Petrologic investigation of Permian metasedimentary rocks in the El Paso Mountains reveals a rock record interpreted to be consistent with the sedimentary pattern of the upper continental plate of a nascent subduction zone, based on geodynamic modeling and comparison with a Cenozoic example (Puysegur Ridge, New Zealand). Facies changes reveal a history of uplift (conglomerate), followed by subsidence (carbonate turbidite deposits) and deeper-water sedimentation (argillite, with portions deposited below the carbonate compensation depth [CCD]), and then gradual shallowing accompanied by the onset of nearby intermediate volcanism (volcaniclastic and bioclastic sediments) and construction of a volcanic edifice (andesitic lavas) in a shallow-marine environment. Comparison with Permian global sea-level curves indicates that initial uplift (relative sea-level fall) followed by deep subsidence (relative sea-level rise) are likely due to tectonic rather than eustatic effects. Shallowing during volcaniclastic sedimentation could have been due to both arc edifice building and global sea-level fall. Sandstone modal analysis suggests that the basin evolved from a tectonic setting involving compressive uplift to an arc basin setting. Geodynamic modeling implies the involvement of a transform/truncation fault in subduction initiation. Magmatic trends based on Permian paleogeography and timing suggest a limited nucleation of subduction in the El Paso Mountains followed by propagation southward. Furthermore, subduction initiation modeling suggests regional lithospheric flexure that may be reflected in coeval basins and uplift in the northern Mojave, Death Valley, and Inyo Mountains regions as well as in coeval facies changes on the western edge of the Colorado Plateau. Overall, the Permian section of the El Paso Mountains may be one of the few preserved Paleozoic sedimentary records of subduction inception along a continental margin.

Recent geodynamical modeling of subduction initiation suggests that this plate-tectonic process should produce distinct sedimentary records on the overriding plate margin depending on the manner of subduction initiation (cf. Marsaglia, 2012; for distinctions between “forced” or “induced” subduction initiation and “spontaneous” subduction initiation, see Gurnis et al., 2004; Stern, 2004; Nikolaeva et al., 2011). However, subsequent magmatism and deformation may obscure or remove this record (Sutherland et al., 2006; Gurnis et al., 2004; Stern, 2004; Hall et al., 2003). We know of no prior work specifically documenting the sedimentary fingerprint of this process in the rock record but have found a good candidate in the Lower Permian section of the El Paso Mountains, southern California. Building on the work of Carr et al. (1997), our more detailed petrologic study in combination with new geodynamic models for subduction inception allow us to interpret the Permian metasedimentary to volcanic succession in the El Paso Mountains as a product of induced subduction initiation. As such, this succession in the El Paso Mountains may be one of the first formally described and best-preserved examples of this tectonic transition.

Regional Tectonic and Depositional Setting

During the Paleozoic, the Cordilleran margin of North America underwent a rearrangement of the margin trend and a tectonic transition from passive margin to active subduction. In the early Paleozoic (Cambrian to Devonian), the Cordilleran margin of North America was a passive margin with a NE-SW trend (Fig. 1), as evidenced by sedimentary patterns (Hamilton and Myers, 1966; Burchfiel and Davis, 1972) and the 87Sr/86Sri = 0.706 line (Kistler and Peterman, 1973). By Triassic time, an active continental margin with subduction had been established along a reoriented NW-SE–trending margin (Walker, 1988; Dickinson, 2000; Stevens et al., 2005). However, many aspects of this passive-to-active transition remain poorly known, and, as a result, much debated in the literature (e.g., Snow, 1992; Dickinson, 2000; Stevens et al., 2005).

One approach to deciphering this tectonic transition is examination of the stratigraphy of sparse basin-fill remnants along the margin. These stratigraphic records, including those in the El Paso Mountains of southern California (Figs. 1, 2, and 3), were previously examined by Carr et al. (1984, 1997) and Martin and Walker (1995), and references therein.

As part of the transition, early Paleozoic passive-margin sedimentation was interrupted in Nevada and northern California by the Antler orogeny as oceanic rocks (the Roberts Mountains allochthon) were thrust over miogeoclinal rocks in latest Devonian to earliest Mississippian time (Dickinson, 2006). Assemblages of Antler-related rocks at Miller Mountain in west-central Nevada belonging to the Roberts Mountains allochthon have been correlated with Lower and Middle Paleozoic rocks of the El Paso Mountains (Carr et al., 1984, 1997), and Upper Mississippian rocks of the El Paso Mountains have been correlated with Antler foreland basin deposits (Carr et al., 1984).

The Lower Paleozoic strata of the western El Paso Mountains (Carr et al., 1997; Stevens et al., 2005) are eugeoclinal rocks (deep-water facies) that appear to have originated as part of the passive-margin eugeoclinal belt that trends from northwestern Nevada into the central Sierra Nevada in California (Fig. 1), but are now juxtaposed against miogeoclinal rocks without intervening facies (Stevens et al., 2005). It is debated as to whether the Lower Paleozoic rocks of the El Paso Mountains were displaced southward from the Antler belt, and, if so, by what means (e.g., Carr et al., 1984; Snow, 1992; Dickinson, 2000; Stevens et al., 2005). Several authors have postulated displacement by a left-lateral transform fault as part of the reorientation of the continental margin (Carr et al., 1997, 1984; Dickinson, 2000; Stevens et al., 2005). The extension of the Antler belt to the latitude of the El Paso Mountains has also been suggested (Stone, 1984); however, no structural indications of the Antler orogeny have been observed at the present latitude of the El Paso Mountains (Carr et al., 1997). Dickinson (1981), among others, proposed that the original Proterozoic rifted margin was irregular, with miogeoclinal trends following a jagged continental margin; therefore, little or no modification of the margin was required prior to later subduction initiation. Stone (1984) similarly proposed that truncation, if it had occurred, was subparallel to an irregular margin. He interpreted the sediment belt from the Antler orogeny to be wide along the southwestern trend of the margin through Nevada and then narrow as it turned southeast in California.

