Reassessment of Characteristics Determining Generic Affinity in Gopherus AND Stylemys (Testudinidae) from the White River Group, Badlands National Park

William P. Wall1 and Dennis Maddox2

Department of Biology, Georgia College & State University, Milledgeville, GA 31061
2Department of Physiology, Medical College of Georgia, Augusta, GA 30901

Abstract—A review of the White River Group's (37.0 to 26.9 MYA) land tortoises is presented. New specimens are examined and described. The generic characteristics of Gopherus, Hesperotestudo, and Stylemys are reexamined. The species Stylemys amphithorax is declared nomen dubium. Based on nuchal morphology, the common White River tortoises should be placed in Gopherus. Stylemys could be retained as a fossil morphotype similar to modern gopher tortoises.


Tortoise shells, family Testudinidae, are among the most common fossils in the White River rocks of Late Eocene through Oligocene age (37_27 million years ago, see Prothero, 1994 for stratigraphic columns) exposed in Badlands National Park. The lowest unit of the White River group is the Chadron Formation. Deposition of this layer began during the Late Eocene, approximately 37 million years ago (Chadronian Land Mammal Age). Overlying the Chadron is the Brule formation, subdivided into the Scenic Member (Orellan Land Mammal Age) and the Poleslide Member (Whitneyan Land Mammal Age) which is overlain by the lower Sharps Formation (Arikareean Land Mammal Age).

The Eocene and Oligocene transition was a time of great change throughout the world, in both climate and species diversity (Clark, Beerbower, and Kietzke, 1967; Retallack, 1992). The general trend was a transformation from an early Eocene rain forest environment to a climate that was cooler and more like modern semiarid regions (Prothero, 1994).

Hutchison (1992) reviewed all herpetofaunal taxa in the late Eocene through early Oligocene of North America. After comparing the stratigraphic variance in the number of genera referable to terrestrial or aquatic habitats, he determined that the number of aquatic taxa was decreasing. Hutchison concluded that the declining diversity of herpetofaunal taxa was due in large part to the dwindling of aquatic environments.

Hay (1908) listed five genera of tortoises, Geochelone, Gopherus, Hesperotestudo, Stylemys, and Testudo, in the White River Group. He stated that the borders where the costals and peripherals meet are more rounded or more weakly sutured in Gopherus and in Stylemys than in Geochelone and Testudo. However, such qualitative characteristics are often difficult to interpret. The extent to which these sutures are more or less strongly interlocked is difficult to definitively measure, but traditionally when the costal/peripheral interface is rounded, Stylemys or Gopherus is indicated. However, suturing of the costal/peripheral interface is not indicative of generic affiliation. The most recent revision of these tortoises (Hutchison, 1996) recognized Gopherus, Stylemys, and
Hesperotestudo as valid taxa in the White River Group.

The question of what tortoise genera are actually present in the White River Group is the subject of this paper.

Materials And Methods

The Georgia College & State University vertebrate paleontology (GCVP) collection provided the majority of specimens studied. Additional specimens from the South Dakota School of Mines and Technology (SDSM) were examined. Modern tortoises from the Georgia College & State University herpetology (GCH) collection were used for comparative purposes. The terminology of Zangerl (1969) is followed for turtle morphology (Figure 1). All material is identified, described, and placed in the correct stratigraphic position.

All measurements were taken with Helios dial calipers to .05mm. The method used to express neural formulae is modified from Hutchison (1996). Neural shape is indicated by a number. For example, the number "6" indicates that the neural is hexagonal. Numbering begins with the first neural and ends with the eighth. Numbers are followed by a parenthesis. An open parenthesis following the number indicates that the neural is widest anteriorly. A closed parenthesis following a number indicates that the neural is widest posteriorly. A neural formula of 4(8)4(8)6(6(6(6( indicates that the first neural is quadrangular and widest anteriorly. The second neural is octagonal and widest posteriorly, etc.

The following characteristics were used in this study: shell characters include nuchal dimensions, neural differentiation, costal alternation, positioning of the inguinal scale in relation to the femoral scale, and the presence or absence of peripheral bridge pits; cranial characteristics involve details of the premaxillary ridge. Hay (1908), Auffenberg (1962), Bramble (1982), Hutchison (1996), and others use nuchal dimensions as the primary means to differentiate among tortoise genera.

Systematic Paleontology

Hutchison (1996) states that Gopherus has a nuchal that is "as wide or wider than long," Hesperotestudo has a nuchal that is "longer than wide," and Stylemys has a nuchal that is "longer than wide." We examined modern specimens of Gopherus aggassizzi and found that all specimens, indeed, had a nuchal that was wider than long. We also examined all fossil tortoises in the GCVP collection and found no reason to discontinue the use of this characteristic as a diagnostic factor.

Figure 1—Generalized testudine carapace (A, C) and plastron (B, D) illustrating bone and scale terminology respectively (from Zangerl, 1969).

Characteristics of the premaxillary ridge are of questionable value. All three genera of the Family Testudinidae known from the White River Group may have a median premaxillary ridge. Stylemys has historically been accepted as differing from Gopherus in having a distinctive lack of the juncture between the longitudinal premaxillary ridge and the masticatory ridges. In skulls of modern Gopherus, the point where the masticatory ridges join with the longitudinal premaxillary ridge is exceedingly faint. We do not think it is practical to try to distinguish fossil remains by a characteristic that is barely observable when present in modern forms. Bramble (1982) and Hutchison (1996) state that Gopherus may occasionally have the median premaxillary ridge not joining the medial maxillary triturating ridge. Thus, the characteristic is sporadic and useless in diagnosing specimens.

Hutchison (1996) stated that the median premaxillary ridge of Hesperotestudo is reduced or absent. Qualitative characteristics are confusing and overlap causes much guesswork. We believe that Gopherus and Stylemys have the median premaxillary ridge, and that this characteristic is variable in Hesperotestudo.

Neural shape is often used in determining the taxonomic placement of a tortoise specimen. Hutchison (1996) has devised a useful method of describing these neurals (see materials and methods), and he gives some neural formulae as follows: G. laticuneus 6)6)4(6(6(6(6(6 or 4(8)4(6(6(6(6(6, Gopherus sp. 4(8)4(8)4(6(6(6 or 4(8)4(6(6(6(6(6, Hesperotestudo brontops 4(8)4(8)4(6(6(6, Stylemys nebrascensis 4(6(6(6(6(6(6(6 or 4(8)4(6(6(6(6(6.

All GCVP specimens available have a neural formula of 4(6(6(6(6(6(6(6(. If the neural pattern were the only defining trait then all testudines in the GCVP collection belong in the genus Stylemys. Nuchal measurements, however, do not support this taxonomic assignment. The nuchal scales in all of these specimens have a width greater than their length, a characteristic considered to be diagnostic of Gopherus.

We examined modern Gopherus specimens in an attempt to resolve this conflict. Modern Gopherus specimens have relatively indistinct neural shapes with the neural formula being closest to 4(6)6(6(6(6(6(6(. The first neurals did not compare well with the one shown by Hutchison (1996). Additionally, the first costal in modern Gopherus contacts the second neural. This condition is also true for all White River testudines in which this character can be observed in the GCVP collection.

Auffenberg (1964) stated that neural differentiation should not be used as a diagnostic characteristic. It is evident, as attested to by the fact that Gopherus and Stylemys have two different neural formulae each, that neural formulae are highly variable. We agree that neural differentiation has little diagnostic value.

According to Hutchison (1996), Gopherus has "costals distinctly alternately narrower and wider laterally",
Hesperotestudo has "costals moderately to distinctly alternately narrower and wider laterally", and Stylemys has "costals only slightly narrower and wider laterally". We believe that qualitative characteristics lead to the potential for overlap, and thereby increasing the probability of an erroneous identification.

Based on specimens in the GCVP collection that we recognize on independent criteria (primarily nuchal shape, see discussion above) as a single species, Gopherus laticuneus, we found a wide range of variation in the degree of costal alternation. The costals obviously alternate in some cases, while they alternate very little in other specimens.

Bramble (1982) used the contact or lack of contact between the inguinal scale and the femoral scale to determine generic assignment of tortoise specimens. Hutchison (1996) stated that Gopherus has an inguinal scale, which is "moderate to small but contacts FEM [femoral scale]." He also wrote that Hesperotestudo has an inguinal scale, which is "enlarged anteromedially and may broadly contact FEM," and that Stylemys has an inguinal scale, which is "moderate to small, never contact[ing] FEM".

We examined modern Gopherus specimens and found that the shape of the inguinal scale is highly variable, but the contact between it and the femoral scale was always present. This character is usually not apparent in fossil tortoises. The entire inguinal and femoral scales are evident in only two of seventeen specimens examined. These two specimens (GCVP 36 and GCVP 37) appeared to be Gopherus laticuneus except for the fact that there was no contact between the inguinal scale and the femoral scale.

Based on examination of the fossil tortoises in the GCVP collection, we believe that the contact, or lack thereof, between the inguinal scale and the femoral scale may be variable in all tortoises from the White River Group.

Leidy (1851) reported the presence of peripheral bridge pits in Stylemys nebrascensis. For many years the presence or lack of peripheral bridge pits was used to distinguish between Stylemys and Gopherus. Auffenberg (1962) re-described Stylemys amphithorax, citing the absence of peripheral bridge pits as one of the major distinguishing factors between S. amphithorax and S. nebrascensis.

Leidy (1851) originally described Stylemys from a specimen that was about seven inches long and four-fifths complete. He reported that this specimen had a gap of one-eighth of an inch (approximately 3 mm.) between the peripheral and costal bones. Later, Hay (1908) retained Stylemys as a valid genus and described it as having ribs at the distal ends of the costal plates which enter pits in the bridge peripherals. We believe the gap between the bones described by Leidy (1851) and Hay (1908) is an age variable trait and therefore not a useful diagnostic tool. Juveniles of some modern turtles (for example, Geochelone radiata), exhibit a gap between the costal and peripheral bones that is absent in adult specimens. Through this gap, the ribs can be seen to insert into grooves along the peripherals. Our observations do not support Hay's (1908) belief that the insertion of the rib heads was along the bridges. All tortoise specimens in the GCVP collection which

wall and maddox—badl, Oligocene tortoises
Table 1—Comparison of characteristics distinguishing S. amphithorax and S. nebrascensis.
shell; GCVP 36, complete shell; GCVP 37, complete shell and skeletal fragments; GCVP 43, complete shell; GCVP 3305, complete shell; GCVP 3501, complete shell; GCVP 3841, partial shell; GCVP 4118, complete shell; GCVP 4452, complete shell; GCVP 4454, complete shell; GCVP 4456, complete shell.

Diagnosis.—Ernst and Barbour (1989) indicate that Gopherus has a nuchal scute that is about as broad as it is long and an inguinal buttress which joins to the sixth costal. According to Hutchison (1996), Gopherus differs from Stylemys in that the inguinal buttress contacts the femoral sulcus and the costals of Gopherus are alternately wide and narrow. Hutchison (1996) says that the first two neurals of Gopherus are widest posteriorly, and that the neurals of Gopherus may be alternately octagonal or quadragonal or they may all be hexagonal.

Age Distribution.—Known: Chadronian through Whitneyan. GCVP Collection: Chadronian through Whitneyan.

Modern Relatives.—There are three modern members of the genus Gopherus (gopher tortoises) found in North America. These are G. agassizii, G. berlandieri, and G. polyphemus (Ernst and Barbour, 1989). All seem to prefer areas with sandy soil and open scrub brush. All extant species of Gopherus are herbivorous, and their food preferences include grasses, leaves, flowers (especially blossoms if available), and fruits. G. agassizii and G. berlandieri, which both dwell in desert or near-desert environments, eat cacti.


Table 2 lists those characteristics we believe may have diagnostic value for members of the Testudinidae from the White River Group. Comparative measurements are often used to distinguish Gopherus from Stylemys. We have previously stated our opinion that some of these measurements are of little diagnostic value. Our taxonomic assignment of the common Badlands tortoises into the genus Gopherus is based primarily on nuchal dimensions because that characteristic appears to us to be the most reliable.

Our interpretation of White River testudines raises the question of the validity of the genus Stylemys as a separate taxon from Gopherus. Both generic names are firmly entrenched in the scientific literature. Which genus is better known depends on the researcher's field. Herpetologists are certainly more familiar with the modern gopher tortoises in the genus Gopherus. On the other hand, paleontologists are probably more familiar with the Badlands fossil tortoises which are typically placed in the genus Stylemys. Gopherus (Rafinesque, 1832) has priority over Stylemys (Leidy, 1851), therefore if the two genera are the same, the badlands tortoises should be placed in the modern genus. We have tentatively adopted that philosophy in this paper. We do, however, propose a possible solution to this taxonomic dilemma. Stylemys could be recognized as a "chrono" genus in a manner similar to that for chronospecies in the paleontological literature.

S. amphithorax S. nebrascensis

*Lacks peripheral bridge pits *Peripheral pits present.

for reception of rib heads.

*Length of posterior lobe of *Length of posterior lobe

plastron equal to or greater of plastron less than

than length of bridge. length of bridge.

#Thinner, more elongate shell. #Thicker, more rounded shell.

#Narrow and elongate cervical. #Broad and relatively short cervical.

#Distinct lateral notch on #Weak lateral notch on

Xiphiplastron. Xiphiplastron.

(*) indicates characteristics shown by Auffenberg (1964) to be juvenile or variable.

(#) indicates characteristics that we believe are individually variable.

exhibit the gap between the peripherals and the costals show a range of insertion points for the rib heads all along the peripherals, they are not confined to the bridges.

Auffenberg (1964) discussed the ontogenetic changes that occur when observing bridge grooves/pits in Gopherus. He demonstrated that the presence of peripheral bridge pits is entirely based on the age of the organism. We do not think bridge pits are a valuable diagnostic character. The weakness of this character places the validity of Stylemys amphithorax (Auffenberg, 1962) in doubt since it was distinguished primarily from S. nebrascensis because it lacked peripheral bridge pits. Other characteristics attributed to S. amphithorax are individually variable (see Table 1). This leaves S. amphithorax with no unique characteristics. We, therefore, suggest that it be synonymized with S. nebrascensis (Leidy, 1851), which has priority.

Based on our analysis, all of the Badlands tortoises in the GCVP collection belong in the genus Gopherus. We believe the following classification best describes our tortoises from the White River Group.

Order Testudines Linnaeus 1758

Suborder Cryppodira Cope 1868

Family Testudinidae Gray, 1825

Genus Gopherus Rafinesque, 1832

Gopherus laticuneus Cope, 1873

Testudo laticunea Cope, 1873

Testudo quadratus Cope, 1884

Testudo thompsoni Hay, 1908

Testudo praeextans Lambe, 1913

Gopherus laticunea (Cope, 1873), Williams, 1950

Geochelone thompsoni (Hay, 1908), Auffenberg, 1963 Scaptochelys laticunea (Cope, 1873), Bramble, 1982

Holotype.—American Museum of Natural History (AMNH 1160)

Selected Reference Specimens: GCVP 1, nearly perfect
Table 2—Potential diagnostic characteristics for the tortoise genera found in the White River Group.

Gopherus Hesperotestudo Stylemys

*Symphyseal groove *No symphyseal groove *Symphyseal groove

*Otolithic mass Nuchal *No otolithic mass *No otolithic mass

scale wider than long Nuchal scale longer than wide Nuchal scale longer than wide

*Phalangeal formula *Phalangeal formula 2-2-2-2-2 Phalangeal formula 2-2-2-2-2

2-2-2-2-1 or less

Premaxillary ridge present Premaxillary ridge may be present or absent Premaxillary ridge present

*Os transiliens present Os transiliens absent Os transiliens absent

*Caudal vertebrae with *Caudal vertebrae with inter- *Some caudal vertebrae lacking

interpostzygopophyseal notches postzygopophyseal notches interpostzygopophyseal notches

(*) Indicates characters not verified in this study and/or rarely observable in fossil tortoises (see Auffenberg, 1961).


We thank Dr. Philip Bjork for access to specimens in the South Dakota School of Mines collection. We also thank Ms. Rachel Benton of Badlands National Park for her extensive support of our research efforts. We both benefited greatly from discussions with Dr. Dennis Parmley on the anatomy and systematics of fossil and recent turtles. We thank three anonymous reviewers for their useful comments. Finally we thank Mr. Vince Santucci for his enthusiasm and support for paleontological research in the National Parks. This research was partially funded by faculty research grants from Georgia College & State University.


Auffenberg, W. 1961. A correction concerning the phalangeal formula of the turtle, Stylemys nebrascensis Leidy. Copeia 1961: 496-498.

———. 1962. Testudo amphithorax Cope referred to Stylemys. American Museum Novitates, Number 2120.

———. 1964. A redefinition of the fossil tortoise genus Stylemys Leidy. Journal of Paleontology. Vol. 38, Number 2.

Bramble, D. M. 1982. Scaptochelys: generic revision and evolution of gopher tortoises. Copeia 1974:102-109.

Clark, J., J. R. Beerbower, and K. Kietzke. 1967. Oligocene sedimentation, stratigraphy, paleoecology, and paleoclimatology in
the Big Badlands of South Dakota. Fieldiana: Geology Memoirs, 5:1-158.

Ernst, C. H. and R. W. Barbour. 1989. Turtles of the World. Smithsonian Institution Press: Washington, DC

Hay, Oliver Perry. 1908. The fossil turtles of North America. Carnegie Institution of Washington: Washington, D.C.

Hutchison, J. Howard. 1992. Western North American reptile and amphibian record across the Eocene/Oligocene boundary, and its climatic implications. Pp. In D. R. Prothero and W. A. Berggren (eds.), Eocene/Oligocene Climatic and Biotic Evolution. Princeton University Press: New York.

Hutchison, J. Howard. 1996. Testudines. Pp. In. D. R. Prothero and R. J. Emry (eds.). The Terrestrial Eocene-Oligocene transition in North America. Cambridge University Press.

Leidy, Joseph. 1851. Stylemys nebrascensis. Proceedings of the Academy of Natural Sciences of Philadelphia. Vol.5, pages 172-173.

Prothero, Donald R. 1994. The Eocene-Oligocene Transition: Paradise Lost. Columbia University Press: New York.

Retallack, Gregory, J. 1992. Paleosols and changes in climate and vegetation across the Eocene/Oligocene boundary. Pp. 382-398. In D. R. Prothero and W. A. Berggren (eds.), Eocene-Oligocene Climate and Biotic Evolution. Princeton University Press: New York.

Zangerl, R. 1969. The turtle shell. Biology of the Reptilia. Academic Press: New York.