias, UNAM, using a Resonetics M050 (now, Applied Spectra) 193 nm excimer laser workstation, coupled to a Thermo ICap Qc quadrupole mass spectrometer, according to the methods reported by Solari et al. (2018). Two of the samples were also analysed employing the same laser ablation system, connected to a Neptune Plus multi-collector inductively coupled plasma mass spectrometer (LA-MC-ICPMS). A 23 μm spot was employed during this study for all the zircon U-Pb analyses measured in the quadrupole, whereas 17 μm were employed for the multi collector analyses, alternating unknown zircon crystals with several standards. The titanite crystals were measured with a 44 μm spot. Standard reference material 91,500 was employed as external reference zircon (ca. 1062 Ma, Wiedenbeck et al. 1995), whereas Plešovice standard zircon acted as secondary (control) standard (ca. 337 Ma, Sláma et al. 2008). For titanite, the BLR-1 titanite of Aleinikoff et al. (2007) was used. Together with those isotopes employed for U-Pb ratios and age calculation (Pbs, Th, U) other isotopes were monitored for trace element concentrations (Si, Ti, Y, Nb, REE, Hf). NIST 610 glass was measured and employed as external standard for trace element concentration calculations, using 29 Si as internal (stoichiometric) standard element. Initial Pb correction was not performed, because the 204 Pb signal is swamped by the isobar 204 Hg present in the ICP carrier gas. Data were carefully filtered, using the cut-off discordance criteria of +30%/-5%. Raw data were reduced offline using Iolite 4 software (Paton et al. 2011), including all the error calculations and propagation, and employing the VizualAge data reduction scheme of Petrus and Kamber (2012). The secondary Plešovice standard zircon yielded a mean 206 Pb/ 238 U age of 338.2 ± 1.7 Ma, in agreement with its accepted age. All the data were plotted employing the free software IsoplotR (Vermeesch 2018), whose radial plot algorithm was employed for calculating the minimum ages. REE data concentrations were normalized against CI Chondrite values of McDonough and Sun (1995). U-Pb results and REE concentrations are reported in the Supplementary tables 2 and 3, respectively.
The Ti-in-zircon thermometry was applied to multiple metamorphic grains from the studied granulite xenoliths, using the calibration of Ferry and Watson (2007,).
Calculations were performed using Ti activities of 0.6 and 1.0, for rutile-free and rutile-bearing assemblages, respectively (Table 1). Since the obtained metamorphic ages are variable in each sample and among them, temperatures were derived based on the HREE concentrations and Th/U values (see above).
The detrital age comparison between the studied xenoliths and the putative sources is readily performed employing multidimensional scaling (MDS), a statistical tool that uses the dissimilarity matrix (D) of the Kolmogorov-Smirnov test to create a map of 'distances' proportional to the dissimilarity, thus the difference, between the age populations (Vermeesch 2013). Unimodal, synthetic samples with normal distributions, corresponding to the main source ages of detrital zircon (from Grenville to the Palaeocene-Eocene), were also employed to aid in visualizing age trends in the MDS diagram. MDS maps are readily created using IsoplotR (Vermeesch 2018).

Results and age interpretations
Sample LXES-1001 is a plagioclase-perthite-garnetsillimanite banded granulite. Fifty zircon grains were selected for dating ( Figure 5A). After filtering, 36 of them were considered useful for interpretation. One zircon yielded a 1,183 ± 30 Ma age, another yielded an Ordovician age of 463 ± 9.7 Ma age ( Figure 5A inset). Some grains, although slightly discordant, are Mesozoic, with remarkable isolated ages at ca. 160, 125 and ca. 80 Ma. Most of the obtained ages (30 out of 36, Figure 5A) are, however, Cenozoic, ranging from 9.6 to 52 Ma. The minimum age obtained in this sample, is of 9.7 ± 0.2 Ma. Chondrite-normalized zircon REEs help to identify the nature of these youngest crystals. In fact, late Miocene zircon minimum age is indicative of the granulite metamorphism, as the corresponding HREE pattern in these zircons are flat and depleted (Suppl. Figure 3A), features that are characteristic of those zircon domains grown together with garnet, during high-grade metamorphism (e.g. Rubatto 2002;Harley et al. 2007).
