PROCEEDINGS OF THE UNITED STATES NATIONAL MUSEUM hsueo' 8^5lV.vi OfJ^ hy the SMITHSONIAN INSTITUTIONU. S. NATIONAL MUSEUMVol. 91 Wathington : 1941 No. 3122A NEW FOSSIL CROCODILIAN FROM COLOMBIABv Ch.\iiles C. Mook^ Fossil remains of a gigantic crocodilian were collected by BrotherAriste (Dr. Maurice Rollot) between Neiva and the River Bache(Colombia) in 1920. The level is not recorded. Dr. J. B. Reeside,Jr., reports on the basis of invertebrates from nearby localities thatthe horizon is probably Lower Cretaceous. These remains consistof six fairly well preserved vertebrae, with parts of ribs, portionsof maxillary and dentar}' bones interlocked, several isolated piecesfrom the posterior portions of the right and left rami of the lowerjaw, and some fragments. The maxillary portion includes part ofthe alveolar series and was evidentlj' situated a short distanceposterior to the maxillo-premaxillary suture. These now constituteNo. 10889 of the collections of the United States National Museum.I wish to thank C. W. Gilmore, of that institution, for the privilegeof describing this material.The incomplete nature of this material makes determination of thei-elationships extremely difficult if not impossible. Several facts,however, may be noted. The vertebrae correspond in general char-acters and somewhat in size with the vertebra described by Gervaisas Dinosuchus terror. The indicated horizon is somewhat lower thanthe level of this form, which Gervais notes as "lower Tertiary orCretaceous."Comparison with the types of Punisaurus hrasiliensis Rodriguezand Brachygnathosuchys hraziliensls Mook shows clearly that theform described has no close relation with either. These species, whilegigantic, have relatively short and broad lower jaws, with largealveoli, while the form described has relatively long and slender lowerjaws and posterior teeth, at least, of relatively small size. ^ Contributions to the Osteology, Affinities, and Distribution of the Crocodilia, No. 35.406803?41 5- 56 PROCEEDINGS OF THE NATIONAL MUSEUM vol.91In view of these facts the material described is referred to a newspecies of the genus Dinosuchus Gervuis {non Holland) , which maybe called Dinosuchus neivensis, named for the city of Neiva nearwhich it was found.Genus DINOSUCHUS Gervais, 1876Generic charactei's?As Gervais never separated the generic char-acters from those of the species D. terror^ the following designationmay be given: Size gigantic, vertebrae procoelian and massivelyconstructed.Relationships.?The genera Dinosuchus Gervais, PurusaurusKodriguez, anc? Brachygnathosuchus Mook have been treated quitedifferently by recent authors. Nopcsa, in 1924, consideredBrachygnathosuchus to be a synonym of Pwrnsaurus, and Dinosuchu^sto be independent. Because of the latter interpretation he proposedthe name Phohosuchus for Holland's Deinosuchus. Mook, in 1934,considered Purusaurus to be a synonym of Dinosuchus^ andBrachygnathosuchus to be independent. Patterson, in 1936, consid-ered Brachygnathosuchus to be a synonym of Dinosuchus, andPurusaurus to be a synonym of Caiman of Spix.At the present time it appears most consistent with the incom-pletely known characters of these forms and with their geologiclevels to consider the Cretaceous Dinosuchus to be valid and inde-pendent, and to consider the upper Miocene or lower PliocenePurusaurus and Brachygnathosuchus to be valid and to be closelyrelated to Caiman. DINOSUCHUS NEIVENSIS, new speciesPlates 4-9Specific characters.?External mandibular foramen unusually largein proportion to the size of the jaw elements surrounding it, jawrelatively long and slender, posterior teeth relatively small and closetogether.Description of ^naterial.?Five maxillary alveoli are visible on thisspecimen. The first is large and is slightly longer than it is broad.The second is larger than the first. Its external border is incomplete ; consequently its proportions are difficult to determine. The lastthree alveoli are approximately equal to the first in size; they appearto be subcircular, although their borders are not entirely visible.Badly mutilated stumps of teeth are visible in these alveoli.The anterior and posterior ends of the lower jaw section that isattached to the portion of the maxillary noted above exhibit sectionsof alveoli 12 cm. deep and fragments of teeth of corresponding size.Another section of the right ramus was located much farther backthan the one noted above. The anterior end of the right external A NEW FOSSIL CROCODILIAN?MOOK 57mandibular foramen is located at the posterior end of this sectionand the posterior end of the alveolar row at the center of the superiorborder locates the position of the section in the ramus. Four alveoliAvith bases of teeth are clearly Aasible, and a fifth or last is somewhatobscure. These alveoli are much smaller than those of the maxil-lary section noted above, and their height, as indicated by theanterior surface of the section, is less than half that of the anterioimandibular teeth. The mandibular cavity, now indicated by matrix,was large, the bony substance being thin.The left ramus is represented by a larger section, about 48 cm. longand composed of two pieces that make clean-cut contacts with eachother. This section is entirely posterior to the alveolar row and in-cludes the external mandibular foramen, of which the superiorboundary is incomplete. The posterior end of this section is near theposterior end of the ramus immediately anterior to the glenoid surface.The sutures separating the elements of which this part of the jaw iscomposed are indistinct, the dentary, angular, and surangular bonesbeing almost indistinguishable from one another.The external mandibular foramen is unusually long and is not veryhigh. The exact relation between length and height cannot be madeout because of the incomplete superior border. On comparing thelength of this opening with that of an 84-cm. ramus of Crocodylusacutus, and assuming that the proportions between the total length andthe length of the foramen are the same in that species and the form nowdescribed, we estimate that the total length of the ramus would be 280cm., or about 9 feet. Comparison with a 32-cm. ramus of Caimancrocodilus indicates a total length of 172 cm., or about 5% feet, whichis more likely.One of the vertebral units is composed of the intercentrum of theatlas, most of the axis, and the proximal portions of the atlas and axisribs in natural positions. The atlas intercentrum is a broad, flat bone,much more distinctly bifurcated posteriorly than in C. acutus. Theatlas ribs attach to the bifurcations and their axes of breadth lie heloicthe axis and the. axis ribs. The atlas ribs are single headed, of course,and are considerably thickened where they attach to the atlas inter-centrum.The characters of the axis are not particularly distinctive except forthe size and strength of the processes to which the ribs are attached.The ribs themselves are distinctly two-headed, the upper element, ortuberculum, being slightly larger than the lower one, or capitulum.The shaft is slender and is situated on edge, at right angles to the posi-tion in which the atlas ribs are situated.Six other vertebrae are preserved, but none of them is complete.Two of these united together, with a fragment of a third, are cervicals. 58 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 91probably 4 and 5. The spines and the postzygapophj^ses are not pre-served. The prezygapophyses, diapophyses, and parapophyses areincompletely preserved. The centrum of the first vertebra of the pairis incomplete. That of the second is complete and is moderately long,rather low vertically and narrow posteriorly but broad anteriorly, ap-parently convex posteriorly, but the degree of convexity cannot bemade out. The prezygapophyses and diapophyses of this vertebraare incomplete, but enough of them is preserved to indicate that theywere very stout. There is a very small median hypapophysial keelnear the anterior end of the centrum. On the whole the vertebraeappear small for the size of the mandible. The capitular and tuber-cular ends of the left rib of the anterior of the two vertebrae are pre-served ; they are very stout, especially the tubercular process.MEASUREMENTS (IN MILLIMETERS)Length of two large contact pieces of left ramus of mandible ?192Maximum height of same 171Length of external mandibular foramen 265Height of same 56Length over four posterior alveoli, right ramus of mandible 82Height of maxillary and dentary fragments in place with each other 211Height of anterior mandibular tooth shov.ii in end of this fragment 94Breadth across atlas centrum, posterior end 88Length of atlas centrum 70Breadth across right atlas rib at proximal end 49Breadth across left atlas rib at proximal end 47Breadth across tuberculum end capitulum of right axis rib 43Breadth across tuberculum end capitulum of left axis rib '46Breadth across axis centrum posterior end 60Length of fifth (?) cervical centrum 83Breadth of fifth (?) cervical centrum anterior end 103Breadth of fifth (?) cervical centrum posterior end '70Breadth of fifth (?) cervical vertebra across prezygapophyses '97 ' Estimate. LITERATURE CITEDGERVAIS, I'ATJL.1876. Crocodile gigantesque fossile au Bresil. Journ. Zool., vol. 5, pp.233-236, 1 pi.Mock, Charles Cbaig.1921. Brachygyiathosuchus iraziliensis, a new fossil crocodilian from Brazil.Bull. Amor. Mus. Nat. Hist., vol. 44, pp. 43-49, 4 figs.1934. The evolution and classification of the Crocodilia. Journ. Geol., vol.42, pp. 295-304, 1 fig.NOPCSA, Fkanz Babon.1924. i'ber die Namen einiger brasilianischer fossiler Krolvodile. Ceniralbl.Min., Geol. und Pal., 1924, p. 878.P.\tte:rson, Bryan.1936. Caiman Jatirostri.s from the Pleistocene of Argentina, and a sum-mary of South American Cenozoic Crocodilia. Herpetologica, vol.1, pp. 43-54, 1 pi.RODRIGUES, JOAO BaRBOSA.1892. Les reptiles fossiles de lo vallee de I'Amazona. Vellosia. vol. 2(1885-88), pp. 41-56, 16 pis. U ? GOVrPNMEtjr PRINTING OTFICE 19