The 36–18 Ma Central Nevada ignimbrite field and calderas, Great Basin, USA: Multicyclic super-eruptions

During the middle Cenozoic, much of southwestern North America (Best et al., 2013a, their figure 1) was flooded by sheets of ash-flow tuff surrounding more than 200 source calderas. This ignimbrite flareup (Lipman et al., 1972; Noble, 1972; Coney, 1978) ranks as one of the greatest global manifestations in the terrestrial rock record of long-lived explosive silicic volcanism in a continental-margin volcanic arc. The ignimbrite flood in the United States aggregates to several tens of thousands of cubic kilometers while that in northern Mexico in the Sierra Madre Occidental likely aggregates to a few hundreds of thousands of cubic kilometers (McDowell and McIntosh, 2012). The subduction-related flareup is the only one known anywhere in the world of its magnitude and of Mesozoic or Cenozoic age that is not related to continental breakup (Bryan et al., 2010; Cather et al., 2009). The magnitude and brevity, chiefly from 36 to 18 Ma, of the middle Cenozoic flareup are all the more remarkable in view of the fact that subduction-related magmatism has persisted more or less continuously and diachronously in western North America since the Triassic.

A major expression of the ignimbrite flareup in the southwestern U.S. took place in the southern Great Basin (Figs. 1 and 2; Best et al., 2013a), the subject of this themed issue of Geosphere. We divide the southern Great Basin ignimbrite province into three individual ignimbrite fields; from west to east they are the Western Nevada (Henry and John, 2013), the Central Nevada, and the Indian Peak–Caliente (Best et al., 2013b). The latter two fields were on the Great Basin altiplano, a high plateau of limited relief underlain by unusually thick crust (50–70 km) resulting from pulses of late Paleozoic through Mesozoic orogenic contractile deformation (Fig. 3). This orogenic plateau has been referred to as the “Nevadaplano” by DeCelles (2004) and others; however, we prefer the appellation “Great Basin altiplano” that carries the connotation of a high plateau as well as a striking volcano-tectonic similarity to the late Neogene central Andean altiplano (Best et al., 2009).

Our investigation of the Central Nevada ignimbrite field and its associated source calderas—the subject of this article—builds upon the pioneering work of E. Bartlett Ekren and associates of the U.S. Geological Survey; we expand on their seminal volcanologic and stratigraphic framework published in the 1970s referenced throughout this article.

In the first comprehensive treatment of the southern Great Basin ignimbrite province (Best et al., 2013a), the Central Nevada field, together with the more voluminous coeval Indian Peak–Caliente field to the east (Best et al., 2013b), stand out in their unusual, even unique, expression of subduction-related explosive silicic volcanism. The magma systems in these two fields were exceptional with respect to the distinctive association of monotonous intermediate and succeeding trachydacitic compositions, their eruptive volumes (totaling 25,000 and 33,000 km³, respectively), and the number of super-eruptions (at least eight in each) that range upwards to 5900 km³.

The term super-eruption was first employed by Rampino and Self (1992; see also Miller and Wark, 2008) for the most voluminous eruptions (>1000 km³) recognized in the global geologic record; a few have occurred within the past 2 m.y., such as at Yellowstone (western U.S.) and Toba (Sumatra). The origin of these colossal events remains elusive. Miller and Wark (2008) proposed that thick continental crust inhibits the ascent of denser mafic magmas that otherwise would trigger eruption before gigantic quantities of silicic magma had time to accumulate in the shallow crust. The suppressed extrusion of intermediate-composition lavas in the middle Cenozoic southern Great Basin seems consistent with their scenario. But can thick crust alone have been responsible for the repeated, multicyclic super-eruptions that occurred at the same focus over 18 m.y. in the Central Nevada caldera complex, and for multiple, clustered, late Neogene super-eruptions of monotonous intermediates on the very thick crust (58–76 km) of the central Andean altiplano (Salisbury et al., 2011; McGlashan et al., 2008)? Or, as emphasized by de Silva (2008, p. 671), does “an elevated basaltic flux from the mantle ... [create a favorable] thermal and mechanical environment supportive of large-scale silicic magma production and storage”? Or, was there a conjunction of thick crust and high mantle flux in the Central Nevada and Indian Peak–Caliente fields during the middle Cenozoic? Did this conjunction play a role in the development of the distinct couplet of super-eruptive monotonous intermediate and trachydacitic magmas?

The purpose of this article is to lay the groundwork for answering these and related questions. We sampled cooling units at hundreds of outcrops, measured their thicknesses, and added hundreds of unit thicknesses measured on published geologic maps (Supplemental File ¹); made hundreds of modal and chemical analyses (Supplemental Files ² and ³); and studied and sampled 9 complete stratigraphic sections of as many as 14 ignimbrite cooling units (Supplemental File ⁴). Hundreds of thermoremanent magnetization directions have been determined and 121 precise single-crystal sanidine ⁴⁰Ar/³⁹Ar ages. Our new data permit the following: (1) rigorous correlation of 17 voluminous ignimbrite cooling units (>200 km³) generally exposed in two or more mountain ranges; (2) determination of unit volumes, showing they range to as much as 4800 km³ with at least eight super-eruptions of more than 1000 km³ and yielding an aggregate for the field of ∼25,000 km³; (3) refinement of source caldera margins; (4) establishment of a detailed eruptive chronology that reveals six eruptive episodes that are separated by five lulls, each lasting from 1.5 to 4.4 m.y., during which time little (<200 km³) or no ignimbrite was deposited; and (5) recognition of two distinct classes of voluminous silicic ignimbrite—monotonous intermediate and trachydacitic—not generally found in other middle Cenozoic fields in southwestern North America manifesting the ignimbrite flareup. This information, together with like data from other fields in the southwestern U.S. that experienced the middle Cenozoic ignimbrite flareup, provides significant constraints on interpretive models for the origin of the flareup, super-eruptions, and the couplet of monotonous intermediate and trachydacitic magmas. Many uncertainties in correlation of units, in perimeters of source calderas, and in related issues in the Central Nevada field are identified to provide direction for further studies. The data presented here will benefit future research in ignimbrite volcanology and its many facets, wherever pursued. On a broader scale, improved understanding of the middle Cenozoic ignimbrite flareups and associated super-eruptions have the potential to provide insights into (1) global climate change (e.g., Cather et al., 2009; Ruddiman, 2010), (2) arc volcanism, (3) the magnitude of transfers of heat and mass at subduction zones from the mantle into the crust, and (4) the origin of magmas in continental-margin volcanic arcs.

Ekren et al. (1974a, 1974b) first used the term Central Nevada caldera complex for overlapping sources of the Windous Butte, Monotony, and Lunar Cuesta ignimbrites, among others (Table 1 and Fig. 2).

Even though the mostly exposed source caldera of the 35.30 Ma Pancake Summit Tuff between Eureka and Austin is isolated well to the north of the main caldera cluster (Fig. 2), the caldera is included as part of the Central Nevada field because the tuff is interstratified with ignimbrite sheets originating from the main cluster to the south.

Approximately 60 km separates the Central Nevada caldera complex from the contemporaneous Indian Peak–Caliente caldera complex to the east (Fig. 2; Best et al., 2013b). No calderas occur in the intervening “outflow alley” where thick stacks of many interfingering outflow cooling units originated from each complex (Supplemental File 1 [see ^{footnote 1}]; Best et al., 2013a, their figures 6 and 7). For example, four cooling units from the Central Nevada complex and five from the Indian Peak–Caliente complex are exposed in the Golden Gate Range, four and ten in White River Narrows, and five and thirteen in the North Pahroc Range (Scott et al., 1992). (Geographic place names cited in this article are shown in Fig. 4; for clarity, place names are omitted from other maps in this article.) Only the most extensive outflow sheets from each complex reach a caldera of the other.

The mostly dacitic Seaman volcanic center (Fig. 4; du Bray, 1993) is the only focus of silicic volcanism between the Central Nevada and Indian Peak–Caliente caldera complexes. This small center is marked by a stratovolcano that was active at ca. 27 Ma, was 10 km in diameter, and had a restored volume of ∼20 km³. No ash-flow deposits extend very far from it.

Only 20 km separates the southern margin of the 27.57 Ma Monotony Valley caldera in the southernmost Central Nevada complex (Fig. 2) from the northern margin of the 13.7 Ma Grouse Canyon caldera, the northernmost caldera in the Southwestern Nevada volcanic field (Sawyer et al., 1994). The earliest activity in the Southwestern Nevada field began abruptly at 16 Ma (D.A. Sawyer, 2003, personal commun.). It is especially noteworthy that volcanic rocks in this field lack the arc chemical signature (e.g., low Rb/Nb ratios) that characterizes the 36–18 Ma, subduction-related ignimbrites and lavas in the Central Nevada field. Outflow sheets from each caldera complex overlap only in the most distal parts of their associated volcanic fields but no sheet reaches a caldera in the other field.

In the southernmost Pancake Range just east of the Twin Springs Ranch, as much as 30 m of the tuff of Bald Mountain crops out between the 24.69 Ma tuff of Buckwheat Rim and 26.36 Ma ignimbrites of the Shingle Pass Formation derived from the Central Nevada caldera complex. This local tuff of Bald Mountain originated from the small (7 × 10 km) Bald Mountain caldera (Fig. 2; Ekren et al., 1973b, 1977) ∼20 km north-northeast of Rachel. Little is known of this caldera and its associated ignimbrite.

The Central Nevada caldera complex merges imperceptibly with calderas to the west and northwest in the Toquima Range and additional ones farther west in the Western Nevada field (Figs. 1 and 2). These Western Nevada calderas and their associated ignimbrites are discussed in Henry and John (2013). Although calderas of the Central Nevada and Western Nevada fields lack clear separation, distributions of associated outflow ignimbrite sheets are somewhat more distinct. Of the 17 regional outflow sheets derived from sources in the Central Nevada caldera complex, only the Monotony Tuff is found with certainty west of the Toquima Range. Calderas in the complex are eccentrically positioned in the western parts of their related outflow sheets (Fig. 2). Because of erosion of the western slope of the Great Basin altiplano and resulting isostatic uplift, a topographically high lip is postulated to have formed on its western margin (Figs. 2 and 3; Best et al., 2009). This roughly north-south topographic barrier that may have existed during the ignimbrite flareup would have blocked most ash flows that originated east of it from advancing westward. Consequently, these ash flows, derived from the Central Nevada caldera complex, swept only to the north, east, and south across this relatively featureless altiplano where we find them preserved today. In contrast, ash flows from calderas in the Toquima Range just to the west of or nearly on the barrier, as well as ash flows from more westerly calderas, mostly flowed down the western slope of the altiplano, in part within stream channels that reached at least as far as what is now the western Sierra Nevada (Henry and Faulds, 2010). Only three of the many 29–25 Ma ash flows from sources in the Western Nevada field advanced eastward and these might have been swept around the north end of the barrier ( Appendix). The existence of the barrier and its effectiveness in governing dispersal of ash flows is problematic in several respects, as will be discussed more fully in the section entitled Topographic Barrier near the end of this article. It is sufficient at this juncture, however, to use the hypothesized barrier as a convenient boundary between the Western and Central Nevada fields in this article.

Without knowledge of the areal extent and volumes of ignimbrites, as well as their composition and age, as detailed below, it is impossible to fully understand the evolution of the explosive magma systems from which they were derived. Ignimbrite areas and volumes in the Central Nevada field are based on more than 500 determinations of the thickness (in meters) of individual ignimbrite units taken from published geologic maps and supplemented by our field measurements (Supplemental File 1 [see ^{footnote 1}]). Data were plotted on a base map for each unit and contoured by eye as isopach lines of uniform thickness at appropriate intervals, including a zero-thickness isopach, i.e., an outer limit beyond which the particular tuff is not seen where it would be expected to lie between older and younger deposits. Because of the progressive smoothing of the landscape during the ignimbrite flareup, contours for most units were fairly uniform, but some units that were the first to be deposited in a particular area reveal the existence of paleohills and paleovalleys.

The area of each ignimbrite unit (Table 2) was determined within the zero isopach using a CAD program as well as ArcGIS software (Environmental Systems Research Institute). The two methods yielded similar results.

Numerical calculation of volume by multiplying the deposit area by its thickness can be conceptually exact, but because of the lack of complete geologic constraints, either one or usually both of these dimensional parameters are uncertain to varying degrees, thus resulting in a volume estimate that is approximate at best (Table 2). Volume calculations were also made in two ways based on the isopachs, again yielding similar results. First is a grossly simplified method using a CAD program:

1. The area of each ring between successive isopachs is calculated and also the area of the source caldera constrained by exposed margin segments with geologically reasonable extrapolations beyond. This step is numerically precise but depends for its accuracy on our subjectively drawn isopachs and especially on the caldera margin.

2. For the outermost ring, i.e., between zero and the next higher nonzero isopach, and for the uppermost of all the inner rings, we account for the bevels by using the simplest possible model, that of a right circular cone, for which the volume is 1/3 times the area of the base times height. For example, between the 0 and 100 m isopachs, the volume (in km³) would be 1/3 times A in (km²) times 0.1 km where A is the measured area of this outermost ring.

3. For all the inner rings below the uppermost, the volume is simply the area times the limiting isopach value.

4. The volume of outflow tuff is given by a sum of the above.

5. The volume of intracaldera tuff is estimated by assuming a uniform thickness of some specified amount multiplied by the measured area of the caldera.

In the second volumetric calculation, we used ArcGIS to fit triangulated irregular networks (TINs) to isopachs.

The volumes were calculated using three idealized caldera models described and justified by Best et al. (2013b, p. 7–9 and their figure 4). Model 1 volumes are calculated using the thickness contours extrapolated smoothly across the central part of the tuff’s distribution. For this “pre-collapse” volume, we assume a uniform central thickness (the “maximum contour” in Table 2). It includes the outflow (“extra-caldera volume”) as well as some amount within the probable area of the caldera (“intracaldera pre-collapse volume”). Model 1 applies to three small cooling units for which no caldera apparently exists and to four for which the source caldera is partly or wholly concealed. Total volumes of ignimbrites in concealed calderas were assumed to be equivalent to the calculated volume of the pre–caldera collapse, or outflow, ignimbrite (Lipman, 1984, 1997; Best et al., 2013a; but cf. Salisbury et al., 2011; Folkes et al., 2011; Hildreth and Wilson, 2007; Self et al., 2010). For the Model 2 and Model 3 volumes, at least a minimum intracaldera ignimbrite thickness is known and enough of the perimeter of the source caldera is exposed to estimate the volume of intracaldera tuff. For Model 2, we assume the caldera has a flat floor; for Model 3, we assume the caldera fill is asymmetric (a trapdoor caldera). Volume estimates based on these three models are cited in this article and are listed in Table 2.

Calculated ignimbrite areas and volumes must be corrected for the amount of east-west crustal extension that took place after deposition (Best et al., 2013a). Four investigations provide insight into this amount of east-west strain. In an east-west transect from ∼113°30′ to 117°20′ W (easternmost Utah to central Nevada) between 39° and 40° N, Smith et al. (1991) determined an overall extension of 55% that resulted from mostly early Miocene (23 Ma) and younger faulting. Our measurement of their present-day cross-sectional length compared to the palinspastically restored section gives 43% overall extension. On the basis of the sedimentary and low-temperature thermochronologic record of upper-crustal extension in an east-west transect from 118°00′ to 115°30′ W between ∼40°00′ and 40°30′ N, Colgan and Henry (2009; see also Henry et al., 2011) concluded that little extension took place until ca. 17–10 Ma. They estimated ∼50–60 km of extension across the 200 km transect, or 30%–43%. McQuarrie and Wernicke (2005, their table 1) have tabulated the amount of extension, mostly after ca. 18 Ma, within individual strain domains in an east-to-west transect from 114°7′ to 117°23′ W between 40°20′ and 38°40′ N. Over the present-day distance of 280 km the extension is 74 km, or 36%. All three of these transects match the east-west extent of the Central Nevada field but lie mostly in the northern part or beyond. To justify extrapolation of the strains from these more northerly transects southward into the Central Nevada field we apply a paleomagnetic study by Hillhouse and Gromme (2011). They demonstrated that the pole of rotation of the Sierra Nevada plutonic block during the Cenozoic was located near the north geographic pole. Hence, the extension of the Great Basin was quasi-rectilinear so that the amount of strain as measured in kilometers is uniform north to south within most of the southern Great Basin ignimbrite province (Fig. 1).

From these four investigations, we feel justified in correcting all east-west dimensions and derivative areas and volumes cited in this article for a uniform 40% post-volcanic, east-west crustal extension. For some ignimbrites, such as the tuff of Palisade Mesa and members of the Shingle Pass (Table 1), present distributions are elongate north-south; applying the 40% extension correction makes them more so.

It must be emphasized that the amount of strain in adjacent structural domains can vary significantly. Smith et al. (1991) noted in their transect that extension ranged from ∼110% in eastern Nevada to ∼40% in central Nevada, and nil in between. Colgan and Henry (2009, p. 939) found that in their transect, extension was strongly partitioned into highly extended domains (50%–100% strain or more) separated by essentially undeformed crustal blocks. Hence, the assumption of uniform extension throughout the entire Central Nevada field cannot be justified in detail, but is a convenience we have adopted in the absence of explicit quantitative information on individual strain domains in the field.

Because of the uncertainties in the extension correction, as well as in perimeters of outflow deposits and calderas and thicknesses within them, ignimbrite volumes cited in this article are, at best, rough working estimates, subject to modification as more data become available. However, as order-of-magnitude estimates—that is, thousands versus hundreds of cubic kilometers—we feel they are meaningful.

An approximate uncertainty for each volume estimate is listed in Table 2. The approximate uncertainties in the volumes are mostly 50%. This uncertainty includes that involved in conversion to a dense rock equivalent volume. Several density measurements on the most densely welded, glassy phenocryst-rich dacites (monotonous intermediates) yield values of 2.5 kg m^–3, whereas less compacted and devitrified samples range to 1.7 kg m^–3. Densities of phenocryst-poor ignimbrites in the southeastern Great Basin are slightly less, according to Cook (1965) and Williams (1967). Their density profiles show that in thick outflow sheets, the density averages to roughly 2.2 kg m^–3. The average density of very thick piles of densely welded intracaldera ignimbrites is obviously higher, to 2.4 kg m^–3 in our measurements. Thus, applying a dense rock correction to the Central Nevada ignimbrites reduces their volumes by roughly 10%.

A series of modal and chemical variation diagrams provide an overview of the composition of 36–18 Ma ignimbrites in the Central Nevada field. The utility of these diagrams lies in their definition of a “main trend” that is occupied by dacites and most rhyolites whose magmas seem to share a common mode of origin. “Off-trend” ignimbrites, which are readily apparent outliers from this main trend, originated in some way different from the rest. For letter symbols used in composition diagrams in this article see Table 1.

Modal proportions of phenocrysts are shown in Figure 5 (see Supplemental File 2 [see ^{footnote 2}] for complete data). Figure 5A clearly shows the off-trend Coyote, Tikaboo, and Egan Tuff Members of the rhyolitic Shingle Pass Formation (letter symbols X, Y, and Z, respectively) that are separate but parallel to the main-trend ignimbrites. For a given plagioclase proportion, all of these Shingle Pass tuffs have relatively high proportions of sanidine and low proportions of quartz (Figs. 5B and 5C). The Coyote Summit Tuff Member has little or no biotite, unlike most other Central Nevada tuffs (Fig. 5C). The phenocryst-poor rhyolite tuff of Clipper Gap (C), another off-trend ignimbrite, has the lowest proportion of plagioclase and among the highest proportions of sanidine in tuffs in the field. The off-trend, phenocryst-poor, trachydacitic Isom-type tuffs (I) have very little or no quartz and sanidine and are unique in having no biotite. Their relatively high-temperature, dry phenocryst assemblage is dominantly plagioclase with lesser pyroxenes and Fe-Ti oxides. The Isom-type tuffs are so named because of their similarity to tuffs of the Isom Formation, a prominent eruptive unit to the east in the coeval Indian Peak–Caliente field (Best et al., 2013b). All of the other rhyolite and dacite tuffs in the Central Nevada field form a more or less continuous main-trend array on these modal diagrams. Among the main-trend ignimbrites, the Monotony Tuff (M, Fig. 5C) stands out because it contains the greatest proportion of biotite.

Complete chemical analyses are in Supplemental File 3 (see ^{footnote 3}) and representative analyses are in Table 3. The International Union of Geological Sciences (IUGS) classification diagram of Le Maitre (1989) that is used throughout this article and shown in Figure 6A indicates that most ignimbrites in the Central Nevada field are rhyolite, fewer are dacite, and still fewer trachydacite. Most of the Isom-type tuffs (I) are trachydacite. Most dacite analyses are of the Monotony Tuff (M) but include some samples of the Windous Butte Formation (W) that is zoned through most of the rhyolite field. The relatively lower total alkalies in ignimbrites of the Stone Cabin Formation (S) correlates with a greater proportion of quartz phenocrysts compared to all other rhyolites in the Central Nevada field (Fig. 5B). Except for the trachydacite Isom-type tuffs (I), which are alkali-calcic, Central Nevada tuffs are calc-alkalic and calcic (Fig. 6B). Only the Isom-type tuffs are consistently ferroan, or tholeiitic, whereas some samples of other units are ferroan but most are magnesian, or calc-alkaline (Fig. 6C). Most ignimbrites are high-K but some are shoshonitic, as are most of the Isom-type (Fig. 6D). The absence of low- and medium-K compositions reflects the unusually thick crust (∼60 km) in which the magmas originated and evolved (Fig. 3; Best et al., 2009).

Chemical variation diagrams (Fig. 7) show that, relative to main-trend ignimbrites, Isom-type tuffs (I) are enriched in TiO₂, Zr, and Ba but are poor in CaO and Sr. Stone Cabin tuffs (S) are enriched in CaO and Sr. The tuff of Lunar Cuesta (L) is enriched in Sr and Ba. Some samples of the Shingle Pass Formation (H, X, Y) are also enriched in Ba. Pahranagat Formation (A) and Stone Cabin Formation (S) ignimbrites, and especially the tuff of Clipper Gap (C), have high concentrations of Nb.

Table 4 and Figure 7F show the Sr isotopic compositions of volcanic rocks in the Central Nevada field. Initial Sr isotopic ratios range from 0.7062 for the ca. 24 Ma trachydacitic Isom-type tuff of Black Beauty Mesa to 0.7123 for the highly evolved upper tuff member of the Stone Cabin Formation—the oldest rhyolite tuff (35.77 Ma; Table 1) for which we have isotopic data. The range of initial Sr isotope ratios is very similar to that of the Indian Peak–Caliente field (Best et al., 2013b, their figure 6D).

The high initial ⁸⁷Sr/⁸⁶Sr ratios show that most of the magmas do not have sources solely in the mantle; significant proportions of old continental crust have been incorporated into most of the magmas. Basement rocks are poorly exposed in the vicinity of the Central Nevada field, but it is presumed to be underlain by crust of the Proterozoic Mojave province (Whitmeyer and Karlstrom, 2007). Wright and Wooden (1991) estimated the Sr isotopic composition of the Proterozoic basement in the Mojave province to range from ∼0.710 to 0.735. Thus, the volcanic rocks from the Central Nevada field with initial ratios higher than 0.710 could include substantial crustal contaminants ranging from 50% to nearly 100% depending upon the composition of the parental mantle-derived magmas and of the contaminating crust, which could have been extremely variable.

As in the Indian Peak–Caliente field, there is little correlation between the initial Sr isotopic ratios of the volcanic rocks and their silica concentrations (Table 4; Fig. 7F). However, ratios are highest for the oldest rocks. The only andesitic lava (54 wt% SiO₂) we analyzed underlies the Stone Cabin Formation (see next section) on the east side of the Pancake Range; it has a moderately high initial ratio of 0.7089. More felsic tuffs, including the rhyolitic 23 Ma Pahranagat and the 27 Ma tuff of Orange Lichen Creek, as well as the trachydacitic tuff of Black Beauty Mesa noted above, all have Sr isotopic ratios lower than this andesitic lava. The ca. 36 Ma rhyolite tuffs of the Stone Cabin Formation and a related rhyolite lava have the highest initial ⁸⁷Sr/⁸⁶Sr ratios (0.7109–0.7123); their magmas must have incorporated large amounts of ancient continental crust.

The oldest of the regionally extensive ash-flow tuffs in the Central Nevada ignimbrite field is a sequence of phenocryst-rich rhyolite cooling units named the Stone Cabin Formation by Cook (1965). These 36.10–35.77 Ma tuffs are restricted principally to the central Pancake Range and Grant Range (Fig. 8) in the central part of the field east of the main cluster of calderas. In the Pancake Range, these ignimbrites are directly underlain by a very local basaltic andesite and by more extensive rhyolite lavas (e.g., Radke, 1992). Because the rhyolite lavas have a very preliminary age determination of 36 Ma and are similar in composition to the tuffs, we include them as informal members of the formation. Thick rhyolite, dacite, and andesite lavas overlie the Stone Cabin Formation.

Hose and Blake (1976) noted the presence of Stone Cabin ignimbrite in eastern Nevada in the Schell Creek Range and in the central Egan Range at our site SP (Fig. 8A) but not in its northern part. (In this article, sample numbers and paleomagnetic sites cited in the text and shown in figures and in Supplemental Files 1 and 2 [see ^{footnotes 1} and ²] are in capital letters, such as SP and WAB, where stratigraphic identity is supported by firm data, or in lowercase letters, such as waj and bg, where data are insufficient to confirm stratigraphic identity and correlation is only permissive.) In the data of Gans et al. (1989) for east-central Nevada, we find no possible correlatives of the Stone Cabin, except in the Hunter district in the northern Egan Range at site sx where three ash-flow tuffs totaling ∼150 m thick are interstratified with andesitic lavas. The composition and K-Ar age of 35.8 Ma of these ignimbrites are consistent with the middle tuff member of the Stone Cabin Formation. However, Gans et al. (1989, p. 14) had “...little doubt that the tuffs represent local extrusive equivalents...” of dikes and a pluton in the same area, a conclusion supported by their remote distance from other known Stone Cabin ignimbrites (Fig. 8A).

The ignimbrite stratigraphy of the Stone Cabin Formation is one of the more complex in the Central Nevada field because it consists of sequences of multiple cooling units that are not entirely the same in the areas of exposure in the Grant and Pancake Ranges. Justification for including these separate units as members of one formation rests on their close spatial and temporal association and systematic compositional variation, implying derivation from a common source. High-precision ⁴⁰Ar/³⁹Ar ages and systematic paleomagnetic variations discussed below confirm the proposed correlations.

