AMERICAN MUSEUM NOVITATES Published by Number 1301 THm AMERICAN MUSEUM OF NATURAL HIsTORY November 6, 1945 New York City

THE HYOID BONES IN PROTOCERA TOPS AND IN. PSITTACOSA URUS1 By EDWIN H. COLBERT INTRODUCTION

During the preparation of a skeleton of in order that the published knowledge con- Protoceratops, Mr. Otto Falkenbach of the cerning Protoceratops, contained in several American Museum Paleontological Labora- papers but especially in the study by tory discovered two small bones which Brown and Schlaikjer (1940), may be as when first seen appeared to be quite unlike complete as possible. At the same time, it any elements hitherto known in this dino- was felt that certain comparisons with other saur. At first it was thought that these dinosaurs of varying relationships to bones might be portions of the sternum, Protoceratops should be included in this but comparisons soon showed that such study. Such is the subject matter that will could not be the case. Indeed, various be treated in the following pages. comparisons with the known osteological The author wishes to express his appre- elements of Protoceratops confirmed the ciation to Mr. Falkenbach, who in the first impression, namely, that these were course of his skillful preparation of the bones hitherto undescribed, so the question specimen of Protoceratops made the discov- arose as to what these bones might be. ery of these new bones. To Dr. Gregory Several possibilities were considered and must go the credit of making the suggestion several opinions were sought, with the re- as to their being hyoids, and to him my full- sult that Dr. William King Gregory of the est acknowledgments are hereby expressed. American Museum suggested that these Subsequently a portion of another bone was might be hyoid elements. The more this found, and this may represent an addi- possibility was investigated, the more tional element in the hyoid apparatus of likely it seemed to be. All subsequent in- Protoceratops. The translation from FiXr- vestigations have tended to confirm this bringer was made by Dr. Otto Haas. view. Finally credit should go to Mr. John C. It has been felt desirable to bring out a Germann, Staff Artist, for the excellent il- description of these supposed hyoid bones, lustrations that appear in this paper.

DESCRIPTION AND DISCUSSION Before making a formal description of the hyoids in modern reptiles was set forth these supposed hyoids of Protoceratops, it by Fuirbringer in his monograph on the min,y be well to consider briefly the homolo- hyoids of the vertebrates (1922). A trans- gies of the hyoid bones in the tetrapods, lation of this statement follows: and their development particularly in mod- "As has been known in the main for a ern reptiles. long time, the hyoid of living reptiles con- A very illuminating statement describ- sists, at its highest stage of development, ing the development and the homologies of of an unpaired body, the corpus (Copula; 1 Publications of the Asiatic Expeditions of the basihyal and basibranchial) [basibranchial American Museum of Natural History, Contribution No. 148. and hypobranchial according to Goodrich] 2 AMERICAN MUSEUM NOVITA TES [No. 1301 tapering in a forward direction into a nar- mellar apparatus of the middle ear, also, rower process, the processus lingualis (proc- an element in the hyoid arch, but this con- essus entoglossus), which varies in length nection has been lost in most of the reptiles and is bound into the posterior part of the and has undergone considerable redqction tongue, and of three pairs of horns con- in many of them. This horn is very sig- nected laterally in various ways with the nificant morphologically and therefore has corpus, an anterior pair, the cornu hyale, been named the cornu principale by Gaupp, a middle pair, the cornu branchiale I, and a but it is inferior in size and strength to the

