Regional Geologic Framework of Mineral Deposits in the Stibnite-Edwardsburg Area, Central Idaho Open Access

Central Idaho is underlain by folded metamorphosed sediments of Mesoproterozoic, Neoproterozoic, and Paleozoic age that have undergone greenschist and, locally, amphibolite facies metamorphism and have been extensively intruded by several pulses of Cretaceous magmatism (Fig. 1). Eocene Challis volcanic deposits locally overlie the older rocks, and Eocene intrusions are also present as dikes, sills, and stocks. The area is complexly faulted with apparent thrust, strike-slip, and normal movement. Several of the structures in the vicinity of Stibnite and Edwardsburg were extensively mineralized and have been exploited to produce gold, tungsten, antimony, and mercury.

In recent years, central Idaho has been the focus of several studies by the Idaho Geological Survey. Access is poor, and previously published detailed geologic maps nonexistent. To provide fundamental geologic data for use in biological studies, mineral exploration, and regional tectonic syntheses, a series of geologic maps have been produced. To the north of Stibnite, regional geologic mapping of the central and eastern parts of the Big Creek drainage culminated in the publication of a 1:75,000-scale geologic map (Fig. 2; Stewart et al., 2013). More detailed mapping of the Big Creek 7.5′ quadrangle was published at 1:24,000 scale (Lewis et al., 2012b). Mapping during the summers of 2012 to 2014 resulted in the 1:24,000-scale Stibnite quadrangle map (Stewart et al., 2016). Reconnaissance surficial and bedrock geologic mapping of the proposed Burntlog Creek access route from Landmark to Stibnite, Idaho, in 2016 resulted in a 1:36,000-scale map (Stewart et al., 2017) and 2019 field work in the Yellow Pine quadrangle west of Stibnite resulted in a 1:24,000-scale map (Stewart et al., 2021). Several previously unmapped faults have been discovered, some regional in extent, and by the establishment of a coherent stratigraphy for both the Precambrian strata and the Challis Volcanic Group, the history of offset on these and other known structures is now better understood. Details of the age and character of mineralization at Stibnite, part of a parallel Idaho Geological Survey effort, was published by Gillerman et al. (2019). Here, we present a summary of our mapping efforts, which were supplemented by U-Pb geochronologic work, to provide a regional geologic framework for the other papers in this issue of Economic Geology, particularly with respect to stratigraphy, structure, and magmatic activity. In addition, new garnet geochronology estimates are presented here that help refine the metamorphic history of the area.

Stratigraphic and geochronologic studies in central Idaho by Lund et al. (2003, 2010, 2015) and mapping by Lund (2004) described a NW-trending series of discontinuous metasedimentary bodies as recording pre-, syn-, and post-Rodinian rifting (Fig. 1). Rifting is thought to have occurred in a series of events from <780 to ca. 720 Ma (Brennan et al., 2021). Some of these bodies are on ridge tops and appear to be roof pendants in the Idaho batholith, but others may be septa and extend to depth. They are composed largely of metasedimentary strata with minor metavolcanic rocks and form a composite section that spans the late Mesoproterozoic to early Paleozoic. The metavolcanic rocks near Edwardsburg were first recognized and described by Leonard (1962), who noted a “family resemblance” to metavolcanic rocks in northeastern Washington and southeastern Idaho. Subsequent U-Pb zircon dating of the interbedded volcanic rocks led to correlation of some of the detrital units with the Neoproterozoic Windermere Supergroup, providing a link to northern and southern segments of the rifted margin (Lund et al, 2003; 2010), a conclusion that was further supported by the discovery of extensive Neoproterozoic through Archean zircon inheritance in the Idaho batholith by Gaschnig et al. (2013).

Quartzite and metamorphosed limestone in the Stibnite area were suspected by early workers to be lower Paleozoic in age (Larsen and Livingston, 1920; Schrader and Ross, 1925; Ross, 1934; Shenon and Ross, 1936). Mapping by Benjamin Leonard incorporated in the Challis 1° by 2° quadrangle assigned a Mesoproterozoic age (Hoodoo and Yellowjacket formations; Fisher et al., 1992), but detrital zircon analyses as young as 500 Ma clearly show the quartzite and metamorphosed limestone to be Cambrian to Ordovician (Stewart et al., 2016; Isakson, 2017). Ma et al. (2016) also recognized Cambrian and Ordovician strata 85 km to the south in the Sawtooth Mountains (Fig. 1). Fossils reported by Lewis and Lewis (1982) from east of Stibnite have not been confirmed by subsequent workers, but a possible bryozoan reported by Gillerman et al. (2019) from north of Salt Creek may indicate potential for future discoveries.

Complicating the stratigraphic story are a series of intrusions related to formation of the Late Cretaceous Idaho batholith. Strata in the Stibnite and Yellow Pine area are in and near the eastern edge of the southern (Atlanta) lobe of the Idaho batholith (Fig. 1). Ages of the Atlanta lobe intrusions range from about 100 to 67 Ma, with the majority younger than 80 Ma (Gaschnig et al., 2010, 2017). Additional magmatism in the Eocene formed the Challis volcanic field (Fisher et al., 1992) and numerous plutonic rocks of this age are present across the region (Lewis and Kiilsgaard, 1991; Lund, 2004).

Ore at Stibnite is hosted in Idaho batholith granitoids and Neoproterozoic to Paleozoic metasedimentary rocks. Important historical references include Schrader and Ross (1926), Currier (1935), White (1940), Cooper (1951), Lewis (1984), Smitherman (1985), and Cookro et al. (1987). Ore is localized along the regional-scale N-S–trending Meadow Creek fault, related structures such as the West End fault, and subsidiary structures. Types of alteration and mineralization are complex, and they span a lengthy period of time. Gillerman et al. (2019, in press) describe a series of events from late Cretaceous to Eocene in age within the district. The main stage of gold mineralization appears to be 66 to 61 Ma. Later epithermal-style mineralization was dated at the West End deposit with samples of hydrothermal potassium feldspar from carbonate-quartz-adularia veins and gold-mineralized schist dated at 52 to 51 Ma (Gillerman et al., 2019, in press). The earlier, deeper, more disseminated and largely intrusive-hosted Au-As mineralizing event was likely the most intense and contains the larger resource (Gillerman et al., in press). It is temporally and spatially distinct from the later ores. Tungsten, as scheelite, crosscuts the gold mineralization. Scheelite dated by U-Pb geochronology constrains the age of main-stage tungsten mineralization to approximately 57 Ma (Wintzer et al., in press) with some evidence for minor scheelite deposition at 47 Ma. Antimony mineralization in the form of stibnite cuts both gold and tungsten mineralization and is intergrown with younger scheelite that yields a U-Pb age of 47 Ma (Wintzer et al., in press).

A unit of the Mesoproterozoic Belt-Purcell Supergroup is overlain by roughly 6,000 m of strata in the Edwardsburg and Stibnite areas (Figs. 3, 4). Most of the strata are nearly vertical and young to the southwest. The oldest strata are exposed to the north near Edwardsburg (Fig. 5), where the Mesoproterozoic Apple Creek Formation of the Belt-Purcell Supergroup is well exposed and characterized by graded beds (Fig. 6A), some of which are carbonate bearing. The Apple Creek is overlain by a thick (~1,350 m) quartzite unit (Square Mountain Formation of Lund, 2004). The Square Mountain Formation differs significantly from Belt quartzite in that it lacks appreciable detrital feldspar and is poorly sorted. Thick bedding and trough cross-beds are characteristic (Fig. 6B). Overlying strata of the Anchor Meadow Formation consist largely of tremolite marble, calc-silicate rocks (Fig. 6C), and dark siltite with a total thickness of about 970 m. Age constraints on the Square Mountain and Anchor Meadow formations are poor, and it is not clear if they are part of the Neoproterozoic Windermere Supergroup or an older stratigraphic sequence. Detrital zircon U-Pb analyses from the Square Mountain Formation show a similarity to zircon populations in Belt Supergroup strata (primary peak about 1740 Ma), but the local presence of ~1360 to 1280 Ma grains indicates that it is younger (Fig. 7; Isakson, 2017) and possibly correlative to the Deer Trail Group in northeastern Washington (Box et al., 2020).

The Anchor Meadow Formation is overlain by the Edwardsburg Formation, which consists of three mappable members (from bottom to top): the Wind River Meadows, Golden Cup, and Placer Creek Members. The basal thin (~75 m) feldspar-poor quartzite of the lower Wind River Meadows Member of the Edwardsburg Formation (Zewrl in Fig. 4) is in turn overlain by a distinctive matrix-supported conglomerate (diamictite; Fig. 6D) of the upper Wind River Meadows Member (Zewru in Fig. 4). The diamictite unit is ~135 m thick on the ridge 6 km northwest of Edwardsburg, and includes clasts of quartzite, siltite, and calc-silicate rocks 1 to 4 cm across within a calc-silicate or, locally, calcitic matrix. Detrital zircon analyses from the quartzite and the diamictite matrix show an abundance of 1.6 to 1.0 Ga grains (Fig. 7; Isakson, 2017), clearly indicating a Neoproterozoic maximum depositional age.

Most of the Golden Cup Member (Zegc in Fig. 4) consists of amphibolite, interpreted as metamorphosed mafic lava flows, along with subordinate, fine-grained volcaniclastic rocks with an aggregate total thickness of 700 m. Locally, plagioclase phenocrysts are preserved (Fig. 6E), as are calcite-filled vesicles. This mafic magmatism likely resulted from continental rifting (Lund et al., 2003). The upper part of the Edwardsburg Formation is a matrix-supported conglomerate (Placer Creek Member diamictite; Zepc in Fig. 4), ~20 m thick, containing spectacular stretched clasts in a mafic matrix (Fig. 6F). The clasts are typically 2 to 8 cm in length and highly varied, consisting of mafic volcanic rocks—quartzite, siltite, and calc-silicate (Fig. 6G). The Placer Creek Member is absent to the southeast near Edwardsburg, where the topmost member of the Edwardsburg Formation is the Hogback Rhyolite Member (Fig. 6H) interpreted by Lund (2004) as tuff and dated at 684 ± 4 Ma by Lund et al. (2003). A slightly older U-Pb zircon age of 694.0 ± 1.6 Ma was obtained more recently by Isakson (2017). Above the Edwardsburg Formation is ~100 m of quartzite of the Moores Station Formation. The bulk of the Moores Station Formation (500 m thick) is phyllite (Fig. 6I) that contains lenses of marble. Cretaceous granitic rocks intrude the upper part of the formation. Detrital zircon analyses from the diamictite matrix and the overlying quartzite show a prominent 1.74 Ga peak and, aside from a few young (~650 Ma) grains in the quartzite, are like ages found in the Square Mountain Formation (Fig. 7; Isakson, 2017).

Fifteen kilometers to the south of Edwardsburg and up section, strata in the Stibnite area (Figs. 3, 8) span the late Neoproterozoic and extend into the early Paleozoic (Fig. 4; Stewart et al., 2016). The stratigraphic link between metasedimentary sections in the Stibnite and Edwardsburg areas is tenuous because of intervening Cretaceous intrusive rocks of the Idaho batholith, faulting, and Eocene volcanic cover (Fig. 3). The most continuous and best-studied part of the section is east of Stibnite, where the lowermost unit defined by Smitherman (1985) is quartzite and schist (Figs. 4, 9A). The lower part of this unit is intruded by granitic rocks of the Idaho batholith, but it is at least 130 m thick. The quartzite and schist unit is overlain by 300 m of calc-silicate rocks (lower calc-silicate unit; Fig. 9B) and, locally, a thin (40-m) dolomitic carbonate unit (Fern marble; Fig. 9C). These rocks are in turn overlain by matrix-supported quartz-pebble conglomerate up to 100 m thick (Fig. 9D) that locally contains interbedded schist layers. Detrital zircon analyses from the quartzite and schist unit and quartz-pebble conglomerate show an abundance of grains from 1.5 to 1.0 Ga (Fig. 10; Stewart et al., 2016; Isakson, 2017) and are thus similar to Neoproterozoic to lowermost Cambrian strata in southeastern Idaho (Yonkee et al, 2014).

A relatively thick (~180 m) feldspar-poor quartzite (lower quartzite unit) that locally contains thin conglomerate and granule beds overlies the quartz-pebble conglomerate (Fig. 9E). Detrital zircon analyses from the quartzite show a much older distribution (mostly 1.9 to 1.7 Ga) than the underlying rocks (Fig. 10; Stewart et al., 2016; Isakson, 2017); these resemble age results from Cambrian quartzite in the region (Gibson Jack Formation in southeast Idaho, Yonkee et al., 2014; Addy Quartzite in northeast Washington, Linde et. al., 2014). Above the quartzite are 125 m of calc-silicate rocks (upper calc-silicate unit of Smitherman, 1985; Fig. 9F), overlain by 180 m of gray marble (middle marble unit) that is massive to locally ribbon laminated. This marble contains local, thin quartzite intervals and is overlain by the thin (~80 m) relatively fine grained and poorly sorted middle quartzite unit of Smitherman (1985). Above this quartzite is a dolomitic marble (Hermes marble) 100 m in thickness that is in turn overlain by about 400 m of feldspar-poor, fine- to coarse-grained quartzite (upper quartzite; Fig. 9G). We believe this uppermost quartzite is Ordovician in age (Kinnikinic or Eureka equivalent) based on the similarity of detrital zircon ages to those reported by Barr (2009) for the type section of the Kinnikinic near Clayton, Idaho, 110 km to the southeast of Stibnite (Stewart et al., 2016; Isakson, 2017). A thin calc-silicate unit (calc-silicate of Midnight Creek) overlies the quartzite, above which a fault repeats some of the older units.

Most of the same stratigraphic sequence outlined above is also exposed north of Stibnite near Sugar Mountain, where the Tamarack Creek anticline repeats the section (Fig. 8) and on Missouri Ridge, where the middle marble (Fig. 9H) and interbedded quartzite are well exposed. Importantly, detrital zircon U-Pb analyses from quartzite in the middle marble unit on Missouri Ridge and near Sugar Mountain show a large population of ~500 Ma zircons (Fig. 10; Stewart et al., 2016; Isakson, 2017). Similar results have been found for the Upper Cambrian Worm Creek Member of the St. Charles Formation in southeast Idaho (Link et al., 2017). The middle marble on Missouri Ridge was mapped by Lund (2004) as the Missouri Ridge Formation and assigned a Neoproterozoic age. We support the retention of the name Missouri Ridge Formation but re-assign it a younger age (Cambrian or Ordovician) based on the young zircon ages. For the purposes of regional mapping, we suggest the strata between the lower and upper quartzite (upper calc-silicate, middle marble, middle quartzite, and Hermes marble units of Smitherman, 1985; Stewart et al., 2016) be assigned to the Missouri Ridge Formation as informal members. South of Sugar Mountain, the Sugar Mountain fault places younger strata (lower quartzite and lower calc-silicate) on the north against older strata (Edwardsburg and Moores Station formations) to the south. Our present interpretation is that this fault is an overturned thrust, as discussed in more detail below.

Less well understood is the relationship of strata northwest of Stibnite, where thick, feldspathic quartzite (quartzite of Profile Creek; >500 m), previously assigned to the Mesoproterozoic Gunsight Formation of the Belt Purcell Supergroup (Lund, 2004), is in fault contact with structurally overlying schist and calc-silicate rocks of the Moores Station Formation (Whiskey Creek fault; Fig. 8). This quartzite contains abundant 1.1 Ga zircon grains and is thus younger than the ~1.45 Ga Belt Supergroup (Fig. 10; Stewart et al., 2016, Isakson, 2017). The quartzite of Profile Creek is darker colored than the other quartzite units in the area (Fig. 9I). It contains significant amounts of biotite and about 15% plagioclase feldspar. At least some of the feldspar is interstitial and in veinlets and thus is secondary. The feldspar and biotite content may reflect the influence of the nearby Idaho batholith, and perhaps a greater degree of metamorphism. We tentatively interpret the quartzite of Profile Creek to be of Neoproterozoic age.

A bimodal sequence of syenite and diorite is well exposed in a series of plutons aligned in a prominent NW-SE orientation northeast of Edwardsburg (Figs. 1, 3). The diorite was interpreted as metavolcanic rocks by Leonard (1962), but as mafic intrusive rocks comagmatic with the syenite by Lund et al. (2010). It includes a fine-grained variety (microdiorite) and subordinate amounts of fine- to coarse-grained gabbro. Perthitic alkali feldspar is the dominant mineral in the syenite, which contains as much as 25% hornblende and biotite (although typically less than 15%). Lund et al. (2010) obtained a U-Pb zircon age of 651 ± 5 Ma from a sample collected on the east side of Ramey Ridge about 12 km northeast of Edwardsburg and a second sample to the east on Acorn Butte yielded an age of 665 ± 6 Ma by sensitive high resolution ion microprobe (SHRIMP) methods and 653 ± 2 Ma by thermal ionization mass spectrometry (TIMS) methods. More recent analytical work by Isakson (2017) on the same set of Acorn Butte zircons yielded a weighted mean date of 652.43 ± 0.17 Ma using chemical abrasion-isotope dilution TIMS (CA-IDTIMS) methods. Chemical data for these Neoproterozoic igneous rocks are consistent with continental rift-related magmatism (Lund et al., 2010). Extension was thus recurrent in the region in the Neoproterozoic, occurring at about 690 Ma with extrusion of the lavas of the Edwardsburg Formation (Lund et al., 2003; Isakson, 2017) and again at about 650 Ma with intrusion of the syenite-diorite sequence.

Early workers noted many different varieties of granitic rocks in the Stibnite area and speculated that “Perhaps the Idaho batholith is more complex than might be judged from available data.” (Schrader and Ross, 1925, p. 143). We concur that the intrusive rocks in the area are much more heterogeneous at outcrop scale than typical Idaho batholith to the south near Lowman and understand why they were mapped as “mixed Cretaceous intrusives” by Fisher et al. (1992). Perhaps the complexity is due at least in part to the fact that the Stibnite-Edwardsburg area lies on the eastern margin of the batholith (Figs. 1, 3). See Gillerman et al. (2019, in press), Box, John et al. (in press), and Lund et al. (2023) for detailed age and chemical information on the various phases of intrusive rocks.

We differentiate four intrusive units based on volumetric predominance at a given locality. See published maps listed in Figure 2 for their distribution; all are shown as “Cretaceous intrusive rocks” on Figures 3 and 5 and “granite and granodiorite” on Figure 8. The oldest batholithic rock is porphyritic granodiorite, most widely exposed in the Burntlog Creek area, 25 km southwest of Stibnite (Fig. 1; Stewart et al., 2017). The porphyritic granodiorite is characterized by 0 to 20% euhedral pink potassium feldspar phenocrysts as much as 4 cm in length (Fig. 11A). Biotite is present as well-disseminated flakes up to 2 mm in size, comprising as much as 15% of the rock. Hornblende is sparse but can be found at most localities and is locally up to 1 cm in length. Magnetite is a common constituent, comprising as much as 2% of the rock, and the unit has higher magnetic susceptibility than younger intrusions of the Idaho batholith near and west of Yellow Pine (Stewart et al., 2021). Its presence is important for interpretation of magnetic surveys of the region (see Anderson et al., in press) as the magnetic signal from this unit can potentially be confused with that from magnetite-rich Eocene plutons. As with the biotite granodiorite, it is typically intruded by leucocratic granite dikes. Typically foliated, and locally lacking phenocrysts, the unit is referred to as foliated biotite granodiorite and porphyritic hornblende-biotite tonalite by Lund et al. (2023). The unit is part of a NW-trending belt of porphyritic granodiorite and tonalite that extends 70 km southeast to Stanley, Idaho (Fisher et al., 1992). Unruh et al. (2008) determined a U-Pb age on multigrain zircon populations of 90.3 ± 1.2 Ma from a sample along Johnson Creek near Whitehorse Rapids, 20 km south of Yellow Pine. Zircons from this sample were reanalyzed by Lund et al. (2023), using SHRIMP single zircon spot ages and an age of 93.4 ± 0.5 Ma was determined. A second sample of hornblende-bearing granodiorite collected northwest of Stibnite, 3.2 km north of the Meadow Creek Lookout (Fig. 3), is probably from rock belonging to this unit. It initially yielded a U-Pb age on multigrain zircon populations of 93.2 ± 1.3 Ma (Unruh et al., 2008) and a more precise SHRIMP age of 95.3 ± 0.5 Ma is reported by Lund et. al. (2023). The porphyritic granodiorite unit may also be present as unmapped plutonic bodies near the Yellow Pine pit. A porphyritic granodiorite with conspicuous potassium feldspar phenocrysts just north of the pit returned a 94.21 ± 0.22 Ma U-Pb TIMS age (Gillerman et al., 2019).

Biotite granodiorite (mapped as quartz monzonite by most previous workers) is the most common plutonic rock in the Stibnite area but is more highly variable in texture than typical batholith granodiorite. Biotite is present as small (<2 mm), well-disseminated flakes that make up 5 to 10% of the rock (Fig. 11B). Mafic material is more common in the Stibnite area than typical for the Idaho batholith and zones rich in mafic schlieren or containing discrete mafic inclusions are widespread. Small, unmapped mafic bodies of quartz diorite are present locally within this unit and they locally exhibit evidence of chilled margins against comagmatic granodiorite (Fig. 11C). A preliminary laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) U-Pb zircon date of 91.2 ± 2.2 Ma reported by Stewart et al. (2016) from biotite granodiorite core near the Meadow Creek mine was later revised to 89.9 ± 1.7 Ma (Gaschnig et al., 2017). Box, Wintzer, et al. (in press) report several SHRIMP U-Pb zircon ages in the 96 to 92 Ma range and one at about 86 Ma from rocks mapped as the biotite granodiorite unit. They also recognize several chemical types based on major- and trace-element compositions.

Two-mica granite near Stibnite occurs primarily as small stocks and is distinctive in containing books of muscovite up to 3 mm thick (Fig. 11D). The two-mica granites have similar composition to the leucocratic granite described below and may be textural variations of a single magma type. Gillerman et al. (2019) report a LA-ICPMS U-Pb zircon age of 86.0 ± 5.4 Ma from a drill hole sample in the Yellow Pine pit that penetrated a granite body possibly correlative with the two-mica bodies mapped at the surface. Alternatively, this sample was from a coarser variety of the leucocratic granite unit.

Leucocratic granite (described as alaskite or aplite by previous workers) is common in the area and occurs primarily as dikes cutting biotite granodiorite, porphyritic granodiorite, or country rock, but also as small stocks. It contains quartz, plagioclase, and potassium feldspar in subequal amounts (Fig. 11E). Muscovite is present as isolated to well-disseminated flakes up to 2 mm in size, locally comprising as much as 10% of the rock. Biotite is less common. Magnetite is locally abundant and some of the highest magnetic susceptibility measurements in the Yellow Pine quadrangle come from these dikes (Stewart et al., 2021). LA-ICPMS U-Pb zircon ages from two localities of the leucocratic granite from Stibnite area are 83.61 ± 0.08 and 84.8 ± 1.3 Ma. (Gillerman et al., 2019).

An outlying intrusive body east of the Yellow Pine pit termed the Stibnite stock (Fig. 8) is compositionally transitional between the biotite granodiorite, two-mica granite, and leucocratic granite (Fig. 11F). Finer than typical biotite granodiorite, it contains minor amounts of muscovite. A preliminary LA-ICPMS U-Pb zircon date on the stock of 84.9 ± 2.0 Ma by Stewart et al. (2016) was later revised to 84.4 ± 2.1 (Gaschnig et al., 2017). Gillerman et al. (2019) report a more precise ID-TIMS age of 85.71 ± 0.10 Ma from the same stock. In contrast, Box, Wintzer, et al. (in press) report a SHRIMP U-Pb zircon age of 97.3 ± 0.65 Ma age on a sample from this body on the edge of the Stibnite open pit, and shallow SHRIMP analyses of the zircon crystal surfaces on the same sample yielded a U-Pb age of 76.8 Ma, possibly for a hydrothermal rim. The reason for this discrepancy in the interior zircon age is not clear.

Contacts between Cretaceous intrusive and metasedimentary rocks are sharp and in the Stibnite area some contacts are subhorizontal with intrusive rocks above, not below, the metasedimentary strata (Fig. 11G). Drilling results and surface exposures like those pictured indicate that some of these intrusive rocks are likely sills. Elsewhere, such as 5 km east-southeast of Yellow Pine, the contact is shallow (there gently east dipping) and granitic rock of the Idaho batholith underlies the metasedimentary rocks (ZYs; Fig. 3).

Rocks of the Eocene Challis volcanic field are widely exposed northeast and east of Stibnite (Figs. 1, 3) and U-Pb zircon and ⁴⁰Ar/³⁹Ar ages of these rocks that range from 50 to 43 Ma are reported in Box, John et al. (in press). The structural history of these volcanic deposits has been unclear, in part because of difficult access. Leonard and Marvin (1982) speculated that a nested set of collapse features (Thunder Mountain caldera, Cougar Basin caldera, and the Quartz Creek caldron) were all centered on Thunder Mountain. Our mapping has shown that likely eruptive centers and collapse features are smaller and more widely distributed (Stewart et. al, 2016; Fig. 3). We term the area a “caldera complex” and propose below a possible sequence of episodes to produce the extensive and partly fault bounded volcanic field seen at present.

Previous mapping east and southeast of Stibnite by Fisher et al. (1992) has established that at least three of the four of the ash-flow deposits associated with the Thunder Mountain vents were produced by large-volume eruptive events. They also assigned informal names to these four units. The oldest is the dime and quarter dacite tuff sequence (Tdq on Fig. 3), followed by the buff rhyolite (Tbr), the lower Sunnyside tuff rhyolite (Tssl) and finally, the upper Sunnyside tuff rhyolite (Tssu). Examples of typical exposures of these units are shown in Figure 12. The buff rhyolite differs in that clear-cut evidence of pumice or welding is lacking and an alternative explanation is that it represents a series of rhyolite flow domes (Box, John, et al., in press). We prefer the ash-flow tuff interpretation because of the distance across which exposures occur (25 km). While it is difficult to establish with precision where the collapse boundaries following the four eruptive events might have been, Figure 3 presents speculative outlines of the four linked calderas that may have produced the present caldera complex. These outlines are based primarily on where thick intervals (100s of meters) of each deposit are preserved and, in the case of the upper Sunnyside tuff, the presence of collapse-associated megabreccia (Tmbx). Following the four eruptions, each forming its respective caldera (with the possible exception of the buff rhyolite), there was a fifth collapse event, the formation of the Big Creek graben, which was not associated with an eruption. This may have taken place soon after the latest caldera collapse, or over an extended period after volcanic activity ended.

The faults bounding the Big Creek graben extend into the Tertiary intrusive rocks to the northeast and into Cretaceous batholith rocks to the southwest. Southwest of the upper Sunnyside caldera the graben has been down-dropped less and exposes a deeper crustal section containing dime and quarter tuff and equivalent Tertiary intermediate intrusive rocks in which are exposed tabular, NNE-striking intrusive bodies that comprise the Pistol Creek dike swarm. This indicates that the Pistol Creek dike swarm either postdates the dime and quarter tuff and predates the upper Sunnyside tuff or, less likely, that the intrusive event is younger than the upper Sunnyside tuff but the dikes ended at depth and did not penetrate the younger, higher-level volcanic rocks. Box, John, et al. (in press) report LA-ICPMS zircon ages for two samples from the Pistol Creek dike swarm at 49.3 ± 1.5 and 48.17 ± 0.57 Ma.

The faults that form the perimeter of the caldera complex are not uniform in nature, reflecting the compositionally heterogeneous and structurally compromised crustal block in which the magma chambers that fed the Tertiary pyroclastic eruptions developed. Some segments of the caldera-bounding fault system are characterized by large (10-m diameter) blocks of collapse debris such as those found in the Stibnite quadrangle north of Sugar Creek (Fig. 8). Portions of the faults bounding the caldera complex are marked by abundant breccias and iron staining as is found west of Murphy Peak and south of The Pinnacles (Fig. 3). Local exposures of mixed volcaniclastic and granite- or quartzite-bearing sediments rest unconformably on the dime and quarter tuff (Fig. 12C), indicating that subsidence was a protracted process, at least during formation of that caldera.

Challis plutonic rocks are present both to the southeast of Stibnite as well as northwest and northeast of Edwardsburg (Fig. 3). Compositions include pink biotite granite and hornblende-biotite granodiorite. Radiometric age determinations are scarce. Dikes of the Challis magmatic episode are widespread in central Idaho, and they are particularly abundant in a NE-trending belt 15 km south of Stibnite (Pistol Creek dike swarm, Fig. 3). Ages of dikes in the Stibnite area range from about 49 to 46 Ma (Gillerman et al., 2019; Box, John et al., in press).

The degree of metamorphism of prebatholithic rocks is varied across the Stibnite-Edwardsburg area and consists of both dynamic regional metamorphism and contact metamorphism (Currier, 1935; Smitherman, 1985). Mesoproterozoic strata northeast of Edwardsburg (Fig. 5) have been metamorphosed to greenschist facies. Contact metamorphic effects adjacent to the Idaho batholith northwest of Edwardsburg produced staurolite (Kirkpatrick, 1974) and staurolite is also present east of Stibnite (Fig. 9A). In addition to effects of regional metamorphism, contact effects east of Stibnite and hydrothermal alteration of the carbonate rocks are pronounced, and detailed descriptions are provided by Cooper (1951) and Smitherman (1985). In the area east of Yellow Pine and northwest of Stibnite, the metasedimentary rocks have reached amphibolite facies and that metamorphism appears to be entirely regional in nature. Sillimanite is common throughout the schist of the Moores Station Formation in this area and garnets are found locally (Stewart et al., 2016). Wintzer (2019) reported midcrustal pressures and temperatures of 7.5 kb and 775°C.

A garnet-rich biotite schist and a garnet-rich felsic dike were collected for garnet geochronology (Lu-Hf method) to investigate metamorphic timing in the Stibnite area. The schist locality is near the mouth of Salt Creek northwest of Stibnite, in the hanging wall of the Whiskey Creek fault (Fig. 8). The schist is a metamorphosed part of the Neoproterozoic Moores Station Formation of the Windermere Supergroup (Stewart et al., 2016). A garnet-rich felsic dike that crosscuts the schist unit about 600 m east-northeast of the schist sample was also collected. Details of the Lu-Hf method used are outlined in Wintzer (2019). A summary is given in Appendix 2.

Garnet ages can be determined using the isochron method, assuming Lu and Hf did not undergo open-system behavior. The garnet in the biotite schist yields an age of 112.8 ± 6.9 Ma (mean square of weighted deviates [MSWD] = 4.8; Fig. 13A). The initial ¹⁷⁶Hf/¹⁷⁷Hf ratio is 0.282420 ± 140 (determined by the y-intercept of the isochron) and has an ε_Hf of –1.29 at the time of metamorphism. This and other initial values were calculated from the measured values of each analyzed fraction and are found in Table 1. The isochron is regressed through three garnet fractions (G1-G3) and one whole-rock fraction. Garnet fraction 1 (G1) is an outlier because it plots far from the line regressed through the three other garnet fractions and was excluded from age calculations. If the outlier is included, the garnet in the biotite schist would yield an age of 114 ± 10 Ma. The whole-rock bomb-dissolved fraction (WRB) was excluded due to significantly elevated Hf concentrations (8.05 ppm; Table 1), which is likely from dissolving Hf-rich zircon grains. Whole-rock bomb methods allow more extensive dissolution, including zircon grains, which is not the case with whole-rock Savillex (WRS) methods. Distribution of two-point garnet dates are shown on Figure 13B, which includes a weighted mean garnet model age. Two-point garnet dates are calculated using the slope of the line from a garnet fraction to the initial ¹⁷⁶Hf/¹⁷⁷Hf ratio, determined from the isochron.

Garnet from the felsic dike yields an age of 99.3 ± 2.0 Ma with an MSWD of 2.6 (Fig. 14A), with an initial ¹⁷⁶Hf/¹⁷⁷Hf ratio is 0.282494 ± 33 with an ε_Hf of –1.03. The isochron is regressed through three garnet fractions and two whole-rock fractions. Figure 14B displays distribution of garnet fraction dates and shows the weighted mean garnet model age. The second garnet two-point date (G2) is an outlier and was not included in age calculations. The isochron age with outlier included would yield an age of 103 ± 11 Ma.

Foliation is readily apparent in the garnet-biotite schist, and garnet inclusion trails are parallel to foliation (Fig. 15A, B). The weak fabric within the garnet and the strong fabric above and below the garnet suggest the garnet grew during development of the fabric. Deformation is also present in the garnet-bearing felsic dike, which cuts the schist. However, deformation is less pronounced in the dike and deformation is not apparent at the mesoscale. Microstructures within the dike that document deformation are subgrain rotation in quartz and some deformation twins in plagioclase (Fig. 15C; Passchier and Trouw, 2005). Metamorphism of the strata northwest of Stibnite occurred at about 113 Ma (Fig. 13A) a time of metamorphism in the vicinity of the arc-continent boundary to the west near Riggins (144 to 90 Ma; Getty et al., 1993; Wilford, 2012; McKay et al., 2017). The ~113 Ma age of metamorphism is older than ages reported for the granitic rocks at Stibnite, indicating that this regional event predated the batholith intrusion. Contact metamorphism like that found east of Stibnite and northwest of Edwardsburg likely occurred some 15 to 30 m.y. later, depending on the specific pluton age.

The deformational history of the metasedimentary rocks in the Stibnite and Edwardsburg areas is complicated and includes episodes of contractional folding and faulting as well as extension. Much of the contractional deformation predated, or was synchronous with, the emplacement of the well-foliated early metaluminous suite of the Idaho batholith (100–85 Ma, Gaschnig et al., 2010). The metasedimentary rocks in the Stibnite area lie along the western boundary of the Thunder Mountain caldera complex, and extension was likely both synchronous with, and subsequent to, the eruption of the Eocene Challis Volcanic Group. This extension also formed a series of NE-striking normal faults within the batholith and metasedimentary rocks (Figs. 1, 3).

Important regional structures include the Johnson Creek shear zone and the Big Creek graben (Fig. 3). The Johnson Creek shear zone is a series of subparallel NNE-striking faults along which there has been ductile deformation, brittle deformation, and local silicification. Lund (2004) notes that ductile shears are in pre-Tertiary rocks and brittle fault segments are present in both Eocene and pre-Tertiary rocks. We have seen similar relationships developed along the northern part of the structure in the area northwest of Edwardsburg where mylonitic Cretaceous rocks are cut by unfoliated Eocene dikes. Lund (2004) reports that mylonitic fabrics superimposed on Late Cretaceous granitic rocks indicate that the fault zone had a Late Cretaceous to Paleocene component of right-lateral movement. In the Yellow Pine area, mylonitic fabrics indicate transpressive, right-lateral motion and later brittle deformation and silicification was focused along the east edge of the shear zone (Stewart et al., 2021). Similar right-lateral transpressive motion is postulated for the boundary between the accreted island-arc terranes and the continental cratonic rocks in western Idaho (McClelland et al, 2000; Giorgis et al., 2008), but that deformation may have ended before (at ~90 Ma) the transpressive deformation in the Stibnite-Yellow Pine area. The Big Creek graben is a NNE-striking structure bounded by the Cave Creek and Cow Creek faults (Fig. 3; Stewart et al., 2013). As discussed above, it offsets Eocene Challis Volcanic Group units and is likely the youngest of the Challis-related extensional structures. Details of the local structures of the Stibnite area, shown in Figure 8, are discussed below. Cross sections illustrating our interpretation of these structures at depth are shown on Plate 2 of Stewart et al. (2016).

The NW-SE–trending Garnet Creek syncline exposed in the Stibnite area was recognized by White (1941) and discussed by Smitherman (1985). It is overturned to the southwest (Fig. 8) and indicates NE-SW compression. It is doubly plunging, suggesting it is refolded along NE-trending axes similar to the Tamarack Creek anticline (see below). Drill data indicates that along the upright lower limb of the Garnet Creek syncline the stratigraphic section is partly cut off by a series of at least two large granite sills (Chris Dail, writ. commun., 2019). Cleavage within the fold is interpreted to be axial planar. The syncline is tentatively interpreted to fold the Cinnabar Peak fault at depth (Stewart et al., 2016). The metamorphic garnet age of 113 Ma given above dates the axial planar cleavage and indicates NE-SW compression of that age.

The Tamarack Creek anticline is exposed in the Missouri Ridge-Sugar Mountain area and is northwest plunging (Fig. 8; Stewart et al., 2016). For most of its exposed trace, the fold is overturned to the southwest. However, near Tamarack Creek the fold becomes upright, and the fold wavelength increases. Cleavage is interpreted to be axial planar to the fold. Smaller NE-plunging folds northwest of the Salt Creek fault refold the sedimentary units as well as the Tamarack Creek anticline, resulting in deflection of bedding 1.3 km west-northwest of Sugar Mountain (Fig. 8). The Tamarack Creek anticline is interpreted to have formed at roughly the same time as the Garnet Creek syncline (late Early Cretaceous), dated by axial planar fabric metamorphism at about 113 Ma. Deformation by the NE-plunging folds postdates that age but predates intrusion of the local plutonic rocks of the Idaho batholith at about 95 Ma (Gillerman et al., 2019; Lund et al., 2023; Box, Wintzer et al., in press).

The Cinnabar Peak fault strikes northwest and dips steeply to the northeast (Fig. 8; Stewart et al., 2016). The fault was originally mapped by Smitherman (1985) as a normal fault, but he did not name the structure. The fault places overturned upper quartzite (Ouq) against overturned quartzite and schist (Zqs) to the southeast. The fault appears to be parallel or nearly parallel to overturned bedding on both the hanging wall and footwall. The fault zone is poorly exposed, kinematic indicators were not observed, and the interpretation of a fault was largely due to the repetition of strata. Fault relations near the Stibnite stock are also uncertain, but the present interpretation is that the Cinnabar Peak fault is cut by the stock. If correct, the fault would be older than the 85.7 Ma age of the intrusion. We tentatively interpret the Cinnabar Peak fault to be an overturned thrust fault, folded by the Garnet Creek syncline (see cross section E-E’ in Stewart et al., 2016). If correct, thrust motion was originally to the southwest. An alternative, less-favored interpretation is that the Cinnabar Peak fault is a late-stage, down-to-the-east normal fault that postdates the Garnet Creek syncline.

The Sugar Mountain fault strikes northwest and dips steeply to the northeast, though it is locally folded by the same NE-plunging folds mentioned above that folded the Tamarack Creek anticline (Fig. 8; Stewart et al., 2016). The fault separates overturned lower quartzite (Єlq) and upper calc-silicate (OЄucs) on the east from overturned Edwardsburg Formation (Ze) and the marble of Moores Station (Zmsm). Both the Edwardsburg Formation and upper calc-silicate units are discontinuous along the fault, although the fault appears to be roughly bedding parallel. The fault zone is characterized by abundant breccia and fracturing, local zones of mylonite, and locally is intruded by Tertiary dikes. As with the other faults in the area, the fault zone is poorly exposed and kinematic indicators were not observed. Iron staining is common along the fault, and garnets as large as 2 cm are abundant within fractures along a section of the fault immediately south of Tamarack Creek. We tentatively interpret the Sugar Mountain fault to be an overturned thrust fault folded by the Tamarack Creek anticline. Like the Cinnabar Peak fault, thrust motion is interpreted to have originally been to the southwest, based on orientation of the overturned Tamarack Creek anticline but we consider this speculative. Some attenuation of the Edwardsburg Formation and the upper calc-silicate units during folding is required for this interpretation.

The relationship between the Sugar Mountain fault and the Cinnabar Peak fault is uncertain. The two structures, as well as strata in the Stibnite and Missouri Ridge-Sugar Mountain areas, are offset by the Meadow Creek fault (see below). Upon-the-west and right-lateral motion on the Meadow Creek fault indicates that the Missouri Ridge-Sugar Mountain area exposes a deeper crustal section than the Stibnite area, a conclusion that is supported by the presence of older rocks in both the hanging wall and footwall of the Sugar Mountain fault with respect to the Cinnabar Peak fault. Assuming both the Sugar Mountain and Cinnabar Peak faults are overturned folded thrust faults, the Sugar Mountain fault may simply be a separate, deeper-level thrust fault relative to the Cinnabar Peak fault. Alternatively, the Sugar Mountain fault may be linked to the Cinnabar Peak fault by a lateral ramp, remains of which may be present along the dismembered stratigraphic zone north of Sugar Creek. This lateral ramp would have been reactivated by the Meadow Creek fault. Additional discussion about the relationship between the two faults is covered below in the description of the West End fault. Alternatively, if the Cinnabar Creek and Sugar Mountain faults are both normal faults as mapped by Smitherman (1985), both may be the same structure, with the Sugar Mountain fault representing the structure at a deeper level.

Several lines of reasoning suggest that the Cinnabar Peak fault and the Sugar Mountain fault are older compressional features (folded thrust faults) rather than younger extensional ones (normal faults). First, the NW-SE trend of both the Cinnabar Peak fault and the Sugar Mountain fault are closely paralleled by the trends of the Garnet Creek syncline and Tamarack Creek anticline, suggesting that all four structures are most likely products of a single, possibly protracted, compressional disturbance. In addition, the NW-SE trend is extremely rare for extensional faults in central Idaho; such a trend is consistent with basin-and-range extension, evidence of which is abundant south and east of Challis, Idaho (Lewis et al., 2012a), but such extension is Miocene or younger in age. Finally, the Cinnabar Peak fault appears to be cut by the Stibnite stock, dated at 85.7 ± 0.1 Ma by Gillerman et al. (2019), which makes it older than this Cretaceous intrusion and most likely a product of Sevier-aged compression. The Sugar Mountain fault is intruded by Eocene dikes and cut by the Missouri Ridge fault, with similar implications.

The Meadow Creek fault is a major structure in the Stibnite area (Fig. 8) and is the controlling structure for mineralization at the Meadow Creek mine (immediately east of the fault) and the Yellow Pine mine (immediately west of the fault). In the southern Stibnite quadrangle, it strikes north-south and appears to bend to the north-northeast at the Yellow Pine pit. We concur with the mapping of Cooper (1951) and extend the NE-trending fault farther northeast, crossing Sugar Creek about 1.2 km northeast of its mouth. The northeast continuation of the fault cuts Cretaceous granodiorite as well as displacing the Stibnite metasedimentary body from the Missouri Ridge-Sugar Mountain strata, apparently about 3 km in a dextral sense. Dextral offset of similar magnitude is also apparent in aeromagnetic data for the Stibnite area (Anderson et al., in press). Although the major structure makes the northeast bend, faults extend to the north and southwest, respectively, from the major fault bend (Stewart et al., 2016). This northeast portion of the Meadow Creek fault was termed the “Yellow Pine shear zone” by White (1940). Cooper (1951) did not have access to the underground workings that intersected the fault zone at the Meadow Creek mine, but he cites Bradley Mining Company level maps that show the fault dipping east at shallow levels in the mine and west at deeper levels. Currier (1935) also noted this change in dip. Recent drilling has confirmed this change in dip, but it appears to be rather abrupt and possibly due to a cross fault (Chris Dail, writ. commun., 2019). Where the Meadow Creek fault bends to the northeast at the Yellow Pine pit, mineralization is extensive, particularly west of the fault (Cooper, 1951).

Cooper (1951) reports that the Meadow Creek fault east of the Yellow Pine pit is 145 ft (44 m) wide, strikes north-northeast, and dips westerly. He states that both slickenlines and cleavage indicate that the west (hanging wall) moved obliquely upward and to the north relative to the footwall. The age of movement is poorly constrained but considerable motion must postdate both the emplacement of the Cretaceous granodiorite as well as the development of the Garnet Creek and Tamarack Creek folds. Some believe that the fault also has some postmineralization movement, and it has been suggested that the mineralized area at the Meadow Creek mine may once have been adjacent to the mineralized area at the Yellow Pine pit (Chris Dail; oral commun., 2012). We did not find evidence, pro or con, regarding this hypothesis. An altered dacite dike just north of the fault trace lacks evidence of significant deformation, suggesting that motion along the fault was largely completed by Eocene time. A possible genetic link between the Meadow Creek and West End faults is discussed in the next section.

The West End fault strikes northeast and is subparallel to the Meadow Creek fault (Fig. 8). Unlike the early work of White (1941) and Cooper (1951), ours shows the southern extension of the fault continuing south-southeast of the Yellow Pine pit rather than ending near the pit. Movement on the West End fault included both normal (southeast side down) and right-lateral components, as shown by drill hole data (Chris Dail, oral commun., 2012). Drill core also demonstrates that, at the West End pit, the West End fault dips steeply (65°) to the east while mineralized zones dip shallowly to the east (Chris Dail, oral commun., 2012).

On first consideration, the West End fault appears to be subsidiary to the Meadow Creek fault; however, study of the rocks caught between the two faults points to a possible genetic relation by way of sequential fault movement. North of Sugar Creek is a zone of structurally disrupted blocks of mostly metasedimentary rocks lying between the northwestward extension of the Cinnabar Peak fault and the southeastern extension of the Sugar Mountain fault that we term the dismembered zone (Fig. 8). Analysis of fault orientations and metasedimentary block displacements within the dismembered zone suggests the West End and Meadow Creek faults are bounding faults of a right-lateral strike-slip fault system that links the Cinnabar Peak-Sugar Mountain thrust fault, detaching and re-orienting segments of the thrust fault to form the dismembered zone (Fig. 16). Displacement within the fault system can be accounted for with several stages of right-lateral movement. Initial movement was on the West End fault. Movement then shifted sequentially northwestward to new fault strands, jumbling and re-orienting the blocks in the dismembered zone. Final movement was on the Meadow Creek fault.

The Whiskey Creek fault strikes north-northwest, dips moderately to the east-northeast, and places the Moores Station Formation over the quartzite of Profile Creek (Fig. 8; Stewart et al., 2016). It occurs only to the northwest of the Salt Creek fault and is presumably intruded by Cretaceous biotite granodiorite to the south. The fault is roughly parallel to bedding in the quartzite of Profile Creek. The fault zone is characterized by a well-developed foliation in both igneous and metasedimentary rocks (Fig. 17).

The Whiskey Creek fault is interpreted to record thrust motion; however, there is uncertainty due to the ambiguity in the stratigraphic position of the quartzite of Profile Creek. We tentatively interpret the Whiskey Creek fault to be a SW-directed thrust that formed as a splay from the Sugar Mountain fault (see cross section C-C’ on the map of Stewart et al., 2016). Although well exposed just north of the East Fork of the South Fork of the Salmon River, the fault has no definitive kinematic indicators and the structure needs additional field study.

The Fern fault is a significant NE-striking fault east of Stibnite (Fig. 8) but kinematics along this structure are uncertain due to poor exposure. Originally mapped by White (1941), the Fern fault dips steeply to the west, and cuts metasedimentary rocks folded by the Garnet Creek syncline. As noted by Cooper (1951), it displays apparent left-lateral displacement of Proterozoic and Paleozoic strata, which could result from sinistral displacement, northwest side-down normal faulting, or prebatholithic, southeast side-up reverse faulting. The fault does not appear to offset the southwestern contact of the metasedimentary strata with Cretaceous batholithic rock, suggesting that its displacement predates the Cretaceous age of the intrusive rocks. It appears to continue at its northeast end into the southeast side-down fault that places the Eocene volcanic rocks against the metasedimentary strata (Stewart et al., 2016). This observation seems to indicate reactivation of a segment of the older Cretaceous fault in Eocene time.

The Missouri Ridge fault in the far northwestern corner of the Stibnite 7.5′ quadrangle strikes northwest and dips to the northeast (Fig. 8; Stewart et al., 2016). It places a package of gently folded strata stratigraphically above the quartzite and schist unit and below the middle marble over more tightly folded rocks of the Moores Station Formation. The fault zone is poorly exposed and kinematic indicators were not observed. We interpret the Missouri Ridge fault to record normal motion, bringing the hanging wall of the Whiskey Creek fault down to the north and east (see cross section A-A’ by Stewart et al., 2016), placing more gently folded Whiskey Creek hanging-wall strata in contact with the more tightly folded rocks adjacent to the Sugar Mountain fault. Movement on the Missouri Ridge fault must postdate thrusting associated with the Whiskey Creek and Sugar Mountain faults, and likely cuts Cretaceous intrusive rocks, but otherwise is poorly constrained.

The contact between the Challis Volcanic Group and older rocks (Proterozoic and Paleozoic strata, the Idaho batholith) near Sugar Creek is a normal fault with the Challis Group downdropped on the northeast side (also as mapped by Lund, 2004) (Fig. 8). Breccia is developed along the fault at three localities near Sugar Creek and two small mafic intrusive bodies were emplaced along the structure (Stewart et al., 2016). Whether this fault represents the margin of a caldera is less certain, but a thick sequence of buff rhyolite is widely exposed to the northeast of this fault and marks a possible source area for this unit, either as an ash-flow tuff or, less likely, a rhyolite flow (buff rhyolite caldera shown in Fig. 3). To the southeast, 3 km north of Murphy Peak, a small mass of Cretaceous granodiorite first mapped by Currier (1935) is within the Challis Group outcrop area that appears to be outside both the Big Creek graben and any of the other caldera collapse features. A similar relationship is found 4 km south-southwest of Edwardsburg where granodiorite of the Idaho batholith is found south of the east-west normal fault marking the northern extent of the dime and quarter tuff (Fig. 3). Areas exposing these Cretaceous plutonic rocks are unlikely to be near the surface within a caldera. Thus, this east-west structure, and many of the other faults that locally mark the extent of the Challis Volcanic Group, are likely not caldera-bounding structures.

A series of NE-striking normal faults cut the metasedimentary strata, Cretaceous intrusive rocks, and in places, the Challis volcanic rocks. All are thought to be young relative to the faults described above, and in this respect, they are like the NE-striking faults that bound the Big Creek graben, east of the Stibnite area. The Salt Creek fault (Fig. 8), striking northeast and dipping steeply to the northwest, is another NE-striking structure. It shows apparent left-lateral displacement of metasedimentary strata that could result from either normal (northwest side-down) or true left-lateral motion, or both. Alternatively, the Salt Creek displacement could be the result of offset on a steep, east side-up, prebatholithic reverse fault. It appears to be displaced by the NW-trending Eocene caldera-complex bounding fault and is therefore older.

The Stibnite-Edwardsburg area contains a wide variety of metasedimentary host rocks that were subjected to regional and contact metamorphism, multiple episodes of intrusion by Late Cretaceous magmas, and a protracted history of structural disruption. The oldest rocks in the region (Mesoproterozoic Belt Supergroup) are greenschist facies metasedimentary rocks about 1.4 Ga in age that are well exposed northeast of Edwardsburg (Figs. 1, 3). These are overlain by Neoproterozoic and lower Paleozoic sequences to the southwest (Fig. 4). The Neoproterozoic section near Edwardsburg includes metamorphosed mafic lavas and dacitic volcanic rocks dated at 685 Ma (Leonard, 1962; Lund et al., 2003) and associated diamictite. Other more common lithologies are quartzite, calc-silicate rocks, schist, and marble. The stratigraphic sequence is interrupted south of Edwardsburg by intrusive and volcanic rocks such that ties with the stratigraphy to the south in the Stibnite area are uncertain. Best understood is the stratigraphic succession east of Stibnite, which extends from the Neoproterozoic into the lower Paleozoic. Importantly, the upper quartzite unit of Smitherman (1985) and Stewart et al. (2016) can be correlated with the Ordovician Kinnikinic Quartzite 110 km to the southeast near Clayton, based on detrital zircon age populations (Stewart et al., 2016; Isakson, 2017). Marble-rich metasedimentary rocks on Missouri Ridge northwest of Stibnite (Figs. 3, 4, and 8) previously assigned by Lund (2004) to the Neoproterozoic contain detrital zircons as young as 500 Ma. Consequently, the Missouri Ridge Formation is reassigned here to the Cambrian and may include some Ordovician strata as well. Less well understood is the relationship of strata northwest of Stibnite, where the quartzite of Profile Creek was found to contain abundant 1.1 Ga zircon grains and is thus younger than the ~1.45 Ga Belt Supergroup (Fig. 10; Stewart et al., 2016, Isakson, 2017). Because this quartzite is in fault contact with structurally overlying schist and calc-silicate rocks of the Moores Station Formation (Whiskey Creek fault; Fig. 8), its relationship to the other strata in the area is uncertain. The Whiskey Creek fault is marked by well-foliated rocks, but the style of faulting is uncertain. Although tentatively interpreted to be a thrust fault (Stewart et al., 2016), additional kinematic work is recommended.

Garnet from the regionally metamorphosed rocks in the area northwest of Stibnite yielded a Lu-Hf age of 112.8 ± 7.2 Ma (Fig. 13A). Significantly older than the intrusive rocks in the area, these garnets yield ages similar to that obtained by McKay et al. (2017) on garnets in the Riggins area, 100 km to the northwest of Stibnite. It corresponds to the hypothesized second-stage ~124 to 112 Ma metamorphic event in the region (McKay et al., 2017) and suggests that the accretionary orogenesis, well documented to the west, may have extended east to the Stibnite area. Garnet from the nearby felsic dike yields an age of 99.3 ± 2.0 Ma (Fig. 14A). This age indicates that some of the felsic dikes in the area are older than the dated leucocratic granite at the Yellow Pine pit, and they predate most if not all the larger intrusive masses in the area.

The wide variety of Cretaceous intrusive phases sets the Stibnite area apart from many areas of the Idaho batholith to the south and west, where large masses of either two-mica granite or biotite granodiorite are common (Fisher et al., 1992). Small amounts of quartz diorite and leucocratic granite, presently unmapped, and larger masses of two-mica granite, biotite granodiorite, and porphyritic biotite granodiorite are intimately intermixed in the Stibnite area. Geochronologic and geochemical work by Gillerman et al. (2019), Box, Wintzer et al. (in press), and Lund et al. (2023) adds to our knowledge of this complex part of the batholith. These Cretaceous plutonic rocks in the Stibnite area are an important ore host, particularly those with a high original mafic mineral content that served as a nucleation site and iron source (Cooper, 1951; Gillerman et al., 2019).

The Eocene Thunder Mountain caldera complex consists of multiple calderas that are likely to have erupted from this region (Fig. 3). Although exact extents are uncertain, these calderas are marked by particularly thick accumulations of welded tuff units. The stratigraphy developed by E. Bart Ekren and Richard Hardyman for the Challis 1° × 2° quadrangle southeast of the Stibnite area (Fisher et al., 1992) can be extended north into the area east and northeast of Stibnite, and four primary units that they defined are recognized (dime and quarter lapilli tuff, buff rhyolite, lower Sunnyside tuff, and upper Sunnyside tuff). Ages for these and additional subordinate volcanic units are given in Box, John et al. (in press).

The Meadow Creek and West End faults are the two most important structures for mineralization in the Stibnite area (Fig. 8). In the southern Stibnite quadrangle, the Meadow Creek fault strikes north-south and is the controlling structure for mineralization at the Meadow Creek mine, most of which is immediately east of the fault. Where the Meadow Creek fault bends to the northeast at the Yellow Pine pit, mineralization is extensive, particularly west of the fault (Cooper, 1951). Although the age of movement on the Meadow Creek fault is poorly constrained, considerable motion must postdate emplacement of the Cretaceous plutonic rocks. The West End fault is subparallel to the Meadow Creek fault (Fig. 8). We propose that the West End and Meadow Creek faults are bounding faults of a right-lateral strike-slip fault system and that initial movement was on the West End fault. Movement then shifted sequentially northwestward to new fault strands. Final movement was on the Meadow Creek fault.

Funding from the Payette National Forest for work in the Edwardsburg area and from Midas Gold Inc. for work in the Stibnite area is gratefully acknowledged. Mark Schmitz oversaw the detrital zircon analyses completed by Isakson at Boise State University and Darin Schwarz assisted with detrital zircon analyses at Washington State University under the direction of Jeff Vervoort. Reviews by Joe Colgan, Scott Giorgis, and Steve Box decidedly improved the manuscript and are greatly appreciated.

Reed Lewis is a research geologist with the Idaho Geological Survey in Moscow. He holds geology degrees from the University of Idaho (B.S), University of Washington (M.S.), and Oregon State University (Ph.D.). His experience includes gold exploration with the U.S. Geological Survey Saudi Arabian Mission, contract geologic mapping for the Montana Bureau of Mines and Geology, and more than 40 field seasons of geologic mapping in Idaho for the U.S. Geological Survey and the Idaho Geological Survey. His most recent research emphasis is on the Belt Supergroup, basement rocks, and lower Paleozoic and Neoproterozoic strata.