THE BOX TOETOISES OF NORTH AMERICA.
By W. E. Taylor.
In the discussion of the genus Terrapene * it is my purpose to pre-sent, in detail, the osteological characters and the geographical distribu-tion of the genus. In doing this I have had the privilege of examininga great number of specimens, representing aijproximately the wholedistribution of the genus in the United States aud Mexico.tHistory and nomendature.?From very early times systematic zoolo-gists have mentioned the common box tortoise, under various names,as occurring in North America. Over sixty authors have given one ormore species as found in various localities. But most of these writersmerely mention the names of the species, while localities, if given, areindefinite. A full discussion of these writings would be out of placeiu this article, and hence I content myself with barely mentioning afew of the most important papers, the authors of which have reportednew species or proposed new names.Edwards, in 1751, gives a good figure of Terrapene Carolina.Linnieus, in his tenth edition of the Systema Naturiv., 1758, mentionsTestudo Carolina as from Carolina. His description was taken fromEdwards.Gray, in 1844, described Emys Mnosternoides, which may be Terrapenetriunguis, Agassiz.Gray, in the Proceedings of the Zoological Society of London, 1849,describes Onycliotria mexicana as from Mexico.Agassiz, in his Contributions to the Natural History of the UnitedStates, 1857, describes Cistudo virginea (Gmelin), Gistudo triunguis,
*Tliis is the generic name, whlcli has to be used for the American hox tortoises.Ciatucla, Fleming, 1822, is a synonym of Terrapene, Merrem, 1820.tThe author is under special obligations to Dr. G. Baur, Assistant Professor, Uni-versity of Chicago, for material aid in the preparation of this paper, the synonymiesbeing, for the greater part, prepared from his manuscripts, and to the authorities ofthe U. S. National Museum for the loan of valuable specimens. He has also to thankthe following gentlemen for specimens from various parts of the country, viz : Prof.H. Garman, Prof. S. S. Maxwell, Prof. Beujamiu B. Pentield, Dr. O. P. Hay, Mr. Gus-tave Kohn, Mr. Julius Hurter, Mr. Roy R. Larkin, Prof. Theo. D, A. Cockerell, Prof.H. B. Duncanson, Dr. Adolph Meyer, and Prof. J. D. Bruner.Proceedings of the TJ. S. National Museum, Vol. XVII?No. 1019. 573
574 THE BOX TORTOISES OF NORTH AMERICA?TAYLOR. vol.xvii.Cisiudo ornata, and Cistudo major, the first being Testudo Carolina ofLinnjpus, while the last three were recognized as new species.Strauch, in his Vertheilnng der Schildkriitcn, 18(35, included all ofour American forms nnder Terrapcne caruiata, Linnams.Cope, in his Check-List of North American Batradiia and Reptilia,1875, recognizes Cistudo clnu.sa, subspecies clausa (Gnielin and Hol-brook), C. clausa, subspecies tnunguis, and C. ornata. He placed TestudoCarolina of Linnaeus under the genus Testudo.True, in Yarrow's Check-List of Korth America Reptilia and Batra-chia, 1883, recognizes Cistudo caroUna, (\ Carolina triunguis, and C.ornata.Boulenger, in his Catalogue of Chelonians in the British Museum,1888, admits but two species, as follows: Cistudo Carolina, C. Carolinavar. Mnosternoides, Cistudo Carolina var. major, C. Carolina var. mcjcicana,and C. ornata.Baur, in the American Naturalist, 1893, after making a study of theosteology of Terrapene, adopts the following species: Terrapene major,T. Carolina, T. mexicana., T. triungrds, and T. ornata.By continuing the work of Dr. Baur, and adding to the collectionsalready made by him, I am able to confirm his observations, and, at thesame time, to add a new species.1 recognize the following species : Terrapene major, T. hauri, T. Caro-lina, T. mexicana, T. triunguis, and T. ornata.Not until within the last few years has the osteology of our formsbeen carefully studied. Gray overlooked the rudimentary quadrato-Jugal in T. Carolina. Agassiz, in his characterization of the genus,speaks of the temporal arch as "either cartilaginous or only partiallyossified," forgetting the fact that