The apparent southward displacement of the El Paso Mountains from about the latitude of Mono Lake (Stevens et al., 2005; Dickinson, 2006) and the apparent displacement of the Caborca block (miogeoclinal rocks similar to the Cordilleran sequence in Death Valley; Burchfiel and Davis, 1981) from their respective Lower Paleozoic facies belts (Stewart et al., 1984, 1990; Ketner, 1986; Ketner and Noll, 1987), along with other evidence (see previous references), have been interpreted as supporting continental margin rearrangement by left-lateral faulting (Fig. 1) (Burchfiel and Davis, 1972, 1975; Dickinson, 2000; Stevens et al., 2005). In this interpretation, outcrops of Antler-related rocks found in the El Paso Mountains, northwest Mojave Desert, and roof pendants in the Kern Plateau of the Sierra Nevada (Dunne and Suczek, 1991) (Fig. 1) were left behind in fault-bound slivers near the truncation fault as the main body of rock was moved into northern Mexico. The parts of this “El Paso terrane” are broadly arranged from west to east, from deepest facies (Kern Plateau) to the shallowest (El Paso Mountains), although individual displacements are uncertain (cf. Carr et al., 1997; Dunne and Suczek, 1991).

The rearrangement of the continental margin was associated with and followed by the initiation of subduction along the now NW- to SE-trending margin. Development of subduction and an associated magmatic arc is indicated by Late Permian through Late Triassic plutons stretched along a NW-SE–trending belt from west-central Nevada through California into what is now Sonora, Mexico (Miller et al., 1995; Barth et al., 1997; Stevens et al., 2005; Barth and Wooden, 2006; Arvisu et al., 2009). Throughout their geographic distribution, the ages of the plutons are mixed. There are fewer Late Permian and Early Triassic plutons, and more Middle and Late Triassic plutons (Stevens et al., 2005). The late Early Permian–age pluton in the El Paso Mountains is one of the few Permian plutons and the oldest in the southern half of California (Miller et al., 1995; Barth et al., 1997; Barth and Wooden, 2006) (Fig. 1).

The plutons also vary in terms of the nature of the lithosphere into which they were emplaced: Northern California Permian plutons were emplaced in intraplate oceanic crust (Dickinson, 2000), a Permian pluton in NW Sonora, Mexico, was emplaced into continental lithosphere (Arvisu et al., 2009), and the El Paso Mountains Permian pluton was emplaced into outer or “thinned” continental lithosphere (Barth and Wooden, 2006; Stevens et al., 2005; Miller et al., 1995). A slightly younger, earliest Triassic pluton (ca. 249 Ma) was emplaced into another part of the El Paso terrane in the southern Sierra Nevada (Dunne and Saleeby, 1993), whereas Early Triassic plutons in the northern Mojave Desert were emplaced into cratonic lithosphere (Martin and Walker, 1995).

Permian plutons in northern California and Nevada are thought to have originated as island arcs that were later accreted to the continental margin (Dickinson, 2000; Arvisu et al., 2009). In contrast, Permian plutons identified in NW Sonora, Mexico, and the El Paso Mountains, California, were associated with subduction along the continental margin (Fig. 1) (Miller et al., 1995; Barth et al., 1997; Arvisu et al., 2009). Regionally, Mesozoic and Cenozoic magmatism and deformation have removed or obscured volcanic records in the northern Mojave Desert and southern Sierra Nevada. The late Early Permian pluton in the El Paso Mountains appears to be uniquely associated with a preserved Permian sedimentary and volcanic rock record.

Paleozoic rocks of the El Paso Mountains were deposited in a marine setting (see references in Carr et al., 1997). The oldest Upper Cambrian to Devonian metasedimentary rocks originated as deep-marine fine-grained clastic sediments, chert, and limestone (Carr et al., 1984); they have been described as outer continental rise facies and correlated with similar coeval facies at Miller Mountain, west-central Nevada (Stevens et al., 2005). Overlying Mississippian through Permian metasedimentary rocks are calcareous to siliceous, consisting of meta-limestone, meta-argillite, meta-sandstone, and meta-conglomerate. Permian bioclastic meta-limestone components may be derived from the western North American carbonate shelf, the Bird Spring Formation (M.D. Carr, 2008, personal commun.). Two beds of well-rounded meta-quartzarenite within the Permian volcaniclastic layers suggest continental affinities. The youngest Permian rocks in the range are volcaniclastic meta-sandstones and meta-andesitic lava flows intruding and capping the Permian metasediments (Carr et al., 1997). The Early Permian age of the metasedimentary rocks is constrained by microfossils (Dibblee, 1952, 1967; Carr et al., 1984). Younger age constraints are the U-Pb age of the meta-andesite (262 ± 2 Ma; Martin and Walker, 1995) and a Late Permian pluton (260 ± 5 Ma; Miller et al., 1995), which intrudes the Upper Cambrian to Devonian deep-marine rocks of the El Paso Mountains. To the west, the Late Permian pluton also intrudes the nearby Bond Buyer sequence of metasedimentary strata and interlayered Early Permian andesite (ca. 281 ± 8 Ma; Martin and Walker, 1995) of uncertain connection with metasedimentary strata of the central El Paso Mountains (Fig. 2).

The stratigraphy of the Pennsylvanian and Permian section in the El Paso Mountains records deposition across “structurally controlled” basins and highs, preceding the development of Permian andesitic magmatism (Carr et al., 1997). Facies relationships suggest links to coeval strata in the Death Valley region (Walker, 1988; Martin and Walker, 1995). This borderland-like topography (Stone and Stevens, 1984; Martin and Walker, 1995; Dickinson, 2000) has been linked tentatively to the Pennsylvanian–Permian(?) displacement of the El Paso Mountains terrane southward from the Antler orogenic belt, possibly along a strand of a cryptic transform/truncation fault (Carr et al., 1997, 1984; Snow, 1992; Dickinson, 2000; Stevens et al., 2005), prior to development of Permian andesitic magmatism (Carr et al., 1997; Stevens et al., 1997).

As the El Paso Mountains basin records the onset of arc volcanism, it may be described as a pre-arc continental margin basin that evolved into an arc-related basin once magmatism began. Although the exact setting (forearc, intra-arc, backarc) is somewhat equivocal given the small outcrop area, fossil content shows that the basin remained marine up until the eruption of extensive lava flows.

Estimates of later tectonic displacement of the succession vary from ∼30 km (Miller et al., 1995) up to a possible ∼100 km (Stevens et al., 2005). Late Permian SW-vergent thrusting associated with emplacement of a Late Permian pluton occurred before emplacement eastward onto the cratonic platform in Triassic or Jurassic time (Miller et al., 1995). Regionally, Early Permian contractional deformation in the Inyo Mountains and conglomeratic deposition both there (Stevens et al., 1997; Stone et al., 2009) and in the El Paso terrane of the Kern Plateau (Dunne and Suczek, 1991) appear to have occurred contemporaneously with uplift in the El Paso Mountains. Martin and Walker (1995) also correlated deformation and Upper Permian–Lower Triassic strata in the northern Mojave Desert with uplift and volcanism in the El Paso Mountains.