C B Ch1

Ch CbI Cb2 C Cb2 Cblt

A E D )CJkCh

Cb2,

Fig. 1. A comparison of the hyoids in the living reptiles. A, Crocodilia: Crocodilus, original. B, Rhynchocephalia: Sphenodon, after Furbringer. C, Squamata: Lacerta, after Goodrich. D, Squamata, Vipera, after Ftrbringer. E, Chelonia: Clemmys, after Forbringer. Not to scale. Key to abbreviations: C-corpus, including the hypohyal or entoglossal. Ch-ceratohyal and epi- hyal; cornu hyale. Cbl-ceratobranchial I; cornu branchiale I. Cb2-ceratobranchial II; cornu branchiale II. posterior pair, the cornu branchiale II, following horn. This horn, the cornu and these are derived from the ordinary branchiale I, which is derived from the first hyoid arch and the first two gill arches. branchial or third visceral arch, is dorsally In the more primitive forms the cornu hy- free and it proves by its permanence and ale is found to be connected with the colu- by the fact that it is ossified when all other 1945] HYOID BONES IN PROTOCERATOPS AND PSITTACOSAURUS 3 portions of the hyoid are cartilaginous, to are pictured. This diagram will, of course, be the most prominent part of the hyoid supplement the summary written by Fiur- complex. The cornu branchiale II, derived bringer and presented above. Homologies from the second branchial or fourth visceral are derived from a comparison with the con- arch, is considerably reduced, and fre- dition in fishes. quently in many reptiles it is split up into a Various names have been applied to the ventral and a dorsal portion, while it may hyoid elements shown in this diagram. It disappear completely in some forms. This may be well, therefore, to include at this horn, too, is as a rule dorsally free, but in place a table to show the system of nomen- some isolated cases it is connected with the clature used for the hyoid elements by dorsal end of the cornu hyale. Starting Futrbringer in 1922, as compared with the with the highest level of development, the system used by Goodrich in 1930, and that hyoid of many reptiles, most of which may followed by Camp in his classification of be considered as advanced forms, is subject the lizards, 1923, and more recent papers.

FtRBRINGER GOODRICH CAMP Corpus Corpus or Copula Corpus Entoglossal Entoglossal process Hypohyal Cornu hyale Hyoid cornu Ceratohyal Epihyal Cornu branchiale I First branchial cornu Ceratobranchial I Epibranchial I Cornu branchiale II Second branchial cornu Ceratobranchial II Epibranchial II to reduction. This may start sometimes In the present work Camp's terminology with the first horn, the cornu hyale, some- is used. times with the last one, the cornu bran- In this present discussion our interest, as chiale II, or it may take place simulta- far as recent reptiles are concerned, is cen- neously at the anterior and posterior ends. tered upon the Crocodilia since these ani- "(For a more detailed survey, reference mals, most closely among the living forms, is made to Gaupp's excellent synthesis of approach the dinosaurs in their morphologi- the ontogeny and phylogeny of the sound cal development. conducting apparatus in vertebrates as The hyoids of the Crocodilia are de- well as of the evolution of the cranial and scribed at some length by Fuirbringer, who the hyobranchial skeleton.) shows that in these reptiles the hyoids con- "Simultaneously the hyoid bone consti- sist of a large corpus, ventrally convex, tutes in various ways the ventral basis from the side of which project large, well- and, when well developed, the support for ossified first ceratobranchials. The corpus the respiratory organs (larynx, i.e., cricoid, is cartilaginous and broader anteriorly with adjacent cartilages, the trachea). than posteriorly. In its dorsal concavity The connection between both is sometimes are the cricoid and arytenoid cartilages. loose, sometimes solid, and it may exist According to Fiirbringer, the existence between hyoid and cricoid or between hyoid of a ceratohyal is not certain in the Croco- and trachea, with all transitions from the dilia. It would seem probable, from the cricoid to the trachea" (Futrbringer, 1922, ontogenetic work of Parker, that it was p. 6). suppressed at an early stage of develcp- For the purpose of orienting ourselves ment. and avoiding a long, detailed discussion, The first ceratobranchial, says Fur- the manner in which the hyoid apparatus bringer, projects from the lateral border of is developed in the modern reptiles may be the corpus and is directed somewhat lat- shown by a diagram in which typical ex- erally and somewhat dorsally. This is a amples of the hyoids in turtles, rhynchoce- very strong, highly ossified element, which phalians, crocodilians, lizards, and snakes often shows, according to the age of the 4 AMERICAN MUSEUM NOVITATES [No. 1301