T. major, described by him, possessesa well-developed zygomatic arch.Boulenger gives the absence of a bony temporal arch as character-istic for Terrapene.Briihl gives a figure of the skull oi Terrapene Carolina, but overlookedthe quadrato-jugal.Baur, in the Zoologischer Anzeiger for 1888, No. 290, first pointed outthe fact that T. Carolina possesses a rudimentary quadrato-jugal, con-trary to the opinion held by Gray and others. Later, in Science, No. 126,189], he gave the osteological characters of Terrapene major, T. Caro-lina, and T. ornata; and still later, in the American Naturalist, 1893,the complete osteological characters and general descriptions of T.major, T. Carolina, T. mexicana, T. triunguis, and T. ornata were given.The geographical distribution of Terrapene has never been com.pletely worked out. Most authors have merely given a few localitiesor the range of each species in a general way.Agassiz, in 1857, gives T. Carolina {Cistudo virginea) as the northeastern type and erroneously states that it "has the most extensiverange " of the genus. He also mentions Terrapene {Cistudo) triunguis
1894. FBOCEEDIXGS OF THE NATIONAL MUSEUM. 575as the western and soutliwestern type, and Terrapene (Cisttido) ornataas the northwestern type. Owing to the limited data accessible hefailed to recognize that the last-named species is the western form, fromthe Yellowstone to the llio Grande. He also mentions Terrapene{Cisiado) major as the southern and soutiieastern type of the genus.Cope, in 1875, states that Terrapcne Carolina is found in the "Easternregion and Louisianian and Floridian districts:" T. triunguis in the
'?austroriparian region to Georgia, eastern Pennsylvania," and T. ornatain the " Central region."Baur, in 1893, gives the geographical distribution of the species asfollows: T. major, " tSouthern States;" T. Carolina, "Eastern States toIndiana;" T. mexicana, "Mexico;" T. triunguis, "Louisiana, Arkansas,Indian Territory, Mississippi, Georgia;" T. ornata, "Central States."The writer's notes on the geographical distribution of TerrapeneyfiW.be given under a separate heading.Genus T E R R A P E N E, M e r r e m.1820.
?
Terrapene, Merrem, Yersucheines Systems der Amphibien, p. 27 (type, Testudoclausa, Gmelin=:T. caroVnia, LiNN.EUS).
?
Baur, Zool. Auz., 1888, No. 296;Science, xvii, 1891, p. 190; Proc. Amer. Pliilos. Soc, 1892, p. 245.1822. Cisiuda, Fleming, Pbilosoi>hy of Zoology, ii, p. 270.1832. DicUda (part), Rafinesquk, Atlantic Journal, i, p. 64 (in Analyse de la Nature,1815, name only).183o. ?Piixideinys (part), Fitzinger, Ann. Wiener Mus., i, p. 123.18-14. Emyoides, Gray, Catalogue of Tortoises in British Museum, ii.27.1849. Onychoiria, Gray, Proc. Zool. 8oc. Loudon, 1849, p. 17.1857. Ciistudo, Agassiz, Contributions to tbe Natural History of the United States,I, p. 444.?BouLEXGKR, Catalogue of tbe Chelouians in British Musenm,1889, p. 114?Baur, Proc. Amer. Pbilos. Soc, 1892, p. 44.KEY TO THE SPECIES OF TERREPEXE.
I. Three digits on the hind foot.1. Zygomatic arch complete. Webs absent. Phalanges on the fore foot 2-3-3-3-2, hind foot 2-3-3-2-1 .... bauri, p. 576.2. Zygouiatic arch iucouiplete. Webs absent.a. Number of phalanges iu the fore foot 2-3-3-2-2, hind foot 2-3-3-3-1.Carapace tectiform mexicana, j). 579.1). Number of jihalanges in fore foot 2-3-3-2-2, hind foot 2-3-3-2-1. Cara-pace not tectiform TRiUNGUis, p. 580.II. Four digits on tbe hind foot.1. Zygomatic arch complete. Wel)s distinct. Phalanges in the fore foot 2-3-3-3-2. bind foot 2-3-3-3-2 , . major, p. 575.2. Zygomatic arch rudimentary. Digits slightly webbed. Phalanges iu the forefoot 2-3-3-3-2, hind foot 2-3-3-3-2. Carapace keeled . . Carolina, p. 577.3. Zygomatic arch absent. Webs absent. Phalanges iu the fore foot 2-2-2-2-2, hind foot 2-3-3-3-1. Carapace not keeled ORNATA, p. 581.TERRAPENE MAJOR (A g a s s i z).1857. Cistudo major, Agassiz, Contributions to the Natural History of the UnitedStates, I, p. 445 (type iu Mus. Comp. Zool., Cambridge, Mass.; locality oftype, Mobile, Ala.).
?
Garman, Bull. Essex Inst., xvi, 1884, p. 10.