Local Structural Setting

Late Permian contractional tectonism and accompanying regional metamorphism affected all Paleozoic strata in the El Paso Mountains, but evidence is not present in younger rocks (Carr et al., 1997). Southwest-vergent thrust faults, overturned tight to isoclinal folds of many scales, and penetrative and spaced cleavage developed throughout the range during this event. The study area lies within a thrust-fault–bounded, ∼2-km-thick, stratigraphically upright panel of rock that appears to be among the least metamorphosed and deformed in the range. Metamorphic grade in the central part of the range, including the study area, is of the quartz-albite-muscovite-chlorite subfacies of the greenschist facies (Christiansen, 1961).

Strata in the study area have been mildly affected by both macroscopic and outcrop-scale, gently north-trending folds and by penetrative ductile deformation. Abrupt changes in amount of eastward dip along the traverse reflect the likely presence of a few large, open, kink-like flexures that are shown schematically in Figure 4. No axial-planar cleavage is associated with these folds. Carr et al. (1997) depicted similarly oriented, although more rounded folds in a cross section drawn ∼750 m south of that in Figure 4. Outcrop-scale kink folds were noted in Cambrian–Ordovician strata near the west end of the study traverse, but not in traverse strata themselves. These kink folds deflect bedding and the bedding-parallel cleavage that is thought to have originated during the regional SW-vergent tectonism, and thus are younger than that tectonism. Very mildly developed penetrative (slaty) and spaced cleavage—the latter expressed by dissolution seams—is present in the study area. The two cleavages are approximately parallel to bedding and to each other, which is a common arrangement in east-dipping upright sections throughout the range (M. Carr, 2011, personal commun.), and this pattern is lithologically controlled. In some but not all outcrops, pebbles in conglomerate are slightly flattened into bedding planes.

One major and several minor faults are present adjacent to and along the study traverse. The east-dipping fault that underlies the study area (Fig. 4) was inferred by Carr et al. (1997) to be a significant thrust fault, although no definitive evidence for this sense of slip was presented, nor was any observed during the present study. Meaningful assessment of stratigraphic throw of the fault is impossible because the fault obliquely transects formations in both footwall and hanging wall as well as a regional mid-Paleozoic angular unconformity that removed significantly different amounts of strata in different locations. Faults characterized by stratigraphic separations of a few meters or less were noted at a few locations along the traverse.

Field Methods

West and south of Mormon Flat (Figs. 3, 4, and 5), 985 m of strata were measured using a combination of Jacob staff, tape and Brunton, and a handheld Garmin eTrex global positioning system (GPS). This measured section extends across the thickest, most nearly homoclinal, and apparently intact section of Permian metasedimentary and metavolcanic rocks, as mapped by Carr et al. (1997), i.e., the so-called “metasedimentary rocks of Holland Camp.” It extends from the unconformable basal contact with Ordovician–Cambrian rocks to the conformable contact with overlying Upper Permian andesitic flows.

Carr et al. (1997) divided these strata into three members, based on lithology and conodont and fusulinid biostratigraphy; this informal terminology has been generally adapted for this study. The lowest, member A (Pha), is late Wolfcampian (ca. 280 Ma) to early Leonardian; the middle, member B (Phb), is late Wolfcampian and Leonardian; and the base of the upper portion, member C (Phc), is latest Leonardian (ca. 270.6 Ma). Phc is capped by the andesite of Goler Gulch (Pgg) with U-Pb zircon ages of 262 ± 2 Ma (Martin and Walker, 1995).

Lithologic and sedimentary features are summarized in a measured stratigraphic column (Figs. 6 and 7). One hundred-and-fifteen representative hand samples were taken at regular intervals in the measured section; selected sample locations were recorded as GPS waypoints using the North American Datum of 1983. Shallow excavations in covered intervals did not reveal the presumably less resistant material (argillite?) beneath, so these rocks are described as “covered” in the stratigraphic column.

Laboratory Methods

For more precise determination of lithologies via petrographic analysis, thin sections were made from 92 of the hand samples for petrographic description. Fifteen feldspar-bearing thin sections were stained for differentiation of plagioclase and potassium feldspar using the method outlined in Marsaglia and Tazaki (1992).

Twenty representative thin sections were point-counted for modal analysis using the Gazzi-Dickinson method (Dickinson et al., 1983; Ingersoll et al., 1984; Dickinson, 1985). The number of points counted varied according to sample grain size: as few as 100 points for a very coarse (meta)sandstone to granule (meta)conglomerate, and between 300 and 500 points for (meta)sandstone. Counted categories and recalculated parameters are defined in Table 1. Only detrital modes are presented here; petrographic point-count data were presented in Rains (2009), as were data from the measured section, GPS waypoints, details of thin-section observations, point-count data, and additional sandstone modal-analysis ternary plots.

Measured Stratigraphic Section

Overall, the section appears stratigraphically intact, although subjected to low-grade metamorphism up to greenschist facies (Carr et al., 1997). However, much original mineralogy can be determined from relict textures, mineral remnants, and pseudomorph shapes. For the purposes of modal analysis, interpretation of original mineralogy was emphasized. Determining metamorphic mineralogy would have required detailed X-ray diffraction studies, which were beyond the scope of this project. Therefore, protolith terminology is used in descriptions and discussion hereafter.

Outcrops of the thin-bedded to massive-bedded strata are variably exposed along the line of section (Fig. 4). Conglomerate forms dark, desert-varnished ridges up to 5 m high. Beds of carbonate, sandstone, silicified limestone, and a section of interbedded siltstone and very fine-grained sandstone form smaller ridges. Isolated beds of siltstone and argillite have more subtle profiles. Beds are commonly laterally discontinuous, owing to variable exposure, and variable in thickness. Throughout the study area, outcrops are locally offset by small (∼5–10 cm) faults. Consistent with Carr et al. (1997), no large fold hinges were observed in the study area. However, given the presence of cleavage, measured thicknesses are likely not original stratigraphic thicknesses.

