Printer Friendly

Coprolites, cololites, and fish fossils in the Mississippian Michigan Formation, Western Michigan.


Fossilized vertebrate excrement (coprolites) and digestive-tract contents (cololites) are identified in the Mississippian Michigan Formation, exposed in old gypsum mines in the Grand Rapids area (Kent County). In western Michigan the formation consists of cyclical gypsum beds, thin shales, dolomitic sandstones/siltstones, and detritus-rich dolomites. These rocks, which contain mud cracks, ripple marks, and terrestrial plant debris, suggest deposition in a shallow, marginal sea with alternate desiccation and stream floodwater input. The coprolites and cololites are in lenses of sandstone at the shale contact with underlying gypsum. Coprolites are common, whereas cololites are rare. The sandstone lenses also contain shark teeth and spines and other fish teeth and scales.

The medium to dark gray or light-gray to brown coprolites range from 2 to 30 mm in length and vary in shape (spherical to elongate to irregular). They were deposited under anoxic conditions as evidenced by a lack of borers, a paucity of other benthic organisms, and the presence of pyrite and glauconite within the specimens. The coprolites were formed by predators, herbivores, and bottom feeders, based on inclusions of bone, plant debris, or silt. The vertebrates that produced the coprolites cannot be identified, but the coprolites are found in association with a variety of fish and shark remains. The rare cololites are recognized on the basis of their coiling. One cololite, 5.2 mm in width, is coiled, has submucosal folds within the coils, and includes a well-preserved fish scale. Another specimen may be spiraled, and thus possibly attributed to the shark or ray families.


The Michigan Natural Storage Company (MNSC) underground facility, a former underground gypsum mine in Kent County, Michigan (Figure 1), provides exposures of the Mississippian-age Michigan Formation. In western Michigan this formation consists of six, cyclically bedded, thick layers of gypsum and thin beds of shale, dolomitic siltstones and sandstones, and detritus-rich dolomite. The strata have been informally divided into six units numbered 1 (youngest) through 6 (South et al. 1995). Generally, each unit consists of a layer of gypsum overlain by shale and then dolomite. The MNSC mine exposes units 1 and 2 in approximately 10 km of tunnels over a 300,000 square meter area at a depth of more than 24 m. The tunnels are generally about 3 m in height. Figure 2 shows the entire unit 2, a portion of the overlying unit 1 strata, and the dolomite and shale of the underlying unit 3. The contact between the top of the gypsum in unit 2 and the overlying shale is a diastem. Lenses of sandstone, which contain the coprolites and cololites of this study, occur along this contact (Figure 2). Using the terminology of McAllister (1985), in this paper coprolites are defined to be preserved excreted fecal material, whereas cololites are defined to be fossilized intestinal contents that were not excreted before the death of the animal.



Although the strata are not particularly fossiliferous, an interesting diversity of fossils is found throughout the section. In addition to the coprolites and cololites, the sand lenses and associated shale contain shark teeth and spines, fish scales, fish teeth, and occasional trace fossils (burrows). Plant fragments occur at the shale-dolomite contact in unit 2. Although Dorr and Eschman (1970) reported the gypsum as being unfossiliferous, within the gypsum in unit 2 we have found small, nodular masses with laminae that could be of algal origin.

This study examines the abundant, small (up to 30 mm), medium to dark gray or light gray-brown, spherical to irregularly shaped structures found within the thin sand lenses at the base of the shale in unit 2 (Figures 2 and 3). Dorr and Moser (1964) first reported these structures to be coprolites, but no published study verifies that they were the fecal matter of vertebrates. Herein we characterise the external shape of the structures, examine their internal fabric and composition, and characterize the associated host sandstone. The results indicate that the majority of these structures are indeed coprolites, but a few are identified and classified as cololites (intestinal structures). Cololites have not been reported previously from the Michigan Formation. These coprolites and cololites offer insights into the paleoenvironments of the Michigan Formation in western Michigan.