Cook (1965) recognized upper and lower constituent ignimbrites at the type section (site SQ, Fig. 8A) in the Grant Range ∼5 km southeast of Currant and just south of an abandoned cabin constructed of blocks of the Stone Cabin ignimbrite. The lower ignimbrite was subsequently referred to as the Calloway [sic] Well Formation by Scott (1965) and Moores et al. (1968). In their discussion of the Trap Spring oil field in Railroad Valley between the Grant and Pancake Ranges, French and Freeman (1979) did not correlate the Calloway Well Formation westward across the valley into the Pancake Range where Quinlivan et al. (1974) mapped three ignimbrite cooling units in the Stone Cabin Formation. In the Pancake Range, the lower ignimbrite is a simple cooling unit as thick as 75 m whereas the middle and upper units are compound cooling units as thick as 460 and 400 m, respectively. A fourth unit of thinly bedded tuff and epiclastic material that is as thick as 75 m lies between the lower two ignimbrite units; this bedded unit may correlate with some of the upper Blind Spring Formation of Moores et al. (1968) in the Grant Range. Based on our paleomagnetic and chronologic analyses together with field work and compositional data from Radke (1992), we have adopted the Quinlivan et al. (1974) subdivision of the Stone Cabin Formation into three members that we here informally designate the lower, middle, and upper tuff members of the Stone Cabin Formation. In so doing we abandon use of the Calloway Well Formation.

The lower tuff member, a small ignimbrite deposited at 36.10 ± 0.10 Ma, appears to be restricted to the Pancake Range (Quinlivan et al., 1974), at about the latitude of Currant (Fig. 8A), where it consists of a simple cooling unit of loosely welded white tuff that contains as much as 25% pumice lapilli and abundant clasts of rhyolite lava as much as 30 cm in diameter. Relative to the other two units in the formation, the lower tuff generally has less total phenocrysts, less quartz and sanidine, and the most plagioclase and biotite (Fig. 9).

The middle tuff member is a thick compound cooling unit or sequence of multiple cooling units in the Pancake and Grant Ranges, respectively. The member is paleomagnetically distinct and has normal polarity with moderate northwest inclinations (Fig. 10). Relative to the lower tuff, the middle has more phenocrysts, to as much as 5 mm in diameter, and the proportions of quartz and sanidine are greater and plagioclase less (Fig. 9). The weighted mean age of seven analyses of the unit is 35.83 ± 0.08 Ma.

A spectacularly exposed section (site SO, Fig. 8A) of the middle unit forms a southeasterly facing steep slope on the eastern flank of the Pancake Range. This so-called Meteorite Crater section (French and Freeman, 1979), visible from the community of Currant across Railroad Valley, is a compound cooling unit more than 280 m thick exposed above alluvium (Fig. 11). It consists of numerous layers defined by contrasts in color (mostly shades of brown and gray), in proportions and size of phenocrysts and pumice clasts, and in degree of welding (but mostly dense) and devitrification. The uppermost ∼50 m of the section is composed of a distinctive “salmon tuff” (the color of baked salmon, sample SNj; Radke, 1992) that overlies a meter or so of fine bedded tuff. The salmon tuff is a simple cooling unit containing conspicuous dark-gray to black quartz phenocrysts and abundant red-brown to gray volcanic lithic clasts to as much as 12 cm in diameter. The paleomagnetic direction of SNj overlaps that of sample SSb (Fig. 10), a dark-gray vitrophyre collected at the base of the middle tuff member to the north in the eastern Pancake Range area mapped by Quinlivan et al. (1974). This overlap implies that the Meteorite Crater section is mostly older than the member exposed in that area (Fig. 12). Still farther north in the Pancake Range, the salmon tuff is still absent and the section of the middle tuff member is ∼120 m thick and consists of a simple cooling unit of ash-flow tuff overlying several meters of porous bedded tuff. Sample SRa in the lower part of this section is somewhat silicified and has perturbed Na₂O and K₂O concentrations (1.76 wt% and 2.38 wt%, respectively) perhaps resulting from zeolitization in a subaqueous environment. This section is overlain by several meters of densely welded black vitrophyre that lies near the base of the upper tuff member (see below), which has been dated and sampled for composition and paleomagnetism at site SGm.

On the east flank of the Egan Range at sites SP and nearby SA (Fig. 8A), the greenish-yellow lower part of the middle tuff member also has perturbed alkalies, possibly as a result of zeolitization; Na₂O in samples SPh and SPj is <0.56 wt% and K₂O is 5.1 wt% and 5.6 wt%, respectively.

At the type section of the Stone Cabin Formation in the Grant Range (site SQ, Fig. 8A), the lower ignimbrite of Cook (1965), or the Calloway [sic] Well Formation of Scott (1965) and Moores et al. (1968), is compositionally like the middle tuff member in the Pancake Range to which it is correlated (Figs. 9 and 13). In the Grant Range, the middle tuff member directly overlies the mostly epiclastic Blind Spring Formation and is as much as 180 m thick. The member consists of nine loosely welded ignimbrites that range from 3 to 60 m thick separated by no more than a meter or so of bedded tuff. Its greenish-gray color reflects analcime and clinoptilolite replacements of vitroclasts. In contrast, a thinner (70 m) section of the unit 11 km to the south is a single moderately to densely welded deposit (Moores et al., 1968, their plate 1). Scott (1965) believed the thicker sequence was created as multiple ash flows were quenched in a water-filled closed basin and covered by tuffaceous sediment before the next ash flow was deposited; in the subaerial setting to the south the rapidly deposited multiple ash flows welded to one another.

Sample SQal collected ∼40 m above the base of sequence at the type section has a significantly steeper inclination but a similar declination as site SNj at the top of the middle member in the Meteorite Crater section (Fig. 10). The lower part (sample SQaug) of the uppermost outcrop in the type-section sequence has a greenish hue, like the tuff below it, and the composition of the middle tuff member, but a still steeper paleomagnetic inclination than site SQal. We interpret these data to mean that most of the type-section sequence is younger than the Meteorite Crater section (Fig. 12). The upper pink part (sample SEaup) of the uppermost outcrop of the type-section sequence is separated by a covered interval from the lower greenish part of the outcrop. This pink part has a reverse polarity, unlike all of the other units in the Stone Cabin Formation (Fig. 10). Because its age of 35.76 ± 0.07 Ma and modal and chemical composition are distinctly those of the upper tuff member, we consider it to be an early precursor of that compound cooling unit. The uppermost greenish to pink outcrop of the type-section sequence is separated by several meters of colluvium from the base of the overlying pervasively oxidized upper tuff member (sample SEb).

Most of the upper tuff member, a compound cooling unit, is normally polarized with a relatively shallow northeast declination (Fig. 10). In some places, this unit has a thick (as much as 10 m) black vitrophyre near the base overlying loosely welded tuff. The upper tuff member generally has more quartz and sanidine and distinctly less plagioclase than the rest of the formation (Fig. 9). Four analyses yield a weighted mean age of 35.77 ± 0.02 Ma. A rhyolite lava (sample BLUE-6C) overlying the Stone Cabin Formation has an age of 35.30 ± 0.02 Ma.

Compared to all other ignimbrites in the Central Nevada field, the Stone Cabin is distinctly enriched in CaO and Nb (Fig. 7B and 7E; also elements Y and V in Table 3). On a total alkalies–silica diagram (Figs. 6A and 13A) many samples have lower total alkalies (chiefly Na₂O) than most other tuffs in the field. Together with the relative enrichment in quartz (Fig. 5B), the high CaO and low alkalies are interpreted to be the result of high-pressure crystallization (Tuttle and Bowen, 1958). Moreover, Radke (1992) showed that rare hornblende is Al rich and consistent with high-pressure crystallization.

The range in chemical composition of the Stone Cabin samples, ∼72–78 wt% silica and sixfold variations in TiO₂, V, Sr, and Ba (Figs. 7 and 13), is also reflected in the wide range of modes (Fig. 9). Considerable scatter in the more mobile alkalies probably is a result of slight alteration, especially zeolitization, in the subaqueous environments in which some of the tuffs were emplaced. Scatter in the modal diagrams may reflect, in part, the large size of quartz phenocrysts that may or may not be intersected and counted in a thin section. These widely varying compositions not only distinguish the three members of the Stone Cabin Formation but also emphasize a systematic trend through time toward more evolved compositions. The three members show continuous trends from the relatively silica-poor, plagioclase-TiO₂-Zr–rich lower tuff member and precursory rhyolite lava to silica- and sanidine-rich, TiO₂-Zr–poor upper tuff member. Plagioclase is progressively poorer in anorthite and sanidine has less BaO from the lower to the upper tuff member.

However, chemical discontinuities within this trend show that the magma system was nonetheless open and not simply differentiating by fractional crystallization. This is borne out by the Sr isotopic data for the Stone Cabin Formation; the initial ⁸⁷Sr/⁸⁶Sr ratio increases from 0.7109 in the lower tuff member to 0.7123 in the more evolved upper tuff member (Table 4). Rhyolite lava underlying the lower member has an intermediate value of 0.7114. These ratios are the highest in the Central Nevada field and indicate that a large component of older continental crust was incorporated in this early magma system.

On the other hand, stratigraphic sections of individual members display no systematic compositional zonation. On any composition variation diagram (e.g., Fig. 11B), the middle tuff member at the Meteorite Crater section discloses significant fluctuations from more to less evolved tuff and back again but no overall trend. Interestingly, the uppermost salmon tuff (samples SNj and SOjuc) is the most evolved whereas a not-too-distant underlying sample SOil is the least evolved; the two samples essentially encompass the entire range of variations through the entire section. Similarly, in the compound cooling unit of the upper tuff member on the east flank of the Grant Range at site SF (same as site SD; Fig. 8B) there is no apparent systematic zonation in modal proportions of phenocrysts (Fig. 14) nor in chemical composition (Scott, 1966). Nonetheless, Scott (1965, his figure 10) documented five compositionally and texturally distinct feldspar groups in the compound upper unit; the anorthite content of the plagioclase alternates from more to less evolved compositions. This lack of any systematic compositional variation in the individual eruptive units but regular compositional variation from the lower to the upper Tuff Member implies extraction of parcels of magma from different parts of a zoned chamber that was evolving through 0.33 m.y.

A negative gravity anomaly of 15 milligals northeast of Currant has been interpreted to be a caldera by Jachens et al. (1996, their figures 2–4) and Ludington et al. (1996, number 29 in their figure 5-1 and table 5-1). This anomaly is marked by an oval ring (Fig. 8B) of felsic and subordinate intermediate-composition lava flows that are undated, but likely about the age of the Stone Cabin (Best et al., 1989, their figure 3). No detailed geologic map exists for this area but the lava ring appears on the 1:250,000-scale White Pine County map (Hose and Blake, 1976) and on the 1:1,000,000-scale map of Stewart and Carlson (1976b). The gravity anomaly might reflect thick intracaldera Stone Cabin ignimbrite buried beneath alluvium inside the ring of lavas, which might represent extrusions from the ring fault bounding the subsided caldera block. However, the location of the oval ring beyond the thickest sections of the Stone Cabin is not consistent with a source for the tuff. Moreover, we have found no paleomagnetic and lithologic evidence for Stone Cabin ignimbrite in the oval ring area.

Because the thickest sections of the Stone Cabin ignimbrite occur in the Grant Range and to the west in the Pancake Range at roughly the latitude of Currant (Fig. 8; see also French and Freeman, 1979), a caldera source in the intervening Railroad Valley would appear to be likely. The occurrence of thick, temporally and compositionally similar rhyolite lavas in these areas further supports this conjecture. But no unequivocal evidence for a buried source caldera in Railroad Valley can be found in subsurface data for oil fields ∼20 km south and southwest of Currant (see articles in Newman and Goode, 1979). However, geologic evidence to the north hints at a concealed source caldera in northern Railroad Valley. On the eastern flank of the Pancake Range, the generally dense welding and vitrophyric zones in the numerous individual flows within the thick compound cooling unit of the middle tuff member in the Meteorite Crater section (Fig. 11) is consistent with a proximal source, but these characteristics might alternatively reflect ponded accumulation in a topographic low. The member exposed to the northwest has more distal attributes. Outcrops of Stone Cabin ignimbrite just south of Duckwater are highly silicified and brecciated. These two occurrences suggested to G.L. Dixon (2013, personal commun.; see also French and Freeman, 1979; Kleinhampl and Ziony, 1985) the possibility of a nearby caldera to the east. A small granitic intrusion across Railroad Valley north of Currant on the west flank of the White Pine Range has been dated at 27, 31, and 36 Ma (uncertainties as much as 5 m.y.) by the K-Ar method on biotite (Kleinhampl and Ziony, 1985, p. 98–100). Based on these relations in flanking ranges, we postulate a buried source caldera for the Stone Cabin in northern Railroad Valley between Duckwater and Currant (Fig. 8B).

Irregularities in the land surface on which the Stone Cabin ignimbrites were deposited (Fig. 8; see also French and Freeman, 1979) compromise estimation of their volume. Some local pre-volcanic basins were filled with more than one kilometer of Late Cretaceous–Eocene sediment of the Sheep Pass Formation (Druschke et al., 2009) and thick rhyolite and andesite lavas were extruded prior to, during, and after deposition of the ignimbrites (Radke, 1992; Scott, 1965). Because the small volume (10? km³) of the lower tuff member had little smoothing effect on this irregular topography, we can only roughly estimate the volume of the middle and upper tuff members at ∼500 km³ each using Model 1. Adding equivalent accumulations in an associated source caldera yield volumes of ∼1000 km³ each (Table 2).

One-half million years after deposition of the youngest Stone Cabin ignimbrite another phenocryst-rich, mostly high-silica rhyolite magma erupted at 35.30 ± 0.03 Ma several tens of kilometers to the northwest. This event created a large-volume outflow ignimbrite and the Allison Creek caldera into which at least one kilometer of intracaldera tuff accumulated (Fig. 15). The outflow ignimbrite was named the tuff of Pancake Summit by Gromme et al. (1972) for exposures at Pancake Summit on U.S. Highway 50, which crosses the northern Pancake Range. Later, Nolan et al. (1974) formalized the name to Pancake Summit Tuff. Throughout the distribution of the outflow sheet, the tuff occurs as a simple cooling unit in which exposures range from a densely welded, locally vitrophyric base upwards to a less welded top.

Samples contain 74.8–76.5 wt% silica (Fig. 13A) and 14%–50% phenocrysts (Fig. 16), mostly quartz and feldspar (to as much as 5 mm); biotite is a minor and smaller phenocrystic phase and Fe-Ti oxides are still less and smaller. A few grains of hornblende, orthopyroxene, and allanite are seen in some samples. The Pancake Summit Tuff typically has higher alkali contents than the older Stone Cabin (Fig. 13A), making it more like other rhyolite ignimbrites in the Central Nevada field.

At two sites, PC and PN south of the source caldera (Fig. 15), phenocryst-rich ignimbrites have the Pancake Summit paleomagnetic direction, but their silica content is distinctly less at 68.8–70.3 wt% and the modes are also distinctly different in that plagioclase is twice as abundant as quartz, only ∼10% of the phenocrysts are sanidine, and there is substantially more (4%) hornblende. The modes of these two samples overlap those of the Windous Butte ignimbrite and overlap appears in all analyzed chemical elements analyzed as well. Further sampling and chemical analyses, particularly in the intracaldera deposit and in the upper parts of the outflow sheet, would be necessary to ascertain compositional variations in the Pancake Summit Tuff.

The source caldera of the Pancake Summit Tuff occupies the northern Antelope Range (Fig. 15). The caldera is shown in Stewart and Carlson (1976a, 1976b) and in Sargent and Roggensack (1984), who cited as a reference Anderson et al. (1967). The name is taken from Allison Creek that drains the range to the east and provides access to the interior of the caldera. The caldera’s northern and southern margins are constrained by extensive exposures of Paleozoic rocks shown on the geologic map of Eureka County (Roberts et al., 1967) and indicate a north-south dimension of ∼20 km; the perpendicular concealed dimension into bordering valleys is conservatively drawn at ∼25 km. On the basis of a brief reconnaissance, the entire mass of this part of the range appears to be underlain by Pancake Summit ignimbrite, from outcrops in the foothills at an elevation of ∼2135 m (7000 ft) above sea level to Summit Mountain at 3191 m (10,461 ft), giving an apparent minimum thickness of ∼1050 m. Some areas of variable alteration were noted inside the caldera but relatively fresh rock is widely exposed. In the southeastern sector of the caldera, somewhat altered, intracaldera Pancake Summit ignimbrite is overlain by brecciated slabs of gray limestone and black chert that represent slide masses of Paleozoic rock that entered the caldera from its margin ∼4.5 km to the southwest where similar in situ rock is exposed (Roberts et al., 1967). These slabs are in turn overlain by at least 150 m of ignimbrite of the 31.69 Ma Windous Butte Formation.

The extension-corrected area of 380 km² for the caldera could be a minimum inasmuch as nearly all of the east-west dimension lies within the mountain horst block in which extensional faulting might have been limited, but of this we cannot be certain. In any case, multiplying this area by the minimal, assumed uniform, thickness of 1050 m yields a minimum volume for the intracaldera ignimbrite of 400 km³ (Table 2).

The topography on which the outflow sheet was deposited was quite irregular because in most places the ignimbrite was the first to be deposited on the erosional surface cut into Paleozoic rocks. For example, in one locale in the northern Toquima Range, it is more than 200 m thick but within 6.5 km to the north, west, and south it pinches out completely on the flanks of an apparent paleovalley (McKee, 1976). The total volume of the Pancake Summit Tuff estimated from Model 2 is on the order of 700 km³ (Table 2), but this value has a large uncertainty.

Besides the Pancake Summit, three other ignimbrite cooling units older than the Windous Butte Formation occur in scattered outcrops east and southeast of the Allison Creek caldera (Table 1; Figs. 15 and 17). Sparse chemical analyses indicate a range in composition from dacite to rhyolite whereas modal compositions appear relatively uniform (Figs. 18 and 19). The oldest of these local units is the Sierra Springs Tuff Member of the Pinto Peak Rhyolite. It underlies the Pancake Summit Tuff and has K-Ar ages of 35.2 ± 1.9 and 36.2 ± 1.9 Ma (Blake et al., 1975; corrected for new decay constants). The Sierra Springs is known only in an area of 130 km² ∼15 km south of Eureka where it consists of three cooling units totaling at least 200 m thick that likely accumulated within a paleovalley (Blake et al., 1975). Two contrasting paleomagnetic directions are represented (Gromme et al., 1972). These sanidine- and clinopyroxene-free, hornblende-rich dacite tuffs are normally zoned upwards with decreasing quartz and increasing orthopyroxene and Fe-Ti oxides.

The tuff of Pritchards Station was emplaced 0.84 m.y. after the Pancake Summit Tuff at 34.46 ± 0.07 Ma. The Pritchards Station is a simple cooling unit as much as 150 m thick north of Pritchards Station (site UD in Fig. 17; Dixon et al., 1972), but it is thicker (215 m) on the east flank of the Pancake Range near Portuguese Mountain (Quinlivan et al., 1974); these occurrences might have accumulated in paleovalleys. Still farther east, the tuff was said by French and Freeman (1979) to constitute the reservoir rock in the Trap Spring oil field in Railroad Valley southwest of Currant. However, they admit the possibility that the reservoir rock might be the Windous Butte ignimbrite, which, according to our data, is a far more voluminous unit. Pumice and lithic fragments are conspicuous in the Pritchards Station ignimbrite. Southeast of Portuguese Mountain, the pumice clasts are as much as 0.5 m across and horizons of megabreccia of older rocks within the unit indicate a nearby source, which might not have been too far distant from the future ring fracture of the Williams Ridge caldera (see Fig. 20) that was the source of ash-flow tuffs of the 31.69 Ma Windous Butte Formation. A single chemical analysis indicates that the Pritchards Station is transitional between dacite and rhyolite (Fig. 18). Plagioclase is the dominant phenocryst, whereas quartz, biotite, and hornblende are subordinate and sanidine, two pyroxenes, and Fe-Ti oxides minor (Fig. 19).

The overlying tuff of Cottonwood Canyon has an age of 32.47 ± 0.10 Ma. It contains 72–73 wt% silica (Fig. 18); phenocrysts of plagioclase dominate over subequal amounts of quartz, sanidine, and biotite along with a few percent of hornblende and Fe-Ti oxides (Fig. 19). The simple cooling unit crops out as a continuous sheet as much as 50 m thick for many kilometers along the west and north flanks of the Park Range (Dixon et al., 1972), on the west and east flanks of the Antelope Range (Fig. 17), and very locally east of Portuguese Mountain (Quinlivan et al., 1974).

All together, the volume of the Sierra Springs, Pritchards Station, and Cottonwood Canyon ignimbrites was likely no more than 100 km³. The latter two are the only known ash-flow eruptions to have occurred in the Central Nevada field during the 3.6 m.y. between eruption of the Pancake Summit Tuff and the Windous Butte Formation. These two ignimbrites, together with local intermediate-composition lavas, are considered to represent minor leakages from the growing gigantic Windous Butte magma chamber.

The ash-flow tuff constituting the Windous Butte Formation is the most compositionally variable of any unit in the Central Nevada field and apparently had the largest volume, and its Williams Ridge caldera source is one of the largest in the Central Nevada field (Fig 20; Table 2). The weighted mean age of seven separate analyses of six samples of the tuff is 31.69 ± 0.02 Ma. The ignimbrite was first described by Faust and Callaghan (1948), named formally by Cook (1965), and subsequently adopted by Gromme et al. (1972) and U.S. Geological Survey geologists mapping in the Central Nevada caldera complex.

In most places, the outflow sheet is essentially a simple cooling unit but compound aspects are apparent in some thicker sections. Paleomagnetic data indicate but one eruptive unit (Gromme et al., 1972). The sheet has a prominent black, near-basal vitrophyre that is locally several meters thick and other zonal aspects normally found in thick ignimbrite deposits. Beginning with Cook (1965), investigators have noted a conspicuous contrast between the lower part of the Windous outflow tuff sheet that contains smaller phenocrysts (none are >2 mm and most are <1 mm) and the upper part in which phenocrysts are larger (as much as 3–5 mm). At site WAP as well as sections WD, WH, and WAB (Fig. 20), a knife-sharp contact between these two grain-size facies lies within densely welded tuff. Paleomagnetic directions above and below the contact at sites W8 and W9 (equivalent to site WD in Fig. 20) near the geographic Windous Butte are identical (Gromme et al., 1972). This implies negligible time difference between successive eruptions of contrasting crystal size, which can be interpreted as eruption from two parts of the magma chamber of contrasting crystal size.

In reconnaissance sampling, Ekren et al. (1974b) noted that the Windous Butte Formation displays a greater chemical variability than any other unit in the Central Nevada field. This astute observation is confirmed by our sampling which shows a range from high-silica rhyolite through low-silica dacite, from ∼76 to 65 wt% SiO₂ (Fig. 21A). There are at least fourfold variations in TiO₂ (0.18–0.67 wt%), Fe₂O₃ (1.39–5.16 wt%), and Ba (285–1180 ppm), and an order of magnitude range in V (13–103 ppm) and MgO (0.17–2.09 wt%) (Figs. 6 and 7). Other elements show lesser but nonetheless significant variations. Two samples of the phenocryst-rich rhyolite portion of the ignimbrite have different initial Sr isotope ratios (0.7099 versus 0.7101; Table 4). These ratios are among the highest in the Central Nevada field—only the older rhyolite Stone Cabin tuffs are higher—indicating incorporation of significant amounts of old crust into the magma.

The variations in chemical and modal composition (Figs. 21 and 22) are rather systematic throughout the outflow sheet, indicating normal zonation. Upper parts of stratigraphic sections, particularly in distal areas of the outflow sheet, are dacites that have greater concentrations of plagioclase and mafic phenocrysts, including sparse clino- and orthopyroxene, lesser amounts of sanidine, and little quartz. Lower parts of distal sections and proximal ignimbrite are rhyolite. Concentrations of TiO₂ + Fe₂O₃ in samples collected high in the unit in distal eastern parts of the outflow sheet (WP and WX sites in Fig. 20) are >5 wt% whereas more proximal samples are <2.8 wt%. Phillips (1989) documented upward increases in TiO₂, CaO, and MgO, and we also found upward increases in transition elements, Zr, and Ba, but a decrease in Rb.

The Windous Butte outflow ignimbrite is distributed in a highly eccentric manner around its Williams Ridge caldera source (Fig. 20). This pattern, as well as variations in thickness, were likely influenced, at least in part, by the terrain cut into Paleozoic rocks on which the Windous Butte ash flow was the first to be deposited over most of its extent. Windous Butte ignimbrite is definitely not present in the Toquima Range, currently within ∼50 km to the west of the caldera, where ash flows were apparently blocked by a north-south–trending topographic high at the time of eruption of the ash flows (Figs. 2 and 3). The southern Monitor Range, between the Toquima Range and the Williams Ridge caldera, is the location of the source caldera of the 25.70 Ma tuff of Luna Cuesta (BL in Fig. 2), which obscures any Windous Butte that might have been deposited there. No certain Windous Butte is known to the south or proximally southeast of the caldera source. In contrast, the outflow sheet currently extends as much as ∼240 km east-northeast of the source into western Utah; sections more than 100 m thick are found in a narrow lobe over one-half of this distance. This elongate pattern appears to have been controlled by a broad valley on the Great Basin altiplano. To the south, in the north Pahroc Range, the Windous Butte is as much as 160 m thick but pinches out just to the north, west, south, and more distant to the east, possibly again reflecting deposition in a paleovalley. A distal section to the north of the source caldera is at least 150 m thick where ash flows accumulated in the depression of the older Allison Creek caldera; sections to the south of this caldera are an order of magnitude thinner and then thicken abruptly to as much as 550 m immediately north of the Williams Ridge caldera source in an area where no older caldera depression existed.

Early workers (e.g., Ekren et al., 1973a) noted that the exceptionally thick (>2000 m) compound cooling unit designated by them as the tuff of Williams Ridge and Morey Peak accumulated within the caldera that formed by eruption of the Windous Butte ignimbrite. Our chronologic and paleomagnetic data confirm the contemporaneity of these two ignimbrite units. Here, we designate the two as the intracaldera and outflow (pre-collapse) ignimbrites, respectively, of the Windous Butte Formation. Compositional data are consistent with a comagmatic relationship; for example, the dacitic composition seen in the upper parts of distal outflow sections is matched by the also late-erupted intracaldera deposit (Fig. 21). Available analyses indicate the intracaldera ignimbrite is less variable in modal and chemical composition than the widely zoned outflow, extending neither to the outflow’s least nor its most evolved compositions.

