Cranial anatomy of Ignacius graybullianus and the affinities of the Plesiadapiformes.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 89:477498 (1992) Cranial Anatomy of lgnacius graybullianus and the Affinities of the Plesiadapiforrnes RICHARD F. KAY, J.G.M. THEWISSEN, AND ANNE D. YODER Department of Biological Anthropology and Anatomy, Duke University Medical Center, Durham, North Carolina 27710 KEY WORDS Primate evolution, Paleoprimatology, Archonta, Dermoptera A nearly complete cranium of Zgnacius graybullianus proABSTRACT vides increased understanding of the cranial anatomy of Plesiadapiformes. In nearly all details of cranial anatomy, Ignacius differs markedly from primates. USNM 421608 exhibits a long tapering snout, small widely spaced orbits, and a complete lack of postorbital process or bar. Large olfactory bulbs are inferred from the wide interorbital space. The marked flare of the zygomatic arches suggests that Ignacius possessed large and powerful temporal muscles. The basicranial region is particularly well preserved and reveals a distinct suture between the petrosal bone and a n entotympanic bulla. This suture is visible on both the left and right sides of the skull and dispels the hypothesis that Ignacius and, by inference, other Plesiadapiformes share the primate synapomorphy of a petrosal bulla. To test the phylogenetic position of Zgnacius, cranial characters were identified and scored for Ignacius, Plesiadapis, Cynocephalus, and a number of primates, bats, and scandentians. Two erinaceomorph insectivores were also included to allow the assessment of archontan monophyly. These characters were incorporated into a maximumparsimony analysis to determine the phylogenetic position of Plesiadapiformes. There are several important phylogenetic conclusions that can be inferred from this analysis: 1)Ignacius and Plesiadapis make up a monophyletic clade; 2) Plesiadapiformes may be the sister group of Dermoptera; 3) Scandentia, not Plesiadapiformes, is the sister group of Primates; and 4 ) Primates. plesiadapiforms. bats, colugos. and scandentians may not form a monophyletic clade Archonta. Consequently, the taxon Archonta is in need of review. 0 1992 Wiley-Liss, Inc Central to debates about primate origins has been the Plesiadapiformes, a Holarctic Paleogene group of mammals. Initial claims for the primate status of Plesiadapiformes were based on shared dental similarities, particularly of Plesiadapis with Eocene primates. With recovery of cranial and postcranial bones, further synapomorphies have been proposed based especially on the structure of the foot skeleton and the anatomy of the middle ear and its vessels (see, e.g., Szalay, 1972,1975). Arguments about the status of Plesiadapiformes (herein sensu Szalay and Delson [19’79] construed to include [but not re0 1992 WILEY-LISS, INC. stricted to] Plesiadapidae and Paromomyidae, but not Microsyopidae) have intensified recently with new recoveries of cranial and postcranial material of Ignacius. Most authors agree that the affinities of Zgnacius are with one of the orders of Archonta (Primates, Chiroptera, Dermoptera, Scandentia), but consensus ends there. Traditionally Zgnacius was considered a paromomyid plesiadapiform closely related to primates (see, e.g., Szalay and Delson, 1979). Recently Beard (1990) proposed on the basis of post~~ Received September 20, 1991; accepted June 6, 1992 478 R.F. KAY ET AL. cranial evidence that Ignacius and other paromomyids were the sister group of the living flying lemur (Cynocephalus, Order Dermoptera) and that the two major plesiadapiform families, Plesiadapidae and Paromomyidae, do not form a monophyletic clade. Some of the evidence for this conclusion has been challenged by Krause (1991). Kay et al. (1990) published a preliminary discussion of the lgnacius skull described in this paper and also concluded that paromomyids were the sister group of Dermoptera, but without proposing that Plesiadapidae were to be excluded from such a relationship. We present here the anatomical details of a particularly well-preserved skull of Ignacius graybullianus and clarify and correct prior cranial reconstructions of this animal based on more fragmentary material. We examine the relations among Primates, Plesiadapiformes, Chiroptera, Scandentia, and Dermoptera. To approach this question, we made detailed cranial comparisons among early eutherians, Ignacius, other Plesiadapiformes, Eocene Primates, and various living mammals, including Chiroptera, Lipotyphla (e.g., hedgehogs and tenrecs), Dermoptera, Primates, and Scandentia. MATERIALS AND METHODS TABLE 1. Taxa and specimens used in phylogenetic analysis (Those specimens with no museum numbers are in priuate collections) Order Proteutheria Asioryctes nemegetensrs (Kielan-Jaworowska, 1981) Order Plesiadapiformes Plesiadapis tricuspidens (per Russell, 1964) Ignaczus graybul2ianu.s (see list in text) Order Primates Leptadapis magnus (per Stehlin, 1916) Adapis parisiensis (per Stehlin, 1916) Notharctus tenebrosus (per Gregory, 1920) Lemur fuluus Loris tardigradus Tarsius hancanus Order Scandentia Tupaia tuna Ptilocercus lowii [USNM (MA) 481103 and 4880681 Order Dermoptera Cynocephalus uariegatus [USNM (MA) 317119 and 4810111 Order Chiroptera Pteropus uanipyrus [USNM (MA) 83277 and 4775871 Nyctzmene albioenter WSNM (MA) 5432371 N major LUSNM (MA) 1246351 Macroglossus lagochilus [USNM (MA) 5431991 Rhcnopoma mzcrophyllum iUSNM (MA) 2824581 Taphozous nudiuentris LUSNM (MA) 3002101 Noctilio albiuenter (USNM (MA) 4071361 Lauza f r o m LUSNM (MA) 164921 and 4528361 Rhanolophus chaseni [USNM (MA) 352581 R euryale [USNM (MA) 4762611 Eptesicus fuscus Myotis lucifugus