The 985 m of sedimentary strata in the measured section (Figs. 6, 7, and 8) include 38 m of Ordovician–Upper Cambrian argillite and cherty argillite at the base, overlain by 947 m of moderately well-exposed Permian strata. However, Tertiary and Quaternary alluvium covers 315 m of the Lower Permian section. The top of the measured section ends at a stream cut that abuts a large outcrop of Quaternary alluvium (Fig. 4), so the line of section was laterally offset to the nearest exposed contact between the Permian sedimentary strata and the overlying andesite. Owing to the irregular nature of the contact between the Permian sedimentary and volcanic rocks and the trend of bedding, there may be some minor repetition of material included in the upper part of the measured section. The stratigraphic section extends 9 m into the overlying andesite flows. These are associated with hypabyssal intrusions near the top of the sedimentary section.

Overall, exposures of Permian metasedimentary strata in the measured section consist of 28% argillite and siltstone, 5% nonvolcaniclastic sandstone, 25% volcaniclastic sandstone, 5% conglomerate, 21% nonvolcaniclastic limestone, 3% volcaniclastic limestone, 3% silicified limestone and partially recrystallized chert, 4% hydrothermally altered intervals, and 6% younger, less-altered igneous intrusions of indeterminate age. Apart from the stream cut, covered intervals between outcrops make up 39% of the measured section. If all of these less resistant, covered intervals were interpreted as argillite, the overall percentage of argillite in the section would be 67%.

Lithologic Descriptions

Fine-grained facies. Within the measured section, variably calcareous argillite and siltstone outcrops make up 51% of member A (Pha), 33% of member B (Phb), and 7% of member C (Phc) (Fig. 8). As also noted by Carr et al. (1997), these very fine-grained sedimentary rocks are thinly to massively bedded, laterally uneven in thickness, and commonly fissile. Fresh rock faces vary from gray to light bluish gray to brownish gray, weathering to light brown. Parallel bedding and lamination are found throughout. In member A (Pha) and member B (Phb), this lithology is commonly exposed in continuous sections; however, in member C (Phc), thin-bedded shale and siltstone layers fine up from, or are interbedded with, more thickly bedded sandstone.

Petrographic analysis shows that several samples from the three members are composed of subequal amounts of carbonate and silica. Much of the siliceous material is recrystallized sponge spicules; carbonate is also recrystallized. These rocks were classified as calcareous siltstones or as calc-siltstones, depending on the relative abundance of siliceous versus calcareous components. In thin section, some argillite and siltstone samples show faint to distinct laminations and local bioturbation(?). Diagenetic effects include stylolites in clay-rich zones.

Most of the fine-grained siliceous facies in member A (Pha) are carbonate-free argillite and siltstone. Uncommon laminated calcareous argillite, light-gray weathering to light brown in outcrop, is found in the lower part of member A, where it is interbedded with carbonate strata. A thick section of light-gray to reddish-brown fissile argillite characterizes the middle of member A. Outcrops of member A lateral to the line of section that crosses the large covered interval (Fig. 4) were found also to be fissile argillite. At least one bed contains Nereites ichnofossils. In thin section, the reddish-brown argillite exhibits finely disseminated iron oxides in a recrystallized mud matrix. Iron-stained grayish-orange, laminated to thinly bedded argillite to sandy siltstone beds form a continuously exposed section near the top of the member. In thin section, one sample contains very angular to well-rounded, fine- to very fine-grained chert lithic fragments (∼20%), quartz grains (∼7%), angular to subangular opaque grains (∼7%), and trace sponge spicules, set in a cherty-silty clay matrix (∼60%) with patchy calcite cement. The opaque minerals define lamination.

Member B (Phb) is more calcareous than member A (Pha) and consists of thinly to thickly bedded, medium- to dark-gray argillite and medium-gray to yellowish-brown calcareous argillite interbedded with dark-yellowish-orange calc-siltstone. Covered intervals between outcrops are more extensive than in member A (Pha). In thin section, all samples from member B (Phb) consist of clay or silt with local sponge spicules, and opaque, (organic?) material.

In member C (Phc), the few exposures of argillite and siltstone are concentrated in the lower half and the very top of the member. Beds are dominantly very thick, with a few thin beds near the base and top of the member. Parallel lamination was observed in two beds. Thin sections of samples from this interval showed diverse lithologies: (1) calcareous siltstone with subequal amounts of siliceous and carbonate silt; (2) cherty argillite with minor (<5% of the rock) lithic clasts of spicular chert and siltstone; (3) volcaniclastic siltstone containing very fine-grained plagioclase and quartz grains; and (4) siltstone from the top of the member, distinguished from the other argillites in member C (Phc) by the presence of dark (organic?) material.

Quartzolithic and quartzose sandstone. Exposures of quartzolithic and quartzose sandstone are sparsely distributed throughout the measured stratigraphic column; they appear to make up ∼3% of member A (Pha), ∼0% of member B (Phb), and ∼5% of member C (Phc).

In member A (Pha), quartzolithic (cherty) sandstone occurs near the base, where conglomerate fines up to thinly bedded, pale-blue to desert-varnished grayish-black, coarse-grained sandstone. In thin section, the sandstone is very poorly sorted and similar in grain composition to conglomerates, containing angular to subrounded clasts of chert, dominantly radiolarian, but also spicular and silty. Uncommon siltstone fragments are present. Silty matrix ranges from 1% to 5%, and there is some local quartz cement.

In member C (Phc), several unique beds of brownish-gray to light-olive-gray, thickly to massively bedded sandstone have a sugary appearance in hand sample. In thin section, they are composed of well-rounded, moderately sorted (bimodal), medium-grained monocrystalline quartz grains cemented by quartz and calcite (Fig. 9E). Grain contacts are point, long, and sutured. These quartzarenite layers are unusual in this stratigraphic section because their only framework grains are monocrystalline quartz; also present are ∼1% –2% of grains of unknown primary mineralogy replaced(?) by carbonate.

Plagioclase-rich arkosic and volcaniclastic sandstone. Plagioclase-rich arkosic and volcaniclastic sandstone, where the dominant sandstone framework grains are plagioclase, are limited to member C (Phc). These grains are universally altered to some extent and locally completely replaced (pseudomorphed) by authigenic/metamorphic minerals (Figs. 9D, 10A, 10C, and 10D). Notably, no potassium feldspar was observed in these sandstones. Plagioclase-rich arkosic sandstone in the basal 60 m of member C comprises ∼3% of the unit and has only trace amounts of possibly volcaniclastic components, as opposed to the more volcaniclastic-rich sandstone in the overlying 300 m of member C (Phc), which has variably altered plagioclase grains and volcaniclastic lithic fragments in an altered matrix. Alteration varies among the samples, with plagioclase grains and volcanic lithic fragments being more easily discerned in less-altered samples.