576 THE BOX TORTOISES OF NORTH AMEBICA?TAYLOE. vol.xvii.1865.-
1870.- -Terrapene carinata (part), Strauch, M6m. Acad. Sci. St. Petersb. (7), viii, No.13, p. 46.-Cistndo Carolina var. major, Guay, Supplement to the Catalogue of ShieldReptiles, p. 19.?BouLKNGEit, Catalogue of Chelonians in British Museum,1889, 1). 117.1891.? Terrtipene major, Baur, Science, xvii, 1891, p. 190; Anier. Natural., xxvii,1893, i>. 677.Quadratojugal well developed; zygomatic arch complete, and alwayspresent and relatively wider and more fully developed than in T. bauri.Maxillft distinctly notched. Upper branch of the.s(;apnla considerablylonger than the inner branch (endo-scapnla); digits with distinct Avell-developed webs. N^umber of phalanges in the forefoot, 2-3-3-3-2 ,'hind foot, 2-3-3-3-2. Number of claws in the hind foot, 4.Carapace oblong in horizontal outline; transverse outline semi-circular. Anterior margin but slightly curved; moderately compressedabove and median ridge distinct. The width of the carapace com-pared to length varies from 10-12 to 10-14. First pleural plate morethan half as wide as long. Ground color varying from black, yellowisholive or dark brown to very light dingy brown. The yellow spots arearranged in radiating lines but are not connected. The plastron isusually of a straw-yellow color with the dermal plates more or lessfaintly bordered with black or dark brown.This species is found in regions adjacent to the Gulf coast of Florida,Alabama, Louisiana, and Texas.JAst of specimens.No.
1894. PROCEEDINGS OF THE NATIONAL MUSEUM. 577The type (No. 8352, U.S.N.M.) was collected in Florida by F. B. Meek.The species is named for Dr. Baiir, who first noticed the peculiaritiesof the type, but having only the one specimen considered it as anexceptional individual of T. triunguis.
678 THE BOX TORTOISES OF NORTH AMERICA?TAYLOR. vol.xvii.1831.?Emys (Cistuda) carolince, Gray, in Griffith's Animal Kingdom, ix, Append., p. 7.1831.
?
Cintuda Carolina, Gray, Synopsis Reptilium, i, p. 18.
?
Holbrook, NorthAmerican Herpetology, 2 ed., i, 1842, p. 31, PI. ii.
?
Ue Kay, Zoology ofNew York, pt. iii, 1842, p. 24, PI. i., tig. 1.
?
Gray, Catalogue of Tortoisesin the British Museum, 1844, p. 30.1835. Cistudo Carolina, Dumkril and Bibron, Erp6tologie G^n^rale, ii, 183.5, p. 210;IV, 1854, p. 224.?DuMERiL, Museum d'Histoire Naturelle de Paris, Cata-logue Methodique de la Collection des Reptiles, 1851, p. 7 (part).
?
Wied,Nov. Acta Acad. Leop. Carol., xxxii, i, 1865, j). 1, PI. i, fig. 1. Gray,Supplement to the Catalogue of Shield Reptiles, 1870, p. 19; Hand-list ofSpecimens of Shield Reptiles in the British Museum, 1873, p. 18. True,in Yarrow, Bull. U. S. Nat. Mus. 24, 1883, p. 37.
?
Boulenger, Catalogue ofthe Chelonians in the British Museum, 1889, p. 115, figs. 32, 33.
?
Strauch,Mom. Acad. Sci. St.-P<5tersb. (7) xxxviii. No. 2, 1890, p. 62 (part).?Baur,Proc. Amer. Philos. Soc, 1892, p. 44.1835. Pyxidemi/8 schneideri, Fitzinger, Ann. Wiener Mus., i, p. 123.1835. Pyxidemys virgalata, Fitzinger, Ann. Wiener Mus., i, p. 123.1835. Pyxidemys clausa, Fitzinger, Ann. Wiener Mus., i, p. 123.1857.?Cistudo virginica, Agassiz, Contributions to the Natural History of the UnitedStates, I, p. 445, PI. iv, figs. 17-19; PL viii, figs. 10-14.1862. Terrapene carinata (part), Strauch, Mdm. Acad. Sci, St.-P^tersb. (7) v. No. 7,p. 96; VIII, No. 13, 1865, p. 45.1875. Cistudo clausa, subsp. clausa. Cope, Bull. U. S. Nat. Mus. No 1, p. 53.188-i.? Cistudo carinata, S. Garman, Bull. Essex Inst., xvi, p. 10.1892. Cistudo Carolina var. Carolina, H. Garman, Bull. 111. State Lab. Nat. Hist., in,p. 219.Qiiadrato-jngal rudimentary, triangular in shape and connected withthe quadrate only. Zygomatic arch absent. Maxillary, distinctlybeaked but not notched. Cervicals anil their processes relatively short.Upper branch of the scapula somewhat longer than the inner branch(endoscapula), but not so long as in T. major. Number of phalanges inforefoot, 2-3-3-3-2 or 2-3-3-2-2; hind foot, 2-3-3-3-2. Digits slightlywebbed. Claws in the hind foot, 4. First pleural plate approximatelytwice as long as wide.Carapace ovoid in outline; ratio of width to length in adults, aboutas 10 to 12. Not compressed above; keel distinct and median ridgeabsent. Color dark brown or blackish, marked by yellowish or brownishradiating sjiots and lines. Often the keel is marked by an interruptedyellow or brownish yellow line.Plastron oval in outline, with distinct shoulders on each side of thehinge area. Ground color dull yellow, varionsly blotched and mottledwith lavendar brown.Kentucky and Tennessee specimens are the most beautiful of ourNorth American forms, and may possibly be entitled to rank as avariety of T. Carolina. The carapace is ovoid in horizontal outline;semiovoid in transverse section. The general color is black or verydark brown, marked by bright yellow, sometimes the latter color pre-vailing. The posterior portions of the second, third, and fourth ver-tebral plates are marked by distinct but irregular quadrate blotcheswhich are broadly bordered b}' bright yellow. The upper portions ofthe first, second, and third costal plates are similarly marked. The
1894. PROCEEDINGS OF THE NATIONAL MUSEUM. 579spaces in front of tliese blotclies on the vertebral plates and below themou the costal plates are marked by spots and radiating blotclies ofyellow. The presence of a distinct keel and the absence of a mediandepression are points readily distinguished. The plastron is yellow,irregularly blotched and mottled with much brown and black, broadestacross the femorals. The upper portion of the head and neck is brown,slightly specked with yellow; a yelloAV line beginning at the posterior ofthe eye runs back over the ear and the posterior of the skull. Mandible,throat, and lower neck light yellow. Ui^per scales of the legs grayish
;
lower scales mostly orange or reddish. The individual described is afemale collected at Midway, Ky., by Prof. H. Garman. Another speci-men very similar but younger was sent me by Prof. S. S. Maxwell. Thisone was collected at the mouth of Billingtons Creek, near Lovelaceville,Ballard County, Ky., in the extreme western portion of the State, wherethe writer has often observed other specimens. I'rof. Benj. B. Peutield,of Nashville, Tenn., sent me two specimens which agree with the Ken-tucky individuals in every point except that a distinct triangular quad-rato-jugal was present.In general form and osteological characters this variety grades into,if it is not identical with, typical specimens of T. Carolina, and may beregarded as the southern extension of this species. Hence the distri-bution of T. Carolina may be given as eastern United States, south toCarolinas, Georgia, Tennessee, and Kentucky, reaching the Missis-sippi, west to eastern Illinois and Wisconsin, and north to Canada.Lint of specimens.
No. Collector.
Dr. G. Baur
....do
....do
....doL. StejnegerJ. D. FigginsProf. S.S.Ma.x well.Prof. H. Garman. .
.
Dr. O. P. HayProf. B.B.Peii field.
....doP. S. TouugFrank BurnsJ . D. Ridgwaydo ."B. A. Bean..Dr. G. Banr.
...do
Collection.
No. 20544, U. S. N. M.
No. 11600, U.S.N.M.No. 14624, U.S.N.M.No. 14670, U.S.N.M.No. 14670, U.S.N.M.No. 8539 (?)No. 13588, U.S.N.M.
Locality
.
Indiana.do.New Haven, Conn.do.Washington, D.C.Kensington, Md.Lovehueville, Ky.Midway, Ky.Kentucky.Nashville, Tenn.do.Bock Creek, D. C.Darlington, S.C.Wheatland, Ind.do.Carlisle, Pa.Bainhridge, Pa.Albany, N.Y.do.TERRAPENE MEXICANA (Gray).1849.
?
Onychotria mexicana, Gray, Proc. Zool. Soc. London, 1849, p. 17, PI. ii (type inBritish Museum; locality of type, Mexico).?Duofes, La Naturaleza (2),1888, p. 11.1855. Cistudo mexicana, Ghay, Catalogue of Shield Reptiles in the British Museum,p. 40; Supplement to the Catalogue of Shield Reptiles, 1870, p. 19.
?
BocoURT, Mission Scientifique au Mexique, Reptiles, fasc. 1, 1870, p. 17(part).?GuNTHER, Biologia Ceutrali-Aniericana, Reptilia, 1885, p. 1.Cope, Bull. U. S. Nat. Mus. 32, 1887, p. 21.
580 THE BOX TORTOISES OF NORTH AMERICA?TAYLOR.1889.
?
Cistudo Carolina var. mexicana, Boulenger, Catalogue of Cheloniaus in theBritisli Museum, p. 118.1891. Terrapene mexicana, Baur., Science, xvii, 1891, p. 191; Amer. Natural , 1893, p.677.Quadrato-jugal very rudimentary; connected with the (luadrate andrather elongated. Zygomatic arch absent. Upper branch of the scapula as in T. major. No webs between the digits and only three clawson the hind foot. Number of phalanges in tlie forefoot li-3-3-2-2; inhind foot 2-3-3-3-1. Maxilla not notched.Carapace oval in horizontal outline; rather triangular in transverseoutline. Ratio of width to length in adults approximately 10 to 13.The additional plate found between the fourth and fifth vertebral platesseems to be characteristic. First pleural plate more than three timesas long as wide.Color of the carapace brownish yellow marked by dark brown radiat-ing lines and irregular spots. Head yellow, irregularly marked byblack or dark brown. Throat yellow, i)osterior neck light brown.Plastron curved in the region of the abdominal and femoral plates,and distinctly mucronate posteriorly. Color, whitish yellow, the pos-terior border of each dermal plate bordered with smoky brown.City of Mexico and Tampico, Mexico.List of specimens.
No.
1
1894. PBOCEEDINGS OF THE NATIONAL MUSEUM. 581Quadrato-jugal rudimeutary and triangular; conuected with thequadrate only. Zygomatic arch absent. Scapula as in T. major. Nowebs between the digits, and only three claws on the hind foot.Phalanges in the forefoot, 2-3-3-L'-2; hind foot, 2-3-3-2-1. Maxillaeslightly notched. Oervicals and their processes relatively short.-Carapace moderately oval, keeled, and slightly compressed. Eatioof width of carapace to length in adults about 10 to 13. First pleuralplate similar to T. hauri. Ground color dark-brown or olive-yellow,much mottled with yellow.Plastron oval. Ground color yellow, much mottled with brown.The Louisiana form seems to be a dwarf variety of this species andis peculiar to Louisiana. They were first noticed by Agassiz whostates: "Had I not noticed a few larger specimens from the OsageEiver and from Georgia, I should not hesitate to consider them as adistinct species." They are characterized by their relatively small size,peculiar markings, and rather full development of the quadrato-jugal.Some times the quadrato-jugal is sufficiently developed to come intocontact, if not uniting, with the jugal. These individuals are readilydistinguished by their small size, three toes, and general color. In colorthey vary from pale yellow or dusky to dark brown, marked by smallradially distributed yellow spots, often only few in number. The factthat they grade into larger forms of other localities prevents them frombeing ranked as a separate species.Mississippi, Louisiana, Arkansas, Indian Territory, southern Mis-souri and Kansas, and Texas.List of spechneti-s.No.
682 THE BOX TORTOISES OF NORTH AMERICA?TAYLOR. vol.xvii.Quadrato-jugal entirely absent, and hence zygomatic arch neverpresent. Postorbital arch much more slender than in other species.Maxilh^ notched. Cervicals and their processes relatively very short.Upper branch of the scapula of the same length as the inner branch(endoscapula). Digits without distinct webs. Number of phalangesin forelimb, 2-2-2-2-2; hind limb, 2-3-3-3-1. Number of claws in thehind foot, 4.Carapace in horizontal outline broadly oval: much compressed above;medium ridge usually present, but the keel always absent, even in theyoung. Interpleural foramina^, between the ribs of the first and seconddorsal vertebrae almost obsolete. Ratio of the width to the length ofthe carapace in adults, never exceeding ten to twelve, rarely more thanten to eleven. The depression of the carapace is usually so great as torender the outline of a transverse section almost oblong in shape.The ground color of the carai)ace varies from very dark-brown,marked with bright-yellow radiating lines with a yellow medium line,to very light colors with no markings. Third vertebral plate less thantwo-thirds as long as wide, being in all other species relatively longer.Plastron widest across the abdominal plates. Ground color brown,marked by irregular yellow lines. Posterior to the hinge ligamentthese lines show a tendency to become longitudinal, while on theanterior of the hinge ligament these lines are mostly transverse.East of the Rocky Mountains to Wisconsin, eastern Illinois, centralIndian Territory, and eastern Texas.List of specimens.No. Collector.Dr.G.KaurdododoDr.O.P.HaydoProf. T. D. A. CockerellRoy R. LarkinProf. J. D. BrunerWilliam Lloyddo ,Dr. KennerlyProf. H. B. DuncanBonHaydeuKennicottHayden
..-.. doBaileyJulius HurlerPalmerW. S. WooddoDr. HaydenDr. Meyer
Collection. Locality.
No. 16484, U. S. N.M .No.20959, U. S.N.M .No. 52, U. S.N. M ....
No. 7541, Ur. S. N.M .
.
No. 7542, U. S. N. M .No. 57, U. S. N.M ....doNo. 15861, U. S. N. M .No. 16491, U. S. N. M .No. 7177, U. S. N.M .,No.l56,U.S.N.M ...No. 156, U. S. N. M .
.
U.S.N.M
Lawrence, Kans.Indiana (?).Kansas; Logan County.do.do.Kansas.Las Cruces, N.Mex.do.Kl Paso, Tex.South of Devil's River,Texas.Corpus Christi, Tex.San Antonio to Fort Inge,Tex.Peru, Nebr.Sand Hills, Nebraska.Illinois.Yellowstone (River).do.Kennedy, Nebr.St. Louis, ]Mo.Old Fort Cobb, Okla.Republican River,Nebraska.do.Fort Laramie, Wyo.Kankakee, 111.GENERAL REMARKS.The geographical distribution of Terrapene offers many interestingphases for study. The relatively fixed habits of the species of the genusrender their variations more or less local in character. Good illustra-
1894. PROCEEDINGS OB^ THE NATIONAL MUSEUM. 583tions of this fact are seen in the T. ornata of Texas and the T. triunguisof Louisiana and southwestern Arkansas and southeastern IndianTerritory. But, notwithstanding these conditions, the specific cbar-acters are maintained throughout the range of each species, as hasbeen shown.Terrapcne major may be said to be strictly a gulf species, having forits range the gulf coast from the Kio Grande to Florida, possibly includ-ing southern Georgia. The specimens examined by me seem to be largerin the average than individuals of other species, and in general osteo-logical characters they certainly represent the primitive form of thegenus. They possess a well-developed quadratojugal, a complete zygo-matic arch, and are distinguished from T. hauri by the number ofphalanges, color pattern, and webbed digits, there being four claws oneach hind foot.Terrapene hauri may be said to belong to the peninsula of Florida,possibly including southern Georgia. It resembles T. major in havinga complete bony zygomatic arch and T. triunguis in having but three toeson the hind foot, while it differs from either in the number of its pha-langes. The color markings of the type, excepting plastron, are almostidentical with T. ornata, a species from which T. hauri is completelyseparated both by osteological characters and geographically. Thespecimens mentioned by Agassiz as three-toed specimens of T. Caro-lina from Xorth Carolina, and T. triunguis from Georgia may belong tothis species.Terrapene Carolina is found in the northeastern United States, extend-ing from the St. Lawrence and Great Lakes south to the Carolinas andTennessee and west to the Mississippi Eiver in Kentucky and to east-ern Illinois.* This species seems to exist in greatest abundance inNew Jersey and adjacent regions, bnt Dr. Hinds informs me that it isvery common at Lebanon, Tenn. It seems to be, in a certain sense, themountain species, being found thoughout the mountains of Pennsyl-vania, as well as elsewhere, and seems to be coextensive with the Alle-gany range southward. It is characterized by its rudimentary quad-ratojugal, the hooked upper jaw, and the presence of a distinct keelon the carapace, the number of phalanges, and its slightly webbeddigits.
* Unfortuuately I am unable to speak detinitely as to the precise western limit ofT. Carolina north of the Ohio Eiver. The specimens reported by Prof. H. Garmanas from various points in southwestern Illinois are inaccessible at the time of writ-ing, owing to alterations now being made in the museum at Champaign, 111. Prof.W. K. Higley, in his catalogue of Wisconsin reptiles, mentions T. Carolina as col-lected at the following places in Wisconsin, namely, Walmouth County, two speci-mens; Milwaukee; Pine Lake; La Crosse; and Green Bay. Not having had anopportunity to examine Prof. Higley's specimens, I have been unable to verify hisidentifications. Mr. Hurter, in his catalogue of St. Louis reptiles, mentions onespecimen of T. Carolina, but since it is the only one which has been found in thatlocality he rather infers that it was brought into St. Louis. This inference is mademore probable by the fact that the individual was found in the elevator yards.
584 THE BOX TORTOISES OF NORTH AMERICA?TAYLOR. vol.xvii.On the south it adjoins or overlaj^s the territory of T. hauri and T.triunguis, while on the west it overlaps the T. triungnis and T. ornatafor a comiiaratively short distance. Its principal characters remainconstant whether the specimens be taken from the Atlantic coast, themountains of Pennsylvania and Tennessee, or the prairies of Ohio andIndiana. Specimens from Kentucky and Tennessee exhibit certainpeculiarities in color markings. These peculiarities have been discussedfully under the specific characters of T. Carolina.Terrapene mexicana is at once distinguished by its tectiform carapace,rudimentary quadrato-.jugal, and the number of phalanges. (Severalauthors have rejiorted this species from the City of Mexico, and Bocourtmentions one specimen as from Tampico, Mexico?the most northerlylocality reported. These meager data would hardly justify me in out-lining its range.Terrapene triungnis in many respects shows a peculiar distribution.Koughly speaking it may be said to occupy the swampy districts of theLower Mississii)pi and bordering territory. It is found in the southernhalf of the State of Mississippi, and passing into the extreme south-western portion of Illinois is found as far north as St. Louis, Mo.;thence west in the Osage River Valley in eastern Kansas; thence south-west to the Arkansas Rivei" and its tributaries in Indian Territory;thence to Matagorda Bay bordering the gulf from the Rio Grande toAlabama. This species is characterized by its rudimentary triangularquadrato-jugal, its number of phalanges and webless digits, three clawson the hind foot, its keeled carapace with its slight median depression.Both Dr. Baur and O. P. Hay inform me that in southwestern Arkansas and central and eastern Indian Territory specimens are often foundwhich are greenish yellow and without spots, but marked individualsare also present in the same region. Specimens from Louisiana arerather small, having a dwarfed appearance, and are somewhat peculiarin their markings. Their osteological characters, also, are somewhatvariable. However, typical specimens of the species are found in thisregion.Terrapene ornata is found from the Rocky Mountains east to LakeMichigan and Indiana, south to the Ozark Mountains, and east to west-ern Indian Territory and central Texas, and from the Rio Grande Rivernorth to the Yellowstone River. This species may be said to belong tothe plains and the table-lands. In Kansas, where it becomes extremelynumerous, Prof. Cragin speaks of it as so abundant as to become anuisance as a cumberer of the ground. It seems to subsist and thrivein our most arid climates, being found in the sand hills of Nebraskaand the barren regions of New Mexico and Texas. In its geographicaldistribution it seems to be governed more by soil, rainfall, and vegeta-tion than latitude. Throughout all its range it is characterized by theentire absence of the quadrato-jugal, the possession of only two pha-langes in each digit in the forefoot, its broadly oval carapace, with amedian ridge but without keel, a i)lastron widest across the abdominal
Fig. 1.SKULL OF CrCLEMMYS AMBOINENSIS.
1894. PROCEEDINGS OF THE NATIONAL MUSEUM. 585plates, and the variegated markings of tlie carapace. Texas specimensare somewhat stouter aud more compact, and often individuals possess-ing no color markings are found, but, notwithstanding these exceptions,the species is remarkably constant throughout its range. It occupies alarger territory than all other si>ecies combined.In the taxonomy of Terrapene, as first noted by Dr. Baur, the modifi-cation of the zygomatic arch occupies au important i)osition. Thequadrato-jugal is well developed in primitive forms of the genus, rudi-mentary in intermediate forms,and absent in T. ornata, the mostspecialized species.The skull of a species belongingto a closely related genus, Cy-clemmys amhoinensis, is representedin fig. 1. In this species the elon-gated quadrato-jugal ih) lies alongthe anterior border of the quad-rate completely separating the a.-PostfromaL S.-Quadratojugal. c.-Jugal.latter, as well as the squamosal,from the jugal (c) and postfrontal {a). The postfrontal is much elon-gated and widened, and with the jugal and quadrato-jugal forms acomplete zygomatic arch.Fig. 2 shows the zygomatic arch of T. major. The postfrontal hasretreated and in this species forms a narrow club-shaped bone justposterior to the orbit and takes no ijart in the formation of the zygo-matic arch. However, a complete arch is present, composed of thesomewhat quadrangular quadrato-jugal and the jugal.Fig. 3 illustrates the structure of the zygomatic arch of T. baio-i, aform, in this respect, closely related to T. major. It will be noticedthat this arch is more slender than in T. major.In T. Carolina (fig. 4) the quadrato-jugal is rudimentary and is notconnected with the jugal, and hence the bony zygomatic arch is incom-plete.In T. mexicana (fig. 5) the zygomatic arch is incomplete, aud thequadrato jugal has been reduced to a very small remnant. The samething may be said of T. triunguis (fig. 6), except that in this speciesthe quadrato-jugal is more nearly triangular. In T. ornata (fig. 7) thezygomatic arch has completely disappeared.In this connection the modification of the phalanges seems worthy ofour attention. In all species there are five digits in each foot, and onthe fore foot of each there are five well-developed claws. However, inthe fore foot the number of phalanges varies, the number being inT. major, T. bauri, and usually T. Carolina, 2-3-3-3-2; in T. mexicana,T. triunyuis, and sometimes T. Carolina, 2-3-3-2-2, and in T. ornata,2-2-2-2-2.The hind foot of T. major, T. Carolina, and T. ornata i^ossesses fourclaws, while in the remaining species but three claws are present. With
586 THE BOX TORTOISES OF NORTH AMERICA?TAYLOR. vol.xvii.reference to the number of phalanges in each hind foot the speciesare as follows : T. major and T. Carolina, 2-3-3-3-2, the same numberas in the fore foot; T. ornata and T. mexicana, 2-3-3-3-1 ; T. hauri and T.triunguis, 2-3-3-2-1.The loss of the phalanges on the hind foot might be accounted for bysupposing that the distal phalange does not develop. But in the caseof the fore foot, where the full number of claws are still present, thedistal phalange evidently does not disappear. Hence the reduction
Figs. 2-7.SKULLS OF TERRAPENE.Showing modificiilion of the zygomatic arch in diflerent species.
a, Postfronlal. h, Quadrato-jugal, c, Jugal.
( For explanation of figures see page 585.
)
must take place either by the coossification of two phalanges, or aphalange and a metacarpal, or by the disappearance of a phalange.Zehntner, 1890, after studying the development of the Alpine Swift,Ctjpselus melba, concludes that the reduction in the number of pha-langes takes place in this species by coossification. He states that inthe fourth digit the first pbalange probably unites with the fqurthmetatarsal, while the third and fourth phalanges unite with each other.
1894. PROCEEDINGS OF THE NATIONAL MUSEUM. 587Also the second and third phalanges of the third digit unite. Hencein the Cypselus, notwithstanding that in the early stages of growth thenormal number of phalanges for birds, 2-3-4-5, is developed, in thisgenus the number of phalanges in adults is 2-3-3-3.Pfitzner, 1890, discusses the reduction of the number of phalangesin the little toe of man. In embryos and young children he found thenormal number of phalanges 3, but in the case of older children themiddle and the end phalanges usually unite, though in some instancesthe middle phalange disappeared.In Terrapene I have not been able to determine the method of thereduction in the number of phalanges since embryological material ofT. ornata has not been obtained. I hope to be able to investigate thisquestion further at a later date. But, in the examination of youngerspecimens, I find no evidences of coossiflcation such as would probablybe indicated by the disproportionate length of a phalange formed bythe union of two phalanges.The distribution of the species of Terrapene presents several interest-ing problems. It seems probable that all of our species are derived fromone form. The closeness of the relations of the species would seem toindicate that our forms are varieties rather than species. However, atleast two difficulties are in the way of this conclusion. First, there canbe no question but that if we take two extremes of development of thespecies of the genus, for instance, T. major and T. ornata or T. Carolinaand T. ornata, we must recognize them as separate species. But sinceother intermediate forms seemingly connect these species, if the close-ness of relations be considered as indicating varietal characteristicsonly, we are forced to consider all species as varieties, a conclusion thatwould seem to be erroneous. Second, while the relations indicatedby a study of the difl'erent species seems close, yet the distinctionsseem definite and fixed, even where the ranges of the species overlap.The study of a number of si)ecimens seems to indicate that the difterentspecies are derived from one form, and that afterwards, by isolation,caused possibly by geological and climatic agencies, they became dis-tinct. When we remember the comparatively fixed abode of theseanimals it seems reasonable to suppose that these changes might havebeen brought about by relatively simple agencies which need not neces-sarily have acted simultaneously. Hence, it would seem proper toclassify each form as a distinct species, each possessing certain fixedosteological characters. If these conclusions be true then it wouhlseem most reasonable to suppose that T. ornata has become more dis-tinct from the other species by its comparatively longer isolation, aidedby the generally arid climate of its habitat.It will at once be noticed that of the species found within the UnitedStates each occurs in a district which, in CL-rtain geographical features,is peculiar to itself, T. ornata occupies a district peculiar in its aridclimate and, for the most part, sandy soil. T. triunguis is found in the
588 THE BOX TOBTOIiiES OF NORTH AMERICA? TAYLOR. vol.xvii.low swampy regions of the Mississippi and its tributaries, the climaticconditions here being the reverse of the district occupied by T. ornata.T. Carolina occurs in both the mountains and plains, yet its climatic sur-roundin;LiS are somewhat uniform and peculiar to its region. T. hauriinhabits tlie peninsula of Florida, while T. major is found adjacent tothe gulf coast, the two species occupying districts each peculiar initself.Another i^eculiar fact in the distribution of Terrapene is that, so farnot a single species has been reported west of the Kocky Mountains,notwithstanding its great abundance on the table-lands east of thesemountains.For an explanation of the cause of this fact we can only surmise. Itis well known that in the Alleghany Mountains T. Carolina ascends toa height of several thousand feet, as high, probably, as any otherspecies of our chelonians. Prof. Cockerell and Mr. Ray R. Larkiuhave sent us specimens of T. ornata, collected at Las Cruces, N.Mexico, between five and six thousand feet above the sea level. Possi-bly the true explanation of the fact that Terrapene has not been foundwest of the Rockies may be accounted for by the great elevation andconsequent climate of the mountains. If this explanation be the trueone, then it is only a question of time when the Terrapene will finds itsway over the Rocky Mountain range.The information extant and the material at our command, we regretto say, does not permit us to accurately outline the limits of the specieseither south or north of the United States boundary lines. It seemsprobable that at least one or more of our species may extend intoMexico, tliough no such record exists. Neither do we have anyauthentic record of the existence of the Terrapene in British America,though it is iiossible, if not probable that specimens will ultimatelybe found there. However, if we judge from the comparatively smallnumber of s])ecimens found on the northern borders of the UnitedStates, the species can not be abundant north of the boundary line.University of Chicago,Chicago, Illinois, June 1, 1894.