The stratigraphy of the Michigan Formation and associated units is summarized by Harrell et al. (1991). The Michigan Formation's extent, lithologies, and thickness throughout the Michigan Basin were mapped by Moser (1963) from subsurface well records and a few locations where the formation can be seen in outcrop or mines. Figure 1 shows the bedrock exposure of the formation below the Pleistocene and Quaternary sedimentary overburden. In the Kent County area the Michigan Formation is approximately 30 m thick, but it reaches a maximum thickness of more than 150 m near the center of the Michigan Basin, where shale is dominant with lesser amounts of sandstone, limestone, gypsum, and anhydrite (Harrell et al. 1991). The formation is conformable with the underlying Marshall Sandstone and overlying Bayport Limestone (Harrell et al. 1991), although these contacts are not exposed in the MNSC mine.

Previous studies have suggested ages for the Michigan Formation ranging from Middle to Late Mississippian. Earlier studies assigned the formation to Os-agian through Meramecian (Monnett 1948; Dorr and Moser 1964; Ells 1979; Shaver 1985; Harrell et al. 1991; Catacosinos et al. 2001). However, Cross and Taggart (2004) assigned a tentative age of Namurian to the Michigan Formation based on a study of fossil plant fragments and associated spores from the MNSC mine. Thus, the formation's age is unresolved.

Dorr and Moser (1964) identified fin spines from ctenacanth sharks and cladodont-and orodont-type shark teeth from what is now the MNSC mine. In addition, acanthodian fish scales and a portion of a fin spine, and numerous actinopterygian and sarcopterygian fish scales and tips of actinopterygian teeth from the MNSC were identified by Kenaga and Sellepack (1995). They also described a portion of the lower jaw of a probable rhipidistian (crossopterygian) fish; the first of its kind found in Michigan. The jaw contains six teeth in its 21 tooth sockets (Figure 4). Unfortunately, the specimen is missing ornamentation necessary to further classify it (M. D. Gottfried, personal communication, 2003). All of these specimens were recovered from the shales and sand lenses in units 1 and 2.


During field work for this study we also recovered four specimens that are thought to be fish cranial material including a head plate, paired and medial elements (possibly a vomer), a tooth socket, and a partial skull with dermal ornamentation (R.F. Stearley, personal communication, 2005). We are currently attempting to verify these initial identifications and the taxa of these specimens.


Individual coprolites and slabs of coprolite-bearing sandstone lenses and shale were collected at the gypsum-shale contact in unit 2 (Figures 2 and 3). One slab was prepared using a rotary tool to expose the coprolites and measure the orientation of their long axes. Three mutually perpendicular axes in 78 loose specimens were measured to obtain length (long axis), maximum width (intermediate axis), and minimum width (short axis). These measurements were then plotted on a Zingg shape diagram (Folk 1974) to determine the dimensional relationships of the coprolites.

Five thin sections were made of the sandstone lenses, and each thin section was point counted (three counts per slide). In each count the first 100 points determined the grains/matrix/cement ratio and identified all grain, matrix and cement types. Each count continued until a total of 100 quartz (Q), feldspar (F), and rock fragment (RF) grains were identified to establish the Q/F/RF ratio. The three counts per slide were then averaged and the sandstone classified according to Folk (1974).

Loose specimens of broken coprolites were analyzed using a dissecting microscope to observe structure and content of the coprolites. For more detailed analysis, 16 thin sections with multiple coprolites per slide were studied in an attempt to relate the internal content and structure of the coprolites to their size, shape, and color. In addition to these visual observations, a powder x-ray diffraction analysis was run on one coprolite specimen to determine the major composition of its groundmass.


Previous studies have not examined the petrography of the sandstone lenses which enclose the coprolites and associated fossils. The composition and character of these sands sheds some light on the source of the sediments, and the conditions which may have favored preservation of the fossils.


The fine-to-medium sand grains are mostly quartz, and range from angular to sub-rounded, averaging sub-angular. Using Folk's (1974) classification for the origin of quartz grains, four types of grains were identified: plutonic quartz (~74% of the total quartz fraction); strongly undulose quartz (~17%); metamorphic quartz varieties (~4%); and composite quartz grains (~5%). The feldspar fraction is dominantly altered orthoclase feldspar and perthitic ortho-clase feldspar with traces of microcline and plagioclase. Rock fragments are dominantly sedimentary and include chert, mudstone, and some coarser, quartzose rock fragments that could either be sedimentary or igneous in origin. Metamorphic rock fragments occur in trace amounts. Other grains include sand-to-pea-sized coprolites resulting in counts as high as 9%, heavy minerals (<1%), and unidentifiable shell fragments. The Q/F/RF ratios indicate these sandstones are sublitharenites (Folk 1974).