In outcrop, plagioclase-rich arkosic sandstone beds are medium to very thick bedded, very fine or medium grained, and very light gray, weathering to pale yellowish orange. Locally, the arkose is white-speckled greenish gray, thickly bedded, and medium to coarse grained. The finer-grained samples are more matrix rich (up to ∼20%), more altered, and dominated by subrounded to rounded, untwinned and unzoned altered plagioclase with common angular to subangular quartz grains and uncommon bipyramidal zircon in one sample. Altered plagioclase grains are difficult to distinguish from a cherty matrix that may consist of devitrified volcanic glass shards. Some matrix, since it has the same distribution and size as other grains, may be pseudomatrix composed of compacted mudstone lithic fragments. Rarely, quartz grains appear resorbed (Figs. 4A, 5B, and 9D) and/or well rounded. Trace amounts of lithic fragments include limestone, siltstone, chert, and microlitic volcanics (Fig. 10A). Also present are calcite and hematite cements, and evidence of compaction.

Volcaniclastic sandstone composes ∼61% of the middle and upper section of member C (Phc). It is very thickly to massively bedded and dominantly greenish gray and less commonly bluish gray or yellowish brown. Black to white clasts give some samples a speckled appearance in hand specimen. In thin section, tuffaceous sandstones are generally medium to very coarse grained, and very poorly to moderately sorted. Clasts are predominantly subangular to subrounded. Distinguishing characteristics of volcaniclastic sandstones include the dominance of angular to subrounded, uncommonly zoned, commonly twinned plagioclase grains and/or plagioclase pseudomorphs surrounded by altered clay-calcite-opaque matrix. Plagioclase varies in degree of alteration/replacement by sericite and carbonate from ∼20% to 100%. Even when completely altered, pseudomorphed plagioclase exhibits relict shape, twinning, and zoning (Figs. 10C, 10D, and 10E).

Lithic clasts in the volcaniclastic sandstones are dominantly subangular to rounded volcanic fragments with uncommon zoned or twinned plagioclase crystals, which are variably altered or pseudomorphed (Figs. 10A, 10C, and 10D). In thin section, the composition of volcanic lithic fragments appears similar to the composition of the general matrix of the sandstones, with a similar ratio of plagioclase grains to groundmass. Volcanic lithic clasts can be distinguished from the matrix by the presence of clay or iron-oxide rims on the clasts or differences in interior plagioclase crystal size. In several samples, grain alteration to clay minerals/sericite obscures boundaries between clasts and matrix. In other samples, iron-oxide alteration is present within the sandstone matrix but not within the volcanic-lithic-fragment groundmass, helping to distinguish the two. The two stratigraphically highest volcaniclastic sandstones show pseudomorphs of vesicular glass in thin section (Fig. 10F). Other rare to common lithic components are chert and sedimentary clasts along with rare carbonate clasts. The rare carbonate clasts observed in thin section are likely intrabasinal, since they either contain volcaniclastic plagioclase crystals or were deposited near the top of the section, where bioclast type suggests shallower water depths. Some “chert” lithics are locally abundant and were likely derived from devitrified volcanic glass.

Monocrystalline quartz grains are present in trace amounts in most member C volcaniclastic sandstone. They are generally angular-subangular to subrounded, but three of the thin sections from the top of the section exhibit a bimodal suite of angular and rounded monocrystalline quartz. One sample has traces of well-rounded monocrystalline quartz that may be silicified biogenic debris, possibly radiolaria.

Trace amounts of altered biotite and muscovite appear in ∼60% of the samples studied from member C. Similar amounts of pseudomorphed amphibole(?), distinguished by its crystal form and reaction rims, are present in a few samples (Fig. 10B).

Bioclasts are found in volcaniclastic sandstone both near the base and the top of member C (Figs. 9C and 9F). Bioclasts in the lower 75 m of member C include sponge spicules and fragments(?), echinoderm fragments, bryozoan fragments, foraminifers, and unknown grains, including possible plant debris (Fig. 9F). Bioclast types in the upper 80 m of member C include algae(?), brachiopod, bryozoan, conodont(?), crinoids, and foraminifer fragments (Figs. 9C and 9F).

Matrix makes up ∼20% to 80% of sandstone samples. Some samples show more compaction than others, with abundant seaming parallel or subparallel to layering in the matrix around plagioclase grains and volcanic lithics. Cherty areas within the matrix of some samples are possibly fragments of devitrified volcanic glass.

Two samples are especially pyroclast-rich, their composition being dominantly plagioclase (or pseudomorphs) and uncommon altered amphibole(?) crystals set in ∼60%–80% matrix with a volcaniclastic texture, including some relict bubble shards. These tuffs(?) have the same general appearance and degree of alteration as the volcaniclastic sandstone, yet they have little or no nonvolcanic clasts (e.g., sedimentary clasts or bioclasts). These samples contain the highest percentage of matrix, suggesting that they were originally poorly sorted.

Conglomerate. Thickly to massively bedded conglomerate comprises ∼11% of member A (Fig. 8). Outcrops of conglomerate are light gray and weather to light brown; desert varnish gives a splotchy dark appearance to the conglomerates from a distance. Conglomeratic outcrops form ridges up to 5 m high. Beds of conglomerate are laterally uneven in thickness and appear somewhat amalgamated. In some outcrops, layering 2–10 cm thick is suggested by variations in average size of tan to dark-gray gravel clasts that range up to cobble size; there are also some laminated sandy interbeds. Massive layers 2 m to 3 m thick include a coarse-grained matrix with both rounded and disc-shaped pebbles and cobbles.

Thin sections of four conglomerate samples revealed them to be very poorly to poorly sorted, consisting dominantly of very angular to well-rounded granules to pebbles (Fig. 9A). Clasts are dominantly chert and argillite and also include trace amounts of polycrystalline quartz and sandstone composed of well-rounded monocrystalline quartz grains. The chert clasts are commonly radiolarian bearing, also spicular, or silty. Coarse-grained angular to subrounded monocrystalline quartz grains are common or occur in trace amounts. Contacts between clasts are commonly long or sutured. Silt to sand cherty matrix generally ranges in abundance from ∼1% to ∼20%, but is up to ∼70% in more altered(?) or more poorly sorted(?) samples.

