- UW Department of Geology and Geophysics
We report and describe abundant, well-preserved, parallel-sided, U-shaped spreite burrows (Rhizocorallium) in the upper part of the latest Cretaceous into early Paleocene Ferris Formation of south-central Wyoming's Hanna Basin. Rhizocorallium typically is a component of the widely represented ‘Cruziana ichnofacies,’ principally involving benthic marine environments seaward of the intertidal zone in shallow to offshore settings. Traditionally, this Ferris section has been interpreted as coal-bearing, continental deposits formed after full withdrawal of the Western Interior Seaway from eastern Wyoming and adjacent areas. The burrowed strata overlie fossiliferous rocks diagnostic of parts of the Puercan Land Mammal Age, early Paleocene. At time of burrow formation, and for several million years thereafter, the vicinity of the future Hanna Basin remained as an undivided, eastern component of an enormous, greater Green River Basin that encompassed almost all of Wyoming's southern half. The Rhizocorallium-bearing marine strata represent westward expansion of a previously more restricted Western Interior Seaway that persisted through latest Cretaceous time in what is now the western Great Plains. Even though tidal influences may have affected rapidly aggrading fluvial systems far upstream to the west in Montana and Wyoming, we regard actual Paleocene marine inundations to have been uncommon and geologically ephemeral events as far west as the Hanna Basin.
Diverse fossil assemblages from strata of the Williston Basin, representing the first five million years of Paleocene time, have led to documentation of fully open, marine conditions as the Cannonball Formation was deposited. Stratigraphic distribution of fossils within the Cannonball shows persistence of the Western Interior Seaway in the northern Western Interior through the Cretaceous, followed by expansion (renamed the ‘Cannonball Sea’) during early Paleocene time. Connections of that seaway to the south, however, remain poorly understood because of later Cenozoic massive erosion of any Paleocene rock record that had existed south of the borderland between the Dakotas. No Paleocene localities in southern Wyoming or Colorado have yet yielded assemblages of marine invertebrate body fossils or microfossils as known from the Williston Basin. No verifiable means, therefore, have been recognized to characterize the Paleocene marine record of the Hanna Basin in terms of species uniquely shared with the Cannonball biota. Short-lived, Paleocene seaway excursions into the Hanna Basin may have been: (1) direct and exclusively from the Gulf Coast; (2) solely from the Cannonball Sea, with seaway contiguity east of the emerging Black Hills; or (3) initiated from a more extensive, midcontinental seaway connecting the Arctic Ocean to the Gulf of Mexico. Substantial structural uplift of the Laramie Mountains prior to mid-Paleocene time would have precluded even brief westward pulses of marine inundation into the vicinity of the presumptive Hanna Basin.
- Cannonball Sea
- Ferris Formation
- Great Plains
- Hanna Basin
- Hanna Formation
- Laramide orogeny
- Rocky Mountains
- trace fossils
- Western Interior Seaway
Diverse forms of Rhizocorallium, U-shaped spreite burrows (Häntzschel, 1975), are widely distributed in Phanerozoic marine strata. Parallel-sided forms are a typical component of the Cruziana ichnofacies (of Frey and Pemberton, 1985, fig. 28, table 7, and p. 99), which principally involves benthic environments seaward of the intertidal zone in shallow (below daily wave base) to quieter offshore settings as interpreted from the rock record. The Cruziana ichnofacies is characteristic of the ‘Marine System, Nearshore or Neritic Subsystems’ as defined by the Federal Geographic Data Committee (FGDC, 2010, p. 19 and 21–22, fig. 5).
We report the occurrence of well-preserved, parallel-sided burrow traces, identifiable as Rhizocorallium, in the upper part of the latest Cretaceous through early Paleocene Ferris Formation of Wyoming's central Hanna Basin (Fig. 1). Its presence in Paleocene strata of south-central Wyoming, in deposits otherwise considered to be continental, implies one of two explanations: (1) Rhizocorallium tracemakers included freshwater as well as marine species; or (2) the host stratum in the Ferris Formation was deposited when ephemeral incursion of the lingering Western Interior Seaway (commonly referred to as the ‘Cannonball Sea’ in application to Paleocene time) influenced aquatic habitats as far west as the central Hanna Basin.
Wroblewski (2002, 2003, 2004a, 2004b, 2005, 2008), Wroblewski and Secord (1998), and Wroblewski and Steel (1998) championed the second option in abstracts and articles emphasizing presence of sedimentary features produced by tidal processes in river-transported, Paleocene valley fills, both within the Ferris and the overlying Hanna Formations. Wroblewski (2003, figs. 20 and 21) envisioned repeated development of long, narrow estuaries extending westward into fluvial systems of the Hanna Basin from their origin in persistent seaways in what today is western Nebraska. On Wroblewski's paleogeographic reconstructions, symbols for ‘marine ichnofossils’ were placed both east and west of the town of Hanna. His text, however, provides little information concerning either the localities or the trace fossils. Wroblewski (2004b, p. 249) referred to Rhizocorallium as a member of a low-diversity marine ichnofossil assemblage, which forms part of his evidence for incursion of tidally influenced brackish water into river channels during deposition of the Ferris Formation. Wroblewski (2004b, fig. 2) shows symbols for occurrences of ‘marine ichnofossils’ at three levels in the Paleocene part of the Ferris Formation, but the traces are neither illustrated nor documented with provision of verifiable stratigraphic or geographic information.
The first report of Rhizocorallium at a specific locality in the upper Ferris Formation is in a master's thesis by Dunn (2003, p. 18). She wrote that the burrows, then tentatively identified by Boyd as Rhizocorallium, are located 22 m above the Ferris Formation's coal bed 24 (of Dobbin et al., 1929). Her photograph of the locality (Dunn, 2003, fig. 2.5) includes the coal, the burrowed bed, and an intervening leaf-bearing layer (formally cataloged as Denver Museum of Nature and Science locality 1–2629). Dunn referred to the leaf-bearing site as ‘Este Lado.’ To distinguish between the leaf-bearing bed and the stratigraphically higher Rhizocorallium bed, we refer to the former as the ‘Este Lado leaf locality’ and to the latter as the ‘Este Lado Rhizocorallium locality.’ Boyd recognized the potential significance of the trace-bearing layer during a brief visit to the locality as a member of a field-trip group led by Lillegraven and Dunn. Study of traces in eight small blocks collected at that time subsequently yielded the morphological information presented below.
The main purposes of our paper are to: (1) document a specific occurrence of Rhizocorallium in the Ferris Formation by describing and illustrating the burrows; (2) provide geographic and stratigraphic information about the fossiliferous bed; and (3) attempt to place this record into a paleogeographic context with evolution of Paleocene midcontinental seaways. In addition, we review taxonomic criteria for identifying Rhizocorallium, summarize the range of environmental settings attributed to the ichnogenus by previous workers, and provide our own conclusions about the depositional environment in which burrows of parallel-sided forms of Rhizocorallium in the Ferris Formation were constructed.
To help clarify the record for relevant fossil-bearing sites, we introduce two corrections (Fig. 2) to leaf-locality data and stratigraphic information reported by Dunn (2003). Following joint examination of the geologic map provided by Dobbin et al. (1929, pl. 27) and coal-bed sections documented in their plate 13, one recognizes that the coal in Dunn's figure 2.5 is ‘locality 341’ of 1929. That locality involves coal-bed 31 of Dobbin et al. rather than bed 24. Coal-bed 31 is markedly higher in the stratigraphic section than bed 24, which is close to the base of the Ferris coalfield of the Seminoe Mining District. Also, the nearby and stratigraphically higher ‘Este Lado leaf locality’ along with the still higher ‘Este Lado Rhizocorallium locality’ are located in T. 23 N., not in T. 24 N. as stated by Dunn (2003, p. 18).
PREPARATION AND DISPOSITION OF SPECIMENS
To prevent crumbling of the eight small blocks collected by Boyd from the Este Lado Rhizocorallium locality, the specimens were gradually hardened by repeated painting with a dilute solution of alvar in acetone. Several burrows were sliced with a rock saw to provide cross sections for study after the surfaces were polished. An XRD analysis of one sample was obtained for determination of composition of clay-sized materials.
The eight outcrop specimens are housed in the Collection of Fossil Invertebrates, Department of Geology and Geophysics, The University of Wyoming (specimen numbers A4473 through A4480). Separated pieces of a sawed block bear different lower case letters following the four-digit number they have in common (e.g., A4473-a).
Locality information and other pertinent data are recorded in the electronic database for the Collection of Fossil Invertebrates. Two compact disks containing an archive of specimen photographs are housed with the numbered specimens.
GEOGRAPHIC AND GEOLOGIC SETTING OF RELEVANT STRATA
Veatch (1907), with field assistance of M. W. Ball and M. A. Pishel, published the first substantive map of surface geology covering the Hanna/Carbon Basins and proximate edges of surrounding uplifts. It is a remarkably accurate map in terms of plotting rock orientations and contacts between major stratigraphic units dominating the basin; Veatch himself modestly referred to his product as a ‘sketch map.’ His efforts set the essential framework to be used in the smaller-scale geologic map (emphasizing interpretations of coal-bed stratigraphy) published in 1929 by Dobbin et al.
The 1929 geologic map applied new formational terminology superimposed by Bowen (1918) upon Veatch's designations for rock units. Bowen subdivided and renamed the principal coal-bearing strata of the Hanna Basin (referred to as the ‘Upper Laramie’ by Veatch, 1907, p. 246) as the Ferris (p. 230) and overlying Hanna Formations (p. 231). The former rock unit is of principal concern for the present paper. It is important to recognize that, without presentation of evidence, Bowen considered the ‘unconformity’ between the Ferris and Hanna Formations to have involved more than three vertical miles of stratigraphic section lost to pre-Hanna erosion. New mapping in the southern Hanna Basin by Lillegraven (in preparation) promises to demonstrate that conclusion to be indefensible.
Although Bowen (1918) defined type sections for neither the Ferris nor Hanna Formations, his description of the Ferris was couched upon where it was best exposed. Inexplicably, he did not utilize the fundamentally sound geographic coordinates provided in Veatch's ‘sketch map’ for purposes of reference. Nevertheless, adequate clues exist within Bowen's discussion to allow confident plotting of his intended defining section for the Ferris Formation. The broad, dashed yellow line connecting the two red stars in Figure 2A indicates that section (between sec. 33, T. 23 N., R. 84 W. and sec. 28, T. 23 N., R. 83 W.).
The modern Hanna Basin (Fig. 1) is tiny in comparison to many other Laramide basins of the Rocky Mountain foreland, but its tectonic setting is amazingly complex (Lillegraven and Snoke, 1996, fig. 4). Indeed, the Ferris and overlying Hanna Formations contribute to “… Upper Cretaceous to lower Eocene marine and nonmarine strata that probably is the thickest syn-Laramide section in the Rockies” (Lillegraven and Snoke, 1996, p. 1). Prior to completion of deposition of the Ferris and Hanna Formations, the general area of the Hanna Basin remained as an eastern component of the greater Green River Basin (Lillegraven et al., 2004, fig. 19). In latest Paleocene time, the still-intact greater Green River Basin dwarfed even the Powder River Basin, encompassing almost all the southern half of what today is known as Wyoming.
Eberle and Lillegraven (1998a, figs. 2 and 3; 1998b) provided the first measured and described section from the general vicinity of the ‘type’ Ferris Formation (see Fig. 2A) and temporally calibrated it using vertebrate fossils and a partial sampling of palynomorphs. Subsequently, Lillegraven and Eberle (1999, figs. 3, 5–8) updated and expanded the formation's vertebrate fossil record of the latest Cretaceous (Lancian Land Mammal Age) through earliest Paleocene (Puercan Land Mammal Age). Grimaldi et al. (2000, figs. 1 and 3) integrated that work into a composite stratigraphic column for the Hanna Basin and temporally linked the Upper Cretaceous–lowermost Eocene record of the basin to the global magnetic polarity time scale, radioisotopic time, and North American Land Mammal Ages.
As indicated in both parts of Figure 2, however, the paleontologically controlled, stratigraphic top of section A–A′ ends at the Seminoe Reservoir's western shoreline. Thus the strata beginning at Pats Bottom (Fig. 2B) and continuing up-section to Bowen's original contact between the Ferris and Hanna Formations (Fig. 2A) remain undocumented in terms of a measured and described section beyond the coal stratigraphy presented by Dobbin et al. (1929, pls. 13 and 27), Glass and Roberts (1980, fig. 15), and proprietary coal-company documents. Despite the best efforts by Dunn (2003) and prospecting teams from The University of Wyoming, no paleontological documentation useful in age control for the upper third of the ‘type’ Ferris Formation exists. All that can be said is that it is younger than strata directly west of Seminoe Reservoir (near the top of section A–A′), dated on the basis of mammalian faunas as very late Puercan (ca. 64 Ma, early Paleocene).
Strata within the lower half of the Hanna Formation as documented in the northeastern corner of the Hanna Basin (‘The Breaks’ in Fig. 1B) have been dated by Higgins (2003) on the basis of mammalian assemblages as Torrejonian (ca. 62 Ma), grading upward into Tiffanian (ca. 60 Ma). In all likelihood, therefore, the uppermost third of the ‘type’ Ferris Formation (which includes Pats Bottom) would be latest Puercan or early Torrejonian in age, spanning roughly 64–62 Ma in Paleocene history. That part of the section, however, has been greatly disturbed by industrial coal mining, and the coal-bearing upper Ferris Formation exhibits widespread faulting (Dobbin et al., 1929, pl. 27; Glass and Roberts, 1980, p. 17). The fact that the peninsula in the Seminoe Reservoir called Pats Bottom remains comparatively intact is most fortunate for present purposes.
Geologically important in the present context is the observation that a major change occurred in depositional history of the Ferris Formation through initiation of regular erosional cutting and sedimentary filling at about 540 feet (165 m) below the top of section A–A′. Eberle and Lillegraven (1998, p. 17) stated:
“Stratigraphically below this point, the pattern of deposition was much more regularly aggradational, with only minor evidence of intraformational scouring. The new cut-and-fill regime continues upward beyond the top of our section A–A′, at least into the levels of minable coal seen east of Seminoe Reservoir. The resulting complex stratigraphy, which intimately involves the coal-beds themselves, is magnificently exposed along the eastern shoreline of the reservoir (i.e., west flank of Pats Bottom), east of section A–A′.”
The recognized depositional change is important because it suggests alterations of base level, hints to a contributing effect upon the accumulation of coal, and is compatible with the possibility of marine influences upon deposition within a long-established, fluvial system.
TAXONOMIC ESSENTIALS OF RHIZOCORALLIUM
The original description of Rhizocorallium, with R. jenense as the type species, was based on specimens from Triassic strata, the ‘Rhizocorallium-dolomit,’ at the base of the ‘Region der bunten Mergels’ in the vicinity of Jena, central Germany (Zenker, 1836, p. 219). Zenker originally identified R. jenense as remains of some sort of sessile invertebrate (using now-archaic terms ‘Zoophyten’ and ‘Pflanzenthiere’), either a sponge or, much more probably in his analysis, a fan-coral.
Since Zenker's work, this distinctive spreite burrow has become one of the best-known ichnogenera in Phanerozoic marine strata. In his review of the state of classification of Rhizocorallium, Fürsich (1974) concluded that only three of its former 15 named ichnospecies warranted continued recognition: R. jenense (short and oblique), R. irregulare (long and sinuous), and R. uliarense (trochospiral). Of the three, our traces from the Ferris Formation are assignable to R. jenense. Nevertheless, although Fürsich's taxonomy has been widely accepted, some authors disagree with his broad interpretation of R. jenense and with his conclusion (Fürsich, 1974, 1975) that suspension-feeding animals constructed all such burrows.
In a recent report on German Middle Triassic trace fossils, Knaust (2007) limited Rhizocorallium jenense to steeply inclined, firmground burrows with poorly developed spreiten. He applied the name R. commune, one of twelve ichnospecies Fürsich synonymized with R. jenense, to mainly horizontally oriented feeding burrows exhibiting pronounced spreiten. Knaust's (2007) figure 8B illustrates an assemblage of R. commune very similar in specimen size, orientation, and crowding to that exposed at the Este Lado Rhizocorallium locality. In the context of Knaust's (2007) revision, the Ferris burrows should be classified as R. commune rather than R. jenense. In order to avoid possible ichnotaxonomic confusion, in subsequent parts of this report we refer to the ichnogenus without species designation, a common practice in the primary literature we have reviewed. In fact, published reports of Rhizocorallium rarely include descriptions of the burrow's morphology. More commonly, the ichnogenus is simply noted as present in a sratigraphic unit under discussion and the reader must assume that even the generic identification is valid.
DESCRIPTION OF SPECIMENS FROM ESTE LADO RHIZOCORALLIUM LOCALITY
At the Este Lado Rhizocorallium locality the burrowed bed is yellowish-gray, very-fine-grained sandstone. X-ray diffraction analysis indicates abundant quartz, low abundance of feldspar, and a clay fraction consisting of kaolinite and illite (Norbert Swoboda-Colberg, personal communication, 2010). The eroded surface of the outcrop exhibits an abundance of diversely oriented and intersecting Rhizocorallium burrows (Fig. 3). The angle of inclination of the burrows typically is less than 25 degrees from originally horizontal strata. Today the locality's strata dip less than 10 degrees to the east-northeast.
Individuals from the Este Lado Rhizocorallium locality are relatively small as judged by comparing burrow width with that reported for the ichnogenus by several authors. Width is a better comparative measure than length because, at least in the Ferris exposure, burrows have been truncated by erosion. Of 18 measurable specimens in a surface area of 470 square centimeters of the Ferris burrowed layer, the greatest width is 26 mm (including the 6 mm-wide marginal tunnel). By contrast, Rodríguez-Tovar and Pérez-Valera (2008) reported an average width of 49 mm for 54 specimens of Rhizocorallium from Middle Triassic strata of Spain; Basan and Scott (1979) noted a range in width of 58 to 113 mm for the Lower Cretaceous ichnogenus in the Purgatoire Formation of Colorado. The narrowest burrow in our sample of 18 measures 11 mm in width with a marginal tunnel only 2 mm in width; specimens narrower than 20 mm are uncommon. Pruss and Bottjer (2004) described a Rhizocorallium assemblage, similar in size (widths of 10 to 24 mm) to that of the Ferris specimens, from the Lower Triassic Virgin Limestone of southern Nevada. They considered the small size of the traces to indicate a stressed environment. Bromley (1996, p. 281) linked the small size of traces to environments that deviated from normal-marine conditions, such as brackish water.
Most of the burrows in our collection from the Este Lado Rhizocorallium locality exhibit the generally accepted criteria (e.g., Fürsich, 1974; Häntzschel, 1975) for the ichnogenus Rhizocorallium (i.e., U-shaped marginal tunnel, spreite, and parallel-to-oblique orientation of the long axis of the burrow relative to bedding; Fig. 4). U-shaped tunnels without a spreite are rare. Although some workers have reported Rhizocorallium with fecal pellets and/or scratch traces (e.g., Rodríguez-Tovar and Pérez-Valera, 2008), we have not recognized either feature in our specimens.
Both the marginal tunnels and spreiten of the Ferris burrows consist of material indistinguishable from that of the host rock. Furthermore, polished surfaces of transverse and longitudinal sections through burrows with sharply defined epirelief fail to show expected meniscate structure beneath the surface. Perhaps the burrowers were active in a substrate too unconsolidated to record details of the burrowing process.
A typical spreite seen in positive epirelief (Fig. 5) is characterized by a series of broad, crescent-shaped mounds separated by sharply defined grooves 4 to 7 mm apart at the spreite midline. The smooth surfaces of the mounds show no evidence of internal structure. Spreiten consisting of closely spaced, arcuate laminae connecting the arms of the marginal tunnel are uncommon. Considering these observations, we suggest that the typical spreite at the Este Lado Rhizocorallium locality reflects a burrowing process in which the tunnel was lengthened by a succession of probes rather than by relatively continuous excavation and backfilling at the tunnel bend.
Most of the spreiten in our collection have a straight medial line in plan view, reflecting no change in direction as the burrowing progressed (Fig. 4). Exceptions represent a gradual shift in direction as the U-shaped tunnel was extended downward (Fig. 6). In any one burrow the tunnel arms are parallel and constant in width, implying no growth of the tracemaker during creation of the tunnel. In our best-preserved specimens (Fig. 4), both arms truncate the spreite; we have not observed a distinction between ‘truncating arm’ and ‘merging arm’ as described by Basan and Scott (1979).
Many authors have described Rhizocorallium as a protrusive burrow with retrusive arms (e.g., Seilacher, 2007, p. 57–59). Typical burrows at the Este Lado Rhizocorallium locality exhibit this combination, which resulted from a change in the tracemaker's behavior from lengthening the tunnel to moving it upward through overlying sediment (Figs. 7B and 8). The latter process involved removal of sediment from the tunnel roof and adding it to the floor, thus producing a stack of concave-upward laminae beneath each arm (Fig. 8). In contrast to the absence of meniscate lamination in longitudinal cross sections of spreiten, the lamination in retrusive arms is very apparent. It is subject to exfoliation at the surface, leaving the marginal tunnel in concave epirelief where the upper members of the retrusive stack have been lost (Fig. 4). The maximum upward shift we have measured for retrusive arms of specimens from the Este Lado Rhizocorallium locality is 16 mm. This is a minimum figure, because the retrusive stack appears to have been truncated by erosion at the outcrop surface.
ENVIRONMENTAL RANGES OF RHIZOCORALLIUM AND SIMILAR ICHNOFOSSILS
Rhizocorallium is a typical member of the Cruziana ichnofacies, interpreted to range from infralittoral to offshore marine (e.g., Frey and Pemberton, 1985). Most workers consider burrows of Rhizocorallium to have been dug in a softground substrate by animals living in water of normal marine salinity. That ichnogenus also is present, however, in strata thought to have originated in marginal marine environments of reduced salinity. Rhizocorallium is known in strata interpreted as representing estuaries, bays, and lagoons along the western margin of the Cretaceous Western Interior Seaway (as reported by Hubert et al. (1972), Basan and Scott (1979), MacEachern and Pemberton (1994), Daly (1997), and Oboh-Ikuenobe et al. (2008)). Additionally, several papers list Rhizocorallium as a member of the Glossifungites ichnofacies, a suite of traces presumably formed on compacted firmground substrates. As such, it is represented in Devonian brackish-water lagoon/bay deposits in the state of New York (Smith and Jacobi, 2001), in Carboniferous deltaic deposits of England (Eagar et al., 1985) and Scotland (Chisholm, 1970), and Cretaceous estuarine deposits in western Canada (Pemberton et al., 2001).
In contrast to the many citations of Rhizocorallium in normal marine and marginal marine settings, we know of only one paper describing it in a completely nonmarine context. Fürsich and Mayr (1981) provided a detailed description of U-shaped spreite burrows in the Miocene Upper Freshwater Molasse of Germany. The traces, assigned to R. jenense, occur in large and dense populations. They are comparable to the Ferris burrows in size and general morphology. Differences include the common occurrence of vertically retrusive arms in the Wyoming collection and presence of individuals with diverging arms (Fürsich and Mayr, 1981, fig. 4) among the German representatives. The Miocene depositional environment was interpreted as partially abandoned river channels, and the authors noted that similar burrows are made in present-day river banks by mayfly nymphs. Although these modern insect traces have been cited as spreite burrows (Seilacher, 1967), our search for illustrations of mayfly burrows has yielded only U-shaped tubes without spreite (e.g., Seilacher, 1967, fig. 1; Chamberlain, 1975, figs. 19.3A, 19.6F, 19.7G; White and Miller, 2008, figs. 3, 4H).
Fuersichnus, a nonmarine ichnogenus, includes forms that might be mistaken for Rhizocorallium. Bromley and Asgaard (1979) coined the name for burrow complexes in Triassic continental deposits of Greenland. Hasiotis (2002, p. 106–107) has provided additional discussion and illustrations. The Greenland Fuersichnus tracemaker produced clusters of J-shaped burrow fills in lacustrine sediments. Where the burrowing involved successive retrusive excavations approximately within a single plane, a spreite-like pattern was formed by the series of distal parts of the J-shaped tubes. In contrast to Rhizocorallium, the behavior evidenced by Fuersichnus is retrusive, and parallel arms do not border the spreite-like pattern.
Based on the above information, we conclude that the presence of Rhizocorallium in the Ferris Formation is strong evidence for marginal-marine, nearshore or neritic conditions during burrow construction. The lack of post-1981 reports of the ichnogenus in continental deposits indicates that the well-documented German Miocene occurrences (Fürsich and Mayr, 1981) are rare exceptions to the preponderance of evidence for Rhizocorallium as a marine indicator.
PALEOGEOGRAPHIC BACKGROUND AND IMPLICATIONS
Historical Background and Early Insights
It is useful to review the current understanding of the youngest phases of evolution of North America's ‘Western Interior Seaway.’ The concept that the seaway persisted in reduced form on the midcontinent well beyond the demise of the dinosaurs is hardly a new one. Recognition of the seaway's persistence through latest Cretaceous and into early Paleocene time was, however, hindered by stratigraphic confusion and initial misconceptions of geologic ages of key rock units. Additionally, inadequate attention had been paid to what we now recognize as prescient observations by some of the earliest geological explorers of the borderland between the Dakotas.
The story had its earliest steps in 1907 with discovery by Leonard (1908, p. 49) of oyster-bearing strata (now known as the Ludlow Member (or Formation) of the Fort Union Formation (or Group), representing early Paleocene time) in Billings County, southwestern North Dakota. Oysters are generally appreciated to be euryhaline animals. Soon thereafter, based on field work undertaken in 1912 and 1913 farther east in southern North Dakota along both sides of the Cannonball River (west of the Missouri River), Lloyd (1914, p. 247) reported up to 300 ft (90 m) of strata “Underneath the Fort Union” bearing remains of previously unknown assemblages of fully marine mollusks. Lloyd stated, these beds “… are now tentatively classified as probably of early Tertiary age and which have been referred to the Lance formation.” He (1914, p. 247–248) continued: “These beds have been mapped separately and are herein designated the Cannonball marine member of the Lance formation. The underlying lower part of the Lance is of fresh-water origin and is composed of alternating beds of sandstone and shale ….” Lloyd (1914, p. 249) also considered this ‘lower part of the Lance formation,’ which holds dinosaurian fossils, to represent the “Tertiary (?) System” because: “Fossil leaves collected near the top of the lower part of the Lance have been identified by Mr. Knowlton as belonging to the Fort Union flora.”
In the following year, as reviewed by Lloyd and Hares (1915, p. 523):
“Field examination by the writers and the paleontological determinations by Drs. Stanton and Knowlton during the years 1912 and 1913 show that in a large region west of Missouri River in North and South Dakota the Lance formation consists of two distinct parts, a lower non-marine part containing a flora very similar to, if not identical with, that of the Fort Union and an upper marine member containing a fauna closely resembling, but not identical with, that of the Fox Hills sandstone. This upper part, on account of its peculiar fauna, has been mapped separately and named the Cannonball marine member of the Lance formation. Farther west non-marine beds bearing lignite and occupying a similar stratigraphic position have been named the Ludlow lignitic member of the Lance.”
The above-quoted recognition by Lloyd and Hares of the correlative stratigraphic positions of the Ludlow and Cannonball rock units (no-matter which formally named rock unit or specific interval of geologic time they might represent) led to their statement (1915, p. 539–540) that: “The oyster beds near Yule in Billings County, North Dakota, first discovered by Leonard and later described by Stanton, may represent the western-most limit to which the Cannonball sea extended.” In other words, Lloyd and Hares (1915, p. 540) clearly had recognized that Leonard's oyster beds represented a “… westward extension … of the strata of marine origin, as the oysters are brackish-water animals and consequently must have had some connection with the open sea.”
With still greater relevance to the present paper, Lloyd and Hares (1915, p. 543) stated: “… the presence in the Dakotas of a marine fauna very similar to that of the Fox Hills sandstone, overlying the fresh-water sediments of the Lance, renders less tenable the theory of an unconformity of any importance at the base of the Lance, since the open sea must have persisted throughout Lance time in a region not very remote from western North Dakota.” Many subsequent workers (including Lillegraven and Ostresh, 1990, map 33 of fig. 4) overlooked Lloyd and Hares' insightful implication that, even though it may have withdrawn markedly, the latest Cretaceous Western Interior Seaway never did actually leave the midcontinent. Rather, within at least part of its former extent (as considered below), it secondarily expanded during early Paleocene time and persisted for about another five million years, to be renamed the ‘Cannonball Sea.’
The next major step was provided by Stanton (1920), who offered a well-documented stratigraphic discussion (focused on what was still called the ‘Cannonball marine member of the Lance formation’), an updated geologic map, and a description of diagnostic bivalves, gastropods, and elasmobranch teeth preserved within that rock unit. This was excellent work for the time. But on the following basis, and in Stanton's opposition (1920, p. 15) to official classification by the U.S. Geological Survey as Tertiary (which still applied Lyell's epochal term ‘Eocene’ in preference to Schimper's split of its older parts as the ‘Paleocene Epoch’), he erred in expressing opinion that the ‘Cannonball and Ludlow members of the Lance formation’ represented Cretaceous time:
“The comparatively small area in which the Cannonball member has been found lies about a thousand miles from the nearest Coastal Plain Eocene at the head of the Mississippi embayment and about the same distance from the nearest Pacific Border Eocene in Oregon and Washington. No other point in the United States is farther from known marine Eocene. Geographically and historically, therefore, the Cannonball marine member of the Lance formation is connected with the Cretaceous rather than the Eocene.”
The directly preceding quotation provides a sobering example of how ‘geographic parsimony’ in part led to a seriously incorrect conclusion. Contributing to the error in geochronology, Stanton put special emphasis on his observation that identity of 40 percent of the Cannonball molluscan fauna agrees with local species from the ‘Fox Hills sandstone’ (which underlies the Lance). But we now appreciate, through use of evidence unavailable to Stanton, that although the Lance and Fox Hills were deposited during very late Cretaceous time, the Cannonball and Ludlow both represent earliest parts of the Tertiary, now referred to the Paleocene Epoch.
Diversity of Known Cannonball Biota
The broad diversity of fossilized animal, plant, microbiological, and ichnological remains known from the main body of what is now called the ‘Cannonball Member (or Formation) of the Fort Union Formation (or Group)’ combines to strongly confirm Stanton's (1920, p. 15) conclusion that it was deposited under fully marine conditions. But the Cannonball also is represented by what must have been laterally more extensive, westerly tongues of sediment laid down under ephemeral and progressively more brackish-water conditions (Kroeger and Hartman, 1997). Some of the extensions of the Cannonball Sea have been recorded as far west as southeastern Montana (Fig. 9) east of the Miles City Arch (Belt et al., 1997, 2004, 2005). We present a sampling of key references to the Cannonball biota as follows:
microbiota (diverse acritarchs, algal spores, diatoms, moss spores, fern spores, dinoflagellate cysts, lycopod spores, vascular-plant pollen, benthonic and planktonic foraminiferans): Fox and Ross (1942); Stanley (1965); Fox and Olsson (1969); Van Alstine (1974); Kroeger and Hartman (1997); Belt et al. (2005).
corals: Vaughan (1920); Wilson (1957).
mollusks (bivalves, gastropods, nautiloids): Stanton (1920); Cvancara (1965, 1966); Feldmann (1972); Van Alstine (1974); Scholz and Hartman (2007).
ostracodes: Swain (1949).
crustaceans (lobsters, crabs): Holland and Cvancara (1958); Feldmann and Holland (1971)
ichnofossils (diverse burrows and trackways): Van Alstine (1974); Belt et al. (1997, 2005); Murphy et al. (2002).
vertebrates (elasmobranchs, chimaeroids, ray-finned fishes, reptiles): Stanton (1920); Cvancara and Hoganson (1993).
Midcontinental Continuity of Latest Cretaceous–Early Paleocene Marine Conditions
The Cannonball Formation's fossil record, linked to efforts by many workers in detailed geologic mapping, provides confirmation of its generally shallow water, marine depositional conditions and its Paleocene age. Evidence from study of the Cannonball's foraminiferal record by Fox and Ross (1942) and later by Fox and Olsson (1969) was particularly influential in affirming its age, principally through comparisons with similar species known from the Paleocene Midway Group of North America's Gulf Coast. Pioneering efforts in magnetostratigraphy and other forms of geochronologic documentation (Lund et al., 2002; Belt et al., 2004; Peppe et al., 2009) are adding to temporally controlled knowledge about expansions and contractions of the Cannonball Sea as occurred in the Williston Basin of the Dakotas and eastern Montana.
Building upon almost forgotten studies from early in the 20th century, subsequent workers have made great progress in the mapping and description of ephemeral expansions of the latest Cretaceous Western Interior Seaway, as recognized from evidence within the correlative and otherwise nonmarine, dinosaur-bearing Lance and Hell Creek Formations of the western Williston Basin. Laird and Mitchell (1942), for example, with additions from Frye (1964), recognized two tongues of rocks within the Hell Creek Formation of North Dakota containing brackish-water or marine fossils. Frye (1964, p. 169) considered the stratigraphically lower strata (near the base of the Hell Creek) to represent “… a large tongue of the Fox Hills Formation thinning to the west ….”
In agreement with that broad concept, Hartman and Butler (1995) stated: “The nonmarine K/T [Cretaceous–Tertiary] section in the Missouri River valley, south central North Dakota, is very thin, with no more than a few meters occurring between the Fox Hills and Cannonball Formations, indicating the continued presence of an interior seaway.” Murphy et al. (2002, p. 9) stated: “These occurrences of marine- or brackish-water-derived strata near both the base and top of the Hell Creek Formation in south-central North Dakota suggest the Fox Hills Sea [i.e., the Western Interior Seaway] may not have withdrawn completely from the area before the advance of the Cannonball Sea in Paleocene time.”
Based on study of brackish and marine mollusks known from upper parts of the Hell Creek Formation, Hartman and Kirkland (2002, p. 271) opined: “… the molluscan record of the last 1–2 m.y. of the Cretaceous indicates environmental settings characteristic of the presence of an adjacent interior sea…. Thus a generally perceived dramatic regression of the Western Interior Sea and a completely nonmarine K–T stratal interval of 1–2 m.y. did not occur in the central interior of North America.” Johnson (2002) contributed a well-documented analysis of macrofloral change in the western Williston Basin through the interval of fluctuating shorelines from very late in the Cretaceous through early Paleocene time.
Summarizing the above discussion, Scholz and Hartman (2007) concluded: “The marine sediments of the main body of the Cannonball Member and its brackish tongues reflect the continued transgression of the seaway and subsequent multiple changes in sea level throughout the first 5 Ma [sic] of the Paleocene. Marine-influenced deposition dominates the early- and mid-Paleocene record of the eastern part of the Williston Basin ….” We wholly agree that the latest Cretaceous version of the ‘Western Interior Seaway’ persisted in North America's midcontinent well into early Paleocene time, renamed within the Williston Basin as the ‘Cannonball Sea.’ As discussed in the following section, however, we know little about detailed limits of the Paleocene Cannonball Sea beyond the Williston Basin.
Paleogeographic Affinities of Cannonball Biota
Case for Marine Connections to Arctic Ocean
Because relevant exposures of marine Paleocene rocks are so limited within the North American midcontinent (Fig. 9A), paleogeographic evaluations of the Cannonball biota depend almost completely upon degrees of similarity with fossilized assemblages known at more distant marine sections. Fox and Ross (1942, p. 668), for example, recognized strong similarity at the species level between foraminiferans collected from the Cannonball Formation in the Williston Basin and the Paleocene Midway Group of the Gulf Coast. Although their research emphasis was to provide age control for the Cannonball, it became widely assumed that the taxonomic similarity implied direct seaway connection in Paleocene time between North Dakota and the North American Gulf Coast. Inadvertently reinforcing that viewpoint was research on marine ostracodes by Swain (1949, p. 174), who recognized that: “All of the [five] species obtained from the Cannonball formation … occur in the Paleocene Midway group of Texas.”
But alternative paleogeographic considerations broadened following later study of Cannonball benthic foraminiferans in north-central North Dakota by S. K. Fox, Jr. As reported (in Lemke, 1960, p. 31), “… these Foraminifera more nearly resembled the Arctic forms described from northern Europe than those of the Midway strata of the Gulf Coast. Therefore, although no marine fauna of Cannonball age has yet been found in the Canadian region, it is likely that the Cannonball sea invaded this area as an arm of the Arctic sea.” Stanley (1965, p. 179), following study of a taxonomically wide suite of microfossils from Cannonball strata in northwestern South Dakota, concurred, stating: “Many of the species of both the plant microfossils and the fossil plankters have also been described from other areas of the world such as western Europe, Siberia, and Australia.”
Similarly, Cvancara (1966) noted that his research on North Dakotan bivalved mollusks showed few species to be in common with the Midway Group of the Gulf Coast. Rather, the Cannonball pelecypods show greater similarity to those of northern Europe. Feldmann (1972) described a fossil from the Cannonball Formation of south-central North Dakota that he identified as one of the commonest nautiloid cephalopods known from the Midway Group. Nevertheless, he asserted: “Although any lithologic evidence of a former connection of the Cannonball seaway with the Arctic Ocean or the North Atlantic Ocean has been removed by erosion, the faunal evidence strongly indicates that such a connection must have existed during the Paleocene. No evidence, lithologic or paleontologic, indicates a physical connection between the Cannonball seaway and the Mississippi Embayment.”
Marincovich et al. (1985, p. 772) studied Paleocene to early Eocene marine mollusks and ostracodes from Ocean Point, northern Alaska, and they concluded:
“The strongest faunal affinities of the Ocean Point mollusks and ostracodes are with the Paleocene Cannonball Formation of North and South Dakota, and with post-Danian Paleocene to early Eocene (Thanetian to Ypresian) faunas of northwestern Europe.”
“The Cannonball deposits are considered by us to represent the most southerly known extension of the early Tertiary circumboreal sea.”
“The Cannonball bivalves and the Ocean Point mollusks and ostracodes are of cool temperate, low-diversity aspect, distinctly different from the subtropical, extremely diverse Paleocene faunas of the Gulf of Mexico.”
Marincovich et al. (1985) suggested that the Arctic Ocean was almost completely isolated from other oceanic basins from very late in the Cretaceous until sometime in the Eocene. They postulated, however, that direct marine connections existed during Paleocene time from the Cannonball Sea across the Arctic Ocean to much of northern Europe, including the London Basin.
Marincovich and Zinsmeister (1991) supported that paleogeographic viewpoint by noting the Paleocene marine gastropod Drepanochilus pervetus (Stanton, 1920) has been found in common among correlative strata of the Cannonball Formation, Ocean Point area in northern Alaska, the southern North Sea Basin, London Basin, and in newly discovered assemblages on Ellesmere Island of northeastern Canada. They suggested free exchange of marine organisms among arms of shallow seas radiating southward from a polar Arctic Ocean. Marincovich and Zinsmeister specifically precluded marine connection of that geographically sequestered boreal complex of seaways with the northwestern Atlantic Basin. However, Cvancara and Hoganson (1993, p. 18) suggested, on the basis of fossilized fishes of the Cannonball found in common with western Greenland, Maryland, and Virginia “… that the Cannonball Sea was connected to the Atlantic Ocean.”
Even though the above information may continue to hold puzzling details, it seems to us that evidence for marine connections to the north by the Cannonball Sea, as classically known from the Williston Basin, is strong. Despite some assertions cited above to the contrary, we also see merit in evidence derived from the fossil record in support of southerly connections of the type Cannonball Sea to the North American Gulf Coast. But just how does the record of Rhizocorallium we presently describe fit into the big picture of seaways that persisted in the midcontinent during early Paleocene history? Are we seeing a record of major southward extension of the Cannonball Sea into the Hanna/Carbon Basin, or are we studying independent effects of a terminally isolated, northern marine embayment from the Gulf of Mexico (Fig. 9B)?
Southern Extent of Cannonball Sea
As shown by the lightened patches in Figure 9A, strata of Paleocene age have been lost through erosion in the southern three quarters of western South Dakota, all of western Nebraska and Kansas, and most of the southeastern and northeastern quarters of Wyoming and Colorado, respectively. Thus, east of the Black Hills, roughly 400 miles (ca. 640 km) of modern landscape separate relevant fossiliferous marine Paleocene strata of the southernmost Williston Basin from their correlatives in Colorado's Denver Basin.
In the Denver Basin, the Laramie Formation directly overlies the coastal-marine, Fox Hills Sandstone. The Laramie is a coal-bearing, temporal equivalent of the very late Cretaceous, Lance and Hell Creek Formations of more northerly latitudes. Weimer and Land (1975) and Weimer and Tillman (1980) provide thorough description and analysis of the Laramie Formation. They interpreted most of it as representing a freshwater-dominated, near-shore delta plain paleoenvironment. However, the shallow-water marine foraminiferan Haplophragmites is found in shale units overlying typical sandstones of the Fox Hills Formation. Weimer and Tillman (1980, p. 37) suggested: “The shales probably represent deposits in low energy marine bays on the delta plain.” As discussed below, brackish-water molluscan faunas also are known from the Laramie Formation of the Denver Basin.
Raynolds (2002) and Raynolds and Johnson (2003) provided insightful summaries of the Laramide history of uppermost Cretaceous and lower Tertiary strata of the Denver Basin. They did not hint to the existence of marine-influenced strata higher in the local section than the Laramie Formation. That, however, does not necessarily preclude a consequential history of shallow-water, marine deposition in the Denver Basin during Paleocene time. A hiatus of about eight million years within the Paleocene sequence, involving a major interval of erosion that interrupted an otherwise aggradational depositional setting, has been recognized across the basin (Raynolds, 2002). Many lesser hiatuses within the sequence also have been linked to Paleocene strata (Farnham and Kraus, 2002; Hicks et al., 2003).
With the above information in mind, one must concede that the geological or paleontological record, as known east of the Black Hills Uplift, provides little direct evidence about seaway distributions during Paleocene time across the landscape from northern South Dakota to south of the Denver Basin. But what of the extensive Paleocene sedimentary column seen in the Powder River Basin in northeastern Wyoming (west of the Black Hills) and southeastern Montana (west of the Miles City Arch; Fig. 9A)?
We have been able to find no evidence for existence of marine influence on the extensive Paleocene outcrop or subsurface record from the Fort Union Formation anywhere within the Powder River Basin. On the contrary, direct geological evidence exists to suggest that latest phases of the Western Interior/Cannonball Seaway would have been precluded from entry into the Powder River Basin by latest Cretaceous–early Paleocene uplift and subaerial erosion of the Black Hills and Miles City Arch. Lewis and Hotchkiss (1981) analyzed massive amounts of data from hundreds of outcrops and assemblages of drill cuttings, well cores, and geophysical logs spaced regularly across the entire Powder River Basin of Montana and Wyoming. They synthesized the data to produce a temporally controlled series of five contoured, sand-distribution maps (showing thicknesses, percent sand, and topographic elevations at bases of specific members or formational boundaries). Sheets 5 (upper Hell Creek Formation—uppermost Cretaceous) and 4 (Tullock Member of Fort Union Formation—lowermost Paleocene) are of special interest for present purposes.
Analysis of sand body contours by Lewis and Hotchkiss (1981) for the upper Hell Creek Formation along eastern parts of the Powder River Basin is important in clearly showing, both north and south of the Montana–Wyoming state boundary, east–to–west-transported (and westward-thinning) lobes of sand derived from what today constitutes northwestern parts of the Black Hills and southern parts of the Miles City Arch. Lillegraven and Ostresh (1988, p. 308 and fig. 4), in attempting paleogeographic reconstructions of Wyoming's geologic evolution through the Laramide orogeny, erred in suggesting that no evidence existed for uplift of the Black Hills in latest Cretaceous time. There can be little doubt that the earliest subaerial appearance and erosion of the Black Hills did, indeed, occur prior to the earliest Paleocene.
Sheet 4 by Lewis and Hotchkiss (1981) provides evidence for progressively stronger uplift of the northern Black Hills and southern Miles City Arch during deposition, directly to the west, of the Tullock Member (or Formation) of the Fort Union Formation (or Group). Flores and Ethridge (1985, fig. 5, p. 113) contributed rose-diagrams for paleocurrent data that confirm the case for east–to–west transport of sands during earliest Paleocene time (Puercan Land Mammal Age) into the Powder River Basin. Lillegraven and Ostresh (1988, fig. 5 and p. 310) erred in the assumption that the Puercan was the time of proto-appearance of the Black Hills. Data from Lewis and Hotchkiss (1981, sheet 4) also show the first clear indications of the characteristic asymmetry across the Powder River Basin through the several-fold thickening of the Tullock from north to south and the progressively greater depth (relative to modern sea level) of the base of the Tullock.
In any case, the evidence provided by Lewis and Hotchkiss (1981), supplemented by Flores and Ethridge (1985), and applied paleogeographically by Lillegraven and Ostresh (1988) has important implications in present context. The possibility would have been slight that marine waters entered either the Montana or Wyoming components of the Powder River Basin from east of the prototypical Black Hills Uplift from the latest Cretaceous (i.e., after deposition of lower parts of the Lance and Hell Creek Formations) through the remainder of Paleocene time.
Uncertainties of Marine Connections to Southern Wyoming
Schlaikjer (1935), provided a summary of his geologic mapping and associated paleontological studies within Goshen Hole, a structural dome eroded to form a small topographic basin in southeastern Wyoming (Fig. 9A, site 13). Schlaikjer (in his fig. 2) recorded several localities containing brackish-water mollusks, which he variously referred to as the ‘Corbicula zones,’ ‘Oyster beds,’ and ‘Corbula zone’ (collectively involving geologically long-lived and geographically widely distributed genera of standard bivalves and gastropods). He tentatively considered all of the sites to be within Upper Cretaceous strata of the Lance Formation. Although queried, Schlaikjer indicated at least two localities with remains of dinosaurs and other terrestrial vertebrates above the ‘Corbicula zones.’ He named that stratigraphically unusually high, fluvial part of the Lance Formation the ‘Torrington Member.’ W. A. Clemens, Jr. (written communication to Lillegraven of December 6 and 10, 2010) confirmed from his own field notes that fossil fish, turtles, dinosaur teeth, and other large bones do indeed exist above the Corbicula-Ostrea beds as proposed by Schlaikjer.
Schlaikjer (1935, p. 54), following transmittal of his mollusk collections for identification to specialists in the U.S. Geological Survey, quoted the following reply from J. B. Reeside, Jr.:
“All of the lots represent a brackish-water fauna. There are no strictly fresh-water shells and no typically marine shells. All of the species occur in the Laramie formation of the Denver Basin, and most of them have been recorded from the lower Lance formation. By themselves they do not indicate an extension of the Cannonball sea into southeastern Wyoming, though they do indicate the presence near by of truly marine waters.”
Thus, although no strata of Paleocene age are known east of the Laramie Basin (Fig. 9A) in southeastern Wyoming, the almost forgotten very Late Cretaceous, brackish-water records provided by Schlaikjer in Goshen Hole and by Reeside in the Denver Basin indicate nearby presence of the Western Interior Seaway in latest Cretaceous time.
Now we shift attention westward into the Hanna/Carbon Basin of south-central Wyoming. Wroblewski (2004, p. 249) cited his own dissertation as authority for the presence of Rhizocorallium, along with eight other marine ichnogenera, within the local Ferris (latest Cretaceous–early Paleocene) and overlying Hanna (remainder of Paleocene into earliest Eocene) Formations:
“… low-diversity marine ichnofossil assemblages (Ophiomorpha, Thalassinoides, Teredolites, Skolithos, Planolites, Teichichnus, Bergaueria, Rhizocorallium, and Diplocraterion) are preserved in key stratigraphic intervals in both formations, indicating at least brackish water (Wroblewski, 2002).”
The only mention of Rhizocorallium within Wroblewski's (2002) monograph, however, is on page 211 within a systematic section devoted to ‘? Teichichnus.’ He stated: “The specimens are similar to Rhizocorallium, but lack the two vertical tubules that form the borders of the meniscate in-filling.” Stratigraphic, temporal, and locality data for the relevant examples of? Teichichnus were given as “middle Ferris Formation (early Paleocene: 64 Ma), central Hanna Basin, east of Seminoe Reservoir (sec. 18, T 21 N R 82 W)” (Wroblewski, 2002, p. 210). But those coordinates are in a southern part of the Hanna Basin some 13 miles (21 km) southeast of where we think Wroblewski's field observations were made and would involve strata traditionally mapped as North Park Formation (?Miocene). We found no other references in other sources to document presence of Rhizocorallium either in the Ferris or Hanna Formations of south-central Wyoming.
It appears, therefore, that the present paper is the first verifiable record of the ichnogenus Rhizocorallium from uppermost Cretaceous or Paleocene strata of southern Wyoming. We do not challenge the possible validity of the other putatively marine ichnogeneric occurrences in the Hanna/Carbon Basin claimed in the various references by Wroblewski and his junior authors. We do, however, assert that the combination of stratigraphic and geographic documentation they provided for each proposed occurrence of trace fossils is either incorrect or inadequately detailed for practical use in field-based verification.
In the historical review presented above, we summarized observations derived from more than a century of research that favor persistence of latest Cretaceous and early Paleocene seaways within the North American Western Interior. We wholly accept the concept that significant depositional influences from marine waters have existed through that interval within fluvial systems that fed the low-lying coastal plain along the western borders of the ever-changing seaways. Additionally, we propose that the Rhizocorallium-bearing, lower Paleocene strata at Pats Bottom provide evidence beyond tidal influences for a westward-directed, ephemeral inundation by shallow seawaters (i.e., true ‘marine’ sensu FGDC, 2010). The transgression briefly affected early Paleocene depositional environments in the eastern greater Green River Basin that had yet to become tectonically subdivided into the Hanna, Carbon, and Laramie Basins (see maps in Lillegraven et al., 2004, figs. 15 and 19). However, those two paleogeographic reconstructions (as also would be the case for the early Paleocene reconstructive maps proposed in figs. 5 and 6 by Lillegraven and Ostresh, 1988) almost surely were flawed in having suggested early Paleocene uplift histories of the Laramie Mountains (Fig. 9A). Earliest Cenozoic (i.e., Puercan or even earliest Torrejonian) elevation of that mountain range probably would have precluded even brief pulses of inundation to the west of its mass by marine waters from the east.
We suggest the Rhizocorallium-bearing strata at Pats Bottom represent the most westerly known marine (sensu FGDC, 2010) inundation of the Paleocene landscape within the state of Wyoming—and very probably within the continent south of the Canadian border (see Fig. 9B). Although tidal influences and increased salinity may have affected rapidly aggrading fluvial systems in Montana and Wyoming even much further to the west throughout Late Cretaceous and early Paleocene time, we recognize actual marine inundations this far to the west as having been transitory, uncommon events during the Paleocene.
To date, interpretations of marine influences within Paleocene strata of the Hanna/Carbon Basin have been based principally upon sedimentological and ichnofossil-based evidence. Not a single Paleocene locality in south-central Wyoming has been reported to contain anything like the taxonomically diverse body fossils of marine invertebrates or microfossils (summarized above) known from the Cannonball Formation or from its short-lived, west-projecting tongues. The absence of non-piscine body fossils exists in spite of common occurrences in exposures of Ferris and Hanna Formations of the kinds of vertebrate fossils and woody fragments that are regularly found elsewhere in association with diverse molluscan assemblages. This situation argues against the persistent existence of stable near-shore marine or estuarine settings.
As a related consideration, what were the paleogeographic relationships of the seaway to the east? Do the Paleocene indications of marine connections within the Hanna/Carbon Basin demonstrate, as concluded by Wroblewski (2002, p. 137), “… presence of a previously undescribed southern embayment of the ‘Cannonball Sea’”? His supposition was founded on presence of putative marine ichnofossils and fossilized teeth of sharks and shark allies from the Ferris and Hanna Formations. Wroblewski (2004b, fig. 2 and table 1; also see Burris, 2001) later refined identifications of the selachians. The relevant ichnogenera reported from the Ferris and Hanna Formations all have extensive geographic and geologic ranges, thus restricting their paleontological utility to indicators of ecological conditions. It is important to recognize that the fossilized selachian teeth are the only body fossils from the local study area that have been used to suggest marine origin of the enclosing strata. The bulk of taxa reported by Wroblewski (2004b), however, are from a single locality in the Ferris Formation (University of Wyoming locality V-72010; see Eberle and Lillegraven, 1998, figs. 2 and 3 and Lillegraven and Eberle, 1999, appendix I), located below the paleontologically recognized Cretaceous–Tertiary boundary. Additionally, none of the three kinds of selachians reported from the Hanna Formation (Paleocene) has been reported from the Cannonball Formation (Cvancara and Hoganson, 1993, fig. 5).
There currently exists no means known to us to characterize the Paleocene record of the upper Ferris or Hanna Formations in terms of species uniquely shared with the Cannonball biota. Although some workers might consider it more parsimonious to simply assume persistence of direct marine connections between the Hanna/Carbon Basin and Cannonball Sea in Paleocene time (e.g., lighter-toned part of hypothesized seaway indicated southeast of Black Hills in Fig. 9B), it seems to us at least equally probable that any marine connection to southern Wyoming was distinct from linkage with the Cannonball Sea and more directly to the south. In absence of verifiable evidence beyond what was summarized above, only speculation can be employed to favor one paleogeographic alternative over another.
SUMMARY AND CONCLUSIONS
Spreite burrows abound within an outcrop in the uppermost third of the Ferris Formation in south-central Wyoming's Hanna Basin. The burrowed strata overlie fossiliferous rocks representing very late parts of the Puercan Land Mammal Age (ca. 64 Ma, early Paleocene). This top third of the formation could be latest Puercan and/or early Torrejonian in age, recognizing perhaps two million years of temporal uncertainty. Traditionally, this section has been considered as coal-rich, continental deposits formed after full withdrawal of the Cretaceous Western Interior Seaway.
The outcrop provides the first verifiable record of the ichnogenus Rhizocorallium either from uppermost Cretaceous or Paleocene strata of southern Wyoming. The trace fossils display all the distinctive features of this common ichnogenus (including U-shaped, parallel-margined tunnels, spreite, and oblique but shallowly descending orientation relative to originally horizontal bedding).
Occurrences of parallel-sided Rhizocorallium typically are common in subtidal-marine, nearshore to neritic strata as members of the softground ‘Cruziana ichnofacies.’ Burrows of Rhizocorallium also are known, but less commonly, in marginal-marine deposits of the firmground ‘Glossifungites ichnofacies.’ Presence of Rhizocorallium in Miocene freshwater strata of Germany is an exception to preponderance of geological evidence for utility of this ichnogenus as an indicator of ancient marine or marine-influenced environments.
We conclude that these early Paleocene traces in the Ferris Formation were constructed by benthic animals inhabiting a marginal-marine, nearshore or neritic environment. The marine pulse represented westward expansion of a formerly far more extensive Western Interior Seaway that persisted in the midcontinent through even latest Cretaceous time. The conclusion is compatible with independently reported sedimentological and trace-fossil claims of tidally influenced fluvial systems stratigraphically above and below this part of the Ferris Formation.
The Rhizocorallium-bearing strata of the Hanna Basin represent the most westerly known marine (sensu FGDC, 2010) inundation of the Paleocene landscape within the state of Wyoming and very probably within the continent south of the Canadian border. Tidal influences may have affected rapidly aggrading fluvial systems in Montana and Wyoming still farther to the west throughout Late Cretaceous and early Paleocene time. Nevertheless, we regard actual marine inundations this far west during the Paleocene as uncommon and geologically ephemeral events.
No Paleocene localities in south-central Wyoming have yet yielded assemblages of body fossils of marine invertebrates or microfossils as known from the type Cannonball Formation (or its various tongues) of the Williston Basin. No verifiable means, therefore, have yet been recognized to characterize the Paleocene record of the Hanna/Carbon Basin in terms of species uniquely shared with the Cannonball biota. Short-lived marine connections to the Hanna Basin may have been solely to the Gulf Coast, solely to the Cannonball Sea, or perhaps contiguity existed with both. Sedimentological evidence analyzed by other workers from across the Powder River Basin strongly suggests that no marine extensions of the Cannonball Sea existed west of the emerging Black Hills Uplift during Paleocene time.
Uplift of the Laramie Mountains served to define the eastern margin of southern Wyoming's Laramie Basin. Emergence of those mountains prior to middle Paleocene time would have precluded westward pulses of marine inundation into what is now the Hanna Basin. The areas recognized today as the Hanna, Carbon, and Laramie Basins, even as late as the Torrejonian, persisted as the undivided, eastern realm of an enormous greater Green River Basin.
We thank Drs. Edward S. Belt, Laurie J. Bryant, William A. Clemens, Jr., Joseph H. Hartman, and Peter Robinson for providing insights to the geological evolution and paleontology of the Williston and Powder River Basins. Bill Clemens deserves special thanks for sharing access to his unpublished paleontological explorations of Goshen Hole. Dr. Robert W. Scott and Scott Schell provided certain papers pertinent to the description of Rhizocorallium. Dr. Susan Swapp and Ted Starns helped with specimen preparation, and Dr. Norbert Swoboda-Colberg provided an x-ray diffraction analysis of a rock sample. Mr. Joshua Satterly was enormously helpful in contributing his photographic skills to specimen documentation. Dr. Michael L. Cassiliano aided in curating the specimens for entry to the permanent collection. Reference librarians at The University of Wyoming, as always, gave superb services. Mrs. Linda E. Lillegraven uniquely contributed with uninterrupted patience and logistical support. We are particularly grateful for the efforts toward improving this manuscript as provided through peer review by Drs. Mary R. Dawson, Anthony A. Ekdale, Robert W. Scott, and Michael O. Woodburne. Finally, we thank Brendon B. Orr, Managing Editor and Sarah R. Garlick, Co-editor of Rocky Mountain Geology, for the many ways in which they facilitated production of this article. The coauthors contributed equally to content and preparation of this paper, so the first name is based on alphabetical priority.
- Received January 19, 2011.
- Revision received March 28, 2011.
- Accepted May 16, 2011.