e-mail: georgeo.g@gmail.com

Early Millennium drilling activity in the Bakken and Three Forks formations of the Williston Basin and “Alberta Bakken” (Exshaw and Banff formations) led to renewed interest in type and reference sections of the Devonian Jefferson, Three Forks, and Sappington formations at Logan Gulch, near Three Forks in Montana (Adiguzel et al. 2012; Doughty et al. 2014a, 2014b). PRISEM Geoconsulting described over 150 outcrops of Devonian–Mississippian strata in western Montana and has hosted field trips there since 2004 (e.g., Doughty and Grader 2010). These stratigraphic sections and their subsurface correlatives were deposited on a carbonate platform flanked to the west by a developing foredeep related to Antler Orogeny. Differential accommodation and facies changes occurred along a Late Devonian hinge zone near the Idaho–Montana border— the Lemhi Arch—which separated the craton and Central Montana Trough from the subsiding Idaho shelf. A shelf break was located near Challis, Idaho, in the Grandview Canyon area.

Low-accommodation cratonic stratigraphic sections and their disconformable surfaces in Montana preserve part of the Late Devonian stratigraphic record. Correlative down-dip strata in Idaho are thicker and more like western Canada as they contain a more complete record of sedimentation (Figs. 1, 2). We have measured or reevaluated over 50 locations, mainly in east–central Idaho, with a focus on the upper Jefferson Formation. This Idaho data set is correlated with western Montana and the Logan Gulch type section (i.e., Jefferson Birdbear Member and Three Forks Formation correlatives). This study (1) evaluates various members of the latest Frasnian and Famennian Jefferson Grandview Dolomite (Ross 1947, Sandberg et al. 1989) and (2) develops a new stratigraphic framework for east–central Idaho that integrates shelf-to-basin stratigraphy and common sequence stratigraphic surfaces formed in differing accommodation settings.

Devonian stratigraphy in western Montana was first studied by Peale (1893), Haynes (1916), Berry (1943), Sloss and Laird (1947), Sloss and Moritz (1951), and Andrichuk (1965) and developed for the Jefferson and Three Forks formations (Holland 1952; Achauer 1959; Gutschick et al. 1962; McMannis 1962; Rau 1962; Sandberg 1962, 1965; Benson 1966; Sandberg et al. 1972; Smith and Dorobek 1989; Dorobek 1991; Di Pasquo et al. 2012; Schietinger 2013). Early stratigraphic studies of Devonian rocks in Idaho included work by Ross (1934, 1947, 1961), Sloss (1954), Scholten (1957), Scholten and Hait (1962), Churkin (1962), Skipp and Sandberg (1975), Loucks (1977), McFaddan et al. (1988), Isaacson and Dorobek (1989), Sandberg et al. (1989), and Grader and Dehler (1999). Devonian through Mississippian rocks were deposited within the Idaho Antler orogenic foreland (Dorobek et al. 1991, Batt et al. 2007, Isaacson et al. 2007) and were subject to unconformities similar to those in Nevada (Silberling et al. 1995, Morrow and Sandberg 2003, Trexler et al. 2004). US Geological Service (USGS) researchers discussed outcrops, biostratigraphy, and paleogeography in multiple publications, although sections in the Lemhi and Lost River ranges received less study (Maple and Sandberg 1968, Sandberg and Poole 1977, Sandberg et al. 1983; Fig. 2).

At the Logan Gulch Devonian type and reference sections in Montana, the Frasnian Jefferson Birdbear Member is 22 m thick. The top of this unit may be a very coarse, bleached white dolostone. The overlying evaporite dissolution breccia of the Famennian Logan Gulch Member of the Three Forks Formation was deposited above a major sequence boundary in restricted environments. It is 34 m thick, but it was 40 to 50% thicker before dissolution of evaporites. The time-correlative upper Jefferson Formation in Idaho is much thicker (Fig. 1). The Birdbear Member and Logan Gulch Member overlie about seven third-order, primarily carbonate sequences of the 180-m-thick Lower Jefferson Member at Logan Gulch (Dorobek 1991; PRISEM, unpublished data). These rocks are time-correlative to the Canadian Fairholme Group, Alexo, and lower Palliser formations of British Columbia and Alberta (Seward 1990, Johnson et al. 2010; Fig. 2).

In Idaho, the mixed siliciclastic and carbonate rocks of the Frasnian and older lower Jefferson D1–D3 members are 392 m thick in the central Lemhi Range; Frasnian and younger upper Jefferson D4 through D6 members are 512+m thick (Hait 1965, Grader and Dehler 1999). This is comparable to thicknesses in parts of the northern Lost River Range. The upper Jefferson Grandview Dolomite is 357 to 320 m thick in the Grandview Canyon area, and in the Borah Peak area this unit is 645 m (Ross 1934, 1947; Wiler 1992; this study). Latest Frasnian–lower Famennian lithostratigraphic D4 through D6 Jefferson members and Grandview Dolomite were deposited in depocycles IId and IIe (Johnson et al. 1985). Early regional arch-to-shelf and craton-to-basin cross sections show interpretations of the stratigraphic geometry of these units in Idaho and western Montana (Figs. 2, 3).

Before the shelf margin broke apart during latest Devonian time, the shelf break and associated bioherms or platform margin grainstone generally moved eastward in the Early and Middle Devonian with rising Kaskaskia sea level (Johnson et al. 1985). These rocks include the lower Jefferson Formation in Idaho and units to the west (Fig. 2). Shelf paleogeography changed during deposition of the Grandview Dolomite. The sandy Grandview Dolomite was first tentatively correlated with the Birdbear Member and part of the Lower Jefferson Formation of western Montana as well as the Bierdneau Formation in southeastern Idaho (Sloss 1954, Benson 1966, Beus 1968, Walters 2012). Correlative Late Devonian sandstone and carbonate in Utah and Nevada include third-order sequences in the upper Guilmette Formation and West Range Limestone (LaMaskin and Elrick 1997, Giles et al. 1999). All of these units were considered part of a larger depositional tectonic unconformity-bound depocycle representing open marine flooding of the Cordilleran miogeocline (Silberling et al. 1995).

Lemhi Range Jefferson D4 Through D6 Mapping Units and the Grandview Dolomite: The Jefferson Formation in the Lemhi Range and Beaverhead Mountains was subdivided into lithostratigraphic members D1 through D6 for mapping purposes (Scholten and Hait 1962, Hait 1965). These units reflect natural sequence stratigraphic breaks and depositional stacking patterns (Grader 1998). Distinctive dark gray to black Euryamphipora beds and common biostromal coral–stromatoporoid intervals and buildups with Amphipora occur in D1 to D5 members (Grader 1998, Grader and Dehler 1999). The last occurrence of these latest Frasnian fossils is in the dark marine to lagoonal rocks of the basal D5 member (Figs. 3A, 4). Because of regional faults and difficult access, these units and their relationship to the Jefferson Formation in Montana and the western Grandview Dolomite in the Lost River Range remain uncertain.

The Grandview Dolomite was defined by Ross (1934) at Grandview Canyon for 357 m of light-colored dolostone and dolomitic, sandy limestone, with very dark gray to black beds, similar to the Jefferson Dark Dolomite below. This formation thinned to the southern Lost River Range and was later identified in the central and northern Lost River area at Borah Peak and at Grouse Creek, where it was significantly thicker (645 m in Ross [1947]). The Grandview Dolomite was considered an upper member of the Jefferson Formation (Sloss 1954, Benson 1966). It was treated as a local Jefferson facies by Mapel and Sandberg (1968), who mapped the Montana Birdbear Member into Idaho, only to discover later that the Montana Birdbear was much older (Sandberg and Poole 1977). Therefore, the Idaho Birdbear became the “False Birdbear,” appearing in chronostratigraphy charts as an upper member of the Grandview Dolomite overlying a lower “Shallow Water Dolomite” member (Johnson et al. 1985, Sandberg et al. 1989).

Loucks (1977) suggested that the Grandview barrier facies can be mapped as a thick dolomitized body through parts of Utah, Nevada, and Arizona and are part of the problematic western USA Frasnian and Lower Famennian depositional complex of Sandberg and Poole (1977). As a narrow linear band, it constituted barrier to marine circulation and sediment transport, demonstrated best in Utah by the high percentage of sand east of it compared with the low percentage of sand west of it (Loucks 1977). Shelf-to-basin relationships suggested that the sandy and cyclic Grandview Dolomite represented an aggradational dolomitic barrier facies and a distinct shelf margin depositional system above the western Dark Dolomite (D3) and was correlative to the D4 through D5 units extending to the Lemhi Arch (Fig. 3A).

D4 limestone beds contain a widely correlated black and white “stromatactis” cavity-like fabric marker bed in the basal Jefferson Formation reaching the Beaverhead Mountains. This implies initial eastern submergence and onlap of the Lemhi Arch during D4 deposition (Scholten and Hait 1962). Figure 3B shows late Frasnian Jefferson Birdbear Member (or “Nisku”) capping the arch under the Three Forks Trident Member. However, this Birdbear Member is younger in Idaho than in Montana and is therefore “false,” as explained above. Additional corrections to early stratigraphic conceptions shown in Figure 3B include (1) the mainly Frasnian age of the Dark Dolomite (not Givetian; Johnson et al. 1985) and (2) the fact that the evaporitic Logan Gulch Member of Montana is correlative with part of the upper Grandview Dolomite in Idaho.

Grandview Dolomite and Tales of Sandstone from the West:Figure 3B shows a strong unconformity west or outboard of the Grandview area. Grandview Dolomite sedimentation was thought to be connected to outer shelf uplift further west and was associated with syndepositional sandstone provenance polarity switch toward the east, onto the Jefferson shelf (cf. Sandberg et al. 1975, Wiler 1992). However, late Mesozoic uplift and Tertiary cover also produced and obscured much of this western unconformity (Skipp and Brandt 2012). There is little evidence for shedding of siliciclastics onto the Jefferson carbonate shelf from the west, although overlying recycled quartz arenite and Jefferson-derived conglomerate of the western Picabo Formation indicates latest Devonian outer shelf tectonic activity (Skipp and Sandberg 1975).

The western unconformity and the Idaho Batholith precludes tracing the Grandview Dolomite barrier system further west, although structural windows exposing Devonian strata suggest this unit is absent (Skipp and Brandt 2012). Thrust fault juxtapositions and elements of this unconformity served as evidence for an off-shelf Antler-age Ancestral Milligen Ridge uplift (Sandberg et al. 1975). Models of a Late Devonian flexural forebulge with a sediment-trapping backbulge basin, as in Nevada, have been applied to the Jefferson Formation (cf. Goebel 1991, Wiler 1992). Likewise, differential subsidence and sediment accommodation over a brittle deforming foreland were postulated by Dorobek et al. (1991). Regional Late Devonian disconformities associated with basinward facies shift and tectonic, forced regressions are widely considered for coeval strata in Nevada and Canada (Giles et al. 1999, Johnson et al. 2010); glacioeustatic controls are also proposed (LaMaskin and Elrick 1997, Sandberg et al. 2002).

The paleogeographic setting for Devonian rocks is complicated by the Mesozoic faulting that propagated through offshore Ordovician to Devonian shale and Late Paleozoic syntectonic turbidites in the Pioneer Mountains (Link et al. 1995; Fig. 2). Thrust fault repetitions or omissions of strata in overthrust and normal faults result in substantial but not insurmountable facies stacking and isopach questions. Imbricate thrust faults and reactivated normal faults in the study area are thought in some cases to thin the Jefferson Formation (Mapel and Sandberg 1968). Faults also locally follow subsurface Famennian evaporites, resulting in complexly overprinted carbonate megabreccias in the upper Jefferson Formation (M’Gonigle 1982).

Facies change abruptly, but conodont biostratigraphy is continuous between the continental rise and active basin shale environments of the Milligen Basin to the west and the eastern carbonate platform environments of the Jefferson Formation (Sandberg et al. 1975, Turner and Otto 1988; Fig. 2). No shale accumulated on the carbonate shelf until later Famennian deposition of the Three Forks Trident Member and Sappington formations during diachronous evolution of the Euramerican west coast from passive Devonian to active Mississippian margin (Trexler et al. 2004). Sand-sized siliciclastics accumulated on the shelf with thick carbonates in the Early Famennian Central Idaho Trough (Grader and Dehler 1999). In the Middle Famennian, sandstone deposition continued in the Wood River Area to the west, and seaway shale deposition occurred in the Central Montana Trough to the east. During pre-Mississippian Antler-induced subsidence and basement faulting, a hiatus developed over the former Idaho Jefferson shelf. In the same area of early onlap onto the Lemhi Arch, the Beaverhead Mountains uplift inverted and reactivated, resulting in complex facies and isopach relationships (Dorobek et al. 1991; Fig. 5).

New Palinspastic Isopach Maps: Palinspastically restored pre-Famennian and Famennian regional isopach maps emphasize Devonian paleogeographic elements in Idaho and Montana (cf. Sternes et al. 1997; Fig. 5). Locations are essentially para-autochthonous in Montana, compared to the significant eastward telescoping noted across the Idaho–Montana border, over the Medicine Thrust. The shelfal strata discussed below, with locus in the Central Idaho Trough, remained on a single structural panel—the Hawley Mountain Plate—carried east in the fold and thrust belt (Link and Janecke 1999). The intrashelf Central Idaho Trough replaces the “foredeep,” as labeled on older Devonian isopach maps (e.g., Sandberg et al. 1983). A broad carbonate factory extended to the west with evaporite deposition on the shelf, before deposition of Three Forks shale, Mississippian inversion, and active basin sedimentation.

Nomenclature disambiguation and genetic solutions began with observations that the thick D4 through D6 Jefferson members could be applied to the Grandview Dolomite in the Borah Peak area. The late Frasnian D4 member and basal D5 members comprise the lower part of the western Grandview Dolomite known as “Shallow Water Dolomite.” These units record sequences similar to those of the Canadian Nisku Formation (Potma et al. 2001) and are part facies and time-correlatives of the thinner latest Frasnian Birdbear Member of Montana (Figs. 1, 2). The western upper Grandview Dolomite is equivalent to the eastern upper D5 and D6 Jefferson members and to the Logan Gulch Member of the Three Forks Formation in Montana. We disassociate the upper Grandview Member from its subordinate False Birdbear member, instead defining an additional thin sequence on top of the Grandview Dolomite and underneath the Three Forks Trident Member. For these rocks, we retain the “False Birdbear” name.

Stratigraphic correlation is based on 50 partial stratigraphic sections that tie Idaho strata to the Montana Logan Gulch type section (Table 1; Fig. 4). Isaacson and DeSantis describe reef geometry and biofacies near Grandview Canyon, in the top of the lower Jefferson Formation (DeSantis 1996). Grader and DeSantis pursued high cirque sections in the northern Lost River Range, also revisiting earlier student-established sections and tracing sequence boundaries into the Borah Peak Area (central Lost River Range). Twenty-four new locations and stratigraphic sections were studied, many in the Lost River Range. These were tied to Lemhi Range sections (Grader 1998) and six wellknown sections in Montana (Table 1). This study focuses on 12 Idaho stratigraphic sections, which update and correct earlier adopted upper Jefferson nomenclatures (Grader and Dehler 1999). Although some in-house conodont work was done at the University of Idaho, we have populated key stratigraphic sections with the locations and stage interpretations of earlier studies. We use the contact with the Middle Famennian Three Forks Formation shale (“trachytera” Zone) as the main datum for stratigraphic sections along east–west transects. Late Frasnian sequence boundaries are labeled lFr0 to lFr2; Famennian sequence boundaries are labeled Fm0 to Fm7 (Table 2).

We reevaluate and compare the Grandview Dolomite with the Jefferson D4 through D6 members. Dark- and light-colored, laminated to bioturbated, variably dolomitized carbonate and mixed sandstone facies of the Jefferson Formation repeat through time, so it can be difficult to interpret in the fold and thrust belt. Data include hundreds of polished hand samples and 50 representative thin sections. Most of these facies are lighter colored peritidal to darker subtidal rocks. Correlation of the upper Jefferson Formation is shown in Figures 4, 6 and 7, with additional data in Pl. 1, 2, and Appendix 1.

Students working the upper Jefferson Formation described but misinterpreted some of these very thick stratigraphic sections, often in structurally complicated areas (Simpson 1983, Wiler 1992). Present work clarifies early work on upper Grandview Dolomite facies associations and the relative age of the Frasnian and Famennian D5 member in the Lemhi Range; D5 was previously tangled in old “False Birdbear” nomenclature (e.g., Grader and Dehler 1999; Fig. 2). Recent field work has discovered and reinterpreted the extent and limits of the False Birdbear and what are likely latest Frasnian Nisku patch reefs (not Famennian reefs). The latter are laterally dispersed within continuous very dark gray to black dolomitic limestone of the basal D5 member under the datum Fm0 (Figs. 4, 7).

Regional chronostratigraphy of Devonian rocks is summarized in Figure 2 and correlated in Figure 4. Incised channel systems and Lower and Middle Devonian carbonate ramp lithofacies flank the Lemhi Arch, below biostromes in the Frasnian Dark Dolomite. The Frasnian–Famennian stage boundary is estimated at the Fm0 sequence boundary in the D5 member, and the Famennian Three Forks Formation Trident Member is overlain by a merged basal Mississippian sequence boundary and flooding surface. The upper Jefferson Formation Grandview Dolomite in Idaho above the Dark Dolomite is predominantly limestone and dolostone and varies from ~300 m thick to the west to over 500 m thick in the Lost River and central Lemhi Ranges (D4–D6 members), thinning to 70 m or less to the east across the southern Lemhi and Beaverhead Mountains (Hait 1965, Grader 1998). Grandview Dolomite strata are simplified in Figure 4 following more detailed explanation and correlation below.

The Devonian succession in the northern Lost River Range and Borah Peak area shows the relationship of lithostratigraphic members D1 through D6, synonymous nomenclature, and proposed sequence boundaries (Fig. 6; Table 2). Common carbonate megabreccia beds with more rarely associated thin red siltstone and pseudomorphs after evaporites are partly interpreted as evaporate–tectonic “rauhwacke” (sensu Bruckner 1941 in M’Gonigle [1982]; Fig. 6A). Most Jefferson Formation breccias, including those in the Logan Gulch Member of the Three Forks Formation at the Montana type section (Fig. 1), are thought to be associated with syndepositional and/or late evaporite dissolution. Upper Jefferson units, especially the upper D5 to D6 members, comprise restricted marine environments that are very thick compared with the correlative Logan Gulch Member in Montana. Yet comparable Devonian breccias occur in the Rocky Mountain front near Choteau, correlative to the thick subsurface Potlatch evaporites of northwest Montana (Fig. 5B). Similar thick interbedded stromatolitic carbonate and breccia of the upper Jefferson Formation in the Lemhi and Lost River Range are interbedded with dark marine subtidal carbonates with rare macrofossils.

Jefferson Formation Sequences Above Late Frasnian Sequence Boundary lFr0 and Below Fm0: The Jefferson Dark Dolomite Member (D3) below lFr0 comprises dark subtidal ramp carbonate (completely shaded in Figs. 4, 7). Decameter-scaled sequences occur continuously across the study area in the latest Frasnian lower Grandview Dolomite interval above lFr0 (Table 2). The shaded units in Figures 4 and 7 represent organic-rich, mostly dark to black marine carbonate; these are subtidal lithologies simplified as transgressive systems tracts (TST). Regressive, restricted platform platform lithologies associated with low and high systems tracts (LST and HST) are not shaded. The lower part of the D5 member is very similar to the D3 member and contains Thamnopora coral, Amphipora stromatoporoid framestone–floatstone, and, commonly, oncoidal gastropod packstone. Likewise, there are dark units in D4, tentatively labeled TSTs that can be correlated from Liberty Gulch in the Lemhi Range to the Borah Peak area (over 50 km apart; Fig. 7). Palmatolepid conodont and Orbiculoidea-bearing calcareous shale of the overlying Three Forks Trident Member also are differentiated into TSTs in western sections, where they crop out well. These facies are thought to represent deeper environments associated with marine flooding events (Figs. 1, 6, 7).

Latest Frasnian third-order sequences in D4 are more symmetrical at Liberty Gulch compared with the asymmetric deepening-upward successions capped by dark biostromal carbonate that characterize older Givetian sequences (Figs. 4, 7). More homogeneous, relatively less cyclic sequences occur in the mostly subtidal D3 rocks (not defined here). Likewise, sequences are only tentatively defined in rare outcrops of the upper D5 to D6 members. Thinner black fossiliferous interbeds and sandy, peloidal to muddy dolostone beds define third- or fourth-order decimeter cyclicity in the western Grandview Dolomite above the Grandview Reef (Fig. 7).

Sequence boundaries between surfaces lFr0 through Fm1 are abrupt in the Lemhi Range and Borah Peak area and more conformable to the west (dashed in Fig. 7). Intense karstification and sandstone influx defines SB Fm0 and occurs directly above a Grandview Dolomite–basal D5 20-m-thick bioherm at Gooseberry Creek. The Gooseberry Reef is a latest Frasnian Nisku Reef equivalent (Figs. 6C, 7). Early work miscorrelated these rocks in the top of the Grandview Dolomite. Instead, we correlate these very dark, petroliferous rocks under the Fm0 sequence boundary to Grandview Hill and to the east. The Gooseberry Reef contains a buildup of Syringopora, Amphipora, Thamnopora, Peneckiella, silicified round, bulbous, thrombolitic and laminar stromatoporoid framestone–rudstone and gastropod packstone. Sandy grainstone and breccia of the Fm0 sequence boundary cross-cut the reef top, in the same way that sequence boundary lFr0 cross-cuts the Grandview Reef, lower in the section at Grandview Hill (Fig. 7). Karsted outer shelf biostromal intervals also occur even lower in the Devonian section on the outer shelf, for example, under SB G2 at Meadow Peak and up-section at Fish Creek, under the Picabo Formation (Fig. 4).

Meter-scale shallowing-upward cycles in the upper Jefferson Formation consist of a dark black bioturbated, partly fossiliferous, faintly laminated dolomudstone overlain by medium gray, traction-laminated dolomudstone with rip-up clasts and low-angle lamination truncations with soft sediment deformation or ripple laminations, overlain by very light gray, burrowed or stromatolitic, finely crystalline dolomudstone (Pl. 1, 2). The ideal meter-scale cycle consists of shallow subtidal to intertidal and less common supratidal environments, typically with sharp bedding contacts between lithofacies. Peritidal successions of the Jefferson Formation through Devonian time have admixed silt to medium sand (“deflation spheres”) and friable to indurated, well-sorted fine to medium sandstone with high-energy planar laminations and ripple to dune cross-bedding. Coherent meter-scale brecciated carbonate cycle caps occur above peritidal units and below bioturbated subtidal units; in some intervals, brecciation obscures cyclicity. The sequence boundary lFr0 marks the transition to an incipient active margin succession with both increased accommodation and sharp base-level changes.

Jefferson Formation Successions Above Sequence Boundary Fm0 and Below Fm4: Upper Grandview Dolomite Famennian successions between Fm0 and Fm4 or Fm5 are defined based on provisional vertical facies stacking. Rocks between sequence boundaries Fm0 and Fm1 represent the first accumulation of sandy, poorly fossiliferous peritidal facies following the black subtidal rocks of basal D5. Upper Jefferson strata in the central Lemhi Range and Borah Peak area thin to the west toward their stratigraphic equivalent, the Grandview Dolomite. Upper Grandview Dolomite strata at Gooseberry Creek include ~200 m of light and dark thinly bedded peritidal cycles, with common laminated and bioturbated, coarsely dolomitized, and muddy lithofacies with stromatolites. Silty carbonate breccia and cross-bedded quartz arenite beds occur throughout this part of the section, as seen in the Grandview area and Lemhi Range (Figs. 7, 8). However, thicker accumulations of these wave- and current-generated facies in the Lemhi Range represent a shift to shoreface and beach environments—for example, at Cedar Run. Few truly black marine carbonate beds occur above the Frasnian–Famennian disconformity—Fm 0—except as dark gray bioturbated mudstone or wackestone in meter-scale cycles.

Lemhi Range Famennian strata above Fm0 thin to the southeast and have fossil-poor, shallow water sandy, stromatolitic to laminated facies associations (Fig. 7). Meter-scale peritidal cycles occur where they are not destroyed by karst or evaporite collapse brecciation effects. Partly time-equivalent upper Grandview Dolomite facies to the west are similar but thinner, with fewer massive evaporite collapse breccias (i.e., D6). These highly laminated rocks are capped by thick outcrops of the Trident Member, which includes local basal sandstone. Higher energy, more peloidal grain-rich western, lower Grandview Dolomite facies with high-frequency transgressive units suggest an inherited, distal shelf shoaling bathymetry (Fig. 8). These depositional environments and isopach patterns seaward of an intrashelf lagoonal subbasin and shoreward of the Milligen Basin represent a carbonate platform edge. These features also may reflect foreland accommodation effects and local hinges (Fig. 9).

Breccias of the D6 member (and parts of the D4 member) are interpreted as remnants after evaporites, similar to the Three Forks Logan Gulch Member of Montana (M’Gonigle 1982). Regionally, strata below the Three Forks Trident Member contain peritidal carbonate shallowing-upward cycles with red shale–siltstone. D6 breccias contain angular stromatolite-bearing clasts and remobilized carbonate and silt matrix (Grader 1998). They are not as thick at Borah Peak, are thin at Gooseberry Creek, and are unlike highly laminated rocks in the Grandview Canyon area (Unit VI cf. Wiler 1992; Fig. 7). Medium to coarse siliciclastic beds occur near the top of the Grandview Dolomite in Grandview Canyon and are interpreted as the base of the overlying False Birdbear unit, defining a sequence boundary involving Fm4 (Pl. 2). Associated argillaceous mudstone overlain by fossiliferous nodular packstone occurs between sequence boundaries Fm4 and Fm5 and represents a different package of rocks (Figs. 4, 7, 8). These rocks were previously included in the upper Grandview Dolomite “False Birdbear Member” at Lower Cedar Creek and Freighter Springs (Sartenaer and Sandberg 1974) and in “Unit VI and the conodont-bearing encrinite bed” at Grandview Hill (Wiler 1992).

Post–Jefferson Formation Crinoidal “False Birdbear” Bench Under Three Forks Trident Member: Famennian Jefferson strata in Idaho lack shelly bioclasts, are dominated by stromatolites, and are difficult to date. They are different from early Middle Famennian open marine limestone locally preserved below discontinuous sandstone and marine shale of the Trident Member. Referred to as the “marginifera” Zone–bearing Birdbear by Sartenaer and Sandberg (1974) and Sandberg and Poole (1977) and as the False Birdbear by Johnson et al. (1985), this nomenclature was overapplied to include much of the upper Grandview Dolomite (Fig. 2). Previous descriptions of this limestone bench correlate it to the upper Jefferson Formation, but these rocks are more similar to the Three Forks Trident Member, especially at Freighter Springs and Long Canyon. The restricted Grandview Dolomite and evaporitic D5 to D6 depositional system was followed by deposition of a relatively thin, distinctive sequence we continue to refer to as the “False Birdbear.”

The False Birdbear, as defined here, has variable lithologies that are either absent, covered, or are not shared at other locations. The unit comprises coarse sandstone grading into calcareous shale overlying an abrupt contact with the Grandview Dolomite on the west side of Grandview Canyon. At Freighter Springs, a 9 m, recessive, medium gray, muddy, calcareous interval overlies karsted Jefferson Formation. This unit is interpreted as a subtidal unit. It is in turn overlain by a prominent light gray, nodular–crinoidal lime packstone representing a high-energy, open marine environment sharply overlain by the tan Trident Member shale (Fig. 7). At the locations in which this muddy unit is not recognized (e.g., Gooseberry Creek), red silt-filled, karst-brecciated Grandview Dolomite with silicified box-work directly underlies the Trident Member (Fig. 8). The False Birdbear is confined to the interval between sequence boundaries Fm 4 and Fm 5 (Fig. 10). The unconformity-bound, open marine facies of this unit in Idaho may correlate to a widely deposited similar stratigraphic interval in Montana at the top of the Logan Gulch Member underlying the “marginifera and trachytera” Zone Trident Member (i.e., “Knoll Limestone”; Doughty et al. 2014). However, the outcrop characteristics, biota, and limestone facies of these units are different from those of the Knoll Limestone, which has not yet been successfully dated.

The “marginifera”-aged False Birdbear at Long Canyon in the Beaverhead Mountains (Sandberg and Poole 1977, Sandberg et al. 1983) contains large cephalopods similar to the cephalopod-bearing Costigan Member of the Palliser Formation, Alberta (Johnston et al. 2010). At this location, the unit looks different than the Lost River Range False Birdbear interval, and it occurs directly below the “expansa” Zone Sappington Formation (Sandberg and Poole 1977; Fig. 7). Although this limestone bench is older, the open marine facies and stratigraphic position matches that of the younger (“trachytera”) “Trident Limestone,” which widely occurs under the Sappington Formation in Montana (Figs. 1, 10). Regionally, nodular packstone at the top of the Trident Member may look similar to False Birdbear lithology (Pl. 2).

Conodont, brachiopod, and vertebrate biostratigraphic data points for the upper Jefferson and Three Forks formations are shown in Figures 2, 4, 7, and 8 (and references therein, e.g., Sartenaer and Sandberg 1974; Sandberg et al. 1975, 1983; Skipp and Sandberg 1975). Conodont data in the Lost River Range (Wiler 1992) indicate zonally indeterminate Frasnian species of Polygnathus and Icriodus that are not correlative with deepwater palmatolepid-based zonations of Ziegler and Sandberg (1990). The “rhenana” Zone was identified at Hawley Mountain (Mapel and Shropshire 1973, Sandberg et al. 1983) based on Palmatolepis semichatovae in what was interpreted as the Dark Dolomite Member. Palmatolepis semichatovae also occurs near the base and top of the Jefferson Birdbear Member at the Jefferson type locality at Logan Gulch (Sandberg et al. 1983; Figs. 1, 4, 7, 8). The late Early “rhenana” Zone “seimicatovae rise” we use in Figure 10 follows the base of the T–R cycle IId of Johnson et al., although this may conflict with other Canadian data (Weissenberger, personal communication, 2016).

The older Birdbear Member data, combined with younger “marginifera” conodonts discovered in what later became known as the “False Birdbear” at Long Canyon in the Beaverhead Mountains, show that the Beaverhead Mountains uplift was a key control on sedimentation (Sandberg et al. 1983). An early and late “marginifera” palmatolepid biofacies from the later-renamed False Birdbear member in the Lost River Range occurs below late trachytera Zone conodont (Scaphignathus subsurratus–Pelekysnathus inclinatus fauna) from the middle and upper part of the Trident Member of the Three Forks Formation (Sartenaer and Sandberg 1974, Fox 1985). Conodonts recovered from 1 m above the base of the Trident Member at Grandview Hill included Palmatolepis marginifera utahensis and suggest a late marginifera age (Wiler 1992). Field work suggests that crinoidal limestone under these rocks are associated with the widely identified False Birdbear interval.

Three Forks Trident Member localities containing ~80 m of calcareous mudstone, shale, abundant rhychonellid brachiopod beds, Cyrtospirifer, crinoidal packstone, and nodular limestone were previously mapped in four units through the Lost River Range (Fox 1985). These are subdivided into systems tracts, with plotted conodont sample points, in Figures 7 and 8. The systems tracts are less easily mapped in the Lemhi Range and thin to a feather edge on the Lemhi Arch. We interpret the underlying False Birdbear marginifera-bearing nodular to encrinite limestone benches as part of an unconformity-bound sequence, as discussed above. Until correlation of these rocks to the craton is better understood, we retain the name False Birdbear for the sequence.

The stratigraphic interval of interest in this article spans the Frasnian–Famennian boundary from the top of eustatic depocycle IIc through IIe of Johnson et al. (1985, 1991; Figs. 1, 2). This interval begins in the upper D3 member of the Jefferson Formation and spans the Three Forks Formation Trident Member. Depocycle IId involves three late Frasnian sequences in Idaho between sequence boundaries lFr0 and Fm0, including the D4 and lower D5 Jefferson members and western Grandview Dolomite. These sequences thin onto the Lemhi Arch and craton (Figs. 7–10). Facies associations and dolomitized meter-scaled Jefferson Formation cycles are similar to those in Montana and North Dakota (Kissling and Ehrets 1984, Wilson and Pilatzke 1987, Smith and Dorobek 1989). The False Birdbear unit, a prominent crinoidal limestone bank previously included in the top of the Jefferson Formation above “unnamed yellow shallow water dolomite” (Johnson et al. 1985), is interpreted as a younger depositional sequence under the Trident Member shale datum. This is summarized in Figures 8 through 10 and highlights both the stratigraphic complexity and the usefulness of mapping erosional sequence boundaries.

Type 1 Late Devonian sequence boundaries show incision or unusual subaerial exposure of subtidal rocks (Sarg 1988, Van Wagoner et al. 1988, Lehrmann and Goldhammer 1999). Good examples of these occur in Givetian or Frasnian–Famennian sequences and can be traced over hundreds of miles in Idaho (Fig. 4; Table 2). Deci- to decameter sequences with sequence boundaries, flooding surfaces, and systems tracts are marked by non-Waltherian vertical facies transitions or are interpreted from estimates of stratigraphic position (Fig. 7). Upper Jefferson early transgressive systems tracts are associated with episodic siliciclastic input, dolomitization below them, and evaporites (with late brecciation, cf. M’Gonigle 1982). As in earlier Devonian LSTs, significant quartz sandstone was deposited along the Lemhi Arch flank. These deposits occur with peritidal cycles, including partly eolian-contributed sand and silt from the craton. Mature quartz sandstone and dispersed deflation spheres in carbonate lithofacies occur in the bottom and top of sequences (LST–HST), but not in the darker, muddier transgressive middle interval (TST).

The most consistently subtidal depositional cycles of the Jefferson Formation occur in the Dark Dolomite (D3). If not bioturbated or biostromal, dark featureless dolomudstone in D3 through D5 has laminations with low-angle truncations formed by low-energy currents in lagoonal to more distal ramp “rhythmite” facies. These lithofacies and their bioturbated, fossiliferous correlatives suggest deposition on wide, low-angle ramps. Carbonate ramp geometry changed abruptly with deposition of shoaling peloidal, intraclastic rocks of the lower Grandview Dolomite (Fig. 7).

Thick latest Frasnian lower Grandview Dolomite strata on the Idaho shelf (late “rhenana” and “linguiformis” zones) result in average undecompacted sedimentation rates of between 60 and 200+ m/my (assuming this interval lasted 1.7 to 2 my; Sandberg et al. 2002, Kaufmann 2006). The upper Grandview Dolomite above Fm0 has subsidence rates of about 33 to 55 m/my, and the shale–limestone cycles of the False Birdbear sequence and Trident Member represent rates of about 25 m/my (Fig. 9). These estimated rates of sedimentation are generally higher than those for underlying passive margin strata.

Similar patterns of subsidence, coastal onlap, and Pilot Shale sedimentation followed by relative Famennian quiescence were interpreted as an initial Antler flexural pulse along the Nevada–Utah margin (Giles and Dickinson 1995). Two sequences were defined in this area and interval by Morrow and Sandberg (2003). Three Guilmette Formation IId depocycle sequences and one in the base of IIe occur under the base of the West Range Limestone in the Egan and Schell Range of Nevada (LaMaskin and Elrick 1997). In Alberta, three sequences of the Southesk Formation equivalent to sequences in the Winterburn Group occur in the IId depocycle (Whalen et al. 2000, Potma et al. 2001). Although there are outstanding conodont range discrepancies (Weissenberger, personal communication, 2016), these western Laurentian sequences are approximately correlative with sequences lFr1 through lFr3 in the D4 and lower D5 members in Idaho (Fig. 8). It remains unclear where the earliest Famennian “triangularis” Zone occurs or whether quartzose sandstone near the Frasnian–Famennian boundary was shed from the west or east. As on the Idaho shelf, the Canadian sequences partly underlie sandstone-bearing rocks of the Graminia and Sassenach formations (Mountjoy and Becker 2000).

Gradually back-stepping to aggradational Frasnian D3 rocks includes deposition of the Leduc Formation–equivalent Grandview Reef on the outer Jefferson shelf during the Early rhenana Zone. Brief drowning of the Grandview coral–stromatoporoid buildup near the top of D3 (McFaddan et al. 1988) was followed by subaerial exposure and deposition of the first of three latest Frasnian sequences (Fig. 8; Pl. 1). The disconformity lFr0 subdivides a lower, first-order lithostratigraphic Siluro–Devonian succession from an upper, latest Frasnian–Famennian succession (simplified in Fig. 9). D4 and D5 member stratigraphic stacking patterns reflect second-order global sea-level maximum, turn-around, and regional switch from a passive to transitional to active margin (Figs. 9, 10). Glacioeustasy controls are generally more accepted for the later Famennian but also influenced Frasnian deposition (Johnson et al. 1985, Sandberg et al. 2002, Haq and Shutter 2008, Isaacson et al. 2008, McGhee 2013).

Three Latest Frasnian Lower Grandview Dolomite Sequences (D4 to D5 Members): Latest Frasnian transgressions and regressions are interpreted from three thick sequences in the D4 and lower D5 members (lFr1–lFr3 in Figs. 1, 8). An aggradational–progradational wedge built away from the Lemhi Arch, with accumulation of peritidal and sandy shoal deposits across the shelf. Early local and late removal of evaporite and collapse breccia resulted in obscured stratigraphic relationships, but up-dip sections near the Lemhi Arch are consistently thin (Sloss 1954). The marker bed with geopetal cavity fabric in the TST of sequence lFr2 in the D4 member is consistently observed and supports late Frasnian foundering of the Lemhi Arch (Scholten and Hait 1962). The origin of the black and white cavity fabric is unknown but is identical to Frasnian facies described in Nevada by Warme and Kuehner (1998).

Abrupt transgressions affected much of the shelf and arch but were of limited extent on the craton. Hypersaline seas and evaporites filled a silled shelf during early transgression and included dysoxic black laminated carbonate (Kellwasser equivalents?). This coincided locally with reestablishment of oxygenated open marine buildups (e.g., Gooseberry Creek Reef). Time estimates for the late Frasnian sequences are either rapid third-order sequences (of ~500-Ka duration) or slower, ~450-Ka fourth-order sequences (sensu Sarg et al. 1999). Tectonic subsidence and overall rising sea level is inferred to have accommodated about 50-m-thick sequences in the central Lemhi Range. Incipient foreland effects overprinted global eustatic–climatic signals (Dorobek et al. 1991). Similar controls are proposed for time-equivalent sequences in Canada (Potma et al. 2001). Thin latest Frasnian transgressions covered the Lemhi Arch, and the widespread recessive to cliff-forming Birdbear Member of the Jefferson Formation in Montana was deposited as a single third-order sequence equivalent to Idaho sequence lFr1.

Without higher resolution biostratigraphic control, dating of Jefferson Formation sequences is uncertain. Previous workers used the highest occurrences of colonial and large solitary corals in shallow water, or Amphipora biostromes in peritidal settings, to approximate the end of the Frasnian in the absence of definitive conodont evidence (Geldsetzer et al. 1993). It is likely that these biostromes disappeared coincident with increased deepening within the linguiformis Zone (Sandberg et al. 2003; Fig. 10). The last Amphipora and coral–stromatoporoid Gooseberry Reef in Idaho occurs below Fm0 and the estimated F–F boundary.

Early Famennian Upper Grandview Dolomite Succession (D5 to D6 Members): Subsidence in the Lemhi Range and Borah Peak area accommodated thick upper D5 and D6 strata that are tied to thinner strata on the outer and inner shelf and to pre–Three Forks Formation unconformities. This geometry suggests simultaneous down-drop and localized uplift and nondeposition during the lower Famennian (Fig. 8). Successions of peritidal to more burrowed subtidal units and preserved lowstand shoreface to beach-barrier quartz arenite in the central Lemhi Range reflect relative accommodation between peloidal grainstone shoals on the distal shelf and uplift along the long-term hinge line/zone to the east. Siliciclastics were trapped against the Lemhi Arch (Sloss 1954) and subsequently active Beaverhead Mountains Uplift (Figs. 8, 9).

Sandstone in the Grandview Dolomite was previously interpreted as having been sourced from a poorly defined uplift to the west in the Pioneer Mountains (Sandberg et al. 1975, Wiler 1992). Thin sandstone and abundant peloidal grainstone facies in the western Grandview Dolomite are interpreted as part of cyclic, distal shelf barrier–shoal environments. Thick shoreface to beach sandstone in three interpreted Famennian sequences in the Lemhi Range above Fm0 suggest (1) long-term cratonic sand sources to the east and (2) entrapment in the mid–shelf basin during overall Early Famennian sea-level lowstands (Figs. 8, 10). Strata in the upper Grandview Dolomite to the west are tentatively correlated as time-equivalent, with more geographically confined intrashelf basin sandstone and evaporite solution breccias of the central Lemhi Range and Borah Peak area.

Western Unconformity and Famennian Intrashelf Basin and Offshore Siliciclastics: The prominence of Devonian carbonate depositional systems on a broad platform or seaward-dipping shelf is juxtaposed with and separated from deepwater shale and siliciclastic rock to the west. An abrupt transition to a western basin, as described in the Alberta Rocky Mountains during the late Frasnian (Whalen et al. 2000), is unknown. Occurrence of the Picabo Formation, absence of the upper Jefferson Formation and Trident Member west of the study area in the White Knob and eastern Pioneer Mountains, suggests western uplift between the Antler highlands and the Devonian shelf (Sandberg et al. 1975). Karst in the Jefferson Formation “Yellow Vuggy Dolomite” at Fish Creek supports this interpretation (Figs. 2, 4). Off-shelf rhenana to expansa Zone rocks of the Milligen Formation Sandstone Member of Independence represent deposits in a separate offshore basin (Turner and Otto 1988; see Fig. 2). Alternatively, these may be unrecognized stacked lowstand systems tracts outboard of the poorly defined shelf break (Fig. 9).

Siliciclastic storage on the shelf and bypass to the western Milligen Basin may have been expedited by progressive eustatic drawdown (cf. Johnson et al. 1985, 1991; Fig. 10). Outboard loading and intermittent tectonic uplift and down-drop of the outer continental margin during the Antler Orogeny may have affected Idaho, as was the case in Nevada, but tectonic paleogeography between Nevada and Canada likely varied considerably (Wilson et al. 1994). The Cove Fort Quartzite in Nevada involves three stacked, geographically confined beach barrier sandstone and lagoonal dolomudstone sequences controlled by forced regression during eustatic sea-level fall and flexurally induced high subsidence rates, focused along a cordilleran hingeline (Giles et al. 1999; Fig. 2). These strata are analogous to the stacked, lowstand sandstones deposited next to the Lemhi Arch in an intrashelf basin east of an eroded shelf break and area of less accumulation (Grandview Canyon and northern Lost River Range).

Late Devonian Sea Level and Idaho Sequences in a Global Context: Three latest Frasnian sequences in Idaho are very similar to the globally recognized IId depocycles of Johnson et al. (1985; Fig. 10). Global transgressive–regressive events at this stratigraphic level encompass both of the Kellwasser anoxic events in Germany and Australia (Buggisch 1991, Joachimski and Buggisch 1991) and linked North American events (Sandberg et al. 2002, Morrow and Sandberg 2003). Intermittent high sea level and destabilized greenhouse climate characterizes late Frasnian–Famennian time (Buggisch 1991). Late Devonian planetary ecosystems were sensitive to orbital forcing and stepwise climatic changes marking evolution from an ice-free to a glaciated state (Sandberg et al. 1988, Isaacson et al. 1999). As oceans and epicontinental seaways deepened with primary controls such as global tectonics, carbonate shelves responded to a stratified water column (Racki 1998, 1999; Sandberg et al. 2002).

Idaho geohistory displays elements of Frasnian–Famennian foreland effects of forebulge/backbulge paleogeography of the Great Basin (Goebel 1991, Giles and Dickinson 1995, Silberling et al. 1995, Trexler et al. 2004). The reconnection of restricted silled basins of the Jefferson shelf and Montana craton with the open ocean resulted in strong temperature and salinity contrasts. These would favor density stratification, stagnation with temperature-induced stress, nutrient loading, and opportunistic plankton blooms (cf. Racki 1998). Juxtaposition of these factors is reflected in the organic-rich black carbonate observed in the transgressive intervals of late Frasnian sequences 1 through 3, subsequently interbedded with artifacts after evaporites and stromatolitic peritidal rocks (Figs. 9, 10).

Evolution of the Technically Transitional Shelf: Grandview Dolomite sequences are more complicated than lower Jefferson Formation sequences (Fig. 4). The distributions of deci- to decameter-scaled Grandview Dolomite sequences represent carbonate factory responses to the first stages of margin transition and glacioeustatic controls (Isaacson et al. 2007, 2008; Fig. 9). Abrupt basinward shoreline shifts are observed, although the shelf break position and bathymetry are uncertain. Similarly scaled Middle Devonian sequences are also laterally extensive, with basal siliciclastics in LST–early TST incised fluvial environments, and glacioeustatic controls were likewise proposed for time-correlative strata in Nevada and the Prague Basin (Elrick et al. 2009). Deposition and erosional surfaces change significantly in overlying latest Devonian False Birdbear and Trident Member sequences, and younger black shale to fine-grained shoreface siliciclastics of the Sappington Formation. These depositional systems represent the last tectonic stage before active margin sedimentation (Figs. 9, 10). Glacioeustasy likely played a role in three latest Frasnian and seven to eight Famennian global sequences (Haq and Schutter 2008, Isaacson et al. 2008; Fig. 10). Although there is tectonic complexity, elements of these sequences are seen across the region.

The Late Devonian shelf transitioned from a passive to an active basin (Roberts and Thomasson 1964, Dorobek et al. 1991). Areas of similar stratigraphic characteristics suggest incipient fault partitioning over inherited transitional basement faults lasting over 18 my (Fig. 9). By contrast, open marine seaway limestone and shale of the Trident Member are evenly and widely distributed in the main part of the study area and show no evidence of subaerial exposure. Missing Famennian conodont zones, variable eastern and western erosion of Trident Member, and nondeposition of upper Sappington units suggest latest Devonian erosion across the shelf and reactivation of the Lemhi Arch (Sandberg et al. 1989). Montania, the Beaverhead Mountains uplift, western offshore uplifts such as the Ancestral Milligen Ridge, and areas of subsidence in the study area represent a broad, differentially deforming and eastward-migrating foreland (cf. Dorobek et al. 1991; Fig. 5). Coeval Late Famennian subsidence in the Central Montana Trough and Williston Basin preserve a comparable or genetically linked stratigraphic record. Overlying Mississippian deepwater turbidites and the deposition of the Madison Group in Montana over subtle Devonian paleotopography indicate that unstable locations such as those in the Beaverhead Mountains uplift finally foundered under intense subsidence and the arrival of a more widespread Antler foredeep.

Regional Correlation with Montana and Remaining Questions: The Idaho stratigraphic and conodont data set and local sea-level curve rely on the global eustatic curve of Johnson et al. (1985; Fig. 10). The least-understood rocks in Montana are the Three Forks Formation Logan Gulch Member, the often solution-collapsed Knoll Limestone, and the lower Trident Member. The False Birdbear and overlying Three Forks Trident Member in Idaho reflect a return to normal, open marine conditions, with more intense paired events of coastal on-lap and regression than are suggested by the global sea-level curve for the marginifera Zone. This conodont zone alone represents a disproportionately long 5 million years—one third of Famennian time (Kaufmann 2006). The least-studied units in Idaho are the highly brecciated D6 member and upper D5 member. Correlation of these units is not certain as a result of differential tectonic controls at this time (Fig. 1; Schwab 2017).

High-amplitude regressive to transgressive Late Famennian to Early Mississippian depositional patterns in Utah and Montana mostly do not occur in Idaho because of broad, tectonic hiatuses on the shelf (Fig. 10). Late Famennian hiatus in Idaho is not accompanied by subaerial exposure in the Trident Member. Thus, subsequent, syntectonic, subaqueous erosion and/or continuous, conformable, condensed intervals must occur in the upper Trident, likely the Sappington, and basal McGowan Creek units. Thin Helminthopsis-bearing black shale and calcareous, offshore rippled siltstone similar to distal Sappington Formation facies in southwest Montana are very similar to lithofacies of the McGowan Creek Formation in the Lost River Range. Such facies occur locally with very thick, black shale and plant-bearing turbidite channels, which cut into or overlie black nodular lime mudstone of the top of the Trident Member in the Lost River Range (Fig. 1). These Mississippian Antler foredeep basin sediments, like those of the Canadian Prophet Trough, are related to strata and disconformities on the foundering continental shelf and craton. Incipient Late Devonian subsidence in different areas along the Antler borderlands and seaways recorded different, genetically connected accumulation histories of the shelf.

Muddy, biostromal to peloidal and stromatolitic mixed carbonate and siliciclastic rocks with evaporite solution breccias characterize the upper Jefferson Formation Grandview Dolomite in east–central Idaho. These shelf strata occur above a late Frasnian disconformity, above the exposed outer shelf Grandview Reef, and below the False Birdbear sequence and Middle Famennian Trident Member of the Three Forks Formation. The Grandview Dolomite varies in thickness from about 70 m thick on the Lemhi Arch in the east to ~500 m thick in midshelf locations to ~300 m thick in the Grandview area in the west. The western Grandview Dolomite is correlative to the eastern D4 through D6 mapping members. Changing depositional systems responded to incipient foreland break-up of the lower Paleozoic carbonate margin. New interpretation of the Grandview Dolomite includes (1) Preliminary identification of major depositional sequences; (2) Geometry of shoreface/beach barrier sandstone facies in the Lemhi Range; (3) Discovery of D5 member patch reefs, such as the Gooseberry Creek Reef in the Lost River Range (a Nisku Formation correlative); and (4) Recognition of regional sequence stacking patterns extending to the Devonian Logan Gulch type section in Montana.

New correlations show a shift in depositional patterns during the latest Frasnian. Low-energy, low-amplitude ramp cycles of the D3 Dark Dolomite were followed by unconformities, brecciation, and sand influx, primarily from the craton. Subsequent transgressions resulted in widespread black (dysoxic) carbonates on the shallow shelf. These include coral–stromatoporoid and Amphipora biostromes that were deposited behind a higher energy, oxygenated distal shelf edge. Three latest Frasnian 50+-m-thick shelf sequences of third-order scale are well developed west of the Lemhi Arch, representing deposition in the late rhenana–linguiformis zones. These units are characterized by (1) basal quartz arenite and reworked carbonate conglomerate, sandy stromatolitic carbonate, and evaporite collapse breccia (LST–early TST); (2) laminated black subtidal to fossiliferous units (TST), overlain by (3) unfossiliferous, sandy peritidal carbonate (progradational HST). Latest Frasnian carbonate ramps aggraded/prograded in response to maximum global eustatic transgressive–regressive pulses, similar to Winterburn Group sequences in Canada. Incipient, crustal loading effects of Antler plate margin tectonics are reflected by increased latest Frasnian subsidence combined with reactivation and subaerial exposure of parts of the study area.

Shelf geometry then changed with overall second-order Lower Famennian regression and regional tectonic activity. Above the late Frasnian sequences, meter-scale peritidal carbonate units, quartz arenites, and stratiform cross-cutting breccias of diverse scale and genetic origin occur throughout the section. Famennian quartz arenite beach-barrier lithofacies were derived from the east or were transported laterally along continental and island shorelines, accumulating in an intrashelf basin on the Lemhi Arch flank. Trapping and preservation of sandstone, red siltstone, and breccias after evaporites occurred with local accommodation in this silled subbasin.

Late Devonian depositional systems prograded and accumulated over a broad shelf well east of the emergent Antler highlands. The paleoshelf developed internal stratigraphic complexity, but was not a foredeep. Shelf evolution can be organized into three transitional tectonic stages between passive and active margins. The lower Grandview Dolomite in the Grandview area represents a barrier shoal on the western shelf edge. Trapping of LST sandstone beds occurred east of this barrier, in the Early Famennian Central Idaho Trough. Quartz arenites may have ultimately bypassed into western basin environments of the coeval upper Milligen Formation. Autocyclic carbonate progradation and shelf-top accumulation of quartz arenites were controlled by global eustasy, persistent paleogeography, and reactivation of local hinge zones.

Submergence of the Lemhi Arch occurred in the latest Frasnian. In the same Idaho–Montana cratonic hinge zone, the Famennian “Beaverhead Mountains uplift” (Sandberg et al. 1975) was reactivated, resulting in complex stacked sequence boundaries there. Unconformities are associated with extensional uplift and subsequent subsidence. Jefferson Formation strata are capped by various expressions of the False Birdbear, the Three Forks Trident Member, and the Sappington Formation. These upper Famennian rocks are characterized by variable disconformities and an incomplete conodont record that represents episodic drowning and submarine erosion of the shelf, arch, and craton. Sequence boundaries show local to regional expressions of base-level changes associated also with glacioeustasy prior to deposition of Mississippian turbidites.

Carmela Garcia helped with creating restored Late Devonian isopach maps. Early versions of this elephantine manuscript were improved by Mike Whalen, Stacey Atchley, and less-impressed reviewers. We are especially thankful to the editors and for early work in Idaho with Carol Dehler, David Elliott, and Paul Link. More recent studies in Montana were encouraged by Murray Gilhooly, John Weissenberger, John Hohman, John Guthrie, Peter Holterhoff, Rick Beaubouef, and Lee Suttner and were made possible by Remuda Coffee in Three Forks. We are thankful to thoughtful ranch owners and for the open spaces of Idaho and Montana public lands.

(a) Palmatolepis marginifera in the cliff-forming, nodular medium-to thick-bedded crinoidal wacke- to packstone with very large cephalopods, similar to facies of the upper Costigan Member of the Palliser Formation of Alberta.

(b) Polygnathus styriacus (expansa Zone) in base of lower Sappington dark shales, where a subcrop of yellow calcareous silts also occurs with unidentified, poorly preserved macrofossils with immature small cephalopods.

(c) Siphonodella isosticha–Kinderhookian in the argillaceous mudstones with calcareous nodules in the McGowan Creek Formation, suggesting missing conodonts, and correlation to the Scallion Member of the Central Montana Trough.

(1) Long-term depositional hiatus and local deep erosion of Lower Paleozoic units occurred at the Lemhi Arch (Sloss 1954). Most sections in the Beaverhead Mountains have basal clastics and a variety of Devonian breccia units that, prior to brecciation, represented a blanket of upper Jefferson Formation limestone and evaporite prone units (i.e., D4 member; Scholten and Hait 1962, M’Gonigle 1982). Oncolitic dark dolostone facies suggests potential presence of common basal D5 member facies. Basal units may include very thin Silurian or Ordovician units.

(2) The cliff-forming “pseudobrecciated” Birdbear Member common at the top of the Jefferson Formation in Montana was once thought to occur at Long Canyon and Black Canyon. It is typically late Frasnian in age (Sandberg et al. 1983), but discovery of marginifera Zone conodonts at Long Canyon and at correlative sections in Idaho (Hawley Mountain and Freighter Springs) led to the problem of the younger “Birdbear-like” outcrops in Idaho and the naming of the “False Birdbear” (Poole and Sandberg 1977). At Long Canyon, reddish-brown, recessive upper Jefferson dolostone-like rocks of the Three Forks Logan Gulch Member occur in subcrop contact with basal blue-gray False Birdbear limestone; they are clearly different units (Pl. 2). The middle and upper Sappington members of Montana are not present at Long Canyon, indicating an unconformity or continued drowning and an undetermined condensed section (with possible subaqueous erosion). The nature of this contact is a key to understanding changing sources of Trident and Sappington clastics in the Central Idaho and Montana troughs east and west of the Lemhi Arch location. There is little direct evidence for locally active Antler fault blocks (Ramspott 1962, Scholten and Hait 1962).

(a) Ordovician conodonts and brachiopods (Ross 1959 personal communication in Mapel and Sandberg 1968; Ross 1961).

(1) Devonian base (Sloss 1954, Churkin 1962). Very similar Lower Paleozoic facies and structural problems have puzzled previous stratigraphers such as Preston Cloud (Ross 1961); correlation problems have remained unsolved (Hait 1987).

(2) Silurian–Devonian contact (Ross 1961, preferred Grader 1998) (Jefferson over Laketown formations).

(3) Silurian–Devonian contact in fault breccia? Mapel and Sandberg (1968) suggested that basal Devonian rocks were fault-shuffled based on breccia zones or discontinuous local breccia “pods.” Certainly, the top of the Devonian section is clearly repeated in a sharp older over younger thrust sliver of Devonian rocks at Black Canyon (removed in Fig. 7). However, a major sedimentary breccia unit near the base occurs with intraformational clasts and reverse-graded fine to very coarse sand in a dolomite matrix. A short distance downslope, clast entrainment is suggested by some imbrication of angular clasts (i.e., a synsedimentary slump?). A 6-m-thick dolostone and quartzite diamictite occurs in sinuous contact with laminated carbonate about 13 m above a basal Jefferson channeled sandstone. It is associated up and down section with bioturbated dolomudstone, sandstone, and laterally linked hemispherical [LLH] stromatolitic dolobindstone. It is not clear if this is a tectonic or sedimentary breccia. These problems in the Lemhi Range were reviewed by M’Gonigle (1982).

The breccia and overlying sandstone facies, if correlated correctly, are followed at other sections across the region by a black subtidal dolomitic limestone with an unusual cavity fabric, well developed at Black Canyon (see #4 below). The same spar-filled fabric may also be observed locally within the breccia unit, so it is perhaps a younger, overprinted fabric. The facies associations with widespread development of the cavity fabric above in dark dolostone appear to be common to one latest Frasnian interval in the Lemhi Range. The cavity fabric also occurs at the same interval throughout the Lost River Range, but it has also been noted higher in the stratigraphic section. Faulting in conjunction with evaporite association provide alternative postdepositional mechanisms for the brecciation, but quartz veins and slickensides are not present, and this unit is more organized and less chaotic than evaporite solution breccias and true fault-associated “rauhwackes” (tectonic, evaporite-associated breccia) seen regionally up section. Given these considerations and breccia reviews of M’Gonigle (1982), this Black Canyon breccia may not represent cataclastic tectonic shear as was previously suggested by Mapel and Sandberg (1968). On the other hand, it is not one of the clearer examples of a Jefferson karst or cave collapse deposit (after brecciation models of Mylroie and Carew, 1995) associated with global Late Devonian forced regressions (Isaacson et al. 1999) and sequence boundaries (this paper).

(4) Cavity Fabric or “Warme Cavity Fabric” (WCF) are temporary names given to a cryptic carbonate cavity fabric in the Jefferson Formation after an identical fabric, compared at a Geological Society of America (GSA) meeting, described and interpreted by Warme and Kuehner (1998). In the study area, at each section, this fabric occurs mainly in one interval, typically in a thick, dark gray to black dolomitic limestone shown above SB lFr1. It is a widely occurring, conspicuous carbonate fabric dissimilar to other “spaghetti beds” (Euryamphipora), which have been variously used as marker beds (Scholten and Hait 1962, Grader 1998, Grader and Dehler 1999). In outcrop and hand sample, cylindrical geopetal-filled cavities show some perpendicular connectivity, but they are mostly horizontally “ramifying” structures, the upper portions of which are filled with blocky, white coarse dolomite. The lower parts of each cavity contain a light gray, fine-grained dolomudstone. At Black Canyon, these dark beds are overlain by broken, transported syringoporoid corals that are lithostratigraphic grounds for correlation with any of the open-marine units of basal D5, D4, or alternatively the D3 member. Like sections in the Beaverhead Mountains (e.g., Long Canyon), strata here are hard to put into stratigraphic context. Based on the occurrence of the cavity fabric at other sections, we have placed this interval in association with a late Frasnian transgression in the D4 member. Initial search for conodonts at Black Canyon was unsuccessful (Grader 1998).

Warme Cavity Fabric at Black Canyon has been previously noted and reluctantly interpreted as bird’s eyes or stromatactis (Measures 1992). Scholten and Hait (1962) identified it as a possible relevant fact, mappable in upper Jefferson limestone from the central Lemhi Range to the Beaverhead Mountains. In accord with correlations suggested here, these authors proposed that this facies was a marker bed in the upper Jefferson and therefore that the thick lower Jefferson thinned dramatically over the southern Lemhi, Beaverhead, and Tendoy mountains. Scholten and Hait (1962) described it as a striking 1 to 2+ m zone packed with curved, discontinuous stringers of white crystalline dolomite, identical to beds described in Jefferson County (Montana), where these occur near the base of the Jefferson Formation. Three expressions of this cavity fabric were described and mapped in Nevada by Warme and Kuehner (1998). Clusters of cavities may be dispersed in a dolostone matrix or stacked, parallel and elongate, offset by small faults and interpreted as pervasive products of thermal shock waves under or near the Nevada Alamo impact crater, at their Tempiute Mountain location. Whereas their interpretation of this impact event occurred in shallow-water shelf strata, this is in conflict with newly discovered radiating channels and environments associated with deeper-water strata (Sandberg et al. 2002). Occurrence of this fabric in Idaho contradicts an interpretation restricted to bolide impact, yet its very similar appearance and limited occurrence remain enigmatic. First, the D4 member is too young to contain either direct impact effects or, for associated distal punctata Zone (Lower Frasnian), Alamo-type tsunami-associated facies or breccia deposits. Second, although the cavity fabric occurs in association with carbonate breccia both in Nevada and Idaho, the associated breccia is below the fabric bed in Idaho, not above as with the Alamo Breccia at Tempiute Mountain. On the other hand, as seen from Figure 7, other breccia units may occur within 10 m above (and below) this fabric.

Jack Wendte and Michael Whalen (personal communication, 1998, 2004) are skeptical of Warme and Kuehner”s bolide interpretation of this cavity fabric. In Idaho, it is a thin (1–5 m), mostly stratiform, secondary feature. Although Warme and Kuehner (1998) entertained a potential metasomatic origin in association with silver and tungsten, they argued that the pervasive fracturing and cavity formation do not occur in the Alamo Breccia or rocks above, and that associated partial bedrock disintegration, impact-formed “propants” (sand-and carbonate spherule-filled fractures), and shocked quartz are indicative of proximity to impact. Although this interpretation cannot be applied to the Idaho case, we agree that the cavity fabric appears to be a conduit for fluids. It may represent dewatering or degassing due to dilation (during seismic shear?) associated with in situ brecciation producing both monomict and polymict breccias, and upward mobilization of underlying loose sands. In the context of an impactor, this is discussed at length in Warme and Kuehner (1998). Similar breccias at or above the same stratigraphic level at the Devil’s Gate section in Nevada (near the Frasnian–Famennian [F–F] boundary), are interpreted as seismic-or storm induced breccias, from-the-forebulge debris flows, and/or carbonate platform margin collapse breccias in association with low sea-level stand with onset of Southern Hemisphere glaciation (Sandberg et al. 2002, Morrow and Sandberg 2003). This was apparently responsible for removal of all triangularis Zone strata from shelf deposits (Sandberg et al. 2002).

Our interpretation of late Frasnian breccias and enigmatic cavity fabrics in Idaho includes the following rational: (1) A less catastrophic sequence stratigraphic interpretation without the requirement of impact associations; (2) a working hypothesis of overprinted global eustasy on late Frasnian stratigraphy; and (3) the possibility of local paleobathymetric slope produced by syndepositional extension. Thus, Late Devonian sedimentation occurred during complex regional passive-to-active margin tectonic changes. Our tentative placement of the WCF into late Frasnian transgressive–regressive sequences gives the impression of a partly time-constrained, syndepositional component (i.e., potential time line). We suggest that the fabric in the Idaho case was associated with pore fluid migration, possibly seismically induced near to the time of deposition. Timing is unconstrained. It is not clear if the basal associated breccia overlying SB lFr1 is related to this fabric or not. It could be interpreted as part of a lowstand systems tract (LST) with possible shelfcollapse/seismicity. We have observed similar facies associations at other western sections, but a full explanation remains elusive (see also Ross 1961).

(5 and 6) Dark fossiliferous cliff-forming marine limestone occurs at the top of what was mapped by Mapel and Sandberg (1968) as massive dolostone of the Birdbear Member of the Jefferson Formation (by our usage, equivalent to D4–lower D5 members). The upper crinoidal limestone shares some stratigraphic and diagenetic commonalities with the “Knoll Limestone” of the Three Forks Formation in Montana (Doughty et al. 2014), as well as the limestone with marginifera conodonts described as the “False Birdbear” at Long Canyon. This unit is altered by partial dolomitization and dissolution under its contact with the yellow calcareous, argillaceous recessive limestone of the Trident Member (a sequence boundary). We suggest this limestone is not part of the older Birdbear Member. Sediment-filled dissolution cavities (microkarst) occur with pyrite-replaced allochems and large frambroidal pyrite below the yellow, iron fleck-stained argillaceous limestone, a common facies of the Trident Member. Mapel and Sandberg (1968) also noted that this argillaceous limestone grades upward into true dark gray limestone ledges up section (typical of a liming-upward Trident depositional pattern). They noted localized development of regolithic limonite-stained claystone or manganese-rich oolite and pisolite at the top of the Trident limestone, indicative of unconformity. Local stacking of flooding surfaces and subaerial unconformities is suggested (Fig. 7), but it is complicated at Black Canyon and in the Beaverhead Mountains by fault omission or repetition in the Devonian section. Altered Trident limestone (i.e., silty regolith) also occurs under the Sappington Formation in Montana.

(a) Halysites corals and Orthoceras cephalopods (known to occur in the Ordovician Saturday Mountain or Fish Haven Dolomite). (b) Palaeophyllum coral (Silurian–Ordovician?; DeSantis, 1996, pers. comm.).

(1) Overturned rugose corals occur in dark dolomite that lithologically resembles the D3 member of the Jefferson Formation, not the characteristic very light gray Silurian Laketown Formation. However, based on experience with local variations and remarkably similar lithologies shared among the Ordovician, Silurian, and Devonian units, we suggest that this unit is probably Silurian. We pick the overlying sharp contact and sandstone unit as the base Devonian (cf. similar questions debated at Black Canyon in Ross 1961).

(2) Lemhi Arch flank sands. Interbedded lagoonal Amphipora lime floatstone (highest occurrence), dolomitic microbial laminites, and high-energy planar laminated to bioturbated quartzite. Shallow-water unconsolidated quartz arenite and quartzite occur throughout the upper Jefferson Formation but are concentrated from Bunting Canyon to Badger Creek. The unprecedented abundance and stacking patterns of eastern siliciclastic rocks (seen also down section in older rocks) and isopach trends suggest that these areas mark the flank of a true paleogeographic feature—the Lemhi Arch (Grader 1998).

(1) The Givetian Jefferson biostromal D1 marker bed and transgressive D2 member (Taghanic onlap) are recognizable as far as Cedar Run and Badger Creek (Fig. 4). Between this part of the Lemhi Range and Uncle Ike Creek, and points south and east, the entire very thick lower Jefferson Formation appears to onlap and thin over a sharp hinge zone (later also strongly impacted by thrust faults). This hinge was a long-term feature supported by unusual Silurian reef facies at Cedar Creek overlain by incised Devonian channels with conglomeratic carbonate debris flows. Interbedded conglomerate and sandstone occur throughout much of the Givetian and Frasnian section in the Badger Creek area (Fig. 4). These are interpreted as transitional Lemhi Arch flank facies (Grader 1998).

(2) Breccia at SB Fm0. An overprinted breccia unit is interpreted to occur in the SB Fm0 sequence boundary zone (Fig. 7). Monomict to polymict breccias are crosscut by complex in situ brecciation with bleaching. This type of brecciation could be due to karst/cave collapse and is one example of a variety of breccia styles encountered in Idaho (M’Gonigle 1982). True karst breccias with yellow and red terra rosa fill have been observed in D1 and D6 members across the shelf.

(3) Beach and bar sands. A sandstone at Cedar Run is over 50 m thick and consists of well-sorted, well-rounded planar and cross-bedded, friable quartz arenite and some quartzite (Grader 1998). These facies are interbedded up section with stromatolitic carbonate and thin capping solution breccias. They are interpreted as beach and bar facies with carbonate peritidal environments. This sandstone is compositionally and texturally mature, typically containing a silica, calcite, or dolomite matrix (Pl. 2). Such sandstone is not different to Lower Devonian arenites derived from the craton, and sandstones from eastern and western sections are similar. Very minor components (< 1%) of feldspar and highly birefringent minerals have been noted in all Lower and Middle Paleozoic sandstones in this area. Floating quartz spheres comprising between 2 and 5%, to as much as 50%, of muddy or peloidal carbonate are also common, indicative of nearby eolian addition or shoreline environments. Minor tourmaline was also reported by Ross (1961) in Devonian sandstones.

(4) Upper D5 to D6 members. Near Cedar Creek at Bunting Canyon, shallow-water facies of the upper part of the D5 member contain meter-scale, shallowing-upward, possibly brining-upward depositional cycles. Thin to very thick breccia units maintain a coherent vertical stratigraphy with other typical Devonian facies, much as reported from cratonic successions of the Williston Duperow Formation (Wilson and Pilatzke 1987). Thick upper D5 and D6 breccias tend to produce a chaotic, yet fundamentally stratiform/strata-bound stratigraphy, which is interpreted as indicating pre-uplift presence of evaporites. Supporting this, time-equivalent evaporites occur in the Montana subsurface (M’Gonigle 1982) and in the Potlatch Evaporites of NW Montana. Salt pseudomorphs in D5 and D6 member outcrops have been noted, seen also regionally in the equivalent Three Forks Logan Gulch Member and in the Bierdneau Formation of southern Idaho. Dark laminated units (generally indicative of deeper subtidal ramp environments) were also noted to occur within incised paleovalleys in the lower part of the Devonian section, suggesting that look-alike subtidal units may occur in many different environments across low-angled pericratonic Devonian ramps.

(1) “Spaghetti rock,” with Euryamphipora. Stromatoporoid-coral biostromes and a 15 m buildup occur in the D3 member in the headwall at Horseshoe Gulch. These depositional environments are eastward equivalents of fossiliferous D3 rocks to the west. Tightly packed long white dolomitized wispy filaments were previously referred to as “spaghetti rock” (Scholten and Hait 1962). At Horseshoe Gulch, these are interpreted to act like encrusting algal filaments, or encrusting bryozoans, and are quite different to more common Amphipora floatstone or planar stromatoporoids that occur in lagoonal, to storm-agitated conditions (e.g., at the Grandview Reef, below). These filaments are a form of stromatoporoid—Euryamphi-pora—and have since been identified associated with Swan Hills correlative biostromes near the base of the Jefferson Formation in Montana (PRISEM, 2012, unpublished data). Their occurrence is very limited throughout the Jefferson Formation of Idaho: They are common in the D1 marker bed down section and only occur rarely in the D2 and D3 members.

(2) Red limestone and quartzite breccia are deformed by soft sediment contorted beds and confused by faulting at this locality. A similar breccia in the same basal D4 member stratigraphic interval at Bear Mountain exhibits association with red mudchips, siltstone, and mudcracks. Whereas previous authors have doubted the origin of Jefferson breccia units, study of their occurrence at multiple localities has shown that they form at predictable intervals in the section and are due to a variety of overprinted processes (M’Gonigle 1982). These include artifacts after evaporites, and rare evidence for subaerial exposure.

(3) Lower D5 member. The very dark, dolomitic, laminated, bioturbated, and partly fossiliferous and biostromal lower part of the D5 member is lithologically similar to the D3 member. Black, thick-bedded laminated to bioturbated subtidal ramp dolostones contain Amphipora floatstone (last occurrence) and some Thamnopora coral debris. Conispiral gastropod beds include oncolites. This unit is tentatively correlated to fossiliferous dark gray subtidal rocks below the cliff-forming unit at the top of the Black Canyon section.

(a) Liberty Gulch, like the Hawley Mountain section mentioned by Sandberg et al. (1983) further west, is one of the thickest Jefferson Formation sections in the region, with thick D5 member breccias marking the depocenter of a Late Devonian intrashelf basin. Because we have few conodont controls, we add here the report of the lower gigas Zone (early rhenana) and the significant Palmatolepis semi-chatovae (the semichatovae rise; Sandberg et al. 2002) 455 m above the Devonian base (349–359 m below the top) at Hawley Mountain. This conodont is also reported 10 m above the base of the Birdbear Member at the Logan type section in Montana (Sandberg et al. 1975, 1983, 1989) and is shown as hachured in Figure 4, where we estimate its occurrence in the transgressive deposits of the first latest Frasnian sequence of the D4 member. Note also that the Logan Gulch Member of the Three Forks Formation rests on the true Birdbear Member in Montana (observed at most SW Montana locations), and that the Sappington Formation occurs without the Trident Member shales at Long Canyon (Figs. 4, 7). Missing strata and conodonts suggested a Beaverhead Mountains uplift and unconformities (Sandberg et al. 1983). As depicted by Sandberg et al. (1989, Map 3, Pa. semi-chatovae) and in Figures 4 and 7, the middle ramp transitioned with and thinned near the reactivated paleogeographic high (Lemhi Arch–Beaverhead Mountains uplift). Similar patterns of depositional omission and merging of sequence boundaries occur on the Beartooth shelf and Central Montana Uplift, sometimes directly on the top of the Knoll Limestone (PRISEM Geoconsulting field work).

(1) Sequences lFr #1 to 3. Using simple vertical stacking patterns, sequences lFr 1 to lFr 3 are defined at Liberty Gulch based on field work there in 1995.

(2) D6 Member. The isopach maximum for the D6 member of the Jefferson Formation occurs at Liberty Gulch, Lemhi Range (Hait 1965), with better-studied outcrops nearby at Mammoth Canyon (M’Gonigle 1982). This evaporite solution breccia with stromatolitic bindstone occurs with dispersed quartz sand grains and some terra rosa-derived yellow and red solution fill. Eolian-derived frosted sand grains occur, but the units and subcrops are chaotic and need further study. D6 is equivalent to the Logan Gulch Member of the Three Forks Formation; however, Sandberg et al. (1989) limited the Logan Gulch Member to the Beaverhead Mountains and points east (Montana). Perhaps the D6 breccias were ignored by them, because they were thought to be indicative of faulting? Hait (1965) and Ruppel and Lopez (1988) mapped them throughout the central part of the Lemhi Range. D6 thins towards Cedar Run to the southeast and correlates with D6 in the Borah Peak area (west). Sandy, unfossiliferous carbonate interbeds in the Grandview Dolomite to the west may or may not be correlative to D6. The D6 member is very likely part of an exposed transitional sequence below the False Birdbear. Both units need further study.

(1) Good outcrop and the Lower Jefferson Formation. Rare 100% outcrop on a rugged section at Borah Peak shows the same depositional sequences and stacking patterns as in the central Lemhi Range. These exposed strata represent the unroofing of a thrust-imbricated, footwall in the Borah Peak horst (a horst within the main range horst). The rest of the Lost River strata to the north and south (Grandview Dolomite) represent more distal ramp sections and may have been eroded during the Famennian. Unlike strata at Borah Peak, they do not fit the member classification of the Lemhi Range. Other pre-Devonian strata and incised channel systems below the D1 marker bed, including the Carey Dolomite, are present in this more distal part of the middle ramp. Detailed meter-scale work has been achieved in excellent outcrops in the Lemhi and Lost River Ranges in basal channel systems and the D1 and D2 Jefferson members. Such outcrops are rare in the upper Jefferson Formation and are always subject to structural complications.

(2) D3–D4 contact. The disconformable contact between members D3 and D4 has variable expression regionally and is abrupt here (Pl. 1). In comparison, subaerial exposure at Grandview Canyon in the underlying buildup is overlain by very thick peloidal, intraclastic cross-bedded sandy “gray dolomite” (sensu Benson 1966). The regionally correlated D4 member (“Shallow Water Dolomite” Johnson et al. 1985, or Lower Grandview Dolomite) shows a switch to shallower-water peritidal cycles with more siliciclastic input (and unusual breccia deposits).

(3) Late Frasnian sequences #1 and #2. Following a truncation of beds at the base of lFr sequence #1, we interpret the D4 member to indicate overall regional progradation (offlap) of inner ramp settings towards the west, with two to three major transgressions and many subordinate black subtidal laminated to fossiliferous units to the west. Two fourth-order, decimeter-scale cycles with black transgressive systems tract (TST) rocks occur in the Borah Peak area (Pl. 1). The second sequence shown is perhaps the Pa. semichatovae rise (Fig. 10). Warme Cavity Fabric occurs in the overlying lFr sequence #2 at Borah Peak, as it does in Lemhi Range sections. Higher-order cyclicity in these third-order sequences deserves further analysis.

(4) Basal D5 member. Reconnaissance suggests D5 may have a “Nisku Reef” developed in the headwall of North Rock Creek, although the underling “Leduc Reef” nearby on East Rock Creek has not been thoroughly mapped; both have similar lithology and biota (Fig. 6B). On the Northeast Ridge, basal D5 is structurally repeated (Fig. 6A).

(5) Complete section of the upper Jefferson and Three Forks section. Excellent exposures in the nucleus of the syncline at the top of the ridge require further detailed work that should be correlated back to the Lemhi Range sections (Fig. 6B).

(a) At Mahogany Hill, polygnathid-icriodid biofacies, including Icriodus costatus, Polygnathus semicostatus, and Polygnathodus perpexus, suggest that Trident Unit 1 shallowed upward to a shallow-water biofacies (trachytera Zone) (Wiler 1992). (b) At Grouse Creek, 245 m below the D3–D4 contact, Pandorinellina insita occurs; i.e., the Frasnian punctata Zone in D3 (Wiler 1992).

(1) The Mahogany Hill channel conglomerate, described by Simpson (1983), was suggested by Wiler (1992) to be distally correlated to the western Picabo Formation, which he argued was evidence for early Antler-associated distal shelf uplift (“Antler welt”). Although Wiler”s hypothesis integrated this with increased late Frasnian subsidence (i.e., generally thicker lower Grandview Dolomite deposits), this interpretation does not fit the stratigraphically higher and younger Famennian Picabo Formation (Skipp and Sandberg 1975), or the Figure 7 correlation. We agree with Wiler on placement of the Mahogany Hill channel and correlative conglomerate/breccia channel at Grouse Creek at the Lower Jefferson Formation–Grandview Dolomite contact (SB lFr0), although we are not familiar with this locality yet. As an alternative paleogeographic hypothesis, instead of interpreting these channels as a distal correlative of the Picabo Formation to the west, we suggest they are LST incisions and backfills, with inferred bypass to the west and out of the plane of correlation. This places the Picabo Formation in a distal rather than proximal paleoshelf setting. Similar to the Picabo Formation, the clast compositions of the basal Grandview Dolomite channels are rounded, shallow-water Jefferson cobble dolostones. These are similar to older Devonian Badger Creek rounded cobble debris flows described in detail near the Lemhi Arch (Grader 1998; Fig. 4). The Picabo Formation occurs closer to the Milligen Basin and fits an off-shelf paleogeography. If the fluvial channel(?) was flowing to the west (like earlier channels of the Lost River Range), it might better fit our regional correlation, and it would call into question the postulation of distal shelf uplift and clastics from the west (cf. Wiler 1992). This requires more study.

(2) Silicified carbonates and abandonment of previous interpretations. Much of the Mahogany Hill section is reported to be brecciated and silicified (jasperoid replacement; Simpson 1983). Finding evidence for the Warme Cavity Fabric (assuming that it is a true marker bed) at other Grandview Dolomite sections may help to better position interpreted stacking patterns. We abandon previous stratigraphic interpretations of Simpson (1983) at Gooseberry Creek and his placement of the Dark Dolomite-Grandview Dolomite contact near Grouse Creek (see item 3 below). Initial interpretations by Simpson at Mahogany Hill were also abandoned by Wiler (1992). Whereas the units in the type locality of the Grandview Dolomite at Grandview Canyon (Ross 1934) and Grandview Hill (Wiler 1992) are clear, complete correlation and reinterpretation of the Grandview Dolomite is needed for the northern through southern Lost River Range—especially upper Grandview Dolomite sequences.

(3) At Grouse Creek, and also reported at Mahogany Hill, abundant dark gray to weathering brown bioturbated–laminated subtidal lithofacies occur with some indications of shallow-water peritidal interbeds. These rocks are similar to our studies in the upper Grandview Dolomite at Grandview Canyon (Pl. 2). The section is composed of mainly calcareous dolostone and seems relatively featureless; much of the upper part of the section here crops out poorly. Dark black carbonate in the 100 to 200 m interval above the Dark Dolomite-Grandview Dolomite contact are thought to be partly correlative to the black, Amphipora-bearing subtidal deposits of the lower D5 member. New field studies suggest good correlation of surfaces and strata down section, with a relative thinning of the Grandview Dolomite. However, a section near here by Ross (1947) suggests the Grandview Dolomite is 645 m thick, rather than our 260 m measurement (PEI and GWG, 2007). Thus, further study is needed.

(1) Reinterpretation: The Gooseberry Creek section was described by Simpson (1983). This and the Mahogany and Grouse Creek sections were initially correlated using the Dark Dolomite–Grandview Dolomite boundary interpretation and paleogeographic reconstructions of Simpson (1983) and Wiler (1992). However, upon study of their fossil content and significant occurrence of black fossiliferous subtidal lithofacies (in association with sections on the northeast ridge of Borah Peak and in the central Lemhi Range), it became unclear how this previous work should be integrated into regional correlation—now reinterpreted in Figure 7.

Simpson (1983) reported thick subtidal successions in the upper Grandview Member at Mahogany Hill and Gooseberry Creek. At Gooseberry Creek, very fossiliferous units were previously positioned ~300 m above the Dark Dolomite-Grandview contact (Simpson 1983). This stratigraphic position was incorporated into tectonic-driven paleogeographic models developed by Wiler (1992) and Wiler and Isaacson (1991). Recent field work confirms that multiple levels of bryozoan bafflestones identified by Simpson (1983) and used by Wiler (1992) to support a backbulge basin east of the Grandview Hill section are silicified Thamnopora tabulate corals (which resemble bryozoans), and they occur much lower in the Grandview Dolomite. These facies occur with excellent exposures of thick stromatoporoid buildups and Amphipora biostromes. New reconnaissance work shows that the entire Devonian section is well exposed at Gooseberry Creek, including clear stacking patterns. The fossiliferous strata previously identified in the upper Grandview Dolomite are actually part of a continuum of fossiliferous cycles starting in the Jefferson Dark Dolomite (D3) and ending in latest Frasnian lower Grandview Dolomite (D4–lower D5). Obvious normal faults perpendicular to strike shuffle the section laterally, and we identified ~200 m of partly sandy, stromatolitic upper Grandview Dolomite facies (upper D5–D6 equivalent) below excellent exposures of the Three Forks Formation (Fig. 6C).

(2 and 3) Gooseberry Reef and western uplifts reconsidered: Eastward deepening away from the seaward Grandview area, Famennian bryozoan bafflestones in the upper part of the section, and other lines of evidence, suggested to Wiler (1992) that uplift of the distal shelf was occurring west of the Grandview area during the latest Frasnian-Famennian (i.e., Bayhorse area). We partly agree with Wiler”s proposal that subsidence in the Central Idaho Trough intrashelf basin led to the pattern of more shallow-water facies in the Grandview area and partly subtidal facies to the east. Wiler”s interpretations followed the earlier model or axiom of Sandberg et al. (1975), Skipp and Sandberg (1975), and Sandberg et al. (1983) regarding uplift and erosion of the basinal Milligen Formation to the west (“Ancestral Milligen Ridge or “Antler welt”). As Simpson (1983) mentioned in his conclusions: This transition eastward away from the apparently deeply eroded, adjacent Pioneer Mountains Ancestral Ridge (“Antler welt”) shows little evidence of associated erosion products in the western Grandview Dolomite. “Sandstone shed from the west” onto the Jefferson shelf (sensu Wiler 1992) is not well supported.

Based on thicker accumulation of sandstone beds to the east (this study), sandstone deposited on the western side of the shelf more likely originated from the east (craton and arch) via more standard and long-term shelf-to-basin pathways controlled by sea level. The Central Idaho Trough intrashelf basin may thus have been silled, but its western rim still faced an ocean and evolving foredeep. Sandstone and associated shallow-water facies in the Grandview area were reworked along the distal shelf. If outer shelf uplift and erosion did occur, it would have eroded the Grandview Dolomite and controlled upper Jefferson facies along this hinge point in the continental profile (Fig. 9). Perhaps pre-Trident Member variable thicknesses of the upper Grandview Dolomite are evidence of Antler Orogeny-induced instability and irregular erosion of the shelf, along with more cryptic, proto-Antler uplifts of the continental slope to the west (i.e., eroded Milligen Formation).

Wiler”s correlation and paleogeographic reconstructions were also partly motivated by misinterpretation of sandstone and dark biostromes (with Thamnopora, stromatoporoids, and rugose corals) on the western slopes of Borah Peak. These rocks are, however, not part of the upper Jefferson Formation, nor are they Famennian. These strata are clearly part of the lower Devonian depositional complex and are probably Eifelian or older (Carey Formation; Grader 2002), and Wiler”s stratigraphic section and interpretation there should be abandoned. We place this part of the Grandview Dolomite into a latest Frasnian Nisku-equivalent sequence (i.e., upper D4 and lower D5) above SB lFr2 and SB Fm0. This solves some correlation problems, but missing thick eastern successions between SB Fm0 and SB Fm4 (upper D5 and D6 members in the Lemhi Range) in most of the Lost River Range are problematic. These rocks require further study: They are either represented by thinner deeper rocks and condensed deposits, or by erosional thinning and hiatus. The False Birdbear was not identified; instead, the Trident Member overlies a karsted top of the Grandview Dolomite.

(a) Sartenaer and Sandberg (1974) suggested the presence of both lower and upper marginifera Zone fauna associations, with occurrence of Palmatolepis marginifera. The top 2 m contain Palmatolepis marginifera marginifera, although other characteristic faunal elements are missing. This sample represents the upper marginifera Zone, which Sartenaer and Sandberg suggested was the top of the Jefferson Formation (i.e., “Birdbear” as later reinterpreted as “False Birdbear”).

(b) Trident Member units 2, 3, and 4 suggest a Late trachytera Zone. Sparse conodonts were assigned to the Scaphignathus subserratus-Pelekysgnathus inclinatus (Scaphignathus velifer Zone in Germany; Sartenaer and Sandberg 1974). Later samples processed by Hell (2010) also found trachytera Zone conodonts.

(c) Reworked Devonian conodonts and the presence of lower and upper Siphonodella crenulata Zone fauna associations suggest a stratigraphic position well into the Mississippian. Fox (1985) also noted Palmatolepis sp., collected from the Trident at upper Lower Cedar Creek, suggesting a deep-water setting, compared to a correlative shallow-water setting in SW Montana. The presence of Palmatolepis glabra lepta is indicative of the late Scaphignathus velifer Zone. A high percentage of Late Devonian species Polygnathus semicostatus (expansa Zone?) in the Mississippian strata are indicative of active erosion of the upper Three Forks Sappington Member in the region. Three missing conodont zones apparently indicate Antler uplift (Sandberg et al. 1989; Fox, 1985; i.e., during the sulcata, praesulcata, expansa, and postera zones). Could the expansa-age conodont represent actual Sappington deposits near the base of the McGowan Formation as suggested at Gooseberry Creek? This needs further study.

(1) “Warme Cavity Fabric.” The Freighter Springs locality was described by Sartenaer and Sandberg (1974). This is the only locality where the Warme Cavity Fabric occurs both in its horizontal, bedding-normal form but also exhibits vertical variations. As at other sections, this lithology follows a poorly outcropping breccia and sandstone unit, yet here it occurs in a medium gray dolomite, versus the more usual association with black dolomite. Its stratigraphic position below the Three Forks Formation represents an anomaly, perhaps indicative of deep local erosion of the upper Jefferson Formation (unlikely). Field work in association with an Idaho State University field camp (2001) resulted in stratigraphic reinterpretation and a new map for this area. As discussed above, we argue that the marginifera-bearing “pseudo-brecciated” (nodular?) encrinite facies discussed by Sartenaer and Sandberg (1974) are not the Jefferson Formation. Instead they appear to be part of an additional 20-m-thick shallowing-upward sequence in the basal Trident Member of the Three Forks Formation. To further test the incomplete Jefferson Formation stratigraphic section here, an accessible section is available nearby, above the Borah Peak Trail Head.

(a and b) Frasnian polygnathid conodonts and Polygnathus-Icriodus biofacies occur within and above the Grandview buildup (bioherm), including Polygnathodus pacificus 10 m below SB lFr0. The Grandview buildup is lower rhenana Zone (Wiler 1992). Unit IV (lower Grandview Dolomite) contains Polygnathus sp., P. pacificus, Palmatolepis sp. indet., and further up section Polygnathodus webbi and Polygnathodus churkini.

(c) Basal Three Forks Formation contains Palmatolepis glabra distorta, Pa. glabra lepta, Palmatolepis glabra pectinata, Pa. marginifera marginifera, Palmatolepis marginifera utahensis, Palma-tolepis marginifera schindewolfi, Polygnathus lagowiensis, and P. semicostatus (Late marginifera Zone; Wiler 1992).

(1) Grandview Reef (see also further Descriptions below). Coral-stromatoporoid biostromes and buildups occur intermittently throughout the D3 member in the Lost River and Lemhi Ranges. At Horseshoe Gulch (above), we described an unusual buildup of stromatoporoids with abundant Euryamphipora. Abundant, correlative, smaller stromatoporoid-Thamnopora biostromes occur in D3 and D4 at Grouse Creek. Thick (15–25 m) biostromes containing large domal and bulbous stromatoporoids along with Amphipora and Syringopora occur at various stratigraphic levels on the north face of Borah Peak and at Gooseberry Creek (ending in the Gooseberry Reef in the basal D5 member). Very large colonial rugose coral heads such as those found at Grandview Hill are rare in the Lost River Range buildups. The best exposed and well-described bioherm occurs at Grandview Hill, with lateral facies correlated to Grandview Canyon (Isaacson and Dorobek 1989, Isaacson et al. 1989, Desantis 1996; see below).

In Grandview Canyon, the eastern flank of the Grandview Hill buildup and overlying Grandview Member of the Jefferson Formation are well exposed. This buildup is notably different from other D3 buildups. First, at 40 m thick, it is much thicker compared to other localities. This is in part because the Grandview buildup contains a series of stacked bioherms with both lateral and vertical biological zonation (Desantis 1996). Primary buildup faunas are Thamnopora, Peneckiella, domal, tabular, and nodular stromatoporoids, Syringo-pora corals, and Amphipora. Less common (but still abundant) are pelmatozoans, brachiopods, and gastropods. Buildup anatomy can be divided into nine biofacies and four developmental stages. Stages I, II, and III each record a discrete period of reef building, whereas Stage IV suggests a return to open ramp carbonate. At least two autogenetic cessations occur in this succession. Three flooding surfaces with four shallowing-upward cycles can be defined (Pl. 1). Reef dwellers persisted in recolonizing the locality several times. Isaacson et al. (1989) suggested that the uppermost part of the buildup was allogenically controlled. Flooding just below the widespread SB lFr0 is suggested by a similar pattern of laminated sediments over biostromal ones at the top of the D3 member in the central Lemhi Range. The uppermost decameter cycle shallows before subaerial exposure indicated by brecciation and breccia-filled vertical fissures and silt-filled karst pipes (SB lFr0). Intraclastic, light-colored, cross-bedded Grandview Dolomite rocks abruptly overlie this surface.

(2) Grandview Dolomite at Grandview Hill (near the type locality in Grandview Canyon; Ross 1934). The lower Grandview Dolomite is characterized mainly by successions of light-colored shallow-water lithofacies with abundant mixed sand and minor intertidal sandstone beds and abundant occurrence of the Polygnathis-Icriodus conodonts of probably latest Frasnian age (Unit IV; Wiler 1992). Cyclicity in Unit IV is suggested by ~12 dark gray to black subtidal units interbedded with these facies. Episodes of significant flood-back and re-establishment of subtidal conditions over the entire region culminated in the accumulation of 10 m of D3-like subtidal bioturbated/laminated dolostones with Amphipora, domal stromatop-oroids, and Syringopora. The last such dark interval is peloidal and contains no fossils, yet it contains the unusual Warme Cavity Fabric, which, as at other locations, is also underlain by a sandstone and cryptic breccia unit. It remains unclear how these rocks should be correlated. Correlation to basal D5 at the top of Wiler”s Unit IV is shown together with potential sequence boundaries. Alternative correlations are possible.

LLH, crinkly, and planar stromatolites are common throughout the Grandview Dolomite. One of the most striking differences of the upper Jefferson Formation here compared with partly correlative upper Jefferson Formation sections in the Lemhi Range is the abundant occurrence of intraclastic and peloidal grainstones (compared with muddier sections to the east). Also, compared with sections immediately to the east (Grouse Creek and Mahogany Hill), there are increased percentages of floating sand grains and sandstone units in the Grandview area (although not as concentrated as those in the Lemhi Range).

Wiler (1992) interpreted this lithofacies distribution as evidence for influx of clastics from the west. We interpret these facies as part of relative sea-level drop and complex late Frasnian progradation/offlap of the carbonate ramp toward the distal shelf. We suggest that G. Loucks (1977) barrier interpretation is partly correct (Fig. 3B); i.e., that restricted conditions occurred behind these distal shelf deposits, except during the early parts (TST) of the most transgressive latest Frasnian flood backs. Westward progradation filled the earlier back-reef area (early extensional intrashelf basin), and aggradation on the distal ramp produced a new ramp geometry. We further propose that the barrier facies faced an open ocean to the west, not an uplifted subaerially exposed continental slope, although this changed later in the Famennian–Mississippian.

(3) Stromatolites. Sandberg et al. (1989) reported stromatolitic structures near the base of the Famennian on the shelf in Nevada. Units IV and V of the Grandview Member contain abundant stromatolitic bindstone with multiple levels of LLH stromatolites and possible disconformable surfaces. Wiler (1992) placed the Frasnian–Famennian boundary in this interval, below a pink to red, very sandy carbonate unit about 220 m above the contact with the D3 member. No Amphipora stromatoporoids were reported from early Famennian rocks in the western United States (Sandberg and Poole 1977). The last occurrence at this section is below the interpreted F–F boundary (SB Fm0), ~35 m lower than where Wiler placed this boundary. We show that the last occurrence of Amphipora in the Lemhi Range occurs in the basal part of D5, below the basal Famennian datum SB Fm0 and an overlying 5 to 10 m thick, fine-to medium-grained, well-sorted sandstone interval with dune to ripple cross-bedding. This unit is also observed ~150 m below the top of the Grandview Dolomite above both sides of the road in Grandview Canyon (Pl. 2). The age and correlation of these rocks and surfaces require further study. Three to four overlying sequence boundary zones occur in Units V and VI at Grandview Hill, below False Birdbear lithologies and marginifera–trachytera zoned Trident Member (Fig. 8). These rocks are approximately correlative with the upper D5 and D6 members in the Lemhi Range.

(4) The upper part of the Grandview Member. Unit IV (cf. Wiler 1992) at Grandview Hill contains sandy and cross-bedded carbonate below subtidal units. Observations at Grandview Canyon suggest that these facies occur locally and that coarse sandstone overlies a sharp disconformity (Fm4) overlain by dark dolomitic, burrowed False Birdbear lithologies (Pl. 2). A 15 m section of rusty quartzite in this interval was reported by Ross (1947) at Grouse Creek. He placed these strata in the base of the Three Forks limestone (109 m thickness overall). At Grandview Canyon, sandstone of similar thickness was observed below Fm5. Based on our regional work, we have reinterpreted Wiler’s section and “Unit VI” at Grandview Hill to include the False Birdbear sequence.

Biofacies 1:Peneckiella with domal, nodular, and encrusting stromatoporoids. This biofacies is laterally continuous across the entire outcrop and thickens from 1 m in eastern sections (east side of Grandview Canyon) to 2 m in western exposures (Grandview Hill, west of Grandview Canyon). Lithologically, it changes upward from basal wackestones to floatstone and bafflestone. Skeletal debris from Peneckiella, Thamnopora, solitary rugose corals, gastropods, brachio-pods, and pelmatozoans is scattered throughout the biofacies. Biofacies 1 is constructed on a hardground with numerous borings and local relief of up to 10 cm. In situ stromatoporoids, exhibiting domal (3–10 cm in height, 3–8 cm in diameter) and nodular growth forms, dominate the bottom of the biofacies. Fragments of Thamnopora occur scattered throughout this zone and as a thin bed of debris at the base of the buildup. Nodular stromatoporoids and some of the domal forms commonly encrust the coral debris. The upper part of the biofacies contains numerous hemispherical Peneckiella colonies. The colonies display distinct size changes from west to east across the study area. The largest colonies, located in section 4, grow to 60 cm in height and 40 cm in diameter. Colonies in western sections are more abundant and generally occur in growth position. In contrast, eastern exposures contain more coral debris, have overturned colonies, and are often sharply truncated and locally overgrown by stromatoporoids.

Biofacies 2: Domal and tabular stromatoporoids. Biofacies 2 is present in all but section 6 and, as with biofacies 1, thickens to the west (1 m in section 4 to 7 m in section 3). Deposits are primarily floatstones. The matrix at the base is mudstone and coarsens upward to wackestone–packstone. Bindstones are present in the upper parts of the biofacies in western sections. The biofacies is characterized by domal stromatoporoids. Other biotas are coral and Amphipora debris, solitary rugose corals, gastropods, brachiopods, and pelmatozoans. Characteristic domal stromatoporoids are most abundant in eastern sections and grade upward into lamellar and tabular forms in western sections. The diversity of non-stromatoporoid taxa increases dramatically to the east, and many are concentrated in localized debris pockets.

Biofacies 3: Stromatoporoids and debris. Biofacies 3 occurs in the two westernmost sections of the buildup. It is an isolated thin (1–2 m) interval separated from under-and overlying units by intervals of poorly fossiliferous mudstone. The deposit is primarily floatstone in a mudstone–wackestone matrix. Thamnopora and Amphipora debris occurs in local concentrations. Brachiopods and solitary rugose corals are scattered throughout. The upper part of the biofacies in section 3 contains a thin stromatoporoid biostrome colonizing a debris bed. Tabular to domal growth forms are up to 10 cm in diameter.

Biofacies 4:Thamnopora. Biofacies 4 is found in all measured sections. Maximum measured thickness (6 m) is in section 5. It thins to < 1 m in section 4 and to 3 to 4 m in western exposures. Deposits are a mix of coral floatstones and rare bafflestones. The biofacies is characterized by branching colonies of the tabulate coral Thamnopora. In situ colonies rarely exceed 15 cm in height and make up massive beds. Debris intervals occur as poorly defined 5- to 30-cm-thick beds containing coral fragments that are 3 to 5 cm (but up to 15 cm) long. Interbedded floatstone and in situ layers indicate that the biofacies was built in stages, but these do not appear to be correlatable. Other biotas are present but are rare compared to Thamnopora.

Biofacies 5: Stromatoporoids. Biofacies 5 occurs in center sections of the buildup exposure. Thickness is 2 to 3 m. Deposits are predominantly floatstone with local bindstone in a mudstone–wacke-stone matrix. The biofacies contains a more diverse fauna than the underlying biofacies 4. Biotas are similar to those of biofacies 2, but in biofacies 5, the taxa are more evenly distributed across the outcrop. Domal and tabular stromatoporoids are in growth position. Colonial coral debris and solitary corals, brachiopods, and pelmatozoans are scattered throughout the biofacies.

Biofacies 6: Tabular stromatoporoids and Amphipora. Biofacies 6 occurs in all sections but 3 and 4 and is about 3 m thick throughout. Deposits are floatstone–rudstone and bindstone in a packstone–wackestone matrix. Tabular stromatoporoids and Amphipora are characteristic. Relative abundance of these varies from section to section. Thin, light-colored hummocky beds of Amphipora rudstone alternate with darker muddy, matrix-supported beds. In situ tabular stromatoporoids associated with the muddy beds bind isolated lenses and pockets of Amphipora debris.

Biofacies 7: Stromatoporoid framework. Biofacies 7 is best exposed in sections 1, 2, and 6. It is 6 m thick in section 6 and thins progressively to the west. Deposits are massive floatstones and bindstones in a packstone to wackestone matrix. Large bulbous (25 cm in height, 5 cm in diameter) and domal (3–10 cm in height, 3–12 cm in diameter) stromatoporoids dominate the lower part, while tabular stromatoporoids constructed a densely intergrown framework at the top. Fossil debris comprises all of the higher-order taxa found within the buildup. These biotas include Thamnopora, Peneckiella, solitary rugose corals, Amphipora, brachiopods, Stachyodes, and pelmatozoans. The unit has the highest concentrated diversity of all the buildup biofacies.

Biofacies 8:Amphipora and solitary rugose corals. Biofacies 8 is limited to the westernmost measured sections and is about 3 m thick. Amphipora and solitary rugose corals in a massive black mudstone dominate the unit. Small Syringopora colonies, brachiopods, and gastropods occur locally. Amphipora fragments are 1 to 3 cm in length, scattered throughout the biofacies, and locally concentrated in thin rudstone beds. Solitary rugose corals are concentrated in thin densely packed beds.

Biofacies 9:Syringopora. Biofacies 9 occurs in well-bedded, bioturbated mudstones and wackestones. Dolostones of this unit form small, less-resistant ridges, contrasting with the more massive cliff-forming units below. The biofacies occurs in all measured sections and ranges from 7 to 17 m in thickness. Delicately branching, unbroken colonies of Syringopora form dense thickets up to 1 m thick.

Stage I: Stage I records the transition from open ramp carbonate to “reef” deposition. Stromatoporoids exhibiting domal and nodular growth forms colonized a hardground strewn with coral debris. The stromatoporoid growth forms suggest an environment with low sedimentation and moderate turbulence (James 1983). Following colonization, the stromatoporoids were replaced by the colonial rugose coral, Peneckiella, at which point the buildup developed lateral zonation. A western “core” area (section 3) is indicated by thicker deposits and large Peneckiella colonies forming bafflestones. A protected “back reef,” supporting a fairly diverse fauna of corals, brachiopods, and gastropods, formed east of the core. Upper parts of Stage I (biofacies 2) continued the previously established zonation. Encrusting stromatoporoids, indicative of high turbulence (James 1983), inhabited the western core. To the east, the back reef remained populated by a more diverse assemblage of corals, domal stromatop-oroids, pelmatozoans, gastropods, and brachiopods. Stage I passes upward into poorly fossiliferous bioturbated mudstone, which separates it from overlying units and apparently represents conditions unfavorable to reef growth.

Stage II: Stage II was a brief recolonization limited to western sections of the outcrop. This area (the core of Stage I) may have remained as a slight topographic high. Initial colonization was by Thamnopora, Amphipora, and solitary corals, species that are attributed to turbid or muddy environments (Wilson 1975, Kobluk 1978). At the top of Stage II, large in situ domal stromatoporoids opportunistically colonized a bed of bioclastic shell debris, forming a small biostrome. This unit is thin, unzoned, and not laterally extensive.

Stage III: Stage III represents a second reactivation of reef growth characterized by vertical and lateral zonations. The stage was initially dominated by Thamnopora. Coral floatstones are interbedded with in situ corals buried in carbonate mud. These deposits indicate background conditions of moderate to high sedimentation and relatively low turbulence, punctuated by storm scouring. The lowest parts of Stage III, then, formed in deeper water, just within storm wave base. Stromatoporoids dominated the upper part of Stage III, where the buildup again developed lateral zonation. In the core area (sections 1, 5, and 6), initial development was by isolated small domal and tabular forms (biofacies 5), suggesting growth in a moderately turbulent environment. Biofacies 5 is muddy and formed as a “back reef” sheltered by biofacies 6. Biofacies 6 was deposited within wave base. Layers of coarse, grain-supported bioclasts interbedded with muddy units suggest periods of higher wave energy. Tabular stromatoporoids also indicate growth in moderate turbulence (Kobluk 1978). The uppermost unit of Stage III was also deposited in turbulent conditions. Domal and tabular stromatoporoids of biofacies 7 constructed an intergrown framework that provided habitat space for a diverse assemblage of reef dwellers. Unlike biofacies 6, this unit is massive, suggesting the turbulence was continuous. Biofacies 7, then, probably grew within fair-weather wave base. Areas away from the core exhibit different patterns. Biofacies 5 and 6 pinch out to the southwest. Also to the southwest is biofacies 8, which likely formed at the same time as parts of biofacies 7. Biofacies 8 (solitary corals and Amphipora) indicates a relatively quiet, muddy environment. This area records accumulation in a protected intrareef lagoon. To the northeast, biofacies 5, 6, and 7 pinch out completely. Coeval deposits in section 4 (Amphipora and gastropods) indicate another protected lagoon.

Stage IV: Stage IV deposits (biofacies 9) contain relatively fewer fossils than underlying units. This stage persists across the entire measured outcrop and blankets units indicative of turbulent (biofacies 7) as well as quiet (biofacies 8) hydrodynamic regimes. Delicately branching Syringopora colonies and a muddy matrix indicate quieter conditions. Stage IV probably marks an increase in water depth and a return to typical Jefferson D3 Dark Dolomite Member conditions.