Limestone and silicified limestone. Limestone in members A, B, and C is dominantly very thickly to massively bedded, light-gray to orange weathered calc-siltstone, although a few thinner beds are present. Variations include pale-yellowish-brown cherty layers in member A, several greenish-gray to bluish-gray to yellowish-brown volcaniclastic beds in member C, and several beds of moderate-dark-gray cherty bioclastic packstone near the top of the section in member C. Interbedded with cherty limestone in member A, there are layers of thickly to massively bedded, medium-gray to dark-yellowish-orange, partially recrystallized silicified limestone, and coarse-grained, light-gray to light-brown, partially brecciated, partially recrystallized chert. A similar partially recrystallized silicified limestone layer is located near the base of member C. Limestone beds are more abundant in the lower two members, A and B, than in member C.

Most samples of calc-siltstone in members A and B, and the lower part of member C (below the first volcaniclastic bed) lack bioclasts in thin section. Rare to common bioclasts in member C include sponge spicules and silicified foraminifers, bryozoan(?), and crinoid(?) fragments. One sample in member B is ∼50% siliceous-sponge-spicule silt and ∼50% calcareous silt.

Several limestone beds in member C are volcaniclastic, with variable amounts of zoned plagioclase, partly replaced by carbonate, and altered clasts (volcanic glass?). The stratigraphically highest volcaniclastic carbonate rock contains rounded plagioclase silt, trace amounts of biotite, sponge spicules, and other unidentified bioclast fragments in a calcareous silt matrix.

The remainder of the nonvolcaniclastic limestone in member C consists of thickly to massively bedded, cherty, coarse-grained bioclastic packstone. There are angular chert granules; bryozoan and sponge(?) bioclasts are visible with a hand lens. In thin section, fragments of brachiopods, bryozoans, sponges, foraminifers, and crinoids are present; some carbonate bioclasts are partially silicified (Fig. 9F).

Hydrothermally(?) altered rocks and igneous intrusions. Hydrothermally(?) altered sedimentary and/or igneous intrusive rocks are present in all three members, but their character varies among the members, with unidentifiable sedimentary protoliths and undetermined mineralogies. In member C, as previously described by Carr et al. (1997), both altered sedimentary layers and unaltered post-Permian igneous intrusions are present.

Andesite. Andesite exposure is poor in the study area. In thin section, the andesite is dominantly composed of very fine- to medium-grained, euhedral to subhedral, plagioclase crystals. Groundmass containing fine opaques composes ∼30% to ∼40% of the rock and is partially altered to chlorite.

From samples taken ∼1 km to the east of the study area, M.D. Carr (2008, written commun.) described the dominant volcanic rock as an andesite porphyry, with plagioclase laths ranging from a few percent to ∼80% of the rock. The remainder of the andesite is a very fine-grained matrix of “plagioclase, minor quartz, and opaque minerals,” along with up to 10% hornblende. Alteration minerals are chlorite and iron oxide. Lesser components of the extrusive and hypabyssal volcanic rocks are “medium-grained hornblende-pyroxene quartz diorite to diorite.”

Broad Stratigraphic Trends and El Paso Mountains Basin History

Stratigraphic trends and the history of the El Paso Mountains basin are based on the interpretation of lithology, fossil content, and sandstone detrital modes in the section studied during this investigation. Depositional environments of the three informal members in this study of the El Paso Mountains (members A, B, and C) are primarily marine, based on fossil content, with the possible exception of the local basal conglomerate of member A. Facies variations indicate changing relative water depth and depositional environments through time that could be a consequence of local tectonic or eustatic effects. Published Permian global sea-level curves show a fairly constant level from ca. 300 to 260 Ma, with a gradual fall in global sea level starting at ca. 260 Ma (Haq and Schutter, 2008). Thus, we interpret sea-level changes implied by facies analysis of the Permian stratigraphic sequence in the El Paso Mountains to be a product of tectonic rather than eustatic controls (Fig. 8). Hence, in the following discussion, we use the terms “subsidence” and “uplift.”

Above an unconformable contact with underlying Ordovician–Cambrian hemipelagic sedimentary rocks, the local base of the Permian section is marked by a distinct conglomerate. The abundant chert and cherty argillite clasts, perhaps originating from rocks similar to the underlying Paleozoic deep-marine units, implies tectonism, uplift, and subaerial exposure of deep-marine rocks in the source area, perhaps a local tectonic high. Rounded clasts indicate likely subaerial origin and transport, but lithification and metamorphism preclude detailed shape analysis to determine whether they are more likely of fluvial or beach origin. Sandstone detrital modes are similar to those associated with subduction complex or fold-and-thrust belt provenance, an interpretation that is consistent with a tectonically uplifted source (Figs. 11 and 12). The basin then deepened, or the source of the coarse clastics was shut off, as evidenced by overlying very fine-grained carbonate and hemipelagic sediments. Further deepening below the carbonate compensation depth (CCD) is suggested by an overlying interval of carbonate-free, fine-grained, red mudstone to siltstone.

The depth of the CCD is dependent on factors such as water temperature, dissolved CO2, and salinity (Berger et al., 1976, 1981). Permian deposition occurred at low latitudes, ∼10°N (Tabor and Montañez, 2002), and equatorial sea-surface temperatures may have been similar to those of today (Kiehl and Shields, 2005), suggesting CCD depths similar to the present equatorial average of 5500 m. However, modeling of Middle Permian conditions indicates that CO2 levels were about four times higher than at present (Winguth et al., 2002), suggesting shallower CCD levels, perhaps 4500 or 3500 m. Although local coastal conditions may affect the CCD, given the low latitude of the El Paso Mountains terrane in the Permian and the likely influence of the shelfal carbonate platform, subsidence below the CCD is a reasonable interpretation of the stratigraphy. Above the carbonate-free argillite, the presence of carbonate beds, including bioclastic turbidites among hemipelagic layers in member B, suggests some slight decrease in water depth (above CCD?). Although carbonate shelf sediment may be carried by gravity flows and deposited below the CCD (e.g., Nichols, 1999), the subsequent increase in coarse clastic input in member C is consistent with a shallowing trend.