Despite the apparent lack of a systematic spatial variation in composition in the dacitic intracaldera deposit, Phillips (1989) found increasing anorthite contents in plagioclase upsection. He believed that bytownite cores in an orthopyroxene-bearing sample of Windous Butte tuff at the top of his sampled section were derived from a high-temperature basaltic magma that possibly underplated the Windous Butte magma body during its evolution. He concluded that chemical variations throughout the entire Windous Butte ignimbrite can be explained by fractional crystallization of the major minerals in the magma system. Investigation of phenocryst-rich cognate inclusions and vitrophyric fiamme that are plentiful in the intracaldera tuff would provide valuable insights into compositional variations in the magma chamber.

The Williams Ridge caldera source (Fig. 20) of the Windous Butte Formation is revealed in one fashion or another on nearly all of the geologic maps of the Central Nevada caldera complex published by the U.S. Geological Survey in the early 1970s. It is also shown on the compilations of Sargent and Roggensack (1984) and Stewart and Carlson (1976b).

Most revealing of the caldera is the geologic map of the Moores Station quadrangle (Ekren et al., 1973a), which includes three cross sections based in part on data from deep drill holes. The intracaldera deposit—their tuff of Williams Ridge and Morey Peak—consists of three or more compound cooling units in drill hole HTH-3, which was collared at an exposure of the tuff (site WAC, Fig. 20) and bottomed at a depth of 1975 m, still in the tuff; at least 120 m of tuff lies above the drill site. Eighteen kilometers to the north, a similar thickness is exposed in a wedge-shaped area of ∼20 km². On the west margin of the quadrangle, and extending into the eastern side of the adjacent Morey Peak quadrangle (John, 1987), the entire mountain beneath Morey Peak, at an elevation of 3123 m (10,246 ft) to exposures near the base at 1952 m (6400 ft), is of the intracaldera facies—a thickness of at least 1170 m (Fig. 23; see also Best et al., 1989, their figure R20). Based on averaged dip and assuming no obscure faults that would repeat the section, the thickness of the ignimbrite is more than 2000 m. This thick massif of intracaldera Windous Butte ignimbrite beneath Morey Peak might be a resurgent block (John, 1987) later exposed as a horst created by basin-and-range block faulting. At the northern margin of the Moores Station quadrangle and extending into the adjacent Pritchards Station quadrangle (Dixon et al., 1972), landslipped and brecciated slices of older rock and of outflow Windous Butte tuff rest on the intracaldera facies just inside the northern topographic margin of the caldera.

To the east, at Black Rock Summit, Quinlivan et al. (1974; see also Best et al., 1989, their figures R24 and R25), found landslide masses of Paleozoic rock that are as much as 3 km in length embedded in variably altered intracaldera-facies tuff that is at least 1200 m thick. They conclude that at this site there was “...coincident emplacement of the landslide masses and the tuff during collapse of the Williams Ridge caldron [sic] in which the tuff accumulated.” (Whereas some geologists use the term “caldron” or “cauldron” for a deeply eroded ignimbrite source lacking topographic expression, in this article we use “caldera” for the structural ignimbrite source, following Lipman, 2000a.) Overlying the landslipped rock slices, additional tuff as thick as 400 m contains abundant zones of breccia and inclusions of Paleozoic carbonate rock ranging to as much as 30 m across. Thus, the thickness of the intracaldera ignimbrite at Black Rock Summit is at least 1600 m.

About 12 km to the south of Black Rock Summit, in The Wall quadrangle (Ekren et al., 1972), an area of ∼10 km² is underlain by altered and fragmented Paleozoic rock and Stone Cabin Formation. However, no intracaldera Windous Butte is exposed here and there are no drill-hole data to indicate if indeed these rocks represent landslide masses shed off the Williams Ridge caldera wall or perhaps the younger Monotony Valley caldera (see below).

In the southern Hot Creek Range ∼40 km south-southwest of Morey Peak at site WAS (Fig. 20), a wedge as thick as 1500 m of intracaldera Windous Butte ignimbrite overlies Ordovician rocks (Quinlivan and Rogers, 1974; note their cross section CC′). Overlying the ignimbrite, as much as 300 m of megabreccia of Mississippian rocks with a thin intercalated tuff is in turn overlain by as much as 600 m of middle Cenozoic sedimentary rocks and an intercalated lava flow. We interpret this 2400 m sequence to be the progressive accumulation of material inside the Williams Ridge caldera: first, 1500 m of ignimbrite deposited during subsidence, then 300 m of landslide debris shed off the nearby unstable caldera wall, followed by 600 m of more slowly deposited sediment and lava within the enclosed basin. Conformably overlying this 2400 m sequence is as much as 900 m of intensely altered ignimbrite designated the tuff of Twin Peaks by Quinlivan and Rogers (1974) that is in turn is overlain conformably by thick Monotony Tuff and tuff of Orange Lichen Creek (Table 1).

Several kilometers north of site WAS in the Hot Creek Range, somewhat altered ignimbrite more than 1000 m thick that is designated as the tuff of Hot Creek Canyon (Table 1) by Quinlivan and Rogers (1974) lies between the tuff of Orange Lichen Creek and Paleozoic rocks. This relationship implies that the margin of the Williams Ridge caldera lies just north of site WAS. However, complicating this relationship are conflicting data: Sample FLGSTF-1A collected in the lower part of the alleged Hot Creek Canyon unit has an age (31.66 ± 0.02 Ma) indistinguishable from the Windous Butte, but the paleomagnetic direction on the same outcrop (sample 8P701) is that of the much younger tuff of Lunar Cuesta (Table 1). Until the discrepancies in stratigraphic identification can be resolved, the exact location of the margin of the Williams Ridge caldera in the Hot Creek Range remains uncertain.

To the south in the Reveille Range, an area of ∼4 km² is underlain by one or more compound cooling units of variably altered ignimbrite more than 150 m thick that is designated as tuff of Williams Ridge and Morey Peak by Ekren et al. (1973b). This altered ignimbrite is shown to be in fault contact with Paleozoic rocks, but Martin and Naumann (1995) showed it in depositional contact on Paleozoic rocks. Several samples of the ignimbrite reveal substantial modal variations in phenocrysts and size of quartz; they resemble the rhyolitic outflow Windous Butte, as well as the rhyolitic Monotony Tuff (Table 1), rather than the dacitic intracaldera facies of the Windous Butte exposed to the north. No intercalated breccias or lithic-rich zones are known. It appears doubtful, then, that this altered ignimbrite in the Reveille Range is truly intracaldera Windous Butte. The alteration could be related to a post-Windous event.

The calculated, extension-corrected volume of the pre–caldera collapse ignimbrite by Model 1 is ∼2200 km³, of which 1700 km³ is outflow (Table 2). We deem this to be conservative inasmuch as no certain Windous Butte ignimbrite is known south of the source caldera where at least some outflow would be expected (Fig. 20).

Adopting the intracaldera Windous Butte deposit at site WAS in the Hot Creek Range as near the southwestern margin of the Williams Ridge caldera, its north-south dimension is ∼50 km and the minimum east-west dimension after correction for extension is ∼40 km, yielding a calculated area of ∼1600 km² (Table 2). For Model 2, assuming the average thickness for the intracaldera tuff is at least 1820 m, the total minimum volume of the Windous Butte ignimbrite is ∼4700 km³. For Model 3, with intracaldera thicknesses ranging from 1500 to more than 2000 m, the volume is at least 4800 km³, which is our preferred volume.

Added to this volume of the ignimbrite is that of the associated fallout ash deposit in the mid-continent, known in western Nebraska as the lower ash bed of the Whitney Member of the Brule Formation (Swinehart et al., 1985; Blaylock et al., 1997). This ash bed contains amphiboles that are indistinguishable in chemical composition from those in the Windous Butte ignimbrite. Sanidine ages of 31.63 ± 0.02 and 31.69 ± 0.02 Ma on samples BAYS-LW and ERD-LW from the bed are close to and identical to, respectively, the mean age of the Windous Butte ignimbrite. James B. Swinehart (1996, personal commun.) conservatively estimated the downwind, elliptical Windous Butte ash plume to have had a long axis of 750 km and a short axis of 200 km for an area of ∼120,000 km². The thickness of the ash layer in western Nebraska is 2–4 m; so using an average thickness of 3 m, the fallout volume in this narrow plume would be at least 350 km³. The total volume of the Windous Butte pyroclastic deposit is therefore apparently more than 5100 km³.

Along the northern margin of the Williams Ridge caldera, a silicic intrusive porphyry and a large rhyolite lava dome to the southeast (Fig. 20) probably represent magma intruded and extruded along its ring fault. The intrusion (sample WAX; unit Trq of Ekren et al., 1973a) has an age of 31.66 ± 0.07 Ma. Although not chemically analyzed, it is surely high in silica because of its abundant large phenocrysts of quartz and sanidine and only sparse plagioclase and biotite. The lava dome (unit Tlp of Quinlivan et al., 1974) consists of three superposed flows whose phenocryst assemblage is increasingly more mafic downwards and is ∼8 km in diameter and 500 m thick. An age of 31.62 ± 0.07 Ma on a sample (WAF) near its base together with an evolved composition (Fig. 21) relative to the Windous Butte ignimbrite indicates the magma that created the lava dome is a late differentiate of the Windous Butte system.

Farther south, at The Wall, along what is possibly the Williams Ridge caldera ring-fracture zone or the coincident(?) Monotony Valley caldera ring-fracture zone (see below) is a small dacitic lava flow (unit Tql of Ekren et al., 1972). It is no more than 300 m thick and has unusually large (to 1.5 cm long) phenocrysts of pyroxene. Though not dated, this lava flow lies stratigraphically between the Monotony (27.57 Ma) and the Palisade Mesa tuffs (29.97 Ma).

The oldest post-collapse, caldera-filling ignimbrite in the Williams Ridge caldera is the tuff of Black Rock Summit. It is exposed for 20 km north-south along the eastern margin of the caldera (Fig. 20; see also Ekren et al., 1972; Quinlivan et al., 1974). The tuff is as thick as 440 m and comprises an upper simple cooling unit and a lower compound cooling unit that all together have a volume on the order of 20 km³. The age of the tuff (31.57 ± 0.02 Ma) is consistent with the erupted magma being a late derivative from the slightly older Windous Butte magma system vented from the eastern ring fracture. Its composition is similar to the Windous Butte ignimbrites (Figs. 21 and 22) but rare earth elements and Y concentrations are higher than most samples and are more like those of the ring-fracture rhyolite dome.

A group of ignimbrites of roughly similar age but diverse paleomagnetic directions that postdate the Windous Butte Formation is exposed in the northern Toquima and Monitor Ranges (Fig. 24). These different cooling units are here informally designated the tuff of Hoodoo Canyon. Although of similar modal composition in containing abundant plagioclase and lesser quartz, sanidine, biotite, hornblende, and pyroxenes (Fig. 25), their chemical composition ranges from latite (61 wt% SiO₂) to trachydacite and low-silica rhyolite (Fig. 26).

The tuff of Hoodoo Canyon of McKee (1976) is widely exposed in the northern Toquima Range and is thickest (125 m) on the west flank in Hoodoo Canyon but thins abruptly elsewhere in the range to more or less continuous exposures no greater than 30 m thick. Gromme et al. (1972) cited a K-Ar age on biotite of 31.4 ± 1.2 Ma (revised to new decay constants) for a sample from the far north end of the range. A more precise but consistent ⁴⁰Ar/³⁹Ar age on sanidine from sample DH collected ∼6 km to the west in Clipper Gap Canyon is 31.51 ± 0.11 Ma. Although this age is indistinguishable from that of the tuff of Black Rock Summit, the paleomagnetic directions differ and the units are separated by 100 km, so they are not correlatives.

McKee (1976) indicated that similar biotite-rich tuffs to the north (e.g., site DB) of the exposures of his tuff of Hoodoo Canyon are likely correlatives. These tuffs were designated the tuff of Bottle Summit by McKee (1968a, 1968b) and Stewart and McKee (1968, 1969). Paleomagnetic directions of the tuff (site DB) near Bottle Summit just west of the Allison Creek caldera (Fig. 15) and at two sites (DA and DH) in the tuff of Hoodoo Canyon in the northernmost Toquima Range are all different from one another (Gromme et al., 1972), indicating that they were emplaced at different times, though possibly in the same reverse-polarity epoch. Furthermore, because only one cooling unit is exposed at each locale, these units might have erupted from more than one local vent in the northern Toquima Range. The total volume of the tuff of Hoodoo Canyon does not likely exceed 200 km³.

The first major explosive activity in the Central Nevada caldera complex after the super-eruptive Windous Butte event ∼1.7 m.y. earlier created many cooling units of the phenocryst-rich, high-silica rhyolite tuff of Hot Creek Canyon. This formation comprises compositionally similar cooling units (Fig. 26) that have similar ages with a weighted mean age of 29.97 ± 0.01 Ma (Tables 1 and 5).

Some of the cooling units included here in the tuff of Hot Creek Canyon are so designated on published maps (Ekren et al., 1973a; John, 1987). Paleomagnetic data at sites TD, TF, and TG (Fig. 27A) show that TF differs in direction from the other two, indicating it cannot be the same cooling unit. The four or more cooling units represented by samples TF and TG west of Morey Peak (John, 1987) have an aggregate thickness of more than 610 m; each cooling unit is normally zoned rhyolite.

Additional units that are designated by different names on published maps lack paleomagnetic information but because of similar ages and compositions are included here in the Hot Creek Canyon unit. West of Morey Peak, the tuff of Sixmile Canyon that is ∼200 m thick underlies the four or more cooling units of the tuff of Hot Creek Canyon (John, 1987). The Sixmile Canyon (sample TE) is an intracaldera unit that contains megabreccia blocks of Paleozoic rocks. On the Moores Station quadrangle (Ekren et al., 1973a; Table 5; see also Supplemental File 4 [see ^{footnote 4}]), correlative cooling units compose a sequence at least 1040 m thick; included are the tuff of The Needles Area (sample TB), the tuff of The Needles (TC), and an ignimbrite between the tuff of Moores Station Buttes and the tuff of The Needles. In addition to these ash-flow tuffs, the sequence includes more than 180 m of intercalated bedded tuffs, layers of coarse breccia (including clasts of intracaldera Windous Butte), and lacustrine sedimentary rocks. This sequence overlies the thick (∼460 m) tuff of Hot Creek Canyon (TA), within the eastern segment of the Hot Creek caldera. The 1680-m-thick accumulation reflects multiple explosive events interspersed with caldera-wall-collapse and sediment deposition during caldera subsidence.

In contrast to the foregoing correlative intracaldera cooling units, the tuff of Halligan Mesa of Snyder et al. (1972) is the only known correlative outside the Hot Creek caldera; its eruption might have been responsible for the initial collapse of the caldera. Nonetheless, the ignimbrite appears to be entirely confined within the older Williams Ridge caldera (Fig. 27A). At site TH, the Halligan Mesa has a similar age (Table 5) and paleomagnetic direction as the tuff of Hot Creek Canyon at sites TD and TG. The Halligan Mesa is also exposed in three other places: farther south on Palisade Mesa, just south of the Hot Creek caldera at Halligan Mesa, and to the east at The Wall (Ekren et al., 1972).

To the southwest of Morey Peak in the northern part of the Tybo quadrangle at ∼38°28′ N and 116°26′ W, a sequence of altered rocks designated by Quinlivan and Rogers (1974) as the tuff of Hot Creek Canyon is more than 980 m thick and rests on Paleozoic rocks and pre–Windous Butte dacite lavas. This sequence appears to be an intracaldera deposit because the upper hundred meters or so contains “at least one horizon rich in lithic inclusions as much as 4 feet [1.2 m] across of intermediate to silicic volcanic rocks, quartzite, sandstone, carbonate rocks, and minor metamorphic rocks and granite.” However, as explained above, this locale is problematic: sample FLGSTF-1A collected in the lower part of this sequence has a Windous Butte age of 31.66 ± 0.02 Ma but the paleomagnetic direction on the same outcrop (sample 8P701) is that of the tuff of Lunar Cuesta.

Overlying the tuff of Hot Creek Canyon are rhyolite ignimbrite units designated by four different names on as many geologic maps, but all have similar paleomagnetic directions (inclination –48° to –55°; declination 140° to 151°) and ages whose weighted mean is 29.97 ± 0.06 Ma (Table 1). These ignimbrites are considered to be a single cooling unit that we call the tuff of Palisade Mesa. On the eastern margin of the Williams Ridge caldera this unit is designated as the tuff of Big Round Valley by Quinlivan et al. (1974) at site EB and the tuff of Palisade Mesa by Snyder et al. (1972) in the central part of the caldera at site ED (Fig. 25B). At Palisade Mesa itself, where the ignimbrite is a multiple-flow compound cooling unit (Snyder et al., 1972), well-developed columnar joints are visible in the 150-m-high, west-facing escarpment alongside U.S. Highway 6 (Fig. 28). Within the eastern part of the Hot Creek caldera, the unit is designated the tuff of Moores Station Buttes by Ekren et al. (1973a) at site EC, where it overlies the tuff of Hot Creek Canyon as defined above; this is the thickest and most mafic part of the tuff of Palisade Mesa. North of the caldera complex at site EE, the deposit is called unit B of the tuff of Crested Wheat Ridge by Dixon et al. (1972).

Another possible outflow site lies at the south end of the Toquima Range 40 km north of Tonopah where no detailed geologic map exists. Although the paleomagnetic direction of the tuff at this site (0P169) is close to others for the unit, this direction is near that of the mean Oligocene dipole and there is accordingly less confidence in its correlation. Moreover, the age of this tuff (sample BAXTER-1A), 29.85 ± 0.07 Ma, is slightly younger than those of Palisade Mesa samples.

Although the tuff of Hot Creek Canyon and the tuff of Palisade Mesa are both high-silica rhyolites, the older unit is more evolved in its modal composition (Fig. 29). But compositions overlap and on element variation diagrams define a fairly coherent evolutionary trend (Figs. 26 and 30). Thus, the two units appear to constitute a normally zoned sequence erupted a short time apart, yielding identical ages within analytical uncertainty, but with sufficient elapsed time to allow a change in their paleomagnetic direction. Analyses of samples in stratigraphic sections (Phillips, 1989) indicate each unit is reversely zoned.

Based on the mapping of Ekren et al. (1973a) and John (1987), the northern margin of the Hot Creek caldera appears to coincide more or less with the northwestern margin of the older Williams Ridge caldera (Fig. 27A). The southern margin of the Hot Creek caldera is not defined geologically but is conservatively drawn so the caldera is an elliptical area of 330 km² after correction for extension. To estimate the total volume of the tuff of Hot Creek Canyon, we assume asymmetric caldera subsidence (Model 3) with accumulation of 460 m of intracaldera ignimbrite in the north, 810 m to the west, and 1500 ( = 460 + 460 +580) m in the east, yielding a preferred volume of 300 km³ (Table 2).

The eruption of the tuff of Palisade Mesa created no recognizable caldera. According to Ekren et al. (1973a), as much as 300 m of bedded tuff, tuffaceous sandstone, and lacustrine sediments in their map unit Tsb overlies the 305 m of the tuff at site EC inside the Hot Creek caldera (Fig. 27B; Supplemental File 4 [see ^{footnote 4}]). In nearby drill holes, the Tsb unit includes tuffaceous conglomerate composed of fragments of Palisade Mesa ignimbrite in a matrix of the same that grades downward into coherent Palisade Mesa ignimbrite (their tuff of Moores Station Buttes). The Tsb unit apparently manifests continued subsidence of the Hot Creek caldera during and after the Palisade Mesa eruption. Based on the isopachs in Figure 27B and using Model 1, the volume of the tuff of Palisade Mesa is 200 km³.

A 2.4 m.y. hiatus in explosive volcanism in the Central Nevada field followed after eruption of the Hot Creek Canyon–Palisade magma. No known ash flows and only a few cubic kilometers of andesitic lava flows were vented until super-eruptions of phenocryst-rich, dacitic magma occurred at 27.57 ± 0.04 Ma (Table 1), creating the Monotony Tuff and Monotony Valley caldera (Fig. 31). This hiatus in explosive activity, though two-thirds that before the preceding super-eruption of the Windous Butte, nonetheless could represent the time required for growth of the gigantic Monotony magma chamber before its eruption. The unit was named by Ekren et al. (1971) and the caldera by Ekren et al. (2011) after the small Monotony Valley near the southern margin of the caldera.

The Monotony outflow deposit consists of as many as three simple cooling units (Quinlivan et al., 1974). No more than two occur in a single stratigraphic section. The lowest unit A, which appears to be the most extensive outflow, is found from northeast of the source clockwise around to the south (Fig. 31A), the middle unit B is superposed on A in some sections, and youngest unit C occurs at one outflow site just south of the source and as an intracaldera deposit. Our sample set reveals no distinguishable differences in the compositions of these three cooling units; even the initial Sr isotope ratios are the same for samples of units A and B (Table 4 and Fig. 7F). ⁴⁰Ar/³⁹Ar ages on sanidines also do not differ significantly. Weighted mean ages are: A, 27.61 ± 0.05 Ma (n = 3); B, 27.56 ± 0.03 Ma (n = 3); C, 27.55 ± 0.05 Ma (n = 2). However, paleomagnetic directions for the three cooling units differ, and the differences are statistically significant. The outflow ignimbrite is locally underlain by lenses of conglomerate a few meters thick consisting of rounded clasts of older volcanic rock. Underlying bedded tuff (<1 m thick) has been noted by Quinlivan et al. (1974).

Possible occurrences of the Monotony are obscured in the southern Monitor Range immediately northwest of the Monotony source by the younger caldera source of the 25.70 MA tuff of Lunar Cuesta (BL in Fig. 2). Much farther west, in the Candelaria Hills, the correlative tuff of Miller Mountain of Robinson and Stewart (1984) is a simple cooling unit less than 60 m thick that generally rests unconformably on Paleozoic rocks or thin lenses of conglomerate. The unusually small phenocrysts (<1.5 mm in maximum length) in Miller Mountain sample MC compared to most samples of the Monotony might suggest an ash-fall deposit; however, lack of sorting and bedding and the presence of sparse lapilli of sedimentary rock in its basal part, together with variable welding—weak to moderate in the basal and upper parts and dense in the central part—indicate it is an ash-flow tuff.

Because the tuff of Miller Mountain lies so much farther west of other sites of the Monotony, we made a special effort to verify its correlation. Sample MC has an age of 27.58 ± 0.04 Ma that is indistinguishable from the weighted mean age of 27.57 ± 0.04 Ma of the Monotony. Its modal composition and concentrations of 11 chemical elements are consistent with other analyzed samples of the Monotony Tuff. However, Si, K, Rb, Nb, Pb, Th, and U are higher than the range of Monotony values, but for all of these elements, except Nb, the values of Monotony sample ML also lie at the high end of that range. Sample ML, collected ∼10 km northeast of the Monotony Valley caldera, has an age of 27.56 ± 0.04 Ma and is similarly fine grained as the Miller Mountain sample MC. Concentrations of TiO₂, Fe₂O₃, CaO, Na₂O, Sc, V, Cr, Ga, Sr, Ba, Ce, and Nd in sample MC are lower than the range of Monotony values. But for all of these elements, except Na, the values of Monotony sample ML also lie at the low end of that range. Thus, samples ML and MC both appear to be chemically extreme samples of the Monotony Tuff unit B. As an additional test of the Miller Mountain–Monotony correlation, we analyzed hornblende and biotite phenocrysts in samples MC and ML and compared them with analyses of Monotony phenocrysts in Phillips (1989). The data in Figure 32 show that the biotites and hornblendes in samples MC and ML are indistinguishable and that these and the other Monotony biotites and hornblendes are distinctive from those in other Central Nevada ignimbrites.

A “biotite lithic tuff” in the Paradise Range north of the Candelaria Hills has a modal composition similar to the Monotony. However, the K-Ar age on biotite is 25.6 ± 0.8 Ma (John, 1992, p. 16) and chemical analyses of the biotite and hornblende (in sample 84-DJ-16) are distinct from the Monotony phases (Fig. 32).

The only known exposure of the outflow Monotony between the Candelaria Hills and the source is in the nearby southern Cactus Range where it is ∼100 m thick (Ekren et al., 1971). To the west, near Goldfield, we considered the possibility that the widespread dacitic tuff of Chispa Hills might be equivalent to the Monotony. However, the K-Ar age is too young (21.7 ± 0.3 Ma, adjusted to new constants; Ashley, 1975) and the paleomagnetic direction (site 1P520) is unlike that of the Monotony. Another possible correlative dacitic ignimbrite is the stratigraphically appropriate unit Tdt of Ashley (1975) that contains “abundant plagioclase and biotite phenocrysts”; it occurs in only three very small exposures more than 100 m thick 12–13 km east of Goldfield. However, no chronologic or paleomagnetic data are available to confirm a correlation. Fridrich and Thompson (2011) identified a tuff in the Death Valley region of California 100 km southwest from the Monotony Valley caldera as Monotony Tuff based only on petrographic characteristics.

The Monotony Valley caldera source of the Monotony Tuff has a north-south diameter of 90 km—the largest known in the southern Great Basin ignimbrite province (Fig. 31). We discuss evidence for such a large caldera in a roughly clockwise route, beginning in the Quinn Canyon Range.

This eastern segment of the Monotony Valley caldera has been mostly obscured by development of younger calderas created during eruption of the 27–26 Ma Shingle Pass Formation. The 1:100,000-scale geologic map of the Quinn Canyon Range by Ekren et al. (2011) reveals several, mostly fault-bounded wedges of Monotony; no intercalated wall-collapse breccias are exposed. The largest terrane of Monotony that rests on the Eocene Sheep Pass Formation and Paleozoic rocks apparently exceeds 900 m in thickness, based on dips and outcrop width and assuming no internal faults. These exposures lie within what they designate as the Cherry Creek caldera segment of the Monotony Valley caldera complex; this segment is bounded on the north and west by high-angle normal faults which juxtapose Paleozoic rocks against Monotony. Locally overlying the Monotony Tuff are evenly bedded sedimentary rocks at least several hundred meters thick that are believed to have accumulated in a caldera lake.

We digress here, noting that the northern margin of the Monotony Valley caldera complex shown by Ekren et al. (2011, their figure 1) corresponds to the northern margin of the Pancake Range caldera shown by Stewart and Carlson (1976b) and Sargent and Roggensack (1984). In view of their extensive field work in central Nevada over four decades, we defer to Ekren and associates in their use of the designation Monotony Valley caldera for the source of the Monotony Tuff, dropping “complex” from the name because the other younger nested or overlapping calderas are simply part of the Central Nevada caldera complex. Abandonment of the Pancake Range caldera name also avoids confusion with the much older and unrelated Pancake Summit Tuff.

