THE PEARCE­SELLARDS Series : TEXAS MEMORIAL MUSEUM THE UNIVERSITY OF TEXAS NUMBER 3 : A New Fossil Tortoise from the Texas Miocene, With Remarks on the Probable Geologic History of Tortoises in Eastern U.S. BY WALTER AUFFENBERG May, 1964 The Pearce-Sellards Series TEXAS MEMORIAL MUSEUM, THE UNIVERSITY OF TEXAS The Pearce-Sellards Series, as other Museum publications, is financed solely bv profits from the Museum’s sales counter. Thetitle seekstocommemorate thefirsttwodirectorsoftheMuseum, both now deceased: J. E. Pearce, Professor of Anthropology, University of Texas, whose efforts were in large part instrumental in establishing the Texas Me­morial Museum; and Dr. E. H. Sellards, who was director of the Museum from the time the doors were opened in 1939 until his retirement in 1957. If these papers can maintain the standards of excellence these men set, the suc­cess of this series is assured. W. W. Newcomb, Jr. Director Table of Contents Geochelone williarnsi new species 3 Holotype 3 Type Locality and Horizon 3 Diagnosis 3 Description of Type 3 Comparisons 6 Discussion 7 SummaryJ and Conclusions 9 Literature Cited 10 Illustrations and Tables Fig. 1 Carapace of Geochelone ivilliamsi 4 Fig. 2 Plastron of Geochelone ivilliamsi 5 Table 1 Measurements of the Type of Geochelone toilliamsi 5 A New Fossil Tortoise from the Texas Miocene, with Remarks on the Probable Geologic History of Tortoises in Eastern U.S.1 WALTER AUFFENBERG Florida State Museum, University of Florida A complete shell of a fossil tortoise in the collection of The University of Texas—Bureau of Economic Geology, from the Miocene Oakville formation of the Texas Coastal Plain, represents an undescribed and distinctive species.It is referred to the genus GeocheJone on the basis of a narrow nuchal scute,first suprapvgal embracing the second, and with the median length of the hvpoplastralgreaterthanthatofthe hyoplastral. Geochelone williamsi new species2 Holotype A complete shell of an adult male tortoise, University of Texas —Bureau Economic Geology, 31084-11; collected by Dr. John A. Wilson in 1949. Tune Locality and Horizon.—Garvin Gullv, 2 mi. north of Navasota, ~ . Jl J Grimes County, Texas, Garvin Gullv local fauna, Lower Oakville member,Oakville formation, Arikareean,Early Miocene. Diagnosis.—A species of Geochelone (subgenus uncertain), apparently most closelv related to G. ducatelli from the Miocene Calvert formation of Maryland. It is distinct from G. ducatelli in the following characters: a rhomboidal, instead of a pentagonal entoplastron; epiplastral projection not truncated, not sculptured, nor swollen ventrally; carapace and plastron more elongate; xiphiplastral notch deeper. Description of Type. —A complete shell of an adult male land tortoise,slightly displaced dorsally and crushed laterally (Figs. 1, 2). Pertinent meas­urements are provided in Table 1. Carapace smooth, without deep growth rings, only moderately broad,slightly truncated anteriorly, rounded posteriorly, with the free peripheralborders not emarginated, and without any recurving or flaring; vetebral region of the carapace forming a smooth arc when viewed from the side,slightly more sloping anteriorly; no keel on the bridge, though this area is linearly swollen; plastron just even with the anterior edge of the carapace SponsoredinpartbvNSF G—17613and G—23562. 2 Named for Ernest E, Williams, Museum of Comparative Zoology, Harvard College,in recognition of his contributions to our knowledge concerning both fossil and Recent land tortoises. Fig. 1. Carapace of Geochelone williamsi new species, Holotype, University of Texas—Bureau EconomicGeology 31084—11, Lower Miocene, Arikareean, Grimes County, Texas; close to G. ducatelli (Collinsand Lynn), Calvert Miocene, Maryland. Top: lateral view. Bottom: dorsal view. Fig. 2. Plastron of Geoche/one williamsi new species, Holotype, University of Texas—Bureau Economic Geology 31084-11, Lower Miocene, Arikareean, Grimes County, Texas. Left: internal view. Right:ventral view. TABLE 1 Measurements (in mm) of the type of Geochelone williamsi new species carapace