The most common matrix is smashed mudstone fragments and silt-sized particles. In a few samples discontinuous laminae of greenish material appear to be clay matrix. Shale clasts occur within this sandstone unit, and very small, deformed shale clasts between sand grains may have been identified as matrix.


Poikilotopic gypsum cement dominates the samples analyzed. The sediment appears to have had a relatively high porosity (well-washed sand) that was subsequently in-filled by the poikilotopic gypsum. Minor amounts of calcite cement are also present. Other diagenetic minerals include trace amounts of glauconite, pyrite, and anhydrite.


Coprolites are here grouped into three categories: spherical, elongate, and irregular (Figure 5). Within each group the coprolites ranged from < 2 to 30 mm in size and were either medium to dark gray, or light gray-brown. Although the majority of coprolites are equant or prolate, the Zingg shape plot (Figure 6) of 78 coprolites shows neither discontinuities nor dominant or preferred shapes that might allow their classification based on shape.



The bedding-plane slab suggests that the coprolites' long axes lie in the bedding plane, but with two orientations which are approximately perpendicular to each other. The sandstone lenses sometimes show poorly developed ripples, and the coprolites' dominant, long-axis orientation is perpendicular to ripple crests, whereas the secondary orientation is the result of coprolites lying in and paralleling ripple troughs.

A number of features are observed in the thin sections of coprolites: (1) clearly defined outer walls (Figure 7A), (2) pyrite replacing plant debris (Clark 1989), (3) bone inclusions (Figure 7B), (4) silt inclusions (Figure 7C), (5) a lack of boring (Figure 7 A and B), and (6) the presence of shell material and glauconite (Figure 7D). In addition, powder x-ray diffraction analysis of a coprolite sample indicates that the groundmass consists of calcium phosphate.



Orientation parallel to the bedding plane suggests that the coprolites are a primary depositional feature. Their verification as coprolites is based on the calcium phosphate composition (Hantzschel et al. 1968; Hunt 1992), the presence of an internal wall structure (Figure 7A), and inclusions of plant debris, bone (Figure 7B), silt (Figure 7C), and shell material (Figure 7D).

Variation in shape and color (Figures 3 and 5), wide range of sizes, and lack of correlation in the shape analysis (Figure 6) imply that multiple taxa were responsible for producing the coprolites. Neither their color nor shape can be used for classification. However, thin-section identification of the type of material found in the coprolites allows placement into three categories, each reflecting the dietary preference of the producing organism: (1) herbivore-produced coprolites contain inclusions of plant debris, partially replaced by pyrite; (2) coprolites produced by predators and/or scavengers contain inclusions of bone (Figure 7B), scales, and whole gastropods; and (3) coprolites produced by bottom feeders contain an abundance of silt (Figure 7C) and shell material (Figure 7D).

These coprolites also provide important paleoenvironmental clues as to the depositional conditions of the Michigan Formation. The lack of borings in the coprolites (Figure 7A and B), along with the presence of pyrite and glauconite (Figure 7D), suggest that the coprolites were deposited in an anoxic, reducing environment. Moreover, the coprolites were produced by nektonic vertebrates, but benthonic, marine invertebrates are not preserved. Together with the black shale, these observations suggest that the water column was most likely stratified, probably with an anoxic bottom. The lenses of sublitharenite sandstone in which the coprolites occur are discontinuous and sometimes rippled. These rippled, sorted sands with coprolites in preferred orientations suggest that occasionally moderate currents prevailed, perhaps due to storms or terrestrial flooding into this marginal marine environment, temporarily oxygenating the waters. The sequence of sand lenses, black shale, dolomite, and gypsum support a variety of possible shallow, marginal-marine environments such as restricted embayments, lagoons, or salinas.