A gradient in gravity defines the southern margin of the Monotony Valley caldera (Fig. 31B). Between Sand Springs Valley and Monotony Valley to the west, the Monotony Tuff is 700 m thick; it rests on Ordovician strata and is overlain by Shingle Pass ignimbrites (Ekren et al., 1971). This type section of the Monotony appears to lie in the collar zone of the source caldera between its outer topographic margin and inner ring fault.

In the southernmost Kawich Range, another collar-zone section of the Monotony is at least 500 m thick but pinches out not far to the south. The wide part of the Kawich Range to the north harbors the caldera source of the 22.93 Ma Pahranagat Formation ignimbrite (A in Fig. 2), thus obscuring the western segment of the Monotony Valley caldera.

In the Reveille Range, a section of ignimbrite that is possibly 500 m thick was designated as Monotony Tuff by Ekren et al. (1972b), but Martin and Naumann (1995) interpreted it to be the similar but younger tuff of Goblin Knobs (Table 1).

Farther north, in the Hot Creek Range, Quinlivan and Rogers (1974) showed a layer as thick as 200 m of angular andesitic boulders and finer clasts embedded within Monotony Tuff that is as thick as 760 m. The northernmost exposures contain numerous fragments of bleached Monotony Tuff and other volcanic rocks in a matrix of mixed andesite and tuffaceous material. We interpret these relations to represent landslide deposits shed off the unstable topographic wall of the source caldera during accumulation of the Monotony cooling units.

Much thicker sections of intracaldera Monotony ignimbrite, similar to occurrences of the Windous Butte inside its Williams Ridge caldera source (Fig. 20), have not been found. However, in the southern Pancake Range and unmapped southern Hot Creek Range, there are broad exposures of massive, altered ignimbrite such as commonly constitute very thick intracaldera deposits. At site MV, we obtained the paleomagnetic direction of cooling unit C of the Monotony in alluvium-surrounded exposures of altered phenocryst-rich dacite tuff that are overlain by Shingle Pass ignimbrite (Table 1). A few kilometers to the southeast and ∼4 km west-southwest of Warm Springs, blocks of fractured Paleozoic rock lie in a terrane more than 400 m thick of massive, argillically and propylitically altered, phenocryst-rich dacite tuff. Some of these blocks of Paleozoic rock are sufficiently large to appear on the 1:250,000-scale map of Kleinhampl and Ziony (1985). This is surely an intracaldera deposit and was so interpreted by Best et al. (1993). Although we have no paleomagnetic or chronologic data to confirm this identification, the proximity to site MV leads us to believe it is Monotony Tuff. At site ME in the southern Pancake Range, we again obtained the paleomagnetic direction of cooling unit C of the Monotony, which appears to be at least 300 m thick.

Farther north in the Pancake Range, on the southwestern side of Palisade Mesa in the Lunar Crater quadrangle (Snyder et al., 1972), Shingle Pass ignimbrites, as old as 27 Ma, abut on a southward sloping surface against the 29.97 Ma tuffs of Palisade Mesa and Hot Creek Canyon; all three units are overlain by the 25.70 Ma tuff of Lunar Cuesta (Fig. 33). Ekren et al. (1974b, p. 607) concluded that this relation defines the “...wall or boundary of the caldera that formed as a result of the extrusions of the widespread Monotony Tuff.” The Shingle Pass ignimbrites represent the uppermost filling of the Monotony valley caldera; they are banked against its topographic wall, which was carved into pre-Monotony ignimbrites by northward caving of the caldera escarpment initially at the ring fault. Subsequent Lunar Cuesta ash flows flooded across the smoothed land surface. The only other ignimbrites younger than the Palisade Mesa and older than the Shingle Pass (Table 1) have caldera sources to the west, leaving only the eruption of the Monotony as a cause for the caldera collapse.

At the opposite, or northeast, end of Palisade Mesa, Shingle Pass ignimbrite is banked against a 120-m-thick exposed layer of Monotony Tuff; this layer is considered to be a pre–caldera collapse deposit that lies on the northeastward extension of the topographic margin of the Pancake Range caldera. Not far to the east, at The Wall (Ekren et al., 1972), the 450-m-thick Monotony, which we interpret to be an intracaldera deposit, is banked against a northeast-trending topographic high formed of pre-Monotony rocks, again indicating a segment of the topographic wall of the Monotony Valley caldera. In the vicinity of The Wall, it is uncertain whether the margin of the Monotony caldera follows or truncates the margin of the older Williams Ridge caldera.

As envisaged by Ekren et al. (2011), with supporting documentation just described, the Monotony Valley caldera (Fig. 31) is ∼90 km north-south and 60 km east-west, after compensation for 40% post-volcanic, east-west crustal extension. The caldera area is calculated to be ∼4800 km² (Table 2). Using the average thickness of the exposed intracaldera Monotony of 540 m in Model 2, the volume of the entire Monotony Tuff is conservatively calculated to be 4500 km³.

The Monotony Tuff is phenocryst rich and mostly dacite but ranges to low-silica rhyolite (Fig. 34). Clasts of pumice are generally inconspicuous except locally (e.g., Quinlivan and Rogers, 1974); lithic fragments are locally evident. Plagioclase is the dominant phenocryst, biotite and quartz are proportionately less, and sanidine, hornblende, clinopyroxene, Fe-Ti oxides, and orthopyroxene are minor constituents (Fig. 35A). Apatite, zircon, and allanite are accessory minerals. Some Monotony samples closely resemble some samples of the Windous Butte but distinction between these two units is generally made on the basis of stratigraphic position, isotopic age, Sr isotopic composition, and paleomagnetic direction. In overall composition, the Monotony Tuff is less variable: silica ranges between 66.9 and 71.4 wt%, TiO₂ 0.35–0.55 wt%, Rb 122–199 ppm, Sr 320–507 ppm, and Zr 149–203 ppm. The Monotony has a significantly lower initial ⁸⁷Sr/⁸⁶Sr ratio than the 4.1-m.y.-older Windous Butte (0.7087 versus 0.7099–0.7101; Fig. 7F). This is consistent with the overall pattern of decreasing ⁸⁷Sr/⁸⁶Sr with time.

The overall uniformity of the super-eruptive, phenocryst-rich dacitic Monotony Tuff was recognized by Ekren et al. (1971) and Phillips (1989). As first conceived by Hildreth (1981; see also Maughan et al., 2002), such relatively uniform and voluminous dacitic monotonous intermediates are a distinct class in the broad spectrum of ignimbrites. He listed the Monotony Tuff as an example and the Needles Range ignimbrites in the Indian Peak–Caliente field to the east (Best et al., 2013b) as another. The three compositionally similar, but older (31.13–29.20 Ma), Needles Range monotonous intermediates (Cottonwood Wash Tuff, Wah Wah Springs Formation, and Lund Formation) occur with the Monotony in some stratigraphic sequences. However, the Monotony Tuff, in addition to its distinctly abundant biotite phenocrysts (Fig. 5C), also tends to have more clinopyroxene and sanidine but less hornblende than the Cottonwood Wash and Wah Wah Springs ignimbrites and it lacks the characteristic titanite of the Lund. Moreover, the Monotony also has a lower initial ⁸⁷Sr/⁸⁶Sr ratio than these three monotonous intermediates in the Indian Peak–Caliente field: 0.7087 versus 0.7092–0.7117. Because little compositional data existed for monotonous intermediates at the time Hildreth (1981) astutely recognized them, we provide additional data regarding the uniformity of the Monotony Tuff in the following paragraphs.

Subtle textural zonation is evident in some stratigraphic sections of the Monotony. Commonly vitrophyric, near-basal parts of the cooling units tend to have somewhat smaller phenocrysts than upper parts, where biotite books and quartz are commonly 3–4 mm but as much as 5–6 mm.

In the 200-m-thick simple cooling unit at site MW a few kilometers north of the source caldera, concentrations of SiO₂ cited by Phillips (1989, his Palisade Mesa section) vary irregularly upward but a slight normal zonation is evident in TiO₂, Fe₂O₃, MgO, and CaO. At approximately the same site, modes in Page and Dixon (1994, their Sandy Summit section) reveal a subtle reverse zonation in quartz and mafic phenocrysts (Fig. 35B). The proportion of total phenocrysts in the MW section increases significantly upwards (not on the basis of dense rock equivalent, however) and virtually in parallel with increasing proportion of quartz. Despite the increasing quartz there is no sympathetic variation in whole-rock silica concentration. Unfortunately, no data are available on possible compositional variation in the phenocrysts through this section, which might explain the puzzling contrasting zonation. However, a possible explanation for the overall compositional variations is withdrawal of increasingly deeper magma from a pre-eruption chamber that had a weak normal chemical gradient and had a sufficient pressure gradient to stabilize increasing quartz with depth. Data on cognate pumice clasts hosted in the Wah Wah Springs tuff indicate a magma chamber possessing increasing quartz with depth (Best et al., 2013b). Experiments by Clemens and Wall (1981) and Johnson and Rutherford (1989) on granitic compositions reveal a broadening in the stability field of quartz at higher pressure.

With slight time lapses between successive eruptions of the three cooling units of the Monotony, it might be expected that there would be some change in the composition of the erupted magmas. Nonetheless, this is not everywhere apparent in stratigraphic sections through two units (nowhere are three directly superposed). In the section at sites MF, MG, and MH through cooling units A and B, no zonation in modes is evident (Fig. 35C). However, samples collected in the basal vitrophyres of units A and B at sites ML and MK, respectively, reveal between-unit normal zoning (Fig. 35D).

Hence, available compositional data reveal a lack of significant and consistent zoning in the Monotony monotonous intermediates. More data, particularly on cognate pumice clasts, are needed to fully evaluate compositional variations in the Monotony magma chamber and its history of evolution and three-stage explosive withdrawal.

The tuff of Ryecroft Canyon crops out in the Toquima and Monitor Ranges (Fig. 36; Shawe, 1999; Shawe and Byers, 1999; Shawe et al., 2000). It is a phenocryst-rich, high-silica rhyolite in which the modal proportions of quartz, plagioclase, and sanidine are roughly equivalent (Figs. 37 and 38A). In the southwest corner of the Corcoran Canyon quadrangle where Shawe et al. (2000) noted the presence of two cooling units separated by a layer of breccia, we found that the lower unit (site QRH) is reversely magnetized whereas the upper unit (site QRA) is normal. An age on the lower unit collected very near our QRH site of 27.43 ± 0.09 Ma (Henry and John, 2013; sample H94-30 in their figure 12) is significantly older than our age on the upper unit (QRA) of 27.14 ± 0.06 Ma. Another area of the upper unit (Shawe and Byers, 1999), from which our sample QRD was collected, was mapped as the 32.9 Ma Northumberland Tuff by McKee (1974, 1976; see also Henry and John, 2013) but later as tuff of Moores Creek by Boden (1986, 1992). The Moores Creek is believed by Shawe and Byers (1999) to be equivalent to their tuff of Ryecroft Canyon because of similar composition. A sample collected very near to QRD has an age of 27.29 ± 0.08 Ma determined by Henry and John (2013, sample H94-30 in their figure 12) that they identify as tuff of Moores Creek and which they correlate with the lower tuff of Mount Jefferson. According to the modal criterion of Shawe and Byers (1999) and Shawe et al. (2000), sample QRD is Ryecroft Canyon because the modal ratio of (quartz + sanidine)/(plagioclase + mafics) = 4.7; ratios of the Mount Jefferson ignimbrites are less than 1.0.

It is noteworthy that the modal compositions of the Ryecroft Canyon, Moores Creek, and silica-rich facies of the Lunar Cuesta (see below) are similar but differ from most other ignimbrites with sources in the Toquima Range vicinity, including the Mount Jefferson ignimbrites, in which plagioclase dominates over subordinate sanidine and quartz (see  Appendix).

A section ∼500 m thick consisting of several ignimbrite cooling units is exposed on the east-facing flank of Georges Canyon Rim at the southeast end of the Monitor Range. The lowest unit above alluvium, represented by sample qrc (Fig. 36), although designated as the tuff of Big Ten Peak on the map by Keith (1987), has an age of 27.24 ± 0.06 Ma, compared to 25.70 ± 0.04 Ma for the Big Ten Peak (see tuff of Lunar Cuesta below and Table 1). This age as well as the modal ratio (quartz + sanidine)/(plagioclase + mafics) of 2.3 is consistent with the upper tuff of Ryecroft Canyon (Shawe, 1999). It is not clear whether the 200 m or so of variably altered tuff overlying this unit is of the same cooling unit or a younger one, such as the high-silica rhyolite tuff of Lunar Cuesta.

The uppermost, 60-m-thick unit on the east flank of the Monitor Range 50 km to the north, represented by sample qrb, petrographically resembles our samples of the upper tuff of Ryecroft Canyon.

Sample qre collected ∼37 km northwest of Tonopah in a unit designated as lower tuff of Cedar Mountains by Whitebread and Hardyman (1987) is a possible Ryecroft Canyon correlative because of its similar composition and apparent age.

The entire tuff of Ryecroft Canyon is more than 200 m thick in the Toquima Range at site QRA/QRH where the two cooling units are separated by a layer of breccia. This section lies to the northwest of a source caldera in southern Monitor Valley postulated by Shawe et al. (2000; Fig. 36); the southern margin of this concealed caldera is believed to be marked by an arcuate string of rhyolite domes. Based on the >700 m thicknesses of the tuff of Ryecroft Canyon at the southern end of Monitor Valley, the total volume of the unit is ∼500 km³ (Table 2). However, if the Ryecroft Canyon is indeed equivalent to the tuff of Moores Creek, as suggested by Shawe et al. (2000), then the source caldera of the correlative ignimbrites would be exceptionally large, extending as much as 50 km to the northwest into the Toquima Range where Boden (1986, 1992) has mapped more than 900 m of the Moores Creek that he believes marks a source caldera. Further study of these ignimbrites, as well as their caldera sources and possible distal correlatives, is clearly needed.

The tuff of Pott Hole Valley and possible correlatives are also problematic. A simple cooling unit of phenocryst-rich, high-silica rhyolite (Figs. 37 and 38B) designated as the tuff of Pott Hole Valley is shown on only three published maps (Dixon et al., 1972; Ekren et al., 1973a; John, 1987) along the northern margins of the Williams Ridge and Hot Creek calderas (Figs. 20, 27, and 39). The unit is thickest (100 m) north of their margins and thinner within these calderas. Its source and volume are unknown.

Like the younger Shingle Pass and Clipper Gap ignimbrites (Table 1) and some of the older Stone Cabin and Pancake Summit ignimbrites, the Pott Hole Valley has a larger proportion of sanidine than plagioclase (Figs. 5A and 38B).

The age (27.31 ± 0.07 Ma) as well as the unusual paleomagnetic direction of the Pott Hole Valley are similar to those of the upper tuff of Corcoran Canyon exposed in the eastern Monitor and Toquima Ranges to the west (see  Appendix). However, our samples of the upper Corcoran Canyon are trachydacite with 65–67 wt% silica, which is far different from the 76 wt% in the single analyzed Pott Hole Valley sample, thus casting doubt on a correlation, unless the unit is highly variable.

The rhyolitic tuff of Orange Lichen Creek (Ekren et al., 1973a), represented in our samples RA, RB, and RD, is here correlated with the tuff of Kiln Canyon (RC) based on the same stratigraphic position (Quinlivan and Rogers, 1974), paleomagnetic direction, composition (Figs. 37, 38, and 40), and age. Sample RB has an age of 27.17 ± 0.06 Ma and RC a weighted mean of three replicate analyses of 27.11 ± 0.04 Ma. The tuff is a high-silica, phenocryst-rich rhyolite that has subequal amounts of felsic phenocrysts, lesser biotite, and still less hornblende and Fe-Ti oxides; some samples have a few grains of titanite.

In the Moores Station quadrangle (Ekren et al., 1973a), the Orange Lichen Creek is a simple outflow cooling unit, but to the west in the southern Hot Creek Range the correlative tuff of Kiln Canyon of Quinlivan and Rogers (1974) consists of two compound cooling units totaling more than 550 m thick. The two cooling units are separated in places by a 6-m-thick layer of breccia of Paleozoic rocks and in the lower 60 m of the lower unit there is as much as 50% clasts, partly of Paleozoic rock, and an intercalated lens 0–75 m thick of brecciated tuff. North of these two compound cooling units are three additional local units: (1) a pyroclastic deposit as much as 500 m thick that contains 10%–50% clasts to as much as 2 m in diameter of Paleozoic and other older rocks, that in places grades downward into the Orange Lichen Creek ignimbrite without a complete cooling break; (2) co-genetic rhyolite intruded into the Orange Lichen Creek ignimbrite and the pyroclastic debris unit; and (3) as much as 120 m of overlying nonwelded and bedded tuff and tuffaceous sediment. This sequence of rocks, which includes 1050 m of ignimbrite, was interpreted by Best et al. (1993) to have formed within the Kiln Canyon caldera source of the Orange Lichen Creek ignimbrite (Fig. 39). It is possible that these three local units mark the caldera margin and that the co-genetic rhyolite is a ring-fault intrusion.

Conformably beneath the intracaldera section of the Orange Lichen Creek in the southern Hot Creek Range are thick sections of intracaldera Monotony (760 m) and Windous Butte (1500 m) shown by Quinlivan and Rogers (1974, their cross section CC′). The intensely altered tuff of Twin Peaks, which is as thick as 900 m, lying between the Monotony and the Windous Butte might be correlative with the tuff of Hot Creek Canyon. Conformable superposition of the Orange Lichen Creek, Monotony, and Windous Butte, all of which contain wall-collapse breccias, implies overlapping of their three source calderas without late resurgence and tilting of the two older ones. The partial overlap of the caldera sources may be especially significant in the light of the initial ⁸⁷Sr/⁸⁶Sr ratios of their associated ignimbrites. The Orange Lichen Creek has a significantly lower ratio (0.7073) than either of the preceding much more voluminous ignimbrites or than those that follow it (0.7087 and 0.710; Fig. 7F). This contrast can be interpreted to have resulted from scavenging of most of the fertile old crustal material from the magma source region where the Orange Lichen Creek magma was generated. The relatively low initial ⁸⁷Sr/⁸⁶Sr ratio in the high-silica rhyolite of the Orange Lichen Creek further illustrates the lack of correlation between silica content and ⁸⁷Sr/⁸⁶Sr ratios. Instead, the Sr isotopic ratios show the open-system character of these long-lived magma systems and reinforce the general trend toward lower Sr isotope ratios with time.

On the basis of an intracaldera ignimbrite thickness of 1050 m in Model 2, the total volume of the tuff of Orange Lichen Creek is ∼600 km³. This conservative volume does not take into account the possible but unknown outflow north, west, and south of the Kiln Canyon caldera (Fig. 39).

In addition to the monotonous intermediate Monotony Tuff, a second distinct class of ignimbrite in the Central Nevada field is made up of several trachydacitic cooling units that contain less than 20% phenocrysts of plagioclase, two pyroxenes, and Fe-Ti oxides (Figs. 41 and 42). Because of their close resemblance to the more voluminous tuffs of the Isom Formation that crop out to the east in the Indian Peak–Caliente field (Best et al., 2013b), they are designated as Isom-type tuffs. Like them, the Central Nevada tuffs are typically dark-red to brown or black and are exposed as densely welded, relatively thin (generally 10–20 m, but locally to 100 m) simple cooling units that typically form cliffs or ledges. A black vitrophyre a meter or two thick is common near the base. A compaction foliation is defined by flattened vugs that are filled with vapor-phase minerals or secondary carbonate as well as by disc-shaped, varicolored pumice (fiamme) as much as several centimeters in diameter. Equally as ubiquitous as the pumice are lapilli and locally blocks of andesite; in some places both pumice and lithic fragments make up as much as one-half of the tuff.

With one exception, Isom-type tuffs in the Central Nevada field have not been dated by the ⁴⁰Ar/³⁹Ar method because sanidine and biotite are nonexistent or at most very sparse and, in some cases, xenocrystic, phases. An anomalously old age, in conflict with stratigraphic relations, for sanidine from an Isom-type tuff in the Toquima Range (unit Ti of Shawe et al., 2000) is consistent with a xenocrystic origin. But their ages can be roughly constrained stratigraphically by dated sanidine-bearing older and younger ignimbrites. On this basis, the age range of the Central Nevada Isom-type cooling units is ca. 27–23 Ma, which is close to the range in age of the tuffs in the Isom Formation to the east of 29.1 to 24.55 Ma. Notably, in both fields, Isom-type tuffs were initially deposited immediately after the close of the monotonous intermediate activity and subsequently for the next 4–5 m.y.

Most Isom-type tuffs in the Central Nevada field occur as a single simple cooling unit but some are found in sequences of two or more relatively thin cooling units that are directly superposed without other intervening deposits. They are commonly intimately associated in space and time with pyroxene-plagioclase andesite and dacite lava flows, at least some of which share the chemical characteristics (e.g., high Zr) of Isom-type tuffs.

Two single cooling units are exposed at sites well to the southeast of the Central Nevada caldera complex (Fig. 43); neither can be unequivocally correlated with cooling units in the widespread Isom Formation to the east. At site IL just south of Coyote Summit (Supplemental File 4 [see ^{footnote 4}]), a one-meter-thick vitrophyre that contains trace amounts of sanidine yielded a weighted mean age on duplicate analyses of 26.87 ± 0.06 Ma, in agreement with the position of the unit between the two oldest members of the Shingle Pass Formation (Table 1) whose ages are 26.82 and 26.98 Ma. A significantly younger Isom-type tuff designated the tuff of Pahroc Valley by Scott et al. (1992) occurs at site II in the North Pahroc Range (Supplemental File 4 [see ^{footnote 4}]). This 5–10-m-thick unit has an age of ca. 23.1 Ma, based on its position between the Bauers and Swett Tuff Members of the Condor Canyon Formation deposited at 23.04 and 24.15 Ma and derived from the Caliente caldera complex (Best et al., 2013b). In the northern Quinn Canyon Range at site IO, Ekren et al. (2011) noted the presence in their Tlc unit of an Isom-type tuff 30 m thick that overlies what they believe to be the 24.69 Ma tuff of Buckwheat Rim (Table 1).

In the Pancake Range near the northern margin of the Monotony Valley caldera (Fig. 31) at site IC, two simple, Isom-type cooling units totaling 30 m thick of the tuff of Black Beauty Mesa lie between the 22.93 Ma Pahranagat Formation and the 24.69 Ma tuff of Buckwheat Rim (Snyder et al., 1972; see also Lunar Crater section in Supplemental File 4 [see ^{footnote 4}]). These two Isom-type cooling units overlie a local pile of andesite and dacite lava flows that are described in more detail below (see the section Tuff of Buckskin Point and Tuff of Buckwheat Rim). A single thin (∼5 m) Isom-type tuff occurs 50 km to the northwest at the IH site; because of finer clasts, it is an apparently more distal paleomagnetic correlative of the lower unit of the tuff of Black Beauty Mesa. At the south end of the Pancake Range a single thin (10 m) Isom-type tuff at the IE site lies between the Tikaboo and Egan Tuff Members of the Shingle Pass Formation (Table 1) and therefore was deposited between 26.77 and 26.36 Ma. This Isom-type tuff is unusual in that it is the only one in the Central Nevada field with a reversed magnetic polarity. To the west, at site IK, in the southern Monitor Range, 6 m of an Isom-type tuff overlies the tuff of Clipper Gap (Gromme et al., 1972, their site C6) deposited at 24.95 Ma.

The thickest accumulations of Isom-type tuffs are found in the southern Monitor Range. On the southeast side at site IF, at Georges Canyon Rim, a sequence 103 m thick of four Isom-type cooling units overlies what might be the 27.14 Ma upper tuff of Ryecroft Canyon (sample qrc) and underlies the 22.93 Ma Pahranagat Formation (Table 1). The lower three of the four units contain relatively abundant andesitic clasts and xenocrysts(?) of quartz and sanidine and might correlate with units at the IB site. At this site and extending northeastward ∼4 km on the southeastern margin of the Toquima Range, two simple cooling units of Isom-type tuff overlie the upper tuff of Ryecroft Canyon but are older than 25 Ma (Shawe, 1998; Shawe and Byers, 1999). At site IA, an Isom-type tuff overlies the 26.93 Ma upper tuff of Mount Jefferson and may correlate with the Isom-type tuff mapped by Shawe (1999) and Shawe et al. (2000) at sites IM and IN and elsewhere in the Toquima Range; this unit (Ti) lies between the Tikaboo or Egan Tuff Member of the Shingle Pass Formation (Table 1) and volcaniclastic rocks of Little Table Mountain whose age is ca. 27.2 Ma (Shawe et al., 2000). In contrast to the 4200 km³ Isom Formation in the Indian Peak–Caliente field, the estimated aggregate volume of the several, widely scattered Isom-type cooling units in the Central Nevada field is ∼600 km³ (Table 2). No caldera has been recognized for any of the Central Nevada Isom-type tuffs.

The off-trend Isom-type tuffs in the Great Basin are distinctive in their trachydacitic composition (Fig. 41) and concentrations of many elements. Compared to other ignimbrites of comparable silica content (mostly 66–69 wt%) erupted from the Central Nevada and Indian Peak–Caliente caldera complexes (Best et al., 2013b), Isom-type tuffs have higher concentrations of TiO₂, K₂O, P₂O₅, Ba, Rb, Ce, Zn, Zr, and Th, and generally higher Nb, Y, and U (Figs. 6, 7, and 44; Table 3; Supplemental File 3 [see ^{footnote 3}]). On the other hand, they have distinctively lower concentrations of many compatible elements including MgO, CaO, Sr, Ni, Cr, and V. These attributes are consistent with an origin for the Isom-type magmas involving relatively high-pressure fractionation of pyroxenes, plagioclase, and Fe-Ti oxides from an andesitic parent (Christiansen et al., 1988).

As a consequence of high alkalies and low CaO abundances, Isom-type tuffs are alkali-calcic to alkalic (Fig. 6B), whereas most other tuffs in the Central Nevada field are calcic to calc-alkalic. Because of low MgO, Isom-type tuffs have high FeO*/(FeO* + MgO) and are thus ferroan, unlike most other tuffs of comparable silica content in the Central Nevada field which are magnesian (Fig. 6C). This, together with the anhydrous mineral assemblage, shows that they crystallized at lower fugacities of oxygen and water than the main-trend dacites and rhyolites.

Although there is an alkaline tendency in the Isom-type tuffs, they are not like A-type (including peralkaline) ignimbrites of the Great Basin that are younger than 20 Ma (e.g., Rowley et al., 1995). Like A-type silicic rocks (e.g., Whalen et al., 1987), Isom-type tuffs have high alkali and Zr concentrations (greater than 250 ppm in Fig. 7E). However, most do not have comparably high concentrations of Y and Nb and do not plot as within-plate granites (Fig. 44C) in the tectonic discrimination diagrams of Pearce et al. (1984). The high Zr concentrations are likely the result of anomalously high eruption temperatures. In addition, Isom-type tuffs in the Central Nevada field are distinct from most A-type silicic rocks in their lower Ga/Al, FeO*/MgO, K₂O/MgO, and (K₂O + Na₂O)/CaO ratios and lower concentrations of Zn and Ce.

We have analyzed the Sr isotopic composition of only one sample of an Isom-type tuff from the Central Nevada field; CHUCK-1DU (ICdu) was collected in the late-erupted part of the tuff of Black Beauty Mesa. Like ignimbrites of the Isom Formation in the Indian Peak–Caliente field (Best et al., 2013b) with initial ⁸⁷Sr/⁸⁶Sr ratios of 0.7076–0.7078, it has a low ratio of 0.7063 (Fig. 7F). This is the lowest ratio of any of the rocks we have analyzed from either field. Apparently, these alkaline tuffs are less contaminated by old continental crust than their more calc-alkaline contemporaries.

Normal chemical zoning is evident at sites IB, IC, and IF where two or more Isom-type cooling units are superposed on one another or where we have multiple samples in thicker vertical sections.

Phenocrysts constitute less than 20% of Isom-type tuffs (on a lithic-clast-free basis) and consist dominantly of plagioclase and lesser orthopyroxene, clinopyroxene, and magnetite plus minor ilmenite (Fig. 42A). Apatite inclusions occur in the pyroxenes. Major phases are commonly in clots whose disaggregation may account for the anhedral to subhedral habits of many isolated phenocrysts. The relatively hotter and drier phenocryst assemblage of off-trend Isom-type tuffs contrasts with the combinations of plagioclase, sanidine, quartz, biotite, hornblende, and Fe-Ti oxide phenocrysts that constitute main-trend middle Cenozoic ignimbrites in the Great Basin (Fig. 5) that were derived from cooler, wetter magmas.

In addition to local xenocrystic sanidine noted above, some samples contain a few anhedral grains of quartz and rarer hornblende, biotite, and zircon, all of which might also be xenocrysts. Isom-type tuffs in the Toquima and Monitor Ranges contain more of these phases than such ignimbrites to the east in the Central Nevada field, which in turn contain more than the Isom Formation in the Indian Peak–Caliente field (Best et al., 2013b).

Samples of the most evolved Isom-type tuff at site IB on the western side of the Monitor Range contain similar types and proportions of phenocrystic phases as the tuff of Mount Jefferson and the upper tuff of Corcoran Canyon (Fig. 42B) to the west in the Toquima Range ( Appendix). This is consistent with a possible genetic link between Isom-type magmas and the magmas that created the broadly contemporaneous Toquima ignimbrites. Most of the less silica-rich Toquima tuffs are alkali-rich trachydacite rather than dacite and possess—relative to main-trend Central Nevada ignimbrites at comparable silica concentrations—lower concentrations of CaO, MgO, Sr, and V but higher TiO₂, P₂O₅, Ba, Sc, and Zr as do Isom-type tuffs. P₂O₅ is especially high in Toquima tuffs (Supplemental File 3 [see ^{footnote 3}]). These compositional similarities indicate that large volumes of Isom-type magmas in the Toquima Range–Monitor Range area might have evolved to varying degrees by crystal fractionation and crustal contamination, resulting in cooler and wetter magmas that erupted to form the Toquima ignimbrites.

The Shingle Pass Formation comprises four widespread, outflow cooling units of rhyolite ignimbrite (Table 1; Fig. 45) that were emplaced from 26.98 to 26.36 Ma contemporaneously with older Isom-type tuffs. A fifth cooling unit included in the formation is found in the intracaldera sequence; a possible outflow correlative lies to the south. Three of the outflow cooling units (X, Y, Z in Figs. 5–7) are quite distinct in appearance and composition from the main-trend rhyolite and dacite tuffs in the Central Nevada field and have chemical affinities with Isom-type tuffs.

Cook (1965) was the first to recognize a distinct, densely welded, generally thin but widespread rhyolite tuff in eastern Nevada that he called the Shingle Pass Ignimbrite. Its type section is west of Shingle Pass in the southern Egan Range (our site XU in Fig. 46A; see also Supplemental File 4 [see ^{footnote 4}]; Best et al., 1989, their figures R27 and R28). Although our data reveal a wider areal distribution of the ignimbrite than shown by Cook (1965), his figure 15 is remarkable in showing the same north-south elongate distribution we have documented. Scott (1965) correlated similar ash-flow tuffs in the Grant Range with the Shingle Pass Ignimbrite and later (Scott, 1966) referred to these as the Shingle Pass Formation. Ekren et al. (1971) and Sargent and Orkild (1973) mapped Shingle Pass cooling units on the Nevada Test and Training Range. Ekren et al. (1972, 1973a, 1974a, 1974b, 1977) recognized two to three units in the Central Nevada caldera complex and in Lincoln County to the east.

Cook (1965, his figures 5 and 7) showed additional rhyolite cooling units overlying his Shingle Pass Ignimbrite. The directly overlying unit is similar in some respects and is coextensive with his Shingle Pass Ignimbrite. We believed (Best et al., 1989) that the unit was derived from the same source, probably in the Quinn Canyon Range, and, therefore, recognized lower and upper units of the Shingle Pass, a designation adopted by Page and Dixon (1994). Our paleomagnetic, chronologic, and compositional data confirm the widespread coexistence of these two separate units. The lower, we here formally designate as the Coyote Summit Tuff Member for its occurrence 1.0 km south-southeast of the Coyote Summit (site XW on Fig. 46A; see also Supplemental File 4 [see ^{footnote 4}]), where it directly overlies the Monotony Tuff and underlies a very thin cooling unit of Isom-type tuff. The upper unit we here formally designate as the Egan Tuff Member for its occurrence in the southern Egan Range as explained above (see also site ZX on Fig. 46D). Both of these members are generally simple cooling units, but in the type section of Cook (1965) and elsewhere they are compound; at the Coyote Summit section (Supplemental File 4 [see ^{footnote 4}]), the Egan Tuff Member consists of two cooling units separated by a few centimeters of bedded tuff.

Another compositionally similar ignimbrite, which occurs everywhere as a compound cooling unit between the Egan and Coyote Summit Tuff Members and is more or less coextensive with them, is here designated the Tikaboo Tuff Member. This name is taken from Tikaboo Valley that lies to the southeast of the stratigraphic section at Coyote Summit (Supplemental File 4 [see ^{footnote 4}]). The thickest outflow section (140 m) is exposed at the south end of the Pancake Range (Figs. 46C and 47).

The paleomagnetic directions of these three members are distinct (Fig. 48). The Egan Tuff Member has reversed polarity, the Tikaboo is normal with a north-northeast declination, and the Coyote Summit is normal with a northwest declination.

A fourth rhyolite cooling unit, here designated the Hancock Tuff Member, lies between the Tikaboo and Coyote Summit Tuff Members in many places. The name is taken from Hancock Summit (Supplemental File 4 [see ^{footnote 4}]) along Nevada State Highway 375 in the Pahranagat Range (Fig. 49). The Hancock, called the tuff of Murphy Gap by Page and Dixon (1994), is widespread and locally quite thick to the north, east, and south of the Quinn Canyon Range. Thinner sections are a simple cooling unit but ones thicker than 50 m are compound. Although petrographically unlike the other Shingle Pass tuffs in being generally more phenocryst and quartz rich, the Hancock shares the same areal distribution (Fig. 46) and the same high Ba/TiO₂ ratio (H on Fig. 7D). Relations observed at section HI 30 km to the southeast of the Quinn Canyon Range are consistent with a proximal source: thickness more than 100 m; the only vitrophyre (∼2 m thick) seen in this unit; the most abundant (10%) and largest (to as much as 3 cm) cognate pumice clasts found anywhere; and slab-like inclusions to as much as 0.3 m long of flow-layered, nearly aphyric rhyolite. In most places the remanent magnetization of the Hancock Tuff Member is very weak to nil, i.e., it commonly has no usable paleomagnetic signature.

Ages of the Tikaboo and Hancock Tuff Members are analytically indistinguishable at 26.77 ± 0.01 Ma and 26.82 ± 0.09 Ma, respectively, whereas the Coyote Summit Tuff Member (26.98 ± 0.04 Ma) and the Egan Tuff Member (26.36 ± 0.06 Ma) are distinct (Table 1).

Our type section for these four members of the Shingle Pass Formation is at Hancock Summit (Fig. 49) and a reference section lies at Coyote Summit. Details of these two stratigraphic sections are in Supplemental File 4 (see ^{footnote 4}).

The four tuff members comprising the Shingle Pass Formation occur in outflow sequences more or less surrounding the Quinn Canyon Range where numerous correlative intracaldera cooling units crop out within a compound(?) source caldera (Fig. 46). According to Ekren et al. (2011), the youngest cooling unit in the Shingle Pass sequence in the range is the tuff of Sawmill Canyon. This ignimbrite directly underlies the 22.93 Ma Pahranagat ignimbrite or the 24.69 Ma tuff of Buckwheat Rim (Table 1) and overlies the Egan Tuff Member of the Shingle Pass Formation. They note that the Sawmill Canyon has an associated plug-like vent in the range and is a petrographic “near twin” of the outflow Coyote Summit Tuff Member, including the occurrence of water-clear phenocrysts of sanidine as much as 5 mm long and deeply embayed quartz phenocrysts. They provide chemical and modal analyses of three samples in a 200-m-thick section on the southeast side of the range. In plots of Sr-Zr-Ba-TiO₂ (Fig. 45), the Sawmill Canyon lies between the Coyote Summit and Hancock Tuff Members. Despite the lack of detailed characterization such as we have on the other four members, we include the Sawmill Canyon as the fifth and youngest member of the Shingle Pass Formation (Table 1). Whatever conditions prevailed in the Shingle Pass magma system to allow creation of the compositionally unique Coyote Summit Tuff Member (see below), they were apparently renewed late in the evolution of the system to create the similar Sawmill Canyon Tuff Member.

South of Coyote Summit, an uncorrelated, “orphan” cooling unit ∼20 m thick overlies the Egan Tuff Member of the Shingle Pass Formation and underlies the high-silica rhyolite facies of the tuff of Lunar Cuesta (Table 1; Supplemental File 4 [see ^{footnote 4}]). Sanidine from sample TEMPMS-1A of the orphan cooling unit yielded a stratigraphically consistent age of 25.99 ± 0.05 Ma. The paleomagnetic direction (site 1P342) is normal, with inclination of 61.8° and declination of 55.1° (compare Fig. 48). The cooling unit might have been derived from the Bald Mountain caldera ∼10 km to the south that was active at ca. 25 Ma (Ekren et al., 1977) but the single mode of the orphan unit is distinctly different from the main ignimbrite unit derived from this caldera. On the other hand, the mode of the orphan closely matches three modes of the Sawmill Canyon (Fig. 50E) exposed in the Quinn Canyon Range ∼50 km to the north. Moreover, phenocrysts of sanidine in the orphan are large and quartz is deeply embayed, as is characteristic of the Sawmill Canyon. Comparison of available chemical analyses of the Sawmill Canyon and the orphan tuff indicate similarity in some constituents but small differences in others (Table 6). Without additional confirming data, we tentatively correlate the orphan cooling unit with the Sawmill Canyon Tuff Member, and suggest that the age of 25.99 ± 0.05 Ma might apply to that member (Table 1). If the correlation is confirmed, the cooling unit south of Coyote Summit would represent the only known outflow of the member.

Two kilometers north of Queen City Summit at 37°46′18″ N and 115°56′36″ W, two ignimbrite cooling units that underlie the Pahranagat ignimbrite (Table 1) are each a few meters thick. The lower unit (sample BLCKTP-1B) is rich in phenocrysts, half of which are quartz and the remainder sanidine, lesser plagioclase, and a trace of biotite (Supplemental File 2 [see ^{footnote 2}]). The upper quartz-free unit (BLCKTP-1C) petrographically resembles the Egan Tuff Member of the Shingle Pass Formation. However, their paleomagnetic directions (sample sites 1P271 and 1P279, respectively) are unlike any of the known Shingle Pass members. They are possibly local outflow cooling units of the formation, located as they are not far south of the Quinn Canyon Range source area. Or, they were possibly derived from the Bald Mountain range (see above).

Ekren et al. (1971, p. 33) described an “alien” tuff between two cooling units of their tuff of White Blotch Spring in areas to the southwest of Queen City Summit; because these two cooling units are the Pahranagat and the high-silica facies of the tuff of Lunar Cuesta (Table 1; see also below), the age of the alien tuff is between 22.93 and 25.70 Ma; its mode is roughly similar to BLCKTP-1C. Ekren et al. (2011) considered the alien tuff to be a very late expression of the Shingle Pass explosive activity.

Except for the Hancock Tuff Member, Shingle Pass ignimbrites share some common petrographic attributes, as noted by previous workers. The Coyote Summit, Tikaboo, and Egan Tuff Members are among the most colorful tuffs in the Central Nevada field because of the white to pastel-colored pumice lapilli surrounded by darker gray, brown, orange, red, or purple matrices. Compared to most rhyolite ignimbrites in the field, these three contain fewer phenocrysts and, for a given proportion of plagioclase, most have relatively less quartz but distinctly more sanidine (Table 1; Fig. 5). These proportions among the felsic phenocrysts cause their modes to plot well off the trend of other Central Nevada ash-flow tuffs. Such low proportions of quartz imply these Shingle Pass magmas crystallized at low pressure, or were water poor, or both (e.g., Whitney, 1988). The low proportions of biotite also show the water-poor character of the Shingle Pass magmas.

In most exposures, these three off-trend ignimbrites possess a prominent compaction foliation defined mostly by conspicuous eutaxitic pumice clasts but also by discoidal lithophysae. In many places, individual, relatively thin members are exposed as pronounced ledges because of their dense welding and erosional resistance. Black or dark-gray vitrophyres, as thick as several meters, are locally conspicuous near the base of cooling units. In these respects, together with the low phenocryst content, the three resemble the Isom-type tuffs, but are readily distinguished because of conspicuous sanidine and less biotite.

Despite the overall similarity of the three members, the Coyote Summit Tuff Member is modally distinct (Fig. 50A; see also below), whereas the Tikaboo and Egan Tuff Members are more similar to one another.

This simple cooling unit is mineralogically nearly unique among all of the middle Cenozoic ignimbrites in the Great Basin in that it contains small amounts of Fe-rich olivine (Fa₈₆). Because of their susceptibility to alteration, the olivines are usually seen only as spots of limonite or hematite. The only other middle Cenozoic ignimbrite reported to contain this phase is the 24.96 Ma Underdown Tuff exposed ∼50 km southwest of Austin (Bonham, 1970); its correlative tuff of Clipper Gap (Table 1) may also have contained Fe-rich olivine, judging from the presence of a few similar limonite/hematite spots. The most abundant olivine (4.5% of total phenocrysts) in any sample of the Coyote Summit is found in the basal bedded tuff at site XV. Most samples of the Coyote Summit also contain a few small, variably but similarly altered phenocrysts of Fe-rich clino- and orthopyroxene that appear to have equilibrated in the magma with the other crystalline phases (Nielsen, 1992). Some samples contain a few phenocrysts of amphibole, which is enriched in Fe, alkalies, and Cl, and of biotite. Compositions of all of these mafic phases are consistent with equilibration in the magma.

Ubiquitous phenocrysts of quartz are typically skeletal, especially in their outer parts. Because quartz phenocrysts in the local basal plinian(?) beds at site XV are not skeletal, we speculate that the presence of such crystals in the overlying ignimbrite is a result of the rapid supersaturation of the erupting magma system as it decompressed, cooled, degassed, and experienced rapid further crystallization. In support of this speculation, we note that Swanson and Fenn (1986) found that a rapid drop in temperature of more than 55 °C in laboratory magma systems yielded skeletal quartz.

An additional characterizing attribute of the Coyote Summit is the abundance (Fig. 50A) of Ba-rich sanidines that occur as robust, water-clear blocky tablets, some as much as 4–5 mm long. Plagioclases are complexly zoned andesine-oligoclase. Magnetite and sparse ilmenite constitute less than 7% of the phenocrysts. Allanite (rather than chevkinite), apatite, and zircon are accessory phases.

As might be expected from its sanidine-rich character, the Coyote Summit Tuff Member is somewhat enriched in Ba and K₂O on a whole-rock basis but is slightly poorer in CaO and MgO, relative to main-trend Central Nevada rhyolite tuffs (Figs. 6 and 7). Many of these sanidine-bearing rhyolites have low Ba because of extensive fractionation of sanidine, whereas the Coyote Summit magma started out more Ba rich, but was on a declining Ba trend, co-varying with TiO₂.

At site XE, the Shingle Pass ignimbrite, called the Wood Canyon facies of the Coyote Summit Tuff Member of the Shingle Pass Tuff by Nielsen (1992), contains no apparent olivine and pyroxene but has other characteristics of the Coyote Summit Tuff Member, including skeletal quartz phenocrysts, appropriate chemical and modal composition, and normal paleomagnetic polarity. This facies might represent late ejecta from the lower part of a zoned magma chamber.

The pastel-colored, moderately to densely welded Hancock Tuff Member is petrographically similar to some main-trend, quartz-rich rhyolite tuffs in the Central Nevada field and differs from other members of the Shingle Pass Formation. Unlike them, it is typically not foliated because of the small proportion of biotite, the large proportion of equant felsic phenocrysts (subequal proportions of quartz, sanidine, and plagioclase; Fig. 50B), and the general lack of pumice clasts. Quartz is typically smoky in the Hancock in contrast to clear in other Shingle Pass ignimbrites. Mafic phenocrysts include a few percent of small biotites and still smaller and fewer magnetites. However, upper and more distal parts of the unit contain in addition no more than a few grains (xenocrysts?) of amphibole and orthopyroxene. Also in distal sections, the Hancock has smaller, less abundant phenocrysts, ∼1 mm or less in diameter, compared to as much as 4 mm in the more proximal HI section in which 37% of the tuff is phenocrysts. In many outcrops, lighter-colored haloes in the matrix surround holes a centimeter or so in diameter that in places contain yellowish inclusions of altered carbonate(?) rock. Thick, compound cooling units that are more than 50 m thick are variably welded, and at site HF, the most densely welded tuff lies near the center of the unit; the top is porous, soft, and partly bedded.

The Hancock Tuff Member at the type section at HF is also normally zoned upwards to a greater proportion of plagioclase and lesser quartz (compare the tuff of Murphy Gap of Page and Dixon [1994, p. 64]).

These two members resemble one another modally (Figs. 50C and 50D) and are difficult to distinguish in the field. Petrographically, they resemble the Coyote Summit Tuff Member in many respects, such as variable matrix coloration, pumice-matrix contrasts, general dense welding, modest amounts of phenocrysts, sparse proportions of mafic phenocrysts, and types of accessory minerals. However, mafic phases are less Fe rich and ilmenite appears to be absent. Furthermore, the two differ in lacking olivine and having typically more plagioclase (zoned from oligoclase to bytownite in the upper member) than sanidine. Whereas the Coyote Summit Tuff Member has little or no biotite, the upper two members contain more. This trend is opposite to that of quartz, which is common in the lower unit and scarce in the younger two members. The contrast in the amount of quartz between the Coyote Summit and Egan Tuff Members (Fig. 51) is not a function of silica concentrations, which are similar, but of more advanced crystallization in the lower unit. However, the small proportion of quartz in some samples of the Tikaboo Tuff Member is likely a result of a lower relative silica concentration and higher crystallization temperature. No consistent modal zoning is evident in the Tikaboo and Egan Tuff Members. However, two samples of the presumed intracaldera correlative of the Tikaboo in the Quinn Canyon Range include a low-silica rhyolite overlain by a trachydacite (Ekren et al., 2011, their tuff of Yurt Buttes and McCutchen Spring).

Ekren et al. (1973b) referred to a mélange of chaotic blocks in the northernmost Reveille Range, near Twin Springs Ranch, some of which contain the same Shingle Pass ignimbrites as exposed to the north in the Pancake Range, but also include “...an olivine-bearing unit present at the type locality...[but which] is not found depositionally in place in the Reveille and southern Pancake Ranges.” This olivine-bearing unit is surely the Coyote Summit Tuff Member, which must have been deposited only very locally somewhere nearby the breccia layer and ∼20 km west-northwest of the caldera source (Fig. 46A).

Both Ekren et al. (1973b) and Martin and Naumann (1995) showed small exposures of apparently a single cooling unit of Shingle Pass ignimbrite 55 m thick on the west side of the central Reveille Range; its mode cited by the latter workers is consistent with either the Egan Tuff Member or the Tikaboo. We tentatively correlate the unit with the Tikaboo. Thicknesses of what we assume to be Tikaboo to the south on the inaccessible Nevada Test and Training Range (Ekren et al., 1971) are queried in Figure 46C. The Tikaboo is as thick as 100 m in the slightly older Mount Jefferson caldera in the Toquima Range. We have no information on a possible more westerly occurrence.

Outflow sheets of the Shingle Pass are eccentric to varying degrees relative to their Quinn Canyon Range source (Fig. 46). Part of this eccentricity may only be apparent because of the presence of the younger source calderas of the Goblin Knobs and Pahranagat ignimbrites (Fig. 2) to the west into which Shingle Pass outflow ignimbrites might have been engulfed. Part reflects the lack of data southeast of Tonopah in, for example, the Cactus Range, where only small-scale mapping is available and widespread alteration obscures the identity of ignimbrite units (Ekren et al., 1971); this area also lies on the Nevada Test and Training Range that has been inaccessible since the 1980s. The origin of the eccentricity of the Shingle Pass outflow sheets is problematic. The apparent absence of outflow to the west for three sheets would imply control by a long-standing topographic high, rather than special eruption dynamics that might not be expected for all three. The postulated high might have existed before and after the deposition of the Shingle Pass, as is evident in distribution maps for other ignimbrite units (see the section Topographic Barrier). However, the eccentricity of the Tikaboo is opposite that of the other outflow sheets.

Another unusual facet of the distribution of the Shingle Pass outflow sheets in conjunction with their eccentricity is their elongate north-south extent. Most extreme in this regard is the Egan Tuff Member, which is 30 m thick 160 km to the north of the margin of the caldera source (Fig. 46D). That this Shingle Pass site (ZD) is indeed the upper unit is confirmed by its reversed paleomagnetic polarity (Fig. 48).

That the Quinn Canyon Range is the source of ignimbrites of the Shingle Pass Formation is based in part on the nearby occurrence of thicker ignimbrites in at least some surrounding quadrants (Fig. 46). Recent 1:100,000-scale mapping by Ekren et al. (2011) has revealed the presence of thick accumulations of numerous cooling units of Shingle Pass ignimbrites and some interleaved, probable wall-collapse breccias of older rocks that confirms the Quinn Canyon Range as the source of the Shingle Pass Formation (for early ideas on the source, see Sargent and Houser [1970] and Ekren et al. [1977]). Ekren et al. (2011) recognized correlatives of the upper and lower units at the type section at Shingle Spring (i.e., Egan and Coyote Summit Tuff Members; see also Best et al., 1989; Page and Dixon, 1994), as well as a third overlying unit—the tuff of Sawmill Canyon. Structural and stratigraphic complexity in the caldera has resulted from overprinting by recurrent caldera collapse, resurgence, and later basin-and-range faulting.

Following Sargent and Roggensack (1984), we use the appellation “Quinn Canyon Range caldera” to designate the composite source of ignimbrites of the Shingle Pass Formation. The northeastern margin of the caldera follows what Ekren et al. (2011) showed as the Cherry Creek and Stairstep Mountain segments of the Monotony Valley caldera (Fig. 31); we approximate the remaining caldera margin around the intracaldera Shingle Pass cooling units (Fig. 46).

The intracaldera Shingle Pass unit in the Quinn Canyon Range that locally rests on the Monotony Tuff is designated the tuff of Stairstep Mountain by Ekren et al. (2011). Its modal and chemical compositions (Figs. 45 and 50) are equivalent to the outflow Coyote Summit Tuff Member. Deeply embayed phenocrysts of quartz occur in both. The characteristic Fe-rich olivine of the Coyote Summit Tuff Member is not reported in the Stairstep Mountain, but some of the widespread altered mafic phenocrysts reported by Ekren et al. (2011) might be relics. On the other hand, because the magma forming the intracaldera ignimbrite was a later eruptive product than that creating the outflow, it might never have contained Fe-rich olivine, as the Wood Canyon facies described above. The Stairstep Mountain ignimbrite crops out in the range in fault-bounded wedges, the largest of which in the southwest covers an area of ∼4 × 15 km designated as the Stairstep Mountain caldera segment. This intracaldera pile is shown on the cross section in Ekren et al. (2011) to have a thickness of at least 1000 m. It seems reasonable to believe that this thick, multiple-flow, compound cooling unit marks the source of the Coyote Summit Tuff Member within the Quinn Canyon Range caldera (Fig. 46A).

Some of the cooling units in their tuffs of Mystery Mountain in the Quinn Canyon Range are believed by Ekren et al. (2011, p. 25) to be like the tuff of Murphy Gap, as the Hancock Tuff Member was called by Page and Dixon (1994). However, the modes of these cooling units are different from our modes of the Hancock (Fig. 50B); moreover, their stratigraphic position between the Monotony Tuff and tuff of Stairstep Mountain (Coyote Summit Tuff Member) also indicate that these cooling units cannot be Hancock. Where the Hancock is expected to lie stratigraphically over the Stairstep Mountain and beneath correlatives of the Egan Tuff Member of the Shingle Pass, Ekren et al. (2011) mapped a multiple-flow ignimbrite that they designate as a middle unit of the tuff of Goblin Knobs. The thickness of this middle unit appears to exceed 2000 m, which, together with included lenses of Monotony and clasts of dacite lava as much as tens of meters in diameter, makes accumulation during caldera collapse likely. However, our analyses of the Goblin Knobs yield a weighted mean age of 25.71 ± 0.05 Ma which is distinctly younger than any Shingle Pass unit (Table 1). Whether an older cooling unit of the Goblin Knobs indeed exists between members of the Shingle Pass Formation or whether the middle unit of the tuff of Goblin Knobs of Ekren et al. (2011) might be an intracaldera correlative of the Hancock Tuff Member found in adjacent, extra-caldera areas can only be resolved by further study. The possibility that the source of the Hancock is concealed in the adjacent valley to the east near site HI cannot be ruled out (Fig. 46B).