The base of member C (608 m above the local base of the measured section) is defined by the presence of medium- to very thick-bedded, moderate- to well-sorted, plagioclase-rich arkose, where it is locally interbedded with lithologies similar to those in member B, suggesting an influx of coarser sediment. Possible sources of the plagioclase grains may be translated crustal blocks along a proposed truncation fault (Fig. 1) or derivation from earlier nearby(?) magmatism, represented by the Bond Buyer sequence roof pendant in the western El Paso Mountains (Fig. 2) (281 ± 8 Ma; Martin and Walker, 1995). The lack of zoning in the plagioclase in these rocks differentiates them from the overlying volcaniclastic sandstones, but we cannot rule out a local volcanic source.

The oldest disseminated volcaniclastic debris occurs at ∼678 m above the base of member A in a very thick interval of cherty calcarenite, and it is the first direct indication of local volcanic activity in this section; thick volcaniclastic beds in the overlying ∼90 m contain a varying mix of poorly to very poorly sorted, subangular to subrounded, volcaniclastic, calcareous and chert clasts. The overlying dominance of thick-bedded volcaniclastic sandstone is consistent with deposition on a submarine volcanic-arc apron, as described by Smith and Landis (1995). Very coarse-grained bioclastic limestone beds near the top of the column suggest shallowing of the basin, owing to rapid filling by volcanic debris or perhaps uplift prior to intrusion and extrusion of the capping andesite flows. This is further evidence for the growth of an intermediate volcanic edifice. Carr et al. (1997) noted that the exposed andesite comprised flows and flow breccia without pillow structures, leaving the depositional setting (marine/nonmarine) equivocal.

Relating Basin History to Subduction Initiation Modeling

We chose the El Paso Mountains study area because it appeared to have a complete stratigraphic record of magmatic-arc development, but as we have outlined herein, it also may record subduction inception along the southwestern U.S. margin. We now compare this record to stratigraphy predicted by numerical modeling of subduction initiation by Gurnis et al. (2004) and Cenozoic records of subduction initiation (e.g., Gurnis et al., 2004; Stern, 2004; Sutherland et al., 2006). These studies imply that the sedimentary record of “induced” subduction initiation in a proto-forearc would begin with rapid uplift (subaerial), followed by rapid subsidence, and then shallowing (Fig. 13). Cenozoic sedimentary records of subduction initiation along Puysegur Ridge, New Zealand, referenced in the modeling, also show an ongoing sedimentary pattern of rapid uplift, rapid subsidence, and then shallowing (Gurnis et al., 2004; Sutherland et al., 2006). In contrast, “spontaneous” subduction initiation should result in a sedimentary pattern of initial subsidence and magmatism in the proto-forearc (Stern, 2004; Marsaglia, 2012).

Numerical modeling by Gurnis et al. (2004) indicates that “induced” subduction initiation requires a through-going fault or lithospheric break such as a transform fault or ridge, buoyancy contrast across the fault (Niu et al., 2003), and development of compressive tectonics before progressing to self-sustaining subduction. In the Permian, the underlying lithosphere in the El Paso Mountains was likely associated with transitional continental crust (outer continental rise; Stevens et al., 2005), and therefore was more buoyant than subducting oceanic crust and lithosphere. As described already, some authors have suggested the El Paso terrane was translated from northern California and left behind on a sliver of the postulated late Paleozoic transform/truncation fault (e.g., Stevens et al., 2005; Dickinson, 2000). Alternatively, as mentioned already, if the Antler belt extended to what is now central California, little or no latitudinal translation of the El Paso Mountains terrane may have occurred in the Pennsylvanian or Permian (Dickinson, 1981; Stone, 1984; Carr et al., 1984; Snow, 1992; Dunne and Saleeby, 1993; Stevens et al., 2005). In any case, it is proposed that left-lateral transform faulting brought oceanic lithosphere into juxtaposition with the El Paso Mountains terrane, enabling the necessary density contrast and creating an opportunity for subduction initiation along a through-going zone of weakness.

Looking only at southern California, the observed pattern of southwestern Cordilleran Permian–Triassic plutons, based on palinspastic reconstruction of Stevens et al. (2005), suggests that magmatism began in a small area (El Paso Mountains) before spreading south (Barth and Wooden, 2006) (Fig. 14). This pattern may be analogous to the Cenozoic initiation and propagation of subduction along the Puysegur Ridge, New Zealand, and consistent with the record of restricted nucleation and progression of volcanism along that margin (Sutherland et al., 2006).

The overall Permian sedimentary pattern of the El Paso Mountains appears consistent with models for “induced” or “forced” subduction initiation, rather than “spontaneous” subduction (Gurnis et al., 2004; Stern, 2004; Sutherland et al., 2006) (Fig. 13). Specifically, initial uplift is evidenced at the base of member A by sandy conglomerate, followed by a deepening marine sequence that suggests subsidence below the CCD. Deep-marine strata occur throughout member B, where renewed carbonate influx suggests progressive shallowing/uplift. Coarsening of sediment and increased carbonate at the base of member C are consistent with further shallowing. This shallowing trend was overprinted by input from the evolving magmatic arc, as first indicated by disseminated volcaniclastic debris near the base of member C, which then becomes more volcaniclastic-dominated upward, consistent with deposition on a prograding volcaniclastic apron. Very coarse-grained bioclastic limestone near the top of the measured section indicates a very shallow-marine setting prior to proximal magmatism (andesite intrusions and flows).

The time span of subduction initiation to magmatism in the Gurnis et al. (2004) model is less than 10 m.y. (Fig. 13). For the Permian section in the El Paso Mountains, the interval between the uplift of deep-marine rocks and the incursion of volcaniclastic sediments was also ∼10 m.y.

Limited Permian subduction initiation in the southern California region contrasts with more extensive Triassic subduction development along the trace of the subsequent Mesozoic arc. Saleeby (2011) suggested that Triassic subduction followed latest Permian compressive tectonics along the oceanic transform-continental margin interface. His detailed study of remnants of the transform within the Kings-Kaweah ophiolite belt (KKO, Fig. 1) gives an age of ≥200 Ma; this transform would have resulted in a large density contrast between outer continental and negatively buoyant transform oceanic lithosphere, i.e., conditions favoring subduction initiation in the Triassic along the length of the transform. However, in the Early Permian, as mentioned herein, a record of contraction occurring at the same time as uplift in the El Paso Mountains is found from the western El Paso terrane in the Kern Plateau (Dunne and Suczek, 1991) through the southern Inyo Mountains and vicinity (Stevens and Stone, 2007; Stone et al., 2009), supporting an earlier occurrence of perhaps “induced” subduction. We speculate that the localization of Permian subduction in the El Paso Mountain region and its apparent southward propagation were associated with a preexisting condition either associated with the lithospheric structure of the soon-to-be subducting oceanic plate (e.g., change in age/structure/thickness across a transform or fossil spreading or other aseismic ridge at that paleolatitude) or the presubduction geometry of the overriding plate (e.g., older indentation or jog associated with a prior tectonic regime).

Regional Ramifications of Subduction Initiation

Whereas we interpret the Permian sedimentary record of the El Paso Mountains to show a proximal sequence consistent with subduction initiation, it is also worth considering potential far-field effects, such as those discussed by Holford et al. (2009), and implied by geodynamic models (Gurnis et al., 2004) (Fig. 1), which may include shoreline migration on the craton and uplift of the Last Chance allochthon (e.g., Snow, 1992).

The numerical models imply that several-hundred kilometers inboard of the trench, on the overriding plate, the rock record would show subtle vertical changes at subduction inception, followed by more substantial subsidence as subduction resistance lessened (Fig. 13). Taking into account the removal of Basin and Range extension and shortening the distance, shoreline regression on the western edge of the Colorado Plateau between ca. 285 and ca. 278 Ma, as determined by sedimentary isopachs (Zahler, 2006, and sources cited within), may be reflecting the lithospheric flexure predicted by modeling (Fig. 1). Between ca. 275 Ma and ca. 270 Ma, both the El Paso Mountains and the Colorado Plateau show evidence of transgression/subsidence in opposition to global eustatic trends (Fig. 8) but consistent with vertical motions suggested by the geodynamic model of Gurnis et al. (2004).

Petrologic investigation of Permian metasedimentary rocks in the El Paso Mountains reveals a rock record indicating uplift (member A) and subsidence (member A, middle and top, and member B) before influx of volcaniclastic sediments and renewed shallowing (member C).

At the base of member A, upward-coarsening conglomeratic strata indicate uplift. A high-energy environment with variable flow energy is suggested by fairly large chert and argillite clasts and variable amounts of matrix. Sharp contacts, upward fining, and lack of sedimentary structures are consistent with gravity-flow deposits. Clast composition suggests derivation from deep-marine sedimentary rocks. However, the lack of marine bioclasts means that the sedimentary environment was not conclusively marine.

The conglomeratic deposits are overlain by fine-grained turbidite facies (middle and top of member A and in member B) that indicate subsidence. Overlying the conglomeratic deposits, there are carbonate and siliceous layers (middle of member A), which are overlain by a carbonate-free layer (middle to the top of member A). Subsidence below the CCD is suggested by lack of carbonate beds and the red color of argillite layers. Abyssal depth is also suggested by Nereites ichnofossils (Prothero, 1998). In member B, the presence of fine-grained carbonate layers among hemipelagic beds suggests a relative decrease in water depth. A marine depositional environment is indicated by local marine microfossils.

Plagioclase-rich and volcaniclastic sediments dominate member C. At the base of member C, two intervals of plagioclase-rich arkosic sandstone interrupt the fine-grained sedimentary pattern. Carbonate beds are thicker up section below the volcaniclastic section. The first volcaniclastic influx, as indicated by rare plagioclase crystals in thickly bedded limestone, signaled the beginning of volcaniclastic input. Overlying layers in the middle to upper part of member C are dominated by plagioclase-rich strata, volcaniclastic sandstone, and resedimented tuffs, with local interbeds of quartzarenite, limestone, and uncommon interbeds of argillite. Layers are dominantly very thickly bedded to massive, parallel bedded, very fine to very coarse grained, and poorly to very poorly sorted, consistent with deposition on a volcanic apron. Intermediate magma composition is suggested by the dominance of plagioclase and sparse amphibole phenocrysts, similar to those in overlying andesite. Marine bioclasts within both volcaniclastic sandstone and local beds of limestone in the lower and upper sections of member C indicate a marine depositional environment. The dearth of volcaniclastic material in bioclastic packstone near the top of the section suggests development of a fringing reef or carbonate platform between eruptive events. The very coarse grain size of the packstone and vesicular remnants(?) in volcaniclastic sandstone near the top of the section suggest shallowing of the volcanic apron or filling of the volcanic basin.

When compared with numerical models and Cenozoic examples of subduction initiation, especially the Miocene to Holocene Puysegur Ridge–Fiordland subduction zone of New Zealand, the interpreted Permian sedimentary record of the El Paso Mountains—of uplift and subsidence followed by volcaniclastic sediment influx—is consistent with the pattern of a developing magmatic arc in a newly initiated subduction zone. Since the El Paso Mountains basin records the onset of arc volcanism, it may be described as a pre-arc continental margin basin that evolved into an arc-related basin once magmatism began. The exact setting (forearc, intra-arc, backarc) is somewhat equivocal, given the small outcrop area, but fossil content shows that it remained marine up until the eruption of extensive lava flows. Reported magmatic trends in central and southern California after ca. 285 Ma (Barth et al., 1997), using palinspastic Permian locations from Stevens et al. (2005), suggest that volcanism nucleated in or near the El Paso Mountains and then propagated southward (Fig. 14).

Subduction initiation modeling and Cenozoic oceanic arc records predict likely near-field and far-field effects in relative water depths for the overriding plate. Depositional environments in the three informal members of the study area of the El Paso Mountains are primarily marine, with the possible exception of the local conglomeratic base of member A. Stratigraphic variations in the Permian sedimentary rocks in the El Paso Mountains indicate changing relative water depths, consistent with predicted near-field effects. Global sea-level curves within the time frame of the Permian section (ca. 280 Ma to ca. 260 Ma) in the El Paso Mountains indicate that the inferred changing relative water depths were likely the result of regional tectonics, especially between ca. 280 Ma and ca. 267 Ma. Far-field effects may be reflected in coeval facies changes on the western edge of the Colorado Plateau and may be related to coeval uplift and basin formation in the Death Valley region and Last Chance allochthon as well. In the Colorado Plateau, facies show local variations in relative water depth; the pattern suggests crustal flexure associated with subduction initiation in the El Paso Mountains and may be consistent with predictions based on subduction initiation models.

We thank Michael D. Carr for providing background information; this study depended in large part on work previously done by him and other investigators of the El Paso Mountains. Graduate research funds from the Department of Geology of California State University–Northridge helped support this work. We thank Richard Rains, Whitney Behr, and Kevin Rivera for field assistance. The manuscript was greatly improved by reviews by Ray Ingersoll and an anonymous reviewer, as well as comments from John Wakabayashi.