Printer Friendly

The Fort Polk Miocene terrestrial microvertebrate sites compared to those from east Texas.

Abstract. -- The small mammal faunas of four Miocene sites from western Louisiana are compared with two sites of similar age from east Texas. All sites are Barstovian and the three most abundant categories of small mammals are the geomyoid, cricetid and heteromyid rodents. Each site differs significantly from the others in the distribution of the most abundant small mammals. These relative differences are probably related to forest/grass cover and soil wetness. Taphonomic factors and mode of fossil recovery are also partially responsible for observed differences, and slight differences in age are considered likely. At the Fort Polk sites, squirrels, beavers, lagomorphs, bats, shrews and hedgehogs total less than 10% of the fauna.


Paleontological research on the Fort Polk Military Reservation (Fig. 1) since 1993 has resulted in the first Miocene terrestrial mammal fauna from Louisiana (Schiebout 1994; 1996). The fossiliferous beds, from the Castor Creek Member of the Fleming Formation, are fluvial overbank deposits including fossil soils (Jones et al. 1995). The best concentrations of small vertebrates (Fig. 2) are recovered from conglomerates formed by concentration of coarse material, including soil-formed nodules (Schiebout 1994). Remains of eight orders of land mammals, including insectivores, chiropterans, a lagomorph, a large carnivore, rodents, horses, a prosynthetocerine, and a gomphothere, have been recovered. Lower vertebrate fossils include fish, alligator, gavial, turtle, snake, and lizard remains. Before the discovery of the Fort Polk Miocene site, there was only a single report of a Miocene terrestrial vertebrate from Louisiana. This find, reported by Arata (1966), consisted of the tips of the lower tusks of a gomphothere (Mammalia: Proboscidea). Arata (1966) estimated that the specimen was from Miocene beds at Fort Polk, but precise stratigraphic and locality data were not available. It could have come from the Castor Creek Member.

The Fort Polk Miocene fauna is early late Barstovian in age, probably between 13 and 11.5 million years ago (Schiebout et al. 1996). At that time, the Fort Polk area was coastal, environmentally different from the long-studied Miocene fossil localities of the Great Plains, for example, Norden Bridge site, Valentine Formation in Nebraska (Voorhies 1990). The two Texas sites at which bulk screening was conducted are the sites most likely to be comparable in fauna. The two Texas sites, Trinity River and Town Bluff in east Texas (Fig. 1), are currently placed in the Burkeville Local Fauna and considered Barstovian (Prothero & Manning 1987).



Laboratory treatment of the Fort Polk Miocene conglomerates involves soaking chunks in approximately 10% acetic acid to partially dissolve and break up the rock, which releases a residue including the fossils. A laboratory area and procedures for bulk acid dissolution have been developed at LSU especially to process material from the Fort Polk Miocene. Calcium carbonate is the main mineral in the nodules and in the cement holding the rock together. It is dissolved in weak acetic acid, while teeth and bones are composed of apatite and are not dissolved. Bulk dissolution takes place in hazardous waste overbarrels. Productivity in preliminary results at one of the most productive Fort Polk Miocene sites (Stonehenge) is 0.2 mammal specimen(s) per pound and 429 mammal specimens per ton.


There are five localities in the Castor Creek Member of the Fleming Formation at Fort Polk where the fossil-bearing conglomerate is exposed. All yield fossils which are Barstovian in age. The exposures are usually in creek beds or gullies, although the site at which Miocene vertebrates were first discovered in 1993, named Discovery site (DISC), is a manmade outcrop. In most places, the exposed conglomerate is a near horizontal, tabular ledge, between 10-25 cm thick, and about 1 to 1.5 m wide, with some trough cross-bedding present (Jones et al. 1995). The conglomerate is usually overlain and underlain by a massive gray clay, often with common CaC[O.sub.3] nodules.

The most striking generalization about the Fort Polk Miocene fauna is that many of the animal remains recovered in screening, from catfish to rodents, are very small representatives of their taxa. Certainly there are biases produced by the deposition and the mode of recovery of the specimens. The most likely explanation would be that the Fort Polk small mammals are smaller than those at most other sites because of sorting in the original deposition and because of the use of fine screens in their recovery. The catfish spines, for example, are smaller than most material used for comparison, and it is difficult to conjecture a single ecological cause for small cricetid rodents and small catfish. Larger bone fragments have been recovered in the screens, however, indicating that the lack of larger rodents cannot be entirely explained by depositional sorting and processing biases. If a beaver the size of Eucastor tortus from Nebraska (Xu 1995) were present, the odds of its teeth being within the range of material concentrated in the fossiliferous conglomerates of Fort Polk are high, and it should have been recovered during this study. Likewise, although many of the Fort Polk Miocene small animals, such as the tiny heteromyids, would pass through the window screen used in some prior bulk screening at other Miocene sites, material such as the Fort Polk Miocene small beaver tooth should have been found. The possibility that much smaller kinds of mammals would be universally available if similar screening techniques were used at other sites is intriguing, but the more likely scenario is that there is also an environmental component to the explanation. The Castor Creek Gulf coastal environments may have been more finely divided into microhabitats than Great Plains wooded or savanna environments, and may have offered less overall food or perhaps more concentrated food sources available to small forms. Black (1963) commented that forest dwelling forms tend to be rare as fossils, and many of the Fort Polk animals are probably forest types.

Intersite variation at Fort Polk is approached herein solely by comparison of the faunas recovered by laboratory processing and screening. The lack of appreciable surface exposure of the claystones from which the Discovery site large fauna appear to be weathering, at the other Fort Polk Miocene microvertebrate sites (TVOR, Gully, Stonehenge), rules out their yielding comparable large material. For further description of the sites, see Jones et al. (1995). In addition to comparing the Fort Polk Miocene sites to each other (Fig. 3), they can be compared to the two Texas Coastal Plain sites from which good faunas of small animals are available (Dorsey 1977). The Texas sites (Town Bluff site and Trinity River site), shown in Figure 4, have the most abundant faunas from a Southern Methodist University program, which Dorsey (1977) reported as having screened 150 tons. The amount screened from Fort Polk for roughly similar amounts of rodents is about 1.5 tons.

Modern heteromyid rodents prefer open areas, and cricetid rodents prefer wooded areas (Dorsey 1977). This type of information may be used to infer the extent of forest cover of individual Fort Polk sites. At three Fort Polk sites, cricetids outnumber heteromyids (Fig. 3), a situation suggesting more forest cover and fewer open areas, and at the fourth site the situation is reversed. Trinity River site has more cricetids than heteromyids and Town Bluff has more heteromyids than cricetids.

The presumed ecology of geomyoids at both Fort Polk and the two Texas sites is less certain, because they may not be closely comparable to modern geomyoids. Dorsey (1977) referred the Texas geomyoids to Texomys and Jimomys. The Louisiana specimens are tentatively referred to Texomys. Korth (1994) called Texomys and Jimomys "problematical taxa". At DISC, TVOR, and Trinity River sites, geomyoids are first in abundance, cricetids second, and heteromyids third. At Gully site on Fort Polk, geomyoids are first in abundance, with heteromyids second, and cricetids third. At Town Bluff site in Texas, heteromyids are slightly more abundant than geomyoids, with cricetids third. Stonehenge is the only site of the six to have cricetids most abundant, with geomyoids ranking second, and heteromyids third.

All of these sites exhibit statistically significant differences in rodent distribution (Table 1). All probably included specimens from both wooded and open areas, probably averaging mixed environments, in addition to time averaging, the result of the mode of burial and recovery. The Texas sites were sandy (Jacobs, pers. comm.), probably deposited in fluvial channels, and much larger amounts of material were reported screened at the Texas sites. These differences suggest that the percentage differences between Texas and Fort Polk sites, which are roughly equivalent in rodent taxa, should not be overemphasized. Possibly, Stonehenge was the most heavily wooded of the Fort Polk sites, DISC, TVOR, and Trinity River had a more mixed environment, and Town Bluff was the most open, but still had a significant wooded component. It appears that the geomyoids thrived in a variety of coastal areas in the middle Miocene.

An animal present at Fort Polk, but absent at the Texas sites, despite the larger amounts screened, is a lagomorph. A single lagomorph tooth, identified by Dawson (pers. comm.) as cf. Hypolagus, has been recovered from the Stonehenge site. Slaughter (1981) has commented on the absence of Miocene rabbits in the Gulf Coast. It now appears that this lack does not include Louisiana.

In addition to analysis of the animals present in the Fort Polk Miocene sites and consideration of variation between sites, it is interesting to consider some of the animals that might be expected to be present at these sites, but which are absent. Of course, in the case of vertebrate fossils, absence of specimens can indicate that the animals were not preserved or have not yet been found. For large members of the fauna, the chance that more kinds remain to be found is good. Their remains are scarce and scattered in occurrence, exactly what might be expected from normal attrition without much effect of concentrating agents. Four genera of rhinoceroses are known from the Texas coastal plain, including two dwarf species (Prothero & Sereno 1982), so the appearance of an unmistakable rhinoceros specimen in the Fort Polk Miocene had been expected. Rare, small fragments of teeth with enamel thicker than that of the horse or prosynthetocerine, but smaller than that of the gomphothere, currently document some type of rhinoceros at Fort Polk. Also, among the missing are small ruminants and the many large rodents reported from other sites similar in age. In some cases, Fort Polk Miocene small fragmentary specimens indicate the presence of forms which cannot yet be identified on the available material. An example is LSUMG 3599, a small artiodactyl right premolar, which differs from AMNH 95489, Blastomeryx from the Trinity River site in Texas, in lacking a distinctive posterolabial heel. Further specimens are needed for more precise identification, but LSUMG 3599 indicates the presence of a Blastomeryx-sized small artiodactyl in the Fort Polk assemblage. Holman (1977) reported four salamanders, four anurans, and nine snakes from the Miocene of the Texas Gulf Coast. Some of these taxa may be present in the Louisiana Barstovian.


Differences between the Texas and Louisiana sites can be partly a result of age differences as well as differences in paleoenvironment, taphonomy, and collecting methods. The Texas sites fit into a long-studied sequence of Texas Miocene vertebrate local faunas (Wilson 1956; Fig. 5) as part of the Burkeville Local Fauna. The Louisiana sites show closest affinity to the overlying Cold Spring Local Fauna (Schiebout et al. 1996). The correlation of the Fort Polk fossils with the Cold Spring Local Fauna is made mainly on the presence of a similar prosynthetocerine, Prosynthetoceras francisi. The Synthetoceratinae have been recently revised by Patton & Taylor (1971). The Trinity River site yields P. trinitensis, considered by Patton & Taylor (1971) to be an incursion of Great Plains stock into the Gulf Coastal region and not closely related to the common Burkeville P. texanus, forerunner of P. francisi. If the correlation of the Louisiana sites with those at Cold Spring on the basis of P. francisi, a large animal, is correct, then there are no Texas micromammal sites of comparable age to those from Louisiana. The similarities between Texas and Louisiana collections of micromammals may or may not span the entire Burkeville and Cold Spring intervals. As additional large taxa are recovered from Fort Polk, the biostratigraphic importance of the prosynthetocerine may weigh less heavily, and the Louisiana sites be linked more closely to the Burkeville Local Fauna. The number of Fort Polk specimens and taxa being recovered is increasing steadily. These finds are expected to refine correlation and paleoecological comparison to Texas Miocene sites and others.
a. Heteromyideae 12.5%
b. Cricetidae 30.4%
c. Geomyoidea 50.6%
d. Castoridae 0.4%
e. Sciuridae 3.9%
f. Erinaceidae 0.2%
g. Soricidae 1.6%
h. Chiroptera 0.4%

Figure 2. Taxonomic distribution of 506 mammalian molar teeth recovered
from acid treatment and screening of the Fort Polk Miocene, as of
October, 1995. The single lagomorph tooth is a more recent discovery.

Note: Table made from pie chart.

Heteromyid (11%)
Cricetid (31%)
Geomyoid (58%)

Discovery Rodent Distribution (total 45)

Heteromyid (11.8%)
Cricetid (28.6%)
Geomyoid (59.6%)

TVOR Rodent Distribution (total: 332)

Heteromyid (30%)
Cricetid (20%)
Geomyoid (50%)

Gully Rodent Distribution (total: 10)

Heteromyid (20%)
Cricetid (51.2%)
Geomyoid (28.8%)

Stonehenge Rodent Distribution (total: 80)

Figure 3. Diagram of cricetid, heteromyid, and geomyoid rodents at
Discovery, TVOR, Gully, and Stonehenge sites at Fort Polk.

Note: Table made from pie chart.

Heteromyid (9.8%)
Cricetid (18.7%)
Geomyoid (71.5%)

Trinity River Rodent Distribution (total: 123)

Heteromyid (40.9%)
Cricetid (19.7%)
Geomyoid (39.4%)

Town Bluff Rodent Distribution (total: 188)

Figure 4. Diagram of cricetid, heteromyid, and geomyoid rodents in Texas
sites, Trinity River and Town Bluff, studied by Dorsey (1977).

Note: Table made from pie chart.

Table 1. Statistical comparison of rodent distributions (Figures 3 and
4) at Fort Polk and Texas sites. All Chi square values are significant
at the 0.001 level. Results were obtained by use of the StatView II


DISC 248.1 25.7 161.5 51.3 112.6
TVOR 1435.7 12206.3 423.6 317.5
Stonehenge 881.6 63.4 83.9
Gully 104.2 168.6
Trinity Rr. 62.8
Town Bluff


I wish to thank individuals at Fort Polk, especially Director of Public Works, Lieutenant Colonel Rory A. Salimbene, and his environmental staff, including Dr. Charles Stagg, James Grafton, Bob Hays and James Hennigan, for their help and cooperation which made this research possible. I appreciate being allowed to study specimens in the collections of the Shuler Museum of Paleontology, Southern Methodist University, and in the Frick Collection of AMNH. Louis Jacobs of Southern Methodist University, Margaret Stevens of Lamar University, and Mary Dawson of the Carnegie Museum of Natural History were among many colleagues providing helpful discussions. Timothy Dalbey, Louis Jacobs, Ruth Hubert, Brett Dooley and two anonymous reviewers provided helpful comments on the manuscript.