Fig. 2. The hyoid apparatus of Crocodilus porosus. Ventral view, natural size. Abbreviations as in figure 1. animal or the species, a longer or shorter Crocodilia, condensed and paraphrased cartilaginous epiphysis at the free end. above, with a few notes regarding the hy- In young embryos this element is cartilagi- oids of Crocodilus porosus as shown by a nous, but it begins to ossify during embry- specimen at hand. Evidently this came onic life. from a fairly large animal, so that ossifica- The second ceratobranchials are con- tion has proceeded to a considerable extent. tained in the two posterior corners of the The corpus or copula, however, is com- hyoid body or corpus. It might be helpful pletely cartilaginous. It is broad, and its at this place to supplement briefly Fur- lateral edges are extended; these may rep- bringer's remarks as to the hyoids in the resent in part at least the ceratohyals, 19451 HYOID BONES IN PROTOCERATOPS AND PSITTACOSAURUS 5

As Fulrbringer has shown, this is a debatable thickened edge of each bone is the lateral question upon the basis of such evidence as one, and the thin, sinuous curved edge the is now available to us. median one. This would accord with the The corpus extends imperceptibly into condition in the Crocodilia, in which the the second ceratobranchials, which are median edges of the second ceratobran- bony. However, the line of separation be- chials thin out so that they merge gradually tween each bony cornu and the cartilagi- with the cartilage of the corpus. Further nous body of the copula is not definite, and support for this orientation of the bones is the two merge into each other. Along the edges the bony cornua are invested by car- tilage. The first ceratobranchials protrude from the corpus near the anterior ends of the second ceratobranchials. These first A ceratobranchials are completely bony and L-shaped, and their anterior ends, articu- lating with the corpus, are rounded into "heads." At the angle of the "L" a distinct point is produced. Posteriorly these bones show cartilaginous connections. This speci- men is illustrated in figure 2. With this orientation of the subject in mind, we may now consider the fossil ma- terials that are the basis for this study. Of first importance are the two paired bones that were associated with the Protoceratops skeleton, A.M.N.H. No. 6471. These bones are shown in figure 3, and reference to this illustration will give an idea as to their form to much better ad- vantage than may be had from any word description. These bones are thin and es- sentially flat, although at their expanded ends they are on one side slightly concave and on the other slightly convex. Each bone is expanded laterally at either end and constricted in the middle, and one side is in the form of a simple, flattened curve, with a somewhat thickened edge. The other edges of the bone are all very thin ex- cept at the ends near the juncture of the rounded end surfaces with the simply curved side. Here the edges are -thick- Fig. 3. A, The hyoid bones (probably the second ceratobranchials) of Limnoscelis, as ened. preserved in position. After Williston, not to It is supposed by analogy with the Croco- scale. B, The supposed second ceratobranchials of Protoceratops. A.M.N.H. No. 6471. Ventral dilia that these flattened bones in Proto- view, natural size. In both drawings the bones ceratops are the second ceratobranchials. are oriented with their assumed anterior ends up. Assuming this to be so, how are these ele- ments to be oriented? This is a difficult afforded by the hyoids of Limnoscelis, fig- question, because unfortunately these bones ured by Williston in 1911. Williston found are shaped somewhat differently from the that in Limnoscelis, a very primitive Per- second ceratobranchials in the Crocodilia. mian cotylosaur, there was a pair of bones It would seem most probable, however, which he interpreted as hyoids, and which that the long, simply curved, and somewhat he described as follows: 6 AMERICAN MUSEUM NOVITATES [No. 1301

"In each skeleton there is a pair of in Limnoscelis, and for the time being this bones .... of the nature of which I am not will have to serve as the guide to their posi- fully satisfied, though there would seem to tion. be little doubt but that they are unusually Finally, it may be rather safely assumed large hyoids. They are . . . . greatly ex- that the convex surface of each bone in panded on their distal, thin end, with a Protoceratops is the ventral surface, and somewhat curved and narrowed shaft that the concave surface faced dorsad, as deeply concave in outline on one side, less is the case in the crocodiles. so on the other, thickened and truncate for In addition to these two flat bones, still articulation at the proximal end. The two another small bony fragment of puzzling bones in each specimen lie with the thin relationships was found in association with ends nearly in apposition, as though they the same Protoceratops, namely, A.M.N.H. had joined each other in life" (Williston, No. 6471. This is a thin, forked bone, of 1911, pp. 36-37). which it would seem that most of the length