The xenolith sample LXES-1003 is a faintly banded plagioclase-garnet granulite, with minor quartz and K-feldspar. Most zircon grains separated from this sample are very small, less than 60 µm in length. Fifty-five zircons were considered, out of 80 initially analysed, because of their high-discordance value. Those filtered ( Figure 5B) are concordant to nearly concordant, revealing the presence of some inherited ages (ca. 1200 Ma for the oldest zircon, but also age distributions of ca. 990 Ma, some Palaeozoic ages between ca. 300 and 400 Ma, and few Mesozoic ones were documented). Cenozoic zircons prevail in abundance, with some slightly discordant analyses, while the youngest has an age of 5.1 ± 0.2 Ma. The latter, together with few zircons of ca. 6 Ma (Table 1) are also characterized by a very low Th/U ratio (0.01-0.02), interpreted as indicative of the high-grade metamorphism that characterized this xenolith.
Xenolith sample XES-1007 is a garnet-plagioclase and sillimanite-bearing sample, considered representative of a para-derivate. Once filtered to meet the discordance criteria, 61 analyses were considered as valid of the original 80 (Suppl. table 1). Few analyses yielded Grenvillian ages, straddling the concordia curve at ca. 980 and ca. 1100 Ma; others are Neoproterozoic to early Palaeozoic ( Figure 5C). Although few zircons yielded Mesozoic ages, most of them, concordant within the analytical error, are Tertiary, mostly ranging between ca. 10 and ca. 28 Ma ( Figure 5C, inset, and Table 1), with a minimum age calculated in 9.8 ± 1 Ma. Unsurprisingly, the latter group shows an HREE metamorphic pattern typical of those zircons grown in the presence of garnet during granulite facies metamorphism (Suppl. Figure 3C).
Xenolith LXES-1008 is mostly made up of plagioclase, perthitic K-feldspar and garnet, although minor sillimanite and some orthopyroxene are also present. After separation, 60 zircon crystals were chosen for U-Pb geochronology, 48 of which passed the discordance criteria. The obtained ages range from the Palaeozoic (Permian to Silurian ages, Figure 5D) to some Mesozoic. Tertiary ages are mostly comprised between ca. 50 and ca 13 Ma ( Figure 5D, inset) with a minimum age calculated in 12.9 ± 0.6 Ma. Suffice to say, Miocene to latest Oligocene zircon REE patterns (Suppl. Figure 3D) are indicative of a substantial depletion in HREE, typical of metamorphic recrystallization under high-grade conditions in the presence of garnet. For this sample, 30 more zircon crystals were analysed at a smaller spot of 17 μm, employing a Neptune Plus multi-collector (see above in Methods). The obtained ages ( Figure 5E and Table 1) mimic the same distribution (without the REE information, though), in this case with a minimum age of 12.7 ± 0.15 Ma.
The sample XES-1009 is a granulite sample made of plagioclase, orthopyroxene and clinopyroxene, perthite, and quartz. Based on its modal composition, it is interpreted as a meta-quartz diorite. Seventy-five zircon grains were selected for geochronology, 59 of which passed the concordia filter. Few of the dated zircon grains are inherited, probably indicative of basement ages ( Figure 5F). They yield Palaeozoic ages (ca. 310, 460 and 530 Ma). There are only few mid-Cretaceous zircons, in the range 90-100 Ma. The remainder zircon grains are Cenozoic, generally younger than 40 Ma ( Figure 5F, inset). The youngest zircons are probably indicative of the metamorphism that affected this sample: although we do not have direct chemical evidence, provided that the REE plot of Suppl. Figure 3E only indicates that the zircon patterns are all igneous, which is a common behaviour for samples that do not host garnet, we hypothesize that the cluster around 23-24 Ma corresponds to its igneous crystallization. The younger, high-grade metamorphism (youngest zircons in this sample are 4.6-6.8 Ma, Suppl. table 2) caused partial reset and Pb loss in some of the older zircon grains.
LXES-1013 is a plagioclase-garnet-orthopyroxene granulite. This sample was also double dated by LA-ICPMS (80 analyses, of which 55 passed the concordia filters) and LA-MC-ICPMS (35 further analyses, aimed to target the outermost zircon rims). Zircon ages straddle the concordia curve from ca. 190 to ca 6 Ma ( Figure 5G). Only few zircons yield Mesozoic ages (ca. 150 and ca. 190 Ma, and a small group of Late Cretaceous ages, Suppl. table 2 and Figure 5G). Most of the analysed zircon grains yield Cenozoic ages, younger than ca. 40 Ma. The minimum age calculated on analyses performed by LA-ICPMS is 6.23 ± 0.3 Ma, which is interpreted as the time of granulite facies metamorphism, also thanks to the REE distribution patterns (Suppl. Figure 3F). The metamorphic age is further refined measuring the outermost zircon rims at 17 μm spot by LA-MC-ICPMS, obtaining a mean age of 4.4 ± 0.1 Ma, interpreted as indicative of the timing of the last high-grade metamorphic peak.
LXES-1015 is a plagioclase-orthopyroxene-quartzgranulite, with a mineralogy consistent with an intermediate-mafic igneous protolith. Twenty-eight analyses, over the analysed 55 zircon grains, yielded useful ages in term of concordance ( Figure 5H). The obtained ages are mostly Cenozoic, with six analyses straddling the concordia around 100 Ma and all the remainder zircon ages younger than 60 Ma. The absence of garnet or other accessory phases growing together with zircon precludes the use of REE patterns, that are all igneous (Suppl. Figure 3G) as an indication of which zircons grew during high-grade metamorphism. The minimum age calculated on this sample, of 10.2 ± 0.5 Ma, is tentatively interpreted as indicative of the time of granulite metamorphism. There are several possibilities to establish the protolith's igneous age of this sample: (1) those zircons around 100 Ma (mean calculated age is of 99.6 ± 1 Ma) are indicative of the minimum crystallization age of the parental magma. In this case, all the sample zircons remained in a hot environment that continuously allowed Pb diffusion, until the high-grade metamorphism at ca. 10 Ma or (2) the ca. 100 Ma zircon dates correspond to crystals inherited from an unknown source and the sample has a Cenozoic crystallization age, probably indicated from the age cluster at around 30-40 Ma ( Figure 5H and Suppl. table 2). The latter is our preferred interpretation as it will be explained in the Discussion.
LXES1016 is a plagioclase-perthite-garnet-quartz and sillimanite-bearing banded granulite. Sixty short and stubby, somewhat rounded, zircon grains were chosen for U-Pb geochronology. Of those, 46 are useful for geological interpretations after concordance filtering. Most of them are concordant, straddling the concordia curve from ca. 960 Ma to the minimum calculated age of 13.7 ± 0.4 Ma ( Figure 5I). This last age, combined with the REE zircon patterns that are indicative of crystallization together with garnet, can be interpreted as representative of the high-grade metamorphism (Suppl. Figure 3H).
XES-1013 is a faintly banded granulite xenolith, made up of feldspar and two pyroxenes, with titanite as a relatively abundant accessory mineral. The absence of quartz suggests it had a mafic igneous protolith. Due to the absence of zircon, titanite was chosen to be dated. Most of the resulting U-Pb analyses are discordant (Suppl . Table 3), yielding a lower intercept age of 4.31 ± 0.65 Ma (n = 17, concordia MSWD 4.7, Figure 6). We don't have any clue about the original igneous age of this sample. The straightforward interpretation is that the obtained age is indicative of the high-grade metamorphism that affected this sample. Just as a tentative hypothesis, we advance the idea that this sample could represent a remnant of the intraplate, alkaline magmas present in the lower crust (e.g. Luhr et al. 1995) and that got trapped during the ascent.
The measured concentrations of Ti (15-140 ppm) in the metamorphic zircon correspond to temperatures of 785-1163°C and are consistent with metamorphic crystallization at granulite-facies conditions (geothermometer of Ferry and Watson, 2007). Temperature data for each sample are provided in Suppl. Table 1.

Discussion
U-Pb zircon studies of individual lower crustal xenoliths usually yield multiple ages (e.g. Rudnick and Williams, 1987;Rudnick and Cameron 1991;Fernández-Suárez et al. 2006), which are interpreted as the record of major tectonic and/or magmatic events that contributed to lower crust formation in the region where the xenoliths were collected. Lower crust evolution of individual regions varies as a function of proximity trough time to: plate boundaries, intra-plate rift zones, orogenic belts, and mantle plumes among other factors. Commonly, a single xenolith, either with igneous or sedimentary protolith, may yield multimodal ages that attest at least part of the ages of tectonomagmatic events previously documented in the surface geology, but there are cases where some events recorded in a granulite are not currently expressed or recognized in the surface geology (e.g. Peltonen et al. 2002;Thakurdin et al. 2019). The study of a set of xenoliths with varied protoliths and mineralogical compositions from a single volcano, or of xenoliths collected at several nearby vents in a volcanic field, increases the chances of obtaining a complete record of the major events recorded in that portion of the lower crust. Age clusters in each xenolith have been attributed to (1) magma emplacement during underplating (in the broad sense used by Thybo and Artemieva 2013) or (2) as products of one or several granulite-facies metamorphic pulses either in igneous or sedimentary protoliths. For instance, Rudnick and Cameron (1991) report one mafic xenolith with a single zircon age cluster that broadly corresponds with the isotopic age of the host lava, age that is taken as evidence of pulse of granulite-facies metamorphism contemporaneous with the volcanism that brought that granulite to the surface. In another example, a granulitefacies metaquartzite yielded several age clusters, where the youngest age was interpreted as the age of metamorphism, whereas other ages were taken as those of detrital zircons that survived the high-grade metamorphism (Arndt et al. 1991).
Several questions arise from the U-Pb ages obtained in the studied set of xenoliths, which clearly are not representative of the whole variety of lithologies in El Toro xenolith assemblage, as they were selected as a function of the probability of containing zircons and on mineral parageneses consistent with a P-T of equilibration at the lower crust. The questions we try to answer in the following pages are as follows: (1) Which is the age and significance of the UHT granulite metamorphism in the Mesa Central? (2) where did the original, now granulitized sediments, accumulated and how they reached the lower crust under the Mesa Central, several hundreds of km away from the trench where they were subducted? (3) Do these xenolith ages help to support any of the geotectonic reconstructions advanced so far for the Phanerozoic assemblage of Mexico (e.g. Campa and Coney 1983;Sedlock et al. 1993;Keppie 2004;Martini et al. 2011;Goodell et al. 2017)? and (4) Was there a ca. 1.2 Ga magmatic pulse that produced the precursors of metaigneous granulites found in other nearby xenolith localities in San Luis Potosí (Figure 2), Mexico (e.g. Schaaf et al. 1994)?
(1) Age of UHT metamorphism beneath the Mesa Central The zircon ages presented in this work, together with the zircon chemistry obtained by LA-ICPMS, allow some inferences within the general context of the tectonomagmatic evolution of Central Mexico. Most of the studied xenoliths, regardless of the nature of their protoliths, yield few Palaeozoic to Neoproterozoic ages, together with more abundant Mesozoic and Cenozoic ages. Based on zircon chemistry, all older zircon crystals (Neoproterozoic to Mesozoic) can readily be interpreted as indicative of igneous pulses, whereas the younger zircon grains, especially those yielding late Oligocene to late Miocene ages, are mostly metamorphic in origin. This is especially true for zircon crystals retrieved from xenoliths whose mineralogy includes Al-rich metamorphic phases such as garnet and sillimanite, which are interpreted as metasediments, where depleted HREE patterns reflect zircon growth at granulite facies conditions. Each of the studied xenoliths yields slightly different ages for the young UHT granulite facies event under the Mesa Central. Two xenoliths produce metamorphic ages of ca. 13 Ma, three of ca. 10 Ma, and four of ca. 5-6 Ma.
Overall, considering all the metamorphic zircon grains characterized by depleted HREE patterns, their normal age distribution varies between ca. 6 to ca. 27 Ma, with a mean age value of ca. 16 Ma. Only three metamorphic zircon ages are interpreted as inherited and correspond to Late Cretaceous, Triassic or Cambrian, and will not be further discussed in this work ( Figure 7A). The ca. 6 to ca. 27 Ma age range is probably indicative of a protracted metamorphic event in the lower crust. It is believed that the underplating caused a thermal pulse that gradually ascended towards the surface and/or, as it is argued by Bryan et al. (2008), the locus of mafic magma intrusion in the lower crust gradually migrated upwards during the Sierra Madre Occidental magmatism in the region. Luhr et al. (1995) based on the study of megacryst and feldspathic granulite-bearing hawaiites from Los Encinos volcanic field (ca. 10-13 Ma, ca. 100-120 km NW of El Toro) argued that mantle-derived basaltic magmas stagnated and/or slowly ascended thru an incipiently extended crust during the Oligocene and Miocene in the Mesa Central. We hypothesize that a similar progression of magma emplacement towards shallower depths may have occurred during the Miocene and Plio-Quaternary intraplate magmatic pulses, which occurred as Basin and Range extension progressed in the Mesa Central. Therefore, zircon ages depend upon the relative equilibration depth of each studied xenolith, as their isotopic systematics closed at different times. All those zircon crystals interpreted as the result of a metamorphic event have closure temperatures in the UHT domain, ranging between ca. 900 to over 1,100°C (Suppl . table 1  and Suppl. table 2 The inferred age of the UHT pulse agrees with the findings of Hayob et al. (1989) who postulated the existence of 'recent' (younger than Oligocene) granulite metamorphism recorded in the lower crust of Central Mexico, as a result of continuous heating caused by mafic magma ascent of Cenozoic to Quaternary age, widespread in the whole region (Luhr et al. , 1995Pier et al. 1989; Aranda-Gómezet al. 2007a). The less abundant xenoliths, analysed for this study, for which an igneous protolith is envisaged (i.e. XES-1009 and LXES-1015), are interpreted to have crystallized from magmas emplaced in the lower crust during the Oligocene-late Eocene (23-24 and 30-40 Ma, respectively) and then metamorphosed during the middle Miocene and/or Plio-Quaternary intra-plate magmatism of central and northern Mexico (Aranda-Gómezet al. 2007a). They are thus deep-seated plutonic rocks equivalent to those volcanic magmas recognized in the surface in the adjacent Sierra Madre Occidental large silicic igneous province (e.g. Orozco-Esquivel et al. 2002;Bryan et al. 2008), which partially covered the Mesa Central (Aranda-Gómez et al. 2003;Nieto-Samaniego et al. 2007). Nearby, at La Herradura, Sieck et al. (2019) reported a garnet-bearing rhyolite with a crystallization U-Pb age of 31.5 Ma, which is interpreted as pertaining to the same magmatic event, which the cited authors associate to partial melting of a parental andesite in the lower crust due to extensional mechanisms. However, it must be stressed that garnetbearing granulites in the San Luis Potosí xenolith localities form between 2% and 16% of the xenolith population (including spinel-lherzolites) and in at least one locality (Joya Prieta maar at the Santo Domingo volcanic field) they are absent (Aranda-Gómez 1982).
The young granulite facies metamorphism is also responsible for generating partial resetting of younger igneous ages recorded in the metasedimentary xenoliths. As most of the studied granulite xenoliths had metasedimentary protoliths, it must be kept in mind that the whole process of igneous zircon crystallization, erosion, transport, and sedimentation requires a significant time-lapse before the sediments reached the physical conditions needed for UHT granulite metamorphism at or near the base of the crust (e.g . Harley 1989;Bohlen 1991;Castro et al., 2013).
2) Origin of the metasedimentary protoliths The presence of granulitized sedimentary protoliths at El Toro, metamorphosed at depths of ca. 35-45 km (0.9-1.2 GPa according to Aranda-Gómez 1982;Hayob et al. 1989) and temperatures of ca. 1000°C, under nearly anhydrous conditions, pose interesting questions about the sediment provenance. Their existence requires a mechanism, suitable to introduce supracrustal material at considerable depth, as compared with their supracrustal depositional environment. While their chemical nature, deeply modified by granulite metamorphism, makes it hard to precisely establish their original protoliths, the abundance of Al-rich mineral phases in the granulite allows to assume that they were originally subaerially derived siliciclastic sediments, very similar to present-day coastal area sediments, produced by weathering and erosion of continental rocks. Thus, we propose that those sediments were subducted and emplaced in the lower crust of Central Mexico, prior to their UHT metamorphism. The Pacific subduction zone of western Mexico, active at least since the Early-Middle Jurassic (e.g. Parolari et al. in press), and until the disruption of the Farallon plate in the Miocene (Boschman et al. 2018), is the best candidate to act as mechanism of sediment recycling and introduction underneath the Mesa Central region. Tectonic erosion is seen as a plausible mechanism envisaged to erode continentally derived sediments either from a forearc basin or from an accretional prism, as the subducting slab is a suitable vehicle to drag them, conveyor-like to an appropriate depth, a process broadly known and described along the whole Meso American trench (e.g. Vannucchi et al. 2004;Straub et al. 2015;Cavazos-Tovar et al. 2020). Another feasible source of at least part of the subducted sediments is the thin veneer of deep-water pelagic sediments riding atop the subducting slab. The processes of tectonic relamination and sedimentary diapirism, are invoked as possible mechanisms of sedimentary incorporation in the lower crust (e.g. Currie et al. 2007;Chapman 2021). Combining the inferred metamorphic age, roughly spanning through the latest Oligocene to all the Miocene time (ca. 26-6 Ma), with the main age groups observed for the sole inherited zircon found in metasedimentary xenoliths (kernel density estimator diagram of Figure 7B) suggests that the original sediments were deposited during the mid-Eocene-Oligocene time, and they were caught in the subduction-relamination process thereafter.
Comparing the zircon ages interpreted as indicative of detrital origin ( Figure 7B), with modern detrital sediments along the Pacific coast of Mexico, one can propose hypotheses that may constrain the possible sources of the subducted sediments. To achieve this task, the similarity among the xenolith detrital ages was tested against selected detrital samples taken from the literature: some belonging to northern (Sharman et al. 2015) and southern Baja California (Fletcher et al. 2007), once the peninsula was restored to its position before the Gulf of California opening in the Miocene time (Lonsdale 2005;Ferrari et al. 2013), prior to the subduction end in the region; and others detrital samples belonging to the Central Mexico Pacific coast, in the region between Puerto Vallarta to Zihuatanejo (Cavazos-Tovar et al. 2020) ( Figure 8A). This comparison is readily achieved with the MDS diagram of Suppl. Figure 4, where similarities can be observed with continuous (first order) and discontinuous lines (second order). El Toro xenolith detrital age distribution (red dot of Suppl. Figure 4) has first-order similarities with samples MICH16-18 and GUE16-01, and second order with nearby samples MICH16-09 and MICH16-13. All these samples were collected along a 100 km long coastal stretch located near Zihuatanejo ( Figure 8A). Although the selected samples are present-day sediments, the comparison is valid, because that coastal portion of Central Mexico lacks recent magmatism and it is isolated from other regional hydrographic systems (e.g. Maderey-Rascón et al. 1990), thus it lacks possible younger zircon sources. Magmatic activity in the nearby Balsas catchment region extinguished since the migration of Chortís Block towards SE during the Cenozoic (e.g. Ferrari et al. 2014;Martini et al. 2016).

Terrane limits in Central Mexico
The last question that must be addressed is about the nature of the continental crust in Central Mexico, underlying El Toro and surrounding areas. The absence of nearby outcrops, in which basement rocks are exposed, only allows to speculate. There is a general consensus that a large portion of Mexico's territory is a complex mosaic of tectonostratigraphic terranes accreted at different times to the southern end of North America during the Phanerozoic (e.g. Campa and Coney 1983). Several suspect terrane maps of Mexico have been published in the past three decades (Campa and Coney 1983;Sedlock et al. 1993;Dickinson and Lawton 2001;Keppie 2004;Centeno-García et al. 2008). Each one displays a variable number of groupings of basement rocks that presumably represent individual terranes. Likewise, the location of the inferred tectonic boundaries (mostly inferred faults) between the terranes varies, sometimes widely. All the currently available terrane maps of Mexico share a few common features, such as (1) the location of North American Precambrian basement rocks and overlying cover rocks in the northwestern part of the country, (2) the existence of the vast Mesozoic Guerrero superterrane in the west-central part of Mexico, and (3) the presence of Grenville-age rocks (ca. 1 Ga) either of the Sierra Madre terrane (sensu Campa and Coney 1983) or in what is been called the Oaxaquia microcontinent (Ortega-Gutiérrez et al. 1995). Oaxaquia is an alleged continuous basement block that forms a 'backbone' broadly underlying the Sierra Madre Oriental Fold and Thrust Belt. The Oaxaquia block extends from NE Mexico to the Pacific coast in the south. It must be noted that the inferred extent of Oaxaquia is only based on (1) the existence of few, widely scattered, isolated outcrops over a broad area; (2) scant reports of basement samples recovered from oil exploration wells; and (3) model ages obtained from lower crustal xenoliths in two widely separated volcanic fields, one in northern Mexico (Camargo Volcanic Field, Chihuahua) and two separated volcanic clusters in San Luis Potosí (Figure 2: Ventura-Espíritu Santo and Santo Domingo volcanic fields: Aranda-Gómezet al. 2007a). Thus, U-Pb ages of zircons recovered from xenoliths of El Toro, a vent in the Ventura-Espíritu Santo volcanic field, offer the opportunity to test some aspects of the terrane models proposed for Central Mexico.
All published terrane maps show a tectonic boundary between the Guerrero and the Oaxaquia Block/Sierra Madre terrane in the general area where Cerro El Toro is located. That boundary is interpreted by Dickinson and Lawton (2001) as a collision/accretion suture formed during the Early Cretaceous. Presence of overlap assemblages atop the basal rocks, which very likely cross terrane borders and hide them in Central Mexico, probably led Sedlock et al. (1993) to show all the terrane boundaries in the region as straight dashed lines.
Igneous and metamorphic ages recorded in Oaxaquia's rocks in an outcrop located east of the Mesa Central include two major Mesoproterozoic igneous suites: 1235-1115 and 1035-1010 Ma, as well as the record of a granulite-facies pulse of metamorphism at ca. 990 Ma. Mafic porphyritic dikes intruded the granulite in the eastern portion of that crustal block at ca. 546 Ma (Trainor et al. 2011). Broadly similar ages have been documented near the southern end of the Oaxaquia block  near the Pacific Coast of Mexico. Thus, it is here assumed that these are the characteristic ages of Grenville-age rocks in Central Mexico. The supposed existence of a Grenvillian basement underneath El Toro at a longitude −101°50' is also maintained by Valencia-Moreno et al. (2021), which in practice extends it for the whole Central Mexico up to the Arperos suture of the Guerrero terrane (c.f. Martini et al. 2011). Goodell et al. (2017) recently proposed the existence of the Western Chihuahua-Mesa Central megablock, a cratonic block inferred from the existence of a gravity minima that extends from Chihuahua to a region near the Trans-Mexican Volcanic Belt and includes El Toro region. In their model, Goodell and collaborators assume that the block is formed by Proterozoic (1.8-0.9 Ga) continental crust.
The Guerrero superterrane, the most extensive suspect terrane in Mexico, is interpreted as formed by the amalgamation of several terranes, including volcanic arc(s) of Pacific affinity (e.g. Centeno-García et al. 2008) and magmatic rocks related to a back-arc basin (Martini et al. 2011). The general ages of its magmatic rocks are Jurassic-Lower Cretaceous, ca. 151-144 Ma: Martini et al. 2011), predating the time of accretion (ca. 113 Ma, Martini et al. 2016). Its evolution is related with a convergent margin outboard of the Pacific coast of Mexico, and the general role it played in the evolution of the Mexican Orogen, a mountain building period that produced shortening in Cerro El Toro region, which is, broadly speaking, in the hinterland of the orogen, whereas the Sierra Madre Oriental Fold and Thrust belt, located farther east and northeast of El Toro represents the foreland.
The data we present in this paper are evidently at odds with the aforementioned interpretations that support the existence of a Proterozoic basement in Central Mexico, because among the samples we studied there is no evidence of Grenville-age or older rocks in the El Toro xenoliths, or evidence of Grenville-age granulite metamorphism or emplacement of igneous rocks at ca. 1250 Ma (e.g. Schaaf et al. 1994, from the Ventura and Santo Domingo cluster of xenolith localities, Figure 2). The existence of a Grenvillian or older crust underneath El Toro and, for what it is known, in the whole region of the southern portion of the Mesa Central cannot be, thus, sustained with the current dataset ( Figure 7B and Suppl. Figure 4).

Major pulses of magmatic activity in Central Mexico
An eastward dipping subduction zone has existed along the Pacific coast of Mexico during the last 220 Ma. The dip angle of the Benioff zone has varied throughout this time and periods of crustal shortening or extension have been correlated with steep and flat angles in the consumed oceanic slab. South of latitude 21° an ESEtrending seismic tomographic cross section displays a continuous record of the consumed slab, whereas north of that latitude the positive wave speed anomaly interpreted as the consumed slab is missing in the upper 400 km of the mantle. This feature corresponds to the end of subduction north of latitude 21° after the arrival of the Pacific-Farallon ridge to the trench during the Miocene and the beginning of a new tectonic regime at and around the Gulf of California (Boschman et al. 2018). The protracted subduction regime in western Mexico, as well as the separation of the Baja California Peninsula from the mainland, paired with crustal thinning and ocean floor generation in the Gulf area contemporaneous with crustal extension in the Mexican Basin and Range province are consistent with magmatic activity since the Late Triassic in the region.
Scattered throughout the Mesa Central, north of the San Luis-Tepehuanes fault zone, exist isolated outcrops of Mesozoic and Palaeogene plutonic (Díaz-Bravo et al. 2021) and volcanic rocks. Some of the Jurassic volcanic rocks have been attributed to the Nazas Arc. Of particular interest because their distance to Cerro El Toro are two rhyolitic ignimbrites exposed at La Ballena range and at El Tepozán. Both yielded a U-Pb zircon ages of ca. 168 Ma (Díaz-Bravo et al. 2021). La Ballena is located 48 km S24W from Cerro El Toro and El Tepozán is 68 km N34E. These rocks have been attributed to the Late Triassic to Middle Jurassic Nazas continental magmatic arc, which developed along the western margin of Pangea (Bartolini et al., 2003).
The Late Cretaceous-Eocene magmatic pulse evidenced by the existence of several plutons in the 72-45 Ma age range in the northern part of the Mesa Central (Díaz-Bravo et al. 2021), which are consistent with the Cretaceous-Eocene Mexican Magmatic Arc, that have been interpreted as formed in a 'typical continental arc' by Valencia-Moreno et al. (2021). Scattered Oligocene (32-29 Ma) subvolcanic intrusions also occur in the same region, broadly equivalent in age to the thick volcanic cover south of the San Luis-Tepehuanes fault system. Again, the bias in the selection of the xenoliths investigated may be the reason that some of these major magmatic pulses are not seen in data presented in this report. Clearly, a broader set of xenolith from El Toro, as well as other nearby localities (Figure 2), must be investigated to solve this problem.

Conclusions
Zircon crystals belonging to granulite-facies xenoliths collected at El Toro, San Luis Potosí, Mexico, testify a young (mostly Miocene to late Oligocene, ca. 6-27 Ma) UHT, protracted metamorphic event in the lower crust. This affected both igneous (quartz diorite of Eocene to Oligocene age) and sedimentary (siliciclastic protoliths of Eocene age, and mostly Mesozoic to Palaeozoic-Proterozoic inherited components) protoliths. The reconstructed evolution suggests that the original sediments were introduced to the lower crust throughout the Pacific subduction zone, and they are like zircon crystals recovered from modern beach sediments found along the Pacific coast of southwestern Mexico (Michoacán to Guerrero coastal stretch). The absence of Grenville-age xenoliths or granulite zircon grains suggests that the unexposed lower continental crust in El Toro is young and mafic, like the Arperos oceanic crust of Mesozoic age, now found as remnant in the Arperos suture zone of the eastern Guerrero terrane.