length 334.0 bridge length 144.5 plastral length 313.0 entoplastron length 64.0 gular length 54.0 entoplastron width 62.5 humeral length 49.0 hyoplastron length 67.0 pectoral length 20.5 hvpoplastron length 78.0 abdominal length 108.0 xiphiplastron length 59.5 femoral length 36.0 1st suprapvgal length 55.0 anallength 40.0 anteriorwidth 1stsuprapvgal 19.5 xiphiplastral notch length 10.5 posterior width 1st suprapygal 73.5 xiphiplastral notch width 38.0 2nd suprapygal length 32.5 xiphiplastral external height 32.0 2nd suprapygal width 42.5 greatest thickness epiplastral lip 29.0 nuchal scute width 13.0 epiplastral lip length 48.0 nuchal scute length 21.0 anterior lobe length 76.5 2nd vertebral length 66.0 posterior lobe length 92.0 2nd vertebral width 65.0 and with the tips of the xiphiplastral projections not extending to the pos­terior carapaceal border. Free borders of peripherals acute; first, third, fifth, seventh and eighth neurals subrectaneular, second, fourth and sixth clearly ? octagonal;firstOsuprapygalembracingOalow,diamond-shaped,second supra­pvgal; pleurals only moderately alternately narrower and broader at their distal ends. Nuchal sente longer than wide; costal scutes low, almost square;first vertebral scute as long as wide, third not wider than the others, fourth longest; epiplastral projection thickened, and slightly concave dorsally, not extending greatly beyond the anterior plastral margin, edge acute, with a slight median notch epiplastral lip long, extending posteriorly for a con­siderable distance, deeplv excavated behind; gular scute entering the ento­plastron; humero-pectoral sulcus extending laterally for about one half its length, then turning forward abruptly in a sharp angle, continuing to the plastral border in a very gentle arc; abdomino-pectoral sulcus extendinglaterally for slightly over half its length, and then turning postero-laterallvto the bridge; abdomino-femoral sulcus and hypo-xiphiplastral articulation not superimposed, former anterior to the latter; both inguinal and axillary scutes present, large and simple, the former in moderately long contact with the femoral scute, and visible in ventral view; hvpoplastron considerablylonger than the hyoplastron; posterior lobe moderate in length, its outer edgewith a distinct angle just anterior to the outer end of the femoro-anal sulcus,and a broad notch at this point; femoral and anal scutes normal; xiphiplastronbroadly notched posteriorly with the tips broadly divergent; ventral surface of the epi-and xiphiplastron without distinctive sculpturing; posterior por­tion of the hvpoplastron and most of the xiphiplastron slightlv concave. Comparisons. —Most of the other Miocene species of Geochelone with which G. williamsi is to be compared can be placed in the subgenus Hes­perotestudo. Miocene members of this subgenus include G. oshorniana, im­pensa, angusticeps, and inusitata. This group is very distinct from that to which williamsi belongs. G. vaga and G. fani are both species with rugoseshells that may belong to the Hespewtestudo line. G. klettiana and G. undata are large tortoises from New Mexico, represented by only the pvgal areas. It is doubtful if these species can ever be properly diagnosed. The specimen onwhichthenameGeocheloneundahuna isbased isprobablyaberrant,and the name may be a synonym of G. primavae. The latter is quite distinct from G. williamsi in possessing all hexagonal neurals. G. niohrarensis and G. ted­whitei seem to belong to the subgenus Caudocheh/s; both differ from G. williamsi in the shape of a number of different parts of the plastron and carapace. The shape of the anterior projection of the plastron, shape of the entoplastron, and the distance between the humeral and abdominal sulci in G. arenivaga are all quite different from those in williamsi. Testudo copeifrom the Deep River Miocene belongs to the genus Gopherus because it has a nuchal scute which is wider than long.The species most closely related to G. williamsi is clearly Geochelone ducateUi of the Maryland Calvert formation. Both forms are characterized bv (in