During hand specimen examination of what appeared to be a coprolite, an apparent folded structure was observed. The broken specimen is 5.2 mm wide, and the enteroltthic structure within it is coiled (Figure 8A). These ropey, primary coils are folded upon themselves and comprise the entire specimen. Although coiled, the structures are not spiraled. Within the primary coils, smaller, secondary folds are visible (Figures 8B and C). The secondary folds lie completely within the walls of the primary coil (i.e., they do not crosscut the primary coil). The secondary folds are oriented perpendicular to the length of the primary coil. Near the center of this specimen, the secondary folds are exposed as three dimensional structures due to the dissolution of the secondary material (gypsum?) which surrounded them (Figure 8C). A well-preserved fish scale is contained entirely within one of the primary coils.


In a sandstone thin section, another apparent coprolite contained a similar internal enterolithic structure (Figure 9). This specimen, however, does not contain the enterolithic structure throughout the specimen, but instead the structure is located on one side. Additionally, the ropey structure is not coiled upon itself, but rather appears to fold in an accordion-like manner similar to the secondary folds observed in the hand specimen.



The internal, enterolithic structures of these two coprolite-like samples were not found in any other specimens. This indicates a difference in either preservation or origin. Preservational differences seem unlikely as these specimens were recovered from the same locations as the coprolites examined in this study. Furthermore, other specimens identified as coprolites preserve internal features, but nothing like the coiled structures seen in these two specimens. This suggests that the two specimens containing the coiled structures are not coprolites (fecal material), but rather preserved gut or intestinal contents. The origin and nomenclature for such structures is debated in the literature (Williams, 1972; Jain, 1983; McAllister, 1985; Clark, 1989; Hunt, 1992), but herein these structures are termed "cololites" following the nomenclature of McAllister (1985).

The secondary folds contained within the primary coils of the broken cololite (hand specimen) are interpreted to be mucosal folds within the whorls of the digestive tract. These features form as the fecal ribbon folds upon itself as it passes through the intestine (Figure 10; McAllister 1985) and around the mucosal lining of the wall (Williams 1972). According to Williams (1972), mucosal folds are not found in coprolites because they are destroyed upon passing through the rectum and sphincter muscle. Additionally, any coiling contained within fecal material would likely dissociate and uncoil after several hours in water (Jain 1983). Finally, the well-preserved fish scale within the broken specimen suggests that the scale did not pass through the complete digestive system, which would have resulted in intense deterioration of the scale.


Many cololites described in the literature are referred to as heteropolar spiral "coprolites" because they were initially believed to be true coprolites and they contain whorls that spiral in an outward fashion from the center of the structure (Williams 1972). Williams suggested calling these fossilized gut contents "enterospirae". He proposed that enterospirae are related to the spiral valves of modern sharks and rays, and, as such, are attributed to fossil taxa in this group. However, the two cololites recovered from the Michigan Formation are coiled, but not spiraled (Figure 8). This indicates that these cololites are not from taxa that contained a spiral valve, and are thus not attributed to sharks or rays.

An additional thin section contained a longitudinal section of an apparent coprolite that displayed an unusual internal structure. This specimen contained a cylindrical structure in the center that had several sharp triangular branches protruding from it. This structure might represent the longitudinal profile of a spiral valvular intestine, and suggests a shark or ray origin.


The question remains as to what aquatic vertebrates produced the coprolites and cololites. Unfortunately, identifying the specific taxa that produced them is not possible, but their association with various fossils from acanthodian, actinopterygian, sarcopterygian, and rhipidistian (crossopterygian) fish suggests that any of these are potential sources of the coprolites and nonspiraled cololites. Cladodont and orodont-type shark teeth and ctenacanth shark fin spines indicate that more than one species of sharks were present as potential producers of the spiral cololite.


In this study we conclude that the majority of the small, elongate to spherical specimens found in sandstone lenses at the base of the shale in unit 2 in the Michigan Formation are coprolites from aquatic vertebrates. This interpretation is based on the largely calcium phosphate composition of the groundmass; the presence of bone, plant, and silt inclusions; and the internal structure of the coprolites. The parallel orientation of the coprolites with the bedding plane also suggests that they are primary in origin.

The presence and preservation of the coprolites allow for interpretation of paleoenvironmental conditions during deposition of the Michigan Formation. The lack of borings in the outer walls of the coprolites, and the presence of pyrite and glauconite, indicate that the coprolites were deposited in an anoxic, reducing environment. Furthermore, the presence of coprolites indicates that vertebrates were in the area, suggesting that the water column was stratified with an anoxic bottom, a hypothesis that is supported by the paucity of benthic (body and trace) fossils in these deposits.

Additionally, the well-washed, sublitharenite sandstone, occasional ripple marks in the sandstone, and the preferred orientation of coprolites on the bedding plane, all point to periodic moderate-to-high energy conditions. Possibly flooding off of nearby land areas was responsible for terminating the evaporitic conditions that favored gypsum formation, and resulted in terrestrial sediment input, which produced the sand lenses overlying gypsum. Such flooding was followed by the establishment of quiet, anoxic conditions in which the overlying black shale was deposited.

The Michigan Formation in what is now western Michigan probably formed in a shallow, marginal-marine, restricted basin or basins (e.g., restricted embayments, lagoons, or salinas). Much of the gypsum is nodular (chicken-wire structure), but the vertically elongate nodules in some beds suggest a relict, initial fabric of crystal masses that grew vertically, typical of subaqueous deposition (Schreiber et al. 1982). The high gypsum to matrix ratios and the thickness of the gypsum also support this conclusion (Warren and Kendall 1985). Periodic influxes of flood water would have ended the hypersaline conditions that favored the formation of gypsum, and reduced salinities to more normal marine or even brackish conditions, favoring an influx of aquatic vertebrates into the shallow, marginal basin. As the basin shoaled with mud fill, restricted conditions were restored favoring first dolomite deposition, and then the precipitation of gypsum as hypersaline conditions once again prevailed.

Although the vast majority of studied specimens are coprolites, three specimens had preserved intestinal whorls, one of which also shows mucosal folds within the whorls. These specimens are interpreted to be cololites (preserved intestinal content). However, these cololites are not heteropolar spiral forms, as are most of those reported in the literature. The cololites described in this study are coiled rather than spiraled, and, with one exception, are not attributed to taxa belonging to sharks or rays. The exception has a similar appearance in thin section to the longitudinal profile of a spiral valvular intestine, which would be attributed to a shark or ray taxon. Mucosal folds described in the literature are almost exclusively recognized only in thin section, but in a hand specimen from the Michigan Formation mucosal folds can be seen using a dissecting microscope, suggesting unusually good preservation.

Other questions regarding the coprolites are under investigation. Mineralogical analyses of different colored and different shaped coprolites, as well as continued petrographic analyses, may lead to more well-defined relations to parent organisms, their diets, and diagenetic alteration of such materials.


We thank Ron Kragt, owner of the Michigan Natural Storage Company, for continually providing access to the gypsum mine. Steve Kenaga and Steve Sellepack collected much of the fish scale and teeth materials, and Steve Sellepack collected the rhipidistian jaw. We also thank Neil Clark for help in locating publications, Figen Mekik and Peter Wampler for drafting figures, and Brian Bodenbender and Reed Wicander for reviewing the manuscript for the Michigan Academician.


CATACOSINOS, P.A., W.B. HARRISON, R. F. REYNOLDS, D. B. WESTJOHN, AND M. S. WOLLENSACK. 2001. Stratigraphic Lexicon for Michigan. Lansing, Michigan: MDEQ Geological Survey Division.

CLARK, N.D. L. 1989. A study of Namurian crustacean-bearing shale from the Western Midland Valley of Scotland. Unpublished Ph.D. thesis, University of Glasgow, Glasgow, Scotland.

CROSS, A. T., AND R. E. TAGGART. 2004. Mississippian plants in gypsum mines, Grand Rapids, Michigan. Michigan Academician 36 (1): 57.

DORR, J. A., JR., AND D. F. ESCHMAN. 1970. Geology of Michigan. Ann Arbor, Michigan: University of Michigan Press.

DORR, J. A., JR., AND F. MOSER. 1964. Ctenacanth sharks from the mid-Mississippian of Michigan. Michigan. Michigan Academy of Science, Arts, and Letters Papers 49:10.5-113.