Fig. 4. A, Portion of a bone found in association with Protoceratops, A.M.N.H. No. 6471, and identified as possibly a part of the first ceratobranchial. B, Outline of the first ceratobranchial of Crocodilus porosus. That portion of the bone supposed to be homologous with the Protoceratops bone above (A) is shaded. External lateral views, both natural size. These bones, which are shown in outline of the two forks has been lost. This bone in figure 3, bear a considerable resemblance has a sort of small hook or process opposite to the supposed hyoids of Protoceratops, as one of the forks. Here again, a figure gives may be seen. Since the hyoids in Limno- a much more adequate idea as to the struc- scelis are in position, with the thin,.sinuous ture of the element than does a word de- edges opposite each other, it is logical to scription. assume that the hyoids in Protoceratops Frequent attempts to homologize this were oriented in a similar manner. bone as a portion of a rib have proved to As to the question of the anterior and be quite unsatisfactory. Nor is there any posterior ends of the supposed hyoids in other element known in the skeleton of Protoceralops, this is a difficult problem to Protoceratops with which this bone is solve. The bones have been oriented as closely comparable. Therefore it is sug- shown in figure 3, with the full realization gested that this may be a portion of a first that further information at some future ceratobranchial. In Crocodilus porosus the date may require their reversal, end for first ceratobranchial, as already mentioned, end. At any rate, the orientation of the is an L-shaped bone, with a point developed bones as shown seems to correspond most at the angle of the L. Perhaps this fossil clos31y with the orienitation of the hyoids fragment comes from the same general re- 1945] HYOID BONES IN PROTOCERATOPS AND PSITTACOSAURUS 7 gion of the angle in the first ceratobranchial right angles to this plane, extends from of Protoceratops with a possible position as near the posterior border of the right flat shown in the figure. This may be an erro- bone to a point beneath the basioccipital neous interpretation, but no other satis- of the skull. This latter bone is closely ap- factory explanation suggests itself at the pressed to the cervical vertebrae. present time. By analogy with the Crocodilia the flat In making this study of the hyoids in bones are here identified as the flattened Protoceratops, a reexamination of the dino- second ceratobranchials, while the elon- saur Psittacosaurus was made. Psittaco- gated bones are considered as the first cera- saurus, described by Osborn in 1924, comes tobranchials. from the Ondai Sair formation of Mongolia, which makes this form somewhat older than Cbl Protoceratops. Moreover, there is every reason to think that Ps'ittacosaurus is very close to, if not actually on, the ancestral line leading to the Ceratopsia, a point that was made by Gregory in 1927, but that has never received the attention it deserves. Indeed, it is proposed at this place to in- clude Psittacosaurus in the suborder Cera- topsia, rather than in the Ornithopoda where it was placed by Osborn, who has been generally followed by subsequent au- thors. Despite Osborn's excellent preliminary paper on this genus, Psittacosaurus has never been fully described. Consequently many various interesting feat,ures charac- terizing this important little dinosaur are still to be recorded in the literature. In the type skull, A.M.N.H. No. 6254, there is a mass of matrix and bones occupying the space between the rami of the mandible, B and because of the information that might be obtained with regard to this question of the hyoids in dinosaurs it was thought expe- Fig. 5. Hyoid bones in Psittacosaurus. A, dient to develop this region in the type A.M.N.H. No. 6254. B, A.M.N.H. No. 6534. specimen. The task was undertaken and Ventral views, both figures natural size. Key accomplished by Mr. Falkenbach with his to abbreviations: Cbl-ceratobranchial I. usual skill. Cb2-ceratobranchial II. As a result of Mr. Falkenbach's further Strictly speaking, the supposed second preparation of the specimen, certain bones ceratobranchials are neither exactly "flat" came to light that seem without much nor truly bony. Careful examination un- doubt to be hyoids. These occupy the po- der the microscope indicates, by reason of sition in which hyoids might be expected, the manner in which they are preserved, and they can be eliminated from considera- that these elements were never fully ossi- tion as cervical ribs because of their shape fied in life but rather were cartilaginous. and size. Therefore these elements may represent They consist of two flattened bones, side the second ceratobranchials and to some by side between the mandibular rami, and extent portions of the corpus, but here a two long, rod-like bones. Of the latter, definite conclusion is not possible. They one lies between the two flat bones and in are somewhat convex ventrally, as might the same plane with them; the other, at be expected, and in each the postero-in- 8 AMERICAN MUSEUM NOVITATES [No. 1301