combination) slender rib heads, the great extent of the gular and pectoral scutes on the inner, anterior surface of the plastron, and a propor­tionately thick posterior lobe at its base. Fortunately, the types of both species are adult males, of about the same size. Although the carapace of ducatelli is not complete, in every character which can be compared these two species approach each other more closelv than either approaches anyotherNorthAmericanfossiltortoisedescribed sofar. In summary, the major distinguishing features of the two species are: G. williamsi G. ducateUi Entoplastron essentially rhomboidal. Entoplastron essentially pentagonal. Entoplastral projection not truncated, Entoplastral projection truncated, not sculptured, not swollen ventrally, slightlv sculptured, and swollen ven­very thick. trally, thinner. Carapace and plastron more elongate. Carapace and plastron less elongate.Xiphiplastral notch deeper (anal scute/ Xiphiplastral notch shallower (anal = = anal notch 3.81). scute/anal notch 0.95). Dorsal and ventral edges of pleurals Dorsal and ventral edges of pleurals more differentiated (dorsal pleural less differentiated (dorsal pleural3/ventral pleural .3-2.00). 3/ventral pleural 3 = 1.52). Anal scute longer (femoral scute/anal Anal scute shorter (femoral scute/anal = scute = 0.90). scute 2.35). Epiplastral lip thicker (length lip/thick-Epiplastral lip thinner (length lip/thick­ness lip 1.66). ness lip 2.37). = DISCUSSION At least two species groups of land tortoises are known to have existed in eastern North America during the Miocene. One of these groups is now known to have been represented by two species in this area. These are Geochelone williamsi and G. ducateUi. The relationships of this group to other fossil or Recent species groups of Geochelone remain unknown. The second species group in the Miocene of eastern North America is representedby Geochelone tedwhitei. This species is a member of the extinct subgenusCaudochehjs (Auffenberg, 1963), which had a long fossil history, extendingfrom the Oligocene to the Late Pleistocene of North America. It is only after the Early Miocene that additional tortoise taxa appear in eastern North America. These include the subgenus Hespewtestudo (Auf­fenberg, 1962) and the genus Gopherus. Theavailable datasuggestthatthesubgenusHesperotestudofirst success­fully invaded eastern United States during the Late Miocene or Early Plio­cene, and that it was probably restricted to the southern part of this geo­graphic area from this time to the Late Pleistocene, when it became extinct. The genus Gopherus also invaded eastern North America from the present Central Plains area, but at an even later time. In the Southeast the earliest known forms occur in the Pleistocene of Florida. Furthermore, the eastern species represents only one of several phyletic lines known to have occurred in western United States at the same time. As far as tortoise evolution in North America is concerned, several basic working hypotheses are suggested: 1. Almost all the major phyletic lines within the North American tortoise genera or subgenera evolved in central and/or western North America. Only one species group (G. williamsi and G. ducatelli) may have evolved in eastern North America, probablyin theLate Oligocene. 2. Of the three fairly well known and widely distributed tortoise groups in western North America during the Miocene ( Gopherus, Hesperotestudo and Caudo­chehjs), only one ( Gopherus ) failed to reach eastern United States by the end of the Pliocene. In all three groups diversification within eastern United States seems negligible. Furthermore, the successful immigrant species represent only a small part of the total phyletic diversification found in western United States during thesame timeperiods. The relationships, or indeed even the validity of Floridemijs nanus—an unusual, diminutive form from the Pliocene of Florida—remain uncertain. The peculiar arrangement of the gular-pectoral sulcus (transverse, and not crossing the entoplastron) may be an example of a variant, unfortunatelypresent