carolina, u.s.a. oligocene pancheloniid sea turtles from...

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, researchlibraries, and research funders in the common goal of maximizing access to critical research.

Oligocene Pancheloniid Sea Turtles from the Vicinity of Charleston, SouthCarolina, U.S.A.Author(s): Robert E. Weems and Albert E. SandersSource: Journal of Vertebrate Paleontology, 34(1):80-99. 2014.Published By: The Society of Vertebrate PaleontologyURL: http://www.bioone.org/doi/full/10.1080/02724634.2013.792826

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, andenvironmental sciences. BioOne provides a sustainable online platform for over 170 journals and books publishedby nonprofit societies, associations, museums, institutions, and presses.

Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance ofBioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.

Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiriesor rights and permissions requests should be directed to the individual publisher as copyright holder.

http://www.bioone.org/doi/full/10.1080/02724634.2013.792826

http://www.bioone.org

http://www.bioone.org/page/terms_of_use

Journal of Vertebrate Paleontology 34(1):80–99, January 2014© 2014 by the Society of Vertebrate Paleontology

ARTICLE

OLIGOCENE PANCHELONIID SEA TURTLES FROM THE VICINITY OF CHARLESTON,SOUTH CAROLINA, U.S.A.

ROBERT E. WEEMS*,1,2 and ALBERT E. SANDERS3

1Paleo Quest, 14243 Murphy Terrace, Gainesville, Virginia 20155, U.S.A., [email protected];2Research Associate, The Charleston Museum, 360 Meeting Street, Charleston, South Carolina 29403, U.S.A.;

3P.O. Box 1318, Folly Island, South Carolina 29439, U.S.A., [email protected]

ABSTRACT—In the Oligocene Ashley and Chandler Bridge formations near Charleston, South Carolina, remains of threespecies of pancheloniid sea turtle are common. In their relative order of abundance, they are Carolinochelys wilsoni, Ashley-chelys palmeri, gen. et sp. nov., and Procolpochelys charlestonensis, sp. nov. Unlike the other two Oligocene South Carolinagenera, Procolpochelys persisted into the middle Miocene. An isolated skull, referable to the Miocene species Procolpochelysgrandaeva, is described here to supplement description of the poorly known skull of the Oligocene species P. charlestonensis.These turtles, all referable to Pancheloniidae, represent an endemic, moderately advanced sea turtle assemblage that occu-pied the Oligocene coastal waters of the southeastern United States. This tendency toward endemism also is seen in someparts of the co-occurring cetacean assemblage, which included at least two endemic families. Regionally restricted endemicassemblages appear to have been the norm among pre-Miocene pancheloniid sea turtles. It is not until the Miocene that crowncheloniids appear and begin to achieve very widespread to global distributions.

SUPPLEMENTAL DATA—Supplemental materials are available for this article for free at www.tandfonline.com/UJVP

INTRODUCTION

A great many papers have been published concerning theabundant fossil vertebrate remains recovered over the centuriesfrom the South Carolina Lowcountry, a region that includesCharleston, South Carolina (Fig. 1A, B). A detailed synopsisof this older literature can be found in Sanders and Anderson(1999). Among these papers, only three have direct bearing onthe abundant sea turtle remains recovered from the OligoceneAshley and Chandler Bridge formations (Fig. 1C). The first wasby Hay (1923a), in which he described, without illustrations,a new genus and species of pancheloniid (sensu Joyce et al.,2004) sea turtle, Carolinochelys wilsoni, from the ‘Eocene marls’of Charleston, South Carolina. Later that same year, after bet-ter preparation of the type skull, Hay (1923b) published anemended diagnosis of this genus and species along with sharpphotographs of the type specimen. In this paper, he also de-scribed and illustrated a humerus that he thought might pertainto the same species. Additionally, he noted that the preponder-ance of geologic thinking had moved toward the conclusion thatthe marls near Charleston were Oligocene rather than Eocenein age. Weems (1974) suppressed the name Carolinochelys andreferred C. wilsoni to the genus Procolpochelys. As discussed be-low under taxonomic descriptions, that designation was correctfor the referred humerus but incorrect for the type skull, and Car-olinochelys is reinstated here.

Although abundant new pancheloniid turtle material has beenfound in recent decades, nothing has been published on it un-til now. In addition to the far more complete material of Car-olinochelys described here, and the recognition of a new genushere named Ashleychelys, rare remains of yet a third kind oflarge Oligocene pancheloniid sea turtle have been found in re-cent years that are referable to a new species of the genus Pro-

*Corresponding author.

colpochelys Hay, 1908. Thus, we now can document and charac-terize three taxa of pancheloniid sea turtles that inhabited thecoastal waters of South Carolina in the mid-Oligocene. Thesethree taxa, along with three taxa of dermochelyid sea turtles yetto be described, indicate that a taxonomically diverse and numer-ically abundant population of sea turtles thrived in the coastalwaters of the southeastern United States during the middle partof Oligocene time (Fig. 1C).

Institutional Abbreviations—ChM, Charleston Museum,Charleston, South Carolina; MCZ, Museum of Comparative Zo-ology, Harvard University, Cambridge, Massachusetts; SCSM,South Carolina State Museum, Columbia, South Carolina;USNM, National Museum of Natural History, SmithsonianInstitution, Washington, D.C.

GEOLOGIC AND GEOGRAPHIC SETTING

Although Oligocene strata are well represented in the GulfCoastal Plain (Fig. 1A), they crop out in the Atlantic CoastalPlain only in a few limited areas (Weems et al., 2004). One ofthese areas is near Charleston, South Carolina, where the Ash-ley and the Chandler Bridge formations are frequently exposed(Fig. 1C). The ‘Ashley marl’ was first recognized and described byToumey (1848) and more fully described and discussed by Sloan(1908). The name was abandoned by Wilmarth (1938), reinstatedand formalized as the Ashley Member of the Cooper Formationby Ward et al. (1979), and then raised in rank to the Ashley For-mation of the Cooper Group by Weems and Lemon (1984). TheChandler Bridge Formation was recognized and named as a dis-tinct and separate stratigraphic unit by Sanders et al. (1982). Bothunits fall within the Trent Supergroup (Weems et al., 2004). Theirdetailed distribution in the Charleston region is summarized inWeems and Lewis (2002).

Most of the South Carolina Lowcountry that surroundsCharleston is blanketed by a veneer of Quaternary sediments.

80

WEEMS AND SANDERS—SOUTH CAROLINA OLIGOCENE PANCHELONIID TURTLES 81

FIGURE 1. A, map showing the regional set-ting of the southeastern United States. Thearea underlain by the eastern Gulf CoastalPlain and the central and southern AtlanticCoastal Plain is shown in light gray; the ap-proximate boundary between them is shownby dashed line. Cape Hatteras and the OuterBanks barrier islands lie just to the west ofthe letter ‘H.’ Small rectangle marks the areaaround Charleston, South Carolina, shown ingreater detail in B. Black circle is locationof Charleston; B, map of the region aroundCharleston showing the geographic areas dis-cussed in the text. Dark gray areas are un-derlain by Eocene strata and light gray ar-eas are underlain by Ashley and ChandlerBridge Oligocene strata (in most areas cov-ered by a thin veneer of Plio-Pleistocene sed-iments). Hachured line marks the southernlimit where Oligocene sediments can be foundnear the surface; to the south of this line theyare present but too deeply buried to be readilyaccessible. A is designated type section of theAshley Formation, CB is designated type sec-tion for Chandler Bridge Formation; C, geo-logic column showing the portion of Oligocenetime represented by the Ashley and ChandlerBridge formations.

Beneath it, occasionally in outcrops but more often in shallowexcavations, there are limited outcrops of older strata. Despitetheir very limited exposures, however, both the Ashley and Chan-dler Bridge formations contain such an abundance of vertebrateremains that a considerable assemblage has been collected andpreserved from the few exposures of both units that are available.This material has been collected from as far west of Charlestonas the Edisto River valley, from where that river swings southat the Great Bight of the Edisto southward almost to Jackson-

boro. Eastward, material has been collected through the head-waters of the Ashley River and then southeastward down its val-ley throughout the old Ashley River phosphate mining district.In the vicinity of Summerville, the productive outcrop belt startsto turn eastward again through Ladson and then turns sharplynorthward through Goose Creek toward Monks Corner, SouthCarolina (Fig. 1B).

The Ashley Formation is late Rupelian (Weems et al., 2006)and the Chandler Bridge Formation is early Chattian (Sanders

82 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 34, NO. 1, 2014

et al., 1982), but even so they are close in age. Both units fallwithin nannoplankton zone NP24 and so accumulated over aspan of no more than 2.4 million years, or about one-quarterof Oligocene time (Fig. 1C). Because the Oligocene is formallysplit into only two stages (Rupelian succeeded by Chattian), itis not acceptable to designate this time interval formally as mid-dle Oligocene, but the informal term mid-Oligocene is acceptableand used here to refer collectively to the age of these two forma-tions. The late Rupelian Ashley Formation represents a some-what offshore continental shelf depositional environment east ofany barrier islands, whereas the Chandler Bridge represents amuch shallower, lagoonal, and barrier-island-to-shallow-offshoredepositional setting (Sanders et al., 1982; Weems and Sanders,1986). The scarcity of land or freshwater animal remains in thenearshore Chandler Bridge Formation suggests that there was alarge lagoonal region to the west of the surviving areas of outcropthat effectively separated this coastal barrier environment fromthe mainland, perhaps in a setting similar to that of the modernCape Hatteras and Outer Banks areas in coastal North Carolina(location immediately west of ‘H’ in Fig. 1A).

TAXONOMIC NOTES

Some significant differences exist between the cetacean as-semblages found in the Chandler Bridge and Ashley formations(Weems et al., 2006). In contrast, the turtle assemblages of bothunits show no apparent differences or changes throughout thistime interval. It should be noted, however, that none of the recov-ered turtle material is known to represent the lower part of theAshley Formation; the turtle remains described here are proba-bly exclusively from the upper Ashley and Chandler Bridge for-mations. Because no differences are apparent between panche-loniid turtles from these two units, the specimens from both aredescribed together.

Hay (1923a) designated without illustration a nearly completeskull (MCZ 1005-A) as the type of the genus Carolinochelysand the species C. wilsoni. This is fortunate, because later in thesame year he described, illustrated, and referred a nearly com-plete humerus to this species (Hay, 1923b). For reasons discussedbelow in the section on Procolpochelys, the humerus pertainsto Procolpochelys and not Carolinochelys. Weems (1974) rec-ognized that this humerus pertained to Procolpochelys and soassigned Carolinochelys wilsoni to Procolpochelys as P. wilsoni,not realizing that the skull and humerus were incorrectly associ-ated. Because the humerus is not the type, Weems’s suppressionof the name Carolinochelys was in error and is corrected here.The holotype specimen of Carolinochelys wilsoni is, therefore,formally restricted to the skull (MCZ 1005-A), and the referredhumerus (MCZ 1005-B) is assigned to Procolpochelys charlesto-nensis, sp. nov., described below. Similar incorrect associations ofskulls, humeri, and shells have occurred all too often in the his-tory of fossil chelonian research (e.g., Osteopygis as corrected inParham, 2005).

SYSTEMATIC PALEONTOLOGY

Order TESTUDINES Linnaeus, 1758Suborder EUCRYPTODIRA Gaffney, 1975

Family CHELONIIDAE Gray, 1825CAROLINOCHELYS Hay, 1923a

CAROLINOCHELYS WILSONI Hay, 1923a(Figs. 2–8)

Carolinochelys wilsoni Hay, 1923a.Procolpochelys wilsoni (Hay, 1923a) (nov. comb.) (see Weems,

1974).

Holotype—MCZ 1005-A, a nearly complete skull.

Type Locality—The two specimens described and figured byHay (1923a, 1923b) were obtained from what then was known asthe ‘Eocene marl’ in the vicinity of Charleston, South Carolina.According to the original label with the type specimen and incor-rectly referred humerus, both specimens came from the commer-cial marl pits at Ingleside in Berkeley County, South Carolina, lo-cated 22.5 km (14 miles) northwest of Charleston (Weems, 1974).The marl exposed in this region northwest and west of Charlestonnow is known as the Ashley Formation of early Oligocene (lateRupelian) age, rather than Eocene age (Weems et al., 2006).Based on Hay’s (1923a, 1923b) description and discussion, thesespecimens very likely came from the upper part of the AshleyFormation. Much of the other material discussed here came fromthe Chandler Bridge Formation.

Referred Material—Ashley Formation: ChM PV8309. Chan-dler Bridge Formation: ChM GPV524, ChM PV2143, PV2144,PV2145, PV2148, PV2151, PV2167, PV2168, PV2171, PV2172,PV2178, PV2179, PV2184, PV2189, PV2190, PV2209, PV2227,PV2297, PV4267, PV4792, PV4367, PV4368, PV4820, PV5927,PV5928, PV6081, and PV7003.

Diagnosis—Large sea turtle with a deep skull and round or-bit; prefrontals form the anterior border of the orbit, whereasfrontals form only a small portion of the dorsal rim of the orbit;quadrate turns back only slightly beneath squamosal; cavum tym-pani not compressed and broadly rounded at its posterior end;secondary palate moderately developed, with the vomer con-tributing only about a third of its length to the secondary palate;maxillary palate furnished on both sides with two sharp ridges;pterygoid flanges only slightly developed; planar joint betweensixth and seventh cervical vertebrae; carapace strongly convexwith a length up to 70 cm, distinctly more elongate than wide,with the neural-costal portion of the carapace widest across thefourth costals; vertebral scutes as long as wide and hexagonal inyounger individuals but becoming longer than wide and almostrectangular in older individuals; costoperipheral fontanelles arelong and narrow but persistently present even in adults; uniquelyamong well known pancheloniids, deep notches are present alongthe outer border of the second peripherals; carapace surface cov-ered by a rugose vermiform pattern; angle between scapular rodsabout 130◦; shaft of humerus strongly curved downward, unlikein Ashleychelys palmeri and Procolpochelys charlestonensis.

Description—In its skull morphology, C. wilsoni is a verygeneralized stem-pancheloniid turtle. Its external surface isrough and covered by a ubiquitous network of shallow pits andvermiform troughs. The holotype skull (Fig. 2C, D), accuratelydescribed by Hay (1923b), clearly shows most of its suturalcontacts on its dorsal and ventral sides, as well as many ofthe sulcal boundaries on its dorsal surface. The holotype issomewhat crushed dorsoventrally, but a skull largely undistortedin its vertical dimension (ChM PV5927) gives a good idea of itsappearance in anterior and lateral views (Fig. 2A, B). The skullcan be restored accurately (Fig. 3). It is relatively high-arched fora pancheloniid turtle, with orbits that face obliquely outward andforward. In dorsal view, the parietals are very nearly pentagonalin shape, being exceptionally wide and sharply angled wherethey meet the suture between the postorbital and the squamosal.The dorsal rim of the orbit may be smoothly curved or it mayproject outward slightly above the eye socket. In ventral view,although the vomer provides a relatively small contribution tothe secondary palate, the vomerine pillar beneath it extendsfar posteriorly. The pterygoids are relatively narrow betweenwide foramina orbito-nasale vacuities and have only a hint ofpterygoid flanges on their external lateral borders. The basisphe-noid extends far posterolaterally in front of the basioccipital, sothat it fully separates the pterygoids from the basioccipital. Thesupraoccipital is relatively short and stout.

A number of partially articulated specimens of C. wilsonihave been recovered in recent decades that consistently associate


FIGURE 2. The skull of Carolinochelyswilsoni. A, anterior view of ChM PV5927; B,right lateral view of ChM PV5927; C, ven-tral view of holotype specimen, MCZ 1005-A; D, dorsal view of holotype specimen, MCZ1005-A.

C. wilsoni skulls with the shells and limb elements described here-after. The common and consistent association of skulls with theseother skeletal elements allows us to assign this material confi-dently to C. wilsoni. Among the carapace material that has beenrecovered, one specimen (PV4792) is especially complete and hasbeen restored by Bruce Erickson and the Science Museum ofMinnesota to as close to its original appearance as can be donegiven the small but pervasive compaction-induced distortions inthe carapace elements (Figs. 4, 5). In adults, the carapace haselongate but narrow costoperipheral fontanelles in its posteriorpart. A decidedly smaller specimen (PV2143) possesses an evenlonger costoperipheral fontanelle that is notably wider for its size,though still relatively narrow for a pancheloniid (Fig. 6B). In an-terior view, the carapace of C. wilsoni is more strongly archedthan in most pancheloniid sea turtles (Fig. 5B), being reminiscentof the modern species Caretta caretta.

The carapace of C. wilsoni is rather generalized with one no-table exception. Within each second costal, on both sides of thecarapace, there is a broad and wide indentation along its externalborder (Fig. 5B). These indentations are pronounced, quite sym-metrical, and fall in a position that would logically serve as hold-fasts for the enlarged, elongate, and curved claws found on thefront flippers of pancheloniid males, which they are known to usein modern species to grasp females during mating. The notchesseem to be natural features in the shell, but the notch on the sec-ond right peripheral element of PV4792 also possesses a scouredgroove that appears to have been abraded into the swale of thenotch (Fig. 5C). This supports the idea that the notches servedas holdfasts during mating. On other parts of this carapace, therealso are elongate (Fig. 5D) to round (Fig. 5E) pits of a very dif-

ferent nature that appear to have been caused by parasites orepizoans such as turtle barnacles.

Much of the plastron of C. wilsoni is known. Typical exam-ples of the hyoplastron, the hypoplastron, and the epiplastron,are shown in Figure 7A–C. These, like the carapace, are coveredby low but ubiquitously developed pits and vermiform troughs ontheir external (ventral) surfaces. At least one example of all plas-tral elements is known except for the entoplastron and xiphiplas-tra (Fig. 7D).

Girdle elements are poorly known, but a large part of a scapulaand most of an ischium are complete enough to be illustrated(Fig. 8D, E). The angle between the scapular rods is about 130◦.None of the preserved humeri are perfect, but they consistentlyshow an exceptionally down-curved shaft, unlike that of mod-ern cheloniid turtles (Fig. 8F). Modern cheloniids have lost theprimitive sigmoidal curvature in their humeral shafts that non-marine turtles normally retain (e.g., Echmatemys, sp. indet. [Hay,1908:fig. 369] and Geochelone orthopygia [Hay, 1908:fig. 594]),and this modification has helped pancheloniids use their fore-limbs as efficient paddles for flying locomotion in their marineenvironment (Hirayama, 1994). The presence of an intermedi-ate shaft curvature in C. wilsoni suggests that it had a humerusdistinctly less well adapted to an open-marine environment thanthe humeri of Ashleychelys palmeri (Fig. S2D in SupplementaryData), Procolpochelys charlestonensis (Fig. S4E in Supplemen-tary Data), and modern cheloniids. Similarly, a nearly completefemur of C. wilsoni, PV6081 (Fig. 8A–C), has trochanters that arenot connected by a bony ridge to the femoral caput as in mod-ern cheloniids, indicating that the rear limbs were not stronglyadapted to be rudders as they are in modern cheloniids.


FIGURE 3. Restoration of the skull of Car-olinochelys wilsoni in A, dorsal, B, ventral,and C, right lateral views. Abbreviations: bo,basioccipital; bs, basisphenoid; eo, exoccipital;FP, frontoparietal scute; fr, frontal; FR, frontalscute; ju, jugal; mx, maxilla; op, opisthotic;pa, parietal; pal, palatine; pf, prefrontal; PF,prefrontal scute; pm, premaxilla; po, postor-bital; pt, pterygoid; qj, quadratojugal; qu,quadrate; soc, supraoccipital; sq, squamosal;SU, supraocular scute; TE, temporal scutes1–3; v, vomer. Lower case labels are for bonesand upper case labels are for scutes.

A sixth cervical vertebra is preserved with PV4267, whichshows that C. wilsoni possessed a planar joint between the centraof the sixth and seventh cervicals (Fig. 8I, J).

ASHLEYCHELYS, gen. nov.ASHLEYCHELYS PALMERI, sp. nov.

(Figs. 9–15, S1–S3)

Holotype—ChM PV7002, a largely complete skeleton includ-ing the skull, lower jaw, much of the carapace and the plastron,girdle elements, humerus, and femur.

Type Locality—East bank of Limehouse Branch, 0.12 km (0.08miles) east of intersection of Ohio Way and Houston Drive(33◦01′30.2′′N, 80◦05′50.7′′W); Berkeley County, South Carolina.

Formation and Age—Upper part of Ashley Formation, up-per Rupelian, early Oligocene, within lower nannoplankton zoneNP24.

Etymology—Genus named for the formation in which thetype and paratype specimens were found; species named forBilly T. Palmer who found the specimen and donated it to TheCharleston Museum.

Paratypes—SCSM 90.19.1, skull, from ditch in Crowfield sub-division (Ashley Formation), Berkeley County, South Carolina;ChM PV6860, skull and greater portion of carapace, from MillCreek near Summerville (Ashley Formation), Berkeley County,South Carolina.

Referred Material—Ashley Formation: ChM GPV523, ChMPV6861, and PV7202. Chandler Bridge Formation: ChM PV2000,PV2140, PV2141, PV2142, PV2153, PV2174, PV2176, PV2181,PV2215, PV2220, PV2247, PV2259, PV2265, and PV7740.

Diagnosis—Moderately large sea turtle with a flattened andelongated skull and an anteroposteriorly elongated orbit; pre-frontals form the anterodorsal border of the orbit, whereasfrontals form only a small portion of the dorsal rim of the or-bit; prominent brow ridges formed by the postorbitals lie abovemost of the dorsal rim of the orbit; quadrate turns stronglybackward beneath squamosal; cavum tympani elongate, narrow,and tightly rounded at its posterior end; secondary palate mod-erately developed, with the vomer contributing about half ofits length to the secondary palate; maxillary palate slightly ru-gose without pronounced troughs or ridges; pterygoid flanges ro-bustly developed; planar joint between the sixth and seventh cer-vical vertebrae; carapace moderately convex with a maximumlength estimated to be about 55 cm, almost as wide as long, andwith the third costals most elongate; vertebral scutes hexago-nal and about twice as wide as long; elongated costoperipheralfontanelles are persistently wide into adulthood; neurals exhibita medial narrow low ridge unlike Carolinochelys wilsoni or Pro-colpochelys charlestonensis; carapace surface covered by an anas-tomosing pattern of grooves; hyoplastron and hypoplastron areangled sharply rather than gently upward away from the midline;


FIGURE 4. Carapace of Carolinochelyswilsoni, ChM PV4792, in dorsal view.

xiphiplastra much narrower and longer than in P. charlestonensis;angle between scapular rods about 130◦; shaft of humerus verystraight.

Description—The holotype skull (Fig. 9) and paratype skull(Fig. 10) are from relatively young individuals that clearly showmost of the suture lines and sulcal grooves. Between them, theypermit a fairly detailed reconstruction of the skull in dorsal, lat-eral, and posterior views (Fig. 11). Like C. wilsoni, A. palmerihas a rather generalized stem-pancheloniid skull, with an exter-nal surface covered by a ubiquitous pattern of shallow pits andvermiform troughs. A distinctive feature is the pronounced browridges (Fig. 11A, D), formed by the postorbitals, which overhangthe orbits more strongly than in C. wilsoni or P. charlestonen-sis. These two skulls do not offer a good view of the palate, butthis portion of the skull can be seen on the paratype adult skullSCSM 90.19.1 (Figs. 12, 13). In ventral view, the secondary palateis moderately well developed, rather rough-surfaced, but with-out any prominent maxillary ridges or troughs. The basisphenoiddoes not separate the basioccipital from the pterygoids as it doesin C. wilsoni. This adult skull is nearly identical in morphologyto the ones in Figures 9 and 10, but about twice as large as theholotype. This indicates that the type was only about half grownat death.

The lower jaw is moderately expanded in its symphyseal regionand has elongate jaw rami (Fig. 9C).

Portions of the fourth (ChM PV2142), sixth (PV6860), and sev-enth (PV2142) cervical vertebrae can be recognized (Fig. S1A–Ein Supplementary Data). The fourth cervical is biconvex, the sixthhas a concave anterior face and a planar posterior face, and theseventh has a planar anterior face and a slightly convex posteriorface. These vertebrae are typical of moderately advanced panch-eloniid turtles.

An imperfect scapula is present in specimen ChM PV2141 (Fig.S1L), and is complete enough to indicate that it had an angle ofabout 130◦ between the scapular rods. The humerus is very sim-ilar to that of modern cheloniids (Fig. S2), but differs strikinglyfrom the humerus of C. wilsoni and P. charlestonensis in that itsshaft is very nearly straight. In this regard, A. palmeri was morespecialized for cheloniid-style aquatic-flying locomotion than ei-ther of the other two species. The femur associated with the holo-type (Fig. S1F–J) has a bony ridge connecting the trochantermajor with the femoral head, which represents a partial devel-opment of the bony ridges found between the femur head andthe trochanters in modern crown cheloniids to improve their rearlimbs as steering organs.

The carapace of the holotype of A. palmeri is missing its cen-tral region (Fig. 14A), but that part of the carapace is preservedin paratype ChM PV6860 (Fig. 14B). The features of these twospecimens can be combined into a composite representation ofthe entire carapace (Fig. 15A). It is moderately convex, attains an


FIGURE 5. Carapace of Carolinochelyswilsoni, ChM PV4792. A, left lateral view;B, anterior view with second costal notchesmarked by arrows; C, groove (trending nearlyleft and right, ends marked by arrows) worninto carapace notch on right second peripheralelement; D, elongate scars on third and fourthleft costals; E, circular scar possibly causedby a sea turtle barnacle. Note pit and troughsculpture on surface of carapace.

estimated length up to 55 cm (based on the size of the adult skullrelative to the holotype skull and its associated carapace), and isslightly more elongate than wide with the neural-costal portion ofthe carapace widest across the third costals. The most notable fea-tures of this carapace are the wide and elongate costoperipheralfontanelles and the very wide vertebral scutes, which are abouttwice as wide as long.

All but the anterior portion of the plastron is preserved inthe holotype (Fig. S3A). These elements can be supplementedby several other specimens (including paratype ChM PV6860and PV9404; Fig. S3B–F), and the epiplastra are preserved withPV2141 (Fig. S1K). The proportions of the entoplastron can bereasonably estimated from the other elements, and a reconstruc-tion of the entire plastron is shown in Figure 15B. The most no-table features of this plastron are the remarkably elongate andthin xiphiplastra and the presence of a distinctly angular, antero-posteriorly trending bend in both the hyoplastra and hypoplastra,giving the plastron an angular rather than smooth shape in trans-verse cross-section.

PROCOLPOCHELYS Hay, 1908

Diagnosis—Large turtle with a deep skull, dorsal and lateralskull surfaces faintly ornamented by ridges, grooves, or pits in itsanterior region; orbit round with prefrontals forming the anteriorborder; frontals form only a small portion of the dorsal orbit bor-der, but medially they project strongly forward along the midlinebetween the prefrontals; supraoccipital process elongate; ventralsurface of skull apparently has a moderately well developed sec-ondary palate with longitudinal ridges on the palatal surface ofthe maxillae; planar joint between the sixth and seventh cervicals;carapace moderately convex, up to 95 cm in length, markedly

longer than wide with the neural-costal portion of the carapacewidest across the second costals; costoperipheral fontanelles wideand persistent throughout life; dorsal surface smooth and un-sculptured or only very faintly sculptured; vertebral scutes hexag-onal and about as wide as long; neurals in the central carapace re-gion usually split transversely into anterior and posterior piecesof unequal size, the smaller piece bearing an internal scar for at-tachment to the tip of an underlying neural spine; xiphiplastrashorter and much wider than in Ashleychelys palmeri; angle be-tween the scapular rods is about 120◦; humerus large and stoutwith a rather short but robust lateral tubercle; femur has a bonyridge connecting the caput with the trochanter major.

PROCOLPOCHELYS CHARLESTONENSIS, sp. nov.(Figs. 16, 17, S4–S6)

Holotype—ChM PV6056, a largely complete carapace associ-ated with a few fragmentary vertebrae, pectoral girdle elements,humerus, and femur.

Type Locality and Horizon—Ditch bank in Turnberry Placesubdivision, 8.04 km (5 miles) north of S.C. Route 7 (SamRittenberg Boulevard) via S.C. Route 61 (32◦51′21.14′′N,80◦04′21.76′′W), Charleston County, South Carolina. ChandlerBridge Formation, lower Chattian, late Oligocene, upper nanno-plankton zone NP24.

Referred Material—The type specimen is supplemented by askull fragment (Fig. S5A, B), consisting of much of the parietal,the frontal, and the prefrontal (ChM PV6055), found in theChandler Bridge Formation near the Mark Clark Expresswayin Charleston County, South Carolina. Although not associatedwith other skeletal elements, the large size of this specimenand its similarities with the skull of Procolpochelys grandaeva


FIGURE 6. Restored carapaces of Carolinochelys wilsoni. A, carapace based on ChM PV4792; B, an immature partial carapace, ChM PV2143(preserved elements shown in gray), provides additional information about the location of the sulci, the existence of variability in the suprapygalregion, and the degree to which the costoperipheral fontanelles shrank during growth of this turtle.

(described below) make its allocation to P. charlestonensisreasonably certain. The humerus (Fig. S4C) ascribed by Hay(1923b) to Carolinochelys wilsoni (MCZ 1005-B) probably camefrom the Ashley Formation; it is nearly identical in appear-ance to the type specimen humerus of P. charlestonensis (Fig.S4D–F) and even slightly larger. Therefore, based both on itsmorphology and its great size, MCZ 1005-B is assigned here toProcolpochelys charlestonensis.

Diagnosis—So far as known, most elements are similar toP. grandaeva. Characteristics that distinguish P. charlestonensisfrom P. grandaeva are the relatively narrower costals and rel-atively wider vertebral scutes in the new species. Both speciesof Procolpochelys have supernumerary neurals and exception-ally thick costals and neurals that are unsculptured or only faintlysculptured on their dorsal surface. In these traits, both speciesdiffer from Carolinochelys wilsoni and Ashleychelys palmeri.

Description—The skull fragment referred to this species (Fig.S5A, B) consists of much of the left parietal, frontal, and pre-frontal. The parietal is essentially complete along its medial edgeand the medial portion of its posterior edge. The frontal is es-sentially complete and the posterior portion of the prefrontal ispreserved. On its ventral side, the anterior portion of the pari-etal preserves the dorsal surface morphology of the brain cavity.This obviously was a very large turtle, with at least the posteriorportion of its dorsal skull surface nearly smooth and with only afaint development of the vermiform grooves and pits that are soprominent on the skulls of both Ashleychelys palmeri and Car-olinochelys wilsoni. A nearly smooth external surface similarly is

found in the parietal region of the Miocene species P. grandaeva(see below), and this supports association of this skull fragmentwith the genus Procolpochelys.

Several vertebrae are preserved with the type. One is the sev-enth cervical (Fig. S5L, M), which has a ventral keel, a posteriorlyconvex centrum, and a planar anterior centrum. Another is a firstsacral (Fig. S5G–I), and the third is an anterior caudal, probablythe first (Fig. 16).

Most of one side or the other of the type carapace is preserved(Fig. 16). The costals and neurals are unusually thick, and this isalso true of P. grandaeva. There is no sign of sculpturing on theexternal surface of the carapace bones as is seen in C. wilsoniand A. palmeri. The sulcal grooves are clearly visible, so the bonesurfaces are not deeply worn or eroded. The carapace elementsin P. grandaeva also are nearly smooth (Zangerl and Turnbull,1955; Weems, 1974), so unsculptured costals and neurals are areal and diagnostically useful characteristic of this turtle. As isalso typical for P. grandaeva, many of the neural elements aresplit transversely into anterior and posterior parts, the smaller ofwhich bears an internal scar where the tip of a vertebral spineattached beneath it. The other element lay in the space betweenvertebral spines. Many of the costals, on their medial ends, havea scalloped indentation into which fit the neural elements thatattached to the vertebral spines. Although not present on all ofthe costals, where this scalloped pattern occurs it is a very char-acteristic feature of both P. grandaeva and P. charlestonensis. InP. charlestonensis, the costals are stout, being less elongate lat-erally and longer anteroposteriorly than the equivalent elements


FIGURE 7. Plastron of Carolinochelys wilsoni. A, right hyoplastron of ChM PV4792 in ventral (external) view; B, left hypoplastron of ChM PV4792in ventral (external) view; C, right epiplastron of ChM PV4367 in ventral (external) view; D, restoration of plastron in ventral (external) view basedon ChM PV4792 and ChM PV4367. Known elements shown in gray.

in P. grandaeva (Fig. 17). Another important distinction betweenthese two species is that the vertebral scutes in P. charlestonensisare relatively wider than they are in P. grandaeva (Fig. 17).

The hyoplastra and hypoplastra are not preserved with thetype specimen, but a left epiplastron and right xiphiplastron arepresent (Fig. S6). The most notable feature in these elements isthe exceptional width of the xiphiplastron.

The pectoral girdle elements (scapula and coracoid) are bothpreserved in the type (Fig. S4A, B). The angle between the scapu-lar rods is about 120◦. A single ilium (Fig. S5F) was the only el-ement of the pelvic girdle recovered. The humerus is large andstout, with a rather short but robust lateral tubercle (Fig. S4D–F);in lateral view, its shaft has a slight but distinct downward curva-ture that is less pronounced than in Carolinochelys wilsoni butmore pronounced than in Ashleychelys palmeri. The femur (Fig.S5C–E) is somewhat damaged but shows typical advanced pan-cheloniid proportions and has a ridge connecting the caput withthe damaged trochanter major, suggesting that the rear limbswere partially adapted to be steering organs.

PROCOLPOCHELYS GRANDAEVA (Leidy, 1851)(Figs. 18, 19)

Comments—The carapace and the plastron of the Miocene seaturtle P. grandaeva were first adequately described by Zangerland Turnbull (1955). Additional material referable to this taxon,including humerus, femur, and part of a basicranium, later wasdescribed by Weems (1974). Since that time, a much more com-plete skull of P. grandaeva (USNM 358862) (Fig. 18) was foundby R.E.W. in the lower Miocene (mid-Burdigalian) Popes CreekMember of the Calvert Formation in Charles County, south-ern Maryland, in the bluffs along the Potomac River south ofPopes Creek. At this long-collected locality, the only two kindsof pancheloniid carapace elements that have been found arethose of Trachyaspis lardyi (= Syllomus crispatus Cope) and P.grandaeva. The skull of T. lardyi is known (Weems, 1980), and it

is smaller and quite different from the more generalized panch-eloniid skull described here. The additional information for P.grandaeva provided by this skull (USNM 358862) offers impor-tant supplemental information for defining and differentiatingthe genus Procolpochelys.

Description—Much of the morphology of the skull of P.grandaeva can be reconstructed from this specimen (Fig. 19).Its parietal, frontal, and prefrontal are quite comparable to theportions of those elements preserved in the only known skullfragment of P. charlestonensis. In dorsal view, the snout of P.grandaeva is short and rather pointed. In ventral view, the foram-ina orbito-nasale are relatively large and the pterygoids are cor-relatively quite narrow. The skull of P. grandaeva is here recon-structed with moderate pterygoid flanges (Fig. 19B), but this isonly inferential because that region of the skull was not preservedon either side of the midline. Similarly, although the squamosalwas not recovered, its bordering suture on the parietal indi-cates that its contribution to the roof of the skull was relativelysmall.

EVOLUTIONARY GRADE OF THE SOUTH CAROLINAOLIGOCENE PANCHELONIID TURTLES

The morphological characteristics of Oligocene pancheloniidturtles from South Carolina are compared and contrasted inTable 1. Only a few of these character traits are known to beuseful for determining the evolutionary grade of these turtles(Hirayama, 1994; Parham and Pyenson, 2010), and these wereused to create the tree in Figure 20. Based on the presenceor absence of these particular character states, Carolinochelyswilsoni is the most primitive of the three pancheloniid turtlesfound in the Ashley and Chandler Bridge formations. All threespecies possess a planar joint between the sixth and seventh cer-vicals and all have a scapular rod angle greater than 110◦. Bothof these characteristics are derived features not found amongprimitive pancheloniid turtles (Hirayama, 1994). Two other


FIGURE 8. Girdle and limb elements ofCarolinochelys wilsoni. A–C, femur of ChMPV6081 in A, proximal, B, posterior, and C,anterior views; D, partial scapula of ChMPV8309; E, ilium of ChM PV8309; F, G, righthumerus in F, anterior and G, ventral views(lateral tubercle broken away); H–J, sixth cer-vical vertebra in H, anterior, I, ventral, and J,posterior views.

derived character traits, typical of more advanced pancheloniidturtles, are found in Ashleychelys palmeri and Procolpochelyscharlestonensis, but not in C. wilsoni. These are the presence ofan attachment between the eighth costal rib and the eleventhperipheral and the presence of a ridge connecting the trochantermajor to the femoral caput. None of these turtles, however,possesses the full complement of character traits that define thecrown Cheloniidae (Hirayama, 1994), which include Trachyaspislardyi plus the six living species Natator depressus, Lepidochelyskempii, L. olivacea, Eretmochelys imbricata, Caretta caretta, andChelonia mydas and any fossil species nested within this group(e.g., Caretta patriciae Zug, 2001). By these definitions, C. wilsonibelongs among moderately advanced members of the stemgroup of Pancheloniidae, but the primitively arched shaft of thehumerus and the absence of the other two advanced character-istics found in P. charlestonensis and A. palmeri place it near thebottom of this grade. Although the notches in the second periph-eral elements of C. wilsoni are an autapomorphy that precludes itfrom being an ancestor to any later or living cheloniid sea turtles,its general appearance nevertheless suggests a convergence inlifestyle with the living loggerhead turtle Caretta caretta.

Procolpochelys charlestonensis is a more advanced panche-loniid than Carolinochelys wilsoni but less advanced than Ash-leychelys palmeri (Fig. 20). Zangerl and Turnbull (1955) con-sidered P. grandaeva to be a primitive carettine turtle, but thatspecies and P. charlestonensis both have remarkably thick neu-rals and costals and attained exceptionally large size. Neithercharacteristic would be expected in an ancestor of modern caret-tine turtles. Moreover, only a few of the characteristics found incrown cheloniids are present in this genus, even though someof these same character states are present in the contemporaryA. palmeri. Therefore, the carettine-like features seen in Pro-colpochelys seem to have evolved independently of the moderncarettine lineage.

Ashleychelys palmeri is the most advanced of the three panch-eloniid turtles found in the Ashley and Chandler Bridge forma-tions. This is demonstrated by its very straight humeral shaftand the presence of a ridge connecting the femoral trochanters,both characteristics reflecting better adaptation to marine lo-comotion than was present in Carolinochelys wilsoni and Pro-colpochelys charlestonensis. An important advanced character-istic not found in A. palmeri is a ‘V’-shaped lateral tubercle on


FIGURE 9. Ashleychelys palmeri, gen. et sp.nov. A–D, holotype skull, ChM PV7002, in A,right lateral, B, anterior, C, ventral, and D,dorsal views. Lower jaws are visible in artic-ulated position in C.

the humerus. The absence of this characteristic indicates that A.palmeri was approaching the level of locomotor sophisticationfound among crown cheloniids but was not quite there. A numberof autapomorphies debar A. palmeri from being ancestral to thecrown cheloniids, including its flattened skull profile, flattenedorbit, rearward tilted cavum tympani, and very wide vertebralscutes.

A SYNOPSIS OF OLIGOCENE PANCHELONIIDSEA TURTLES

Only two early Oligocene (early Rupelian) sea turtles areknown (Fig. 21). These are Glarichelys knorri from Germany(Karl, 2007) and possibly Poland (Mlynarski, 1959; but this ma-terial may be late Eocene) and Allopleuron lipsiensis (Karl,2007) from Germany. ‘Chelonia’ caucasica, originally reported as

FIGURE 10. Ashleychelys palmeri, gen. etsp. nov. A–D, partial posterior skull ofparatype, ChM PV6860, in A, posterior, B,right lateral, C, dorsal, and D, ventral views.


FIGURE 11. Ashleychelys palmeri, gen. etsp. nov. A, restoration of skull of holotype(ChM PV7002) in dorsal view. Position ofpostorbital brow ridge shown by arrow; B, C,restoration of skull of paratype, ChM PV6860,in B, dorsal and C, posterior views; D, com-posite restoration of skull in dorsal view basedon holotype skull, ChM PV7002, and paratypeskull, ChM PV6860. Position of postorbitalbrow ridge shown by arrow; E, restorationof skull of holotype, ChM PV7002, in rightlateral view. Abbreviations: bo, basioccipital;bs, basisphenoid; eo, exoccipital; FP, fron-toparietal scute; fr, frontal; FR, frontal scute;IP, interparietal scute; ju, jugal; mx, maxilla;op, opisthotic; pa, parietal; PA, parietal scute;pf, prefrontal; PF, prefrontal scute; pm, pre-maxilla; po, postorbital; qj, quadratojugal; qu,quadrate; soc, supraoccipital; sq, squamosal;SQ, squamosal scute; SU, supraocular scute;TE, temporal scutes 1–3. Lower case labels arefor bones and upper case labels are for scutes.

coming from the early Oligocene of Russia (Riabinin, 1929; Lap-parent de Broin, 2001), instead comes from the early Miocene(Chkhikvadze, 1983, 1989) and currently is considered to be anomen dubium (Averianov, 2002).

The record of mid-Oligocene (late Rupelian–early Chattian)sea turtles is more diverse and widespread. It includes the threespecies described here from South Carolina plus six nominalspecies from Europe and one species from western Asia. A num-ber of other species have been named, but most or all are indeter-minate, invalid, or junior synonyms of other species (Lapparentde Broin, 2001).

In Belgium, four nominal species have been described from themid-Oligocene: Chelone (Bryochelys) waterkeynii, Chelone van-benedenii, Chelyopsis littoreus, and Oligochelone rupeliensis. Ch.vanbenedenii, Ch. waterkeynii, and Cl. littoreus were described bySmets (1886, 1887a, and 1887b, respectively), who soon after pub-lished another paper concluding that Cl. littoreus was a chimerawith vertebral and girdle elements coming from Psephophorusrupeliensis and other elements coming from Chelone vanbene-denii (Smets, 1888:194). As the type species of Chelyopsis is Cl.littoreus, Chelyopsis became a nomen nudum. Similarly, Dollo(1909) announced a new genus and species, Oligochelone ru-peliensis, about which he was planning to publish in the then nearfuture. This apparently never happened, however, and Lapparent

de Broin (2001:183) has concluded that this taxon also is a nomennudum. Thus, only two valid species of Oligocene sea turtlesare presently recognized from Belgium. Neither of these speciescan be validly assigned to Chelonia or to the defunct genus‘Chelone,’ so they are here designated Bryochelys waterkeyniiand ‘Chelone’ vanbenedenii. Although both species are based onvery fragmentary material, they appear to be valid. Character-ization of these two species has been complicated because Smets(1888:208) concluded that the snout that he originally referred to‘Ch.’ vanbenedenii actually belonged to B. waterkeynii. He dis-tinguished between the carapaces of these two species by notingthat Bryochelys waterkeynii has “expansions foliacees,” whereas‘Chelone’ vanbenedenii lacks them (Smets, 1886:123–124). Thisseemingly means that B. waterkeynii has a sculptured or reticu-late pattern on its carapace, whereas ‘Ch.’ vanbenedenii does not.

Five nominal species have been described from the mid-Oligocene of Germany and Switzerland: Glarichelys knorri(Gray, 1831) (see Zangerl, 1958), Chelyopsis holsaticus Dames,1894, Chelonia gwinneri Wegner, 1917, Chelyopsis halleriKarl, 1989, and Rupelchelys breitkreutzi Karl and Tichy, 1999.Recently, Karl (2007:29, 38) synonymized Cl. holsaticus, Ch.gwinneri, and Cl. halleri with G. knorri, leaving only G. knorriand R. breitkreutzi as recognized species from this region. Anearly complete pancheloniid skull and other material from the


FIGURE 12. Ashleychelys palmeri, gen. etsp. nov. A–C, paratype skull of adult individ-ual, SCSM 90.19.1, in A, right lateral, B, dor-sal, and C, ventral views.

mid-Oligocene of Azerbaijan was referred to G. gwinneri (nowG. knorri) by Aslanova et al. (1979), who also reported remainsof Glarichelys from Romania and probably Kazakhstan. Thisspecies therefore was widespread in central Europe and westernAsia.

Two other species have been named from the mid-Oligocene.One of these, ‘Chelone’ acuticostata from Italy (Bergounioux,1954), is very poorly known and cannot be meaningfully diag-nosed or compared with the other species (Lapparent de Broin,2001). The other species, ‘Chelonia’ aralis from the Aral Sea areaof Kazakhstan (Khosatzky, 1945), likewise is very poorly knownand presently is considered to be Cheloniidae indet. (Averianov,2002).

Identifiable pancheloniid sea turtles are all but unknownfrom the later Oligocene (middle to upper Chattian) except forGlarichelys knorri from Germany (Karl, 1989, 2007), which isthe youngest known occurrence for that genus. A species of Pro-colpochelys must have been present, because P. charlestonensisof the mid-Oligocene gave rise to P. grandaeva of the early andmiddle Miocene (Weems, 1974) across this interval. Similarly,durophagous stem pancheloniids must have persisted in the east-ern Pacific region from the Paleogene into the Miocene (Parhamand Pyenson, 2010), and the rootstock of modern crown cheloni-ids must have existed and persisted from the Paleogene into theMiocene somewhere outside of the areas that have been stud-ied. Therefore, at a minimum a durophagous stem-pancheloniidghost lineage and a crown-cheloniid ghost lineage must have beenpresent across the Oligocene–Miocene boundary along with theProcolpochelys lineage.

COMPARISON WITH OTHER KNOWN OLIGOCENEPANCHELONIID ASSEMBLAGES

There are three areas from which significant collections ofpancheloniid turtles have come. These are the southeasternUnited States, Belgium, and southern Germany/Switzerland.Each area has a pancheloniid fauna distinctly different from theothers, because each of the constituent described species is lim-ited in its regional distribution.

Neither of the Belgian Oligocene pancheloniid sea turtlesappears to be particularly close to any of the South Car-olina Oligocene species. The plastron of Bryochelys waterkeynii(Smets, 1888) differs from that of Ashleychelys palmeri in thatA. palmeri has a much more anteroposteriorly elongate hyoplas-tron (Fig. 15B). The deep pit in the premaxillae of B. waterkeyniifor reception of an upturned tip of the dentaries (Smets, 1886)also is quite different from A. palmeri. Bryochelys waterkeyniihas a plastron that obviously differs from the plastron of Car-olinochelys wilsoni, for C. wilsoni has a much wider sutural bor-der between the hyoplastron and hypoplastron (Fig. 7D). Thefourth left costal of B. waterkeynii, though not extremely elon-gate, still is about twice as long as it is wide (Smets, 1888). In thischaracteristic, it differs obviously from the fourth costal of Pro-colpochelys charlestonensis, which is only slightly wider than it islong (Fig. 17B).

‘Chelone’ vanbenedenii (Smets, 1886) has wide vertebral scuteson its carapace that are similar to the scutes on Ashleychelyspalmeri (Fig. 14B) but quite unlike the much narrower verte-bral scutes found on Carolinochelys wilsoni (Fig. 6) and Pro-colpochelys charlestonensis (Fig. 17B). In other regards, ‘Ch.’vanbenedenii is quite different from A. palmeri. Its nuchal hasan elongate neck that projects posteriorly about halfway backbetween the first costals (Smets, 1886), and this characteristic isquite unlike the nuchal of A. palmeri (Fig. 15A) that does notproject at all rearward between the first costals. Similarly, thesixth costal of ‘Ch.’ vanbenedenii (Smets, 1886) is relatively lesselongated than the sixth costal of A. palmeri (Fig. 15A). Theproximal end of the third costal (Smets, 1886) shows no hintof a scalloped emargination to accommodate a supernumerarysplit neural such as is characteristic of P. charlestonensis and P.grandaeva (Fig. 17B, C).

None of the German/Swiss Oligocene pancheloniid turtlesshows any obvious close relationship to either the South Carolinaor Belgian turtles. Rupelchelys breitkreutzi is based on a partialcarapace that has unusually narrow neural elements (Karl andTichy, 1999); it is quite unlike any of the South Carolina turtlesdescribed here. The prefrontals of Glarichelys knorri are quiteelongate and its humerus has a very poorly developed lateral


FIGURE 13. Ashleychelys palmeri, gen. etsp. nov. A–C, restoration of adult skull, basedon SCSM 90.19.1, in A, dorsal, B, ventral, andC, right lateral views. Abbreviations: bo, ba-sioccipital; bs, basisphenoid; eo, exoccipital; fr,frontal; ju, jugal; mx, maxilla; op, opisthotic;pa, parietal; pal, palatine; pf, prefrontal; pm,premaxilla; po, postorbital; pt, pterygoid; qj,quadratojugal; qu, quadrate; soc, supraoccip-ital; sq, squamosal; v, vomer.

FIGURE 14. Ashleychelys palmeri, gen. et sp. nov. A, holotype carapace, ChM PV7002, in dorsal view; B, carapace of paratype, ChM PV6860, indorsal view. Characteristic surface texture is readily visible.


FIGURE 15. Ashleychelys palmeri, gen. et sp. nov. A, composite restoration of the carapace based on Figure 14A–B; B, composite restoration of theplastron of Ashleychelys palmeri based on the holotype ChM PV7002, paratypes ChM PV6860 and ChM PV9404, and epiplastra from ChM PV2141.Entoplastron is unknown.

FIGURE 16. Procolpochelys charlestonensis, sp. nov. Carapace of holotype, ChM PV6056.


FIGURE 17. A, carapace of Eochelone brabantica from the middle Eocene of Belgium; B, Procolpochelys charlestonensis from the Oligocene ofCharleston, South Carolina; C, P. grandaeva from the Miocene of New Jersey. Eochelone drawing based on a photograph provided by the InstitutRoyal des Sciences Naturelles de Belgium (preserved areas in gray). Procolpochelys charlestonensis drawing based on type (ChM PV6056, preservedareas in gray). Composited restoration of Procolpochelys grandaeva after Zangerl and Turnbull (1955).

tubercle (Zangerl, 1958). Both of these traits debar G. knorrifrom either the ancestry of or synonymy with any of the SouthCarolina species, all of which have a robust and elongate lateraltubercle and prefrontals with normal pancheloniid proportions.Additionally, the short and wide snout of G. knorri (Zangerl,1958) with its very short secondary palate (if the specimenillustrated by Aslanova et al. [1979] is correctly assigned to G.knorri) is quite different from the snouts of all of the SouthCarolina species. The carapace has very large fontanelles, evenin adults, and its surface is nearly smooth. In this trait it issomewhat reminiscent of the carapace of P. charlestonensis, butthe carapace of G. knorri is even more reduced and shows no in-dication of the split neurals that are so characteristic of the genusProcolpochelys. Similarly, Allopleuron lipsiensis (Karl, 2007)also has very large fontanelles as in Procolpochelys grandaevaand P. charlestonensis, but again there is no hint of split neurals.

DISCUSSION AND CONCLUSIONS

The turtle material found in the Ashley and Chandler Bridgeformations shows that three species of pancheloniid sea turtlesinhabited coastal North American waters in what is now SouthCarolina during mid-Oligocene time. In terms of the numbers ofspecimens found, the most common of these was Carolinochelyswilsoni. Less common but still abundant was Ashleychelyspalmeri, whereas Procolpochelys charlestonensis was relativelyrare in this area. The carapace of A. palmeri is estimated to havereached a length of about 55 cm, the carapace of C. wilsoni isknown to have reached a length of about 70 cm, and the carapaceof P. charlestonensis was about 95 cm long.

Most of the specimens of C. wilsoni and A. palmeri representjuvenile individuals that are less than half grown; adult specimens

are uncommon. In contrast, only three specimens, all adult, areknown of P. charlestonensis. This disjunct distribution of growthstages among the discovered remains strongly suggests that C.wilsoni and A. palmeri nested in or near the coastal South Car-olina region. The hatchlings stayed in the coastal region untilabout half grown before heading out to open waters, returningthereafter only to lay eggs. In contrast, P. charlestonensis ap-parently was only a seasonal adult visitor that nested and grewup elsewhere. The occurrence of half-grown specimens of P.grandaeva in the Miocene Calvert Formation (Weems, 1974),suggests that P. charlestonensis may have nested in Oligocenetime well north of the South Carolina region.

Notable in the Oligocene of South Carolina is the absenceof any turtle remains referable to durophagous stem cheloniids.Turtles of this sort were present in eastern North America (whatis now the eastern U.S.A.) throughout the Paleocene (Weems,1988), but after that their remains are unknown from the AtlanticCoastal Plain. In contrast, a durophagous stem-pancheloniid tur-tle (Pacifichelys) has been documented from deposits as youngas the middle Miocene in California (Lynch and Parham, 2003)and western South America (Parham and Pyenson, 2010), indi-cating the persistence of this lineage at least through the middleMiocene along the Pacific coasts of North and South America.The complete absence of durophagous stem pancheloniids in theOligocene and Miocene deposits of the Atlantic Coastal Plain(and conversely the complete absence of Procolpochelys alongthe Pacific Coast of North America) provides compelling nega-tive evidence that, by the mid-Oligocene, the pancheloniid seaturtles of the southeastern United States had become an endemicassemblage that was isolated from other Oligocene sea turtle as-semblages found elsewhere in the world. This sense of geographicisolation seemingly is reinforced by the cetacean assemblage that


TABLE 1. Comparison of the morphological characteristics of Ashleychelys, Carolinochelys, and Procolpochelys, with crown Cheloniidae(Trachyaspis, Natator, Lepidochelys, Caretta, Chelonia, Eretmochelys).

No. Character Carolinochelys Procolpochelys Ashleychelys Crown Cheloniidae

1 Skull profile (height/length) Normal (0.5) Normal (0.5) Flattened (0.4) Flattened or normal(0.4–0.5)

2 Orbit shape (height/width) Subround (0.7) Subround (0.8) Flattened ellipse (0.5) Subround to nearlyequant (0.7–0.9)

3 Skull surface texture Sculptured Smooth Sculptured Sculptured to smooth4 Cavum tympani tilt Near vertical Near vertical Rearward Near vertical to rearward5 Supraorbital ridge Slight Slight Prominent Slight to prominent6 Prefrontal location on orbital rim Anterodorsal Anterodorsal Dorsal Dorsal7 Frontal: max. external length/max. midline

length1.6 2.2 1.6 1.2–3.0

8 Narial height/width 1.0 1.0 0.7 1.09 Snout shape Angular Angular Rounded Angular to rounded

10 Premaxilla on palate Long Short Long Short to long11 Pterygoids Very narrow Very narrow Narrow Narrow12 Maxillary palatal trough Deep Deep Shallow Shallow to deep13 Platycoelous articulation between 6th and

7th cervicalsYes Yes Yes Yes

14 Carapace surface texture Sculptured Faintly sculptured Sculptured Faint to strong sculpture15 Carapace thickness Moderate Thick Moderate Moderate16 Carapace width/Carapace length 0.9 0.7 0.9 0.7–1.117 Location of greatest width across neurals

and costalsAcross fourth costals Across second costals Across third costals Across third or fourth

costals18 2nd peripherals notched Yes No No No19 Neurals Single Split Single Single or split20 Medial neural ridge No No Yes Yes or no21 Costoperipheral fontanelles Narrow Wide Wide Absent to wide22 Vertebral scutes Narrow Narrow Wide Narrow or wide23 8th costal rib attaches to 11th peripheral No Yes Yes Yes24 Humeral shaft Strongly curved Slightly curved Straight Straight25 Humerus with ‘V’-shaped lateral tubercle No No No Yes26 Scapular rod angle >110◦ Yes Yes Yes Yes27 Trochanter major connected to femoral

caput by ridgeNo Yes Yes Yes

28 Femoral trochanters connected by ridge No No Yes Yes

Abbreviation: max., maximum.

FIGURE 18. Skull of Procolpochelysgrandaeva, USNM 358862, from the lowerMiocene of Maryland. A, left lateral, B,posterior, C, dorsal, and D, ventral views. In Ba clayball is visible holding the skull roof in itscorrect relative position to the basicranium.


FIGURE 19. Restoration of the skull of Procolpochelys grandaeva based on USNM 358862. A, dorsal, B, ventral, C, posterior, and D, left lateralviews. The central palatal region of P. grandaeva is unknown and thus shown by dashed lines. Abbreviations: bo, basioccipital; bs, basisphenoid;eo, exoccipital; fr, frontal; ju, jugal; mx, maxilla; oc, occipital condyle; op, opisthotic; pa, parietal; pal, palatine; pf, prefrontal; pm, premaxilla; po,postorbital; pt, pterygoid; qu, quadrate; soc, supraoccipital; sq, squamosal; v, vomer.

FIGURE 20. Cladogram constructed byHennig argumentation, showing the relativerelationships of Carolinochelys, Ashleychelys,and Procolpochelys to each other and toearlier and later pancheloniids. The numericdesignations for the character states used inthis cladogram are shown in Table 1. Charac-ter states 13 and 26, common to all three taxa,are not present in primitive pancheloniidsand thus serve to place all of them abovethat evolutionary grade. Character state 25 ispresent in all crown cheloniids (living generaplus Trachyaspis) and thus serves to place allthree below that evolutionary grade.


FIGURE 21. Distribution of Oligocenepancheloniid sea turtles (sensu Joyce et al.,2004), showing known ranges of taxa. Of thesepancheloniid species, only Procolpochelyscharlestonensis is known to have persistedfrom the Oligocene into the Miocene togive rise to P. grandaeva. However, at leasttwo ghost lineages not yet found in thefossil record must have given rise to thedurophagous Miocene genus Pacifichelys(Parham and Pyenson, 2010) and to the livingcrown cheloniids.

co-occurs with these turtles, because two of the most primi-tive odontocete families (Xenorophidae and Agorophiidae) arepresently known only from the Carolina region, and xenorophidsare abundant on a scale similar to that of Carolinochelys wilsoni(Whitmore and Sanders, 1977).

Except for the advanced Miocene cheloniid Trachyaspis lardyi(senior synonym of Syllomus aegyptiacus; Lapparent de Broin,2001), the pre-Pliocene fossil record of pancheloniid turtles sug-gests that regionally localized distributions of coastal-dwellinggenera and species, incompletely adapted to high-seas life, werethe norm among these turtles. Truly cosmopolitan high-seas ch-eloniids are rare before the Pliocene and completely unknownbefore the Miocene. This change in habitat from predominantlycoastal forms to high-seas forms closely corresponds to (andis probably directly related to) the Neogene decline of der-mochelyid turtles, which were represented in Oligocene time byat least three species in South Carolina and today are representedby only one worldwide. It is not at all clear, however, whether ch-eloniids have moved into an increasingly vacated dermochelyidecological niche or whether cheloniids have in some way con-tributed directly to the decline of the dermochelyids.

ACKNOWLEDGMENTS

We thank J. Johnson who visited the Institut Royal des Sci-ences Naturelles de Belgique and brought back excellent pho-tographs of specimens of Eochelone brabantica for us to use.His efforts were fruitful thanks to the considerable help he re-ceived there from P. Godefroit, J.-J. Blairuacq, and W. Miseur.We thank W. Joyce and I. Danilov for their very helpful reviewsof this paper, J. G. Maisey and N. D. Pyenson for their commentson an early version of the manuscript, and R. T. J. Moody and J.F. Parham for helpful discussions on Eochelone brabantica. Wealso owe a debt of gratitude to the library staff of the U.S. Geo-logical Survey, who skillfully ran down many old and obscure ref-erences that were difficult to locate. We offer our sincere thanksto all of these people for their efforts on our behalf.

LITERATURE CITED

Aslanova, S. M., V. G. Pronin, and V. M. Chkhikvadze. 1979. [Remains ofsea turtles from the Oligocene deposits of Apsheron and northernUstyurt]. Vestnik Zoologii 6:17–22. [Russian]

Averianov, A. O. 2002. Review of Mesozoic and Cenozoic sea turtlesfrom the former USSR. Russian Journal of Herpetology 9:137–154.

Bergounioux, F. M. 1954. Les Cheloniens fossiles des terrains terti-aires de la Venetie. Memoire degli Istituti Geologica e Mineralogiadell’Universita di Padova 18:1–115.

Chkhikvadze, V. M. 1983. The Fossil Turtles of Caucasus and NorthernBlack Seaside Region. Metsniereba Publishers, Tbilisi, Georgia, 149pp.

Chkhikvadze, V. M. 1989. The Neogene Turtles of USSR. MetsnierebaPublishers, Tbilisi, Georgia, 102 pp.

Dames, W. 1894. Die Chelonier der norddeutschen Tertiarformation.Palaeontologische Abhandlungen (n.s.) 2:197–220 + 4 pls.

Dollo, L. 1909. The fossil vertebrates of Belgium. Annals of the NewYork Academy of Sciences 19:99–119.

Gaffney, E. S. 1975. A phylogeny and classification of the higher cate-gories of turtles. Bulletin of the American Museum of Natural His-tory 155:391–436.

Gray, J. E. 1825. A synopsis of the genera of reptiles and Amphibia, witha description of some new species. Annales of Philosophy 10:193–217.

Gray, J. E. 1831. Synopsis reptilium; or Short Descriptions of theSpecies of Reptiles. Part 1, Cataphracta, Tortoises, Crocodiles, andEnaliosaurians. Treuttel, Wurz Company, London, viii + 85 pp.

Hay, O. P. 1908. Fossil turtles of North America. Carnegie Institution ofWashington Publication 75:1–568.

Hay, O. P. 1923a. Characteristics of sundry fossil vertebrates; PartVII, New fossil turtle from Eocene marl of South Carolina. Pan-American Geologist 39:119–120.

Hay, O. P. 1923b. Oligocene sea turtles of South Carolina. Pan-AmericanGeologist 40:29–31.

Hirayama, R. 1994. Phylogenetic systematics of cheloniid sea turtles. Is-land Arc 3:270–284.

Joyce, W. G., J. F. Parham, and J. A. Gauthier. 2004. Developing a pro-tocol for the conversion of rank-based taxon names to phylogenet-ically defined clade names, as exemplified by turtles. Journal of Pa-leontology 78:989–1013.

Karl, H.-V. 1989. Die Schildkroten aus dem marinen Mitteloligozander DDR (Reptilia, Testudines). Mauritiana (Altenburg) 12:225–242.

Karl, H.-V. 2007. The fossil reptiles (Reptilia: Chelonii, Crocodylia)from the marine early Oligocene of the Weisselster Basin (cen-tral Germany: Saxonia). Studia Geologica Salmanticensia 43:25–66.

Karl, H.-V., and G. Tichy. 1999. Zur Taxonomie eines neuen Tribusvon Seeschildkroten aus dem Oligozan von Deutschland (Tes-tudines: Chelonioidea). Joannea Geologie und Palaontologie 57:61–77.

Khosatzky, L. I. 1945. Remains of a sea-turtle discovered in Oligocenedeposits of the Near-Aral region. Comptes Rendus (Doklady) del’Academie des Sciences de l’URSS (n.s.) 49:52–54.

Lapparent de Broin, F. de. 2001. The European turtle fauna from theTriassic to the present. Dumerilia 4:155–217.


Leidy, J. 1851. [Several fossils from the Greensand of New Jersey].Proceedings of the Academy of Natural Sciences of Philadelphia5:329–330.

Linnaeus, C. 1758. Systema Naturae per Regna Tria Naturae. Tenth edi-tion, Volume I: Regnum Animale. L. Salvii, Stockholm, Sweden,824 pp.

Lynch, S. C., and J. F. Parham. 2003. The first report of hard-shelledsea turtles (Cheloniidae sensu lato) from the Miocene of Califor-nia, including a new species (Euclastes hutchisoni) with unusuallyplesiomorphic characters. PaleoBios 23:21–35.

Mlynarski, M. 1959. Glarichelys knorri (Gray)—a cheloniid fromCarpathian menilitic shales (Poland). Acta PalaeontologicaPolonica 4:177–192.

Parham, J. F. 2005. A reassessment of the referral of sea turtle skulls tothe genus Osteopygis (Late Cretaceous, New Jersey, U.S.A.). Jour-nal of Vertebrate Paleontology 25:71–77.

Parham, J. F., and N. D. Pyenson. 2010. New sea turtle from the Mioceneof Peru and the iterative evolution of feeding ecomorphologies sincethe Cretaceous. Journal of Paleontology 84:231–247.

Riabinin, A. N. 1929. Une tortue du Paleogene des environs deVladikawkaz. Memoires de la Societe Russe de Mineralogie58:193–199. [Russian]

Sanders, A. E., and W. D. Anderson Jr. 1999. Natural History Investiga-tions in South Carolina from Colonial Times to the Present. Univer-sity of South Carolina Press, Columbia, South Carolina, 333 pp.

Sanders, A. E., R. E. Weems, and E. M. Lemon Jr. 1982. ChandlerBridge Formation; a new Oligocene stratigraphic unit in the lowercoastal plain of South Carolina; pp. 105–124 in Stratigraphic Notes1980–1982. United States Geological Survey Bulletin 1529-H.

Sloan, E. 1908. Catalogue of the Mineral Localities of South Carolina.South Carolina Geological Survey Bulletin 2, Series 4, 505 pp.

Smets, G. 1886. Chelone Van Benedenii. Annales de la Societe Scien-tifique de Bruxelles 10:109–128.

Smets, G. 1887a. Chelone (Bryochelys) waterkeynii. Annales de la SocieteScientifique de Bruxelles 11:291–302.

Smets, G. 1887b. Chelyopsis littoreus. Annales de la Societe Scientifiquede Bruxelles 11:303–307.

Smets, G. 1888. Les chelonees Rupeliennes. Annales de la Societe Scien-tifique de Bruxelles 12:193–214.

Tuomey, M. 1848. Report on the Geology of South Carolina. A.S. John-son, Columbia, South Carolina, vi + 293 pp. (plates publishedseparately with ‘Pleistocene fossils of South Carolina’ by MichaelTuomey and F. S. Holmes, 1857).

Ward, L. W., B. W. Blackwelder, G. S. Gohn, and R. Z. Poore. 1979.Stratigraphic revision of Eocene, Oligocene, and lower Miocene for-mations of South Carolina. Geologic Notes 23:2–23.

Weems, R. E. 1974. Middle Miocene sea turtles (Syllomus, Pro-colpochelys, Psephophorus) from the Calvert Formation. Journal ofPaleontology 48:278–303.

Weems, R. E. 1980. Syllomus aegyptiacus, a Miocene pseudodont sea tur-tle. Copeia 1980:621–625.

Weems, R. E. 1988. Paleocene turtles from the Aquia and Brightseat for-mations with a discussion on their bearing on sea turtle evolutionand phylogeny. Proceedings of the Biological Society of Washing-ton 101:109–145.

Weems, R. E., and E. M. Lemon Jr. 1984. Geologic map of the Stallsvillequadrangle, Dorchester and Charleston Counties, South Carolina.United States Geological Survey Geologic Quadrangle Map GQ-1581, 1 sheet, scale 1:24,000.

Weems, R. E., and W. C. Lewis. 2002. Structural and tectonic setting ofthe Charleston, South Carolina, region: evidence from the Tertiarystratigraphic record. Bulletin of the Geological Society of America114:24–42.

Weems, R. E., and A. E. Sanders. 1986. The Chandler Bridge Formation(upper Oligocene) in the Charleston region, South Carolina. Geo-logical Society of America Centennial Field Guide (SoutheasternSection) 6:323–326.

Weems, R. E., J. M. Self-Trail, and L. E. Edwards. 2004. Supergroupstratigraphy of the Atlantic and Gulf Coastal Plains (Middle? Juras-sic through Holocene, Eastern North America). Southeastern Geol-ogy 42:191–216.

Weems, R. E., W. B. Harris, A. E. Sanders, and L. E. Edwards. 2006.Correlation of Oligocene sea level cycles between Western Europeand the Southeastern United States; pp. 205–206 in G. Camoin, A.Droxler, C. Fulthorpe, and K. Miller (eds.), Sea Level Changes:Records, Processes and Modeling—SEALAIX’06 Abstract Book,Giens, France, 25–29 September 2006, Association des Sedimentol-ogistes Francais, Paris, Publication 55.

Wegner, T. 1917. Chelonia gwinneri Wegner aus dem Rupelton vonFlorsheim: Abhandlungen Herausgegeben von der Senckenbergis-chen. Naturforschenden Gesellschaft 36:361–372.

Whitmore, F. C., and A. E. Sanders. 1977. Review of the OligoceneCetacea. Systematic Zoology 25:304–320.

Wilmarth, M. G. 1938. Lexicon of Geologic Names of the United States(Including Alaska). United States Geological Survey Bulletin 896,Parts 1–2, 2396 pp.

Zangerl, R. 1958. Die oligozanen Meerschildkroten von Glarus. Schweiz-erische Palaontologische Abhandlungen 73:1–56.

Zangerl, R., and W. D. Turnbull. 1955. Procolpochelys grandaeva(Leidy), an early carettine sea turtle. Fieldiana: Zoology 37:345–384.

Zug, G. R. 2001. Turtles of the Lee Creek Mine (Pliocene: North Car-olina); pp. 203–218 in C. E. Ray and D. J. Bohaska, (eds.), Geologyand Paleontology of the Lee Creek Mine, North Carolina. Smithso-nian Contributions to Paleobiology, Volume 90.

Submitted September 17, 2012; revisions received March 8, 2013;accepted March 30, 2013.Handling editor: Sean Modesto.

carolina, u.s.a. oligocene pancheloniid sea turtles from...

Documents