C

Cbl

C

B

Cbl

Fig. 6. Restoration of the hyoid apparatus in (A) Psittacosaurus and (B) Protoceratops. Ventral views, natural size. Key to abbreviations: C-corpus. Cbl-ceratobranchial I. Cb2-cerato- branchial IL. 1945] HYOID BONES IN PROTOCERATOPS AND PSITTACOSAURUS 9 ternal corner is lower than the major por- viously homologous with the first cerato- tion of the element. It is an interesting branchials as they are developed in the fact that a small portion of this postero- Crocodilia; indeed this conclusion seem- internal corner shows a high degree of os- ingly was reached by Jaeckel, because sification in just the region where os- he speaks of the "sogenannten Zungenbein- sification of the second ceratobranchial horner," a term that has been applied to might be expected. In shape each of these the first ceratobranchials by several Ger- bones has a straight, internal, median edge man writers. and a curved, external, lateral edge, and In this connection it might be pointed the antero-internal and postero-internal out that Gilmore, in 1925, described very corners are somewhat produced. briefly some hyoid bones associated with a The supposed first ceratobranchials are skull and lower jaws of Camarasaurus, in elongated, rod-like bones and are fully ossi- the collections of the Carnegie Museum, fied, as might be expected. These bones Pittsburgh. are comparatively straight, and their ends "That there is a well developed hyoid are rounded. apparatus in the Sauropod dinosauria is Another specimen of Psittacosaurus, shown by the preservation in the matrix A.M.N.H. No. 6534, also shows two loose beneath the lower jaws of three rod-like bones in the basicranial region that are simi- bones, which undoubtedly represent ele- lar in shape 'to the long bones described ments of the hyoid arch. The longest of above in A.M.N.H. No. 6254 and for this these measures 165 millimeters in length. reason are identified as the first cerato- Two of them are paired and probably repre- branchials of the hyoid complex. These sent the thyrohyal bones" (Gilmore, 1925). bones, as exposed in the specimen, are The two long paired bones mentioned by shown in figure 5. Gilmore are almost certainly the long, bony, Jaeckel, in his paper of 1913 describing first ceratobranchials, the homologues of Plateosaurus, mentions and figures two the thyrohyals in the mammals. What large hyoid bones which were found be- the third bone may be is a debatable ques- neath the lower jaws in one of the speci- tion. From Gilmore's figure (pl. 15) this mens under his observation. These are appears to be a rod-like bone, considerably very long, rod-like bones, somewhat broad- shorter than the paired bones. Perhaps it ened at their posterior ends. They are ob- is a second ceratobranchial.

CONCLUSIONS From the evidence at hand it would ap- ossified and are generally the bones pre- pear that the hyoid bones in the dinosaurs served, were attached at one end near the show. a development rather similar to that antero-lateral borders of the second cera- in the modern Crocodilia. In other words, tobranchials and were directed postero- it is probable that in the Saurischia and the laterally into the region of the throat. If Ornithischia the hyoid apparatus took the this picture of the hyoid development in form of a broad corpus, perhaps for the the dinosaurs is correct it is in accord with most part cartilaginous, in the posterior other morphological characters which show portion of which there were enclosed the the relationship between these two extinct paired, flattened, second ceratobranchials. orders and modern Crocodilia. These elements were ossified to a greater or In figure 6 the hyoid apparatus is re- lesser degree. From this broad, plate-like stored in Psittacosaurus and in Protoceratops fusion of the corpus and cornua there pro- on the basis of the known condition in truded the long, rod-like, first ceratobran- Crocodilus. chials. These elements, which were highly 10 AMERICAN MUSEUM NOVITATES [No. 1301

REFERENCES BROWN, BARNUM, AND ERICH M. SCHLAIKJER per. Geol. und Paleont. Abhandl., 1940. The structure and relationships of new ser., vol. 15, no. 2, pp. 139-179, Protoceratop8. Ann. New York Acad. pls. 7-13. Sci., vol. 40, art. 3, pp. 133-265, pls. JAEKEL, OTTO 1-13. 1913. Uber die Wirbeltierfunde in der oberen CAMP, CHARLES Trias von Halberstadt. Palaeont. Zeit- 1923. Classification of the lizards. Bull. schr., vol. 1, no. 1, pp. 155-215, pls. Amer. Mus. Nat. Hist., vol. 48, pp. 3-5. 289-481. OSBORN, HENRY FAIRFIELD FtRBRINGER, MAX 1924. Psittacosaurus and Protiguanodon: two 1919. Uber das Zungenbein der Reptilien. Lower Cretaceous iguanodonts from Bijd. Dierk., no. 21, pp. 195-212. Mongolia. Amer. Mus. Novitates, 1922. Das Zungenbein der Wirbeltiere ins- no. 127, pp. 1-16. besondere der Reptilien und Vogel. OWEN, RICHARD Abhandl. Heidelberg Akad. Wiss., div. 1850. Communications between the cavity B, vol. 11, 164 pp., 12 pls. of the tympanum and the palate in the GILMORE, C. W. Crocodilia. Phil. Trans. Roy. Soc. 1925. A nearly complete articulated skeleton London, pp. 521-527, pls. 40-42. of Camarasaurus, a saurischian dino- 1859. On some reptilian fossils from South saur from theDinosaur National Monu- Africa. Proc. Geol. Soc. London, vol. ment, Utah. Mem. Carnegie Mus., 16, pt. 1, pp. 49-63, pls. 1-3. vol. 10, no. 3, pp. 347-384, pls. 13-17. PARKER, W. K. GOODRICH, EDWIN S. 1883. On the structure and development of 1930. Studies on the structure and develop- the skull in the Crocodilia. Trans. ment of vertebrates. London, Mac- Zool. Soc. London, vol. 11, pt. 9, millan and Co., pp. 440-448. pp. 263-310, pls. 62-71. GRANGER, WALTER, AND WILLIAM K. GREGORY SUSHKIN, P. P. 1923. Protoceratops andrew8i, a pre-ceratop- 1927. On the modifications of the mandibu- sian dinosaur from Mongolia. Amer. lar and hyoid arches, and their rela- Mus. Novitates, no. 72, pp. 1-9. tions to the brain case in the early GREGORY, WILLIAM K. Tetrapoda. Palaeont. Zeitschr., vol. 1927. The Mongolian life record. Sci. 8, pp. 263-321. Monthly, vol. 24, pp. 169-181. WILLISTON, SAMUEL W. GREGORY, WILLIAM K., AND CHARLES C. MooK 1911. American Permian vertebrates. Chi- 1925. On Protoceratops, a primitive ceratop- cago, University of Chicago Press, 145 sian dinosaur from the Lower Creta- pp., 38 pls. ceous of Mongolia. Amer. Mus. Novi- ZAVATTARI, EDOARDO tates, no. 156, pp. 1-9. 1908. Materiali per lo studio dell' osso ioide HUENE, F. VON dei Sauri. Atti. Accad. Sci. Torino, 1926. Vollstandige Osteologie eines Plateo- vol. 43, pp. 1138-1145. sauriden aus dem schwfibischen Keu-