in the only individual of this species known. The remainder of the specimen bears a considerable similarity to some of the species of Sti/lemysfrom the John Day beds of Oregon. If related to Stijlemys, then Floridemijsisprobably aPliocenezoogeographiccounterpartoftheamphisbaenid genusRhineura, of which extinct species are known from the Oligocene of Colo­rado, whereas the only living species is restricted to Florida. The peculiarsnake genus Stdosoma (Pliocene to Recent of Florida) may eventuallv be foundtofitthe same chronogeographicpattern. The discoveries of fossil tortoises during the past two decades in the far West, the northern and central plains states, and southeastern United States have greatly extended our knowledge of the history of these interestingvertebrates. Conclusions of a preliminary and general nature regarding the phylogeny and paleogeography of tortoises are now possible. This contribu­tion, as well as those of Hibbard (1960) and Brattstrom (1961) are examples.Nevertheless, data are still lacking from certain geographic areas importantto such considerations. Northeastern United States is an example, since there are no terrestrial deposits of Tertiary age in this area that contain fossil tortoises.As aresult, ourknowledgeofthehistoryoffossiltortoisesin eastern UnitedStates isbased onremainsthathavebeenfoundonlvintheSoutheast. Another important area for which there are very few paleozoogeographicdata is that encompassing Arizona, New Mexico and all of Mexico. This is a critical region in tortoise evolution in view of its role as a potential source area for new taxa, an area traversed by both east-west and north-south highways of dispersal, and (at least as far as Mexico is concerned) a pre­sumably important tortoise refngium (as it is for Gopherus at the presenttime). SUMMARY AND CONCLUSIONS 1. A new species of land tortoise, Geochelone williamsi, is described from the Lower MioceneOakvilleformationoftheTexasCoastalPlain. 2. With G. ducatelli of the Maryland Miocene it seems to form a small species group confined to southeastern United States. Together they may be part of the Miocene equivalent of an Austroriparian fauna; probably ultimately derived from the Caudochelys phyletic line, which is more characteristic of western United States during this same geologic period. 3. Throughout the Tertiary there were several eastward migrations of tortoise groups. The earliest of these migrations was effected by the subgenus Hodri­anus, a primitive group that apparently reached southeastern United States in theEarly orMiddleEocene. According to the available data the second major eastward migration from western United States was effected by the subgenus Caudochelys, whose earliest southeastern records are from the Lower Miocene of Florida. By this same time, a typically (?) eastern group, including both williamsi and duca­telli, had alreadv evolved in eastern United States, presumably from a pre-Miocene ancestral form that moved eastward during the Oligocene. No Oligo­cene tortoises are known from eastern United States. 4. During the Miocene there were a number of distinct phyletic lines of tortoises in western North America that either never reached eastern North America, or did so only after the Miocene. One of the groups which apparently never reached eastern United States was the genus Sfylemys (unless Floridemys nanus was derived from Stylemys). A distinct group present in the Miocene of central and western North America, but that did not reach eastern North America until a later date, is the suhgenus Hesperotestudo of the genus Geo­chelone. By the Miocene, Hesperotestudo had already broken into a number of species groups. Two of these reached Florida in the Pliocene. The Orthopygiaseries is represented by G. hayi, and the Turgida series by an undescribed species. The genus Gopherus, with a long fossil history and considerable diversi­fication in western North America, did not reach the eastern half of the con­tinent until the Middle Pliocene. 5. The unique eastern United States genus Floridemys may be related to Sty­lemys. If so, it is a relict Pliocene form with essentially the same chronogeo­graphical pattern as the genus Rhineura. 6. During the Pleistocene there were three groups of tortoises in eastern United States: Caudochelys, Hesperotestudo, and Gopherus. The presumed typicallyeastern groups found in this area during both Miocene and Pliocene times had alreadv disappeared. Both Caudochelys and Hesperotestudo became extinct over most of North America during the Wisconsin.3 The genus Gopherus is still found in southeastern United States. 7. The peninsula of Florida is often regarded as both a center of speciation and a refugium for more northern or western animal groups. However, it does not seem to be particularly important in either as far as tortoise evolution is con­cerned. The only major tortoise evolutionary lines that may have evolved in Florida are those represented by G. williamsi and/or G. ducatelli, and Flori­demys nanus. As far as is known, no tortoise group existed in Florida for anylonger period of time than the same group did in western United States (unlessFloridemys is a synonym of Stylemys). Though several major evolutionary lines areknowntohaveextendedtheirranges tosoutheasternUnited States,themi­grations were apparently accomplished only long after these same groups had already madetheirappearancein westernUnitedStates. LITERATURE CITED Auffenberg, W. 1962. A new species of Geochelone from the Pleistocene of Texas. Copeia,1963 (3): 627-636. 1963. ThefossiltortoisesofFlorida. GeneraGeocheloneandFloridemys. Bull. Florida State Museum, 7(2):53-97. Brattstrom, B. H. 1961. Some new fossil tortoises from western North America with remarks on the zoogeography and paleoecology of tortoises. Journ. Paleon., 35(3):543-560. HireAim, C. W. 1960. An interpretation of Pliocene and Pleistocene climates in North America. Pres. Address, 62nd Annual Report, Mich. Acad. Sci., Arts & Letters, pp.1-30. Geochelone ( Hesperotestudo) tvilsoni is known from post-Pleistocene deposits of southeastern Oklahoma (C14 date of 10,000 b.p. fide Slaughter—pers. comm., 1962).Specimen examined by the author. Museum Publications Funds for publicationofthePearce-Sellards Series aswellasall otherpamphletsand booklets are derived solely from the proceeds of the Museum’s sales counter. Profits, if any, from Museum publications are used to issue additional papers. THE PEARCE-SELLARDS SERIES No. 1. Fossil Bears from Texas, by Bjorn Kurten 25 No. 2. Post-Pleistocene Raccoons from Central Texas and their Zoo-geographic Significance, by Thomas Wright and E. L. Lundelius, Jr 40 No. 3. A New Fossil Tortoise from the Texas Miocene, with Remarks on the Probable Geologic History of Tortoises in Eastern U.S., by Walter Auffenberg 25 BULLETINS OF THE TEXAS MEMORIAL MUSEUM No. 1. Mylohyus nasutus. Long-nosed Peccary of the Texas Pleistocene, by Ernest L. Lundelius, Jr 1.00 No. 2. Part 1. The Friesenhahn Cave, by Glen L. Evans. Part 2. The Saber-toothed Cat, Dinobastis serus, by Grayson E. Meade . . . 1.00 No. 3. A Bibliography of the Recent Mammals of Texas, by Gerald G. Raun 1.00 No. 4. Handbook of Texas Archeology: Type Descriptions. A Cumula­tive Series (Published jointlyby the Texas Archeological Societyand the Texas Memorial Museum. Not available in exchange) . . 6.00 No. 5. Salvage Archeology of Canyon Reservoir: The Wunderlich, Footbridge, and Oblate Sites, bv Leßoy Johnson, Jr., Dee Ann Suhm, and Curtis D. Tunnell 2.00 No. 6. The Ethnography and Ethnology of Franz Boas, by Leslie A. White 2.00 No. 7. Fossil Vertebrates from Millers Cave, Llano County, Texas, byThomas Hudson Patton 2.00 No. 8. A Bisexual Fish Species by Clark Hubbs 2.00 Texas residents must add two per cent sales tax for orders over 24 cents. Contributions and Bequests Contributions and bequests to the Texas Memorial Museum may be made in securities, money, books, or collections. They may, if desired, take the form of a memorial to a person or a cause, to be named by the giver. Contributions to the Museum are allowable as deductions in computing net income for the federal income tax.