ELLs, G. D. 1979. The Mississippian and Pennsylvanian (Carboniferous) Systems in the United States - Michigan. In The Mississippian and Pennsylvanian (Carboniferous) Systems in the United States. U.S. Geological Survey Professional Paper 1110:J1-J17.

FOLK. R. L. 1974. Petrology of Sedimentary Rocks. Austin, TX: Hemphill Publishing Co.

HANTZSCHEL, W., F. EL-BAZ, AND G. C. AMSTUTZ. 1968. Coprolites, an Annotated Bibliography. Geological Society of America Memoir 108.

HARRELL, J. A., C. B. HATFIELD, AND G. R. GUNN. 1991. Mississippian System of the Michigan Basin; stratigraphy, sedimentology, and economic geology. In Early Sedimentary Evolution of the Michigan Basin, edited by P. A. Catacosinos and P. A. Daniels, Jr. Geological Society of America Special Paper 256:203-219.

HUNT, A. P. 1992. Late Pennsylvanian coprolites from Kinney Brick Quarry, central New Mexico, with notes on the classification and utility of coprolites. In Geology and Paleontology of the Kinney Brick Quarry, Late Pennsylvanian, Central New Mexico, edited by J. Zidek. New Mexico Bureau of Mines & Mineral Resources Bulletin 138:221-229.

JAIN, S. L. 1983. Spirally coiled coprolites from the Upper Triassic Maleri Formation, India. Palaeontology 26:813-829.

KENAGA, S. G., AND S. M. SELLEPACK. 1995. Fossils from the gypsum mines under Grand Rapids, Michigan. Michigan Academician 27:272.

MCALLISTER, J. A. 1985. Reevaluation of the formation of spiraled coprolites. University of Kansas Paleontological Contributions 114:1-12.

MICHIGAN DEPARTMENT OF ENVIRONMENTAL QUALITY. 1987. Bedrock geology of Michigan. Geological Survey Division, Lansing, Michigan.

MONNETT, V. B. 1948. Mississippian Marshall Formation of Michigan. American Association of Petroleum Geologists Bulletin 32:629-688.

MOSER, F. 1963. The Michigan Formation: A study in the use of a computer oriented system in stratigraphic analysis. Unpublished Ph.D. thesis, University of Michigan, Ann Arbor, Michigan.

SCHREIBER, B. C., M. S. ROTH, AND M. L. HELMAN. 1982. Recognition of primary facies characteristics of evaporites and the differentiation of these forms from diagenetic overprints. In Depositional and Diagenetic Spectra of Evaporites-a Core Workshop, edited by C. R. Handford, R.G. Loucks, and G. R. Davies, 1-32. SEPM Core Workshop No. 3, Calgary, Canada.

SHAVER, R. H. (Coord.). 1985. Midwestern Basin and Arches Region: Correlation of Stratigraphic Units of North America (COSUNA) Project. Tulsa, Oklahoma: American Association of Petroleum Geologists.

SOUTH, D. M., P. E. VIDETICH, AND V. OCHOCINSKI. 1995. Stratigraphy and depositional environments of the Mississippian Michigan Formation, western Michigan. Michigan Academician 27:277.

WARREN, J. K., AND C. G. ST. C. KENDALL. 1985. Comparison of sequences formed in marine sabkha (subaerial) and salina (subaqueous) settings-modern and ancient. American Association of Petroleum Geologists Bulletin 69:1013-1023.

WILLIAMS, M. E. 1972. The origin of "spiral coprolites." University of Kansas Paleontological Contributions 59:1-19.


Western Michigan University


Grand Vally State University



Grand Valley State University
COPYRIGHT 2008 Michigan Academy of Science Arts & Letters
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2008 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Regenmorter, John Van; Videtich, Patricia E.; Neal, William J.
Publication:Michigan Academician
Article Type:Report
Geographic Code:1U3MI
Date:Mar 22, 2008
Previous Article:Community design and community health: issues in building a holistic model in Michigan.
Next Article:Recent activity of the retrotransposable B2 element in hamsters and its use as a phylogenetic marker.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters