sign in sign up
<<
Home , Aceratheriinae, Acerorhinus, Aphelops, Chilotherium, Coelodonta, Elasmotherium

AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 3143, 25 pp., 2 figures, 5 tables August 14, 1995

Cladistic Analysis of the Family Rhinocerotidae (Perissodactyla)

ESPERANZA CERDENO'

ABSTRACT A cladistic analysis ofthe family Rhinocerotidae otinae and are questioned. Within encompassing 45 taxa and 72 characters (referred the former, the elasmotheriines are separated into to craniomandibular [20], dental [26], and post- two groups: a new subtribe, Iranotheriina, is pro- cranial [26] features), and using Hyrachyus and posed, while and Ninxiatherium two subfamilies of Hyracodontidae as outgroups, appear more closely related to and discovered 104 equally parsimonious cladograms. . The content of the subtribe Rhino- When the analysis was run with Hyrachyus as the cerotina is revised. The tribe Teleoceratini is re- only outgroup (removing the hyracodontid sub- moved from the Rhinocerotinae and included families), only six equally parsimonious trees were within the subfamily Aceratheriinae. A new ac- discovered. The discovered trees of the two anal- eratherine tribe, Alicornopini, is proposed for Al- yses are compared, which suggests a reinterpre- icornops, , and Chilotheridium. The tation ofthe phylogenetic relationships within the analysis indirectly supports some synonymies pre- Rhinocerotidae. The subfamilies Diceratheriinae viously suggested, such as "Begertherium" = His- and Menoceratinae are not supported, as the gen- panotherium, and "Dicerorhinus" schleiermacheri era included within them appear as a paraphyletic = schleiermacheri. group. Phylogenies of the subfamilies Rhinocer-

INTRODUCTION The investigation offossil rhinocerotids has data on other groups of perissodactyls (e.g., long been considered a difficult task because horses), investigations of fossil of their great intraspecific variation, as well are relatively scarce, at least at the family as the general homogeneity of the group at a level. Some of the most recent monographic higher taxonomic level. Unlike the body of works on rhinos are those of Guerin (1980),

I Fulbright Postdoctoral Scholar, Department of Vertebrate Paleontology, American Museum of Natural History; Departamento de Paleobiologia, Museo Nacional de Ciencias Naturales, Jose Guti6rrez Abascal, 2; 28006 Madrid, Spain.

Copyright © American Museum of Natural History 1995 ISSN 0003-0082 / Price $3.40 2 AMERICAN MUSEUM NOVITATES NO. 3143

TABLE 1 Genera and of Hyracodontidae and Rhinocerotidae Studied, Geographical Distribution AF = Africa; AS = Asia; EU = Europe; NA = . HYRACODONTIDAE S. kirchbergensis EU Hyracodon Leidy, 1856 Coelodonta Bronn, 1831 H. nebraskensis NA C. antiquitatis EU/AS H. leydianus NA Crusafont & Villalta, 1947 Ardynia Matthew & Granger, 1923 H. matritense EU/AS A. praecox AS "Begertherium" Beliajeva, 1971 A. kazachstanensis AS "-B." grimmi AS Prohyracodon Koch, 1897 "Beliajevina" Heissig, 1974 P. meridionale AS "B." tekkayai AS P. orientale EU Ringstrom, 1924 Eggysodon Roman, 1911 L mongoliense AS E. osborni EU I. cf. longirhinus AS RHINOCEROTIDAE Elasmotherium Fischer, 1808 Aymard, 1886 E. shansiense AS R. fdlholi EU E. lagreli AS Protaceratherium Abel, 1910 Gaindatherium Colbert, 1934 P. albigense EU G. browni AS P. minutum EU G. vidali AS P. platyodon EU Gray, 1867 P. fahlbuschi EU C. neumayri EU P. mirallesi EU C. simun AF Pleuroceros Roger, 1898 Subhyracodon Brandt, 1878 P. pleuroceros EU S. occidentalis NA Dietrich, 1931 S. mitis NA D. lemanense EU S. tridactylum NA D. aurelianense EU Marsh, 1875 Roger, 1904 D. armatum NA B. brachypus EU D. niobrarense NA D. annectens NA Heissig, 1974 P. germanicus EU Penetrigonias Tanner & Martin, 1976 P. douvillei EU P. hudsoni NA P. dakotensis NA Alicornops Ginsburg & Guerin, 1979 A. simorrense EU Trigonias Lucas, 1900 A. alfambrense EU T. osborni NA T. wellsi NA Hoploaceratherium Ginsburg & Heissig, 1989 H. tetradactylum EU Amphycaenopus Wood, 1927 A. platycephalus NA Kaup, 1932 A. incisivum EU Menoceras Troxell, 1921 M. arikarense NA "Mesaceratherium" Heissig, 1969 Floridaceras Wood, 1966 "M." gaimersheimensis EU F. whitei NA Lartetotherium Ginsburg, 1974 L. sansaniensis EU "F:AM 95544" Prothero (in press) NA Dicerorhinus Gloger, 1841 Peraceras Cope, 1880 "D." schleiermacheri EU P. superciliosum NA "D." pikermiensis EU/AS P. profectum NA P. hesei NA Stephanorhinus Kretzoi, 1942 Cope, 1873 S. miguelkrusafonti EU A. megalodus NA S. megarhinus EU A. malacorhinus NA S. etruscus EU A. mutilus NA S. hemitoechus EU 1995 CERDENO: RHINOCEROTIDAE 3

TABLE 1-(Continued) Hatcher, 1894 Linnaeus, 1758 T. medicornutum NA R. unicornis AS T. meridianum NA R. sondaicus AS T. major NA Diceros Gray, 1821 T. fossiger NA D. bicornis AF T. proterum NA T. hicksi NA

Cerdefno (1989, 1992), Heissig (1989), For- However, duplication of these results is not telius et al. (1993), and Prothero (in press). possible because published data matrices are During the last decade, several cladistic phy- not available in the three cladistic analyses logenetic analyses ofrhinoceroses (sensu lato) noted above. The phylogenetic hypotheses have provided new hypotheses on their re- proposed by these authors are reflected in the lationships (Heissig, 1981, 1989; Groves, current classification of rhinoceroses (Proth- 1983; Prothero et al., 1986). The work of ero and Schoch, 1989). The research reported Prothero et al. (1986) is the most ambitious here is an attempt at a more rigorous phy- attempt; it includes the whole superfamily logenetic analysis of the family Rhinocero- Rhinocerotoidea and provides a good com- tidae. pilation of previous phylogenies of rhinos.

MATERIAL AND METHODS Specimens ofmost genera were studied di- bridge); National Museum of Natural History- rectly by the author for more than 10 . Smithsonian Institution (Washington). The material is housed in the following in- Extant species were reviewed from the col- stitutions: lection in the Department of Mammalogy, AMNH. Table 1 summarizes the species of Spain: Museo Nacional de Ciencias Naturales the reviewed by the author and indi- (MNCN), Museo del ITGE, Departamento de cates their geographical distribution. Paleontologia de la Universidad Complutense, Taxa in quotation marks correspond to and Museo Arqueologico Nacional (Madrid); those forms previously considered or sug- Instituto de Paleontologla M. Crusafont (Sa- gested to be synonymous, although a general badell), Museo Paleontologico (Valencia), So- consensus has not been reached: Hispano- ciedad de Ciencias Aranzadi (San Sebasti'an). therium = "Begertherium" and "Beliajevi- Portugal: Departamento de Estratigrafia e Paleo- na," Antunes and Ginsburg (1983)-Cerde-no biologia, Universidade Nova de Lisboa. (1989)-Fortelius and Heissig (1989). "Dicer- France: Institut de Paleontologie, MNHN (Paris; orhinus" schleiermacheri = ?Lartetotherium Departement des Sciences de la Terre de l'Uni- schleiermacheri-Cerdeino (1992). "Mesa- versite Claude Bernard (Lyon). ceratherium gaimersheimensis" = Acerathe- Netherlands: Instituut voor aardwetenschappen, rium paulhiacensus-Bonis (1973), "Mesa- Rijksuniversiteit (Utrecht). ceratherium gaimersheimensis" = Protacera- Germany: Universitlits-Institute fur Paliiontolo- therium gaimersheimensis, author's opinion. gie und historische Geologie (Munich), Senck- One specimen in particular ("F:AM 95544," enberg Museum (Frankfurt). a in the Frick collection, AMNH) has Italy: Museo di Geologia e Paleontologia (Flor- not been formally defined taxonomically, al- ence). though Prothero et al. (1986) and Prothero USA: American Museum of Natural History (in press) consider it to be a new genus. For (AMNH) (New York); Museum ofComparative "Begertherium" I have considered the Zoology (MCZ) (Harvard University, Cam- species B. borrissiaki as well as B. grimmi, 4 AMERICAN MUSEUM NOVITATES NO. 3143 following the suggestion of Fortelius and ACKNOWLEDGMENTS Heissig (1989). "Dicerorhinus" pikermiensis (Geraads, 1988) seems to be a species very I am very grateful to Dr. R. H. Tedford for close to "Dicerorhinus" schleiermacheri, but the facilities he provided at the AMNH. I it is not included in the analysis because I also thank Drs. M. C. McKenna, R. H. Ted- have examined just a few ofits remains, and ford, K. Heissig, C. Guerin, S. Roig, and G. Geraads (1988) did not establish clear skel- C. Gould for the critical reading ofthe manu- etal characteristics for it. Groves (1983) dis- script, and their useful comments. The Ful- cussed the affinities of several spe- bright scholarship of the author is included cies assigned to the genus Dicerorhinus. Spe- in the Research Project PB91-0082, DGI- cies such as D. leakeyi (Hooijer, 1966), D. CYT, Spain. orientalis (Ringstrom, 1924), or D. piker- miensis (Geraads, 1988) are closely similar CHARACTER ANALYSIS to Lartetotherium sansaniensis and "Dicer- Dental morphology has been classically orhinus" schleiermacheri, and their detailed used as the basis for taxonomical studies, study would probably show their closer re- while postcranial characters have often been lationship with the genus Lartetotherium, as neglected. This analysis is mostly based on already suggested (Groves, 1983; Cerdeino, craniodental features, although a number of 1992). postcranial elements are also considered (ta- Some poorly defined genera such as Ken- ble 2) in order to achieve a more complete yatherium (Aguirre and Guerin, 1974), Shen- anatomical characterization. All segments of nongtherium (Huang and Yan, 1983), and the limbs were considered; many characters Tesselodon (Yan, 1979), based solely on den- refer to the general shape of the bones, but tal material, were excluded from the study. some elements, like the astragalus, the cal- This analysis was executed using Hennig86, caneum, or the scaphoid, provide good fea- version 1.5 (Farris, 1988). The evolution of tures in the articular facets that can be con- characters has been examined with the CLA- sidered separate characters. Detailed expla- DOS program (Nixon, 1992). The close phy- nation ofcommonly used terminology in rhi- logenetic relationships between the families noceros anatomy, as well as the general shape Rhinocerotidae and Hyracodontidae (Proth- of teeth and bones, and the position of the ero et al., 1986, 1989) led to the choice of different articular surfaces, can be obtained the latter as the outgroup, represented by the in Heissig (1972), Guerin (1980), and Cer- subfamilies Eggysodontinae (= Allaceropi- de-no (1989). Some characters are present with nae) and Hyracodontinae (Heissig, 1989). different character states within species ofthe Heissig (1989) considered the subfamilial same genus (sometimes they even vary with- name Allaceropinae to be valid for the Eg- in a species), which is reflected by question gysodontinae even though Allacerops Wood, marks in the matrix (table 3). The amount of 1932, was synonymized Eggysodon Roman, missing data is also increased by the lack of 1911. Nevertheless Eggysodontinae was first knowledge ofsome characters ofcertain taxa. established in 1923 (Breuning, 1923), thus Data are missing for several postcranial fea- having priority. With respect to the family tures of some genera because published de- Hyracodontidae, I follow Heissig (1989) in scriptions are not sufficiently detailed, even considering Hyracodon and Ardynia within when the bones are known, and I have not the Hyracodontinae, and Eggysodon and observed them. Detailed characters and char- Prohyracodon within Eggysodontinae (see acter states can be obtained in table 2, and Dashzeveg [1991] for another perspective). their distribution among the 45 taxa is given The primitive Rhinocerotoidea Hyrachyus in table 3. Most characters are ordered (38 has also been added as a third outgroup in are binary); only 6 of the 34 multistate char- order to clarify the polarity of some charac- acters were left unordered (1, 2, 11, 47, 50, ters that appear with different character states and 60), as directionality could not be estab- or with missing data among the hyracodontid lished. The polarity of characters was estab- subfamilies. The matrix has also been run lished following the outgroup comparison with Hyrachyus as the only outgroup, and the methodology (Watrous and Wheeler, 1981; respective results are compared. Nixon and Carpenter, 1993). 1995 CERDENO: RHINOCEROTIDAE 5

With respect to the evolution ofcharacters, noceros and three "elasmotherine" forms; the the present cladistic analysis makes it evident derived state 2 is an autapomorphy of Elas- that many of them are quite homoplastic. motherium. Despite this behavior that diminishes some 4. The ossification of the nasal septum is a global value of these characters, they are synapomorphy at node 20 (fig. 1), and the valuable in defining some lesser groups with- most derived state (septum totally ossified) in the whole ingroup. As a matter offact, the is reached independently in Coelodonta and removal ofsome characters only leads to more Elasmotherium. unresolved cladograms. Taking into account 5. The dorsal profile of the skull is pre- the high number of taxa considered, I have dominantly present with character state 1, tried to use as many characters as possible, but several changes occur. There are reversals without judging them a priori. This first at- to the plesiomorphic state in the Subhyra- tempt at a rigorous cladistic analysis of the codon clade (fig. 1, node 05), excepting Men- whole family Rhinocerotidae provides a ba- oceras, and independently in sis for a reevaluation of the usefulness of the and Dicerorhinus. The apomorphic state 2 is characters, the ingroup size, and the generic/ achieved independently in Pleuroceros, Te- specific level considered. leoceras, Diaceratherium, Diceros, and the Some characters used by other authors have Rhinoceros clade (fig. 1, node 17). been ignored in this analysis mainly because 6. The deep nasal opening is a derived char- they are not known (at least to me) for most acter of most Aceratheriinae, with a reversal of the considered taxa (e.g., position of the for the Teleoceras group (fig. 1, node 31). The lacrymal bone, articulation tibia-fibula- derived state is also a synapomorphy at node Groves, 1983; the dorsal notch of the atlas- 20 (Coelodonta and Elasmotherium sub- Prothero et al., 1986; the direction ofthe pre- clades). maxillae, the articulation femur-fibula, or the 7. The retraction of the anterior border of median lower crest of the mandibular sym- the orbit occurs independently in Floridacer- physis-Heissig, 1989). I consider others- as, Iranotherium, Elasmotherium, and gets such as shape and depth ofthe postfossete of its maximal degree in Ninxiatherium (aut- upper or the divergence of the apomorphy). parastyle of the upper teeth (Fortelius and 8. The relative projection of the orbit hap- Heissig, 1989)-not useful, at least at a ge- pens in Iranotherium and the Coelodonta and neric level. Elasmotherium subclades, less markedly in The following characters are used in this Coelodonta and Stephanorhinus than in the study (table 2): other genera. 1-3. Presence of horns. Nasal and frontal 9. The sagittal crest (fig. 1, node 04) has horns are considered separately, since they been lost by most of the Rhinocerotidae, but arise from different bones, and they are not several reversals to the pleisomorphic state homologous. Nasal horns have been treated occur independently, as well as the achieve- as two characters depending on whether they ment of the apomorphic state 2 from the are single or paired, to avoid a character with state 1. many different states which apparently 10. The relative position of the postglenoid evolved independently. and posttympanic apophyses appears to be a The analysis shows that the development rather homoplastic character. The character ofa nasal horn evolved independently to the state 1 is a synapomorphy at node 06 (fig. 1; apomorphic states (on the tip of the nasals table 4), but both reversals and changes to or well developed in the middle ofthe nasals), the state 2 occur. The change from character and there are reversals from both to the ple- state 0 to 2 seems to happen in Menoceras. siomorphic state (absence). The development Variation ofthis character among the species of paired nasal horns also evolved indepen- of a genus is known for Subhyracodon, Ron- dently to the apomorphic states, only once zotherium, Teleoceras, and Diaceratherium. to rounded bosses, and twice to lateral ridges. Heissig (1989) established that no reversals The development of the frontal horn consti- occur in the evolution ofthis character, which tutes a synapomorphy of the rhinocerotine disagrees with the present results. group, with independent reversals in Rhi- 11. The inclination of the occipital face is 6 AMERICAN MUSEUM NOVITATES NO. 3143

TABLE 2 Characters and Character States Considered for the Cladistic Analysis Plesiomorphic state = (0) unless otherwise noted. See comments in the text about the polarity of certain characters (*). A. Skull-mandible 1. Unique nasal horn: absent (0), small, on the tip of nasals (males) (1), well developed in the middle of the nasals (2). Unordered. 2. Paired nasal horns: absent (0), rounded bosses on the tip (1), lateral ridges (2). Unordered. 3. Frontal horn: absent (0), well developed (1), hugely developed (2). 4. Nasal septum: not ossified (0), partially ossified (1), totally ossified (2). 5. Cranial dorsal profile: flattened (0), slight occipital elevation (1), great occipital elevation (2). 6. Nasal opening: short (posterior edge between P1P3) (0), deep (posterior edge above P4M1) (1). 7. Anterior border of the orbit: P4M2 level (0), M3 level (1), behind M3 level (2). 8. Border of the orbit: continuous with the zygomatic arch, not projected laterally (0), projected laterally (1). 9. Sagittal crest: present (0), parietal crests very little separated (1), parietal crest clearly separated (2). 10. Postglenoid and posttympanic apophyses: separated (0), in contact (1), fused (2). 11. Occipital face: vertical (0), inclined backward (1), inclined forward (2). Unordered. 12. Occipital outline: high and narrow (0), roughly squared (1), low and broad (2). 13. Zygomatic width: normal (0), very broad (1). 14. Skull: dolicocephalic (0), brachycephalic (1). 15. Nasal length: very short, retracted (0), long (1). 16. Apophysis on the lateral border of the nasal bone: present (0), absent (1). 17. Mandibular symphysis: narrow (0), broad (1), very broad (2). 18. Mandibular ventral profile: straight (0), with upraised symphysis (1), clearly convex (2). 19. Posterior edge of the symphysis: short (plp2 level) (0), long (p3p4 level) (1). 20. Ascending ramus: inclined forward (0), vertical (1), inclined backward (2). B. Dentition 21. Upper 13-C: present (0), absent (1). 22. Lower i3: present (0), absent (1). 23. Lower C: present (0), absent (1). 24. Upper I2: present (0), absent (1). 25. Lower il: present (0), absent or minimum development (1). 26. Upper Il, shape: incisorlike (0), small chisel-shaped, laterally compressed (1), large chisel-shaped (2), absent (3). 27. Lower i2, shape: incisorlike (0), small tusk (1), large tusk (2), absent (3). 28. Lower pl: present (0), absent (1). 29. Lower p2: present (0), absent (1). 30. Cheek teeth: brachydont (0), subhypsodont (1), hypsodont (2). 31. Upper M3: quadrangular (metacone developed) (0), triangular (metacone lost) (1). 32. Posterior cingulum on upper M3: long, somewhat projecting, lengthening the base of the (0), short, little developed (1). 33. Metacone rib on upper premolars: well developed (0), slight or absent (1). 34. Hypocone on upper P2: united to the protocone, not to the metaloph (0), united to the metaloph and with a "bridge" to the protocone (1), hypocone and protocone separated (molarized premolars) (2). (*). 35. Hypocone on upper P3-P4: united to the protocone, not to the metaloph (0), united to the metaloph and with a "bridge" to the protocone (1), hypocone and protocone separated (molarized premolars) (2). 36. Protocone on upper premolars: not constricted (0), slightly constricted (1), very constricted (2). 37. Protocone on upper molars: not constricted (0), slightly constricted (1), very constricted (2). 38. Secondary folds on upper premolars: absent (0), simple (1), multiple (2). 39. Enamel of cheek teeth: little folded (0), quite folded (1), very much folded (2). 40. Lingual cingula on upper premolars: strong (0), weak (1), absent (2). 41. Lingual cingula on upper premolars: strong (0), weak (1), absent (2). Plesiomorphic state = (1). 42. Lingual cingula on lower premolars: strong (0), weak (1), absent (2). Plesiomorphic state = (1). 43. Labial cingula on lower premolars: strong (0), weak (1), absent (2). 44. Metaconid of lower p2: very prolonged backward (0), normal (1). 45. Labial groove on lower cheek teeth: deep (0), shallow, faded (1). 46. Cement on cheek teeth: absent (0), moderate (1), abundant (2). 1995 CERDENO: RHINOCEROTIDAE 7

TABLE 2-(Continued) C. Postcranial skeleton 47. Metapodials: long and narrow (0), long and broad (1), short but not massive (2), short and massive (3). Unordered. 48. Metacarpal V: functional (0), reduced (1). 49. Metacarpal IV, proximal facet: trapezoidal outlined (0), triangular outlined (1). 50. Metacarpal II, lateral facet for the McIII: anteriorly and posteriorly developed, with or without medial union (0), continuous, without marked medial narrowing (1), only anteriorly developed (2), very reduced or absent (3). Unordered. Plesiomorphic state = (2). 51. Metacarpal II, trapezium-facet: present (0), absent (1). 52. Scaphoid and semilunate: with short proximolateral and without a posterolateral facet (0), with long proximo- lateral and without posterolateral facet (1), with posterolateral facet (2). 53. Semilunate: with proximal facet for the ulna (0), without it (1). (*). 54. Pyramidal, medial distal facet: simple (0), bilobed or L-shaped (1). 55. Astragalus: high and narrow (0), short, squared (1), low and broad (2). 56. Astragalus, anteroposterior diameter (APD): normal (0), large, with facet 2 outstanding from posterior face (1). 57. Astragalus, trochlea: very oblique to the distal articular zone (0), slightly or not oblique (1). 58. Astragalus, facet 1 (proximoextemal): without distal prolongation (0), with narrow prolongation (1), with wide prolongation (2). Plesiomorphic state = (1). 59. Astragalus, facet 1: very concave (0), more or less flattened (1). 60. Astragalus, facet 2: isolated (0), united to facet 3 (1), elongated to proximal border (2). Unordered. Plesiomorphic state = (1). 61. Astragalus, facet 2: high and narrow (0), roughly rounded or oval (1), transversely elongated (2). 62. Calcaneum, tibial facet: present (0), absent (1). 63. Calcaneum, fibular facet: present (0), absent (1). (*). 64. Calcaneum, tuber: short (0), long (1). 65. Calcaneum, tuber: smooth unevenness (0), strong unevenness (1). (*). 66. Calcaneum, sustentaculum: at obtuse angle (0), at right angle (1). 67. Humerus, laterodistal epicondyle: short (0), high (1). 68. Humerus, laterodistal epicondyle: slightly laterally projected (0), well projected (1). 69. Femur, third trochanter: little developed (0), very developed (1). 70. Radius, posterior articular facets: united (0), separated (1). (*). 71. Tibia, anterior groove: wide (0), narrow (1). 72. Long bone epiphyses: narrow (0), wide (1).

rather homoplastic, with independent rever- dependent reversals in Subhyracodon, Am- sals to the plesiomorphic state (vertical face) phicaenopus, Aceratherium, Peraceras, and from the apomorphic state 1. The character Chilotherium. Nasal shortening also occurs state 2 is achieved independently from the in some species of Teleoceras and Aphelops. states 0 and 1. 16. The lateral apophysis of the nasal bone 12. The occipital outline also shows a ho- appears as a derived character (absent) for moplastic behavior in its evolution. Changes most of the family, with retention ofthe ple- occur from the plesiomorphic state to the siomorphic character state in Teletaceras, Pe- character state 1, and from state 1 to state 2, netrigonias, and Subhyracodon. with several reversals from state 1 to state 0. 17. The width of the symphysis is present 13, 14. Broad zygomatic width and brachy- with the derived state 1 in most acerather- cephaly. The derived state ofthese characters iines, related to the development of the i2, is achieved independently by different taxa. reaching the most derived state 2 in Chiloth- Both derived states co-occur in Peraceras and erium; there are reversals to the plesiomor- most teleoceratines. Therefore, these char- phic state in Prosantorhinus and Chilotheri- acters are not necessarily related, as previ- dium. Most ofthe rhinocerotines are also ple- ously claimed by Heissig (1989). siomorphic for this character, except Gain- 15. The nasal length appears as a derived datherium and Rhinoceros. character (long) in the family, with some in- 18. The mandibular ventral profile has var- a C\, 0 OO0 C-O-OOO"-4 C--C-"-4""4 O.-.0O" o C- 0 C- C- C- C- C- O -4 C- 0 0 0 o 4 o-. C- "4 0 "40-- 0 4" -" 4" 4 C- - C- 0 0 0 (D O) O " 4O C-F C-0O F H0000 O C C OOO0 PO D 0 00C-C-0C-00OC C-.--4 -4O0C- 0 C- 0 0Oa rm (D t- O c- c - C-0oo0 0 oC 40o -coC ---44 "40o a (D(D O-0 "4 "4 C- "4 "4 "4 "4O4"4O.C- C- -4 "4 C-O4C- 4 "4 "4 (O L) O "4 "4C-O 4 C- "4 "40O "4O 0 - C-4O C- C- O4 0" '"4"4"

(DY O00 C C- 0 0 0 0 0 c-00 " -C- -o -C - - o CD K) o"4 4-"4C--0 -C-"4 O0OC)CC- C- 00" 4 -4C- 00()C-0" '4 c O C O C- 0 0 C- O 4 0 O- - - C-" 4O4 0 - PO C 0 -4 - - C- C- 0 O C\2 O4COO0O0 0 C- O4O C- OO O OO O4O O-4 C-C4 C OO4O-- 0- C-C- O4 0C-H0- 0 OO CDa 00C-OC- oOOo"4(D0C-"4" -"40C-00"4"4 -4" LO OD -- O- -- 0- -A - c -c b- -4 o - t c\ c\ o. c, - c- c- oc-o X 1-a o o o o C C- 0000C-oo 4 o "4 0 C-O -4 0 4 " -4 = 10O0 0 a-C C-C-00000C- 0-4" q0 000 -4 -4 " 4 O )LLO 0 0 00 C- 4 0 0 O - C- 4 O4 \ 0 OC\ \ C § U1V0O " --IO00 0C- 0O-C-C-00000C-OOOOOOOOOOOOC- - 0H vC\2 0 0 0 Cm2C C\2 C\2 L) K) O O4 O4 C C-C-O C- " " " C 0 "4 " C-IO "0 "4 "4 O 0"4 O 4C-I CC-C- 0 0 0 "4 0 "4 C- 0 0 0 0 U) C NC\- O OOC-c- ONv )0--IO 4 C-\ )O OC- C- Ci- C- a) 0 C- C- C- C- 000 O C- O O OC- 0 0 0 00 1OD "4 C- CO- C- 0"" "40 C- 00 4a000"40 = _ - 0 0 C- "4 0 0 0o40 - "4) 000 o)0 o o)"400 (D 0 "40O-CO O O" O N O O O O "404 0 0 0 0 C- -I4 " 00O " "400 MO "4 000 C 0 000 C- "4 "4 "4 0 "4 C- "4 " C "4 K) o" N 0O4O \j O- 00 N C- 00 N C- C-O0 C- 4 C- 4 0 02CNM °o 0 4 C °-C\ - -° 0 -° C- 4 C-0 C°- C\q"4 C\2 "4000°C\ C\2 C0 000000000000 C- 0 0 o C) K)O OOOOOOOOOOOo O OO c 0O0 C OD 0 "-C- C- 00 N" "4 "4 N" 0tO 0 N 0. 0 4 4 4 "NN 0 N " C- N N N N _ )CD 0 0 "4 0 00o0 0 "4NI"4 0 0 "4 0 N N 0 X tO O 0 °C-C -0 - 0 0 ° °\4 ° °\O\N " I\- O"- O\IC 4 4 N F U F °C -I~ N-NNC-N-C\"N CQN"CQ CQ 4" 4 "4N " 4NNNC\2 C\ NCQ toK)t 00 o" a 0 a " 4 "4 "4 " 4-4 " 44 0 4 "4 "4 " 4 -.I

K0) OOOOOO0"4 4" 000 "4 o"4 -o "

)C\"00 ° 4 OC -4 O O ° °""4° °44""4° O"""444 NO1 ) 0 0 0 000 0 0 C 0 " 0 0 o C\lNC%lCY) 00"CC-0 O O O) O 0"4 0 40 00CH " "4 4 "4 00 0" " N\l - 00"4"4N-4-q "4 4 4 C N N" " " " N N CVNIL 00 0 0 0 00C 00"4 C-" 4-4 0"4000 Cq (D 0 O 0 j_ C- O O O t O\) cm O O cvO -4 CM~FNILO 0C-00 0 -OC0004"4C O O -"44"-4b- 4 --0 "4"Cb- O O~ "44O~ 0 0 4"4O NK) 0 000"4"4""4""4""4O) 4""-4"4""44""4 4" N ( 0 C- 0 0 "4" 4" 4" 4" 4 "4 "4 "4 "4 "4 "4 "4 "4 "4 "4

wSN"4N OOOO-OCOOOOC-0"C"400 "4 "4 "4 " C-"4C0-40-0"4-"4 oD 0o b-"40o 4 "4 "4 "4 " "4 "4 n 00"4)0C)0 °0 0 00 0 0 0 °--0O4C 0 "4 " "4 g4 0"4"400 0 000 C- 0 " 0" "4 00"C\ "4- "4 - C-1- 0 C-1 C-. C- 0-0 00"4"0 0 0"4 "4 "4 "4 "4"4"4C- "4- "4 "4 "4 "4 -4 "4 "4 "410 C-o oggg0"o00 0 o"o" o\ o o C C- C-" o4 C\" C\0"-

OD1o-°00C°oo 00o0o o--o C" O C- O

o o o O o - t- o b. o o) o ooooo> "4K)O) 0"0C- 00 0 " 0 " C- 4 (D 0 ) ) 0o0 " 4 0

F4N 0 00 C000 0 " C-0 4 N C - " t- 0 C- 0 00 0 0 4 0 0 - 0 0 0 O) O ooOO4C- C\ -IoooC- C-" C.-0- C- NOO

O C-O0000000'O A L"44"400C-"4006 0" 0 O O O O O O O O O O - 00 0 K)6OOO O O O O O O O O O O O O -00 00 N OO O ONSt0O O 000 -O N O O "4OOOO OOOOOOOO"4OC40"4Q4"4

r.~~~~~C';cd9ZSEXXoX .; s NV 0 0a-~ 0 O- -4 C- 0 -4 1- -4 C-- C-- -4 -4 - -4 0 O00O 4-4 C- 0C- 00OC) - - 0C- 0OC- -0 C- C- C- 0o C- - C- C- - C C- - C- C- C- 0 4 C- 0 C..

(D O 0 O C- 0 0 O - O O C- - C-o C- C- C- -4 -4 -40 CD 0 0 C- 000 C- 4 0 C- C- C- 0 C- C- C- C- -4 -4 - (D LmO O O0 -4 O o o 1- c- C - o 0- -4 4 *-- c- o - 4 O C- CD(oN- 00000C-0- 0 C -4 -C- - C-00-0 C- C-0000C)C)-4 (D- - -I O C- -4 ~-4 -4 C-- C- C- C- C4 4 CO- C- 000O O

O) 0 O H C C- C-- C-- -.CC O C- OC C-- C-0 0 OC

LO CY) a O 0 O0 O0 O O O O _I 1D00 C-O C-- IC--4-C-- 0 C- co 0 0 0 0-0-OOO0 --OC0c I4 c C--- - 1-_I C- 0C-_1-0O Cc C- L)@O OvO- O O-O4O\O- C-- -4O - C- O- C- 1) D 0 O O 0 _I O O O C O O O- O4- C-- C-- 0 s 0 0--

10CQ ~~ 000000C- C -C-b C C- M-4-0 0 - 00 C- -4 -4 - 1O O C-C-00N-C--4 C-O 10C\ O OC-0O00N NN0CCORC\l 000 0C-O C- C-C--0-C-C- 0 OC- -C)C-0-CC- C-C--C)0C--4 C-C-

Vt(0 )_l00004c-O00O O O -NNNO O NNNC- C C-C- 0.-I*-O-0NN--00 C- --O I ~-IN- C-

LO O O 0 0 0 -4 0 0 - 0 O C- 0 0 0 C- b- .. 0 0 0 -4 0 1DNoNNjCQ C>2 C- CN C N2 C\ cN N N 0 c- C- N C- \ N N C-. -4 V C\ 02 v CQ mO \O Ol " " Cv w CV C\l C\ C\ C\ 04 cv O C\ 0 0o O\ OcC4C\l\l CQ C\ CQ Ofl C\l O Ol -4 OC\al Ca2-| -1 C\l 0- -4 N N N C-. C- 0 OO\N IN C- N N- C-OO C\ N N 0 N O N\ C- C-N\ N\ N H-N\l C-. V)OD -4 -4 I4 O-0 Ol0 C\l 020 l 4 _ - 4o cl -4 _-I -_-4- V)t- -I Cl CQ --I <-4 Cl cm C2 Cv Csl Cqi C\ OC\l? C\ilC-2- qN c Cf2 CQ -|C\z C.2o C\ C\zN >C\zN NCs2-4 Cvl NCfi-4CAN O N-\a2C -4 C\ C\22 \1 Cv l C'Il C\ V) IV C\l C0 CC\lC22 C\2CQ CQ CO.. qCQ C\CO NO- ol> C\l 02 Oa C\l Oa C\i c)0toV) lO CC2C C CQC -4 -4 -40 C-IO I 0 0 04 -

w) CM O C> CQ -4 -4 CQ CQ CfiC C-2 CfiC- Ca Ca C-2 04 -4_ -I O4 O4 O -I U t' 0 D OO O-N N .-I-4 -I O-0C O C- N Nl NO N N N

Oan t. C\2 C-4 C>2 0-2C"~H O.- C\j ca2Ci2C2C C\ C\ 2 C>242 CQ4)p NO ) 0 00 00 0 00a00 C-CO- 0 O 00 0C00 00 000 OO O O C N0O 0 0 0 - = -I C- - C- -4 -I C l C- -I C\ -4 -I C\l OO O- to to O) O4 -4 -O - - tO toO

IIV r00 .- C *.-C-4 -IN 0Nt.to- t 0tottI-I- 4

\ 0 N C-C - _ - 0O- N O- N 0- _ C I l H O) O N O . NNC-a2e o NNo- I NNN-N F o a cqN ca2 N- m - 000 0C C - C 0 C-.N0000> 0 - - 0 00( - C0C NNN C -C-0000HHC H~ 0NNNN0000H ~H -4 C\ 0 0 -4O O N C O. O- O a -4 - H q O b- 0 O4 O4 O4 C\ -4 9 ~ ~~~ ~~~~ 0 0O 02 O O 0- O. 4 0 b- 0 Q -4 CQ 0 C\ q 0 -4 4 C) 00 004-400 -40C\2a -0C-.I.-4.A-I000..-4CD 0 000\0\0O O) 0 O\ 0 -- Cm ol C- 0- C\ C\ C\b-4 COQ - cq cq cv cq ol C\ C (0 ~ 0~-4C4-414- 414-000 ~- I.414- C-4-4-4-I004- 00 t- O O O O O OC- 0 b- O O O\ O C- O O 0 O O O O (Oto --I O --O C4 - O O O4 O4 _ O O. O O OD O O O O LO_I O O4 O4 C\z_I 4 e- C.- 0-- -4 O4 O -O O -4 C-I O C\j C\ 0 -\| -I O V O Cq0 CQ0 CV C C- 0 a Oa O- O 0 O CQ CQC 0 0aa 0 0

C\ O O O a CD 0= 0u O O _4 1 Cf 0 0 0 0 O a O O O O O N 0000000000000000000000O0 _I N 0 C N 0 C 0 0FC C000C\ 0C>2H0000 to 0 0 --I C.-C\l C' O ONNONN0CO-NN0NC-0NO0OONONNONNONNOO -l - O O O O Oj

toO0111 1 Cs O O O C\l O O O O O ~ O -i 1 0-4 CF OOOOO OOOOOO H 1 0 C2ClClCi - ClCi ClC-O C X s x \ Q ClC 10 AMERICAN MUSEUM NOVITATES NO. 3143 ied several times within the group, with in- to the state 1 in Alicornops, and to the state dependent reversals from the upraised sym- 2 in some teleoceratines. physis condition to the straight one; the de- 28. The presence or absence of a lower pl rived state 2 (very convex) characterizes the is rather variable even among the outgroups group of rhinocerotines at node 18 (fig. 1), (absent in Hyracodontinae). It is absent in with a reversal to the plesiomorphic state in most aceratheres (derived state), while it is Stephanorhinus. present in the Rhinocerotinae except Gain- 19. Length of symphysis. A long symphysis datherium and Lartetotherium sansaniense is present in Menoceras and all the Acerath- (it is absent in L. schleiermacheri). eriinae, while it remains short in most Rhin- 29. The absence of p2 happens in Teleo- ocerotinae except Lartetotherium sansan- ceras (although the p2 is actually present in iense (it is short in L. schleiermacheri, so it T. proterus), and it is variable in Elasmo- varies within this genus) and the dicerotine therium. group (also variable within Ceratotherium). 30. Tooth hypsodonty. Most rhinoceroses 20. The ascending ramus appears with the have brachyodont teeth, although the devel- derived state 1 (vertical) in Protaceratherium opment of a certain hypsodonty (state 1) oc- and most of the Aceratheriinae, with rever- curs several times as a response to a more sals in Peraceras and some teleoceratines. This siliceous diet, and some taxa reach a higher reversal also occurs within rhinocerotines in degree of hypsodonty (state 2) like Chilo- Iranotherium and Rhinoceros, while the de- therium or Teleoceras, and mainly Elasmo- rived state 2 (inclined backwards) is a syna- therium. pomorphy of the clade at node 18 (fig. 1), 31. Shape of the M3. The triangular shape with reversal to the state 1 in Stephanorhinus. is a derived character ofthe family, excluding 21-23. Third and canine. These Teletaceras, which retains the plesiomorphic characters have been assumed as character- quadrangular shape with the metacone rib state 1 (absence of third incisor and canine) developed. Only one reversal happens in for some forms whose anterior dentition is Coelodonta. actually unknown, based on the fact that these 32. Shape of posterior cingulum on M3. teeth are only documented among the most The considered character states (table 2) re- ancient species of Rhinocerotidae (Teleta- flect the main observed patterns, although ceras, Penetrigonias, and Trigonias; table 3). variation within these patterns occurs even Teletaceras is, in fact, the only one with com- intraspecifically. The well-developed poste- plete dental formula, while the other two taxa rior cingulum of the M3 is reduced several have already lost the i3 and lower canine. times in the Subhyracodon group, some ac- Character 21 gathers I3 and upper canine since eratheres, and all rhinocerotines except Coe- both teeth behave in the same way. lodonta (reversal). 24, 25. 12 and iH. The absence of 12 and il 33-38. Morphology ofthe upper premolars. is a common derived feature among the As in the previous case, the character states which in- Rhinocerotidae; however a secondary devel- reflect the main observed patterns, clude a greater variation that can occur main- occurs in L. opment of I2 Gaindatherium, ly at specific level. schleiermacheri, and Punjabitherium, and the The fading or loss of the metacone rib on presence of il is more common in the acer- the ectoloph (character 33) occurs in the Sub- athines as well as in the rhinocerotines. hyracodon group and the aceratheres, while 26, 27. The complex chisel-shaped Il/tusk- it is present in most Rhinocerotinae, being like i2 is a characteristic ofthe whole family. lost in Dicerotina, Rhinoceros, and Coelo- The large i2 (272) is present in all Acerath- donta. eriinae, while most Rhinocerotinae present Molarization of the premolars is achieved the character state 1 for both teeth. The loss (34, 352) several times independently (Sub- of Il/i2 (state 3) is characteristic of the Di- hyracodon group, some aceratheres and most cerotina and the Coelodonta and Elasmo- rhinocerotines). Some genera present a mo- therium clades (fig. 1, node 18). The loss of larized P2, while the P3-P4 remain pre- or the upper I1 is also documented among the submolariform (character states 0, 1). Heissig aceratheres (table 4, node 24), with reversals (1989: fig. 21.1) considered five different de- 1995 CERDENO: RHINOCEROTIDAE I1I grees of molarization with which I do not dently, but the most derived state (2) char- agree. The premolariform and submolari- acterizes the Rhinocerotinae, except Gain- form stages established by Heissig seem to datherium and Lartetotherium. reflect minor variations within a premolari- 47. Type of metapodials. The apomorphic form stage equivalent to the state 0 consid- states 2 and 3 are achieved independently ered herein, and the paramolariform stage from state 1 (long and broad); state 2 (short would be included in the molariform state but not massive) is present in Alicornops, Per- (state 2). aceras, and Aprotodon. The short and mas- The greatest constriction of the protocone sive metapodials are characteristic of the te- (36, 372) is developed among the aceratheres leoceratine group, except Aphelops (471). In (more on molars than on premolars) and in some cases the difference between states 0 the elasmotheriine (sensu lato) group. and 1 can be somewhat subjective, condi- Concerning the development of secondary tioned by the size of the bones, and a pos- folds on upper premolars (character 38), a sibility for further analysis might be the quan- certain variation occurs mainly referred to tification of the character based on the gra- multiple folds (state 2); I consider this state cility index. only when most of the premolars present 48. The McV is functional in Hyrachyus multiple folds (crochets and/or cristae), even while it is reduced in the hyracodontid sub- if some teeth with simple ones are docu- families. The plesiomorphic condition re- mented within a same sample. There is an mains in most aceratheres, becoming re- independent evolution from character state duced in Chilotherium, Teleoceras, and Bra- 0 to both apomorphic states. State 2 is present chypotherium. The derived state is also pres- at node 05 (fig. 1), except Subhyracodon. State ent in all Rhinocerotinae. 1 appears as a synapomorphy at node 08 (fig. 49. The outline of the proximal facet of the 1; table 4), changing to character state 2 in McIV is unknown for many taxa, including Prosantorhinus, Punjabitherium, the Iran- the basal ones. The derived state (triangular) otherium clade, and the Elasmotherium and is present in one of the outgroups (Hyraco- Coelodonta subclades. dontinae), in Menoceras, and in the Rhino- 39. Folding of dental enamel is a derived cerotinae, except Elasmotherium in which the state characteristic of the Iranotherium and trapezoidal outline appears as a reversal. Elasmotherium clades, reaching its maxi- 50. The lateral McIII facet of the McII mum degree in Elasmotherium. shows an independent evolution from the 40-43. Development of lingual and labial plesiomorphic state (2: anteriorly developed) cingula is very variable, even among the out- to the different apomorphic states, with re- groups. The derived condition (state 2: ab- versal from state 1 (continuous facet). The sence) is constant in Rhinocerotinae; only the reduction ofthe facet (state 3) occurs in Men- lingual cingulum on upper premolars appears oceras and Protaceratherium. State 1 appears in Paradiceros and Diceros. It can also be in some aceratheres and the Dicerotina, while present in Stephanorhinus. For characters 41 most Rhinocerotinae present the plesiomor- and 42 the plesiomorphic state seems to be phic state as reversal from the state 1. state 1 (weak development), conditioned by 51. The loss of the trapezium facet on the variation among the outgroups. McII appears as derived state in one out- 44. The metaconid of p2 is not prolonged group (Eggysodontinae), and independently backward in most members ofthe family (ex- in the Subhyracodon group (except Menocer- cept the oldest ones) or in Eggysodontinae. as), in Hoploaceratherium, and the Dicero- Reversals happen in Floridaceras, Aproto- tina. don, and Gaindatherium. 52. Articulation scaphoid-semilunate. The 45. The shallow labial groove on lower cheek evolution of this character shows that pos- teeth is independently reached in some taxa session of a posterolateral articular facet be- (Peraceras, Aphelops, Diaceratherium, Bra- tween the scaphoid and the semilunate (state chypotherium, Ceratotherium, and Coelo- 2) is achieved independently from the ple- donta). siomorphic state 0 (only one short proxi- 46. Presence of a certain quantity ofcement molateral facet) or from state 1, which im- on the teeth occurs in different taxa indepen- plies a previous elongation of the proximo- 12 AMERICAN MUSEUM NOVITATES NO. 3143 lateral facet. The posterolateral facet is pres- ceros, Floridaceras, and many rhinocero- ent in Ronzotherium, most teleoceratines (fig. tines, with reversals from state 0 to 1 in Apro- 1, node 31), and some rhinocerotines (Lar- todon and Hispanotherium. The apomorphic tetotherium and Coelodonta; it varies in Ste- state 2 is present in "Beliajevina," Hoploa- phanorhinus). ceratherium, and the teleoceratine group, al- 53. Ulnar facet. The polarity of this char- though the character varies in some genera acter is conditioned by the different state of of the latter. the outgroups. Taking both analyses into ac- 59. Flattening of astragalar facet 1 occurs count, the plesiomorphic state is considered several times from the primitive concave to be the presence of an ulnar facet (0). Both condition. Within aceratheres it happens in hyracodontid subfamilies, some basal rhi- Floridaceras and the clade at node 25 (fig. 1), noceroses and all aceratheres present state 1. while within the rhinocerotines it appears in Reversals to state 0 occur in Mesacerather- Rhinoceros and the dicerotine clade. ium, Aphelops, Brachypotherium, and the 60. Astragalar facet 2 is primitively united Rhinocerotinae, changing in Lartetotherium to facet 3 (state 1), although one of the out- sansaniense. groups, Eggysodontinae, presents the apo- 54. Presence of a bilobed or I-shaped in- morphic state 0 (isolated facet). At node 06 feromedial facet on the pyramidal is the de- (fig. 1) there is a change within rhinocerotids rived condition present in all Rhinoceroti- to this same condition, but there is a reversal nae, with reversal to the plesiomorphic state in the rhinocerotine group (node 10), within in the Hispanotherium group. However, a which only Diceros reaches the isolated con- certain posterolateral extension of that facet dition. "Beliajevina" is the only taxon with has been observed on one specimen of His- character state 2 (proximal elongation of the panotherium matritense (Cerdefio, 1992). facet), although this condition can also ap- 55. General shape of the astragalus. Most pear (maybe less markedly) in some taxa such ofthe family presents the apomorphic states. as Diceratherium, where other states are also The roughly squared shape is present in Tri- present and are predominant. From the de- gonias, and as a synapomorphy at node 06 scription of Borissiak (1938) it can be as- (fig. 1). The broadening ofthe astragalus (state sumed that the four known astragali of Be- 2) occurs independently within the Acerath- liajevina caucasica present the same charac- eriinae, in Alicornops, and in most teleocer- teristic offacet 2, without variation. This spe- atines (node 31); Aprotodon also has this state. cies also presents character state 1, since facets The rhinocerotines are derived with charac- 2 and 3 are united. A possibility could be the ter state 1, with two reversals in Dicerorhinus division of character 60 into two, one with and Paradiceros. the character states 0 and 1 established here 56. Astragalus development. The special (table 2), and the other considering the pres- anteroposterior development, with outstand- ence or absence of a proximal elongation. ing facet 2, is a derived condition achieved However, the results would be similar since by most aceratheriines (fig. 1 node 25), with such elongation in other taxa does not seem a reversal in Peraceras. to be constant, and this would be reflected 57. Obliquity of trochlea. The loss of obli- by missing data. Therefore the presence of quity with respect to the distal articulation this feature would still be an autapomorphy is a derived feature for Aceratheriinae and of "Beliajevina". Rhinocerotinae, with reversals in Aphelops 61. Shape of astragalar facet 2. Most ofthe andAprotodon, in which the obliquity is quite family Rhinocerotidae is derived for this marked. character with the state 1 (rounded or oval 58. Shape of astragalar facet 1 is a rela- astragalar facet-2) (fig. 1, node 04), although tively variable character, which is reflected several reversals occur in Subhyracodon, by missing data for different character states Floridaceras, Alicornops, Rhinoceros and the in a genus. From the plesiomorphic state (1: dicerotine group. The evolution to character with narrow prolongation), this character state 2 (transversely elongated) takes place evolves to both apomorphic states. The state within the teleoceratine clade (Teleoceras, 0, without prolongation, appears in Pleuro- Diaceratherium and Brachypotherium). 1 995 CERDENO: RHINOCEROTIDAE 13

62, 63. Tibial and fibular facets on calca- of the femur takes place in the outgroup Eg- neum. In both cases Hyrachyus presents the gysodontinae, in Peraceras, the teleoceratines state 0 (presence), and the hyracodontid sub- (except Teleoceras), and all rhinocerotines families the state 1 (absence) (table 3). The except the Hispanotherium group. loss of the tibial facet occurs several times 70. Posterior facets of the radius. The po- within Rhinocerotidae, in Mesaceratherium, larity of this character depends on the differ- Pleuroceros, Chilotheridium, Diaceratherium ent state present on the outgroups. The de- and "Begertherium." Here the latter taxon is rived character state (posterior facets of the definitely considered a synonym of Hispan- radius united) is present in Eggysodontinae otherium, in which the tibial facet is present, and some basal taxa; data are missing for what would reflect a variation ofthe character Hyracodontinae and Teletaceras. The char- within the genus. acter appears to be very homoplastic; it varies With respect to the fibular facet, its absence within a number of genera. occurs in the Alicornops group, in Larteto- 71. The anterior tibial groove is present therium, Hispanotherium, and "Begerthe- with the apomorphic state (narrow) in the rium." Many taxa present missing data for clade at node 05 (fig. 1) except in Subhyra- this character. codon, and in the Hispanotherium group. 64. Development of a long tuber calcis oc- 72. The acquisition of wide long bone curs in Ronzotherium and as a synapomor- epiphyses happens independently in Flori- phy of the clade at node 06 (fig. 1). Within daceras, Alicornops, and the teleoceratines the Aceratheriinae there is a reversal at node within the acerathere group, and in most 25, with a new change to the apomorphic rhinocerotines except the Hispanotherium state in Brachypotherium. Within the Rhin- group, Lartetotherium sansaniense, and Par- ocerotinae the reversal takes place in the Ir- adiceros, which present reversals to the ple- anotherium clade (node 16), Rhinoceros (node siomorphic state. 17, but unknown in Punjabitherium), and the Elasmotherium and Coelodonta subclades (node 20). RESULTS AND DISCUSSION 65. The unevenness of the tuber calcis is a One hundred and four equally parsimoni- more variable character. The derived state ous trees with a length of 497 (consistency (strong) appears in Diceratherium and as syn- index 22, and retention index 59) were dis- apomorphy at node 06 (fig. 1; table 4). With- covered in the first analysis including three in the aceratheres, there is a reversal at node outgroups. Figure 1 shows the strict consen- 26 (Alicornopini and Teleoceratini). Among sus tree ofthese 104 trees, from which several the rhinocerotines several independent interesting reinterpretations of rhinocerotid changes occur. phylogeny are proposed. The different clado- 66. The angle of the sustentaculum of the grams correspond to alternative topologies calcaneum is present with the derived state for the polytomies ofthe consensus tree. Fig- (at right angle) in one of the outgroups (Eg- ure 2 presents the consensus tree correspond- gysodontinae) and most rhinocerotids, with ing to the second analysis of the data matrix a reversal for the rhinocerotine group at node including just Hyrachyus as outgroup. This 10 (fig. 1), within which the derived state is second analysis provided six equally parsi- reached independently in the Hispanother- monious trees with a length of 473 (ci = 23; ium group (node 16), Ceratotherium, and ri = 59). Comparison of the two analyses re- Rhinoceros. veals the stability of some of the obtained 67, 68. The laterodistal epicondyle of the groups, while relationships of other taxa, humerus becomes high and well projected mainly the geologically oldest ones, vary from (derived states) in the teleoceratine clade, with one analysis to the other. The following dis- reversals for character 67 in Aphelops, and cussion will focus on the first analysis (fig. 1). for character 68 at node 33 (fig. 1). There are The basal node presents a trichotomy for several independent changes within the rhin- Hyrachyus, the two hyracodontid subfami- ocerotines. lies as a monophyletic group, and the family 69. The development ofthe third trochanter Rhinocerotidae as another monophyletic 14 AMERICAN MUSEUM NOVITATES NO. 3143

Hyrachyus Eggysodontinae Hyracodontinae Teletaceras Penetrigonias Trigonias Ronzotheriun "MesaceratheriunW' Subhyracodon Diceratherium Menoceras "F:AN 95544" Protaceratherium Pleuroceros Anvhicaenocus ELasmotheriun Ninxiatherium CoeLodonta StePhanorhinus Paradiceros Ceratotheriun Diceros Puniabitheriun Rhinoceros x,

"BeLiaievinall 0mrn Hisoanotheriun m "Beaertheri un' -4 Iranotheriun Dicerorhinus m Lartetotherium "D."schleiermacheri Gaindatheriumn ADrotodon Floridaceras Aceratheriun HopLoaceratherium Al icornors Peraceras m Chilotheridiun AoheLogs m Chilotherium Teleoceras Prosantorhinus re Diaceratheriun Brachypotheriun Fig. 1. Strict consensus tree of the 104 obtained cladograms for the family Rhinocerotidae, with Hyrachyus and the hyracodontid subfamilies Eggysodontinae and Hyracodontinae as outgroups. Char- acters at each node in table 4. "F:AM 95544" = new genus in Prothero (in press). group (fig. 1, nodes 00-03). Due to this basal state is considered to be state 0 (hypocone trichotomy and to the fact that the outgroups united to the protocone on P2; presence of sometimes present different character states, ulnar facet on the semilunate; presence of the polarity of certain characters (34, 53, 63, fibular facet on the calcaneum; smooth un- 65, and 70) differs from the first analysis to evenness of the tuber calcis; posterior facets the second. In the first one, these characters of the radius united). appear to be derived for Hyrachyus (with The clade of the hyracodontid subfamilies character state 0), but the second analysis is supported by three synapomorphies (481, shows these five characters as derived with 500, 621): McV reduced, McII with lateral character state 1 for all Rhinocerotidae ex- McIII facet anteriorly and posteriorly devel- cept Teletaceras (fig. 2; table 5, node 01). Thus oped (missing data for Eggysodontinae), and considering both analyses, the plesiomorphic the absence of tibial facet on the calcaneum. 1995 CERDENO: RHINOCEROTIDAE 15

Hyrachyus Teletaceras Penetrigonias Trigonias Ronzotherium "Mesaceratherium" Subhyracodon ,Diceratherium "F:AM 95544" Protaceratherium Menoceras PLeuroceros ELasmotherium Ninxiatherium CoeLodonta Steohanorhinus Paradiceros Ceratotherium Diceros Puniabitherium = Rhinoceros -4 0 Dicerorhinus m "Beliaievina" m HisDanotherium z "Begertheriun" - Iranotheriun Lartetotherium "D. 'schLeiermacheri Gaindatherium Anvhicaenoous Aorotodon =m Floridaceras Aceratherium HooLoaceratheriun Acerorhinus ALicornoos m Peraceras ChiLotheridium ADheLovs m ChiLotherium TeLeoceras Prosantorhinus Diaceratherium BrachyDotherium Fig. 2. Strict consensus tree of the six obtained cladograms for the family Rhinocerotidae, with Hyrachyus as the only outgroup. Characters at each node in table 5. "F:AM 95544" = new genus in Prothero (in press).

The monophyly of Rhinocerotidae is sup- characters 22 and 23 (see below). This result ported by four synapomorphies: 1 51, 261, 271, agrees with Hanson (1989: 379), who already and 661: nasal long, chisel-shaped I1, i2 de- discussed the primitive condition retained in veloped as a tusk, and sustentaculum of the Teletaceras, and regarded its phylogenetic calcaneum at a right angle. However, this position to be between Hyrachyus and all oth- monophyly is not totally supported in the er rhinocerotids. On the other hand, char- second analysis (fig. 2), which shows the acters 261 (chisel-shaped Il) and 271 (tusklike primitive condition of Teletaceras, and its i2) unit Teletaceras with the rhinocerotids. closer relationship with the Hyracodontinae. Penetrigonias is also a primitive taxon Teletaceras retains the plesiomorphic states which has already lost the lower i3 and canine for characters 210, 220, and 230 (presence of (221, 231), but retains the upper 13-C (210) I3/i3 and canines) while all other rhinocer- like Teletaceras and Trigonias. otids exhibit the apomorphic state at least for The derived state is shared by most of the 16 AMERICAN MUSEUM NOVITATES NO. 3143

TABLE 4 Distribution of Character States at Each Node of the Consensus Tree in Figure 1 Synapomorphy = S; parallelism = P; parallel reversal = PR; reversal = R. Node 01: 48', 500, 62' = P Node 02: 15' = S; 26', 17', 66' = P. Node 03: 16' = S; 5', 11', 22', 23', 31', 37' = P. Node 04: 9' = S; 21', 33', 35', 61' = P. Node 05: 50, 420, 630 = PR; 32', 342, 352, 382, 48', 51', 71' = P. Node 06: 580 = S; 630, 650 = PR; 10', 12', 24', 43', 50', 55', 600, 64' = P. Node 07: 272, 47' = S; 17', 18', 412, 432 = p. Node 08: 120 = PR; 25', 38', 57' = P. Node 09: 640 = PR; 102, 342, 52', 72' = P. Node 10: 60' = R; 530, 502, 660 = PR; 32', 422, 522, 69' = P. Node 11: 240, 330 = PR; 12, 352, 402, 48', 49' = P. Node 12: 3', 54' = S; 170, 250, 27' = PR; 44', 63' = P. Node 13: 10', 180 = PR; 12', 20', 24', 36', 67' = P. Node 14: 100, 52', 610, 630 = PR; 92, 28', 46', 50', 68' = P. Node 15: 30, 120, 34', 35', 502, 520, 540, 680, 690, 720 = PR; 30', 372 382 39', 61', 64' = P. Node 16: 470, 58' = PR; 63', 66', 71' = P. Node 17: 200, 700 = PR; 52, 102, 112, 30', 64' = P. Node 18: 182, 202, 273 = S; 700 = PR; 25', 263 = P. Node 19: 360, 370, 40' = PR; 19', 33', 51', 59' = P. Node 20: 4' = S; 502, 670, 680 = PR; 6', 8', 30', 382, 61', 64' = P. Node 21: 102, 302, 522 = p Node 22: 110, 34', 35', 490 = PR; 7X, 372, 39' = p. Node 23: 110, 58' = PR; 6', 20', 262, 372, 44' =p. Node 24: 41', 43' = PR; 263 = P. Node 25: 56' = S; 100, 430, 640 = PR; 19', 28', 30', 59' = P. Node 26: 65', 72' = P. Node 27: 342, 352, 62', 63' = P. Node 28: 90 =PR; 46', 472 = p. Node 29: 92, 10', 12', 14', 36', 582, 68', 69' = P. Node 30: 473 = S; 13', 321, 422, 431, 671 = P. Node 31: 60, 250, 262 = PR; 1', 46', 522, 552 = p Node 32: 200, 410, 680, 700 = PR; 122, 342, 352, 362 = p. Node 33: 280, 320 = PR; 451, 612 = P family Rhinocerotidae and one of the out- aceratherium gaimersheimensis (Heissig, groups for character 31, the shape of M3, 1969) and Protaceratherium minutum (Cer- which appears as state 1 (triangular) in Eg- denio, 1989). However, the cladograms (figs. gysodontinae and all Rhinocerotidae except 1, 2) do not support this view, but relate it Teletaceras and Coelodonta (reversal). more closely to Ronzotherium. The polytomy at node 04 in figure 1 in- Node 05 of figure 1 gathers most of the cludes the genus "Mesaceratherium," whose genera included in the subfamilies Dicerath- position varies in the alternative topologies eriinae and Menoceratinae of the current owing to the large amount of missing data. classification (Prothero and Schoch, 1989), This is also true for "F:AM 95544," which except Pleuroceros. The main feature used to appears in the polytomy at node 05 (fig. 1), define these subfamilies is the presence of although in the second analysis it constitutes paired nasal horns (character 2), but the apo- the sister group ofDiceratherium (fig. 2, node morphic states of this character appear as an 05). "Mesaceratherium" seemed to be closer independent acquisition in the genera Men- to Protaceratherium as supported by a num- oceras (21), Diceratherium (22), and Pleuro- ber of similarities between the species Mes- ceros (22). On the other hand, the synapo- 1995 CERDENO: RHINOCEROTIDAE117

TABLE 5 Distribution of Character States at Each Node of the Consensus Tree in Figure 2 Synapomorphy = S; parallelism = P; parallel reversal = PR; reversal = R.

Node 01: 221, 23' = S; 281, 341, 531, 631, 65', 701 = P. Node 02: 161, 211, 31' = S; 111, 351, 371 = p. Node 03: 110, 420 = PR; 32', 331, 342, 481, 511 = P. Node 04: 630 = PR; 382, 61', 71' = P. Node 05: 650, 700 = PR; 11', 352 = P. Node 06: 280 = PR; 24', 42' = P. Node 07: 510 = PR; 5', 9', 10' = P. Node 08: 580= S; 38' = R; 650, 710 =PR; 431, 551 =P. Node 09: 272, 47' = S; 102, 171, 412, 432 = p. Node 10: 330, 530, 660 = PR; 18', 422, 522, 65', 69' = P. Node 11: 58', 700=PR; 11', 12' =P. Node 12: 240 = PR; 12, 352, 402, 491, 57, = P. Node 13: 3', 54' = S; 170, 27' = PR; 11', 44', 631 = P. Node 14: 101, 180, 650 = PR; 201, 241, 361 = P. Node 15: 100, 521= PR; 92, 281, 462, 671 = P. Node 16: 30, 341, 351, 520, 540, 690 = PR; 301, 372, 382, 391, 641 = P. Node 17: 470, 581 = PR; 661, 71' = P. Node 18: 610, 630, 700 = PR; 12', 50', 721 = P. Node 19: 500, 68' = P. Node 20: 200 = PR; 52, 102, 1 12, 30', 64' = P. Node 21: 182, 202, 273 = S; 25', 263 = P. Node 22: 401 = PR; 19', 331 = P. Node 23: 360, 370 =PR; 511, 59', 68' = P. Node 24: 4' = S; 502 = R; 670 = PR; 6', 81, 301, 382, 611, 641 P. Node 25: 102, 302, 522 = p. Node 26: 110, 341, 351, 490 = PR; 71, 372, 391 = p. Node 27: 320, 480 = PR; 251, 50', 571, 600 = P. Node 28: 341, 581 = PR; 61, 181, 201, 262, 372, 441 = p. Node 29: 101, 411, 431 = PR; 263 = P. Node 30: 561 = R; 100, 430 = PR; 191, 28', 30', 59' = P. Node 31: 65', 72' = P. Node 32: 342, 352, 62', 631 = P. Node 33: 90 = PR; 46', 472 = p. Node 34: 92, 10', 12', 141, 361, 582, 68', 69' = P. Node 35: 473 = S; 13', 32', 422, 431, 671 = P. Node 36: 60, 250, 262 = PR; 11, 461, 522, 552 = P. Node 37: 200, 410, 680, 700 = PR; 122, 342, 352, 362 = P. Node 38: 280, 320 = PR; 451, 612 = P. morphy used by Prothero et al. (1986) to de- nor the subfamilies Diceratheriinae and fine the subfamily Menoceratinae (terminal Menoceratinae are supported, since their nasal horn bosses) would not include Pro- genera appear as a paraphyletic group. There- taceratherium, as later stated by Prothero and fore, it is better to maintain all these genera Schoch (1989). Heissig (1989: 405-406) con- as primitive Rhinocerotidae without gath- sidered the tribe Trigoniadini within the sub- ering them at any other taxonomic level. "F: family Diceratheriinae for the genera Tri- AM 95544" is added to them, closer to Di- gonias, Ronzotherium, and Amphicaenopus ceratherium (fig. 2), and thus is removed from (Prothero and Schoch, 1989, in the same vol- the aceratheres (Prothero et al., 1986; Proth- ume, did not follow Heissig). Looking now ero, in press). at the cladograms (figs. 1, 2), neither the tribe At node 06 of figure 1, Pleuroceros is the 18 AMERICAN MUSEUM NOVITATES NO. 3143 sister group of the remaining rhinocerotids, Floridaceras appears at node 9 in the first sharing eight synapomorphies and three par- analysis (fig. 1) with six apomorphies and four allel reversals (table 4). reversals (71, 90, 180, 352, 370, 500, 591, 610, All other taxa comprise a monophyletic 62', 67'). In the second analysis, where it group (fig. 1, node 08), which comprises in appears related to the aceratheres, it shows turn two monophyletic clades, which to some one more apomorphy (72') within the clade degree support the subfamilies Rhinocerotin- (fig. 2, node 27). In this case, the character ae and Aceratheriinae of the current classi- states shared at node 27 are 25' (presence of fication, although there are exceptions. The i 1), 320 (long posterior cingulum of M3), 480 second analysis (fig. 2) also shows two main (functional McV), 57' (astragalus trochlea not clades, with two main differences: (1) in the oblique), and 600 (astragalus facet 2 isolated), first analysis (fig. 1, node 07), Amphicaenopus while those shared at the node 9 of the first is the sister group of the great monophyletic cladogram (fig. 1, table 4, node 9) are 102 group while it is included in the clade of the (fused posttympanic and postglenoid apoph- Rhinocerotinae in the second one (fig. 2) re- yses), 342 (molarized premolars), 52' (scaph- lated to Aprotodon; (2) Floridaceras is in- oid with long proximolateral facet), 640 (short cluded in the rhinocerotine clade in the first calcaneum tuber), and 721 (narrow long bone case, while it is within the aceratheres in the epiphyses). These latter character states (ex- second one. cept for 52') are also present in some acer- Amphicaenopus is considered a primitive athere taxa. On the other hand, among the rhinocerotid by Prothero and Schoch (1989), character states shared with the aceratheres although the present analysis suggests that it at node 27 (fig. 2), characters 320, 480, and is more derived than previously thought, 600 are not present within the rhinocerotines sharing the synapomorphies of the node 07 (only 320, posterior cingulum ofthe M3 long, (fig. 1; table 4). Within the clade, it presents is a reversal in Coelodonta). Therefore it can characters 150, 330, 350, 502, and 700 as re- be said that a close relationship of Florida- versals to the plesiomorphic state, and 131, ceras is better supported to the aceratheres 342, and 422 as apomorphic. (fig. 2) than to the rhinocerotines (fig. 1), even With respect to Aprotodon, it is evident when the most parsimonious option in the that its current ascription to the teleocera- first analysis relates it to the latter group at tines (Prothero and Schoch, 1989) is ques- the base of the clade. tioned. As can be seen in the matrix (table The rest ofthe Rhinocerotinae, apart from 3), a number of characters are unknown for Amphicaenopus, Floridaceras, and Aproto- this genus (14 of them corresponding to the don, is roughly equivalent to the tribe Rhin- skull, mandible, and dentition), but they do ocerotini ofProthero and Schoch (1989), but not affect its position in the cladogram. Apro- it differs in the internal relationships of its todon shares all the synapomorphies of the genera, except for the subtribe Dicerotina. In rhinocerotine clade (fig. 1; table 4; nodes 09, both cladograms (figs. 1, 2) Gaindatherium 10), without missing the corresponding char- appears to be closer to Lartetotherium than acters, except for characters 34 and 42. With to Rhinoceros, to which it was supposed to regard to the eight synapomorphies shared be directly related (Colbert, 1934; Prothero by the teleoceratine clade (fig. 1; table 4; node et al., 1986; Prothero and Schoch, 1989). Out 29), Aprotodon shares with them characters of the four characteristics (92, 110, 370, and 121 and 691 only, with missing data for char- 380) of Gaindatherium within the clade at acters 9, 14, and 68. Similarly, with respect node 11 (fig. 1), only one (92: parietal crests to the seven synapomorphies shared by the clearly separated) is shared with Rhinoceros. Aceratheriinae (fig. 1; table 4; node 23), Apro- Characters 11, 37, and 38 in Gaindatherium todon is missing two characters, and only are reversals to the plesiomorphic state (ver- shares the characters 372 (protocone of the tical occipital face; protocone of the upper upper molars very constricted) and 581 (facet molars not constricted; absence ofsecondary 1 ofthe astragalus with narrow prolongation), folds on upper premolars). which are apomorphies for Aprotodon within A close relationship that is also supported the clade united at node 10. by both analyses is that of the species "Di- 1995 CERDENO: RHINOCEROTIDAE 19 cerorhinus" schleiermacheri with the genus apomorphies (fig. 2, node 23), concerning Lartetotherium (L. sansaniense) (figs. 1, 2; the constriction of the protocone (not con- table 4, node 12; table 5, node 13), as sug- stricted, plesiomorphic), the absence of tra- gested in a previous paper (Cerdeino, 1992). pezium facet on the MclI, the flatness of the Therefore, I propose to formally include "D." astragalar facet 1, and the strongly projected schleiermacheri as a second species of Lar- lateral epicondyle of the humerus. tetotherium, together with L. sansaniense. Paradiceros shows characters I 10, 550, 68°, This species presents several plesiomorphic and 720 as plesiomorphic with respect to the states with respect to L. schleiermacheri such two other dicerotines, and it is derived for as characters 90 (presence of sagittal crest), characters 361 and 371 (protocone of upper 101 (postglenoid and posttympanic apophy- teeth slightly constricted). The monophyly of ses in contact), I 10 (vertical occipital face), these three genera is supported in any case, 180 (straight mandibular ventral profile), and and the subtribe Dicerotina is thus main- 720 (narrow long bone epiphyses). In turn, L. tained. The close relationship between Di- schleiermacheri retains the plesiomorphic ceros and Ceratotherium is considered well condition for characters 19 (short posterior supported, as shown in figure 2. edge of the symphysis), 36 (protocone of the Other monophyletic clades can be estab- upper premolars not constricted), and 53 lished within the rhinocerotine group. In both (semilunate with ulnar facet), also plesiomor- analyses, a monophyletic clade gathers two phic in Dicerorhinus. Both Lartetotherium smaller groups as a sister group of the sub- species appear close to the extant Dicero- tribe Dicerotina. One of them (fig. 1, node rhinus (D. sumatrensis) in the first analysis 21)joins the genera Stephanorhinus and Coe- (fig. 1, node 14), and to a lesser degree in the lodonta, and the other one (fig. 1, node 22) second one (fig. 2, node 19); in both cases unites Ninxiatherium and Elasmotherium. Dicerorhinus appears characterized within the This implies the separation of the elasmo- respective clades by reversals to the plesio- theriine genera as currently considered. These morphic states 50, 11°, 360, 370, and 550. four genera were previously placed in two The genus Punjabitherium is here consid- different subtribes, Dicerorhinina and Elas- ered the sister group ofRhinoceros, support- motheriina (Prothero and Schoch, 1989). ed by both analyses (figs. 1, 2; tables 4, 5; Ninxiatherium was considered by these au- nodes 17 and 20, respectively), contrary to thors within Rhinocerotoidea as incertae Prothero and Schoch (1989: 536) who placed sedis, although since its original description the former within the Rhinocerotoidea as in- its relationships with the elasmotheriines were certae sedis. However, their relationship with clearly established (Chen, 1977). Present re- respect to other rhinocerotines is still not well sults lead to a modification ofboth subtribes supported. In the first analysis (fig. 1, node Elasmotheriina and Dicerorhinina. The group 14), they appear as part of a polytomy to- (Elasmotherium plus Ninxiatherium) ap- gether with two other subclades and the genus pears to be much closer to other Rhinocero- Dicerorhinus, while in the second analysis tini than to the Iranotherium group, which both are the sister group of Dicerorhinus, appears as a well-defined monophyletic group forming a monophyletic clade at node 19 (fig. in both analyses (figs. 1, node 15; fig. 2, node 2). The current subtribe Rhinocerotina 16). Elasmotherium and Ninxiatherium share (Prothero and Schoch, 1989) is not supported five synapomorphies (110 and 490 as plesio- by these analyses, since it gathers the genera morphic, and 341, 351, and 372 as apomor- Rhinoceros and Gaindatherium. phic). Character 7 (anterior border of the or- A group well supported by both analyses bit) is apomorphic in both genera, but with within Rhinocerotinae is that formed by the state 1 in Elasmotherium (border over M3), genera Paradiceros, Diceros, and Cerato- and state 2 in Ninxiatherium (border behind therium, the subtribe Dicerotina (Prothero M3). Similarly, character 39 appears with and Schoch, 1989). In figure 1 (node 19) they character state 1 in Ninxiatherium (enamel constitute a trichotomy, while in figure 2 of the cheek teeth quite folded), and with (nodes 22, 23), Diceros and Ceratotherium character state 2 in Elasmotherium (very appear more closely related, sharing five syn- much folded). In this later feature Ninxiathe- 20 AMERICAN MUSEUM NOVITATES NO. 3143 rium is coincident with the whole Iranothe- iine" species is Hispanotherium matritense rium group. The separation of Elasmothe- from Spain. Recent discoveries (Inigo and rium from the other Miocene "elasmotheri- Cerde-no, in prep.; unpubl. data) have greatly ines" is not so surprising, taking into account increased knowledge of this species, includ- several features, such as the huge frontal horn ing the morphometrical variation within a or the extremely hypsodont teeth, that render large single population. Nevertheless, an im- this genus a very peculiar rhino. But it is not portant character is not yet clear, the presence the same for Ninxiatherium, whose dental of nasal or nasofrontal horn, that was estab- characteristics and the presence of a nasal lished as a synapomorphy ofthe elasmotheri- horn are much closer to Iranotherium. ine group by Fortelius and Heissig (1989). The main synapomorphy shared by (Nin- Personal observation of the nasal fragments xiatherium plus Elasmotherium) and (Ste- of H. matritense (unpubl. data) and "Beger- phanorhinus plus Coelodonta) refers to char- therium" grimmi from Sof9a (Turkey; Heis- acter 4 (nasal septum partially or totally os- sig, 1976: 29) did not reveal any rugosity that sified), while many of the other synapomor- could demonstrate the existence of a horn, phies (table 4, node 20; table 5, node 24) are although Heissig (1976) included its presence also shared by the Iranotherium group (table as a diagnostic character of"B." grimmi. The 4, node 15; table 5, node 16; 301, 382, 502, polytomy of the Hispanotherium clade (fig. 611, 641), out of which Ninxiatherium has 1, node 16; fig. 2, node 17) renders evident missing values for characters 50, 61, and 64. the very close relationships among them. It There are also three other synapomorphies appears that "Begertherium" is characterized to be considered, which gather the subclades only by the apomorphic character 62' (cal- (Ninxiatherium plus Elasmotherium) and caneum without tibial facet), which indirectly (Stephanorhinus plus Coelodonta) with the supports the suggested synonymy with His- Dicerotina: 182 (convex ventral mandibular panotherium not only at the generic level profile), 202 (ascending mandibular ramus in- (Antunes and Ginsburg, 1983; Cerdeino, 1989; clined backwards), and 273 (absence of i2), and contrary to Fortelius and Heissig, 1989), although Stephanorhinus presents reversals but also at the specific level as stated by Inigo for characters 180 and 20', and Ninxiathe- and Cerdeino (in prep.). This synonymy would rium has missing values for the three char- imply a possible regard- acters. So the inclusion of Ninxiatherium in ing the presence ofa horn, since B. borissiaki this group is really supported by the presence does have a well-developed nasal horn boss. of nasal septum (41) as well as the derived "Beliajevina" presents six apomorphies state of character 6 (deep nasal opening) in within the group at node 16 (fig. 1): absence contrast to the Iranotherium group. In sum- of I1 (263), lateral facet of McII anteriorly mary, the closer relationship of Ninxiathe- and posteriorly developed (500), facet 1 of rium to (Coelodonta plus Stephanorhinus) is the astragalus with wide prolongation (582), than that of Elas- facet 2 elongated to the proximal border (602), more weakly supported strong unevenness of the tuber calcis (65'), motherium. Further studies on this "elas- and short laterodistal humeral epicondyle motheriine" group, especially on some Mio- (670). These features support its separation cene remains from China and Spain (Cer- from Hispanotherium and "Begertherium" de-no, in progress), will likely provide new (with character states 261, 502, 581, 60', 650, data to establish more accurate relationships. and 671), in agreement with Fortelius and The Miocene age of Ninxiatherium (Chen, Heissig (1989), although Iinigo and Cerdeiio 1977) implies an early acquisition ofthe nasal (in prep.) establish itjust as a different species septum with respect to the other three Plio- of Hispanotherium, keeping their synonymy genera. The development of the at generic level. All these characters as well nasal septum may be an independent acqui- as most of the synapomorphies at node 16 sition within two different evolutionary lin- (fig. 1, table 4) are missing data in Iranothe- eages ifNinxiatherium is proven to belong to rium. the Iranotherium group. The phylogenetic analysis of Fortelius and At present the best known "elasmother- Heissig (1989) only dealt with the elasmo- 1995 CERDENO: RHINOCEROTIDAE 21 theriine (sensu lato) species. The synonymy 411 and 431 (weak labial cingulum on upper of Caementodon oettingenae and Hispano- and lower premolars). Hoploaceratherium, in therium matritense was already justified turn, presents six apomorphies for characters (Cerdeiio, 1989; Inigo and Cerdefio, in prep.). 1', 342, 352, 51', 582, and 64'. Node 30 (fig. Fortelius and Heissig (op. cit.: 227) identified 1) unites the remaining genera, sharing five two monophyletic clades of elasmotheriines, apomorphies and two reversals (table 4). Ac- recognizable at a suprageneric level, but con- erorhinus is apomorphic for character 412 cluded that "to do so formally serves no use- (absence of labial cingulum on upper pre- ful purpose." Looking now at the global anal- molars), and presents reversals for 250 (pres- ysis of the family, this recognition seems to ence of il) and 420 (strong lingual cingulum be useful, since the Elasmotherium-Ninxia- on lower premolars). therium group appears quite well differenti- The remaining aceratheres form a mono- ated from the Iranotherium clade. So, the phyletic group with two clades in the same subtribe Elasmotheriina of Prothero and way in both analyses (fig. 1, node 26; fig. 2, Schoch (1989) could be transformed into two. node 31). These two clades share two syna- This view would really correlate with the old pomorphies: strong unevenness of the tuber concept ofthe subfamilies Iranotheriinae and calcis (65') and wide long-bone epiphyses Elasmotheriinae (Kretzoi, 1943; Viret, 1958) (72'). The first clade (fig. 1, table 4, node 27) being placed at a lower taxonomic level. gathers Chilotheridium with Alicornops plus However, this would not be real, since Elas- Peraceras, sharing the following synapomor- motherium and Ninxiatherium gather with phies: molarized premolars (342_352), and other quite different genera in the present absence oftibial and fibular facets on the cal- cladograms (figs. 1, 2). Trying to fit these re- caneum (62'-63'). Within this clade, Chil- sults into a systematic classification, I pro- otheridium has characters 12, 92, 11', 302 and pose the subtribe Iranotheriina (new rank) for 401 apomorphic, and characters 170, 180, and the clade including Iranotherium, Hispano- 410 as reversals. Alicornops and Peraceras therium (= Begertherium), and "Beliaje- share the moderate presence of cement on vina." Elasmotherium and Ninxiatherium, the cheek teeth (461) and the short but not in turn, remain in the subtribe Elasmotheri- massive metapodials (472), as well as the ina that also includes Coelodonta and Ste- presence of a sagittal crest as a reversal (90). phanorhinus as two different genus groups, This monophyletic clade of three genera is but caution must be used with respect to Nin- proposed as the new tribe Alicornopini. xiatherium as explained before. The other monophyletic clade (fig. 1, node The second large clade previously referred 29; fig. 2, node 34), with Aphelops, Chilo- to (fig. 1, node 23; fig. 2, node 27) includes therium, Teleoceras, Prosantorhinus, Diacer- most taxa assigned to the subfamily Acera- atherium, and Brachypotherium, includes theriinae. The basal node of this clade varies most teleoceratines of the current classifica- from one analysis to another and, as com- tion (Prothero and Schoch, 1989) and two mented above, the inclusion of Floridaceras other aceratheriine genera, sharing the fol- within the Aceratheriinae is considered better lowing synapomorphies: parietal crests clear- supported than within the Rhinocerotinae. ly separated (92), postglenoid and posttym- The remaining aceratheres share six syna- panic apophyses in contact (10'), occipital pomorphies and two reversals (fig. 1; table 4, outline roughly squared (121), brachycephalic node 23): short nasal opening, upraised man- skull (141), protocone ofthe upper premolars dibular symphysis, ascending ramus inclined not constricted (36'), facet 1 ofthe astragalus forward, large chisel-shaped Il, protocone of with wide prolongation (582), laterodistal upper molars slightly constricted, metaconid epicondyle of the humerus well projected of p2 not prolonged backward, vertical oc- (68'), and third trochanter of the femur very cipital face and astragalar facet 1 with narrow developed (69'). Excluding Aphelops, the re- prolongation. The following node 24 (fig. 1; maining teleoceratines are gathered at node table 4) presents an apomorphy in character 30 (fig. 1) with six synapomorphies among 263 (absence of I1) with respect to node 23, which the acquisition of short and massive and it is also characterized by two reversals, metapodials (473) is unique within the Rhin- 22 AMERICAN MUSEUM NOVITATES NO. 3143 ocerotidae. The others refer to: the great zy- letaceras together with the other Rhinocer- gomatic width (131), also present in Peracer- otidae; Teletaceras is included when the three as; the short posterior cingulum of M3 (321) outgroups are considered, while it remains with reversal to the plesiomorphic state in separated from the rest when only Hyrachyus Diaceratherium and Brachypotherium; the is the outgroup. absence of lingual cingulum on lower pre- The previous recognized subfamilies Di- molars (422) with reversal to state 1 (weak ceratheriinae and Menoceratinae are not sup- cingulum) in Prosantorhinus; the presence of ported by the analyses, since their genera form a weak labial cingulum on lower premolars paraphyletic groups. (431) with two changes to apomorphic state The cladograms exhibit two main clades 2 (absence) in Chilotherium and Teleoceras; which roughly correspond to previous con- and finally the high laterodistal epicondyle of cepts of the subfamilies Rhinocerotinae and the humerus (671). Aceratheriinae, with several exceptions. At node 31 (fig. 1; Table 4) the group shares Within the former, the tribe Rhincerotini is the presence of a little horn on the tip of the modified: the new subtribe Iranotheriina is nasals (I1), the presence of a third postero- proposed, including some of the "elasmo- lateral facet on the scaphoid (522), and a low theriine" taxa of previous classifications; the and broad astragalus (552). The presence of subtribes Elasmotheriina and Rhinocerotina cement on the teeth is moderate (46'), as in are reconstituted; and the subtribe Dicerorhi- Peraceras and Alicornops, and becomes nina is not supported. Punjabitherium and abundant in Teleoceras (462). Prosantorhinus Ninxiatherium, previously considered as is plesiomorphic for this character (460). The Rhinocerotoidea incertae sedis, are included group (fig. 1 , node 31) is also characterized in the Rhinocerotini: the former as sister tax- by two additional reversals: the short nasal on to Rhinoceros, the latter as sister taxon to opening (60) and the presence of i 1 (250). Elasmotherium. The subtribe Dicerotina is The tribe Teleoceratini was included with- well supported in both analyses. in the subfamily Rhinocerotinae by Prothero The suggested synonymy between "Be- and Schoch (1989: 535), but not Heissig gertherium" and Hispanotherium is sup- (1989: 406), who, in the same volume, ported, as well as the close relationship ofthe grouped the teleoceratines with the acerath- species "Dicerorhinus" schleiermacheri with eres. The first authors clearly followed the the type species ofthe genus Lartetotherium, previous opinion of Prothero et al. (1986) becoming Lartetotherium schleiermacheri. when they discussed Heissig's viewpoint. The Other Miocene species usually ascribed to Di- present cladograms support the inclusion of cerorhinus might belong to Lartetotherium as the teleoceratines within the Aceratheriinae, well. and implies an extension of the tribe Teleo- The second main clade unites the acera- ceratini to include Aphelops and Chilothe- there and teleoceratine genera; so that the rium. tribe Teleoceratini (with additional number Qiu et al. (1988) established a new tribe, of genera) is removed from the subfamily Chilotherini, within Aceratheriinae, includ- Rhinocerotinae, and included in the Acer- ing Acerorhinus and Chilotherium. This view atheriinae. A new tribe of aceratheres, Ali- was not followed by Prothero and Schoch cornopini, is proposed. (1989), nor is it supported by the present The phylogenetic relationships of the analysis. Rhinocerotidae lead to the classification list- ed below if the following conditions are met: (1) differences between the two analyses per- CONCLUSIONS formed here, and (2) evident homoplasy of The cladistic analyses of the family Rhin- certain characters, as well as the incomplete ocerotidae, using two subfamilies of Hyra- knowledge of many taxa. codontinae and/or Hyrachyus as outgroups, Family Rhinocerotidae Owen, 1845 show a basal polytomy which differs from one Teletaceras Hanson, 1989 analysis to the other in the inclusion of Te- Trigonias Lucas, 1900 1995 CERDENO: RHINOCEROTIDAE 23

Penetrigonias Tanner and Martin, 1976 Subfamily Rhinocerotinae Owen, 1845 Ronzotherium Aymard, 1886 Amphicaenopus Wood, 1927 Mesaceratherium Heissig, 1969 Aprotodon Forster Cooper, 1915 Subhyracodon Brandt, 1878 Gaindatherium Colbert, 1934 Diceratherium Marsh, 1875 Lartetotherium Ginsburg, 1974 Pleuroceros Roger, 1898 Tribe Rhinocerotini Owen, 1845 Protaceratherium Abel, 1910 Dicerorhinus Gloger, 1841 (=Plesiaceratherium) Subtribe Rhinocerotina Owen, 1845 Menoceras Troxell, 1921 Punjabitherium Khan, 1971 "F:AM 95544," new genus Rhinoceros Linnaeus, 1758 in Prothero (in press) Subtribe Dicerotina Ringstrom, 1924 Subfamily Aceratheriinae Dollo, 1885 Paradiceros Hooijer, 1968 Floridaceras Wood, 1966 Diceros Gray, 1821 Aceratherium Kaup, 1932 Ceratotherium Gray, 1867 Hoploaceratherium Subtribe Iranotheriina Kretzoi, 1943 (new Ginsburg and Heissig, 1989 rank) Acerorhinus Kretzoi, 1942 Hispanotherium Crusafont and Villalta, Tribe Alicornopini NEW 1947 Alicornops Ginsburg and Guerin, 1979 (= Begertherium; Caementodon) Peraceras Cope, 1880 Beliajevina Heissig, 1974 Chilotheridium Hooijer, 1971 Iranotherium Ringstrom, 1924 Tribe Teleoceratini Hay, 1902 Subtribe Elasmotheriina Bonaparte, 1845 Prosantorhinus Heissig, 1974 Group Elasmotherium-Ninxiatherium Diaceratherium Dietrich, 1931 Ninxiatherium Chen, 1977 Brachypotherium Roger, 1904 Elasmotherium Fischer, 1808 Teleoceras Hatcher, 1894 Group Coelodonta-Stephanorhinus Aphelops Cope, 1873 Stephanorhinus Kretzoi, 1842 Chilotherium Ringstrom, 1924 Coelodonta Bronn, 1831

REFERENCES Aguirre, E., and C. Guerin Familie der Rhinocerotidae. Palaeon- 1974. Premiere decouverte d'un Irantheriinae tol. Z. 5: 118-121. en Afrique: Kenyatherium bishopi nov. Cerdeiio, E. gen. nov. sp. de la formation vallesienne 1989. Revision de la Sistematica de los rino- (Miocene superieur) de Nakali (Kenya). cerontes del Neogeno de Espafia. Co- Est. Geol. Inst. Inv. Geol. Lucas Mal- leccion Tesis Doctorales NO 306/89. lada 30: 229-233. Edit. Univ. Complutense de Madrid: Antunes, M. T., and L. Ginsburg 429 pp. 1983. Les rhinoc'erotides du Mioc'ene de Lis- 1992. Spanish Neogene Rhinoceroses. Pa- bonne-systematique, ecologie, paleo- laeontology 35(2): 297-308. geographie, valeur stratigraphique. Chen, G. Ciencias Terra (UNL) 7: 17-98. 1977. A new genus of Iranotheriinae of Nin- Bonis, L. de xia. Vertebr. PalAsiat. 15(2): 143-146. 1973. Contribution a l'etude des Rhinocero- Colbert, E. H. tides aquitaniens d'Europe occidentale. 1934. A new rhinoceros from the Siwalik beds C.R. Acad. Sci. Paris 270: 764-766. of India. Am. Mus. Novitates 749: 13 Borissiak, A. pp- 1938. A new Dicerorhinus from the Middle Dashzeveg, D. Miocene ofNorth . Trav. Inst. 1991. Hyracodontids and rhinocerotids Paleozool. Acad. Sci. URSS 8(2): 1-69. (Mammalia, Perissodactyla, Rhinocer- Breuning, S. otoidea) from the Paleogene of Mon- 1923. Einige Worte zur Stammesgeschichte der golia. Palaeovertebrata 21(1, 2): 1-84. 24 AMERICAN MUSEUM NOVITATES NO. 3143

Farris, J. S. R. M. Schoch, (eds.), The evolution of 1988. Hennig86, version 1.5. Computer pro- perissodactyls, pp.399-417. New York: gram and documentation. Port Jeffer- Oxford Univ. Press. son Station, New York. Hooijer, D. A. Fortelius, M., and K. Heissig 1966. Miocene rhinoceroses of East Africa. 1989. The phylogenetic relationships of the Bull. Br. Mus. Nat. Hist. Geol. 13(2): Elasmotherini (Rhinocerotidae, Mamm.) 117-190. Mitt. Bayer. Staatsslg Palaont. Hist. Geol. Huang, W., and D. Yan 29: 227-233. 1983. New Material of Elasmotherini from Fortelius, M., P. Mazza, and B. Sala Shennongjia, Hubei. Vertebr. PalAsiat. 1993. Stephanorhinus (Mammalia, Rhinocer- 21(3): 223-229. otidae) of Western European Pleisto- Inigo, C., and E. Cerdefio cene, with a revision ofS. etruscus (Fal- In Hispanotherium matritense from the coner, 1868). Palaeontogr. Italica 80: 63- prep. Middle Miocene of Corcoles (Guada- 155. lajara, Spain): contribution to the Geraads, D. knowledge of the Miocene Elasmoth- 1988. Revision des Rhinocerotinae (Mam- eriines (Rhinocerotidae). malia) du Turolien de Pikermi. Com- Kretzoi, M. paraison avec les formes voisines. Ann. 1943. Gobitherium n. g. (Mammalia, Rhino- Paleontol. 74(1): 13-41. cerotidae). Foldt. Kozl. 73: 268-271. Groves, C. P. Nixon, K. C. 1983. Phylogeny of the living species of Rhi- 1992. CLADOS 1.1 IBM pc compatible char- noceros. Z. Zool. System. Evol. 21(4): acter analysis program. Ithaca. 293-313. Nixon, K. C., and J. M. Carpenter Guerin, C. 1993. On outgroups. Cladistics 9: 413-426. 1980. Les rhinoceros (Mammalia, Perissodac- Prothero, D. R. tyla) du Miocene terminal au Pleisto- In press. Rhinocerotidae. In C. Janis et al. (eds.), cene superieur en Europe occidentale. Tertiary of North America, Comparaison avec les especes actuelles. vol I. Cambridge, England: Cambridge Doc. Lab. Geol. Lyon 79: 1182 pp. Univ. Press. Hanson, C. B. Prothero, D. R., and R. M. Schoch 1989. Teleteraceras radinskyi, a new primi- 1989. Classification of the Perissodactyla. In tive rhinocerotid from the late Eocene D. R. Prothero and R. M. Schoch (eds.), Clarno Formation, Oregon. In D. R. The evolution of Perissodactyls, pp. Prothero and R. M. Schoch (eds.), The 530-537. New York: Oxford Univ. evolution of Perissodactyls, pp. 379- Press. 398. New York: Oxford Univ. Press. Prothero, D. R., E. Manning, and C. B. Hanson 1986. The phylogeny of the Rhinocerotoidea Heissig, K. (Mammalia, Perissodactyla). Zool. J. 1969. Die Rhinocerotidae (Mammalia) aus der Linnean Soc. 87: 341-366. oberoligozanen Spaltenfullung von Gai- Prothero, D. R., C. Guerin, and E. Manning mersheim bei Ingolstadt in Bayern und 1989. The history of Rhinocerotoidea. In D. ihre phylogenetische Stellung. Bayer. R. Prothero and R. M. Schoch (eds.), Akad. Wissensch. 138: 1-133. The evolution of Perissodactyls, pp. 1972. Pala6ntologische und geologische Un- 322-340. New York: Oxford Univ. tersuchungen im Tertiar von Pakistan; Press. 5: Rhinocerotidae aus den unteren und Qiu, Z., J. Xie, and D. Yan mittleren Siwalik-Schichten. Abh. Bay- 1988. A new chilothere skull from Hezheng er. Akad. Wissen. Math. Naturw. KI. Gansu, China-with special reference 152: 1-112. to the Chinese "Diceratherium". Sci. Sin. 1976. Rhinocerotidae (Mammalia) aus der 31(4): 493-502. Anchitherium-Fauna Anatoliens. Geol. Ringstrom, T. J. Jahrb. 19: 1-121. 1924. Nashorner der -fauna Nord- 1981. Probleme bei der cladistischen Analyse Chinas. Palaeontol. Sin. 1(4): 1-159. einer Gruppe mit wenigen eindeutigen Roman, F. Apomorphien: Rhinocerotidae. Pa- 1911. Les Rhinocerides de l' d'Eu- 1 inont. Z. 55: 117-123. rope. Arch. Mus. Hist. Nat. Lyon 11: 1989. Rhinocerotidae. In D. R. Prothero and 1-92. 1995 CERDENO: RHINOCEROTIDAE 25

Viret, J. Wood, H. E. 1958. Perissodactyles. In J. Piveteau, Traite 1932. Status of Epiaceratherium (Rhinocero- de Paleontologie, VI, 2: 368-475. Paris: tidae). J. . 13(2): 169-171. Masson. Yan, D. Watrous, L., and Q. Wheeler 1979. Several species offossil mammals from 1981 The out-group comparison method of the Miocene ofFangxian County, Hubei character analysis. Syst. Zool. 30(1): 1- Province. Vertebr. Palasiat. 17(3): 189- 11. 199.

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates, Bulletin, and Anthropological Papers published during the last five years are available free of charge. Address orders to: American Museum of Natural History Library, Department D, Central Park West at 79th St., New York, N.Y. 10024. TEL: (212) 769-5545. FAX: (212) 769- 5009. E-MAIL: [email protected]

® This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).