Ekren et al. (2011, p. 21–22) believed three map units exposed in the eastern Quinn Canyon Range (Fig. 46D) correlate with the upper Shingle Pass unit at the Shingle Pass type section (Page and Dixon, 1994). One chemical analysis of a sample whose mode is a “clone” of the upper unit matches our analyses for the Egan Tuff Member found in outflow sections (Fig. 45). Modes of at least nine samples of these three units match our modes for the Egan Tuff Member (Fig. 50D); especially characteristic is the dominance of plagioclase over sanidine and the absence, or very small amounts, of quartz. Their tuff of Yurt Buttes and McCutchen Spring may include as many as five or six separate cooling units, nearly all of which are thin, mostly less than 30 m. Overlying these multiple cooling units is their tuff of Cherry Creek, which directly underlies the tuff of Sawmill Canyon, is at least 250 m thick, and contains blocks up to 10 m across of Monotony Tuff. Their third map unit, which they designate the flow-layered upper Shingle Pass tuff, consists of two cooling units at least 1200 m thick that contain megabreccia similar to that in the tuff of Cherry Creek. There can be little doubt that these three map units constitute firm evidence for remnants of a caldera source—shown as the Cherry Creek caldron segment by Ekren et al. (2011)—for the Egan Tuff Member of the Shingle Pass Formation.

Ekren et al. (2011) were unaware of the Tikaboo Tuff Member that we have found in outflow sections north, west, and south of the Quinn Canyon Range (Fig. 46C). The modal composition of the Tikaboo is indistinguishable from the Egan Tuff Member but the Tikaboo is distinguishable in outflow areas on the basis of contrasts in paleomagnetic direction (Fig. 48), age (Table 1), and distinct chemical composition (Fig. 45). Three chemical analyses of ignimbrites in the Quinn Canyon Range listed in Ekren et al. (2011) are consistent with the Tikaboo. One analysis is of a multiple-flow, compound cooling unit of the tuff of Yurt Buttes and McCutchen Spring that contains rounded granite clasts of unspecified origin as much as 1 m in diameter. Thus, evidence is tantalizing but inconclusive that an intracaldera correlative of the Tikaboo Tuff Member lies in the Quinn Canyon Range caldera.

Volumes of the individual members in Table 2 are rough estimates because of the uncertainties in the entire distributions of the outflow sheets as well as the complexities of the composite Quinn Canyon Range caldera and its approximate perimeter, which we assume to be the same for all members. Model 2 is used for the Coyote Summit and Egan Tuff Members based on the thickness of intracaldera correlatives. For the Hancock and Tikaboo, either no intracaldera ignimbrite is known or the intracaldera correlation is uncertain, so the calculated volume uses Model 1; obviously, these are only very rough estimates.

Compositional data on phenocrysts provide information regarding intensive variables in the magmas that erupted to form the Coyote Summit and Egan Tuff Members of the Shingle Pass Formation (Nielsen, 1992; corresponding data on the Hancock and Tikaboo Tuff Members are lacking). Average feldspar, pyroxene, biotite, hornblende-plagioclase, and Fe-Ti oxide temperatures in the lower unit range from 750 to 770 °C but temperatures in the upper unit are not as well constrained and only hint at a somewhat lesser temperature. Hornblende and feldspar barometers suggest a higher pressure for the lower than the upper, perhaps ∼3 and 2 kb, respectively. Biotite-magnetite-sanidine equilibria for the lower unit indicate a water fugacity of ∼0.35 kb using a 3 kb pressure and 770 °C; this indicates definite water undersaturation in the magma. For the upper magma, water near-saturation may have prevailed but the constraints are weak.

Variations in water fugacity can explain the contrasting mineral proportions in the different magmas. The lower with its low calculated water fugacity contains Fe-rich olivine, abundant sanidine, but little biotite. At higher water fugacity, olivine and sanidine react with melt to make biotite, yielding a mineral assemblage more like the upper which lacks olivine and has more biotite and less sanidine.

The resemblance between rhyolitic Shingle Pass and Isom-type trachydacitic ignimbrites is especially close for the Coyote Summit Tuff Member; its unusual occurrence of Fe-rich clino- and ortho-pyroxene and little biotite resembles Isom-type tuffs. Similarly, the Tikaboo unit also shares some chemical kinship with Isom-type tuffs. Because of lower incompatible-element abundances, the rhyolitic Shingle Pass magmas could not have formed by fractional crystallization of trachydacitic Isom-type magmas. Nonetheless, the Shingle Pass magmas might have incorporated similar source components as the Isom-type magmas. Alternatively, Shingle Pass magmas might have mixed with Isom-type magmas evolving in contemporaneous magma systems.

In spite of the chemical and petrographic differences among the Coyote Summit, Tikaboo, and Egan Tuff Members of the Shingle Pass Formation, all share the same initial ⁸⁷Sr/⁸⁶Sr ratios when rounded to four decimal places (0.7086; Table 4 and Fig. 7F). If considered significant at five decimal places, the Egan, with its hydrous phases and higher oxygen fugacity, has a slightly higher Sr isotope ratio than the less hydrous and reduced Coyote Summit (0.70864 versus 0.70859, respectively). This contrast is consistent with relatively more and less, respectively, interaction between their associated magmas and felsic crust. The isotopic composition of the Shingle Pass tuffs also marks a reversal of the general trend toward lower ratios with time as the ⁸⁷Sr/⁸⁶Sr ratios of the Shingle Pass Formation are higher than that of the slightly older (by 0.15 m.y.) rhyolitic tuff of Orange Lichen Creek which erupted on the western side of the field.

Each of the four cooling units of the Single Pass Formation is normally zoned with respect to many elements (e.g., Fig. 52). Variation patterns in each can generally be explained by fractionation of their observed phenocrysts. Rb-Sr relations show that none of the Shingle Pass units can be related to the others by closed-system fractional crystallization, despite the short time periods that separate the eruptions and their apparent common source area. The magma chambers feeding the ash-flow eruptions apparently evolved independently of one another.

The Hancock magma appears to have been generated in a manner more like high-silica, main-trend rhyolite magmas in the Central Nevada field, which it more closely resembles chemically and modally. Nonetheless, the Hancock shares the same high Ba/TiO₂ ratio of other members of the Shingle Pass Formation (Fig. 7D). Although the Hancock is generally more depleted in compatible elements than other members, it does not contain higher concentrations of several key incompatible trace elements like Rb, Nb, and Pb.

This rhyolite ignimbrite (Fig. 53) was first recognized by Ekren et al. (1973b) in the northern Reveille Range (Fig. 54) where the relatively massive ash-flow tuff has weathered into pinnacles and knobs. The weighted mean age of six analyses of the tuff of Goblin Knobs is 25.71 ± 0.05 Ma. It resembles the mostly dacitic Monotony Tuff but some contrasts were seen by Ekren et al. (1973b). Generally, the Goblin Knobs has more sanidine and less plagioclase, biotite, and hornblende, and no pyroxene (compare Figs. 35A and 55). Modally and chemically, the Goblin Knobs is indistinguishable from the rhyolitic parts of the Windous Butte Formation; however, unlike that much older tuff, the Goblin Knobs contains abundant clasts of pumice—generally 3–8 cm, locally 18 cm long—and of a variety of volcanic and sedimentary rock—generally 0.5–6 cm, rarely to as much as 60 cm (Martin and Naumann, 1995). Thin lenses of fine-grained bedded tuff occur about 4 km southwest of sites GC and GD within the thick ignimbrite (Fig. 54), indicating at least two episodes of ash-flow emplacement.

The considerable thickness (>1500 m) of the ignimbrite in the Reveille Range together with the locally large size and abundance of lithic and pumice clasts led Ekren et al. (1973b; see also Martin and Naumann, 1995) to suggest it accumulated within a caldera, the Goblin Knobs caldera. A mélange of chaotic and exotic blocks of tuff and lava lies within the tuff of Goblin Knobs in its most northerly exposures. Although this mélange was interpreted by Ekren et al. (1973b, their map unit Tde) to be related to “intense strike-slip faulting,” it appears to us that at least some of the mélange, which is mostly composed of older Monotony and Shingle Pass ignimbrites, could be wall-collapse debris shed off a nearby caldera wall. Southward into the Reveille Range the caldera perimeter shown in Figure 54 appears to be its structural margin; scabs of landslipped(?) mélange lie on pre-caldera rocks as much as 3 km west of this ring fault.

Additional tuff of Goblin Knobs has been confirmed by ⁴⁰Ar/³⁹Ar and paleomagnetic data at sites GA and GB across Railroad Valley to the east in a north-trending horst just west of the main Quinn Canyon Range. About 4 km to the north of these sites, the compound cooling unit is more than 1000 m thick and still farther north in the horst, a thickness of 600 m is shown in Ekren et al. (2011, their cross section AA′). We have opted to include this compound cooling unit within the eastern segment of the Goblin Knobs caldera (Fig. 54), but admit the possibility that it could be caldera filling in the older Quinn Canyon Range caldera.

We base the volume estimate for the Goblin Knobs ignimbrite on the area of its caldera multiplied by the average intracaldera thickness of 990 m, yielding ∼560 km³. The small contribution of a possible outflow in the central Reveille Range that is 120 m thick is ignored.

The tuff of Lunar Cuesta has been described by Snyder et al. (1972) and Ekren et al. (1974b), and the tuff of Big Ten Peak by Keith (1986, 1987, 1993). Weighted mean ages of these two rhyolitic ignimbrites that we sampled in the areas studied by these workers are analytically indistinguishable at 25.67 ± 0.05 Ma (n = 2) and 25.72 ± 0.03 Ma (n = 4), respectively. Moreover, their paleomagnetic directions are similar and unique among all the tuffs we have sampled in the middle Cenozoic Great Basin, viz., perpendicular to the geocentric axial dipole field for Oligocene time. Therefore, the two deposits are contemporaneous within available limits of resolution. Their overall weighted mean age, including analyses of additional samples designated in our present study as belonging to this unit, is 25.70 ± 0.04 Ma (n = 9).

Although the tuff of Goblin Knobs has an age of 25.71 ± 0.05 Ma indistinguishable from that of these two deposits, the Goblin Knobs is not a correlative because it has a distinctly different paleomagnetic direction. This appears to be another instance of eruptions from separate magma chambers at nearly the same time.

Sampling through stratigraphic sections of the tuff of Lunar Cuesta reveals a reverse compositional zonation from high-silica dacite upwards into low-silica rhyolite that extends systematically on much the same trend into the more evolved high-silica rhyolite tuff of Big Ten Peak (Fig. 56). All these data, together with apparent proximate sources of both units in the southern Monitor Range, are consistent with a single, co-magmatic, reversely zoned deposit.

The Lunar Cuesta name was first used by Snyder et al. (1972) in the Pancake Range (site LB in Fig. 57). It has priority over the Big Ten Peak name first used by Keith (1986). Accordingly, we here apply the name tuff of Lunar Cuesta for the entire reversely zoned deposit that was formerly designated as tuff of Lunar Cuesta and tuff of Big Ten Peak. In the following paragraphs (see also Table 1), we use the appellation low-silica rhyolite facies (<75.0 wt% SiO₂) for ignimbrites previously designated as tuff of Lunar Cuesta and high-silica rhyolite facies (>75.0 wt % SiO₂) for ignimbrites previously designated as tuff of Big Ten Peak. In composition diagrams, the former is shown by letter symbols such as L, LK, LCm, etc. and the latter by B, BA, BH, etc.

The basal tuff (sample qrc) on the east side of Georges Canyon Rim, shown by Keith (1987) as tuff of Big Ten Peak, appears to be continuous in outcrop to his intracaldera Big Ten Peak at our BD site (Fig. 57). However, our age of this basal tuff, 27.24 ± 0.06 Ma, is much too old for Big Ten Peak; chronologically it is close to the age of the upper tuff of Ryecroft Canyon (27.14 ± 0.06 Ma) or the tuff of Orange Lichen Creek (27.13 ± 0.04 Ma). To add to the stratigraphic confusion: (1) Keith (1987) indicated the tuff of Big Ten Peak underlies the tuff of Orange Lichen Creek but this is inconsistent with our chronologic data (Table 1); (2) Larry Garside (2002, personal commun.) cited a K-Ar age on an ignimbrite overlying the tuff of Big Ten Peak southeast of Big Ten Peak that is very similar to the age of the tuffs of Orange Lichen Creek and upper Ryecroft Canyon; and (3) Shawe (1998) indicated the Big Ten Peak caldera formed more than 27 m.y. ago. Thus, previous work had indicated an older age for the tuff of Big Ten Peak than what our analyses disclose, emphasizing the need for detailed mapping of the ignimbrites in the southern Monitor Range supplemented by extensive sampling and laboratory analyses.

According to Bonham and Garside (1979), an outflow equivalent of the intracaldera tuff of Big Ten Peak is the rhyolitic tuff of Rye Patch exposed on the western flank of the Monitor Range. The Rye Patch is ∼80 m thick and consists of a lower tuff cooling unit that contains blocks as much as 15 m in diameter of densely welded tuff and Paleozoic quartzite; this unit is separated from an overlying simple cooling unit of lithic-free tuff (sample BH) by several thin lenses of tuffaceous sediments. The contrast in amount of Paleozoic fragments in these two units corresponds to the upward diminishing amount of such fragments in the intracaldera Big Ten Peak tuffs described by Keith (1987).

Stratigraphic position together with our paleomagnetic, chronologic, and compositional data indicate that an additional high-silica rhyolite facies of the tuff of Lunar Cuesta occurs in the Belted, Groom, and Reveille Ranges at sites BA, BE, BF, and bg (Fig. 57). These sites are as far as 130 km to the southeast of the Big Ten Peak caldera source. Site BA has the appropriate paleomagnetic direction and a composition similar to sample BD collected in the high-silica rhyolite facies in the Monitor Range. Site BE has a stratigraphic position and composition appropriate to the high-silica rhyolite facies. At site bg in the central Reveille Range, the tuff of Reveille Range (Ekren et al., 1973b) that overlies the tuff of Goblin Knobs is a possible correlative.

Site BF is in a lower cooling unit at White Blotch Springs (Ekren et al., 1971). It has the composition of the high-silica rhyolite facies as well as appropriate age and paleomagnetic direction for the Lunar Cuesta. Overlying this cooling unit is 60 m of Pahranagat tuff (see below) that was emplaced at 22.93 Ma. The two cooling units at site BF compose the tuff of White Blotch Spring of Ekren et al. (1971), which is widespread in the southernmost part of the Central Nevada field on the Nevada Test and Training Range. They write (p. 33; see also Ekren et al., 2011) that “...marked differences in nonbasal accumulations of lithic fragments indicate that the [White Blotch Spring] unit as mapped contains ash flows from different centers. Each center seemingly stamped its identity on its ash flows by means of lithic fragments that are representative of the crust through which the magma was erupted.” Eighty kilometers to the west of site BF, in the northern Cactus Range which lies due south of the Monitor Range, the lower tuff of White Blotch Spring is hundreds of meters thick and is exceptionally rich in pumice and lithic clasts (Ekren et al., 1971). The pumice clasts are as much as 30 cm in diameter and occur in a horizon ∼200 m thick. The lithic clasts that locally make up one-half of the tuff are commonly a couple of meters in diameter and consist of older tuffs, granite, and Paleozoic sedimentary rocks. This Cactus Range unit resembles the high-silica rhyolite facies within the Big Ten Peak caldera, but it lies some 60 km farther south of the southern margin of this source caldera (Fig. 57). It could be an exceptionally thick outflow, but the size of the lithic clasts makes an outflow origin unlikely, so we have not included it as part of the tuff of Lunar Cuesta. It may be an intracaldera unit that we have not elsewhere recognized but with the same characteristic lithic clasts as the high-silica rhyolite facies. Bart Ekren (1996, personal commun.) suggested that a similar tuff of about the same age as the Big Ten Peak had a source in the Cactus Range area. He also suggested that the Vindicator Rhyolite of Ransome (1909) that is widespread in the Goldfield district to the west may be this look-alike tuff. We could not pursue these suggestions on the inaccessible Nevada Test and Training Range.

Ignimbrites of the low-silica rhyolite facies in the Pancake and Hot Creek Ranges contain conspicuous lithic inclusions of the Shingle Pass and other volcanic rocks as well as sparse Paleozoic(?) sedimentary rock, amounting to as much as 10% of the basal part of the simple cooling unit. These ignimbrites contain a high proportion of plagioclase phenocrysts together with, in essentially decreasing abundance, quartz, sanidine, biotite, hornblende, and Fe-Ti oxides (Fig. 58A). Sanidines are conspicuously zoned, which is unusual for the more than 100 middle Cenozoic Great Basin ignimbrites we have examined. Broken phenocrysts (or phenoclasts; Best and Christiansen, 1997) of plagioclase are prominent together with more euhedral grains that have strong oscillatory zoning but, overall, are normally zoned from rims as low as An₂₅ to cores as much as An₈₃ (Phillips, 1989). The relatively small labradorite-bytownite cores, many of which are highly corroded, are subhedral and in fairly sharp contact with the oscillatory zoned outer parts; this texture suggests the cores had an exotic source, possibly basaltic, and experienced subsequent overgrowth of more sodic composition after partial resorption. In support of this hypothesis of hybridization with a basaltic component is the fact that, compared to other ignimbrites in the Central Nevada field, the low-silica rhyolite facies contains unusually high concentrations of Sr, Zr, and Ba but low concentrations of K (and Rb, not shown in Figs. 6 and 7).

Phillips (1989) documented reverse zoning in the simple cooling unit of the low-silica rhyolite facies at the Palisade Mesa and The Wall stratigraphic sections (Fig. 56). In these sections, there are upward decreases in TiO₂ and Fe₂O₃ but an increase in SiO₂. These variation trends are directed toward, but are separated from, the clusters of points for the high-silica rhyolite facies. In the low-silica rhyolite facies section at The Wall, the most evolved uppermost ignimbrite, which contains 74.4 wt% SiO₂ and 0.25 wt% TiO₂, has proportionately less plagioclase, biotite, and hornblende but more sanidine and quartz than the lowermost ignimbrite (Fig. 58B). This most evolved ignimbrite near the top of the reversely zoned section is modally similar to the high-silica rhyolite facies (Fig. 58C). Modal compositions of the two facies partially overlap.

In the southeastern Toquima Range (Fig. 57), Shawe and Byers (1999) described two simple cooling unit members of a crystal-rich welded ash-flow tuff, their Tos unit. No chronologic or paleomagnetic data are available but it overlies the 27.14 Ma upper tuff of Ryecroft Canyon. The upper member of unit Tos consists of two chemically and modally distinct parts, the upper part of which is represented by our BJ sample and the lower part of which is represented by our two LK samples (Figs. 56 and 57). The lower member appears to be more modally uniform but no chemical analyses are available. All together, these two members are remarkable in that they are reversely zoned, are compositionally much like our newly designated tuff of Lunar Cuesta, and bridge the compositional gap between our analyses of the low- and high-silica rhyolite facies. Sedimentary rock fragments are locally present but are smaller (<3 cm) and less abundant compared to ignimbrites in many exposures in the thick intracaldera pile of high-silica rhyolite facies in the southern Monitor Range. This thick ignimbrite pile contains the most clasts of exotic sedimentary and granitic rock of any ignimbrite in the Central Nevada field (Keith, 1986, 1987; Whitebread and John, 1992).

Stewart and Carlson (1976b, their sheet 2) showed a “probable caldera” ∼35 km in diameter in the southern Monitor Range; Ekren et al. (1976, their plate 1) labeled it as a “probable resurgent cauldron complex.” Sargent and Roggensack (1984; see also Bonham and Noble, 1982) referred to it as the Big Ten Peak caldera, as did Bonham and Garside (1979, p. 40), who indicated it to be the source of their outflow tuff of Rye Patch that crops out to the west. The latter workers noted that the intracaldera high-silica rhyolite ignimbrites in the Big Ten Peak caldera are at least 500 m thick, probably substantially greater, and contain embedded blocks of limestone and snow-white quartzite, several of which are over 0.5 km long to as much as 2 km. These “rafts floating” in ignimbrite are large enough to appear on the 1:250,000-scale map of Kleinhampl and Ziony (1985). Subsequent mapping by Keith (1986, 1987, 1993) disclosed a broad arcuate zone of xenolith-rich (to as much as 50% by volume) tuff and megabreccia in both of his Big Ten Peak tuff units on the western and southern flanks of the Big Ten Peak massif; this zone he believed marks the approximate boundary of the caldera. To the north, Shawe and Byers (1999) indicated a segment of the structural margin of the caldera barely crossing the southeast corner of the Belmont East quadrangle.

South of the Big Ten Peak caldera, at site LC (Fig. 57), a compound cooling unit of the low-silica rhyolite facies has the age and unique paleomagnetic direction of the Lunar Cuesta. Sample LCv of vitrophyre was collected at the base of the unit above altered lavas(?), LCm a few tens of meters above that, and LCu near the top of the exposed 240-m-thick section; no younger overlying rock is exposed. As much as one-half of the unit is made up of fragments of exotic volcanic rock to as much as 12 cm in diameter. Exposures continue for a few kilometers to the north along the east flank of the Monitor Range. We interpret these exposures to be a segment of the caldera source of the low-silica rhyolite facies of the tuff of Lunar Cuesta; its margin cannot be delineated without detailed mapping. Ekren et al. (1976, their plate 1) showed a caldera encompassing this segment but provide no details as to what ignimbrite is associated with it.

Ekren et al. (1974b, 1976) believed the source of the tuff of Lunar Cuesta to be a hypothesized Lunar Lake caldera ∼60 km to the northeast of our compound Big Ten Peak caldera (Fig. 57). The existence of this caldera was based partly on an arcuate topographic ridge—The Wall (site LD)—along its eastern margin. Quinlivan et al. (1974) also mapped a large (several cubic kilometers) rhyolite lava dome that they believed to have erupted along the buried ring fracture of the Lunar Lake caldera. An obsidian nodule from near the base of the lava dome yielded a K-Ar age of 25.8 ± 1.3 Ma, which is indistinguishable from the age of the tuff of Lunar Cuesta. Ekren et al. (1974b, p. 607) noted “...that the tuff is no thicker in the caldera [61–105 m] than outside; conceivably it is thinner. Eruption of the tuff of Lunar Cuesta, therefore, was apparently completed before caldera subsidence began.” We note that in and around the postulated Lunar Lake caldera, the tuff of Lunar Cuesta is a simple cooling unit that has fewer and smaller lithic inclusions than at our proposed intracaldera LC site 60 km to the southwest. The Wall could represent an elevated block bounded by arcuate faults of basin-and-range vintage, possibly influenced by the buried structural margin of the older Williams Ridge caldera or Monotony Valley caldera along which a lava dome was subsequently formed, rather than representing the preserved topographic rim of a 25.70 Ma caldera.

The eruptive pattern of the outflow sheet of the Lunar Cuesta is unusual in three respects (Fig. 57):

1. The source caldera lies in a highly eccentric position within the sheet. Dispersal of the ash flows to the west was apparently blocked by a topographic high immediately to the west on the lip of the Great Basin altiplano (Figs. 1–3; Best et al., 2009; see also the section Topographic Barrier).

2. The low-silica rhyolite facies forms an outflow lobe to the east beyond the intracaldera high-silica rhyolite facies in the Big Ten Peak caldera, whereas a high-silica rhyolite facies outflow lobe extends to the southeast beyond the apparent source of the low-silica rhyolite facies. Conceivably, ash flows creating these two main lobes of the outflow could have been constrained by very broad paleovalleys tens of kilometers wide cut into older ignimbrites or separated by a paleoridge. Whether outflow occurs to the south of the source caldera could not be evaluated because of widespread alteration of the volcanic rocks and lack of access to the Nevada Test and Training Range.

3. The outflow deposit is reversely zoned. If the magma chamber had been zoned in a normal manner with the more evolved magma at the top, then the eruption somehow must have been initiated in the lower less evolved part of the chamber and subsequent withdrawal progressed upward. Input and mixing of more mafic magma into the lower part of the chamber, as suggested above, could have triggered the eruption as exsolving volatiles from the cooling mafic magma created sufficient buoyancy for it to rise (e.g., Sparks et al., 1977). But rather than rising through the overlying more evolved magma, it may have ascended and erupted on the north and east of the chamber. The more evolved high-silica rhyolite erupted from new vents that tapped the upper part of the chamber (Fig. 59) to create ash flows that traveled mainly toward the southeast.

Using an intracaldera thickness of 1000 m in Model 2 yields a total volume for the tuff of Lunar Cuesta of ∼1100 km³.

The relatively thin tuff of Clipper Gap emplaced at 24.95 ± 0.07 Ma was first recognized and designated as the youngest of five rhyolite cooling units in the Bates Mountain Tuff of Sargent and McKee (1969). Subsequently, Gromme et al. (1972) found this unit to have a different distribution from the four underlying Bates Mountain cooling units, including the immediately underlying unit D, which has since been redesignated as the Nine Hill Tuff (Deino, 1985; Henry and John, 2013). They consequently designated the youngest of the five Bates Mountain units as the tuff of Clipper Gap. An excellent and relatively thick exposure of more than 30 m is at the mouth of Clipper Gap Canyon at the northwest end of the Toquima Range (site CH, Fig. 60) where the ignimbrite overlies the Nine Hill and Bates Mountain B and C units (McKee, 1976).

Most known exposures of the Clipper Gap are in the northern Toquima and Monitor Ranges. An outcrop (site CC) only ∼10 km from the Utah state line is ∼5 m thick with an eroded top; the narrow outcrop band eastward toward Utah suggests confinement to a paleovalley, but this may only be an artifact of the lack of information on the distribution of this young ignimbrite to the north and south. Among the numerous ignimbrites sampled for paleomagnetic direction in the Great Basin, the Clipper Gap is unique in having a near zero inclination; its declination is south-southeast (Gromme et al.,1972).

The wholly devitrified Clipper Gap ignimbrite is mostly densely welded with flattened lighter-colored pumice lapilli constituting no more than ∼10% of the rock, but grades to nonwelded at the base. Densities of paleomagnetic-sample cores decrease eastward in the area of its distribution. Shades of gray, brown, and red are typical in the commonly ledge-forming exposures. Lithic clasts are inconspicuous. Phenocrysts are small, mostly less than 1 mm, and compose only 4%–7% of the tuff (Fig. 61A). Rare, tiny, “rusty” pseudomorphs could be after Fe-rich pyroxene or olivine. Alkali feldspars are complexly zoned and include sodic sanidine, anorthoclase, and andesine-oligoclase.

In addition to its unusual mineralogical composition, limited chemical analyses confirm the off-trend character of the Clipper Gap, contrasting it with more common main-trend rhyolite ignimbrites in the Central Nevada field (Figs. 5, 6, and 7E). It is a relatively uniform high-silica, high-alkali rhyolite that contains 75.6–76.8 wt% SiO₂ and 8.4–8.9 wt% Na₂O + K₂O (Tables 3 and 7). The Clipper Gap contains more Nb than any other Central Nevada ash-flow tuff. Chemically, mineralogically, and petrographically, it resembles the somewhat older 25.48 Ma Nine Hill Tuff mentioned above (Deino, 1985) that crops out chiefly to the north and west of the Central Nevada field. The Clipper Gap and the Nine Hill are the only ignimbrites we are aware of in the middle Cenozoic of the Great Basin that contain three feldspars. The Clipper Gap is not quite as extreme as the Nine Hill in its concentration of incompatible elements and plots in the field of within-plate or anorogenic granites on a Rb versus Y + Nb diagram (Fig. 44C) (Pearce et al., 1984). It has a high Ga/Al ratio like within-plate granites (Whalen et al., 1987) and is the only Central Nevada tuff to do so consistently. The Clipper Gap also has unusually high concentrations of Zn, Y, and Zr and low MgO, CaO, and Sr (Figs. 6 and 7); thus it shares many of the chemical features of the Isom-type tuffs and most tuffs of the Shingle Pass Formation, but these tuffs plot cleanly in the field of arc rocks in the Rb-(Y + Nb) diagram. Despite its within-plate signature on this diagram, the Clipper Gap has a typical “spikey” signature in a spidergram (Fig. 62). It has a small negative Nb anomaly and a positive Pb anomaly that are typical of subduction-related arc rocks. We believe that these apparently conflicting chemical attributes indicate that the Clipper Gap magma was a highly differentiated arc-type magma.

About 40 km west of the Toquima Range in the Shoshone Range is a thick deposit of the Underdown Tuff (Bonham, 1970) that appears to be a correlative of the Clipper Gap. The age of the Underdown (24.96 ± 0.05 Ma) is indistinguishable from the Clipper Gap and has a similar mode (Fig. 61A). Chemical data provide additional strong support for a correlation (Table 7; Figs. 7E, 44C, and 61B). Element concentrations are the same or close; similarly high Y and Nb are distinctive, at 30–45 ppm and 26–32 ppm, respectively. The Ca/K value in distinctive sodic sanidines of both units are similar and higher than for sanidines in any other ignimbrite in the Great Basin, except for the Nine Hill and the 23.04 Ma Bauers in the southern Indian Peak–Caliente field (Best et al., 2013b). Preliminary microprobe analyses failed to reveal anorthoclase in our single sample of the Underdown.

The thick deposit of the Underdown in the Shoshone Range marks the Underdown caldera. The base of the ignimbrite is not exposed, but it is at least 490 m thick, comprising two lithologically similar, variably lithic, cooling units separated by a meter or so of sandstone and conglomerate, according to Bonham (1970). Columnar joints are well developed in the lower, densely welded exposures that display prominent laminar flowage structures, including folded and stretched and lineated pumice clasts to as much as several tens of centimeters long. Field relations indicated to Bonham (1970) that the thick ignimbrite “was deposited against a ridge” of Permian rocks. Although he didn’t conceive of a caldera margin, this interpretation is nonetheless consistent with it. The Underdown is overlain above an erosional contact by as much as 500 m of a variety of epi- and volcaniclastic deposits and these in turn are overlain by as much as 460 m of a massive phenocryst-rich tuff designated the Toiyabe Quartz Latite by Bonham (1970). This tuff resembles the 23.31 Ma tuff of Toiyabe, an ignimbrite derived from a caldera ∼25 km to the south (Henry and John, 2013), but it has opposite paleomagnetic polarity (S. Gromme and M.R. Hudson, 2013, personal commun.). This 960-m-thick deposit can reasonably be interpreted as post-collapse filling of the Underdown caldera after accumulation of the intracaldera Underdown ignimbrite.

To the west of the Underdown caldera, no correlative outflow Underdown Tuff occurs in the southwestern Clan Alpine Mountains at ∼39°30′ N, 118°0′ W (John, 1997) nor in the Fairview Peak area 25 km to the southwest (Henry, 1996). However, the tuff of Brunton Pass and the “red tuff” in the Paradise Range might be correlatives, based on similar modal and chemical composition, including relatively high Y and Nb (John, 1992). Dilles and Gans (1995) tentatively correlated an ignimbrite southeast of Yerington (Fig. 2) with the “red tuff.”

Although the Underdown caldera source of the tuff of Clipper Gap lies in the Western Nevada field, the ignimbrite has been described in this article because of the occurrence of the outflow tuff in stratigraphic sequences with Central Nevada ignimbrites, including the Pancake Summit, Windous Butte, and Isom-type tuffs.

The total volume for the combined known outflow tuff of Clipper Gap and intracaldera Underdown Tuff calculated according to Model 2 is at least 180 km³.

The compositionally heterogeneous tuff of Buckskin Point and tuff of Buckwheat Rim are the youngest dacite ignimbrites (Figs. 63 and 64) at 24.75 ± 0.02 Ma and 24.69 ± 0.10 Ma, respectively, in the Central Nevada field. They crop out in the southern Pancake Range (Fig. 54; Supplemental File 4 [see ^{footnote 4}]; Snyder et al., 1972) where their combined volume is probably no more than 50 km³. A possible distal correlative of the younger Buckwheat Rim was found by Ekren et al. (2011, p. 16, their unit Tlc) in the Quinn Canyon Range 50 km to the southeast at site je; it is 60 m thick, overlies the Shingle Pass Formation, and is overlain by an Isom-type tuff ∼30 m thick. If this unit is indeed the Buckwheat Rim, then the volume would be more than 50 km³.

In the southern Pancake Range, the Buckwheat Rim and Buckskin Point lie stratigraphically above and below, respectively, a local pile of andesitic and dacitic lava and debris flows that is as thick as 600 m. Directly overlying the younger tuff of Buckwheat Rim are two cooling units totaling ∼30 m thick of trachydacite Isom-type tuffs designated the tuff of Black Beauty Mesa by Snyder et al. (1972). Overlying these two tuffs is the 22.93 Ma Pahranagat rhyolite ignimbrite. Because this entire sequence of intermediate-composition lavas and dacite and trachydacite tuffs was vented in a restricted area and time from 24.75 to earlier than 22.93 Ma, we believe them to be a co-magmatic assemblage that represents the products of a small evolving magma system near the ring fault of the Monotony Valley caldron that collapsed ∼2.8 m.y. earlier. No caldera apparently resulted from this small volume of erupted magma.

The older Buckskin Point is a compound cooling unit. Resembling Isom-type tuffs, the lower to middle part of the Buckskin Point has relatively few phenocrysts of which plagioclase and two pyroxenes are prominent, whereas the upper part has more phenocrysts—including quartz, biotite, and hornblende—and less plagioclase and pyroxenes like typical dacites of the Central Nevada field (Fig. 63A). The occurrence of black and light-pink pumice clasts in the middle, more anhydrous part of the unit that has a greater concentration of alkalies and incompatible elements attests to the heterogeneity of the apparently small magma chamber from which the Buckskin Point erupted.

The two simple cooling units that compose the younger tuff of Buckwheat Rim are similarly heterogeneous in composition but neither contains much pyroxene (Fig. 63B). The lower unit, like the lower part of the Buckskin Point unit, contains abundant sanidine, to as much as 43% of the phenocrysts, and a high concentration of incompatible elements and alkalies; its composition plots in the trachydacite field on a total alkalies–silica diagram (Fig. 64). The upper unit again has more phenocrysts and proportionately more biotite and hornblende. Like other phenocryst-rich, biotite- and hornblende-bearing dacites as old as ca. 32 Ma, the Buckwheat Rim initial ⁸⁷Sr/⁸⁶Sr ratio is 0.7082, which is significantly higher than that of the younger Isom-type Black Beauty Mesa of 0.7062 (Fig. 7F). Overall, the Buckwheat Rim is chemically less evolved than the Buckskin Point emplaced 0.06 m.y. earlier. If the two ignimbrites erupted from the same evolving magma system, there must have been a replenishment of more primitive mafic magma between them; some of this postulated replenishment could be manifest in the 3 km³ of extruded intermediate-composition lavas.

The evolution of the small magma system in the southern Pancake Range serves as a model for larger-scale systems explosively erupting dacite and Isom-type trachydacite magmas earlier throughout the southern Great Basin province, together with lesser extruded intermediate-composition lavas.

This high-silica rhyolite ignimbrite (Fig. 65) was first recognized by Ekren et al. (1973b) in the northern Reveille Range (Fig. 54) where it consists of three cooling units. Included slivers of other volcanic units and abundant small clasts of dacitic lava and Shingle Pass and Monotony tuffs characterize the tuff. Martin and Naumann (1995) found, in addition, blocks of granite and tuff of Goblin Knobs to as much as 1 m in diameter. They concluded that the tuff of Northern Reveille Range, which may locally be more than 300 m thick, is an intracaldera accumulation within its source—the Northern Reveille Range caldera. The extent of this postulated caldera and its relation to the older Goblin Knobs caldera in the range is uncertain but they may be more or less coextensive, at least in the northeastern segment of the Goblin Knobs caldera. Our only two modally and chemically analyzed samples of the tuff of Northern Reveille Range, as so designated by Ekren et al. (1973b), are NA and NC. Paleomagnetic data are not available for either but the age of NA is 23.70 ± 0.01 Ma.

An ignimbrite sample (nb) collected during a brief reconnaissance of the horst west of the main Quinn Canyon Range (Ekren et al., 2011) is chemically and modally very similar to sample NA of the tuff of Northern Reveille Range; both are characterized by intensely embayed quartz phenocrysts. Another sample (nd), collected 4 km to the northwest of sample nb, is identical in almost every respect, including the presence of phenocrysts of sanidine that have very thin, conspicuous rims of greater birefringence; this feature has not been seen in any other middle Cenozoic tuff we have examined in the Great Basin. Ekren et al. (2011, p. 25) noted such unusual zoning in sanidine in some of their tuffs of Mystery Mountain in the Quinn Canyon Range. Both samples nb and nd were collected just below the 22.93 Ma Pahranagat ignimbrite near paleomagnetic site 8P733. The age of sample nd of 23.88 ± 0.05 Ma is different from that of sample NA.

The stratigraphy, geochronology, paleomagnetism, and petrology of the high-silica rhyolite ignimbrite (Fig. 66) constituting the Pahranagat Formation has been described by Best et al. (1995). Only a brief summary of this unit emplaced at 22.93 ± 0.02 Ma is included here. The outflow sheet (Fig. 67) is a normally zoned, simple cooling unit that locally overlies a meter or so of bedded tuff; it has been designated by four different stratigraphic names in different areas. Additional names have been applied to the intracaldera ignimbrite (Gardner et al., 1980).

The Pahranagat ignimbrite contains subequal proportions of quartz, plagioclase, and sanidine phenocrysts (Fig. 68) that are as much as 5 mm across. Quartz phenocrysts are deeply embayed and plagioclases are pervasively corroded; both were mostly blown apart during eruption, forming phenoclasts (Best and Christiansen, 1997). Lower parts of the outflow sheet that are more proximal to the caldera source contain the highest proportions of quartz and the least plagioclase together with less than 4% total mafic phenocrysts, mostly biotite, minor titanomagnetite, and rare ilmenite and amphibole. In contrast, upper and more distal parts of the outflow sheet as well as the intracaldera tuff have lower quartz/plagioclase ratios and slightly more mafic phenocrysts that include more amphibole and sparse clinopyroxene. Bulk-rock chemical variations complement these vertical and lateral modal variations; the more evolved tuff that has lower TiO₂ concentrations occurs in lower and more proximal parts of the outflow sheet.

Four compositional types of pumice fragments hosted in the tuff provide further characterization of the zoned magma chamber from which the ash flows were erupted (Table 8). Clasts of the most evolved pumice occur throughout the outflow sheet whereas progressively less evolved types occur progressively higher and in more distal parts of the sheet. The spatial manner in which these four pumice types occur within their zoned host tuff is consistent with progressive withdrawal of gas-charged magma from a compositionally layered chamber (Best et al., 1995, their figure 16). During the progress of the eruption, physical mixing of pyroclasts from the top two layers, then the top three, and finally all four layers created the vertically and laterally zoned outflow sheet and the intracaldera deposit that closely resembles the least evolved, late outflow.

The uppermost part of the pre-eruption Pahranagat magma chamber was among the most differentiated that erupted in the Central Nevada field. We measured initial ⁸⁷Sr/⁸⁶Sr ratios for four samples of the Pahranagat (Table 4). Three ratios (0.7071–0.7079) follow the overall pattern found in the Great Basin in that ratios for younger ignimbrites are generally lower than for older ones (Fig. 7F). However, the most highly differentiated sample AH has an anomalous ratio of 0.7092, making the Pahranagat more isotopically variable than any other Central Nevada ash-flow tuff.

The Kawich caldera source of the Pahranagat occupies most of the Kawich Range and the southern part of the Reveille Range (Gardner et al., 1980) where it might truncate the southwestern margin of the older Goblin Knobs caldera (Figs. 54 and 67; see also Stewart and Carlson, 1976b; Sargent and Roggensack, 1984). The Kawich caldera appears to be nested within the western part of the 5-m.y.-older Monotony Valley caldera (Fig. 31). Because of the lack of large-scale mapping, the location of the southern margin of the Kawich caldera in the Kawich Range is uncertain, but is drawn conservatively to exclude a paleomagnetic site where the thickness is more than 230 m. The eastern margin is also drawn conservatively on the west side of Railroad Valley. The intracaldera compound cooling unit is locally more than 1000 m thick.

At two sites in the southeastern Quinn Canyon Range (Ekren et al., 2011), the Pahranagat is 350 and more than 225 m thick. Although a compound cooling unit, we conservatively interpret this accumulation to lie within the older Goblin Knobs caldera, rather than being an eastern segment of the Kawich caldera. Tuffaceous sediments are locally hundreds of meters thick beneath the Pahranagat in the Goblin Knobs depression. To the northwest, an entirely fault-bounded outcrop area of the Pahranagat is structurally complex because of reversed dips but the Pahranagat here could be more than 400 m thick.

Because of only small-scale mapping and widespread alteration obscuring the identity of ignimbrite units, almost no information is available regarding the possible occurrence of the Pahranagat on the now inaccessible Nevada Test and Training Range southwest of the source caldera (Ekren et al., 1971). However, paleomagnetic data at two sites constrain the outer limit of the outflow ignimbrite sheet (Fig. 67).

Using an average, and conservative, intracaldera thickness of 600 m in Model 2, the total volume of the Pahranagat ignimbrite is estimated to be ∼2100 km³. This is smaller than our previous estimate of 2600 km³ (Best et al., 1995) because we have used a more conservative perimeter for the Kawich caldera.

The stratigraphic unit Fraction Tuff has had a tangled history and went through several name changes in the Tonopah mining district where it was first recognized. Stratigraphic problems still persist in the district, in large part because of poor exposures and widespread alteration. We have not attempted to resolve the ignimbrite stratigraphy, but did obtain an ⁴⁰Ar/³⁹Ar age on the lower tuff of the King Tonopah Member of the Fraction Tuff (Bonham and Garside, 1979) of 20.02 ± 0.05 Ma (Table 1; sample FE collected out of area shown in Fig. 69).

Ekren et al. (1971) applied the Fraction Tuff name to a compound cooling unit of zoned rhyolitic ignimbrite on the Nevada Test and Training Range ∼100 km southeast of Tonopah. Because of compositional and apparent age differences, Bonham and Garside (1979) concluded that this unit is not equivalent to the Fraction Tuff around Tonopah. Our three analyses of the intracaldera facies of the Fraction Tuff of Ekren et al. (1971) yielded a weighted mean age of 18.57 ± 0.01 Ma.

An additional complication is the occurrence of a mapped “Fraction Tuff” farther south on the Nevada Test and Training Range that has a similar paleomagnetic direction as the Fraction Tuff of Ekren et al. (1971) but is distinctly younger at <15.9 Ma (Hudson et al., 1994; M.R. Hudson, 1995, personal commun.). This younger unit is actually the tuff of Twin Peaks according to David A. Sawyer (2003, personal commun.; not the same tuff of Twin Peaks in the Hot Creek Range cited above).

Rogers et al. (1967), Stewart and Carlson (1976b), and Sargent and Roggensack (1984) all indicated that the Cathedral Ridge caldera is the source of the Fraction Tuff of Ekren et al. (1971) on the Nevada Test and Training Range (Fig. 69). The caldera is recognized on the basis of a compound cooling unit that is 2200 m thick and has multiple vitrophyres and abundant lithic fragments, commonly 25 cm and rarely nearly 1 m in diameter making up as much as 30% of the ignimbrite. Using Model 2 and an assumed uniform intracaldera thickness of 2200 m, the total volume of the ignimbrite is ∼700 km³.

The Fraction Tuff of Ekren et al. (1971) is similar in age and composition, including trace amounts of titanite (Fig. 70), to the 18.51 ± 0.05 Ma Hiko and the 18.57 ± 0.03 Ma upper Racer Canyon Tuffs, which are the youngest ignimbrites in the Indian Peak–Caliente field to the east (Best et al., 2013b; Gromme et al.,1997). This is yet another example of near-simultaneous eruptions, in this case of similar magmas, from independent and well-documented sources in the southern Great Basin ignimbrite province; they are as little as 100 km from one another, after correction for extension.

The eccentric location of source calderas in the western part of the Central Nevada outflow ignimbrite field is one of its most striking features (Figs. 2 and 71; John H. Stewart pointed out this peculiar attribute to the senior author in the 1980s). Whereas some ash flows might have been expected to be dispersed to one side of the source caldera because of eruptive dynamics or the configuration of the vent(s), the fact that all but one of the most extensive, regional ignimbrites do not extend west of ∼117° W longitude implies some sort of topographic control on dispersal of the ash flows. And, as detailed in the  Appendix, only three ash flows of many from sources to the west of 117° W longitude are known to have moved east of this meridian, and these are in the north. These relations indicate that a northerly trending topographic barrier impeded dispersal of middle Cenozoic ash flows westward from eastern sources and vice versa.

The hypothesized topographic barrier approximately coincides with the present-day Toquima Range and with the Toiyabe uplift zone of Speed et al. (1988) that is marked in the southern Toquima Range by once deeply buried, early Paleozoic metasedimentary rocks and Cretaceous granitic intrusions that were exposed at the surface when the ignimbrites were deposited (Shawe, 1998, 1999a, 1999b; Shawe and Byers, 1999; Shawe et al., 2000). The barrier might have been residual topography from a mostly eroded Jurassic contractile belt that sloped gently eastward (Best et al., 2009, their figure 1). The barrier also lies near the western edge of the Precambrian basement (Figs. 1–3 and 71). During the middle Cenozoic ignimbrite flareup, the barrier might not have been a continuous high ridge but rather a broad swell modified by deposition of ignimbrites derived from the east and eroded from the west by headward incision of streams that ultimately fed into the ancestral Pacific Ocean (Cassel et al., 2012). Erosion would have been counterbalanced by uplift in response to denudation of the western slope of the altiplano and consequent isostatic rebound (Best et al., 2009, their figure 15), thus contributing to the topographic lip on the western margin of the Great Basin altiplano (Fig. 3).

In Figures 15 and 71, the distribution of the 35.30 Ma Pancake Summit Tuff is consistent with a barrier to the west of its Allison Creek caldera source; a paleovalley at least 200 m deep apparently headed in this barrier and was filled with the tuff deposited from ash flows from east of this barrier. A barrier might have been present to as far north as 39°30′ N or so at the time of the dispersal of ash flows of the 29.01 Ma tuff of Campbell Creek from a source to the west of the Toquima Range (Henry and John, 2013). But by the time of eruption of the 25.48 Ma Nine Hill ash flows (Henry and John, 2013; Deino, 1985) and the 24.95 Ma Clipper Gap ash flows (Fig. 60), apparently no barrier existed—at least north of ∼39°30′ N—to impede their dispersal almost to Utah from sources to the west (Fig. 2). Their sources at that time may have lain on previously deposited ignimbrites that effectively modified the western slope of the altiplano (Fig. 3) so as to allow eastward dispersal of the ash flows. Alternatively, western ash flows might have swept around the north end of the barrier and flowed to the southeast. South of ∼38° N, a 27.57 Ma Monotony ash flow from its source in the Central Nevada field traveled far to the west into the Candelaria Hills (Fig. 31). Whether this reflected a breach in the barrier created by a valley eroded from the west or the absence of the barrier south of 38° N cannot be answered.

But between ∼38° N and 39°30′ N the existence and nature of a barrier is especially problematic, to say the least. In this interval (see  Appendix), ash flows from caldera sources west of the Toquima Range traveled down the western slope of the Great Basin altiplano (Fig. 3), in part in stream valleys (e.g., Henry and Faulds, 2010; Henry and John, 2013). Ash flows of the 25.18 Ma tuff of Arc Dome from a caldera west of the Toquima Range not only went west but also might have traveled eastward a short distance into the Toquima Range, and possibly farther east, via a drainage that had breached the barrier. Ash flows from sources in the Toquima caldera complex (Fig. 36) were apparently not dispersed far eastward into the Central Nevada field. So far as we are presently aware, of the many Central Nevada ash flows from proximal sources, only the Tikaboo and Lunar Cuesta might have flowed west of the barrier (Figs. 46C and 57).

How high must the hypothetical barrier have been to block dispersal of ash flows from proximal sources? Ash flows that surmounted topographic barriers hundreds of meters in height tens of kilometers from the source have been described from other locales (e.g., Sheridan, 1979; Lipman, 2000b). Fisher and Schmincke (1984) indicated ∼600 m is the upper limit over which ash flows can be dispersed. The mobility implied by these field observations can be compared with model experiments (e.g., Woods et al., 1998; Andrews and Manga, 2011). In view of the still poorly understood dynamic complexities of ash flows and probable differences in their mechanics from one to another depending on eruptive conditions, temperature, ash-flow density, etc., it is difficult to draw firm conclusions regarding how effective a hypothetical barrier might be for dispersal of many ash flows erupted over time on either side from proximal and distal sources.

The focus of this article is on the ignimbrites and their sources in the Central Nevada field of the middle Cenozoic ignimbrite flareup. However, scattered extrusions of contemporaneous lava flows are a part of the history of the field and deserve to be mentioned here. Of particular significance is the relative volume of intermediate-composition lava through time with respect to the volume of silicic ignimbrite. This volume reflects the efficiency of unusually thick crust and the silicic magmas generated in it blocking the ascent of mantle-derived basaltic magmas and derivative magmas to the surface.

Andesitic lavas in the Central Nevada ignimbrite field that were extruded during the flareup are clearly subordinate to silicic ignimbrite, as first documented by Stewart and Carlson (1976b, their sheet 2). Here, we update their 34–17 Ma compilation (Fig. 1), make minor adjustments for the 36–18 Ma ignimbrite flareup, and subdivide andesitic activity during three time divisions of the flareup. Many geologic maps published since 1976 have been used to make this time-divided revision, especially by Crafford (2007), but the overall 34–17 Ma distribution of lavas and ignimbrites shown by Stewart and Carlson (1976b) remains essentially unchanged.

An inventory of andesitic extrusions on published geologic maps indicates an apparent hiatus from 34 to 32 Ma prior to the most voluminous eruptions in the Central Nevada field of the Windous Butte, Monotony, and Shingle Pass ignimbrites. To compare andesitic activity roughly corresponding to the inception, maximum expression, and demise of explosive silicic activity in the southern Great Basin ignimbrite province, we examine andesitic lavas that were extruded in the Central Nevada field in three time periods: 36–34 Ma, 32–25 Ma, and 24–18 Ma. The extrusive activity generally migrated southward in the Great Basin (Fig. 72; see also Best and Christiansen, 1991) but stalled at ∼37°30′ N so no andesitic lavas are found through a roughly 100-km-wide, east-west amagmatic corridor to the south (Stewart and Carlson, 1976b).

In her compilation of 376 samples of 42–17 Ma lavas across the Great Basin that contain <69 wt% silica, Barr (1993; see Best et al. [2009] for a link to the database) found that most are high-K andesite; dacite is subordinate. Basalt in the International Union of Geological Sciences (IUGS) classification (Le Maitre, 1989) appeared only after 20 Ma and until 18 Ma was rare. Because of sparse chemical data, it was impossible to cleanly distinguish the outcrop areas of andesitic (and lesser latitic) rocks having <63 wt % silica from those of more silica-rich dacitic rocks. Consequently, the areas displayed in Figure 72 may be excessive because of the inclusion of dacitic rocks, but, on the other hand, some andesitic rocks may have been excluded in dominantly dacitic occurrences.

Late Eocene (36–34 Ma) lava flows and lesser flow breccias extend in a southeast-to-northwest swath across Nevada mostly north of the Central Nevada caldera complex. Within the Central Nevada field, most intermediate-composition lavas occur in a few thick piles (Hose, 1983; McKee, 1968a) northwest and southeast of the Allison Creek source caldera of the 35.30 Ma Pancake Summit Tuff (Figs. 2, 15, and 72). After correction for post-volcanic extension, these flows occupied an area of 2900 km² and had a volume of ∼650 km³, based on average thicknesses from published maps. Andesitic lavas extruded 32–25 Ma during peak explosive activity appear to be far less widespread (600 km²) than those from the earlier time period and had an estimated volume of only 70 km³, an order of magnitude less. They crop out in a WNW-ESE–trending band ∼50 km wide that spans the central part of the caldera complex. During the waning of the middle Cenozoic ignimbrite flareup in the southern Great Basin, 24–18 Ma, andesitic lavas were extruded in the southern part of the Central Nevada field; they occupied an area of ∼1200 km² and had a volume of ∼250 km³.

To correct these lava volumes for that buried beneath Neogene sediment in fault-bounded valleys, we simply doubled the volume. Another correction might be made for lavas buried within calderas, especially for the 32–25 Ma time period. However, in ranges tilted by subsequent basin-and-range faulting, little lava is apparent; had major edifices existed within the calderas, volcanic debris flows would be exposed in the extracaldera sections, but such deposits are virtually absent.

Doubling the andesitic rock volumes gives ∼2000 km³ for all three time periods, which is 8% of the total volume of contemporary silicic ignimbrite of 25,000 km³ (Table 2). Because of the area over which we have made the lava inventory—within a roughly 50 km distance from the Central Nevada caldera complex (Fig. 72)—some lavas actually lie in the Western Nevada field, but we have not included any volume of ignimbrite in this field in the comparison.

Despite the several uncertainties in our estimate of the volume of andesitic lavas, there can be no doubt of their small volume relative to the colossal volume of silicic ignimbrite. Notwithstanding the enormous input of mantle-derived basalt magma into the crust that must have been required to provide the necessary energy and mass to drive the crustal magma systems generating the silicic ignimbrite magmas during the flareup, virtually no basalt and only a relatively small volume of andesitic magma differentiated from the mass of basalt penetrated all the way through the crust to the surface. And this penetrating volume was apparently the least between 32 and 25 Ma, during which time most of the ignimbrite volume was erupted. While the ignimbrite-forming magma systems were operative in the unusually thick crust (Fig. 3), bodies of silicic magma, or zones of crustal partial melt, effectively blocked the ascent of denser andesitic magmas. The circumscribed area over which volumes were inventoried might correspond to the source region, or “circle of influence,” in the crust where silicic magmas were being generated and from which they ascended into the shallower crust and erupted.

The Central Nevada ignimbrite field occupies the central part of the southern Great Basin ignimbrite province, which in turn is a part of a broad swath of southwestern North America stretching far into Mexico where the middle Cenozoic ignimbrite flareup flooded the landscape with ash to depths as much as a kilometer or so. Calderas in this swath probably number in the hundreds, are tens of kilometers in diameter, and were filled with thicknesses of kilometers of ignimbrite and wall-collapse breccia accompanying subsidence. Our investigations of the Central Nevada ignimbrite field, coupled with like studies of nearby fields resulting from the flareup (Henry and John, 2013; Best et al., 2013b), provide a baseline of information that constrains interpretive models of how caldera-forming super-eruptions develop during ignimbrite flareups and how distinct types of magma originated in this continental arc regime. Compared to more normal arc volcanism on thinner crust, including that of oceanic character, the world-class explosive volcanism of the middle Cenozoic in southwestern North America lies at the opposite end of the arc spectrum.

In this way, the southern Great Basin ignimbrite province is similar to the late Neogene Altiplano-Puna volcanic complex in the central Andes (de Silva et al., 2006; Salisbury et al., 2011). Both had super-eruptions of monotonous intermediates and subordinate contemporaneous extrusions of andesitic lavas. Both ignimbrite terranes developed on a high orogenic plateau founded on unusually thick crust as thick as 70 km that overlay a subducting oceanic plate.

In the Central Nevada field, 11 partly exposed source calderas range from 12 to 78 km in diameter. They harbor intracaldera tuffs that are at least as thick as 2000 m with intercalated landslide material that sloughed off the unstable caldera walls during continuing explosive eruptions and progressive subsidence of the caldera floors. Hundreds of meters of post-collapse sediment accumulated within the Williams Ridge, Hot Creek, and Monotony Valley calderas. In spite of substantial post-caldera extension, erosion has not exposed major intracaldera intrusions that characterize some other calderas. Local rhyolite and dacite lava flows extruded along ring faults are exposed in some calderas.

Eight eruptions of more than 1000 km³ each have been documented in the Central Nevada field (Table 2). The two largest had volumes of 4500 km³ (Monotony) and at least 4800 km³ (Windous Butte). Four of the largest eruptions broadcast ash flows over extension-corrected areas of more than 20,000 km² each and 160 km from their sources. Single outflow cooling units as thick as 550 m occur in sequences of as many as a dozen units; the aggregate thickness of sequences locally reaches 1000 m; stacks a few hundred meters thick are common. Outflow sequences are almost everywhere conformable and lack substantial intervening erosional debris and angular discordances, testifying to the lack of significant crustal extension during the ignimbrite flareup; most east-west extension occurred later.

Bedded tuff deposits produced by precursory plinian activity and/or reworking between ignimbrites are relatively thin (a few meters) and only locally evident and may in fact be coignimbrite ash deposited in advance of related pyroclastic flows. Limited plinian activity has been noted by de Silva et al. (2006, p. 55) in the geologically similar central Andean ignimbrite province. High eruption rates probably cause continuous column collapse, as ejecta boils over from the vent. Instead of a lack of plinian activity, it may be that Central Nevada caldera-forming ash flows were of sufficient vigor and mass to have swept away and engulfed the unconsolidated precursory ash-fall deposits. On the other hand, fallout ash as far as 1500 km away in Nebraska has been correlated with Central Nevada ignimbrites (Blaylock, 1998). Given the near absence of proximal plinian fallout, these distal ash beds probably formed from coignimbrite ash clouds.

In the Central Nevada ignimbrite field, six caldera-forming eruptive episodes are separated by five lulls in activity during which time little (<200 km³) or no ignimbrite was erupted; these lulls lasted from 1.7 to 4.4 m.y. and averaged ∼3 m.y. (Table 9; Figs. 73 and 74). Earliest and latest eruptive episodes (I and V–VI) were of phenocryst-rich rhyolite and each episode had an aggregate volume of ∼2700 km³. Inside these “bookend” episodes were super-eruptions of zoned dacite-rhyolite with volumes of 4800–1100 km³. Further inside the bookend episodes (episode IV from 27.57 to 25.70 Ma), a broad spectrum of magmas erupted with unusual frequency; no more than a few hundred thousand years separated these eruptions. This spectrum that composes episode IV includes ∼1700 km³ of phenocryst-rich main-trend rhyolites, 4500 km³ of three cooling units of dacitic monotonous intermediates, 600 km³ of several cooling units of Isom-type trachydacitic tuffs, and 5600 km³ of four cooling units of phenocryst-poor, off-trend rhyolite.

Long lulls (3.6 and 2.4 m.y.) preceded the eruption of 4800 km³ of the zoned dacite-rhyolite Windous Butte Formation and of 4500 km³ of the three cooling units of the monotonous intermediate Monotony Tuff, respectively (Fig. 73). These long lulls likely reflect the time for accumulation of magma in a shallow chamber before the eruption was triggered in some manner. Long lulls (2.8 and 4.4 m.y.) also preceded the last two single eruptive episodes, but these lulls were the result of the gradual demise of the greater magma-generating system prior to the transition to non-arc magma production to the south in the Southwestern Nevada volcanic field (Sawyer et al., 1994).

Central Nevada ignimbrites have typical arc chemical signatures and are mostly calc-alkalic and high-K. A spectrum of chemical compositions ranges from high-silica (78 wt%) rhyolite to low-silica (63 wt%) dacite. Most ignimbrites are rhyolite, from the earliest to the latest eruptions in the field, and most of these are phenocryst rich. Main-trend rhyolite and dacite ignimbrites were derived from relatively low-temperature (700–800 °C), water-rich magmas that equilibrated a couple of log units more oxidized than the QFM oxygen buffer at depths of ∼8 km to possibly as much as 15 km. They possess a phenocryst assemblage of plagioclase, sanidine, quartz, biotite, Fe-Ti oxides with or without hornblende, pyroxene, and titanite. Apatite and zircon occur in trace amounts. Silicic magma generation took place in thick crust (∼60 km; Fig. 3; Best et al., 2009) as large-scale inputs of mantle-derived basalt magma powered partial melting, assimilation, mixing, and differentiation processes.

Bachmann and Bergantz (2004) advanced the hypothesis that rhyolite magmas world-wide in continental settings originate as derivative phenocryst-poor melts segregated from more mafic, crystal-rich mushes, citing as an example a pyroclastic unit in the Great Basin ash-flow province. The monotonous intermediates in the Central Nevada field may correspond to the crystal-rich mush component of these crustal magma systems. However, rhyolites in the field are dominantly phenocryst rich (∼8300 km³) rather than phenocryst poor (∼5600 km³ of the ca. 27 Ma Shingle Pass). The common phenocryst-rich rhyolites are probably the result of inefficient separation of crystals from viscous rhyolitic magmas which differentiate by wall crystallization (e.g., de Silva and Wolff, 1995; Christiansen, 2005). Moreover, ignimbrites zoned from rhyolite to dacite are rare in the Central Nevada field, with only the Windous Butte and Lunar Cuesta tuffs showing this type of strong zonation.

Off-trend ignimbrites include those of the phenocryst-poor trachydacitic Isom-type that contain phenocrysts of plagioclase, pyroxene, and Fe-Ti oxides. Although there is an alkaline tendency in these ignimbrites derived from drier and hotter magmas, they nonetheless possess an arc chemical signature (e.g., high Ba/Nb ratios) inherited from a subduction-related magma system. These attributes are consistent with an origin for the Isom-type magmas involving deep-crustal fractionation of pyroxenes and plagioclase from an andesitic parent magma. The initial ⁸⁷Sr/⁸⁶Sr ratio of 0.7063 for one of these units is the lowest we have found in either the Central Nevada field or the Indian Peak–Caliente field (Fig. 7F). This implies less contamination by old continental crust than in the calc-alkaline, main-trend rhyolite and dacite contemporaries.

Also off trend are the younger, phenocryst-poor, rhyolitic Shingle Pass and Clipper Gap ignimbrites that possess some of the unusual chemical attributes of the Isom-type tuffs. These rhyolitic ignimbrites appear to have had some involvement with Isom-type magmas, perhaps via contamination or by differentiation from them. At least some of the off-trend magmas had lower oxygen fugacities (e.g., Coyote Summit Tuff Member of the Shingle Pass Formation, the tuffs of the Stone Cabin Formation, and the Isom-type tuffs).

In the southern Great Basin ignimbrite province, the caldera-forming eruptions and relatively minor extrusions of lava migrated southwestward during the 36–18 Ma ignimbrite flareup as a result of rollback of the subducting oceanic plate beneath the continent. Basalt magmas created in this subduction regime invaded the thick crust, providing heat and mass to drive the production of voluminous silicic magmas.

Despite what must have been prodigious volumes of mantle-derived basalt magmas invading the crust to drive ignimbrite-forming magma systems, virtually none was extruded during the 36–18 Ma ignimbrite flareup in the southern Great Basin ignimbrite province. Extruded andesitic lavas derived from the basalt magmas constitute only about one-tenth of the volume of silicic ignimbrite (Figs. 1 and 72). The smallest volumes erupted from 32 to 25 Ma during peak explosive activity. The unusually thick crust (∼50–70 km; Fig. 3; Best et al., 2009), a widespread partial-melt layer, and flat, shallow crustal magma chambers must all have played roles in blocking ascent of denser basaltic magmas to the surface and severely limiting extrusion of derivative andesitic magmas. In striking contrast, in the three contemporaneous fields to the east on the margins of the Colorado Plateau (Best et al., 2013a, their figure 1), intermediate-composition lavas clearly dominate over ignimbrite, including in the Marysvale (Fig. 1; Cunningham et al., 2007), the Southern Rocky Mountain (Lipman, 2007), and the Mogollon-Datil (McIntosh et al., 1992) fields.

The southern Great Basin ignimbrite province spans across a major transition in the continental crust. To the east of the buried margin of the Precambrian continent (Figs. 1–3), the middle Cenozoic crust was likely 50–70 km thick and was composed of a thick sequence of Phanerozoic sedimentary rocks overlying a granitic and metamorphic basement. This orogenically thickened crust was capped by a high-standing plateau of subdued relief we call the Great Basin altiplano (Best et al., 2009), because of its remarkable volcano-tectonic similarity to the late Neogene central Andean altiplano. The Central Nevada ignimbrite field and its contemporary Indian Peak–Caliente field (Best et al., 2013b) to the east developed on this orogenic plateau during the flareup. To the west of the Precambrian continental margin, apparently somewhat thinner crust was composed of accreted Phanerozoic oceanic rocks. The Western Nevada field (Henry and John, 2013) developed on this terrane incised into the western slope of the Great Basin altiplano by westward-draining streams.

As a result of these variations in crustal architecture, significant contrasts distinguish the three fields composing the province (Fig. 2), especially the Western Nevada from the two fields to the east. The Western Nevada field has more ignimbrite cooling units (>100 versus 76–55). Source calderas appear on the whole to be smaller, but this conclusion must be tempered by the preliminary nature of the mapping of many. There is a notable lack of the super-eruptive monotonous intermediate ignimbrites as occur to the east; nothing like them in any volume is found in the Western Nevada field. And only a few very small (<50? km³ total) cooling units of Isom-type trachydacitic tuff have been found, unlike the several hundred to thousands of cubic kilometers of such tuffs succeeding the monotonous intermediates to the east. Like the Mogollon-Datil field (Best et al., 2013a, their figure 1) and possibly the vast Sierra Madre Occidental field in Mexico (Mcdowell and McIntosh, 2012), ignimbrites in the Western Nevada field are dominantly rhyolite (Henry and John, 2013). The Southern Rocky Mountain field is more like the two fields in the central and eastern Great Basin in possessing a super-eruptive monotonous intermediate ignimbrite. Isom-type trachydacitic tuffs are absent but some ignimbrites and lavas have similar alkali- and Zr-rich compositions (Lipman, 2004).

For volcanic fields that experienced multiple super-eruptions during an ignimbrite flareup, we propose a tentative model in which the spectrum of ignimbrite compositions and volumes develops in relation to crustal thickness. In the southern Great Basin province, the Indian Peak–Caliente field, founded on what appears to have been the thickest crust (70 km), is an extreme end member with its three colossal monotonous intermediates (2000, 4400, 5900 km³) and 4200 km³ of subsequently erupted Isom-type trachydacitic tuffs; these two distinct types of ignimbrite make up about half of the total volume of the field (33,000 km³). The Central Nevada field on somewhat thinner crust is less extreme in volumes of monotonous intermediates (4500 km³) and Isom-type trachydacitic tuffs (600 km³) and total volume (25,000 km³). The Western Nevada field on still thinner crust has no monotonous intermediates and very sparse Isom-type trachydacitic tuffs, and consists dominantly of rhyolite ignimbrite. The character of other fields that experienced ignimbrite flareups seems consistent with this model.

Compared to later times, more voluminous intermediate-composition lavas were extruded during the early history of the Central Nevada field—in roughly the same space and time as the Stone Cabin and Pancake Summit ignimbrites (Fig. 72). This implies that factors blocking ascent of these magmas to the surface were less effective at the initiation of the ignimbrite flareup. Apparently, early, relatively smaller rhyolite magma chambers were built by accumulation of larger fractions of partial melt from old crust, as indicated by the highest initial Sr isotope ratios in the field (Fig. 7F). As the magma systems in the Central Nevada field grew and matured, larger fractions of mantle basalt or derivative andesite were incorporated into the systems, rather than ascending all the way to the surface. This hybridization of mantle and crustal components resulted in super-eruptions of monotonous intermediate and zoned dacite-rhyolite magmas (Table 9, Fig. 74) having lower Sr isotope ratios (Fig. 7F).

After the monotonous intermediates of the Monotony Tuff and the zoned dacite-rhyolite of the Windous Butte Formation were erupted and by the time of the Shingle Pass and earliest Isom-type activity, an unknown fraction of a huge volume of silicic magma in the crust underlying the Central Nevada caldera complex had erupted to form ∼14,000 km³ of ignimbrite. Generation of this huge volume had scavenged so much of the felsic components from the magma source region by partial melting that as additional basaltic magmas from the mantle entered the crust, less silicic magma could be generated at this focus. Moreover, the insertion of large quantities of mafic mantle-derived magma had modified the overall composition of the crust, as implied by the Sr isotopic compositions. However, there was continuing differentiation of andesitic magma from the input of primitive basalt in the deeper crust. Chiefly to the west of the residual silicic magma chambers, some of these andesitic magmas differentiated into Isom-type magmas and some Isom-type magmas apparently differentiated (and combined with crustal partial melts) to form the trachydacite-rhyolite Toquima magmas ( Appendix). In the southeastern part of the Central Nevada caldera complex, differentiating andesitic magmas apparently assimilated fertile crust to form the voluminous off-trend Shingle Pass rhyolites, which have higher Sr isotope ratios than the kindred Isom types (Fig. 7F).

As a result, initial Sr isotope ratios for Central Nevada ignimbrites and lavas are not correlated with their silica contents, but they do correlate with age. The general trend toward lower Sr isotope ratios with time reinforces the interpretation of an open-system character of these long-lived magma systems. Ratios are generally highest for the oldest rocks, implying more extensive incorporation of ancient continental felsic crust into the early magmas. With time, less of the fertile felsic material was available and the mantle-derived basalt component played an increasing role in the compositions of the volcanic rocks. The Isom-type trachydacites, like those in the adjacent Indian Peak–Caliente field (Best et al., 2013b), have the lowest initial Sr isotope ratios showing that crustal interaction was less extensive for them. The distinctive couplet of large-volume monotonous intermediate magmas and succeeding lesser Isom-type trachydacitic magmas in the Central Nevada field is an important facet of the southern Great Basin ignimbrite province.

Our research in the Central Nevada field relied heavily on the pioneering mapping, stratigraphic correlations, and volcanological insights of geologists with the U.S. Geological Survey, including R. Ernest Anderson, Frank M. Byers, William J. Carr, Gary L. Dixon, Edwin H. McKee, E. Bartlett Ekren, Richard K. Hose, Paul Orkild, William R. Page, William D. Quinlivan, Kenneth A. Sargent, Daniel R. Shawe, R.P. Snyder, and John H. Stewart. Early efforts were an outgrowth of the Nevada Test Site project of the U.S. Atomic Energy Commission. As our study of the Central Nevada field was initiated, Bart Ekren and Jack Stewart offered valuable encouragement. Gary Dixon and Rick Page gave us access to several hundred modal analyses. Dan Shawe sent copies of his quadrangle maps of the Toquima caldera complex. He and Dave John provided critically important hand samples. Chris Henry provided information and ages on select ignimbrites. For copies of unpublished geologic maps we are indebted to Will Carr, Gary Dixon, Bart Ekren, Don Hudson, Dave John, Rick Page, and Pete Rowley. Microprobe analyses were made by Michael J. Dorais at Brigham Young University. The comments of Carol Frost, Shan de Silva, and an anonymous reviewer helped us craft a more sharply focused manuscript.

Financial support for the Great Basin project was provided by the National Science Foundation through grants EAR-8604195, EAR-8618323, EAR-8904245, EAR-9104612, EAR-9706906, and EAR-0923495 to M.G. Best and E.H. Christiansen. The U.S. Geological Survey and Nevada Bureau of Mines and Geology supported quadrangle mapping. The continuing financial and material assistance of Brigham Young University is gratefully acknowledged.

Although we do not include calderas in the Toquima Range (Fig. 36) as a part of the Central Nevada caldera complex, we made a brief reconnaissance study of ignimbrites in the range that included modal, chemical, and ⁴⁰Ar/³⁹Ar analyses of several samples and determination of paleomagnetic direction at several sites. Our intent was to ascertain which Central Nevada ignimbrites are found in the Toquima Range and which ignimbrites derived from calderas in and west of the Toquima Range in the Western Nevada field occur to the east in the Central Nevada field. For additional data and interpretations, see Henry and John (2013).

Figure 71 is a synopsis from maps in this article showing the distributions of outflow ignimbrite sheets derived from the Central Nevada caldera complex. Only the Monotony Tuff has been verified west of the Toquima Range where it correlates with the tuff of Miller Mountain in the Candelaria Hills 75 km west of Tonopah (Fig. 31; Robinson and Stewart, 1984). Other than for the Monotony, distributions of major regional ignimbrites fall into three categories. Central Nevada ignimbrite outflow sheets (1) are not exposed in the Toquima Range (Stone Cabin, Windous Butte, Hot Creek Canyon, Palisade Mesa, Orange Lichen Creek, Coyote Summit, Hancock, Goblin Knobs), or (2) pinch out in the Toquima Range (Pancake Summit, Ryecroft Canyon, Egan, Pahranagat), or (3) could possibly extend beyond the Toquima Range but large-scale mapping of individual ignimbrite units in ranges immediately to the west is not available to be certain (Tikaboo, Lunar Cuesta). Many Central Nevada source calderas are eccentrically positioned in the western part of their associated outflow sheets (Figs. 2 and 71).

Geologic investigations of ignimbrites in the Toquima Range (Appendix Table 1) began with McKee (1974, 1976) who mapped its northern part and delineated the 32.91 Ma Northumberland Tuff and its source caldera (Fig. 36; Henry and John, 2013). The Northumberland Tuff is chiefly an intracaldera tuff with intercalated wall-collapse breccias, and outflow is not known beyond the northern Toquima Range. A possible correlative is the nearby, very local tuff of Stoneberger Canyon. To the south, Boden (1986, 1992) provided geologic maps and numerous modes and K-Ar ages of the younger tuffs of Moores Creek and Mount Jefferson exposed in what he called the Toquima caldera complex. We extend his designation to include the Northumberland caldera to the north and additional ones to the south described by Shawe (1998, 1999a, 1999b), Shawe and Byers (1999), and Shawe et al. (2000), including the Manhattan caldera (source of Round Rock Formation) and Round Mountain caldera (source of tuff of Round Mountain). They included in their descriptions of ignimbrite units numerous analyses, including major elements, modes by Frank M. Byers, Jr., and ⁴⁰Ar/³⁹Ar by L.W. Snee. Henry et al. (1996) provided ⁴⁰Ar/³⁹Ar ages and a description of the Round Mountain mining district sited on the Round Mountain caldera.

The upper and lower tuffs of Corcoran Canyon are the oldest ignimbrites exposed in the southeastern Toquima Range (Shawe et al., 2000). Identical ages on the lower unit of 30.01 ± 0.06 Ma and 30.01 ± 0.08 Ma were obtained by us and Henry and John (2013), respectively. A considerable thickness (2000? m) of the lower unit in a small area indicates the presence of a source caldera but it has been obscured by development of subsequent calderas. Contrasting ages on the upper tuff unit of 27.29 ± 0.05 Ma (sample QCG), 27.52 ± 0.05 Ma (Henry and John, 2013), and 27.53 ± 0.05 (Shawe et al., 2000) indicate the possibility of at least two cooling units.

We have found possible correlatives of the Corcoran Canyon to the east on the east flank of the Monitor Range in a section about 600 m thick composed of several cooling units of uncertain stratigraphic identity. Sample qcd from this section has an age of 30.00 ± 0.06 Ma that is indistinguishable from the lower Corcoran Canyon. However, its high-silica trachydacite composition differs from our single sample QCG of the lower Corcoran Canyon, a high-silica rhyolite. The cooling unit could have a variable composition. Complicating the possible correlation is the fact that the tuff of Hot Creek Canyon exposed to the east (Fig. 27) also has an indistinguishable age of 29.97 ± 0.01 Ma (Table 1). We have no paleomagnetic data on the lower Corcoran Canyon to compare with the Hot Creek Canyon, but the chemical and modal compositions of our sample of the lower Corcoran Canyon differ from those of the Hot Creek Canyon. The intracaldera nature of both of these units that are separated by an extension-corrected distance of about 30 km east-west is an additional reason for doubting a correlation. We tentatively conclude that these tuffs were the coincidental result of independent eruptions at the same time, within limits of the uncertainty of analyses. Sample QCC from the overlying 60-m-thick cooling unit in the section on the east flank of the Monitor Range section has an age of 27.40 ± 0.05 Ma and has modal and chemical compositions that are similar to the upper tuff of Corcoran Canyon, which yielded ages from 27.53 to 27.29 Ma cited above.

Our age on the tuff of Pott Hole Valley (sample VA) of 27.31 ± 0.07 Ma (Table 1) is indistinguishable from our age on the upper Corcoran Canyon. Because these two ignimbrites also share similar and unusual paleomagnetic directions, they would seem to be correlative. However, the high-silica rhyolite composition of our single sample of the Pott Hole Valley contrasts with the trachydacite composition of the upper Corcoran Canyon. The possibility of a highly variable unit cannot be ruled out. No source caldera is known for either the upper Corcoran Canyon or the Pott Hole Valley, but, if indeed they are correlative, the source would most likely be in the area of the southern Monitor and Hot Creek Ranges, that is, in the Central Nevada caldera complex.

Detailed mapping and analyses of individual cooling units in the thick ignimbrite deposits in the south-central Monitor Range are lacking (Kleinhampl and Ziony, 1985) and our reconnaissance sampling was very restricted, so any number of additional correlations could exist for other ignimbrites whose sources are in the Toquima Range, or beyond. Given the compositional similarity of most ignimbrite units, some correlations deemed only possible or unlikely on the basis of composition could prove to be valid if additional data were available. We are not aware of any certain or provisional correlations of Toquima ignimbrites farther east of the Monitor Range.

Three ignimbrites, whose source calderas lie north of 39° N and from about 50 km to possibly 175 km west of the Toquima Range, are definitely found in the Toquima Range and farther east in the northern part of the Central Nevada field. The first two discussed here lie in stratigraphic sections with the Pancake Summit and Windous Butte ignimbrites derived from the Central Nevada caldera complex. The 24.95 Ma tuff of Clipper Gap (Figs. 2 and 60) crops out in the northern part of the range and eastward almost to the Utah state line. The 25.48 Ma Nine Hill Tuff (Deino, 1985; Henry and John, 2013) is a compositionally unique ignimbrite apparently derived from a concealed caldera probably east of Reno. Probable correlatives are 0–40 m thick in the map area of Shawe et al. (2000), more than 35 m in Shawe and Byers (1999), and 0–30 m in Shawe (1999). Additional exposures are found to as far as 100 km northeast of Eureka and 35 km southwest of Ely and into the Grant Range (Fig. 2). The 29.01 Ma tuff of Campbell Creek and its probable correlative Bates Mountain Tuff unit C had a source about 75 km west of the Toquima Range; this unit is exposed mostly north of 39° N as far east as Elko (Figs. 1 and 2; Henry and John, 2013).

John (1992) showed that the regionally extensive tuffs of Toiyabe, Arc Dome, and Gabbs Valley do not extend eastward into and beyond the Toquima Range. The known source calderas for the first two tuffs lie about 50 km to the west of the Toquima Range. The titanite-bearing Toiyabe superficially resembles titanite-bearing rhyolite ignimbrites that were sampled in two places in the Central Nevada field in the Monitor Valley area where the units are 8–15 m thick: Georges Canyon Rim (sample GRGES-1C), and in low hills to the southeast (sample STONECR-2CL) (Supplemental Files 1–3 [see ^{footnotes 1–3}]). Stratigraphic control is inadequate to be sure whether these samples are from the same cooling unit but their compositions are similar. Both differ in several element concentrations from the Toiyabe, which has an age of 23.31 ± 0.05 Ma (Henry and John, 2013); this age is significantly younger than the age of GRGES-1C of 25.14 ± 0.05 Ma. This latter age is close to the age of 25.01 ± 0.06 Ma on our sample ELLSW-1ED of the tuff of Arc Dome collected in the northern Paradise Range. The Arc Dome has a similar composition to samples GRGES-1C and STONECR-2CL, except that titanite is not present. The 240-m-thick Diamond King Formation in the western Toquima Range “probably correlates” with the Arc Dome, according to Shawe (1998). Whether the ignimbrite in the Monitor Valley area is indeed a more distal correlative of the tuff of Arc Dome cannot be answered definitely without additional data. Other western Nevada ignimbrites are unknown in the Central Nevada field, but this conclusion must be tempered by ignorance of the Cenozoic geology of much of the southern Monitor Range.