Work is currently supported by U. S. Army FORSCOM, on a contract administered and managed by Timothy S. Dalbey, Fort Worth District, U. S. Army Corps of Engineers, issued under Contract No. DACW64-94-D-0008, Delivery Order No. 006, entitled "Paleofaunal Field Survey, Collection, Processing, and Documentation at Two Locations on Fort Polk, Louisiana" to Prewitt and Associates, Inc. Parts of the discussion comparing Texas and Fort Polk Miocene sites were presented in a report of the same title to the U.S. Army Corps of Engineers, submitted for review in August, 1995. Support has also been provided by the LSU Museum of Natural Science.


Arata, A. A. 1966. A Tertiary proboscidian of Louisiana. Tulane Studies in Geology, 4:73-74.

Black, C. C. 1963. A review of the North American Tertiary Sciuridae. Bull. Mus. Comp. Zool. (Harvard), 130(3):109-248.

Dorsey, S. L. 1977. Biostratigraphy of the Mio-Pliocene of the Texas Gulf coastal plain. Unpub. Masters thesis, Southern Methodist Univ., Dallas, 65 pp.

Holman, J. A. 1977. Amphibians and reptiles from the Gulf Coast Miocene of Texas. Herpetologica, 33(4):391-403.

Jones, M. H., J. A. Schiebout & J. T. Kirkova. 1995. Cores from the Miocene Castor Creek Member of the Fleming Formation, Fort Polk, Louisiana: relationship to the outcropping Miocene terrestrial vertebrate fossil-bearing beds. Gulf Coast Assoc. Geol. Soc. Trans., 65:293-301.

Korth, W. W. 1994. The Tertiary record of rodents in North America. Plenum Press, New York, 319 pp.

Patton, T. H. & B. E. Taylor. 1971. The Synthetoceratinae (Mammalia, Tylopoda, Proceratidae). Amer. Mus. of Nat. Hist. Bull., 145:119-218.

Prothero, D. R., & E. M. Manning. 1987. Miocene rhinoceroses from the Texas Gulf Coastal Plain. Jour. Paleontology, 61(2):388-423.

Prothero, D. R., & P. C. Sereno. 1982. Allometry and paleoecology of medial Miocene dwarf rhinoceroses from the Texas Gulf coastal plain. Paleobiology, 8(1):16-30.

Schiebout, J. A. 1994. Fossil vertebrates from the Castor Creek Member, Fleming Formation, western Louisiana. Gulf Coast Assoc. Geol. Soc. Trans., 44:675-680.

Schiebout, J. A. 1996. A Miocene hedgehog (Mammalia: Erinaceidae) from Fort Polk in western Louisiana. Occas. Papers LSU Mus. Nat. Science, 70:1-9.

Schiebout, J. A., M. H. Jones, J. H. Wrenn & P. R. Aharon. 1996. Age of the Fort Polk Miocene terrestrial vertebrate sites. Gulf Coast Assoc. Geol. Soc. Trans., 46:373-378.

Slaughter, B. H. 1981. A new genus of geomyoid rodent from the Miocene of Texas and Panama. Jour. Vert. Paleo., 1(1):111-115.

Voorhies, M. R. 1990. Vertebrate paleontology of the proposed Norden Reservoir area, Brown, Cherry, and Keya Paha Counties, Nebraska. University of Nebraska Technical Report 82-09, prepared for the U. S. Bureau of Reclamation, 138 pp. plus 593 p. appendix.

Wilson, J. A. 1956. Miocene formations and vertebrate biostratigraphic units, Texas coastal plain. Amer. Assoc. Pet. Geol. Bull., 40:2233-2246.

Xu, Xiaofeng. 1995. Phylogeny of beavers (Family Castoridae): applications to faunal dynamics and biochronology since the Eocene: Unpublished Ph.D. dissertation, Southern Methodist Univ., Dallas, 287 pp.

Judith A. Schiebout

Louisiana State University, Museum of Natural Science Baton Rouge, Louisiana 70803

JAS at:
COPYRIGHT 1997 Texas Academy of Science
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1997 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Schiebout, Judith A.
Publication:The Texas Journal of Science
Geographic Code:1U7LA
Date:Feb 1, 1997
Previous Article:A late Pleistocene (Sangamonian) vertebrate fauna from eastern Texas.
Next Article:Assigning numbers to the nodes on a ring: the first step to improved performance of a ring network.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |