Gold Open Access: This paper is published under the terms of the CC-BY license


The Ignacio Quartzite—exposed in the San Juan Mountains of southwestern Colorado—is composed of red and brown arkose and subarkose sandstones and minor interbedded shales. The formation is newly divided here into the Tamarron Member (0–24 m) and the overlying Spud Hill Member (0–21 m). The Spud Hill Member has a greater abundance of sandstones with shale clasts, weakly fissile shale beds, and trace fossils than the Tamarron Member.

The McCracken Sandstone Member of the Elbert Formation, which overlies the Ignacio, is chiefly white and off-white quartz-cemented quartzarenites. The McCracken is divided for the first time into the Mill Creek facies (0–12 m) to the south of Coal Bank Pass and the Sultan Creek facies (0–36 m) to the north of the pass. The Sultan Creek facies contains dolostone-sandstone parasequence tidal-flat cycles up to 70 cm thick with a composite thickness of 14 m.

Eastward transgression across the western edge of the Transcontinental Arch permitted the accumulation of fluvial deposits of the lower Tamarron Member in the deepest channels incised into the craton. As sea level continued to rise, fluvial channels evolved into estuaries dominated by sandy tidal flats (upper Tamarron Member, Mill Creek facies, and Sultan Creek facies) and mixed sand and mud tidal flats (Spud Hill Member).

Sandstone composition and ages of detrital zircons indicate that sand grains were derived from a complex terrain that included granitoid plutonic rocks (∼0.46 to >2.4 Ga), metamorphic rocks, and well-rounded quartz sand from eolian ergs. The area between the present Coal Bank and Molas passes was a boundary between a northern fluvial source with an abundance of superbly rounded quartz grains of eolian erg origin from a southern fluvial source with few such grains, but large amounts of K-feldspar.

The Ignacio and McCracken units are, at least in part, coeval and of Late Devonian age as shown by the stratigraphic distribution of rocks resting on the basement, the presence of an Ordovician zircon in the Ignacio, and the presence in the same Ignacio sample of oboloid brachiopods of questionably late Cambrian age with well-dated Late Devonian placoderm fish plates.


The purpose of this paper is to describe the stratigraphy, depositional history, petrography, and provenance of the lowermost two Paleozoic stratigraphic units exposed in the western part of the San Juan Mountains, southwestern Colorado (Fig. 1). The units are the Ignacio Quartzite and the overlying McCracken Sandstone Member of the Elbert Formation. Although some workers have referred to these units as the ‘Ignacio Formation’ and ‘McCracken Quartzite Member,’ the present paper follows the National Geologic Map Database's usage of ‘Ignacio Quartzite’ and ‘McCracken Sandstone Member’; National Geologic Map Database (2016). Refining the stratigraphy of the two units within the present study has led to dividing the Ignacio Quartzite into two members (the Tamarron and overlying Spud Hill Members) and dividing the McCracken Sandstone Member into two geographically separated facies (the Mill Creek and Sultan Creek facies).

The Ignacio and McCracken stratigraphic units are of interest because (1) they record the earliest stages of transgression across Precambrian basement rocks of the Transcontinental Arch in southern Colorado; (2) their lithic contrast with the much thicker overlying carbonates of the Upper Devonian and Carboniferous; and (3) their possible outcrop equivalency to hydrocarbon-exploration targets in the Paradox Basin to the west (Fig. 2). In addition, a century of controversy exists over their distinction and the lack of a recognizable regional disconformity between alleged late Cambrian (Ignacio Quartzite) and Late Devonian (McCracken Member) rocks.

The study area is situated in the San Juan Mountains on the southwestern margin of the Uncompahgre Uplift, an Early–Middle Pennsylvanian feature of the Ancestral Rocky Mountains (Baars, 1965; Thomas, 2007; Fig. 1, this paper). The study area is bordered on the west by the northeastern margin of the Paradox Basin (a Permian salt basin [Baars and See, 1968; De Voto, 1980]), on the south by the San Juan Basin, a Laramide feature (Tweto, 1980), and on the east by Proterozoic basement rocks of the Needle Mountains. Nearly 3,000 m of Phanerozoic sedimentary rocks are exposed on the southeastern margin of the San Juan volcanic complex (Fig. 2). A large-scale northwest-trending fault block—in which late Precambrian through Mississippian rocks are exposed—forms the core of the San Juan Mountains. The area was overprinted by Tertiary volcanics that formed the San Juan Dome (Condon, 1995). Uplift and deformation occurred during development of the Ancestral Rocky Mountains (e.g., Baars and Ellingson, 1984; Thomas, 2007) and then later by Laramide deformation (Condon, 1995).

The Proterozoic history of the Needle Mountains records two orogenic events, both involving regional metamorphism to greenschist and amphibolite facies and intrusion of dominantly felsic igneous rocks (Gonzales and Van Schmus, 2007). At the end of the first orogeny (Boulder Creek, 1.7–1.6 Ga), more than 1,000 m of quartzarenite and shale of the Uncompahgre Formation were deposited (Tewksbury, 1985). The second orogeny (Silver Plume, ∼1.4 Ga) involved polyphase deformation and intrusion of granitic melts. Probably the last depositional event of probable Precambrian age was the patchy accumulation of up to 23 m of cobble and boulder conglomerates dominated by clasts of Uncompahgre quartzite. Although the conglomerate unit has been considered the basal member of the Ignacio Quartzite for decades, Campbell (1994a, 1994b), who named it the Weasel Skin Member of the Ignacio, thought that the conglomerate might be older than the Ignacio—that is, Precambrian and early or middle Cambrian in age. Condon (1995), Evans (2007), and I agree with the Precambrian age.

Baars (1965) proposed that a structurally high NW–SE-trending fault block of Precambrian rocks existed between Coal Bank and Molas passes that he called the Grenadier Horst or Highland. He and others (e.g., Block, 1986; Wiggin, 1987) considered this element to be the source of much of the sediments in the Ignacio and McCracken units. The history of faults that frame the hypothetical highland and other faults along the western margin of the Animas Valley were discussed by Baars (1958, 1965), Baars and See (1968), Spoelhof (1976), Weimer (1980), Tewksbury (1985), Block (1986), Wiggin (1987), and Thomas (2007). In contrast to Baars (1958), James Evans (personal communication, 2011, 2015) reports that mapping shows that all stratigraphic units are present across the highland proposed by Baars (1965), which negates the existence of a positive element in this area during Paleozoic time.

Cross and co-workers (Cross, 1901, 1904; Cross et al., 1905a, 1905b; Cross and Hole, 1910) first mapped the area and named the stratigraphic units, but provided very sketchy descriptions of them. The oldest Paleozoic unit, dominated by sandstones and named the Ignacio Quartzite, was dated from a single oboloid brachiopod genus as being either Cambrian or Ordovician (Cross et al., 1905a) or middle or late Cambrian (Cross et al., 1905b) (Fig. 3). Overlying the Ignacio with no universally recognized discordance are dolostone, shale, and sandstone with rare Late Devonian fish fossils (Eastman, 1904); this package of rocks was named Elbert Formation by Cross (1904). As noted by Wiggin (1987), little interest was shown in these rocks for 45 years until Read et al. (1949) suggested that the age of the Ignacio was possibly Devonian. Shortly after this, the McCracken Sandstone was named as a lower member of the Elbert Formation in the subsurface in the Paradox Basin by Knight and Cooper (1955). And Cooper (1955, p. 63) reported that the McCracken Sandstone reached the surface making up the upper part of what was identified as the Ignacio Quartzite in the San Juan Mountains. Since then a large literature has appeared concerning the Ignacio Quartzite and its relationship to, and its distinction from, the overlying McCracken Sandstone Member of the Elbert Formation.

Stratigraphic relations between the two sandstone-rich units are complicated by poor exposures near their contact and by the presence of 65 m of relief on the Precambrian surface. Between the southernmost and northernmost localities, there are places where rocks of the Ignacio Quartzite, Elbert Formation, and Devonian Ouray Limestone lap onto the paleosurface of basement rocks (Baars, 1965; Evans, 2007). Thus, the basal Paleozoic stratigraphic unit in the study area is not everywhere Ignacio.

Distinguishing the Ignacio from the overlying McCracken has been controversial for more than a century. Some outcrops and road cuts have been identified as either stratigraphic unit on different field trips and publications. Baars and Knight (1957, p. 117), Baars (1965, p. 24–25), and Baars and See (1968, p. 340) vacillated on whether the two units could be distinguished and whether an unconformity between the two units could be identified. Baars (1965) said McCracken sandstones were whiter, better cemented, harder, more quartzitic, better sorted, and better rounded than those in the Ignacio. He presented 12 measured sections in the study area where he placed a contact between the Ignacio and McCracken. Wiggin (1987) recorded 25 measured sections in the study area where he established contacts between the two units based on the contrast of white sandstones he assigned to the McCracken versus red or maroon sandstones he assigned to the Ignacio. He also illustrated one locality where he reported an angular discordance between two sandstone beds and two localities where he reported evidence of karst features on top of dolostone beds that he assigned to the Ignacio. Rarely do Wiggins' formation contacts agree with Baars' (1965) picks or, more recently, with those of Maurer (2012). Baars (1965) identified the section at Bakers Bridge (my locality A) as a place where the contact between the two units could not be recognized; Wiggin (1987) and Maurer (2012) placed the contact between the two stratigraphic units at that site at different places. I assign all but 4 m of the exposure to the Ignacio Quartzite. As the stratigraphic units have most commonly been identified, exposures of both units are barren of dateable fossils except for five localities where Wiggin (1987) found Devonian placoderm fragments in rocks that he identified as the McCracken Member.

Adding to the controversy are patchy pebble to boulder conglomerates up to 43 m thick (Baars, 1965, p. 149), largely of clasts of Precambrian Uncompahgre quartzite that rest on the Precambrian basement. The age of the conglomerates is disputed (Precambrian, Cambrian, or pre-Elbert Paleozoic), as is their environment of deposition (talus, fluvial, or marine) (Cross, 1910; Baars and Knight, 1957; Block, 1986; Wiggin, 1987; Campbell, 1994a, 1994b, 1996; Evans, 2007). My cursory examination of the conglomerate supports the conclusions of Wiggin (1987), Campbell (1994a, 1994b), Condon (1995), and Evans (2007) that most of them are Precambrian, largely because, as noted by Evans (2007), illite matrix in them has been recrystallized to sericite—a feature not found in Ignacio sandstones.

Disputes over the identities of the Ignacio and McCracken stratigraphic units exist for four reasons: (1) no reference section for the McCracken Member of the Elbert Formation in outcrop was established prior to this study; (2) the nature of the contact has not been described in stratigraphic terms other than unconformable or gradational prior to this study; (3) the contact is generally poorly exposed owing to the weathering of shale beds; and (4) no well-established age exists for the Ignacio Quartzite. Items 1, 2, and 4 have been addressed herein.

Building upon the work of previous investigations, stratigraphic, sedimentologic, and petrographic data were obtained to clarify the stratigraphy of the two units and the distinction between them and also to reconstruct their depositional environments, provenance, and sedimentary history. My study was confined to the rocks above the Precambrian basement exclusive of the basal conglomerate and below the unnamed member of the Elbert Formation or, where the latter is absent, the Ouray Limestone.


For this study, exposures mentioned by previous workers were examined and sampled in addition to previously unmentioned exposures. Several sandstone samples from localities that I did not see were provided by James Evans and David Gonzales. Current limitations of stratigraphic analysis and sampling stem from deterioration of once-pristine road cuts and natural exposures since much of the earlier work was published, especially during the 1950s and ’60s.

Representative sandstone and carbonate samples were taken at each locality. Samples at 14 localities were tied to stratigraphic position based on rough measured sections, whereas grab samples were taken elsewhere. To clarify the distinction between the Ignacio and McCracken, I provide expanded descriptions of the two units at the outcrop and microscopic scale and describe their contacts with each other and superjacent and subjacent units. Two formal members for the Ignacio Quartzite and two lithofacies of the McCracken Sandstone Member were established based on outcrop characteristics and, to a lesser extent, on thin section petrography.

During field reconnaissance for this study, the distinction between the Ignacio Quartzite and McCracken Member was readily apparent based on differences in the color, resistance to weathering, and sedimentary structures of sandstones in addition to stratigraphic position above basement; thus, the stratigraphic units that my samples came from were decided in the field.

Conventional optical study was made of 220 thin sections and 20 polished sections in addition to use of scanning electron microscopy (JEOL JSM-T330A), conventional cathodoluminescence (CL; luminoscope), and scanned CL (Philips/FEI XL-30 with an Oxford Instruments MonoCL system). Point counts (300/sample except where noted) were made of thin sections of sandstones using a petrographic microscope. Samples were stained for K-feldspar using sodium nitrocobaltate. “Dust lines” on detrital quartz grains are absent or incomplete in many samples viewed in transmitted light, so data on the relative abundance of quartz cement versus detrital grains are of variable reliability. Many samples, however, have either iron-oxide grain coats or “ghost dust lines” of former iron oxide grain coats that are visible in reflected light. These samples were point counted using a combination of transmitted and reflected light.

For each sandstone thin section, the diameter of a representative grain was measured to characterize average grain size; sorting was estimated from the graphic charts of Harrell (1984). For bimodal samples, the size of each mode was estimated.

Detrital clay occurs in the sandstones as millimeter-scale drapes, clay clasts, and matrix. All categories of these clays are optically pure illite and lack organic matter and silt grains. If a patch of clay could be recognized as a mashed clast, it was tabulated as a clay clast. If the clay patch could not clearly be so identified, it was tabulated as matrix. Matrix and clay clasts were probably derived from clay drapes. Matrix was either introduced during bioturbation or was produced as strongly mashed clasts (= pseudomatrix).

To aid in provenance interpretation of sandstones, the age of detrital zircons was determined for two samples of Ignacio and two samples of McCracken sandstones. Standard mineral separation techniques, including heavy mineral and magnetic susceptibility separation, were employed to extract zircon from sandstone samples. Separated zircons were sprinkle-mounted onto double-sided tape on one-inch acrylic discs and analyzed at random using depth-profiling laser ablation, inductively coupled plasma–mass spectroscopy, uranium-lead geochronology (LA–ICP–MS U-Pb geochronology). For each sample at least 100 zircons were analyzed to obtain a detailed provenance data set that resolves any age component >5%. The analyses were completed using a Photon Machines Inc. Analyte G2 Excimer laser (30 μm laser spot size) with a large-volume Helex sample cell and a Thermo Scientific™ ELEMENT 2™ ICP–MS. GJ1 was used as the primary reference standard and a secondary in-house zircon standard (Pak1 with a TIMS 206Pb/239U age of 43.0 Ma). Depth profiling of non-polished, tape-mounted zircons enables the resolution of multiple zircon growth zones evident from core and rim ages. The data from the analyses were then reduced using the Iolite data reduction software and VizualAge. For analyzed detrital zircons, the 206Pb/238U age is used for grains younger than 850 Ma, and 207Pb/206Pb age is used for grains older than 850 Ma. Mineral separation and LA–ICP–MS analyses were completed at the UTChron facilities at the Jackson School of Geosciences at The University of Texas at Austin.

Six samples of dolostone from the Sultan Creek facies of the McCracken Member were processed for conodont samples in an attempt to establish a paleontological age of the member. Dolostone samples were dissolved in buffered acetic acid, and the acid-insoluble residues were treated in methylene iodide to separate the light and heavy mineral fractions. This procedure preserves conodonts, which are composed of apatite.

The paleomagnetic signature of the Ignacio was also examined to try to date the member. Twenty-five one-inch-diameter plugs of red sandstones from the Ignacio Quartzite at localities G and H were analyzed for their remnant magnetism. The analytical technique is described by Holt et al. (2000).

The term “bed” as applied to sedimentary rocks has so many different uses (e.g., McKee and Weir, 1953; Ingram, 1954; Campbell, 1967; Brandt, 1986) that the word is ambiguous. I use “bed” here for a layer distinctively different from neighboring layers in composition (e.g., sandstone versus shale), grain size, or bedding structure. These differences are generally manifested by differences in resistance to weathering and result in either recessive or non-recessive layers, i.e., beds are chiefly “weathering units.” Beds in the study area range from barely recognizable clay films to layers of sandstone or dolostone 2 m thick. The term “sedimentation unit” (Otto, 1938) is used here for a layer that formed as a single depositional event under essentially uniform conditions, e.g., clay drape or single trough cross-bed. Some beds are single sedimentation units, but most encompass more than one sedimentation unit (they are amalgamated sedimentation units).


General Comments

The Ignacio Quartzite was named by Cross (1901) for quartzite and interbedded shale up to 65 m thick that comprises the basal Paleozoic clastic rocks in the San Juan Mountains. No description, however, was given until Cross et al. (1905a, p. 3) noted: “The greater part is fine grained, white, gray, or pinkish, and highly indurated. The lower portion is commonly a massive quartzite of prevalent pink or reddish color, while the succeeding strata are nearly white.” The white quartzite layers in the upper part of the sequence have since been assigned to the McCracken Member of the Upper Devonian Elbert Formation, i.e., Baars and Knight, 1957; Baars, 1958, 1965; Wiggin, 1987. Cross et al. (1905a, p. 3) designated the stratotype of the Ignacio Quartzite to be exposures along Elbert Creek near Ignacio Lake. They misidentified Electra Lake as Ignacio Lake because good exposures of sandstones occur only along the west flank of Electra Lake, which is slightly west of the much smaller Ignacio Lake.

Baars and coworkers (Baars and Knight, 1957; Baars, 1958, 1965; Baars and See, 1968; Baars and Ellingson, 1984), Rhodes and Fisher (1957), Wiggin (1987), Campbell and Gonzales (1996), and Maurer (2012) greatly expanded stratigraphic details and descriptions of the rocks targeted for this study, but almost no information was given on the nature of contacts of the stratigraphic units. Baars and co-workers (Baars and Knight, 1957; Baars, 1958, 1965; Baars and See, 1968; Baars and Ellingson, 1984), Block (1986), and Wiggin (1987) reported rapid lateral facies changes and thickness variations of the Ignacio (scale of hundreds of meters) and attributed much of these variations to topographic elements of structural origin (horsts and grabens) developed during late Precambrian tectonics. Spoelhof (1976), Weimer (1980), Baars et al. (1987), Condon (1995), Evans (2007), and Thomas (2007) also explored the relation between tectonics and sedimentation. Maurer and Evans (2011) and Maurer (2012) attributed some of the rapid facies changes to be the result of backfilling a valley incised into the Precambrian basement during the Late Devonian.

Campbell (1994b) introduced the name Weasel Skin Member for the basal conglomerate of the Ignacio, and Stag Mesa Member for other Ignacio rocks. Campbell and Gonzales (1996) and Gonzales et al. (2004, p. 15) described the Stag Mesa Member as chiefly “mottled reddish-brown to grayish-brown, white, or light pink coarse-grained quartzarenite and pebble conglomerate.” Stratotypes and stratigraphic details of the two named members were not described in the Campbell report (1994b) or later publications, so those names have no validity according to the North American Stratigraphic Code (North American Commission on Stratigraphic Nomenclature, 2005).

New Members of the Ignacio Quartzite

The members of the Ignacio Quartzite proposed by the authors cited above do not provide a useful subdivision of the Ignacio as characterized in this study; thus, I establish here two new formal members of the Ignacio Quartzite to provide a means of displaying the three-dimensional geometry of the units and aid in interpreting environments of deposition. The members are named after geographic features near which they are well exposed and include the Tamarron Member (Figs. 4–6) and overlying Spud Hill Member (Fig. 7). Names of the two new members have not been used previously in North America and have been approved by the secretary of the Geologic Names Committee of the National Geologic Map Database project (N. Stamm, personal communication, 2016). The geometry of the members is shown on cross sections and in a map view (Figs. 8–10).

Tamarron Member of the Ignacio Quartzite

Well-exposed beds along the west margin of what is now Electra Lake were apparently identified by Cross et al. (1905a) as the type section of the Ignacio Quartzite. Construction of the dam for the lake drowned Elbert Creek in this area, but about 23 m of a distinctive and widespread facies of the Ignacio is exposed on the lake margin (locality E) except for its base, which is below lake level. I establish the type section of the Tamarron Member about 12 km south of Electra Lake on the east-facing slope of the hill where La Plata County Road 250 crosses the Animas River at Bakers Bridge (lat 37.458549 N, long -107.800852 W). The location of coordinates cited in this article can be viewed at At this site, an essentially complete section of the member is exposed between the underlying Bakers Bridge Granite (Precambrian) and overlying Elbert Formation (Upper Devonian). The member is named for the resort community of Tamarron, La Plata County, Colorado, where the member has numerous exposures along streets in the community. (Coordinates for the Glacier Club in the center of the community are lat 37.500261 N, long -107.807357 W). A complete continuous section of the Tamarron Member, however, is not exposed in the resort village itself.

Both new members are characterized in outcrop by well indurated reddish-brown sandstones except where bleached (Figs. 5–6), dominance of planar laminations and cross-beds, and thin red or green shale interbeds. All sandstones have >5% feldspar (Fig. 11; Table 1), bimodal grain size, poorly sorted texture, abundant polycrystalline quartz, dominance of subangular to subrounded quartz grains, and scarcity of well-rounded spherical quartz grains that are typical of the McCracken Member. Twenty-five percent of samples also contain more than 5% rock fragments, nearly all of which are shale rip-up clasts. Extrabasinal rock fragments rarely exceed 2% and include sandstone, siltstones, weakly foliated equivalents of the sandstone and siltstone grains plus schist/phyllite and metachert (rare). Biotite, muscovite, chlorite, zircon, tourmaline, and opaque heavy minerals are present in small amounts in fine-grained beds (data from Goldstein inParker and Roberts, 1963; this study). Oboloid brachiopods occur as transported clasts in sandstone beds at localities H and Q. Quartz is the dominant cement except in the uppermost meter or so where the Tamarron Member becomes dolomitic.

The Tamarron Member extends from the southernmost locality (A, Bakers Bridge) nearly to Coal Bank Pass, although there is a 12-km gap with no data (Fig. 9). This member comprises the bulk of the Ignacio Quartzite.

The Tamarron Member is composed chiefly of variegated (brown, purple, white, red) sandstones (90–95%) in sharp-based beds from 30 cm to 2 m thick separated by shale (5–10%) beds from drapes to beds 15 cm thick (Table 2; Figs. 4–6). Sedimentation units in sandstones, however, range from 1 to 40 cm. Most sandstone beds are laterally continuous for the width of each outcrop (Fig. 4). Scour at the base of sandstone beds reaches 10 cm, but generally is much less, and meter-scale channeling is absent. Sedimentation units are composed chiefly of laminated beds and tabular and trough crossbeds with lesser structureless beds and rare current ripple cross-beds and tidal couplets (Figs. 5 and 12). The bedforms that generated cross-beds were nearly all decapitated subsequently by erosion, such that their original thickness is rarely preserved. Current ripples and current-ripple cross-beds are rare. Cross-beds seen during this study chiefly have bimodal dips to the W and E with the W direction dominant. Maurer (2012) found chiefly NW dips of cross-beds that I assign to this member. Wiggin (1987) reported hummocky cross-beds to be the dominant bedding type in this member, and Maurer (2012) reported them to be common, but I recognize none (Table 2). Many bedding planes, especially those that can be traced laterally more than 2 m, are stylolites that developed at clay drapes. Stylolite amplitudes do not exceed 2 mm. Mud cracks are rare as are curled mud flakes that likely had microbial binding (cf. Eriksson and Simpson, 2012). Rare horizontal trace fossils (Planolites) occur in centimeter-thick shale layers, but only in beds at least 5 m above the base of the member.

Sandstones are mostly bimodal with grain-size modes in the fine and coarse/very coarse sizes. Owing to the bimodality, overall sorting is poor to very poor (Table 1), although individual laminations tend to be moderately well sorted. Very poor sorting is a common result of “cryptic bioturbation” (see below).

The sandstones are strongly cemented by quartz and fit the field term “quartzite,” but generally they lack the hardness of the quartzarenite sandstones typical of the McCracken Member. Of the several populations of detrital quartz described later, well-rounded spherical grains rarely exceed 3%. Conglomerates 5 cm thick with angular and rounded vein-quartz granules and pebbles occur in beds scattered throughout the facies; some units also contain pebbles of pink Uncompahgre quartzite and shale rip-up clasts. Pebbles of milky vein quartz up to 10 cm are most abundant near the base of the member, but there is no overall upward fining in grain size except for a slight increase in shale beds up-section.

Pink K-feldspar is the diagnostic framework grain component of this member, and all grains have well-developed overgrowths. Feldspar abundance decreases upward from 25% (rarely 40%) to 4% of framework grains (Table 1) and is most abundant in the fine-sand fraction (cf. Odom et al., 1976) as noted by Wiggin (1987). Clay rip-up clasts of pebble size are rare, but several percent of sand-size clasts of rip-ups are common, and there are rare sandstone beds composed of greater than 40% red clay clasts. Compaction of the ductile clasts in these beds plugged all pores to the exclusion of cement. These beds weather recessively owing to their high clay content and consistently have been misidentified as shales. Minor grains of extrabasinal origin include siltstones, shales, schist, muscovite, biotite, and ultrastable heavy minerals (zircon, tourmaline) plus lesser ilmenite, magnetite, and leucoxene (Goldstein inParker and Roberts, 1963).

The Tamarron Member has a distinctive color mottling in which various shades of red (weak red, 10R, 5R 5/2, 2.5YR 4/2; dark reddish brown, 5YR 4/2) to purplish red color (pale red, 7.5R 6/4; reddish gray, 5R 6/1) are common. The strongest pigmentation is in clay clasts and drapes. Although the “red” color is not as bright as that of prominent Mesozoic sedimentary rocks exposed in the study area, unbleached sandstones of the Tamarron Member qualify to be labeled “red beds.” The red/purple/brown sandstones of this member have prominent lenticular white mottles (pinkish white, 5YR 8/2; 20% of beds) that are elongate in the plane of bedding and which also cut across bedding (Figs. 5–6). Thin sections show that iron-oxide grain coats that are present in the red beds are absent from the white lenses and mottles. This color pattern is typical of sandstone beds that have undergone the uneven removal of iron-oxide grain coatings during a diagenetic bleaching event resulting from the invasion of hydrocarbons and hydrogen sulfide (H2S) into “red beds.” In some outcrops of this member, bleached beds comprise 80% and may be the reason that some Ignacio outcrops have been misidentified as McCracken.

Planolites, the only trace fossil that I recognized in the Tamarron Member, is sparse in the clay drapes and interbeds. At the thin section scale, however, many sandstone laminations are discontinuous, and coarse and fine grains are randomly scattered. This indicates that there was destruction of laminations by “cryptic bioturbation” (cf. Miller, 1977; McBride, 2012), where animals mine bedding planes, but do not leave recognizable burrows. Trace fossils and disturbed laminations have not been found in the lower 5 m of the member. Planolites occurs bimodal at the millimeter- and centimeter-wide scales (Tables 2–3). No skeletal fossils were found in this member.

At most localities, only 4–6 m of the member are exposed, generally within a meter or two of underlying basement rocks or, rarely, the patchy basal conglomerate of apparent Precambrian age. At the stratotype (locality A, Bakers Bridge), the member rests with several meters of relief on the Bakers Bridge Granite. The basal contact is covered by talus and water at the type section of the Ignacio Quartzite at Electra Lake. In the southern part of the study area the Tamarron Member comprises the entirety of the Ignacio Quartzite. But at locality Q (Canyon Creek), it is abruptly overlain by the overlying Spud Hill Member of the Ignacio Quartzite and at locality O (Deadwood Creek Gulch) grades into the overlying Sultan Creek facies of the McCracken Member of the Elbert Formation. At locality B (Shalona Lake railroad cut), 2 m of pebble conglomerate (Mill Creek facies) form the base of the exposure. Lenses of conglomerate are clearly interbedded with feldspathic sandstone of this member at this site. This is the only locality where conglomerate shows such interbedding with feldspathic sandstone that I assign to the Ignacio. Pebbles are well-rounded Uncompahgre quartzite and lesser vein quartz.

The upper contact of the member is generally covered, presumably by weathered shale of the upper member of the Elbert Formation, but the Tamarron Member has diverse stratigraphic relations at its upper contact as shown in Figures 13A and 13B. At locality A and north to B (Bakers Bridge and Shalona Lake railroad cut, respectively), typical Tamarron Member sandstone is capped by about 4 m of sandy dolostone (G. Gianniny, personal communication, 2012) with probable pedogenic fabric (Fig. 14) and brecciated sandy dolostone clasts; at locality C (Glacier Club), 6 m of Tamarron Member is overlain by 2 m of Spud Hill Member, which is followed by 3 m of interbedded sandy limestone and calcitic sandstone assigned to the Mill Creek facies of the McCracken Member. The latter, in turn, is capped by mottled limestone assigned to the upper member of the Elbert Formation before losing the outcrop to cover. At localities D and D’, Tamarron sandstone is capped by a dolomitized carbonate mudstone bed and then cover. The upper centimeter of the sandstone bed at locality D is a lag deposit containing quartz granules and abraded fragments of placoderm plates bored by Trypanites. It is a matter of preference to which stratigraphic unit these carbonate beds at the top of the Tamarron Member should be assigned. I assign the stratigraphically lowest occurrence of shale and carbonate beds thicker than 1 m to the Elbert Formation or to the Ouray Limestone as appropriate. And, as noted here, at localities F (Cascade Creek) and Q (Canyon Creek), the Tamarron Member is overlain by the Spud Hill Member, but the Tamarron Member is overlain by the McCracken Member at locality O (Deadwood Creek Gulch).

The upper 1.5 m of the member at Bakers Bridge (locality A) consist of light brown (2.5YR 5/4) sandstone beds containing lenses and pods of dolostone (dolomitized grainstone) at the decimeter scale. The lenses are depositional layers, but the pods are clasts. This bed and the sandstone bed below it have uneven stratiform patches of dolostone with nebulous contacts with the host sandstone (Fig. 14A). Crinkly anastomosing lamination and concretion-like bodies in the dolostone resemble pedogenic features reported in caliche (Brewer, 1964, p. 318; James, 1972, his fig. 4B). The dolomite overprint on original limestone textures has obliterated microscopic pedogenic textures that may have existed originally (Fig. 14B). I assign this carbonate unit to the McCracken Member on the basis of the extremely well-rounded texture of the quartz grains. This thin McCracken sandy carbonate is overlain abruptly by the poorly exposed upper member of the Elbert Formation. The maximum thickness of the Tamarron Member at Bakers Bridge is 23 m.

Spud Hill Member of the Ignacio Quartzite

The Spud Hill Member is a redder, shalier, more trace-fossil-rich, and generally finer-grained version of the Ignacio than the Tamarron Member. It is named for Spud Hill (also known locally as Potato Hill; elevation = 3,612 m; GPS = lat 37.672501 N, long -107.766733 W) in La Plata County, Colorado. This modest peak is 2.5 km east of Mill Creek Lodge, the latter of which is located on U.S. Highway 550, known as the “Million Dollar Highway.” The type section is established on the south-facing road cut on U.S. 550, 3.7 road kms south of Coal Bank Pass and begins where the highway crosses the rivulet of Mill Creek (lat 37.680041 N, long -107.785256 W) (locality I [Mill Creek road cut]; Fig. 7). The Spud Hill Member here rests on the Precambrian Twilight Gneiss and is overlain gradationally by the McCracken Member of the Elbert Formation (Upper Devonian). Features of its contacts elsewhere are shown in Figures 8, 13A, and 13B.

It is a heterolithic member composed of 75–90% feldspathic sandstones, 8–23% weakly fissile sandy shales, and 2% carbonates, the latter of which are restricted to its top. Shale beds occur evenly distributed in places and clustered elsewhere; they are most abundant in the upper few meters. Thin carbonate beds occur in places only at the gradational contact with the overlying McCracken Member. Although the color chart shows the beds are mostly reddish brown (2.5YR 4/2, 5Y 4/2), most outcrops easily fit the field term “red beds” and exposures are noticeably redder than the Tamarron Member.

Sandstones are subequal amounts of hard, well-indurated beds (Fig. 15) and poorly resistant “shaly” beds that weather recessively (Fig. 16). The recessive layers are either sand and clay layers that have been mixed during bioturbation to form clayey sandstone or compacted sandstone composed of more than 40% clay rip-up clasts. Clay clasts in the latter beds have been mashed during compaction to fill all pores in the form of “pseudomatrix.” Except for the greater abundance of clay clasts, modal composition of sandstones in this member is similar to the feldspathic sandstones of the Tamarron Member, and K-feldspar content generally exceeds 10% (Table 1). Bleached beds are far less abundant in the Spud Hill Member. The upper half of the Spud Hill Member at the reference locality I (Mill Creek road cut; Fig. 7) has been thoroughly bleached. Here, however, removal of iron-oxide pigment permits the green (5G 7/2) color of detrital illite to dominate the beds.

Unusual but distinctive rip-up clasts of phosphate nodules occur at locality I, where millimeter-thick phosphate crusts occur also in shale (Fig. 17). The rip-up clasts include phosphate-cemented sandstones, where the phosphate is largely a passive cement, to clasts whose grains of K-feldspar and micas have been replaced to various degrees by it.

Sandstone beds in general are thinner than those in the Tamarron Member. Beds range from 5 to 50 cm thick and are mostly continuous within an outcrop; sedimentation units are from 1 mm to 20 cm thick. Owing to the overprint of bioturbation (10–100%), the identity of most sedimentation units thinner than 2 cm has been destroyed, but vestiges of heterolithic interbeds of sandstone-shale at the centimeter and decimeter scale typical of intertidal deposits (Reineck and Wunderlich, 1968; Fig. 16 this paper) survive. Because of the high degree of bioturbation, individual traces are not identifiable within beds. Present locally are event beds—those with sharp bases and bioturbated tops (c.f. Brandt, 1986) up to 20 cm thick. Sedimentation units are texturally mottled, laminated, structureless, and current rippled, or they display rare tabular or trough cross-beds, HCS (hummocky cross-stratification), and flaser bedding (Fig. 16).

In contrast with Wiggin (1987) and Maurer (2012), I recognized only rare examples of HCS, all with wavelengths of 80–100 cm and amplitudes less than 10 cm, which is similar to the Tamarron Member.

Most current ripples migrated northeastward, but some are bipolar (NE–SW; Fig. 16). A few beds have decimeter-size phosphate nodules (Fig. 17). Meter-scale channeling is absent, and scour at the base of sandstone beds is less than 10 cm. Many bedding surfaces are indistinct owing to a bioturbation overprint, whereas those resting on shales have sharp bases and abundant trace fossils (Figs. 18–20). Mud cracks are present, but they are rare (cf. Baars, 1965; Wiggin, 1987; Table 2, this report).

Similar to the Tamarron Member, most sedimentation units in the Spud Hill Member are bimodal in grain size with modes of fine and coarse/very coarse sand, so overall sorting is poor to very poor (Table 1). The majority of thin sections show laminations, although many laminations have been disturbed or destroyed by bioturbation. Granule- to pebble-conglomerate sedimentation units are scattered throughout. They reach 15 cm thick, but most are less than a centimeter. Some pebbles occur as lags on top of sandstone beds (Maurer, 2012). Vein quartz, including angular and rounded grains, comprises 95% of coarse clasts with Uncompahgre quartzite clasts comprising the remainder. Although the coarsest pebbles occur near the base of the Spud Hill, the member does not fine upward perceptibly. A sandstone from locality H (Mill Creek Lodge) contains fragments of both lingulids and fish plates. Rare shale rip-up clasts reach 8 cm. No stylolites occur in this member.

Shale beds are mostly less than 5 cm thick and poorly fissile owing to considerable fine sand and silt that were mixed with clay during bioturbation. Only clay drapes thicker than 2 mm extend more than 2 m laterally. Shales are red except where they were bleached green like at locality I (Mill Creek road cut; Fig. 7). Illite is the only clay mineral recognized. Quartz, feldspar, biotite, and muscovite are scattered throughout. Black phosphate oboloid shells 1.5 cm in length occur in a greenish-gray sandy shale and are preferentially aligned with their long dimension in the bedding plane, indicating that they are not in growth position. Scraps of phosphatic brachiopods of unidentifiable genus occur in thin sections of sandstone beds from localities G and H. Syndepositional phosphate occurs as millimeter-scale crusts and pea-size nodules in strongly bioturbated shale and as pea-size rip-up clasts at localities I and at P (Fig. 17).

Nearly all exposed bedding planes developed at shale drapes, and beds display trace fossils. Planolites of two sizes (mm and cm diameters) dominates the trace fossil assemblage (Fig. 18). Trichophycus (Fig. 19), Rusophycus (Fig. 20), Monomorphichnus, and Lockeia are rare, though Trichophycus is locally abundant (Table 3). The first four named genera were probably produced by arthropods, such as trilobites, but no skeletons were found. Lockeia, a brachiopod resting trace, may have been made by the oboloids. Wiggin (1987, p. 118) reported the presence of Corophoides and its bedding-plane analog Arenicolites in the Ignacio at Mountain View Crest, possibly from the Spud Hill Member. Maurer (2012) reported a broader suite of trace fossils, possibly from this facies, but he included taxa (e.g., Ophiomorpha) not known to occur in rocks as old as these.

Rhodes and Fisher (1957) measured a stratigraphic section at locality I when the road cut was new and nearly continuously exposed from Precambrian Twilight Gneiss to the Devonian Ouray Limestone. At present, 21 m of cover separate the member at this site from the underlying Precambrian phyllite. Two meters of shale that separate it from the overlying McCracken Member are poorly exposed. At this locality, Rhodes and Fisher (1957, p. 2,511) described a scoured surface 2 m deep that was situated 80 m above the contact with the Precambrian basement. Baars and Knight (1957) surmised that this surface was the disconformity between the Cambrian Ignacio Quartzite and Devonian McCracken Member. This surface is no longer visible.

The Spud Hill Member is exposed in two belts. The southern belt connects localities C and Q (Glacier Club and Canyon Creek, respectively), while the northern belt occurs between locality G and I (Lime Creek Road and Mill Creek road cut, respectively; Fig. 10). The member has a maximum thickness of 21 m. It is not exposed north of Coal Bank Pass. It apparently rests on the Precambrian basement at localities G, H, and I, but overlies the Tamarron Member at localities C and Q. The Spud Hill Member is separated from sandstone of the overlying McCracken Member by 1 m of shale at the type locality (I) or by interbeds of sandstone, carbonate, and shale of the Mill Creek facies at localities G, H, and Q (Fig. 8). At locality H, a mixed sand-carbonate bed has isopachous quartz overgrowths that are common in pedogenic silcretes.

Basal Contact

The base of the Ignacio has been placed at the contact of sandstone or, where present, of pebble/cobble/boulder quartzite conglomerate, with the underlying Precambrian igneous or metamorphic rocks. The quartzite conglomerate is, as noted, of variable thickness and patchy distribution. For years the quartzite conglomerate has been considered to be the basal Ignacio. Campbell (1994a, 1994b) and Campbell and Gonzales (1996) considered the conglomerate might possibly be Neoproterozoic to middle Cambrian in age. Condon (1995) and Evans (2007) endorsed the probable Precambrian age—as do I. Wiggin (1987) suggested that there are coarse conglomerates of two different ages: Precambrian and pre-Devonian Paleozoic.

Upper Contact

Cross et al. (1905a, 1905b) did not describe the contact of the Ignacio Quartzite with the overlying Elbert Formation. Barnes (1954) reported quartzite of the Ignacio to be interbedded with shale, quartzite, and limestone of the Elbert Formation within 7 m of section in the vicinity of the Tamarron Member. Although various other workers have mapped and recorded measured sections of the two controversial units (Baars, 1965, nine measured sections; Block, 1986, nine sections; Wiggin, 1987, 25 sections; Maurer, 2012, 11 sections), only Baars (1965, p. 24) and Wiggin (1987) described the lithologic criteria that they used to distinguish between the top of the Ignacio versus the base of the McCracken. However, the stratigraphic details of the contact have not been described.

Inspection of the measured sections of these workers, except for Maurer (2012), indicates that the contact with the McCracken Member was commonly placed where non-white (brown, red, or maroon) sandstones (Ignacio Quartzite) are overlain by white, generally better-cemented sandstones (McCracken Member). Wiggin (1987) placed all placoderm-bearing sandstone beds near the contact at five localities where he found them in the McCracken.


Thickness values given for the Ignacio Quartzite vary because of disagreement among workers in identifying the contact with the overlying McCracken Sandstone Member, because of disagreement on which exposures are Ignacio Quartzite versus McCracken Member, and because the sandstones were deposited on a paleosurface of considerable relief (Baars, 1965; Evans, 2007). Most workers report the Ignacio Quartzite to be less than 25 m thick, to thin toward paleohighs, and, according to Baars (1958, 1965), to be absent on paleohighs. Gonzales et al. (2004) reported a maximum thickness of 30 m in the Electra Lake Quadrangle; Baars (1965) reported a maximum thickness in the study area of 46 m; and I found the maximum thickness to be 32 m.

Age and Correlation

The only known skeletal fossils recovered from the Ignacio prior to this study are three genera of phosphatic brachiopods. A single taxon collected by Cross et al. (1905a) from Overlook Point on Mountain View Crest south of the Needle Mountains was identified as Obolus sp.?, similar to those found in middle to late Cambrian rocks of the Western United States. A larger collection of brachiopods found by Rhodes and Fisher (1957) from my locality I (Mill Creek road cut) was identified for these authors by G. A. Cooper as equivocally similar to Obolus of late Cambrian to Early Ordovician age. The inability to examine muscle scars inside fragile shells of the oboloids prevented a reliable age determination of the taxon. Baars (1965, p. 19) identified a few examples of Lingulella (middle Cambrian–Late Ordovician) and Dicellomus (late Cambrian), also lingulid brachiopods, from localities F and I. He reported uncertainty in distinguishing between the two genera and, consequently, uncertainty about their usefulness in limiting an age for the Ignacio.

An inferred correlation of the Ignacio with other basal Paleozoic sandstones in the Western U.S. (e.g., Dresbachian Sawatch Quartzite, Ross and Tweto, 1980; early–middle Cambrian Tapeats, Tweto and Lovering, 1977; Cambrian Tintic Quartzite, Baars, 1965, p. 20) and the tentative age assignment of Obolus sp. led to the general acceptance that the Ignacio is of late Cambrian age (e.g., Baars and See, 1968, p. 339). Read et al., 1949, however, gave strong reservations about the Cambrian age. Barnes (1954) suggested that the Ignacio was probably Devonian because Ignacio sandstones graded upward into dolostone and shale of the overlying fossiliferous Elbert Formation, and Baars and Knight (1957) suggested that both Cambrian and Devonian rocks might be present in the Ignacio as originally defined. Recently, Evans (2007), Maurer and Evans (2011), and Maurer (2012) proposed that the Ignacio Quartzite is a transgressive unit of an Upper Devonian stratigraphic sequence that includes the Ignacio Quartzite-McCracken Member-upper Elbert Member-Ouray Limestone. The chronologic changes in age assignments for the Ignacio Quartzite over the years are shown in Figure 3.

Wiggin (1987, p. 16–43) marshaled the strongest arguments in support of a Cambrian age for the Ignacio Quartzite and for the presence of a regional unconformity between the Ignacio Quartzite and the McCracken Sandstone. Wiggin's evidence in support of the Cambrian age of the Ignacio Quartzite includes: (1) the age assignments of the Obolus sp. brachiopods as described here; (2) subsurface correlation with the better-dated Tintic and Tapeats sandstones; and (3) surface correlation with the better-dated Sawatch Quartzite. Evidence for a major unconformity between the two units includes: (1) the differences in ages of the Ignacio brachiopods (Obolus and Dicellomus genera) and the Late Devonian placoderm fossils of the McCracken Member; (2) the contrast in sandstone composition between the red feldspathic Ignacio sandstones and the white feldspar-poor McCracken sandstones; and (3) field criteria for an unconformity at the Ignacio-McCracken contact based on his work. The latter includes: (1) an angular discordance at the base of a sandstone bed at locality N (Sultan Creek north); (2) sandstone of the McCracken infilling a decimeter-scale cave formed in Ignacio dolostone; and (3) McCracken sandstone infilling a dissolution-enlarged joint formed in the Ignacio dolostone at locality L. (The location of locality L is shown in Figure 1, but no samples were collected at that site, so there are no data for the locality in Table 1.) If interpreted correctly, these latter three features are evidence of an unconformity. Without chronostratigraphic control, however, the temporal significance of an unconformity is unknown.

Wiggin (1987, p. 33) pointed out the weaknesses of his evidence, including the critical point that a fully reliable age for the Oboloid brachiopod taxon has not been possible. In regard to finding a discordance at the base of a sandstone bed, some degree of erosion can be expected at the base of any estuarine sandstone bed. Such discordances need not be regional unconformities. I searched for Wiggin's examples of karst features, but was able to find only the alleged dissolution-enlarged joint feature he described. It is certainly an unusual feature, but an alternate tentative interpretation is that it is a clastic injection into underlying beds rather than an open-cavity filling.

The search for conodonts in limestone beds at the top of the Ignacio (Tamarron Member) during this study was unsuccessful as was the attempt to date several red sandstone beds from the Spud Hill Member using their paleomagnetic signatures.

An Ignacio sandstone from locality H contains fragments of both oboloid brachiopods and fish plates. Both taxa are too fragile to have been reworked from older beds, so the conclusion is that the two taxa are coeval and, therefore, Late Devonian in age. In addition, of the 100 detrital zircons from the Ignacio that were dated by U-Pb geochronology (see below), one grain yielded an Ordovician age (460 Ma). These data discredit the late Cambrian age of the Ignacio and adds credence to the Late Devonian age of the Ignacio revived most recently by Maurer and Evans (2013).



Cross (1901, 1904) named the Elbert Formation and designated exposures along Elbert Creek north of Rockwood, Colorado, as the type area. The only measured section and rock description that Cross provided of the Elbert are from exposures 16 km to the east of the type area at “Devon Point.” The presence of fish plates and scales established a Late Devonian (late Frasnian–Famennian) age for the Elbert in outcrop (Eastman, 1904; Cross and Larsen, 1935; Thomson and Thomas, 2001).

More than half a century later, the McCracken Sandstone Member of the Elbert Formation was proposed as a new stratigraphic unit for 34 m of sandstone with interbeds of dolostone recovered in a core in the Paradox Basin, San Juan County, Utah (Cooper, 1955, p. 63; Knight and Cooper, 1955, p. 56).

Cooper (1955) and Baars and Knight (1957) noted that the McCracken Sandstone of the Paradox Basin might reach outcrop in Colorado's San Juan Mountains, but Baars (1965) argued against designating a type section owing to the uncertainty of this correlation. The failure to establish a proper reference section in outcrop has subsequently led to chaos. Although Baars (1965) later doubted that the subsurface McCracken Member reached the outcrop, the use of McCracken Sandstone Member or McCracken Member for an outcrop stratigraphic unit at the base of the Elbert Formation has prevailed in subsequent reports.

Descriptions of the McCracken by different workers differ depending on their assignment of an outcrop to the Ignacio Quartzite versus the McCracken Member. Baars (1965, p. 24–25) described sandstone beds in the McCracken to be whiter, better cemented, harder, more quartzitic, better sorted, and better rounded than those in the Ignacio.

Neither Wiggin (1987) nor Maurer (2012) summarized the character of the McCracken Member nor described its contacts, but both provided descriptions of McCracken rocks as they identified them in their measured sections. Wiggin (1987) reported 14 m as the maximum thickness of the McCracken Member.

As noted earlier, my criteria for assigning rocks to the McCracken Member versus the Ignacio Quartzite in the study area are based on clear differences in outcrop features. In addition, there are similarities between the McCracken Member in the subsurface and surface as described by Cooper (1955), Knight and Cooper (1955), Cole and Moore (1996), and Peterman and Cole (2001). I also examined 19 thin sections of the McCracken Member from the B-614 well in the Lisbon Field, Colorado (Appendix B), to compare with samples from the Animas Valley. Cole and Moore (1996) describe the stratigraphy and sedimentary history of the McCracken in this field, located about 85 km northwest of Durango, Colorado. Characteristics of McCracken Member sandstone include white/off-white color, extreme hardness owing to pervasive quartz cement, near restriction of trace fossils to Planolites, absence or near absence of feldspar and polycrystalline quartz, abundance of well-rounded and spherical quartz grains, and local presence of placoderm fossils.

Facies of the McCracken Member

In order to best display the three-dimensional stratigraphic features of the McCracken Member, I establish here two lithostratigraphic facies: the Mill Creek and the Sultan Creek facies. I use “facies” as defined in the Glossary of Geology: “(b) a mappable, areally restricted part of a lithostratigraphic body, differing in lithology or fossil content from other beds deposited at the same time and in lithologic continuity. Cf: sedimentary facies (Neuendorf et al., 2005).” The reference section for the Mill Creek facies (Figs. 21–22) is the same road cut as the stratotype for the Spud Hill Member as identified above (locality I; Fig. 7). The Sultan Creek facies reference section of the McCracken Member is the east-facing slope east of U.S. Highway 550 that is 100 m south of Sultan Creek (locality M [Sultan Creek south]). With the exception of the Sultan Creek facies, my distinction between the Ignacio and McCracken units agrees closely with Wiggin (1987), moderately well with Baars (1965), but not at all with Maurer (2012).

Sandstones of both facies of the McCracken Sandstone Member are nearly all quartzarenites with less than 5% detrital K-feldspar (Fig. 11). Three dominant and several minor populations of quartz sand comprise the sandstones. The largest population is fine- to medium-grained, subangular to subround quartz with up to 5% polycrystalline grains. Next in abundance are well-sorted, medium-grained, rounded to very-well-rounded (“billiard-ball”) grains (Fig. 23) that are nearly all monocrystalline and non-undulose quartz. This population dominates in some beds of the Mill Creek and in a few beds of the Sultan Creek facies. Least abundant of the major populations are pebbles of milky vein quartz showing a range of roundness values. Lesser populations are sliver-shaped fine grains, some highly angular and others with rounded edges; well-rounded coarse/very coarse sand; and well-rounded very-fine to silt-size grains.

The abundance of the three main quartz populations varies both regionally and stratigraphically without predictability. A percent or two of feldspar occurs in some samples, particularly in the finest grain-size fraction. Sand-size clay rip-up clasts are widespread minor components and were a source of pseudomatrix. Rare detrital components include rock fragments (schist, weakly foliated shale, and sandstone), zircon and tourmaline, muscovite, placoderm fish plates, and glaucony (see also Wiggin, 1987). (Glaucony is the preferred term for green clay pellets because many such grains are not composed of the mineral glauconite [Odin and Letolle, 1980].)

Mill Creek Facies of the McCracken Member of the Elbert Formation

This facies is named from exposures in the same road cut (locality I; Fig. 21) that exposes the Spud Hill Member, which it overlies. At this reference site, 10 m of white quartzite/sandstone with partings and thin interbeds of shale (10% of the facies) transition downward over 1.5 m into the underlying Spud Hill Member and from the overlying unnamed member of the Elbert Formation by 2 m of sheared and boudined sandstone, shale, and dolostone.

The Mill Creek facies is made up predominantly of hard white (N9, 5YR 9/1; Figs. 21–22) and off-white sandstones (light gray, N7, and light yellowish brown, 2.5Y 6/4) with shale partings (drapes) and beds up to 15 cm thick. In places, interbeds of sandy dolostone form the base of the facies where it overlies the Spud Hill Member. The abundance of shale at different localities ranges from 2% to 15%, but increases up-section at locality I (Mill Creek road cut). At localities J and K, shale is present only as drapes.

Sandstone sedimentation units are 2 mm to 15 cm thick, and beds, which generally extend the limit of each outcrop, reach 1.5 m thick. Both sedimentation units and beds are demarked by contrasts in grain size and in places by millimeter-scale stylolites (Fig. 22), clay drapes, and centimeter-thick shale beds. The clarity of bedding varies by locality, but laminated and structureless sedimentation units comprise 90% of beds, while current-ripple laminations, burrow-mottled beds, wave-ripple laminations, planar cross-beds, and hummocky cross-beds (HCS; Fig. 21) in decreasing abundance make up the remainder. Wiggin (1987) and Maurer (2012) report HCS to be abundant, but I nowhere found more than two HCS beds per outcrop. My examples of HCS have wavelengths of a meter and amplitudes up to 30 cm. Structureless beds have a random or clustered distribution of coarse and fine sand grains typical of cryptic bioturbation.

Laminations well displayed at localities I, J, and K have two features ascribed to tidal bedding (Reineck, 1967; Davis, 2012). Sand laminations ranging from 1 to 5 mm are separated in places by clay films (Fig. 24) that are locally pigmented by iron oxide. These heterolithic couplets are typical of single tidal cycles (Davis, 2012). Other sand laminations have bimodal thicknesses, with thin laminations from 1 to 2 mm and coarse laminations from 3 to 5 mm. In places thicker laminations (1–3 cm) are displayed by differential quartz cementation, where fine-grained laminations are completely cemented by quartz in contrast to incompletely cemented coarse-grained laminations (Fig. 25). Such couplets are examples of monolithic tidal bedding (Reineck, 1967; Davis, 2012). Stylolites with spacing of 1–50 cm, relief of 1 mm, and lateral extent to 3 m developed on clay drapes in some tidal couplets.

Spherical patches 1–3 cm in diameter of partly quartz-cemented sandstone make up 30% of some beds (Fig. 26). These patches, which weather as depressions, are typical of those formed where former calcite-cemented concretions have been leached of their carbonate cement in outcrop (e.g., Ozkan, 2001; McBride, 2012).

Like beds in the Ignacio Quartzite, bimodal texture dominates and leads to poor and very poor sorting (Table 1); however, about one-third of the beds are well or moderately sorted unimodal samples. Silt-free clay matrix, introduced where burrowers mixed clay drapes into clean sand, is common locally. A distinctive and characteristic feature of both facies of the McCracken, in general, is a population of extremely well-rounded, medium-size spherical (“billiard- ball;” Fig. 23) grains of quartz. Other quartz grains display a range of roundness values.

The trace fossil Planolites is abundant along clay-draped bedding planes (Wiggin, 1987; Maurer, 2012). The larger size-mode is dominant, but vertical burrows (Arenicolites, Wiggin, 1987; Skolithos, this study) are present, but rare (Tables 2–3). Millimeter- to centimeter-thick shale beds are intensely burrow mottled. Like shales in the Spud Hill Member, they are composed of illite with variable amounts of sand and silt composed of quartz and mica. Wiggin (1987) found placoderm plates of Late Devonian fish at five localities in this facies. These distinctive, purple-white punctate phosphate shells occur as disaggregated pieces up to 6 cm long in the sandstone.

Conglomerate beds of white vein-quartz granules and pebbles (up to 10 cm in length) a centimeter or two thick are scattered throughout the facies. Winnowing of these sedimentation units when they were exposed to currents left the coarsest grains at the top of layers, and local pebbles were stranded among sand like glacial dropstones.

South of locality E (Electra Lake), the Mill Creek facies is a meter or so of interbedded sandstone and carbonate and has a patchy distribution. From its type area, where it overlies the Spud Hill Member, it thins northward toward Coal Bank Pass from 10 m to only half its thickness in 2.5 km (Fig. 27). At the Coal Bank Pass turnout, it rests on basement or on the basal conglomerate. It is only 3 m thick, but it increases to 8 m within 80 m north and south of this point. North of the turnout at the pass, the uppermost Mill Creek sandstone has a lag deposit of vein-quartz pebbles and cobbles up to 20 cm long. This bed is overlain by shale (largely concealed), which I assign to the upper member of the Elbert Formation. The facies is absent in the area of Molas Pass. South of locality I (Mill Creek road cut), the facies is sandwiched between sandy carbonate of the unnamed member of the Elbert Formation or Ouray Limestone above and the Spud Hill or Tamarron facies below. At locality H, sandstone of the Mill Creek facies is interbedded with carbonates assigned to the unnamed member of the Elbert through 5 m of section. I place the contact between the two stratigraphic units at the base of the lowest carbonate bed thicker than 1 m. The stratigraphic section at locality F (Cascade Creek; Figs. 8–9) is based on data from Baars (1965, his section B), and my pick for the facies contacts from his data is an estimate.

Sultan Creek Facies of the McCracken Member of the Elbert Formation

This facies—the most heterogeneous of the two in the McCracken Member—includes sandstone, shale, dolostone, and limestone. Relative abundance varies laterally. Much of the unit is poorly exposed owing to the abundance of shale, and many of the exposed beds are visible only in joints coated by manganese and iron oxides. In the vicinity of Silverton, Colorado, rocks have been overprinted to various degrees by hydrothermal fluids from the Silverton volcanic province with the development of various calc-silicate and sulfide minerals.

The reference section of this facies is the east-facing slope 100 m south of Sultan Creek (locality M). This reference section exposes 2 m of pebble-cobble conglomerate of probable Precambrian age resting on the Precambrian Twilight Gneiss. The succeeding 20 m is interbedded sandstone and shale (50%). This, in turn, is overlain by 14 m of cyclic beds of dolostone, sandy dolostone, and shale (Figs. 8, 28, and 29), which is capped by 2.5 m of white quartzarenite. Several meters of covered shale separate the rocks from the upper member of the Elbert Formation. Exposures are inadequate to document the lateral geometry of beds of this facies in detail, but Figure 30 shows the geometry of beds based on four measured sections within a lateral distance of 1 km. This facies comprises all of the sequence above the Precambrian basement and overlying conglomerate to the upper member of the Elbert Formation north of Molas Pass, except for 4 m at locality O (Deadwood Creek Gulch).

At the reference section for the Sultan Creek facies, the beds from the Twilight Gneiss to the top of the dolostone were traditionally assigned to the Ignacio Quartzite (J. Sprinkle, personal communication, 1985), and only the uppermost white quartzite bed 2.5 m thick was assigned to the McCracken Member. In contrast, I assign the entire 36-m section, exclusive of the basal conglomerate, to the McCracken Member. Evidence for the assignment of this facies to the McCracken Member includes similarity of the sandstone in color, hardness, and composition to sandstone of the Mill Creek facies, which other workers agree is McCracken, and the occurrence of carbonate beds, chiefly dolostone, in the McCracken in the Paradox Basin (Cole and Moore, 1996; Appendix B this paper). Diagnostic features of sandstone in the Paradox Basin include white and off-white colors, abundance of “billiard-ball” quartz grains and polycrystalline quartz, and near absence of K-feldspar.

Sandstone of this facies, with a few exceptions, is white (N9, 5Y 8/1; Fig. 31) or off-white (5Y 8/1) quartzarenite (<5% detrital feldspar). Rare beds are light red (7.5R 6/6) or red (5R 5/6; Fig. 32), owing to hematite-coated quartz grains. Sedimentation units range from 1 mm to 20 cm thick; beds range from 50 to 100 cm thick and are laterally continuous within an outcrop. In disagreement with Baars (1965), sandstone beds do not show channeling or abundant cross-beds. The lack of continuity of sandstone beds in closely spaced measured sections (Fig. 30) is best explained, however, by lenticular beds, such as those that fill broad channels. Millimeter-scale stylolites are spaced from 10 to 50 cm apart and extend less than 2 m laterally. They are present only at clay drapes.

Sandstone beds closely resemble those of the Mill Creek facies in their dominance of laminated and structureless beds. Burrow-mottled beds are also common. Planar cross-beds and HCS are rare. Laminated beds have features similar to those of the Mill Creek facies, including rhythmic thick–thin laminations and rhythmic clay drapes. In exposures north of Sultan Creek (locality N), fracture surfaces of sandstone and limestone are heavily stained by manganese dioxide (MnO2) such that bedding structures there are obscure. “Billiard-ball” quartz grains are common in most beds (Fig. 23). Several percent of illite matrix—clay drapes dismembered by bioturbation—are present in the heavily bioturbated beds. Bedding planes exposed at shale beds have several generations of Planolites and formless burrows; Skolithos is rare, but comprises 90% of bedding typical of “pipe-rock” in places (Table 2). Table 3 summarizes the distribution of trace fossils in the units studied and their ecologic significance.

Some sandstone beds have scattered porous spherical patches from 1 to 3 cm in diameter similar to the leached carbonate-cemented concretions of the Mill Creek facies (Fig. 26).

The Sultan Creek facies exposed at locality M (Sultan Creek south) reveals the thickest section of carbonate beds (17 m) in this facies. Carbonate cycles range from 10 to 70 cm thick and are demarked by millimeter- to centimeter-thick shale drapes and stylolites developed on them (Figs. 28–29). Cycles are composed of brown sparry dolostone beds capped by uneven layers of white to purple/pink quartz-cemented quartzarenite beds 5 to 15 cm thick.

The ideal dolostone cycle begins with a clay drape or bed and ends with a sandstone layer a few centimeters thick. Intense bioturbation of clay layers left Planolites and formless burrows weathered out on the top of the sandstone bed that caps the underlying cycle. The following two-thirds of the cycle is strongly burrow-mottled brown dolostone with scattered well-rounded quartz sand grains. This is followed by several centimeters of irregular wispy microbial laminations/stromatolites (Wiggin, 1987), which, in turn, are capped by an uneven, largely burrow-mottled layer of quartzarenite of thickness mentioned above (Fig. 29). The top surface of the sandstone layer is irregular and patterned by Planolites. Where the sandstone layer is absent from the top of the cycle, a dolostone breccia (pedogenic?) several centimeters thick is present locally. The uneven contact of dolostone and quartz sand, textural mottles in the sand in places, and vestiges of sand beds in dolostone indicate that, prior to dolomitization, originally discrete carbonate and quartz sand layers were mixed to variable degrees by bioturbation.

Dolostone beds are composed of spar grains (0.05–0.20 mm) with variable amounts of well-rounded, poorly sorted quartz grains. Many of the dolomite grains have slightly curved crystal faces typical of saddle dolomite, and some crystals are zoned. Patches of pure dolomite within sand-rich intervals are either filled burrows or former carbonate clasts. The purple/pink color of the quartzarenite layers is imparted by iron-oxide coats on detrital quartz grains. Quartz grains are well rounded, are variably sized and sorted, and include “billiard balls.”

Limestone beds are less than 1 m thick and demarked by clay drapes and centimeter-thick shale beds; they comprise 10% of carbonate beds. They are structureless to burrow or color-mottled carbonate mudstones (micrites) with sand-sized ghosts of skeletal grains. Most beds are light gray (N8).

At most localities the Sultan Creek facies rests on the basal conglomerate or basement; however, at locality O (Deadwood Creek Gulch) it rests abruptly on a few meters of the Tamarron Member. It is overlain by the upper member of the Elbert Formation or, where absent, by the Ouray Limestone. The contact ranges from abrupt, where overlain by shale beds of the upper member of the Elbert Formation, to gradational, where interbedded with carbonate beds of the Elbert Formation or Ouray Limestone. Where sandstones of McCracken facies are interbedded with carbonates of the latter two stratigraphic units, I place the contact at the base of the lowermost carbonate bed thicker than 1.5 m. The maximum thickness I measured for the facies is 36.5 m, which is at the reference locality.

Upper Contact

As noted by Baars (1965, p. 28), the upper member of the Elbert Formation “is a poorly resistant, slope-forming sequence of both shales and shaly dolomites…. The dolomites typically are thin bedded to fissile and often display stromatolitic laminations and salt casts…. The shales are usually gray or light green in color, but may be tan to maroon, and almost always are fissile and soft.” Wiggin's (1987) sections document that white sandstones assigned to the McCracken locally are overlain abruptly by shales as described by Baars (1965), but in places they are overlain abruptly by pure dolostones or by a sequence of interbeds of dolostone, limestone, and sandstone in decimeter- to meter-scale beds.


The maximum thickness reported for the McCracken Sandstone in outcrop is 15 m by Wiggin (1987) to 30 m in this study. Wiggin (1987) notes that the McCracken is absent in some places even where the upper member of the Elbert is present.


The Late Devonian age assignment of the McCracken in the Paradox Basin is based on its position between the Aneth Formation below and the upper member of the Elbert Formation above. The Aneth Formation has fish remains dated as Late Devonian (Knight and Cooper, 1955), whereas the basis for the age of the Elbert Formation is described below.

Correlation of the white quartz-rich sandstone beds of the subsurface McCracken Member with hard, white sandstone beds exposed in the San Juan Mountains was endorsed by Baars and Knight (1957), Baars (1958), Parker (1961), and Parker and Roberts (1963). Baars later (1965, p. 26) doubted that the subsurface McCracken reached the outcrop. Nevertheless, subsequent studies, including this one, find the outcrop McCracken Member to be an essential stratigraphic unit and that it is the basal part of the Elbert Formation. Fish fossils identified from the undifferentiated Elbert Formation (Eastman, 1904; Cross, 1905a, 1905b) were identified as Late Devonian. Review of the taxonomic status of the genus (Bothriolepis), which was found in the Elbert Formation, indicates that the taxon is known only from the late Frasnian and Famennian = Late Devonian (Thomson and Thomas, 2001). Fossils in the overlying Ouray Limestone indicate that the Elbert can be no younger than mid-Famennian in age (Armstrong and Mamet, 1976).

Wiggin (1987) subsequently found placoderm fragments in white sandstones that I place in the McCracken Member. This supports the subsurface to surface correlation.


The contact between the Ignacio Quartzite and McCracken Member, as defined herein, with their subjacent and superjacent stratigraphic units was examined at 25 localities. Schematic stratigraphic sections showing the distribution of facies of the Ignacio and McCracken and their bounding units for 20 sites are shown in Figure 8, and lithologic logs for 12 localities are shown in Figures 13A and 13B. The latter sections show the types of rocks exposed at each contact, the position of major covered intervals, and the position where I placed the facies contact for this study.

Contacts of the Ignacio and McCracken with nonconformable basement rocks are straightforward. And in agreement with Wiggin (1987), Condon (1995), Maurer and Evans (2011), and Maurer (2012), the basal quartzite conglomerate has, in most places, such a sharp contact with Ignacio sandstones and textural features to suggest that it is older than the Ignacio. Contact relations of the Ignacio and McCracken stratigraphic units with each other and with adjacent sedimentary rocks are diverse, owing to lateral facies changes within both stratigraphic units and in the overlying upper member of the Elbert Formation.

Contacts are commonly poorly exposed, owing largely to the weathering of shale beds at the site. In addition, the upper member of the Elbert Formation is not everywhere present (Gonzales et al., 2004, p. 15), whether owing to erosion or facies changes is undetermined, and it is unclear whether certain carbonate beds are best assigned to the upper member of the Elbert Formation or to the Ouray Limestone.

Of the 17 stratigraphic sections shown schematically in Figure 8 that are from my data, the Ignacio-McCracken contact is present at only eight and visible to some degree at only six. Of the six, four contacts (localities A, B, C, and H) have carbonate beds that I assign to the McCracken resting directly on Ignacio sandstone beds, and one (locality Q [Canyon Creek]) has a shale bed of the McCracken resting on an Ignacio sandstone bed. None of the contacts show any angular discordance, but at the contact or within a meter or so of the contact there are anomalous features at five sites. Pedogenic features are present at localities A and H; decimeter-scale stratiform breccias are present at localities B, C, and H; and a lag deposit of a probable flooding surface is present at locality D’. The latter lag deposit has fragments of fish fossils characteristic of the McCracken Member. Lag deposits and breccias are also present near the top of the McCracken in some places.

Within a meter or so of the contact between the Ignacio Quartzite and McCracken Member as defined herein, I recognize: (1) pedogenic features at Bakers Bridge locality A (caliche) and Mill Creek Lodge locality H (probable silcrete and caliche); (2) thin stratiform breccias of probable karst origin at localities B and C (Shalona Lake railroad cut and Glacier Club, respectively); and (3) razor sharp contacts between sandstone below and carbonate above at locality D' (Chris Park road). These features, if correctly interpreted, clearly indicate an episode of subaerial exposure at the contact for the first two types of contacts and an episode of erosion at the third type of contact. On the other hand, these features were not recognized at other localities, where the contact is gradational. Such a localized unconformity indicates either syndepositional faulting or a eustatic drop of sea level of only a meter or so, both of which are difficult to document. In the absence of dateable fossils, the time duration of the hiatus cannot be determined and the erosional surface may be only a parasequence boundary representing a few thousand years.


U-Pb ages were determined on detrital zircons for one sandstone sample from each of the members and facies. Results are shown in Figure 33 and Table 4. The Ignacio (Tamarron and Spud Hill) samples are quite similar and have the most uniform age distributions. Both have sub-equal age peaks at 1.75 and 1.42 Ga. The sample from the Tamarron Member has one Ordovician-age grain and no Archean grains. The sample from the Spud Hill Member has one Paleoproterozoic grain and one Archean grain.

The two McCracken samples are also quite similar to each other, but have a wider spread of zircon ages than Ignacio samples. Both have the same dominant 1.75 and 1.42 Ga peaks found in the Ignacio samples, but the peaks have considerable noise, and both contain three Cambrian-age grains (535–520 Ma). The sample from the Mill Creek facies has five Archean grains, and the Sultan Creek sample has one. The Mill Creek facies has about 10% of grains with ages between 1.9 and 1.8 Ga, and the Sultan Creek facies has a similar amount of grains with ages between 2.1 and 1.9 Ga.


Regional Depositional Setting

Decades ago Read et al. (1949) and Barnes (1954) interpreted the Ignacio-McCracken-upper Elbert-Ouray succession to be the depositional product of the basal Paleozoic transgression across the southern part of Laurentia in the Late Devonian, an interpretation supported by this study. Expanding on this theme, Maurer and Evans (2011) and Maurer (2012) proposed that the transgressive deposits of the study area record the infilling (Ignacio-McCracken) and eventual overfilling (upper Elbert-Ouray) of a single valley approximately 20 m deep that was cut into Precambrian bedrock during the Late Devonian. Measured sections of my study indicate, however, that the depositional topography and history of the rocks studied likely were more complex than this scenario. The combined thickness of the Ignacio and McCracken units ranges from 10 to 45 m along a N–S transect 25 km long (localities A–F; Fig. 1), but the lateral variation in the thickness of the two units does not fit the geometry of a single major filled channel. Instead, the thickness pattern suggests that the transgressing paleoshoreline was cut by numerous channels/embayments of different depth. Accordingly, the rocky paleoshoreline envisaged by Campbell and Gonzales (1996) to explain the source of the conglomerate beds of debatable age would have been highly irregular with numerous small estuaries during initial stages of deposition until deposition smoothed the shoreline. The presence of pedogenic and possible karst features at or near the Ignacio-McCracken contact indicates that there was a period of subaerial exposure, but there is no evidence that much sediment was eroded during this event nor an indication of how much time this hiatus represents.

Recent paleogeographic reconstruction for the Late Devonian ( in Laurentia indicates that transgression was southeastward across a northeast-trending peninsula or landmass that has commonly been labeled the Transcontinental Arch (e.g., Carlson, 1999). Myrow et al. (2003), however, raised the question whether the Transcontinental Arch existed this early in the Paleozoic. Whatever name be applied to this positive element, it was an important source of clastics in the Ignacio and McCracken in the study area and probably also the clastics of the Ignacio in the Paradox Basin. The eastern source is clearly shown by the presence of coarsest clastics in the eastern part of the study area (Maurer, 2012) and paleocurrent data of Seeland (1968) and Maurer (2012).

Processes and Environments of Deposition

Most interpretations of environments of deposition of the rocks of interest to this study made before the 1980s were made before much was known about near-shore coastal processes and deposits, especially of estuaries and tidal flats. Nevertheless, there was general agreement that the deposits were deposited close to shore in waters of strong marine influence. Previous interpretations of environments of deposition and provenance are summarized in Table 5.

The coarseness of clastic rocks in the eastern part of the study area, scarcity of shale, and apparent absence of landward-directed cross-beds and trace fossils in the minimally exposed eastern facies of the rocks studied (Maurer, 2012) suggest that these beds are fluvial deposits, probably of braided streams. The lower few meters of the Tamarron Member, which have unidirectional cross-bed dips and no evidence of bioturbation, are interpreted as fluvial/estuarine deposits also by Maurer (2012) and endorsed by this study. All the other rocks except shales have features, highlighted in Table 2, indicating that they were deposited under tidal regimes and, thus, are interpreted as tidal-flat deposits. Key features of tidal activity include landward-directed paleocurrents, heterolithic beds, several types of tidal couplets, association of marine/brackish-water trace fossils, and, although rare, desiccated surfaces (cf., Reineck, 1967; Dalrymple, 2010; Davis, 2012; and articles therein). The broader context, however, indicates that the tidal flats were integral parts of estuaries that were strongly influenced by ebb- and flood-tidal currents (Maurer, 2012). The lagoonal interpretation of the Tamarron Member by Wiggin (1987) is broadly compatible with an estuarine setting.

Baars (1965) and particularly Wiggin (1987) concluded that a topographic high (the controversial Grenadier Highland of Baars, 1965) effectively separated two different depositional domains. The area north of the area between Coal Bank and Molas passes, dominated by the Sultan Creek facies of this study, received less terrigenous sediment than south of the area of the passes, such that carbonate banks and tidal flats flourished between episodes of terrigenous sand and mud supply. The dolomitization overprint on limestone beds destroyed most original carbonate textures, but carbonate mud survives in a few limestone beds, and the mixed sand and sparry dolostone beds suggest that many of these beds were originally grainstones. The stromatolites first reported by Wiggin (1987) record shallow subtidal to supratidal episodes in the Sultan Creek area. The abundance of carbonate and near absence of Ignacio facies in this northern domain indicate that this area was either farther seaward from the paleoshoreline than the area south of the passes, or at a site where there was much less input of terrigenous sediment, which permitted carbonate-producing organisms to flourish. The presence of a topographic barrier to explain the differences between two sedimentary domains is not necessary because sediment input from two different river systems could produce the same sediment pattern. The existence of two river systems also better explains the distribution of “billiard-ball” quartz grains. The geometry of facies of the McCracken indicates that the facies are coeval.

The Tamarron Member has characteristics (Table 2) of estuarine-tidal sand flats whose surfaces were populated at different places or times by upper flow-regime sand flats (e.g., Reineck, 1967) and sand waves/megadunes; desiccation cracks indicate a supratidal setting at times. Although bidirectional cross-beds are present, paleocurrent data indicate that sand-wave bedforms migrated chiefly west and northwest—that is, seaward (Seeland, 1968, his fig. 28; Maurer, 2012; qualitative data, this paper). This shows that ebb tides were stronger than flood tides. Maurer (2012) proposed that the lower few meters of this facies are fluvial deposits that pass upward to estuarine deposits. This interpretation is compatible with the tidal-flat hypothesis proposed herein because tidal flats are integral parts of many estuaries (Dalrymple, 2010). The presence of Planolites in mud drapes and cryptic bioturbation within sandstone beds does not closely constrain the salinity during the deposition of this facies. The absence of tidal-channel deposits is somewhat problematic to the tidal-flat interpretation; however, the near absence of mud beds to stabilize channel walls (e.g., Reineck, 1967) could have led to the development of braided tidal channels of shallow depth. The lateral migration of such channels would be difficult to document in the small lateral exposures available in the study area. The scarcity of mud and uniquely fluvial deposits support the conclusion that the rivers supplying this facies had both little discharge and little sediment load.

The scarcity of slack-water mud beds among the sand beds and absence of reactivation surfaces suggest a setting of exceptionally slow generation of accommodation space during deposition. Under such a regime, most mud drapes would be eroded by subsequent tides or storms as would the upper part of sand waves with reactivation surfaces. Neap deposits would be selectively preserved, and there would be lots of opportunity for cryptic burrowers to destroy laminations. The low diversity of trace fossils in the Tamarron Member (Table 3) may reflect salinities less than seawater owing to river input (Maurer, 2012) or, alternately, scarcity of food in active tidal flats.

Heterolithic beds, bi-directional and landward-migrating cross-beds, and association of marine trace fossils with desiccated surfaces of the Spud Hill Member (Table 2) are characteristic of mixed sand-mud tidal flats (Reineck, 1967), which, at times, were also supratidal. Tidal couplets have not been recognized in this member, but bioturbation has destroyed bedding details in many beds where they may have once been present. The intertidal to supratidal environments proposed here for the Spud Hill Member were quiet enough to permit much more mud to accumulate than found in environments of either the Tamarron Member or Mill Creek facies. The presence of pebble lag deposits and HCS, however, indicate that the tidal flats were subjected to periodic storms. Storms or perhaps tidal currents were responsible for eroding mud beds and generating sand-size mud grains that, after compaction, produced the pseudoshale sedimentation units. Of the several tidal-flat facies preserved in the Ignacio-McCracken rocks, the environment in which the Spud Hill Member was deposited supported the most diverse mud-burrowing organisms as shown by the trace fossil assemblage. Although no organic matter apparently survives in the rocks today, there was obviously enough in the mud and in suspension to support a significant population of detritus-feeding organisms. The intense bioturbation record also suggests a slow rate of sediment accumulation (Buatois and Mángano, 2011).

A 3-m-thick sequence of upward-thickening heterolithic beds at locality Q (Canyon Creek) is interpreted as the deposits of a bayhead delta (Maurer, 2012; Maurer and Evans, 2013). This is at the transition between what I interpret to be fluvial deposits of the Tamarron Member (mud poor) and the overlying estuarine facies (relatively mud rich, abundant Trichophycus) of the Spud Hill Member and an expectable site for such a deposit. On the other hand, these authors interpret the upper part of the Tamarron Member at Bakers Bridge (my locality A) as the deposits of a strand/barrier complex, whereas I interpret them to be fluvial-estuarine deposits. I believe that these same beds of the Tamarron Member of disputed origin are beds that Wiggin (1987) interpreted to be storm-deposited shelf deposits characterized by abundant HCS, a bedding type that I find to be rare.

The tidal couplets and abundance of horizontal laminations that characterize the Mill Creek facies indicate that deposition took place on tidal-flat sands dominated by upper-flow regime transport such as reported from modern tidal flats in India (Chakrabarti, 2005) and in the Western Interior Seaway during the Campanian (Steel et al., 2012). Wave and tidal currents, the latter shown by landward-directed current ripples, were vigorous enough to prevent much mud accumulation. Disturbed laminations typical of cryptic bioturbation seen in thin sections and structureless beds are chief lines of evidence for bioturbation. Storm events are recorded by sharp-based sandstone beds, pebble lags, and rare HCS. Maurer (2012) and Maurer and Evans (2013) interpret the Mill Creek facies as identified here to be deposits of a strand/barrier complex, an interpretation at conflict with arguments presented here.

The Sultan Creek facies, essentially restricted to a position north of Coal Bank Pass, records episodes of deposition of tidal-flat sands and carbonates and the lateral shifting of environments; covered shale beds remain an environmental enigma. Where sandstone beds are well exposed, they have features similar to those of the Mill Creek facies except that there are also some megascopically bioturbated beds. The lack of lateral continuity over distance of hundreds of meters suggests that some beds filled estuarine channels, although the degree of tidal influence is unclear. Skolithos, typical of “high-energy” loose-sand settings (Gingras and MacEachern, 2012), is rare. Many thin mud drapes that may have once identified tidal rhythmites were mixed into adjacent sand beds by bedding-plane burrowers. Centimeter-scale mud drapes were thoroughly mined by Planolites burrowers. Maurer (2012) noted that Planolites and companion burrowers suggested the presence of firmgrounds.

At locality M (Sultan Creek south), the 12 m of cyclic deposits of sandstones and dolomitized limestones document the interplay of carbonate and terrigenous sediments. Although the cycles have to be viewed through the overprint of dolomitization, they have the shallowing-upward characteristics of aggradational parasequence cycles, those produced by relative sea-level changes (Jones, 2010). The highly variable thickness of cycles suggests that there were local autocyclic events—such as lateral shifting of facies and storm erosion—that modified allocyclic climatic effects. The microbial laminites/stromatolites in the carbonate cycles can be either intertidal or supratidal, but are typical of humid rather than arid settings (Pratt, 2010). The well-sorted sand that was introduced into the carbonate environment either washed into the tidal flats during storm events or was wind-blown material (Wiggin, 1987) from nearby coastal dunes. The superb rounding of the quartz certainly indicates an eolian abrasion history prior to deposition, but there is no bedding suggesting that the thin sandstone layers accumulated as dune deposits. The sea-level changes that generated the parasequences at locality M (Sultan Creek south) cannot be recognized elsewhere in the Sultan Creek facies or in facies south of Coal Bank Pass. Maurer (2012) assigns the dolostone beds to fining-upward cycles of an estuary complex.


Composition and roundness of sand and gravel clasts and ages of the detrital zircons indicate that the clastics were derived from a broad terrain that included granitoid plutonic rocks ranging in age from ∼0.46 to >2.4 Ga, micaceous metamorphic rocks (Maurer, 2012), metaquartzite, sand from mature eolian ergs, and older sandstones. Although the immediate source of detritus may have been the geographic site of the Uncompahgre Uplift as proposed by several previous workers (i.e., Baars, 1965; Wiggin, 1987; Maurer, 2012) the ultimate source lands for the sandstone beds was a much larger and much more diverse terrain than the uplift.

The dominant age of zircon grains in all samples (1.8–1.7 and 1.4 Ga) are similar to ages of igneous basement rocks throughout the south-central part of Laurentia composed of the Mazatzal (1.7–1.62 Ga) and Yavapai (1.8–1.7 Ga) Provinces, the 1.48–1.42 Ga mid-continent Granite-Rhyolite Province, and scattered 1.48–1.4 Ga granites like the local Eolus Granite (Fig. 34). These conclusions are based on a synthesis of basement rock ages by Gonzales and van Schmus (2007) and Gehrels et al. (2011). The Amarillo-Wichita Uplift of Texas-Oklahoma has Cambrian granites that match the 535–520 Ma ages of McCracken zircons. As noted by Gehrels et al. (2011), however, there is no stratigraphic evidence that these granites were exposed during Devonian time (Gilbert and Denison, 1993). The oldest grains (>2.5 Ga) were likely derived from the Archean craton, of which the Wyoming portion was the closest part. Seven grains from McCracken samples have values (2.1–1.87 Ga) that are assigned to an unidentified terrain previously recognized by Gehrels et al. (2011).

The Ignacio zircon ages are nearly identical to those in the Temple Butte Formation (Middle–Upper Devonian) in the Grand Canyon, including the signature of the Amarillo-Wichita Uplift in the Tamarron sample (Gehrels et al., 2011). Unlike Cambrian and post-Devonian sandstones in the Grand Canyon, however, Ignacio and McCracken samples lack a Grenvillian signature. This suggests that much Ignacio-McCracken sand came from a more northerly transport route than pre-Mississippian Grand Canyon sands. This suggestion fits the evidence of a contribution of detritus to the study area from the Wyoming craton. The presence of three Ordovician grains in the Temple Butte Formation led Gehrels et al. (2011) to suggest a possible contribution from the Appalachian or Franklinian orogens. The single 460 Ma grain in the Tamarron Member raises the same possibility, although the possibility of recycled grains must be considered. Both McCracken facies contain some zircons of unknown provenance.

The zircon data, then, together with other sand grain types and textures, indicate that the target sandstones had a remarkably diverse provenance and disprove the majority of earlier interpretations that local Precambrian rocks supplied all the detritus. The relative abundance of zircons of a certain age, however, cannot be used to determine the relative volume of different source rocks owing to variable zircon yields of different plutons (Dickinson, 2008). Thus, the relative abundance of the ages of zircons listed in Table 4 cannot be equated to the relative outcrop area of the different terrains. The zircon data also do not imply that all the inferred source terrain were exposed during the Late Devonian because recycling cannot easily be assessed. The absence in my samples of quartz grains with abraded overgrowths indicates that there was no significant recycling of quartz-cemented sandstones, but recycling of sandstones cemented by other minerals (carbonates, sulfates, etc.) cannot be ruled out. Maurer (2012), however, reported finding quartz grains with reworked overgrowths in some of his samples, which provides evidence of some recycled sandstones.

The provenance diversity indicated by the indestructible zircon grains is not documented in the composition and texture of grains recognized in thin section. My data indicate that K-feldspar is the only feldspar species to survive the rigors of weathering and abrasion. The coarse grains, especially the angular ones, could have been supplied largely from local granites as first-cycle detritus. The well-rounded K-feldspar grains are the product of extensive fluvial or eolian abrasion, but the location of their granitic parent and transport history are unknown. The metamorphic rock fragments have little diversity, and their composition is compatible with derivation from local source rocks exposed in the Needle Mountains to the east of the outcrop belt. The same is true of the micas. The micas and micaceous metamorphic rock fragments are certainly first-cycle grains because micaceous grains do not readily survive the rigors of weathering and transportation more than once. The Uncompahgre Formation clearly was a major local source of gravel clasts, but, like most metaquartzites, likely did not contribute much quartz sand owing to the absence of weatherable silicate minerals that permit disintegration of the mother bedrock into sand-size grains.

The bulk of quartz grains do not provide specific information on provenance or transport history, but certain populations provide clues. The combination of bimodal textures and well-rounded, spherical to subspherical quartz sand grains, many with hematite coats, indicate that these grains were derived directly from a mature eolian erg or recycled from such a deposit. Kuenen's (1959, 1960) experimental abrasion studies of sand-size grains indicate that eolian abrasion is likely the only process that can produce well-rounded and spherical quartz grains of sand size in abundance. The “tidal race-track” hypothesis of Klein (1970) as a means of generating well-rounded, sand-size quartz must be considered, given the evidence of tidal-flat deposits presented herein, but evidence corroborating the “race-track” hypothesis as a means of producing large amounts of well-rounded quartz sand is lacking.

Iron-oxide grain coats on sand grains are typical of many Quaternary ergs (Folk, 1976; G. Kocurek, 2016, personal communication), whereas bimodal textures typify interdune sand flats (i.e., regs, Folk, 1968). There is no record of Devonian eolian deposits east of the study area, but the well-rounded, wind-blown quartz grains in the cyclic carbonate beds of the Sultan Creek facies indicate the nearby presence of an active or recently denuded dune complex. None of the rounded quartz grains have abraded quartz overgrowths that would indicate that the grains were reworked from older quartz-cemented sandstones. Although first-cycle quartzarenites are known (e.g., Johnsson et al., 1988), none are composed of well-rounded spherical grains. Thus, the likelihood is strong that the quartz population of well-rounded quartz typical of the Mill Creek facies has been recycled (cf. Suttner et al., 1981). A variety of metamorphic rocks could have supplied polycrystalline quartz grains and those with Boehm lamellae.

Except for the eastern outcrops, the basal Tamarron Member has the greatest population of angular–subangular quartz and feldspar, which may be expected of “granite wash” being stripped off the source area. Nevertheless, Ignacio sandstones have up to an estimated 15% of the well-rounded quartz population that characterizes McCracken sandstones. The geographic distribution of the two stratigraphic units indicates that sediment input from the river(s) south of locality F (Cascade Creek) had only a minor amount of quartz with an eolian signature in contrast with sand north of locality F.

Are The Ignacio Quartzite and McCracken Member Coeval?

Stratigraphic data and the presence of both oboloid and placoderm fossils in the same sample (albeit only one sample!) suggest that the Ignacio Quartzite and McCracken Member are coeval, at least in part. Nowhere did I find the two stratigraphic units interbedded, and rocks that I assign to the McCracken everywhere overlie the Ignacio where both units are present. However, as shown in the cross-section of Figure 8, the basement and basal (Precambrian?) quartzite conglomerate are overlain by members of the Ignacio Quartzite south of the area between Coal Bank Pass and Molas Pass, whereas north of this area the basal conglomerate is overlain by facies of the McCracken Member. Although sandstones of the Ignacio and McCracken differ markedly in composition, their similar stratigraphic position on Precambrian(?) conglomerate and basement suggests that they are at least partially coeval. The likely reason for the marked compositional difference between the Ignacio and McCracken, as noted above, is that they are composed of deposits from two different coeval river systems, where the northern river had a copious supply of “billiard-ball” quartz grains and little feldspar in contrast with the southern river, which had a copious supply of granite-derived K-feldspar and few “billiard-ball” quartz grains.


The Ignacio Quartzite and the generally conformably overlying McCracken Sandstone Member of the Elbert Formation record the first phase of a Late Devonian transgression across Precambrian rocks of the Transcontinental Arch of North America in Colorado. The absence of a stratotype for the McCracken Member in outcrop and sketchy descriptions of the Ignacio Quartzite have led to decades of uncertainty about how to distinguish the two stratigraphic units. The red and brown feldspathic sandstones of the Ignacio, however, are easily distinguished in the field from the generally better indurated white quartzarenites of the McCracken Sandstone Member. The feldspar-rich Ignacio Quartzite dominates the stratigraphy south of the Coal Bank Pass and Molas Pass areas, whereas the feldspar-poor McCracken Sandstone dominates the stratigraphy north of the area. These strong mineralogical contrasts reflect sediment sources of two different river systems that provided sediment to the Late Devonian coastline. The quartz-dominant McCracken contains a large population of superbly rounded quartz grains that were derived from an eolian dune complex, whereas the feldspathic Ignacio is dominated by detritus derived from granitic source rocks.

Two new members were established for the Ignacio Quartzite to better display the geometry of major rock packages. The Tamarron Member (0–24 m), the first to transgress across much of the Transcontinental Arch, was deposited in fluvial and sandy tidal flats. The overlying Spud Hill Member (0–21 m) and much of the succeeding McCracken Sandstone Member were deposited in sand and mud tidal flats in marginal estuarine settings. The McCracken Member, chiefly white and off-white quartz-cemented quartzarenites, is divided here for the first time into the Mill Creek facies (0–12 m) to the south of Coal Bank Pass and the Sultan Creek facies (0–36 m) to the north of the pass. The Sultan Creek facies alone contains dolostone-sandstone parasequence tidal-flat cycles up to 70 cm thick with a composite thickness of 14 m.

Sandstone composition and ages of detrital zircons indicate that sand grains in the units studied were derived from a complex terrain of igneous basement rocks similar in age (1.8–1.7 and 1.4 Ga) to basement rocks throughout the south-central part of Laurentia, the 1.48–1.42 Ga mid-continent Granite-Rhyolite Province, and scattered 1.48–1.4 Ga granites like the local Eolus Granite.

Stratigraphic relations of the Ignacio Quartzite and McCracken Member, occurrence of oboloid and Late Devonian placoderm fossils in the same bed in the Ignacio Quartzite, and presence of a detrital zircon of Ordovician age in the Ignacio indicate that the two stratigraphic units are, at least in part, coeval and of Late Devonian age.

Field support was provided by the late Peter Carver, Gary Gianniny, Suzanne McBride, L. J. Salyer, Douglas Smith, James Sprinkle, and Mark Helper; lab support was provided by Jaime Barnes, James Barrick, Dan Breecker, Rob Reed, Toti Larson, Donggao Zhao, and Daniel and Lisa Stockli. Some samples and thin sections were provided by the late Peter Carver via Gary Gianniny, Rex Cole, David Gonzales, and Joshua Maurer via James Evans; drafting is the work of Jeffrey Horowitz; counsel on various topics was provided by R. A. Davis, James Evans, R. L. Folk, Gary Gianniny, David Gonzales, Richard Kyle, Fred McDowell, M. Dane Picard, Douglas Smith, and James Sprinkle. Forrest Dean Brown provided access to exposures at Electra Lake. Rex Cole, James Evans, M. Dane Picard, James Sprinkle, and Roger Wiggin suggested improvements to the manuscript. I flag James Evans for his detailed suggestions for improving the manuscript, and I thank Jason Lillegraven (Rocky Mountain Geology co-editor) for his encouragement in revising the original document. In addition to Jay Lillegraven, I thank RMG co-editor Art Snoke, RMG managing editor Brendon Orr, and RMG copy-editor Robert Waggener for their editorial assistance. Financial support was provided by the Geology Foundation of the Jackson School of Geosciences at The University of Texas at Austin.

Appendix A: Locality Descriptions.

Satellite images of localities can be found at Plotted positions on Global Positioning System devices (such as Garmin), Google Maps, and other maps on the Internet may differ slightly from those shown on the website identified above, owing to different formulas used for the shape of the Earth.

A) Bakers Bridge: lat 37.458549 N, long -107.800852 W; elevation = 2,066.5 m (6,779.9 ft). East-facing cliff where La Plata County Road 250 crosses the Animas River; 27.1 m of section. Stratotype of Tamarron Member (23.1 m) rests on granite and is capped by 4 m of sandy dolostone of the poorly exposed Mill Creek facies of the McCracken Sandstone Member of the Elbert Formation.

B) Shalona Lake railroad cut: lat 37.485926 N, long -107.804596 W; elevation = 2,272.3 m (7,458.3 ft). East-facing exposure. About 24 m of Tamarron Member of the Ignacio Quartzite rests on granite and conglomerate. The upper 11 m is not exposed in the railroad cut, but it is exposed 100 m south in a high cliff, where it is overlain by 2 m of Mill Creek facies, the latter of which has an abrupt upper contact with dolomite of the upper member of the Elbert Formation.

C) Glacier Club (Glacier Club Trail): lat 37.508281 N, long -107.804596 W; elevation = 2,311.6 m (7,584.0 ft). East-facing road cut exposes uppermost 6 m of Tamarron Member capped abruptly by 2 m of typical Spud Hill Member of the Ignacio Quartzite, which is overlain by 3 m of interbedded sandy limestone and calcitic sandstone assigned to the Mill Creek facies of the McCracken Member. The latter, in turn, is capped abruptly by mottled limestone assigned to the upper member of the Elbert Formation. Contact with the Precambrian is not exposed.

D) Haviland Lake (Haviland Lake): lat 37.540457 N, long -107.817558 W; elevation = 2,505.5 m (8,220.0 ft). About 2 m of Tamarron Member are exposed in an east-facing poor exposure approximately 400 m north/northwest of the north end of Haviland Lake. Covered above and below exposure.

D’) Haviland Lake (Chris Park Road [Forest Route 791]): lat 37.522272 N, long -107.809141 W; elevation = 2,446.7 m (8,027.2 ft). About 8.5 m of Tamarron Member are exposed along the Chris Park Road in a southwest-facing road cut; covered below the exposure and capped by soil. The Haviland Lake south exposure is 2 km south/southeast of the north exposure.

E) Electra Lake: lat 37.564616 N, long -107.813722 W; elevation = 2,648.3 m (8,688.6 ft). Low east-facing cliff that forms the west bank of Electra Lake just west of the large island in the middle of the lake. About 22.5 m of Tamarron Member; talus from the cliff conceals the contact with the Precambrian; capped by soil. This is almost certainly the type section of Ignacio identified by Cross et al. (1905a).

F) Cascade Creek: lat 37.655343 N, long -107.807508 W; elevation = 2,679.0 m (8,789.4 ft). The section is approximately 0.5 km southeast of where Cascade Creek crosses under U.S. Highway 550. The section between the basement and top of the Mill Creek facies of the McCracken Member is 90% covered. About 12.5 m of Tamarron Member is intermittently exposed above the Precambrian to its contact with the Spud Hill Member, here approximately 20 m thick. The top of the section sampled, directly above the Spud Hill Member, is 10 m of Mill Creek facies, which here is 20% white quartzite, 77% carbonates, and 3% green shale. Baars (1965) recorded a measured section at this site that had 90% exposure, but the section has not been mentioned since his work.

G) Lime Creek Road (Forest Route 591): lat 37.654644 N, long -107.804603 W; elevation = 2,707.2 m (8,881.9 ft). South-facing road cut along Lime Creek Road. About 15 m of Spud Hill Member and 7 m of Mill Creek facies are exposed; contact with the Precambrian is concealed.

H) Mill Creek Lodge: lat 37.676895 N, long -107.792358 W; elevation = 2,933.5 m (9,623.3 ft). This is the road cut on U.S. 550, approximately 1 road km north/northeast of the Mill Creek Lodge turnoff. About 17 m of Spud Hill Member are overlain by 6 m of Mill Creek facies with anomalous carbonate beds. Rocks are exposed in a west-facing road cut; base concealed; uppermost 2-m-thick sandstone bed is overlain by interbeds of off-white sandstone and light gray dolostone assigned here to the upper member of the Elbert Formation.

I) Mill Creek road cut: lat 37.680041 N, long -107.785256 W; elevation = 2,955.5 m (9,696.5 ft). South-facing road cut on U.S. 550 about 3.7 road kms south of Coal Bank Pass. This is milepost 54 locality of Maurer (2012). There are 2 m of cover separating the base of the exposure from phyllite of the Precambrian Twilight Gneiss. The section exposes 11.2 m of dominantly red beds and 10 m of green beds, both of which are the type Spud Hill Member, and then 1 m of poorly exposed shale. This is followed by 10 m of type Mill Creek facies. About 2 m of sheared shale/sandstone beds separate the uppermost sandstones from micritic dolostones of the unnamed member of the Elbert Formation.

J) Coal Bank Pass south: lat 37.697985 N, long -107.777534 W; elevation = 3,245.2 m (10,647.0 ft). Exposure is 30 m southwest of the parking lot at the pass extending southward for 100 m. Phyllite basement is overlain by 0.5–6 m of probable Precambrian conglomerate followed by 12 m of white quartzite of the Mill Creek facies. Capped by soil.

K) Coal Bank Pass north: lat 37.700206 N, long -107.777262 W; elevation = 3,243.0 m (10,649.8 ft). East-facing exposure in the gully approximately 150 m north of the outhouse at roadside turnout. About 5 m of white quartzite of the Mill Creek facies rests on Precambrian argillite; capped by soil developed on the unnamed member of the Elbert Formation.

L) Molas Lake: lat 37.741225 N, long -107.680950 W; elevation = 3,190.3 m (10,466.9 ft). Isolated exposure of 3 m of section in timbered area approximately 250 m southeast of the south end of Molas Lake: paleokarst feature of Wiggin (1987) at/near the upper contact of Ignacio. This is Wiggin's Section V (1987, p. 218–219).

L’) Molas Lake: Refer to above locality for coordinates and elevation. Exposure 20 m downhill from locality L = 14 m of light-colored carbonate beds of the Sultan Creek facies of the McCracken Member and some covered beds underlain by conglomerate. Section V of Wiggin (1987, p. 218–219).

M) Sultan Creek south: lat 37.760137 N, long -107.674636 W; elevation = 3,158.0 m (10,360.9 ft). Approximately 450 m south of Sultan Creek; descends east of U.S. 550. About 36.5 m of Sultan Creek facies measured above Precambrian schist and a lens of conglomerate; capped by 2.5 m of cover below the upper member of the Elbert. These beds have been assigned variously to the Ignacio Quartzite and McCracken Member of the Elbert Formation. I assign them all to the McCracken Member on the basis of sandstone type and presence of interbedded dolostones. Dolostones with minor sandstone interbeds reach 17 m thick.

N) Sultan Creek north: lat 37.773994 N, long -107.67065 W; elevation = 3,176.0 m (10,420.0 ft). Approximately 1.5 km north/northeast of Sultan Creek locality M. Section 18 m thick begins a few meters downslope of U.S. 550 then goes west across the highway to the east-facing road cut. Sultan Creek facies rests on 8 m of Precambrian(?) cobble conglomerate and is capped by fine-grained dolostone of the Ouray Limestone.

O) Deadwood Creek Gulch: lat 37.784984 N, long -107.672614 W; elevation = 3,023.9 m (9,920.9 ft). East-flowing creek below level of U.S. 550. The exposures along this drainage described by Baars (1965) have been mostly covered by rock debris dumped when the highway was widened in 1973. About 1.2 m of Precambrian(?) quartzite conglomerate rests on Precambrian basement, which is followed by 3 m of Tamarron Member, 6 m of cover, then 7.5 m of off-white quartzite of the Sultan Creek facies of the McCracken Member; capped by road fill.

P) Deadwood Gulch road cut: lat 37.786205 N, long -107.672206 W; elevation = 3,019.0 m (9,905.0 ft). East-facing road cut along U.S. 550 approximately 80 m north of Deadwood Creek. Measured section of Sultan Creek facies = 17 m of white quartzite, calcite-cemented sandstone, and dolostone. Contact with Precambrian not exposed; section overlain abruptly by carbonates of the upper member of the Elbert Formation or Ouray Limestone; 80% of joint surfaces are covered by manganese dioxide (MnO2).

Q) Canyon Creek: lat 37.518158 N, long -107.730861 W; elevation = 2,957.3 m (9,702.4 ft). South-facing cut along La Plata County Road 253/Forest Route 682 (Missionary Ridge Road) 1.2 km west of intersection with Canyon Creek. About 20 m of section. Lower 5 m are Tamarron Member, followed by 13 m of Spud Hill Member—both members here are red beds. Top 2 m are off-white Mill Creek facies of thin interbeds of sandstone and carbonates capped by structureless thick dolostones assigned to the upper member of the Elbert Formation.

R) Lemon Reservoir: lat 37.425008 N, long -107.671804 W; elevation = 2,619.3 m (8,593.5 ft). About 5 m of Tamarron Member are exposed approximately 1 km north/northwest of the reservoir, adjacent to La Plata County Road 243. Isolated exposure with cover above and below.

S) Vallecito Reservoir south: lat 37.431539 N, long -107.540097 W (please note: coordinates for this locality are approximate); elevation = 2,380.2 m (7,809.0 ft). Samples provided courtesy of David Gonzales. I identified the samples as McCracken sandstone on the basis of sandstone composition (unknown thickness).

T) Vallecito Reservoir north: lat 37.502609 N, long -107.554691 W; elevation = 2,488.7 m (78165.0 ft). About 6 m of coarse–very coarse sandstone with scattered white quartzite clasts approximately 10 km north of the north end of Vallecito Reservoir. Ignacio Formation of Maurer (2012), but my Mill Creek facies of the McCracken Member. Though Maurer formally referred to the Ignacio as Ignacio Formation, I am using Ignacio Quartzite (National Geologic Map Database, 2016). The McCracken is capped by a few meters of Ouray Limestone. Samples were collected by Joshua Maurer, but provided courtesy of James Evans.

Appendix B. Description of McCracken thin sections: B-614 Well, Lisbon Field, Paradox Basin, southwest Colorado.

Samples lent by Professor Rex Cole of Colorado Mesa University, Grand Junction, Colorado, and formerly with Unocal in Brea, California.

Nineteen samples of sandstone and dolostone were examined from 2,716.2–2,744.5 m (8,911.4–9,004.3 ft) depth. Twelve samples are composed of from 10 to 70% dolosparite with the remainder of the sample chiefly quartz sand; the other seven samples are quartzarenite sandstones that have only traces of dolomite or none. Ninety-eight percent of the non-dolomite grains are detrital quartz, most of which have overgrowths. Most thin sections display disturbed fabric from bioturbation. Grain size of quartz ranges from fine to coarse; sorting is moderate to well.

Ghosts of carbonate allochems, including skeletal grains, are detectable among the dolosparite grains. Traces of K-feldspar survive in a couple samples with dolomite cement, but feldspar was likely replaced by dolomite in most sandstone samples. I have not seen K-feldspar in dolomite-free samples, but these generally have some dissolution pores that could have been K-feldspar. All K-feldspar has overgrowths similar to sandstone beds in the study area. Many of the medium and coarse sand grains are very-well rounded, and some are spherical. These grains are similar to the well-rounded population in the McCracken Member of the Elbert Formation north of Durango, Colorado. No iron-oxide grain coats survive if they were originally present. One sample has some mashed clay clasts. Half the pores formed by the dissolution of carbonate. No red pigment was noted.