PRELIMINARY REPORT ON A RECENTLY DISCOVEREDPLEISTOCENE CAVE DEPOSIT NEAR CUMBERLAND,MARYLAND. Bt James Williams Gidlet,Assistant Curator of Fossil Mammals, United States Natiomil Museum.INTRODUCTION.The recent fortunate discovery of Pleistocene mammal remains incave deposits near Cumberland, Maryland, adds one more to therather limited number of such occurrences and promises to be ofgreat importance in working out the comparatively little knownPleistocene mammahan life of the eastern United States. It mayalso aid in the proper correlation of these and similar deposits ofthe East with the better known Pleistocene beds of other parts of thecountry.The preliminary investigation of the Cumberland Cave depositmade last October produced most encouraging results. Over 100specimens were secured, consisting principally of jaws and jaw frag-ments, which represent 22 recognizable genera, including one genusnow exclusively African, and at least 29 species, most of which areapparently extinct, or are now living in remote localities. The work ofexploration again taken up by the writer in May of the present yearhas already added several other forms to the list, and is yieldingbetter material of many of the forms represented in the collectionof last autumn. This material will be pubHshed with the final resultsand conclusions at a later date when the exploration is completed.The location of this important find is at the bottom of a deep cutof the Western Maryland Railway where it passes through the northend of a spur or ridge of hmestone near the little village of Corrigans-ville, at the mouth of Cash Valley, about 4 miles northwest of Cum-berland. The ledge is here upturned at an angle of about 90?, theroadway cutting it nearly at right angles, and the excavation isabout 100 feet deep at the point where the fossil-bearing depositswere exposed. When first observed the workmen naturally regardedthe bones as those of animals now living in the neighborhood, andbeyond exciting their curiosity at finding them buried in the rocksand debris of a small cavern at so great a depth, no particular interestProceedinqs U. S. National Museum, Vol. 46?No. 2014-
94 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.46.was aroused. I was told that quantities of the material weredestroyed by the steam shovel and djmamite in making the excava-tion, while many specimens were picked up by the workmen andothers and carried away as curiosities. After the cut had been com-pleted the locality was visited by Mr. Raymond Ai'mbruster, of Cum-berland, Maryland, and Mr. George Roeder, of Swetnan, Virginia, who,obtaining a few specimens from the still undisturbed deposits out-cropping at the side of the excavation, recognized the possiblescientific value of the material and reported it to the United StatesNational Museum. The credit of the discovery therefore belongsentirely to these gentlemen, and especial praise is due to Mr.Armbruster for the subsequent interest and for the assistance he hasgiven in securing this material for science. It seems appropriate inthis connection to express my obligation to these gentlemen and alsoto Mr. G. H. Friend, principal assistant engineer in charge of theCumberland division of the Western Maryland Railway, and Mr.Martin Gallagher, industrial commissioner, for their interest in thework and their assurance of hearty cooperation in continuing theinvestigation of the still unexplored deposits. Thanks are also dueMr. G. C. Hendrickson, of Cumberland, for placing at my disposal aportion of a skull taken from the railroad cut, representing an extinctspecies of the dog famUy.GEOLOGIC HISTORY AND AGE OF DEPOSITS.The upturned ledge of rock in which the cave deposits occur repre-sents the lower Helderbergian division of the Devonian. It forms asegment of the much eroded west wing of a great anteclinal fold havinga nearly north and south axis. The total thickness of the formationat this place is about 900 feet, but the cave chamber containing thefossil bones and other caverns in the immediate vicinity seem to beconfined to a single stratum not more than 20 feet thick. Thesmall size and peculiar disposition of the caverns in a single planesuggest a true fissure cave, although it can not properly be so called.The cause for this resemblance is probably due to the upturned con-dition of the strata which brings the bedding plane and hence theline of cleavage to a nearly perpendicular position. This has forages given free access to the corrosive action of the surface watersalong the line of strike which easily following the lines of cleavage ofthe rocks, would spread downward to great depths and laterally onlyalong the line of outcrop, without forming caverns of any great size.PROBABLE MANNER OF ENTOMBMENT.In making the railway cut, several small chambers at higher levelsthan the one containing the bones were encountered, and before thework of excavation began there was said to have been an opening tothe surface on the crest of the hill directly above the middle of the
NO. 2014. ON A PLEISTOCENE CAVE DEPOSIT?GIDLEY. 95present roadbed. This opening so nearly overhead probably at onetime served as a trap through which were introduced the animalswhose remains are now in the deposits of the bone cavern. Thereare other openings along the line of outcrop of the ledge, one of themat about the same level with the bone-bearing deposits, appearing atthe north end of the ridge where it slopes abruptly down into theWills Creek Valley. These openings may or may not have communi-cated at one time with the caverns intersected by the railroad cut,but probably had nothing to do with the accumulation of materialin the latter.From Brown's * account of the Conard Fissure, it would seem thatthe conditions governing the accumulation of material in the Cumber-land Cave were quite similar. The bones for the most part are muchbroken, yet show no signs of being water worn. They are foundscattered fairly uniformly throughout the entire mass of unstratifiedaccumulations which consist entirely of cave clays and breccias,unevenly hardened and more or less cemented together by stalactiticmaterials. There is an almost entire absence of admixture of sandor gravel, or in fact anything that would suggest the possible aid ofstream currents in sorting or placing the material during the processof accumulation. It seems probable therefore that this little fossil-bearing pocket represents the accumulation of a great number ofyears in which the conditions were such that animals, both large andsmall, sometimes by accident, sometimes by being dragged thereby carnivores, occasionally became entrapped in the upper cham-bers of the cave. Thus carcasses of the larger animals were proba-bly caught and held in crevices not far beneath the surface of theground and remained there until the bones were sufficiently maceratedto allow them to fall apart by their own weight, when the separatedbones would work their way by gravity to lower and lower levels untilthey finally came to rest at the bottom of the cavern then a hundredfeet or more below the surface of the ground. The broken and scat-tered condition of the bones found in the deposits would be accountedfor in this way.The mammals represented in the collection are undoubtedlyPleistocene and probably pre-Wisconsin in age; a more exact geo-logical horizon of the deposits, however, can not at present bedetermined. From this preliminary study they appear to be aboutthe equivalent of the Port Kennedy cave deposits, the fauna ofwhich was described by Cope ' and is now regarded as early Pleisto-cene. The Cumberland Cave fauna may represent a somewhat laterphase. But this supposition can be verified or disproven only by acareful comparison of the material with that from Port Kennedyand other localities.
? Memoirs Amer. Mas. Nat. Hist., vol. 9, pt. 4, 1907, pp. 163.? Joum. Acad. Nat. Sci. PhUa., vol. 11, 1899, pp. 194-267.
96 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.46.LIST OF FAUNA REPRESENTED.Equus sp.Tapirus cf . liays^i ( ?) Leidy.Taurotragus americanus Gidley.Platygonus cf . vetus ? Leidy.TJrsus (Euarctos) cf. americanus ? Pallas.Ursus (Euarctos) vitahilis, new species.Canis armhrusteri, new species.Canis sp.Vulpes (?) sp.Mustela cf . mson Schreber.Lepus americanus ? Erxleben.Lepus sp.Ochotora cf. princeps.Synaptomys sp.5'. (Myctomys) cf. borealis (Richardson).Microtus cf. chrotorrhinus Miller.Microtus sp.Neotoma sp.Napseozapus sp. probably new.Peromyscus cf. leucopus (Rafinesque).Erethizon, two new species.Marmota cf. monax Linnaeus.Sciuropterus cf. alpinus Richardson.Sciurus Jiudsonicus Erxleben.Myotis, new species (?).Vespertilio grandis Brown.Vespertilio sp.Blarina cf. hrevicauda (Say).DESCRIPTION OF NEW SPECIES AND NOTES ON CANIDS.CARNIVORA.Genus URSUS (EUARCTOS) Gray.This subgenus is represented by at least two species, as shown bymany specimens consisting of foot and limb bones, and a few upperand lower jaw portions containing teeth. At least one of these formsis new and is described below.URSUS (EUARCTOS) VITABILIS, new species.Type.?Lower jaws, nearly complete, lacking incisors (Cat. No.7665, U. S. Nat. Mus.), see figs. 1, la, p. 97.Description.?About the size of U. (Euarctos) americanus, but differsfrom that species in (1) comparatively larger canines; (2) wider space
ON A PLEISTOCENE CAVE DEPOSIT?GIDLET. 97between the anterior cheek-teeth, combined with a less wide branchingof the horizontal rami in general; (3) a relatively larger symphysis,which IS more sharply constricted and more flattened laterally behindthe canines; and (4) longer diastemabetween canines and cheek-teeth.A second specimen, a portion of theright maxillary (Cat. No. 7664, U. S.Nat. Mus.), is probably referable to thisspecies. It contains the two molars,which do not differ materially, exceptin their somewhat smaller size, from thecorresponding ones of U. americanus.
Fig. I.-Ursus (Euarctos) vitabilis. Type-specimen. Outer view or right lower jaw, 2-3 nat.Superior view, nat. size.size.95278??Proc.N.M.vol.46- -13-
98 PROCEEDINGS OF THE NATIONAL MUSEUM.Genus CANIS.There are at least two carnivors represented in the collections whichare referable to Canis as that genus is at present understood. Onespecies is here described.CANIS ARMBRUSTERI.i new species.r?/pe.?Portion of a left lower jaw (Cat. No. 7662, U. S. Nat. Mus.),containing three teeth, p^ to m^. (See figs. 2, 2a.)Paratypes.?Portion of a right lower jaw (Cat. No. 7661, U. S. Nat.Mus.) containing four teeth, 2>2 to m^ (see figs. 3, 3a), and the alveolifor Pi, and the canine; and portions of the right and left lower jaws
Fig. 2.?Canis akmbkusteri, type-specimen, portion of left lower jaw. 2. Outer \iew, 2-3 nat.SIZE. o. Superior view, nat. size.of another mdividual (Cat. No. 7482, U. S. Nat. Mus.), containingteeth (see figs. 5, 5a), which include the carnassials of both sides,and 7)12 and the posterior half of p^ of the left side.Description.?Size slightly less than that of C. occidcntalis, as thatspecies has been defined b}^ Miller, ^ but the tooth characters indicatean animal quite distinct from any of the true wolves. Its principaldifferences are seen in the greater relative depth of jaw, smaller canine,more simple p^ and p^, the presence of a posterior basal tubercle onp^, and in the relatively larger heel of the carnassial. The paraconid
' This species is named in honor of Mr. Raymond Armbruster, through whose efforts the CumberlandCave deposits were first brought to scientific notice.> Smiths. Misc. CoU., vol. 59, No. 15, 1912, p. 2.
ON A PLEISTOCENE CAVE DEPOSIT?GIDLEY. 99also is less expanded at base, with more perpendicular anterior face.The metaconid is larger and higher placed, while the protoconid isless broad and full, as seen from the inner side. The carnassials as awhole suggest those of the jackal, fox, or coyote rather than those ofthe wolf. The anterior functional premolars are relatively small andhave no accessory tubercles, while 7)4 is full}^ as heavy and robust asin the wolves and carries, besides the usual secondary cusp, an extraposterior basal cusp in addition to the cingulum, as in the jackals andcoyotes. In the wolves and dogs (see figs. 4, 4a) p^ has but onesecondary cusp and a cingulum heel, but P2 ^^^^ Pa usually have a well-developed posterior secondary cusp.
Fig. 3.?Canis armbrusteki. Cat. No. 7661, portion of right lower jaw.SIZE. a. Superior \qEW, nat. size. Outer view, 2-3 nat.ADDITIONAL NOTES ON THE LOWER TEETH OF THE CANIDS.The carnassials in the canids, except within the narrow limits ofindividual variation, are very constant in character and presentcertain modifications which for the most part readily determine thegroup to which they belong. These, taken together with the com-bined characters of the other teeth, are clearly diagnostic, not only ofthe various larger groups of the family, but even of groups nowincluded in the genus Canis. Thus in the true wolves and domesticdogs the heel of the lower carnassial is short (being less than one-fourth the total length of the crown) and is narrower than thetalonid; the paracoid is relatively large, with antero-posteriorlylengthened base, so that the anterior face slopes backward at a con-
100 PROCEEDINGS OF THE NATIONAL MUSEUM.
4 4aFigs. 4-8.?Lower fourth premolars and carnassials of Canids. All natural size. 4, 4a,Canis occidentalis.cat. no. 1006,u.s.n.m. 5,5a, Cants armbrusteri.cat. no.7482, u.s.n.m. 6,6a,Cants (lysciscus) latrans, cat. no. 3618, u.s.n.m. 7, 7a, Cants aureus, cat. no. 181500, u.s.n.m.8,8a, VuLPES, CAT. NO. 7183, u.s.n.m.
NO. 2014. ON A PLEISTOCENE CAVE DEPOSIT?GIDLEY. 101siderable angle; the body, or column, of the protocoid is full androunded, and the metacoid is greatly reduced, appearing in the livingspecies as a small tubercle on the inner posterior angle of the proto-coid near its base (see fig. 4). Similar and fully as importantdiflFerences are observed in the upper carnassial. The character ofthe premolars have been stated above, page 98.The constancy of these characters seems to mark the wolves anddogs as closely related members of a natural group, and tends tosupport the behef held by many, and recently especially expressedby Miller,^ that all the domestic breeds of dogs were originally derivedfrom some species of true wolf, and not from the jackal as has beenheld by some other authorities.The subgeneric distinction of the coyotes {Lyciscus) is well sub-stantiated by their tooth characters (see figs. 6, 6a). The teeth inthis group are all relatively narrower and less robust than in thewolves, while the carnassial has more the general proportions ofthose of the j ackal or fox (see figs. 7, 8) . Thus the heel is less reduced,with the two principal cusps more trenchant and more nearly sub-equal; the metaconid is more prominent; the paraconid shorter;and the bodies or columns of the paraconid and protoconid are lessfull and rounded, leaving the cutting blades of the trigonid muchsharper. The p^ has two posterior tubercles and a posterior basalcingulum, and p2 is usually simple.Both p2 and p^ are usually simple in the jackals while the crownsof all the premolars are relatively higher and shorter than are thoseof the coyotes.The foxes differ from the other canids in having relatively lowercrowned, smaller carnassials as well as in the greater relative lengthof the canine, as pointed out by Miller.-In Cuon and Lycaon the lower carnassials have a completelysingle-cusped, trenchant heel, which distinguishes them from all theother living canids. There are differences likewise in the uppercarnassials and other teeth, especially the molars, which separatethis group from Canis and seem to ally it to some of the extinctforms of the late Ohgocene, of the Temnocyon, or Hyxnocyon type.Some of the characters mentioned above have been recognizedand used by various investigators, but others of seemingly equalimportance seem to have been overlooked.Herewith is a fist of the principal characters of the lower teethwhich seem to be diagnostic for some of the groups of living canids.*Canis armhrusteri is also included.
' Catalogue of Mammals of Western Europe, 1912, p. 313.2 Idem, p. 326.* This is not a complete classification and is onl}' given to show some of the more important tooth char-acters, especially of the carnassials. The South American fox-like dogs are not included here.
102 PR0CEEDTX08 OF THE NATIONAL MUSEUM. vol.46.
a'. Lower camassials with trenchant heel. Cuon, Lycaon.a?. Lower camassial with heel more or less basin-shaped with two prominent cusps.6*. Heel of camassial reduced, narrower than trigonid, length contained in totallength of tooth about four times; main cusps of camassial heel, and an-terior pair of mj as well, verj^ unequal in size.c*. Protoconid and paraconid of camassial very large and full, subconic tosummit. Premolars robust, with but one posterior tubercle each onPj, p^, and usually on p^- Each premolar has besides a posterior basalshelf formed by the cingulum. Canis (wolves and domestic dogs),c^. Anterior premolars greatly reduced. Epicyon (extinct).h^. Heel of camassial less reduced, about equaling triconid in width, length ofheel contained in total length of tooth three and one-half times or less.Protoconid and paraconid of camassial less full and more bladelike.Innercusps of camassial heel and anterior pair of m^ relatively larger,sometimes almost subequal with the opposing outer cusps.c^ Molars and premolars relatively narrow or compressed, 'p^ with two pos-terior tubercles and basal cingulum, ^3 and usually p2 '"^^h a singleposterior tubercle. Canines large, but relatively longer than in thewolves. Canis {Lysciscus) (coyotes).c2. Premolars relatively shorter as in the wolves but with higher more pointedcusps. P4 with two posterior tubercles and a cingulum as in the coyotes,but with no posterior tubercles on the other premolars.Canis aureus, etc. (jackals).c*. Camassials proportionally small with relatively lower crowns, p^ with oneposterior tubercle and a posterior basal cingulum cusp, p2 and p^with posterior tubercles obsolete or wanting.(P. Tooth cusps prominent, trenchant; heels of premolars short.Vulpes (red fox).(F. Tooth cusps less prominent; heel of camassial broad with posteriorinner tubercle (entoconid) small; heels of premolars long.Alopex (Arctic fox).c*. Camassial moderately robust with short paraconid, and long, broad heel;premolars relatively long, low and simple, except p^, which has twowell-defined posterior tubercles and a posterior basal cingulum; canineand anterior premolars relatively small, jaw of great relative depth inregion of camassial. Canis armhrusteri (extinct).In the new species from the Cumberland Cave, C. armhrusteri, thegeneral form of the camassial is more Hke that of the coyote andjackal, and in the heavy three-cusped p^ combined with the relativelysmall, simple, single-cusped p^ and f^ and small canine it resemblesthe jackal despite its much greater size. The relatively deeper jawand broader more basinhke heel of the camassial would, however,scarcely warrant considering it a member of that group. It prob-ably represents an extinct group of dogs which when better knownmay be referred to a nQW genus.