The sample inspection was carried out by a Hitachi TM-1000 scanning electron microscope with an energy dispersive spectrometer (SEM-EDS). Further mineralogical determinations were carried out by means of shortwave infrared (SWIR) using a portable LabSpec Pro Spectrophotometer (Analytical Spectral Devices, Inc.). Visible and near-infrared reflectance of samples, for the spectral range between 350 and 2500 nm (with a sampling interval of 2 nm and a 0.1 s single scan), was measured using an internal light source and sensor. The spectral resolution was 3 nm in the 350-1000 nm range and 10 nm in the 1000-2500 nm range. The SWIR wavelength region (1300-2500 nm) was used for our determinations in order to attain the necessary sensitivity to OH, H₂O, CO₃, SO₄, CH, and NH₄ bonds (Thompsonet al., 1999,2009). Mineral identification was based on the wavelength of absorption and the shape of spectra by using the available spectral libraries and tables (Spectral International Inc., 1994). Both SEM and SWIR equipment are available at the Instituto de Geofísica of the Universidad Nacional Autónoma de México (UNAM). The images obtained through the petrographic study by means of SEM are shown inFigures 2 and3.

Figure 2
Microstructural, mineralogical, and paleobiological aspects of structures associated with gas venting in advanced argillic associations within a steam-heated ground environment in the Ixtacamaxtitlán epithermal deposits. In all the pictures, fungal hyphae and other possible biological structures are replaced by opal or cristobalite. (A) Upper view of the structures in hand sample showing the high porosity of vents, as tubular cavities, at the paleosurface. (B) Lateral view of (A); notice the columnar arrangement of crystal aggregates (mostly alunite) within each tubular cavity on a relatively low porosity kaolinite + alunite association hat includes mm-sized alunite-lined vugs. (C-F) Secondary electron images of the mineralogical and fossil biological content within tubular cavities in (A) and (B). (C) Fungal hyphae attachment parallel to earlier rhombohedral alunite crystals. (D) Close-up image from (C); notice dissolution pits on the surface of alunite crystals where fungal hypha were attached. (E) Hyphae tramlines that show typical structures of fungal networks as septa or anastomosis in septate hyphae for a possible reproduction; as in (D), notice pitted surfaces on alunite crystals due to dissolution. Also notice collapsed hyphae conterminous to bright white, non-collapsed hyphae; this means that collapsed hyphae did not contain cytoplasm because they were already dead before their fossilization whereas non-collapsed hyphae were fossilized while still containing cytoplasm. (F) Fine fungal hyphae spread between opal-covered alunite crystals (hence the smooth surfaces) and on a delicate cobweb-like microstructure that is suggestive of bacteria or archaea biofilms. In all these pictures, fungal hyphae and other possible biological structures are replaced by opal or cristobalite.

Figure 3
Secondary electron images of microstructural, mineralogical, and paleobiological aspects of structures associated with gas venting in advanced argillic associations within a steam-heated ground environment in the Ixtacamaxtitlán epithermal deposits. In all the pictures, fungal hyphae are replaced by alunite or kaolinite. (A) Basal region of a gas-vent tubular structure where a mesh of fungal hyphae developed among fine and particulate kaolinite and alunite debris. (B) Close-up image of (A) showing surficial concave depressions or grooves on rhombohedral alunite crystals due to hyphae-driven dissolution; notice the good cleavage of alunite on {0001}, highlighted by dissolution. (C) Mycelial cord structures formed by a parallel assemblage of hyphae; notice curved surfaces developed on alunite crystals due to partial dissolution. (D) Close-up image of the mycelial cord in (C) showing a set of no less than four hyphae and a euhedral alunite microcystal on them, as part of a second generation of alunite. (E, F) Dense networks of interconnected mycelia that show their characteristic particle aggregation (kaolinite fragments due to biobrecciation). (G) Kaolinite aggregate with a corrosion bay adjacent to a mycelial tramline that retains alunite debris (biobrecciation) over superficial hyphae; notice the flattening of some hyphae due to their collapse and seeFigure 2E for an explanation. (H) Chalcopyrite grain on kaolinite; the corrugated surface in the upper part of the picture, signaled with an orange arrowhead, could be due to remains of a fossilized biofilm. In all these pictures, fungal hyphae are replaced by alunite or kaolinite.

The pH values for the hydrothermal waters can be assessed using the present geothermal waters from the Los Azufres field (Michoacán, Mexico) as an analog for the fossil system. The likelihood of such waters for this use relies on the variety of their geochemical characteristics (particularly in composition and pH), origin, and evolution, as they were classified among steam-heated, mature, and peripheral waters (González-Partidaet al., 2005). Therefore, a full set of chemical composition of hot spring waters located in and around the Los Azufres geothermal field presented byGonzález-Partidaet al. (2005) has been used for this purpose. Water samples have been equilibrated with kaolinite and alunite at temperatures between 25 and 150 °C using the code PHREEQC v.3 (Parkhurst and Appelo, 2013) and the Lawrence Livermore National Library thermodynamic database. The range of calculated pH and amorphous silica saturation index (SI) values for each of the samples are shown inTable 1. The pH ranges thus calculated and represented inFigure 4 for the different types of waters (all of them, in principle, likely to have occurred in the epithermal paleosurfaces of Ixtacamaxtitlán) allowed us to choose the most representative samples for each type. The Cumbres II and Azufres I samples were selected to represent the bimodal behavior of steam-heated waters at the Los Azufres geothermal field. Their representative character was determined upon the most frequent range of calculated pH, as shown inFigure 4. The Zimirao and Casa Lázaro Cárdenas samples were chosen as the geochemically closest representatives of mature and peripheral waters (with respect to a steam-heated water system), respectively. The four selected representative samples were then used to correlate their pH and temperature in order to obtain likely estimations of the saturation conditions for amorphous silica (opal), in the understanding that alunite and kaolinite were in equilibrium, as shown inAppendix 1 andFigure 5.

Table 1
Calculated pH and silica saturation index ranges for the geothermal waters from the Los Azufres geothermal field (Michoacán, Mexico) in equilibrium with kaolinite and alunite at temperatures between 25° and 150 °C, after data fromGonzález-Partidaet al. (2005).

Figure 4
Range of calculated pH for actual steam-heated, mature, and peripheral waters form the Los Azufres geothermal field (Michoacán, Mexico) in equilibrium with kaolinite and alunite. Data on which the modeled waters are based were obtained byGonzález-Partidaet al. (2005). These data are used as a reasonable approximation to the environment that prevailed during the formation of the epithermal paleosurfaces in the Ixtacamaxtitlán area. Data derived from hydrogeochemical calculations are displayed inTable 1. Among all samples, those indicated in bold italic typeface were selected for use inFigure 5. Two samples from steam-heated waters were selected as the most representative for the bimodal distribution of pH data; the most recurrent pH ranges are indicated with dashed lines. In contrast, two samples from mature and peripheral waters were selected as most representative for the purpose of this study because they do not stray too far from the common range of pH for acidophile and thermophile fungi (between 3 and 5) and their pH is similar to at least that of another sample.

Figure 5
Correlation between the saturation index of amorphous silica [SiO₂(am) SI] and temperature (ºC) (upper left), between SiO₂(am) SI and pH (lower left), and between temperature and pH (right), which displays the calculated amorphous silica saturation index for each selected sample inFigure 4. The subsaturation/supersaturation boundaries of amorphous silica are indicated with black dashed lines. The yellow field represents the most reasonable range of temperatures for the installation of fungi in this environment, and the blue shade brackets are the most reasonable range of pH. Numbers 1, 2, and 3 on black circles occupy tentative spaces that are reasonable representatives of the three stages in the steam-heated paleosurface. Number 2 on gray circles represents non-unequivocal possibilities for stage 2. The possible main paths (A, B, C, D, and AB) for the trajectory between a starting scenario with water supersaturated in amorphous silica into another with water subsaturated in amorphous silica (determined by means of petrographic criteria) are indicated by white empty arrows. Such paths allowed us to constrain the most likely pH-temperature field for water that supported fungal consortia during stage 2 (box dashed in red). As stage 3 represents a return to essentially the same conditions of stage 1 (supersaturation of water in amorphous silica), its position in the SiO₂(am) SI-pH-temperature field is not further addressed.

The alunite samples for geochronological determinations were crushed in a ring mill, washed in distilled water and ethanol, and sieved when dry to −40+60 mesh. Appropriate mineral grains were picked out of the bulk fraction. The samples were wrapped in aluminum foil and stacked in an irradiation capsule with similar-aged samples and neutron flux monitors (Fish Canyon Tuff sanidine [FCs], 28.201 ± 0.046 Ma;Kuiperet al., 2008). The samples were irradiated in October 2017 at the McMaster Nuclear Reactor in Hamilton, Ontario, Canada in a shielded can for 6 MWH in the medium flux site 8E. Analyses (n = 39) of 13 neutron flux monitor positions produced errors of <0.5% in the J value. The samples were analyzed at the Noble Gas Laboratory, Pacific Centre for Isotopic and Geochemical Research, University of British Columbia, Vancouver, British Columbia, Canada. The mineral separates were step-heated at incrementally higher powers in the defocused beam of a 10 W CO₂ laser (New Wave Research MIR10) until fused. The gas evolved from each step was analyzed by a VG5400 mass spectrometer equipped with an ion-counting electron multiplier. All measurements were corrected for total system blank, mass spectrometer sensitivity, mass discrimination, radioactive decay during and subsequent to irradiation, as well as interfering Ar from atmospheric contamination and the irradiation of Ca, Cl and K. The isotope production ratios were: (⁴⁰Ar/³⁹Ar)_K = 0.0005 ± 0.00006, (³⁷Ar/³⁹Ar)_Ca = 1048 ± 0.9, (³⁶Ar/³⁹Ar)_Ca = 0.3952 ± 0.0004, Ca/K = 1.83 ± 0.01(³⁷Ar_Ca/³⁹Ar_K).

Details of the analyses, including plateau (spectrum) and inverse correlation plots, are presented inTable 2 andFigure 6. Initial data entry and calculations were carried out using the software ArArCalc (Koppers, 2002). The plateau and correlation ages were calculated using Isoplot v.3.09 (Ludwig, 2003). Errors are quoted at the 2-sigma (95% confidence) level and are propagated from all sources except mass spectrometer sensitivity and age of the flux monitor. The best statistically justified plateau and plateau age were picked based on the following criteria: (1) three or more contiguous steps comprising more than 60% of the³⁹Ar, (2) probability of fit of the weighted mean age greater than 5%, (3) slope of the error-weighted line through the plateau ages equals zero at 5% confidence, (4) ages of the two outermost steps on a plateau are not significantly different from the weighted-mean plateau age (at 1.8 σ six or more steps only), and (5) outermost two steps on either side of a plateau must not have nonzero slopes with the same sign (at 1.8 σ nine or more steps only).

Table 2
⁴⁰Ar/³⁹Ar step-heating data for the IXT02-6 alunite sample from steam-heated ground-type advanced argillic alteration associated with the low-sulphidation epithermal deposits at Ixtacamaxtitlán, Puebla.

J = 0.00015610 ± 0.00000023; Volume³⁹ArK = 1.924 x E-13 cm3 NPT Integrated Date = 2.86 ± 0.41 Ma Plateau Age = 2.86 ± 0.41 Ma (2s, including J-error of .2%), MSWD = 1.7, probability = 0.087, 100% of the³⁹Ar, steps 1 through 10 Inverse isochron (correlation age) results: Model 1 Solution (±95%-conf.) on 9 points Age = 2.79 ± 0.62 Ma Initial 40Ar/36Ar = 257.9 ± 9.7, MSWD = 0.64, Probability = 0.72

Figure 6
⁴⁰Ar/³⁹Ar age spectrum and inverse isochron for an alunite sample from a steam-heated ground-type advanced argillic alteration associated with the low-sulfidation epithermal deposits in the Ixtacamaxtitlán area, Puebla.

Most of the study area is strongly kaolinized and shows partly eroded patches of silica sinter on top (Figure 1; seeMorales-Ramírezet al., 2003;Tritllaet al., 2004). The sample examined in this study was collected from a surface exposure within ~10 m of one of the remaining patches of silica sinters in the Ixtacamaxtitlán area. Sinter outcrops occur at different heights across the mineralized area, but almost exclusively near the 2400 or 2500 m contour lines between showings that are ~1000 m distant (Figure 1). However, no faults, fractures, or other features that would have significantly disturbed their original position were detected among sinter outcrops. Therefore, it is interpreted that sinter outcrops and patches of alunite-rich layers occurin situ and on a contemporaneous paleosurface. The sample was taken from a ~1 m² beige to bright white sponge-like subhorizontal patch in which the spongy appearance (Figure 2A) is due to a few cm long vertical tubular structures (Figure 2B), which are constituted almost entirely of alunite. Additional minor minerals are kaolinite, opal, and, to a lesser extent, chalcopyrite. The walls of the remaining porosity in the tubular structures are essentially covered by a tapestry of euhedral (rhombohedral) alunite crystals (Figure 2C to2E); those, in turn, can be covered by opal, thus developing smooth surfaces (Figure 2F). Such rhombohedral alunite crystals normally show micron-sized cavities on their surface (Figure 2C to2E) or other evidence for inorganic dissolution, either favored by the cleavage of alunite or not (Figure 3). Kaolinite constitutes the floor of the tubular structures, and chalcopyrite crystals or aggregates up to a few tens of microns in diameter are common on the floor or the lower portion of the structures (Figure 3H).

A remarkable characteristic of these tubular structures is the occurrence of thread-like microstructures of likely organic origin that have been replaced by opal (Figure 2), alunite, or kaolinite (Figure 3). Most of the organogenic microstructures are compatible with the architecture of fungi (hyphae) whereas some have a dubious origin despite an organic resemblance. The latter are <1 µm thick individual filaments interwoven into a cobweb-like microstructure in contact with fungal structures (Figure 2F). Fungal structures are filaments (hyphae) with irregular borders, sometimes bifurcated, about 2.5 µm wide with contrasting superficial densities. Some transects look darker and have been flattened, as fossilized empty structures (Figures 2C to2E and3G). Such an observation is relevant because it means that part of the mycelium was already dead when fossilization processes occurred. However, each cellular compartment divided by septa allowed fungi to survive in stretches where some cells were dead while others were still alive. Hyphae in this study show compelling evidence for septa (Figure 2E) that divide otherwise cellular content, as well as interconnections (i.e., anastomosis;Figure 2E and2F). Hyphae grew in three-dimensional networks (Figures 2C to2F and3) where sometimes hyphae aggregated in cord-like structures (Figure 3C to3D). Some hyphae actually developed boring structures or dissolution grooves on alunite crystals (Figure 3B), which constitutes petrographic evidence that suggests that organic acids segregated by hyphae were capable of dissolving alunite. Besides the replacement of hyphae, mycelial cords, and other fungal structures by alunite, a later crystallization of euhedral (rhombohedral) alunite microcrystals was visibly produced on fungal structures (Figure 3B to3D). Both replacements and new euhedral alunite stand for, at least, two generations of alunite, thus pointing to a recurrence of the phenomena that gave way to its crystallization, separated by favorable periods for fungi to thrive.

The amorphous SiO₂ saturation index (SiO₂(am) SI)vs. temperature, amorphous SiO2(am) SIvs. pH, and temperaturevs. pH diagrams (Figure 5) show the modeled curves for each selected water sample that represent the equilibrium loci for kaolinite and alunite. In other words, these are the curves at which the SI of these minerals equals zero at variable pH, temperature, and amorphous SiO₂(am) SI. If we accept such water samples from the Los Azufres geothermal field as representative for the different types of water that would be dominant in the Ixtacamaxtitlán area while hydrothermally active, we may characterize the evolution of the dominant fluids in this location in terms of temperature, pH, and SiO₂(am) SI through time: (1) steam-heated waters; (2) a lull in the generation of acidic vapors and “invasion” of cooler and more alkaline water (i.e., partly meteoric or mature water) that favored fungal-bacterial consortia; and (3) back to steam-heated waters, and fossilization of fungi and biofilms. This allows us to tentatively position these three stages in the evolution of the advanced argillic alteration assemblages in the Ixtacamaxtitlán mineralized area. During stages 1 and 3, the steam-heated water would be supersaturated in amorphous silica (opal), and thus such water would be located somewhere above the amorphous silica equilibrium (i.e., SiO₂(am) SI > 0). However, during stage 2, opal was redissolved, either inorganically or due to the action of fungi, during the incursion of mature water that can be located somewhere below the amorphous SiO₂ equilibrium (i.e., SiO₂(am) SI < 0). Modeling results (Figure 5) indicate that the stability of the alunite and kaolinite assemblage relies on significant variations in temperature, pH, and saturation conditions for amorphous silica. This means that even slight variations in pH may easily lead to crossing the amorphous silica supersaturation-subsaturation boundary.

The⁴⁰Ar/³⁹Ar age determined for alunite in the tubular structures is 2.87 ± 0.41 Ma (integrated age;Figure 6A), which is very similar to the inverse isochron age of 2.79 ± 0.62 Ma (Figure 6B) and corresponds to the late Pliocene, or Piacenzian.

The⁴⁰Ar/³⁹Ar age yielded by alunite from tubular structures at 2.87 ± 0.41 Ma corresponds to an advanced argillic alteration assemblage overprint that postdates the deep hypogene epithermal alteration assemblages (dated at 4.3 ± 0.1 Ma;Poliquin, 2009). Nevertheless, this age is still relevant as an indicator of possible contemporaneous, deeper epithermal mineralization (discussed in section 5.2 below). Both ages bracket a minimum range of ~1.3 My for the formation of the low-sulfidation epithermal deposits at Ixtacamaxtitlán in their entirety. Such a range is similar to the general ~1 to 3 My bracket determined in several intermediate- to low-sulfidation epithermal deposits of different sizes and ages in Mexico (namely, the Pachuca-Real del Monte, Fresnillo, Guanajuato, Zacatecas, Taxco, Tayoltita, and Temascaltepec deposits;Langet al., 1988;McKeeet al., 1992;Camprubí et al., 2003;Camprubí and Albinson, 2007;Veladoret al., 2010;Farfán-Panamáet al., 2015;Martínez-Reyeset al., 2015;Enríquezet al., 2018;Zamora-Vegaet al., 2018). High-sulfidation epithermal deposits, however, were formed in shorter time spans (Valenciaet al., 2005,2008;Jansenet al., 2017). Therefore, the age difference between the alunite layer and the deep phyllic alteration in the low-sulfidation epithermal mineralization at Ixtacamaxtitlán is not long enough to allow claiming more than one epithermal deposit. On the contrary, such ages are compatible with the existence of different stages of mineralization within the same deposit.

The geological evidence (Carrasco-Núñezet al., 1997;Gómez-Tuenaet al., 2000,2003,2005,2007;Morales-Ramirezet al., 2003) and all the available ages of hydrothermal deposits in the Ixtacamaxtitlán area (Figure 7;Tritllaet al., 2004;Poliquin, 2009; and this study) show a tight linkage between the magmatic activity of the Trans-Mexican Volcanic Belt (TMVB) and both the porphyry-type and epithermal deposits. Such ages and those in other articles in this issue (Fuentes-Guzmánet al., 2020a,2020b) stand collectively for a metallogenesis of the TMVB, as already indicated byCamprubí (2009,2013) andPoliquin (2009). Specifically, the Cu-Mo-Au porphyry deposits would be associated with the middle to late Miocene arc (first stage of the TMVB, as ofGómez-Tuenaet al., 2005,2007) and the low-sulfidation epithermal deposits would be associated with bimodal volcanism that occurred at the end of the third stage in the evolution of the TMVB.

Figure 7
Sequence of magmatic and hydrothermal events in the Ixtacamaxtitlán area since the Miocene with the available ages. The range of ages for the volcanic rocks associated with the Cerro Grande stratovolcano was determined with the data provided byCarrasco-Núñezet al. (1997),Gómez-Tuena and Carrasco-Núñez (2000), andGómez-Tuenaet al. (2003).

There is a notorious ~12 My gap between the available ages for the Cu-Mo-Au porphyry and the low-sulfidation epithermal deposits (Figure 7). Earlier works described the spatial association between the Cu-Mo-Au porphyry/skarn and low-sulfidation epithermal deposits in Ixtacamaxtitlán as a case for telescoping (i.e., as a genetic association;Morales-Ramírezet al., 2003;Tritllaet al., 2004). Despite the clear overlapping in space of these deposits, such a time gap makes it difficult to describe this occurrence as true telescoping (sensuSillitoe, 2010), as this term implies aprogressive thermal decline of the (mineral)systems combined with synmineral paleosurface degradation (that)results in the characteristic overprinting [sic,Sillitoe, 2010]. Such a definition entails a continuum of mineral processes from the deep to the shallow environments, which is necessarily combined with some degree of exhumation that is synchronous to hydrothermal activity. Without the exhumation ingredient, porphyry-type and epithermal environments would be found stacked, not overlapped (therefore, not telescoped; see discussion inCamprubí and Albinson, 2007), although a continuum between both types would still exist. None of these is the case in the Ixtacamaxtitlán deposits. Besides, the nearby Cerro Grande stratovolcano was formed (Figure 7) between the porphyry-type and epithermal deposits (between ~11 and 9 Ma). It would have forcefully disrupted hydrothermal activity in the area and, if anything, set the course for a new mineral system on its own (not the case, though). Documented (or claimed) cases for telescoping between porphyry-type or skarn and epithermal deposits are not uncommon in the literature (e.g.,Simpsonet al., 2004;Catchpoleet al., 2011;Cookeet al., 2011;Camprubíet al., 2015;Dillet al., 2015;Franchiniet al., 2015;Imeret al., 2016;Penget al., 2017). The present geochronological study demonstrates that overlapping does not necessarily imply telescoping, not even in cases in which the types of deposits involved are likely to be part of the same specific mineral system (sensuHronsky and Groves, 2008;McCuaiget al, 2010;Hronskyet al., 2012). Therefore, high-resolution geochronological studies remain an essential tool to determine the actual linkage between different types of ore deposits. In the Ixtacamaxtitlán case, Cu-Mo-Au porphyry/skarn and low-sulfidation epithermal deposits are not related in time, and thus it is most unlikely that they are genetically related despite their spatial association. Such an association would then indicate a structural association between these deposits, as it was described, for instance, between volcanogenic massive sulfide and skarn types in the Francisco I. Madero deposit in Zacatecas (Camprubíet al., 2017b). No structural association has been explored for Ixtacamaxtitlán and would make a case for further research-that is, the persistence of structural features through time that would favor the emplacement of ore deposits in the same areas.

The mineralogy of shallow hydrothermal alteration in the Ixtacamaxtitlán epithermal deposit consists of kaolinite and alunite, with minor quartz, opal, and uncommon and less abundant smectite or illite-smectite. A mineral assemblage vastly dominated by kaolinite and alunite constitutes an advanced argillic alteration assemblage. Such assemblages may occur in any epithermal settings (1) as deep hypogene alteration, that is, in magmatic-hydrothermal environments; (2) as shallow hypogene alteration, that is, in steam-heated grounds; or (3) in supergene environments (e.g.,Sillitoe, 1993,2015). The vertical tubular structures found on anin-situ spongy layer are mostly constituted by aggregates of euhedral (rhombohedral) alunite crystals. These structures are very similar to fossil subaerial gas vents (as in Milos island, Greece;https://www.bgs.ac.uk/research/bufi/photosGallery.html, as of October 2019). These similarities, along with the nature of the fossil remains (discussed in section 5.3 below) within the alunite-rich tubular structures, and the high porosity shown in such structures suggest that their formation was produced in the subaerial part of a steam-heated environment. Both the spongy layer and the hot-spring silica sinters are subhorizontal and subconcordant with the underlying kaolinite + alunite ± opal ± smectite alteration envelope (seeMorales-Ramírezet al., 2003;Tritllaet al., 2004). In our view, the sinters and the alunite spongy layer would likely have formed simultaneously and thus share the same paleosurface. Therefore, they are interpreted as formed in a steam-heated environment where the spongy layer would be due to acidic gas venting on the paleosurface (i.e.,Sillitoe, 2015) peripherally to the silica sinters. Such an environment requires the occurrence of boiling at depth, which is corroborated by the occurrence of adularia in the veins (Poliquin, 2009). Then, boiled-off acidic vapors would have condensed in a shallow paleoaquifer, as demonstrated by the resulting broad subhorizontal kaolinite + alunite ± opal ± cristobalite ± smectite blanket below the hot-spring silica sinters (Figure 1;Morales-Ramírez et al., 2003;Tritlla et al., 2004) that resulted from the pervasive alteration of host rocks by acidic fluids. Evidence for boiling in the epithermal deposits at Ixtacamaxtitlán are the occurrence in mineralized structures of (1) pseudorhombohedral adularia, (2) bladed calcite phantoms (Poliquin, 2009), and (3) coexisting vapor- and liquid-rich fluid inclusions within the same fluid inclusion assemblages in vein minerals (Poliquin, 2009). Eventually, acidic vapors could travel relatively unscathed toward the paleosurface, thus generating the alunite-dominated tubular structures shown in this study (Figure 2A and2B). However, no chalcedony blanket that landmarks the paleo-phreatic level-a typical feature in this environment (Hedenquistet al., 2000;Sillitoe, 2015)-was found during our surveys in the area.

Such an environment may naturally include periods of quiescence, which are generally favorable for input of water with different temperatures and geochemical characteristics, such as meteoric water or upwelling water from deep sources. The latter may come from different possible sources and evolutionary paths, and thus different geochemical characteristics as well. The modern analog used in this paper for a steam-heated environment is the Los Azufres geothermal field in Michoacán. The reason for such a choice is thatGonzález-Partidaet al. (2005), among other characteristics that are shared with the epithermal deposits at Ixtacamaxtitlán, provided evidence for the concurrence of (1) boiled-off acidic water; (2) meteoric water; (3) fluids with different degrees of chemical equilibration with host rocks; (4) mixing phenomena between upwelling hydrothermal fluids and meteoric water; (5) prevailing advanced argillic alteration assemblages of the shallow hypogene type; (6) the occurrence of sinters; and (7) some degree of space zonation of the different types of waters, which range from steam-heated to mature and peripheral waters. Therefore, we use that temperature and geochemical data obtained by González-Partidaet al. (2005) in our geochemical modeling, as well as a conceptual analog to the subject of study.

In the present study, the fossils of fungal remains are exceptionally well preserved, allowing us to identify mycelia, hyphae with septa, and anastomoses between branches, among other characteristically fungal features (Figures 2 and3). Such remains are fossilized by opal, kaolinite, and alunite at Ixtacamaxtitlán, although fungi have been described to be much less prone to silicification than other organisms (Joneset al., 1999;Konhauseret al., 2004). Silicification of microbes in hot-spring environments would have occurred rapidly, within a few days after their demise and before their soft tissues started to collapse or organic matter started degrading at the temperatures that are typical for these environments (Joneset al., 2004). Although we are unaware of fungi fossilization by kaolinite or alunite in other locations, and silica minerals are the common preservers of fungal remains, fungi by a geyser in Lake Ngakoro in New Zealand were fossilized by both silica and jarosite (Jones et al., 2000) and show strong similarities with the remains described in this paper (compareFigures 2 and3 in this paper withFigures 7 and9 inJoneset al., 2000).

Relatively regular-shaped dissolution pits in alunite that were not formed by living organisms (Figure 2C to2E) probably denote cooling of fluids in this environment once it was starved from boiled-off acidic vapor, as alunite destabilizes at temperatures <200 °C and at pH > 3 (Hedenquist and Taran, 2013;Aceroet al., 2015). Once the temperature was low enough to allow fungi to thrive, their activity led to boring on their substrate, which was mostly a tapestry of alunite crystals. In fact, fungi are very effective at degrading silicates, which is an ability that would account for the notorious corrosion bays developed on kaolinite (Figure 3E and3F);e.g.,Aspergillus niger degrades kaolinite and several other silicates (Sterflinger, 2000). In subaerial hydrothermal systems or deposits, fungal evidence can be interpreted in terms of paleoecological changes or as cycles in hydrothermal activity (Figure 8). Provided that fungi are not autotrophic organisms, it can be deduced that the hydrothermal environment would have necessarily had to cool down to allow bacteria or thermophilic archaea to thrive as well, but also algae, which would supply nourishment for fungi. Then, upon the reactivation of hydrothermal activity, new boiling at depth and gas venting in steam-heated grounds would have killed fungi and fossilized them with opal, alunite, or kaolinite (as seen inFigures 2,3, and8). The reactivation of acidic-gas venting by subsequent generations of boiled-off vapors is suggested by both the replacement of fungal microstructures by alunite and a second (at least) generation of rhombohedral alunite crystals on fossilized hyphae (Figures 3D and8). The presence of heterotrophic fungi in extreme nutrient-poor habitats have been attributed to three possible explanations: (1) they are nourished by sediments rich in organic remains, (2) they are nourished by abiotic mineral-fluid reactions, or (3) they are nourished by symbiotic relationships with chemoautotrophic prokaryote biofilms that served as a carbon source for anaerobic fungi under anoxic conditions (Bengtsonet al., 2014;Ivarssonet al., 2015,2016). In this sense, the cobweb-like structure inFigure 2F might represent the fossilized skeletal remains of biofilms (such possibility is further discussed below). All the same, the corrosion bays in kaolinite aggregates (Figure 3E to3G) and the generally “bio-brecciated” appearance of the mineral-fungal ensemble (Figures 3A,3F,3H and8) are not to be ignored, as they are far more developed than boring on alunite. The intensive dissolution of kaolinite could be explained by the relatively poor nutrient potential of this mineral, which compels the microorganisms that live on kaolinite-rich substrates to be particularly aggressive in order to obtain sufficient metal nutrients (Cuadros, 2017). Fungi are also described to occur on open-air microstromatolitic kaolinite laminae in the Te Whakarewarewatangaoteopetauaawahiao geothermal system in New Zealand (Joneset al., 2001a).

Figure 8
Conceptual sketch of the low-sulfidation epithermal deposits at Ixtacamaxtitlán (right) and of the general evolution of the spongy alunite blanket on steam-heated grounds in the area (left).

In extreme environments, cavities, cracks, and crevices are suitable sites for fungi development (Ehrlich, 1998;Viles and Gorbushina, 2003;Gorbushina, 2007). Over 80 fungal species have been listed as acidophilic or acid-tolerant (Gross and Robbins, 2000). Although such inventory refers to extant species, a <3 Ma paleohydrothermal system like the one at Ixtacamaxtitlán may well have harbored similar species, albeit episodically. In fact, fungi (or reasonably suspected to be so) have been found in association with (paleo-)hydrothermal manifestations in the geological record of a broad variety of ages and geological settings (Joneset al., 1999,2001a,2001b,2004;Rasmussen, 2000;Sterflinger, 2000;Fayers and Trewin, 2004;Gadanho and Sampaio, 2005;López-Garcíaet al., 2006;Van Doveret al., 2007;Connellet al., 2009;Le Calvezet al., 2009;Chiacchiariniet al., 2010;Ivarssonet al., 2012,2019;Massiniet al., 2012,2016;Tayloret al., 2015;Dekovet al., 2016;Taksavasuet al., 2018). Such settings include both epithermal and extant geothermal environments.

Another interesting feature in the tubular structures found in Ixtacamaxtitlán is the occurrence of abundant chalcopyrite crystals and aggregates (Figure 3H). The precipitation of sulfides (most noticeably, pyrite) in steam-heated environments is common by inorganic means (Stoffregenet al., 2000;Sillitoe, 2015). However, the occurrence of boring on alunite crystals by hyphae and the subsequent reduction of the released sulfate due to the dissolution of alunite by organic acids could also contribute to the formation of sulfides. Rocks and minerals are indeed altered by fungi (e.g.,Gómez-Alarcónet al., 1994;Hirschet al., 1995;Sterflinger, 2000). Also, the microbial weathering of rhyolitic obsidian may produce quartz and alunite (Cuadroset al., 2012). Usually, fungi attack minerals by two main mechanisms: by acidification of the environment, and by mechanically disaggregation or aggregation of particles (Sterflinger, 2000). There are two mechanisms of fungal mineralization: (1) controlled, which is mediated by selective exudates depending upon the mineral; and (2) induced, which is mediated by indirect metabolic activity as excreted polymers. Such polymers serve as nucleation sites that promote mineral crystallization, or simply by wall surface charges as adsorption sites that also promote the nucleation of minerals even in dead cells. This mechanism has been demonstrated in jarosite biomineralization at pH values ~2 with acidophillic fungi whose hyphae were completely covered by precipitated jarosite (Oggerinet al., 2013). This case is relevant because jarosite [KFe³⁺₃(SO₄)₂(OH)₆] and alunite [KAl₃(SO₄)₂(OH)₆] are isostructural end-members in the alunite supergroup. Further, their stability conditions and geological occurrences can be similar, although jarosite is a rare hypogene mineral because it requires more extreme acidic and oxidizing environments to form (Joneset al., 2000, and references within;Stoffregenet al., 2000). Therefore, relatively slight increases in pH may result in the dissolution or precipitation of either jarosite or alunite in acidic and oxidizing environments, which makes the behavior of these minerals comparable. Fungi-driven acidification is originated by pumping H⁺ and excreting metabolites like carboxylic acid, among other acids (i.e., oxalic, citric, carbonic, phosphoric, aromatic and aliphatic, etc.;Mülleret al., 1995), CO₂, siderophores (Renshawet al. 2002), and extracellular polymeric substances, among other substances. For example, acidophilic fungi can precipitate jarosite by decreasing pH down to the range between 2.5 and 2 by controlled biomineralization in merely 10 days (Oggerinet al., 2013). Such values are similar to the pH at which alunite destabilizes (>3;Aceroet al., 2015). The sulfate thus released would then possibly be reduced by prokaryotic organisms (archaea or bacteria) that formed consortia with fungi, which can be found in a variety of geological environments (Konhauseret al., 1994;Chiacchiariniet al., 2010;Bengtsonet al., 2014;Ivarssonet al., 2015), even at great depths within the continental crust (Drakeet al., 2017). In fact,one of the most successful means for fungi to survive in the extreme sub-aerial environment is underpinned by their symbiotic associations with algae and cyanobacteria… [sic] (Rangelet al., 2018). For instance, strains ofLeptospirillum ferrooxidans,Acidithiobacillus ferrooxidans,Acidithiobacillus thiooxidans, andAcidianus spp., among others, along with different species of theAspergillus andPenicillium genus (at pH between 3 and 3.5) among filamentous fungi, yeast, and archaea consortia were determined for the Copahue-Caviahue geothermal system in Argentina (Chiacchiariniet al., 2010). The occurrence ofAspergillus andPenicillium was also argued in the Te Whakarewarewatangaoteopetauaawahiao geothermal system in New Zealand (Joneset al., 2001a), asin waters with pH < 5 … fungi become dominant because they are adept at surviving in acidic water [sic] (Joneset al., 2001a). Therefore, the occurrence of fungi in a hot-spring environment can be tentatively constrained at pH between 3 and 5.

Bacterial sulfate reduction could account for the precipitation of chalcopyrite, as inFigure 3H. Could the cobweb-like structure inFigure 2F be fossil evidence for bacterial consortia as biofilms? Similar structures at similar scales have been thus characterized both in the fossil record (e.g.,Figure 1 inSchopfet al., 2015; also,Campbellet al., 2015a,2015b;Fadelet al., 2017;Schopfet al., 2017) and as extant or recent consortia (Jannaschet al., 1994;Joneset al., 1999,2001a,2001b,2004; also, seeFigure 6 inMaranoet al., 2016). Despite being prone to obliteration due to opal dehydration or recrystallization, acid etching, and other phenomena, bacteria fossils can even be preserved in silica sinters (e.g.,Campbellet al., 2015a,2015b). Also,Metallogenium bacteria involved in the oxidation of Fe or Mn may grow on the hyphal network as a result of indirect fungal biomineralization mediated by fungal exudates (Emersonet al., 1989;Furutaet al., 2007). Associated bacteria may nourish fungal mycelium, thus resulting in active bioweathering that is able to mobilize Si, Fe, Mn, and Mg, and stimulate the neoformation of clay minerals, among other silicates. The accumulation of the associated bacteria on hyphae may well lead to their preservation (Peckmannet al., 2008;Ivarssonet al., 2018). Bacterial communities associated with arbuscular mycorrhizal fungi assist the latter to complete the functions required for this association to succeed (Turriniet al., 2018). Arbuscular mycorrhizal fungi would then recruit specific bacterial populations capable of dissolving P from relatively insoluble sources when it is associated with Al and Fe in either acid or alkaline substrates (Turriniet al., 2018). However, there is a large gap in our understanding of the possible role of fungi in surficial environments where only Ascomycota and Basidiomycota, including strains of anaerobic fungi, have been identified in vent fluids (López-Garcíaet al., 2007) and Chytridiomycota in hydrothermal vents (Le Calvezet al., 2009). All in all, the occurrence of archaea or bacteria in consortium with fungi is likely even in steam-heated environments.

Fungi and fungal remains and their interactions with minerals in epithermal deposits or geothermal systems, like those in the deposits at the El Deseado massif in Argentina (e.g., Coelomycetes, Microthyriales, Chytridiomycota, etc.;Massiniet al., 2012,2016), are barely subjects of thorough examination. As discussed earlier, there is room for research on sulfate release by fungal activity and subsequent sulfate reduction and precipitation as sulfides mediated by prokaryotes. Besides the obvious interest in characterizing such processes for microbiological disciplines, there is a reasonable possibility that significant concentrations of key metals occur in similar hydrothermal environments. The efficiency of such processes would then be dictated by the specific prokaryote-eukaryote consortia that could actually form in each case. Interestingly, as mentioned above, such consortia can be established in a broad range of depths, from the very paleosurface down to hundreds of meters deep. This means that the role of living organisms in mineral precipitation (and dissolution) can be more widespread than what is generally recognized in ore deposit studies. It is widely believed that extreme shifts toward very low δ³⁴S values in sulfides account for bacteria-mediated precipitation in a broad variety of types of mineral deposits or mineral systems (e.g.,Miranda-Gascaet al., 1998;Camprubíet al., 2001;Conlyet al., 2006;Tornoset al., 2008;Arninget al., 2009;Carrillo-Rosúaet al., 2014;Bonettiet al., 2015;Drakeet al., 2015,2017;Simpsonet al., 2017;Zhaoet al., 2018;Fazliet al., 2019;Holleyet al., 2019). In many cases, mineral precipitation was produced hundreds of meters below the paleosurface and in contrasting geological environments, but the role of bacteria in them is not disputed. However, fungal contributions are undetectable unless it is illustrated by compelling petrographic evidence or experimental work. Might hydrothermal systems that can bear bacteria and archaea also bear fungi? As discussed earlier, it is most likely that they do, and they have been described together as agents for mineralization in geothermal sinters (Joneset al. 1999). Then, could fungal activity be held accountable for part of reactive sequences that are so common in sulfide associations? Chemolithotrophic bacteria such asThiobacillus ferrooxidans andThiobacillus thiooxidans are very effective in solubilizing sulfides, hence their industrial use in bioleaching technologies, whereas fungi are more effective in bioleaching of non-sulfide ores (Bosecker, 1997;Weiet al., 2013). Experimental studies withAspergillus niger produced variable (if not contradictory) results in the dissolution of sulfides like sphalerite or galena, although the fungal dissolution of zinc and lead oxides, carbonates, or phosphates can be much easily and effectively achieved (Sayeret al., 1997,1999;Sutjaritvorakulet al., 2013;Weiet al., 2013). Also, some metal-tolerant strains of mycorrhyzal fungi were actually able to solubilize some galena (Fominaet al., 2005) despite its recalcitrance. Most studies are directed to bioremediation and thus to assessing the ability of fungi to leach, accumulate, and immobilize metals (e.g.,Gadd, 2010;Sabraet al., 2011;Weiet al., 2013;Cecchiet al., 2019). Such is the reason for focusing on the subsequent precipitation of other solids following the dissolution of sulfides, such as oxalates, instead of setting environments in which sulfur metals would be released into an aqueous phase. None of these is an easy task, asfungal interactions with metal sulfides have been much less studied and these are generally regarded as quite recalcitrant materials for fungi [sic] (Weiet al., 2013). Bioweathering of zinc sulfide minerals (sphalerite or wurtzite) by saprotrophic fungi (Aspergillus niger,Penicillium roqueforti,Beauveria caledonica,Serpula himantioides,Trichoderma versicolor andTrichoderma viride) is possible at room temperature and under specific conditions nonetheless (Weiet al., 2013). All in all, the definition of the role of fungi in the dissolution of sulfides is, to say the least, problematic. Studies in sulfate minerals are even scarcer than in sulfides, althoughAspergillus niger has been documented as an agent of gypsum dissolution (Gharieb, 2000). However, the available experimental studies on this matter are carried out at surficial temperature because these organisms are envisaged as agents of industrial bioleaching or bioremediation. What if those experimental studies were run at temperatures compatible with those of mineralizing hydrothermal systems as long as fungi, which are only moderately thermophilic, can endure them?

The activity of water further constrains the activity of microorganisms for it ranges between 0.611 and 0.755 for halophilic archaea or bacteria, while it ranges between 0.585 and 0.632 for fungi (Oren, 2013;Stevensonet al., 2015;Leeet al., 2018;Merinoet al., 2019). This means that water activity could be narrowed down to a range between 0.611 and 0.632 upon the coexistence of fungi and bacteria or archaea, as suggested earlier. The microstructural and mineralogical evidence may allow us to establish two feasible scenarios for the colonization of the alunite layer: (1) fungi would have established alone (water activity between 0.585 and 0.632), or (2) fungi would have been forming consortia with bacteria or archaea (water activity between 0.611 and 0.632). We envisage both possibilities because plausible evidence for such consortia are not abundant and may correspond to very localized conditions. As discussed above, the occurrence of fungi in this environment can be tentatively constrained at pH between 3 and 5 and at temperatures between 50 and 100 °C. Such pH and temperature brackets are represented inFigure 5 as blue and yellow shades, respectively. As presented in section 4.2, we may position water in the three stages of evolution of the advanced argillic alteration assemblages in the Ixtacamaxtitlán area with respect to the subsaturated or supersaturated character of each mineral. Amorphous silica (opal) occurs as an early mineral in this association (stage 1), also fossilizes fungal remains and biofilms, and occurs as a late coating (stage 3), but it does not occur while the fungal consortia were alive (stage 2), and thus we assume that the water during that period was subsaturated in amorphous silica. Alunite and kaolinite occurrences are about the same as for opal, but both were equilibrated in the modeled water samples and are being dissolved both inorganically and organically during the lifespan of fungi and bacteria or archaea. Therefore, the evolutionary trajectory of water in this environment would (1) start in a steam-heated environment in which water was supersaturated in amorphous silica, kaolinite, and alunite (inside the amorphous silica supersaturation box and above the amorphous SiO₂ equilibrium inFigure 5); (2) continue in a more mature environment, subsaturated in opal (outside the amorphous silica supersaturation box and below the amorphous SiO₂ equilibrium inFigure 5); and (3) end back in a steam-heated environment in which water was supersaturated in opal.

The evolution from stage 1 to stage 2 can be achieved through different paths, and all of them lead to the subsaturation in amorphous silica in water. Path A inFigure 5 involves increasing pH at nearly constant temperature, which can be due to the incursion of upwelling hydrothermal fluids that underwent no boiling and perhaps with various degrees of interaction with preexisting acidic water. Path B involves increasing temperature at nearly constant pH, which can be due to the influx of water similar to that of stage 1, only significantly hotter. Path C involves decreasing pH at nearly constant temperature, or dilution by acidic water, which is unlikely because such an extremely acidic environment would prevent the installation of any fungi or other organisms. Path D involves increasing pH while decreasing temperature, which would be compatible with an incursion of meteoric water during a momentary shutoff in the generation of boiled-off vapors. A combination of paths A and B (path AB) involves increasing both pH and temperature, which would be compatible with an incursion of upwelling mature and hotter water. While remaining geologically plausible, those paths that lead to a hotter environment (paths B and AB) would be hostile to the installation of fungi and other organisms unless the temperature remained below ~100 °C, but then such paths would not fulfill the requirement that the subsaturation in amorphous silica was achieved. Then, the most geobiologically plausible paths that led to stage 1 to stage 2 would be paths A and D (Figure 5). Finally, the hydrothermal system would evolve into a steam-heated system once more (stage 3), subsequently killing the life forms of stage 2, with water supersaturated in opal, alunite and kaolinite that fossilized the fungal and bacterial/archaeal remains. Then, it would result that the most likely pH during stage 2 and the installation of fungal consortia between ~3.2 and ~3.6, at temperatures between 53 and 75 °C (Figure 5).

It is necessary to clarify that 75 °C is taken only as a reference value that represents the mean between the temperature span that would generally allow fungi to live, but it is in no way a maximum value. The minimum pH and temperature conditions correspond to the lowest values calculated for steam-heated waters at subsaturated conditions of amorphous silica, whereas the maximum pH and temperature conditions are the minimum pH calculated for peripheral waters and the referred temperature value, respectively.

Therefore, pH remained relatively low during stage 2, but its moderate increase would have sufficed to induce the subsaturation in amorphous silica (Figure 5) and the subsequent dissolution of kaolinite and alunite, and would favor the colonization of this environment by living organisms. Then, although steam-heated acidic waters remained the dominant type in the paleosurface, their mixing -albeit limited- with meteoric or upwelling less acidic waters could account for the necessary pH and temperature variations.

*Corresponding author: (A. Camprubí)camprubi@comunidad.unam.mx