Order Lipotyphla Echinosorex gymnurus lUSNM (MA) 4488601 Erznaceus europaeus Tenrec ecaudatus [USNM (MA) 328589 and 3416181 Institutional abbreviations The following abbreviations are used: ANS, Academy of Natural Sciences, Philadelphia; MNHN, Museum National d’Histoire Naturelle, Paris; UM, University of Michigan, Museum of Paleontology; USNM (MA), National Museum of Natural History, Smithsonian Institution, Division of Mammals; and USNM (VP), National Museum of Natural History, Smithsonian Institution, Division of Vertebrate Paleontology. Materials Most of the observations reported here are based on USNM (VP) 421608, a nearly complete skull (Kay et al., 1990).This skull was found at Houde site 24 (near University of Michigan fossil vertebrate site SC 1251, SW 1/4 of NW 1/4 of section 24, T56N, R102W, Clark Quadrangle. The site is in the Cardiolophus radinskyi range zone described by Gingerich (19911, Wasatchian, Sand Coulee beds, Clarks Fork Basin, Park County, Wyoming. Another specimen consulted in making the reconstructions that appear here [USNM (VP) 2147771 includes a lower jaw and parts of the face of a n immature animal from Houde site 14 (within UM site SC 26), N 112 of section 4,T55N, RlOlW, from the same range zone. Additional information comes from UM 65569, the rostra1 part of a skull with most of the dentition from the Sand Coulee beds, Willwood Formation, early Eocene, described by Rose and Gingerich (1976). Table 1lists the comparative material used in this study. Cranial anatomy of lgnacius The following description of the cranial anatomy of Ignacius graybullianus is based on USNM (VP) 421608 unless otherwise stated. Cranial dimpnsions of this specimcn are given in Table 2. Terminology follows IGNACIUS AND PRIiMATE ORIGINS TABLE 2. Skull dimensions of Ignacius graybullianus’ Dimension Interorbital breadth Prosthion-nasion Internasal suture length Dental arcade length Preorbital rostrum length Dental arcade width (maximum) Prosthion-inion USNM(W) 421608 UM 65569 19.5 21.@ 16.4 22.12 16.3 14.4 48.2’ - 21.1 28.1 - - ‘All measurements are in millimeters. Measurements and measuring techniques follow those of Kay and Cartmill (1977). a Based on the skull reconstruction. 479 They may be scratch marks of a small predator or scavenger. Dentition The upper canine through M3 are preserved in this specimen. Dental morphology of Ignacius graybullianus is described by Rose and Gingerich (1976) and Bown and Rose (1976). Face and orbit The facial skeleton of the new specimen IS very similar to that described in UM 65569 (Rose and Gingerich, 1976). The snout tapers markedly rostrally. The nasals are long and convex mediolaterally. They widen State of preservation caudally to form a broad contact with the The skull (see stereophotographs in Figs. frontals at a point caudad to the anterior 1 4 and reconstruction in Fig. 5) was re- orbital margin and at a similar level a s the moved from a limestone matrix by acid etch- maxillofrontal suture. Although the preing. It is slightly crushed in the interorbital maxillae were lost before fossilization, their region, with the face displaced dorsally rela- outlines and sutural contacts are clear. The tive to the braincase. The right and left pre- premaxillae overlapped the nasals medially maxillae are not preserved. The medial and butted against the maxillae caudally. sides of both orbits are badly fragmented They tapered caudally and are widely sepaand their sutures obscured. The zygomatic rated from the frontals by a maxillonasal arches, dorsal cranial vault, and occipital re- contact. This is unlike the condition region are complete. The midline basal ele- ported in UM 65569 where the premaxilla ments of the skull and lateral pterygoid contacts the frontal (Rose and Gingerich, plates are well preserved, as are the bony 1976).Reexamination of this specimen leads walls of the middle ear and the articulations us to conclude that the location of the suwith surrounding bones. The premaxilla tures in this area is confounded by cracks in was reconstructed in Figure 5 based on UM the critical areas. The maxilla makes a broad contribution 65569 (Rose and Gingerich, 1976). The lower jaw is reconstructed based on USNM to the orbital margin and wall. Its sutural (VP) 214777, and the palate is well pre- contacts in the orbit are obscured by breakserved on the skull. The canines and cheek age. The maxilla excludes the lacrimal from teeth are preserved on both sides. The post- contact with the zygomatic. The suture becanines were acid etched during prepara- tween the maxilla and zygomatic is broad tion, slightly obscuring some anatomical de- and extends obliquely from rostrodorsal to tail. The teeth appear worn due to slight caudoventral. There is a large infraorbital acid etching, leaving the mistaken impres- foramen dorsal to P3.The orbital foramen of sion that the skull was of a n old individual, the infraorbital canal is roofed by a shelfwhich clearly is not the case, as the distinct- like process of the maxilla. The lacrimal is ness of the cranial sutures demonstrates. confined to the antorbital rim and lacks a Part of the right auditory bulla was removed distinct facial process. The lacrimal has a in the course of preparation to reveal details rugose orbital rim with a lacrimal tubercle. The lacrimal foramen opens posterolaterally of the middle ear (Figs. 6,7). A number of more or less perpendicular within the orbital rim. A foramen medially scratches occur on the dorsal aspect of the adjacent to M3 could be the caudal palatine right parietal. These match a similar set on or sphenopalatine foramen, which would the orbital aspect of the same bone (Fig. 8). mark the presence of the palatine in the or- the Nomina Anatomica Veterinaria, a s cited in Miller2 Anatomy of the Dog (Evans and Christensen, 1979). 480 R.F. KAY ET AL. Fig. 1. Ignacius graybullianus [USNM (VP) 4216081. Stereopair of skull in ventral view bit. A similar foramen occurs in UM 65569. ble between the foramen ovale and the orHowever, the sutural contacts of the pala- bital fissure, indicating that neither a n alar canal nor a foramen rotundum was present. tine cannot be observed in either specimen. The orbits are widely separated by a broad The latter is probably confluent with the orinterorbital space. The orbit was small but bital fissure. A foramen rotundum has been its precise limits are unclear. Laterally and described for Plesiadapis (Russell, 1964) but ventrally, the raised orbital margin merges we believe that this foramen in Plesiadapis smoothly with the sharply edged dorsal (MNHN CR 965) is more likely the suboptic margin of the zygomatic arch. Medial and fnramen. The alisphenoid IS not strongly in dorsal t o the lacrimal tubercle, the orbital Slated laterally in this region, implying that margins merge smoothly with temporal the middle cranial fossa was relatively crests. There is no postorbital process o r bar. small. No optic foramen is visible in USNM There is no trace of a n orbital process of the (VP) 421608 but a small foramen, possibly zygomatic bone. the optic foramen, is present in the ventroThe precise limits of the bones of the or- caudal part of the orbit on UM 65569. The bital wall are not discernible in USNM (VP) small size of this foramen in Ignacius would 421608. However, the absence of sutures in imply that the eye must have been small like well-preserved orbital fragments when that of Erinaceus and much smaller than other cranial sutures are so clearly defined seen in living or fossil primates. A foramen suggests that the frontal must have ex- that probably transmitted the frontal diploic tended far down the lateral wall of the vein occurs in the right orbit of UM 65569; braincase in the infratemporal fossa and breakage obscures this region in USNM probably contacted the maxilla broadly, as (VP) 421608. in Plesiadapis (Russell, 1964). The foramen ovale is lateral t o the root of the lateral Neurocranium pterygoid plate and faces ventrally. Further anteriorly is a large, oval, forwardly directed The postorbital constriction is extreme in orbital fissure. There are no foramina visi- the region of the frontal just rostra1 t o the IGNACIUS AND PRIMATE ORIGINS 481 Fig. 2. Ignacius graybullianus [USNM (VP) 4216081. Stereopair of palate and basicranium from ventrolateral view back of the skull, where it is confluent with sharply raised nuchal crests. The point a t which the temporal lines converge appears to be further rostra1 than in UM 65569. Above and below the suture between parietal and squamosal on the side wall of the braincase are several vascular foramina that open into shallow, caudally directed surface grooves. These probably connected with intracranial dural venous sinuses. A large suprameatal foramen (Novacek, 1986) is present on the lateral side of the zygomatic arch above the mandibular fossa, posMasticatory anatomy sibly for a vein that was confluent with the The zygomatic arches are board dorsoven- intracranial venous sinuses. trally and bowed laterally. The zygomatic The mandibular fossa of the squamosal is articulates with the squamosal along a n ob- large, transversely concave, and somewhat lique suture. Ventrally the suture reaches elongate mediolaterally. Medially, the caudally to the caudolateral border of the squamosal laps onto the craniolateral side of mandibular fossa. Lateral flaring of the zy- the auditory bulla. The mandibular fossa is gomatic arches indicates that the temporal bordered caudally by a weak retroarticular muscles were large. Temporal lines con- (=postglenoid) process and the cranial parts verge strongly to form a single sharp mid- of the auditory bulla and external auditory sagittal ridge beginning on the frontal bone meatus. A well-developed retroarticular a t the level of the maximum postorbital con- (=postglenoid) foramen occurs medial to the striction. This crest extends distally to the retroarticular process. The meatal surface of suture between the frontal and parietal. This marks the level of the most posterior extension of the olfactory bulbs and the most anterior extension of the cerebral hemispheres. The olfactory bulbs were probably large, as indicated by the great breadth between the orbits. The cranial cavity is partly filled by matrix. However, a n ossified tentorium cerebelli is visible through the foramen magnum, as is a large subarcuate fossa on the left side. 482 R.F. KAY ET AL. Fig. 3. lgnacius graybullianus [USNM (VP) 4216081. Right: Skull in dorsal view. Left: Facial skeleton in coronal view. Fig. 4. Ignucius gruybullianus [USNM (VP) 4216081. Stereopair of braincase in posterolateral view the squamosal is extremely short mediolaterally. The squamosal laps caudally onto the mastoid process. The palate is broad between the molars and lacks palatal fenestrae. The palate is delineated by a sharply raised edge rostral to the cheek teeth that extends rostral and medial to a small single-rooted canine. Several transversely oriented palatal ridges are visible in front of the cheek teeth. The palatine bone reaches forward to P4. Palatal processes extend distally from the palate behind M3, bordering the greater palatine notch. On the left side of the skull, this notch is bridged by the broken and displaced lat- eral pterygoid plate. Anterior and middle palatine foramina are indistinct. The posterior margin of the palate is nearly straight and buttressed by a strong rounded postpalatine torus. The vomer is extended caudally into a high midline keel in the fossa between the medial pterygoid plates. The sphenoidal processes of the palatine are strong. Medially they are continuous with short and poorly defined medial pterygoid processes. Strong and deep lateral pterygoid plates occur laterally. On each side they reach caudally t o the medial side of the foramen ovale and the lateral side of the auditory bulla. IGNACIUS AND PRIMATE ORIGINS 483 Fig. 5. Reconstruction of t h e cranium and mandible of Ignaczus graybullzanus based on USNM (VP) 421608 and USNM (VP) 214777. A Lateral view. B: Dorsal view. C: Ventral view. The lateral pterygoid plates diverge widely, producing large fossae for the medial pterygoid muscles. A foramen is present in the lateral pterygoid plates opposite the foramen ovale, which may have transmitted the nerves to the medial pterygoid muscle. Basicranium and braincase The auditory bulla and circumbullar bones and foramina are well preserved. The bulla is large and completely ossified. It is continuous laterally, with a fully ossified ec- R.F. KAY ET AL. 484 , Foramen ? for newes to medial pterygoid muscle Gutter for "auditory" tube Promoiitoriuni Epityiupanic recess PeLrosal-entotympanic suture Tubular ectotympanic Promontory shield Entry for Carotid foramen Hypoglossal foramen Fig. 6. Ignacius graybullianus [USNM (VP) 4216081, ventral view of right auditory bulla and basicranium totympanic that is prolonged laterally into an ossified external auditory meatus. The ectotympanic is fused to the bulla, but parts of its sutures are visible. In Plesiadapis, the middle ear cavity is prolonged laterally ventral to the ectotympanic to form a subtympanic recess (Fig. 15B in MacPhee and Cartmill, 1986), whereas, in Ignacius, the tympanic ring is fused with the bulla, and no subtympanic recess is apparent. The bulla extends medially close to the midsagittal plane. Caudally, it balloons out over the basioccipital and obscures the lateral parts of the occipital condyles. The bulla is composed of an entotympanic bone and lacks basisphenoid, basioccipital, ectotympanic, or petrosal contributions (Fig. 6). Three observations indicate this. 1) Sutures separate medial, anterior and posterior bullar walls from the basisphenoid and basioccipital. 2 ) Rostrally and caudally, partial sutures are visible between the ectotympanic (which forms the external auditory meatus) and the bony bulla. The more cranial of these shows a small depression for the tympanohyal. 3) In both bullae, on the roof of the tympanic cavity, the promontorial part of the petrosal is separated from the bulla medially by a suture (Fig. 7 ) . The petrosal partly overlaps the bone of the bulla in this area. Both petrosal and entotympanic thin out at their junction rather than butting up against one another as might appear in postmortem breakage. The position of the suture is identical on right and left sides. These observations suggest that there is a true suture be- ZGNACZUS AND PRIMATE ORIGINS Fig. 7. Scanning electron micrograph of a cast, of the inside of the right auditory bulla of USNM (VP) 421608 showing part of the right promontorium (petrosal) and entotympanic bones. The orientation is similar to that in Figure 6. The ventral overlap of the petrosal onto the Fig. 8. Scratch marks on the right side of USNM (VP) 421608. See text for discussion, Midsagittal line is towards top of page. tween the petrosal and bulla and that the auditory bulla is composed of the entotympanic. 485 entotympanic is matched by a depressed area on the central surface of the entotympanic. This and the thinning of the pretrosal edge indicate that this is a suture rather than a crack. The overall morphology of this suture is the same in the left auditory bulla. The principal blood supply to the brain was probably from the vertebral arteries via the foramen magnum. A significant supply from the internal carotid artery (ICA) is ruled out by the small size of the carotid foramen. This foramen is located ventromedial to the “dual” stylomastoid foramen (see below) on the caudolateral side of the bulla just medial to the suture between the entotympanic and ectotympanic (Fig. 9). The diameter of the right carotid foramen of USNM 421608 is -0.17 mm, comparable to that or lorises (Table 3, Fig. 10) in which the. internal carotid artery degenerates during development and only the plexus of nerves that travels with the internal carotid artery is enclosed by a bony canal. In contrast, in a ‘pecirnen Of Erinaceus Of length, the diameter of the internal carotid groove that transmits a patent ICA is 0.43 mm. The small size of the foramen in Ignacius suggests that it transmitted only the nerve plexus (Fig. 10). R.F. KAY ET AL. 486 Mastoid foramina al stylomastoid foramina Suprameatal foramen Entotympanic ' I Entotympanic- ectotympanic suture Fig. 9 Ignacius graybullianus (USNM VP 421608). Right side of neurocranium, auditory bulla a n d zygomatic a r c h viewed posteriorly. T A B L E 3 . Internal diameter of carotid foramen in selected primates and other mammals Taxon Specimen Carotid foramen Skull length ~~~ Ignacius sp. Erinaceus sp. Tupaia glis (n = 3) Tupaia tuna Nycticebus coucang Perodicticus potto Galago senegalensis Galago demidouii Lemur fuluus Lemur sp. USNM (VP) 421608 Cartmill coll. Cartmill coll. Cartmill coll. USNM (MA) 488076 USNM (MA) 450056 Cartmill coll. Cartmill coll. Cartmill coll. Kay coll. Tarsius sp. CaZlithrix argentata Callicebus sp. (Bolivia) Saguinus mystax Saimiri sciureus Aotus tnuirgatus Cebus apella Pithecia pithecia Ateles geoffroyi Alouatta pigra Kay coll. USNM (MA) 239457 A N S 14306 USNM (MA) 397877 USNM !MA! 398676 USNM [MA) 396724 USNM (MA) 261319 USNM (MA) 545891 USNM (MA) 211240 USNM (MA) 544508 0.17 0.43' 0.62 0.81 0.453 0.18 0.20 0.14 0.68 0.75 0.56 0.74 1.05 0.99 1.24 1.30 2.00 1.37 2.42 2.50 ~ 48.2 45.0 49.1 59.4 59.2 61.9 46.8 37.3 87.3 83.7 39.3 45.7 59.0 50.7 59.7 62.7 95.5 82.6 102.9 104.4 'All foramen measurements were made using Wild M5 microscope fitted with an eyepiece reticle a t either 25 x or 50 x magnification. 'This is a measure ofthe breadth ofthe internal carotid sulcus immediately proximal to Its bifurcation into stapedial and promontory arteries 'The internal carotid foramen 1s slit-like; the maximum diameter is 0.45 mm. A significant blood supply to the brain from the ascending pharyngeal artery, proposed by MacPhee et al. (19831, is also unlikely. There is no groove on the basioccipital or basisphenoid along the cranial base that would indicate a n extrabullar course (sensu Wible, 1984) for ICA, nor is there a petrooccipital fissure (=middle lacerate foramen) to transmit the ascending pharyngeal artery from the neck to the inside of the braincase as in lorises (Cartmill, 1975). Foramina and grooves identified by MacPhee et al. (1983) for these structures may be for a ventral petrosal venous sinus extending intracranially in the fissure between the medial side of the petrosal and the basisphenoid and basioccipital (Szalay et al., 1987). The carotid foramen is situated lateral to the middle ear cavity so that the sympathetic nerve plexus associated with the invo- IGNACIUS AND PRIMATE ORIGINS I . , I6 , . . , 17 . , 1.8 , 19 . . . . 2 Log Skull length Fig. 10. Lug-lug bivariate plot of size of the carotid canal versus skull length based on the taxa listed in Table 3. The line above the points represents a slope of isometry (1.0). Filled circles represent small to mediumsized anthropoids, Tarsius, and lemurs; open squares represent Erinaceus and two species of Tupaia; open circles represent four galagines and lorisines. Ignacius is the solid circle within the lorisid cluster. A polygon is drawn around species whose carotid canal transmits an internal carotid artery (above); another polygon is drawn around the lorises and galagos whose carotid foramen transmits only a sympathetic nerve plexus in the adult. luted internal carotid artery entered the bulla posterolateral to the fenestra cochleae. The course of the carotid canal close to the carotid foramen is obscured by matrix and a lamina of bone that projects ventrolaterally from the promontorium obscuring also the fenestra cochleae in ventral view. It is possible that this lamina formed ontogenetically when a medially expanding middle ear cavity enveloped the short bullar portion of the carotid canal. This distinctive feature was previously reported for Ignacius, Plesiadapis, and Phenacolemur (Szalay et al., 1987). The free edge of the lamina may contain a small carotid canal in Plesiadapis (Szalay et al., 1987; but see MacPhee and Cartmill, 1986). Matrix obscures the lateral aspect of the promontorium, and it is unclear if there is a stapedial groove. Medially, there is no groove on the promontorium for a promontorial artery. A septum of bone (longitudinal septum; Szalay, 1972) stretches rostrally from the promontorium to a point just medial to the entrance of the auditory tube. Szalay (1972) suggested that the promontorial artery may 487 have traveled through a similar septum in Phenacolemur. However, this septum shows no lumen rostrally where it is broken and is thin caudally. Phenacolemur also lacks a lumen (MacPhee and Cartmill, 1986). Moreover, in Ignacius, rostrally this longitudinal septum projects laterally away from the position of the cerebral arterial circle where the promontory artery commonly ends. In Ignacius and other Plesiadapiformes, the auditory bulla is extremely large. Right and left bullae are separated medially by only 1 4 mm. and the basisphmoid and basioccipital bones are narrow. Caudally, the bulla extends ventrally over the jugular foramen and hypoglossal foramen. The promontorium is laterally positioned within the auditory bulla owing to the enormous medial expansion of the hypotympanic cavity a s in Phenacolemur, Plesiadapis, and adapids (Szalay et al., 1987). Plesiadapiformes differ from adapids in having a narrow basisphenoid across which the right and left bullae nearly touch, whereas in adapids the basisphenoid is much broader and the bullae are widely separated (compare Fig. 1 ofIgnacius with Adapis, Fig. 8, in Gingerich and Martin L198ll). A single hypoglossal foramen is adjacent to the occipital condyles. A large jugular foramen is posterior to the bulla and lateral to the hypoglossal foramen. The former transmitted the internal jugular vein and cranial nerves IX, X, and XI. Further lateral is a recess that contains two foramina. One of these is the stylomastoid foramen, and the second is associated with a vascular trace. It could be the opening for the stylomastoid artery connecting the occipital or posterior auricular artery with the ramus posterior of the stapedial artery (Wible, 1987). A similar “dual” stylomastoid foramen is reported in Plesiadapis (Russell, 1964). Mastoid foramina are present dorsolaterally in the nuchal plane between the mastoid and occipital. This foramen may have transmitted the large diploic artery (Wible, 1987) or a vein (Novacek, 1986). Either or both the stylomastoid artery and the large diploic artery could have served as a n additional source of blood for the brain or other areas supplied by the distal branches of the stapedial artery. 488 R.F. KAY ET AL. The braincase bears several crests and surfaces for the attachment of the nuchal muscles. A midline sagittal crest merges with strong nuchal crests on the parietal bones. Laterally, the occipitals extend onto the lateral wall of the braincase from the nuchal crests. Still further laterally, the parietals again reach caudad to the nuchal crest. Here, they contact broad nuchal exposures of the mastoids and do not extend on the caudal aspect of the skull. The foramen magnum is broadly oval and faces directly posteriorly. The diameter of the foramen magnum is 4.8 mm midsagittally and 6.8 mm mediolaterally. Strong paroccipital processes occur immediately adjacent to the dorsal occipital facets. formal phylogenetic analyses because the focus of our study is the skull of Ignacius. The outcome of the analysis is thus based only on a portion of the available evidence for most taxa. Our results for the affinities of all other taxa included in our analysis remain tentative and are open to testing on the basis of noncranial material. Version 3.0 of the PAUP program (Swofford, 1989) was used to analyze the data matrix of Table 4 based on the characters of Appendix A. Because the full character/ taxon matrix exLeedo ihe limits of the exhaustive search and branch-and-bound search algorithms of PAUP, we approached the analysis in several steps. 1. As a starting point for the analysis, we accept the monophyly of the following To investigate the phylogenetic position of groups: Primates, Scandentia, MicrochiIgnacius, we surveyed the cranial anatomy roptera, Megachiroptera, and Erinaceomorof extant and some fossil archontans and pha in accordance with the consensus view lipotyphlan insectivores (Table 1).We based current in the recent literature of the monoobservations of extant taxa on original ma- phyly of these clades. Next, we condensed terial but relied primarily (but not exclu- the taxa composing each of these clades into sively) on the descriptions of Gregory (1920) a single representative. The scores for these for scoring Notharctus, Stehlin (1912) and hypothetical ancestral representatives were Forsyth-Major (1901) for scoring Adapis, those character state distributions at each of and Russell (1964) for scoring Plesiadapis. the basal nodes in a n initial heuristic To determine if Ignacius is a member of Ar- search. The character matrix for these hypochonta, we also studied two erinaceomorph thetical taxa (HTUs) is given in Table 5. In insectivores (Erinaceus and Echinosorex) addition to the five clades mentioned, Chiroptera (Megachiroptera and Microchiand a tenrecomorph (Tenrec). We chose Asioryctes a s the single out- roptera) and Lipotyphla ( Erinaceomorpha group for our analysis because it is cranially and Tenreci were also found t o be monophjilwell-known and was probably close to the etic in the heuristic analysis, but we did not ancestry of all later placental mammals use condensed HTUs for these possible (Kielan-Jaworowska, 1984). Several skulls clades in later analyses because their monohave been described ofAsioryctes and closely phyly has been challenged in the literature related Kennalestes (Kielan-Jaworowska, (see Fig. 12). 2. An exhaustive search was executed on 1981). These taxa are the only proteutherians for which adequate cranial material is Ignacius, Cynocephalus, and Tenrec, toknown. Palaeoryctids, once considered to be gether with HTUs representing Erinaceoancestral placentals, are now generally morpha, Primates, Scandentia, Microchithought to be lipotyphlan insectivores (Mc- roptera, and Megachiroptera. This search Kenna et al., 1988; Butler, 1988; Thewissen resulted in a single most-parsimonious claand Gingerich, 1989). The great antiquity of dogram of 61 steps and a consistency index Asioryctes (Santonian-Coniacian, late Cre- of 0.69 (Fig. 11).We found 21 cladcgrams of taceous) is consistent with its position as 62 steps. 3. Plesiadapis was included in a second close to the ancestry of placentals. Each of these taxa was scored for 33 cra- exhaustive search resulting in ten most-parnial characters (Table 4). Dental and post- simonious cladograms of 67 steps and a concranial characters do not form part of the sistency index of 0.59. Plesiadapis and IgPHYLOGENETIC ANALYSIS 0 0 ~ 0 0 0 0 ~ 0 3 - ~ 3 0 0 0 0 ~ - 0 0 0 0 -- - 0 0 0 ~ 3 3 3 3 ~ 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 3 0 0 3 m 3 C . 3 0 0 0 3 3 ri i 0 4 c. m c? --- ri 0 3 3 r( 3 o i3 3 3 3 3 3 NC. N 3 3 ri ~ 3 0 0 0 1 r i 3 0 0 3 ~ 0 0 ~ 0 3 3 3 0 ~ 3 ~ c.e.3 - 3 ri ri c-. 0 0 0 0 0 000 0 0 3 0 0 0 0 0 - 0 0 0 0 0 - 0 0 0 0 0 - 0 0 0 0 0 0 0 13 - 3 4 3 0 ri 0 0 0 0 0 3 0 ri a w I 0 0 0 r- 0 ?. c-, 3 000 3 0 0 00 c-. C. c. 0 0 0 c C-. c-0 c . 0 e -o c -0 0 0 0 C. 0 0 0 - - 0 0 c 0 N N N N N N N N c. 3 ri ri N N N 3 C - N 3 3 0 ri 0 e -0 3 3 3 i 3 R.F. KAY ET AL. 490 ,Ignacius P i c 0 0 0 - ori100c 0 0 O O 0 t Cynocephalus Tenrec % Erinaceomorpha Primates Scandentia - 0 c 0 0 c 13000c , ,Microchiroptera Y Megachiroptera mm.lmi-* 1 0 0 0 o c 0 3 0 - 1 c Fig. 11. Maximum-parsimony cladogram of the characters in Appendix A. This cladogram is discussed in the text as analysis 2. 0 0 0 0 3 c 0--00c 1 1 d d N C 0 0 0 - o c 0 - c 0 0 c 1 0 0 0 o c C N N r i N C 0 3 0 0 c c O N N O - I C 0 0 0 0 0 - 0 0 0 0 o c c c, -- - i oc000r 3 0 0 - 1 c d d - - N N 3 0 0 0 0 0 1 i i 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 1 0 0 3 1 1 1 1 nacius formed a monophyletic clade in nine of these ten most parsimonious cladograms. In the remaining cladogram they formed a paraphyletic base taxon to all other included groups except Dermoptera. There seems to be overwhelming evidence for monophyly of Paromomyidae and Plesiadapidae to the exclusion of Cynocephalus. This is borne out not only by the cranial evidence subsumed within our analysis but also by dental synapomorphies described in the literature but not used in our analysis (Gingerich, 1976; Kay and Cartmill, 1978; Szalay and Delson, 1979). Therefore, a n additional HTU was created from the character distribuliuris "1' the basal node that unites Plesiadapiformes. Using this HTU with the others mentioned in this paragraph, a further PAUP exhaustive search yields a single most parsimonious cladogram of 59 steps and consistency index of 0.71. This cladogram is identical to the cladogram in Figure 11 (with Plesiadapiformes substituted for Ignacius) and represents in our opinion the best supported hypothesis of the relations of Ignacius and Plesiadap is. 4. Several hypotheses have been proposed for the relations among Archonta. These are tested with our data by calculating the minimum number of transformations based on our character matrix when these tree topologies are forced. We evaluated the cladograms proposed by Gregory IGNACIUS AND PRIMATE ORIGINS C ynocephalus Microchiroptera Megachiroptera Primates Scandentia Tenrec Erin aceomorph a Gregory, 1910 length: 56 shortest length: 54 Microchiroptera Megachiroptera Scandentia Erinaceomorpha Primates Miyamoto and Goodman, 1986 length: 46 shortest lenght: 39 49 1 Cynocephalus Microchiroptera Megachiroptera Primates Scandentia Tenrec Erinaceomorpha Novacek and Wyss, 1986 length: 56 shortest length: 54 Cynocephalus Megachiroptera Primates Scandentia Microchiroptera Pettigrew et al., 1989 length: 48 shortest length: 41 Fig. 12. A test of four published cladograms representing the phylogenetic relationships proposed by Gregory (1910; Fig. 31), Novacek and Wyss (19861, Miyamoto and Goodman (19861, and Pettigrew et al. (1989).Each cladogram is presented as a user tree in PAUP, and the characters in Table 4 are plotted onto it. Below each is given the number of steps (length)of each tree compared with the shortest tree using our characters and the taxa presented. (1910), Novacek and Wyss (19861, Miyamoto et al. (1986), and Pettigrew et al. (1989; Fig. 8). These authors did not include all of the same taxa in their analyses a s we did, so we also executed a n exhaustive search for the most parsimonious cladograms for their subsets of taxa. Cladograms, numbers of steps, and consistency indices are summarized in Figure 12. DISCUSSION OF CLADISTIC ANALYSIS We will restrict our discussion to the single most parsimonious cladogram in the exhaustive search without Plesiadapis (search 2), the 21 cladograms of the same analysis that were one step longer, the ten most parsimonious cladograms of the analysis that included Plesiadapis (search 3), and the single most-parsimonious cladogram based on the HTU for Plesiadapiformes. Our most significant results are that 1) Ignacius and Plesiadapis are sister taxa, 2) Plesiadapiformes may collectively group with Dermoptera, and 3) Ignacius and Plesiadapis are never linked with Primates. We will also discuss 4)the sister-group relation of Primates and Scandentia and 5) the concept of Archonta. R.F. KAY ET AL 492 lgnacius and Plesiadapis form a clade, Plesiadapiformes ply the brain and that the ventral petrosal sinus is endocranial. Ignacius does not differ in these features from Plesiadapis. One potential difference between Ignacius In nine of ten cladograms of search 3, Plesiadapis was the sister group to Ignacius, and Plesiadapis is the composition of the whereas, in the remaining cladogram, these bulla. The auditory bulla (character 24) of two taxa formed a paraphyletic root group to Ignacius is completely ossified by the entoall taxa in the analysis (except Cynoceph- tympanic bone (interpreted a s a synapomoralus). This corroborates the monophyly of phy of Ignacius and Plesiadapis in 11 of 22 the two core families of Plesiadapiformes: cladograms). Plesiadapis reportedly has a Plesiadapidae and Paromomyidae. Depend- petrosal bulla, but the detailed similarities ing on which of the ten trees is selected (each between the two taxa in the middle ear arvarying in the remaining topology of the cla- teries hugged that Plesiadapis actually had dogram), the following are possible shared a n entotympanic bulla. If the bulla of Plederived features of Plesiadapiformes: maxil- siadapis is entotympanic, then the suture lary-frontal contact in orbit (character 6; between petrosal and entotympanic must two of ten cladograms); suboptic foramen have been obliterated-a common phenomepresent (character 16; nine cladograms); os- non in ontogeny (MacPhee and Cartmill, sified external auditory meatus (character 1986). 23; eight cladograms); promontory artery absent (character 25; eight cladograms); staPlesiadapiformes are the sister pedial artery absent (character 26; eight clagroup of Dermoptera? dograms); mastoid tubercle strong (character 29; one cladogram). Our analysis weakly supports the notion Several other characters not used in the of a monophyletic Plesiadapiformes as a sisanalysis appear to further support the hy- ter group to Dermoptera but does not suppothesis of plesiadapiform monophyly. Ad- port the hypothesis of Beard (1990) that ditional cranial similarities of Ignacius to Paromomyidae (but not Plesiadapidae) and Plesiadapis, the cranially best known paro- Cynocephalus (Order Dermoptera) are sismomyid and plesiadapid respectively, in- ter groups and its corollary that Plesiadapiclude the carotid foramen being situated in a formes is a paraphyletic group. Three basic posterolateral position of entry into the patterns emerge from our analyses of the bulla in both taxa and the fenestra cochleae relations of Plesiadapiformes. In the analybeing shielded ventrally by a bony shelf (see, sis without Plesiadapis (search 21, Dere g , Gingerich, 1976; Kay and Cartmill, moptera is the sister group ofIgnacsus in ? I 1977; Szalay and Delson, 1979; and Szalay of the 22 most-parsimonious cladograms (toet al., 1987).Ignacius and Plesiadapis, as do pology A) or of Lipotyphla in nine claall other known paromomyids and ple- dograms (topology B). In the remaining two siadapids, also share apomorphies of the cladograms, Ignacius is the primitive sister dentition including the presence of a Nanno- to a clade consisting of all other taxa except pithex fold on the upper molars, reduction in Cynocephalus (topology C). Topologies A the number of anterior teeth and a n en- and C occur only once in the analysis that larged lanceolate lower incisor (Gingerich, included Plesiadapis, while topology B was 1976; Kay and Cartmill, 1977; Szalay and common (seven cladograms). Thus these results are ambiguous with respect to the conDelson, 1979). Moreover, the skull oflgnacius, unlike the clusion of Kay et al. (1990) that Plesiadapiprevious reconstruction suggestions of formes and Dermoptera are closely related. MacPhee et al. (1983), demonstrates that Possible characters supporting monothe arteries of the basicranium are very sim- phyly of Plesiadapis, Ignacius, and Cynoilar to those of Plesiadapis. The new mate- cephalus are ossified auditory meatus (charrial demonstrates that Ignacius has no acter 23; 11 of 22 cladograms), promontory petrooccipital fissure and implies that the artery absent (character 25; 11 of 22 claascending pharyngeal artery does not sup- dograms), and stapedial artery absent IGNACIUS AND PRIMATE ORIGINS (character 26; 11of 22 cladograms). Another possible cranial synapomorphy not included in this parsimony analysis also suggests a dermopteran-plesiadapoid clade. While many living or extinct mammals have or had a n ectotympanic bulla, Cynocephalus and Ignacius share one apparently unique feature for this structure. In both taxa, the entotympanic bone apparently contacts the basioccipital medially (see cross-sectional diagram in Hunt and Korth, 1980, for Cynocephalus). In other mammals with entotympanic bullae, the entotympanic contacts only the petrosal, which in turn contacts the basioccipital. This hypothesis of the affinities of Ignacius is also being debated with reference to postcranial evidence not used in our analysis (Beard, 1990; but see Krause, 1991). In summary, the link between Dermoptera and Plesiadapiformes is plausible but weak: Dermoptera was the sister group in one-half of the most-parsimonious trees when Plesiadapis was excluded, and Lipotyphla was the sister group to Plesiadapiformes in most analyses that included Plesiadapis. This link needs to be further evaluated on the basis of new fossil material for Plesiadapiformes and Dermoptera. A Plesiadapiformes-Primate link? In our analyses, Plesiadapiformes was never the sister group to Primates, a widely held view in the past (see, e.g., Szalay and Delson, 1979). For more than 100 years, paleontologists have recognized Eocene primates on the basis of synapomorphies with living primates, including especially a postorbital bar, a n orbital convergence, a welldeveloped internal carotid artery in a promontorial position enclosed in a bony canal, a n auditory bulla composed of the petrosal, a Nannopithex fold on the upper molars, nails rather than claws on some digits, and a n opposable hallux. There is continued debate about which nonprimate group is sister to Eocene-Recent primates, and Plesiadapiformes are often nominated to this position. One thorough recent analysis (Szalay et al., 1987) lists nine shared-derived features uniting Plesiadapiformes with adapids and omomyids. Six characters are cranial or dental: 1)auditory bulla inflated and formed by 493 the petrosal bone, 2) ectotympanic expanded laterally and fused medially to the wall of the bulla, 3) promontorium centrally positioned in the bulla, and large hypotympanic sinus widely separating promontorium from the basisphenoid, 4)internal carotid entering the bulla posteriolaterally and enclosed in a bony tube, 5 ) Nannopithex fold on the upper molars, and 6) loss of one pair of incisors. This new information on the cranial anatomy of Ignacius puts the similarities between Plesiadapiformes and Primates in a new light. The bulla of Ignacius is entotympanic and not petrosal, suggesting that the bulla of Plesiadapis is also entotympanic. The ectotympanic is expanded laterally (tubular) and fused to the bulla only in omomyids. Adapids show the more primitive mammalian condition with the ringlike ectotympanic. This suggests that the omomyid-plesiadapiform similarity is a homoplasy. The internal carotid artery is absent in Plesiadapiformes, with the principal blood supply to the brain probably coming from the vertebral arteries. In primates, the ICA is well developed. The reconstructed last common ancestor of Plesiadapiformes and Primates had a n arterial arrangement similar to that of the inferred primitive condition for eutherians (ICA and its promontory and stapedial branches well developed) (Wible, 1987). Reduction of dental formula seems to be a common parallel occurrence among many mammalian groups and is therefore a poor guide in phylogenetic inference. Of the other features of similarity between Plesiadapiformes and Primates, the central position of the promontorium in a large hypotympanic sinus is a n interesting similarity but probably occurred by different developmental pathways. In plesiadapiforms, the brain is relatively small and the promontoria appear to be situated closer to the midsagittal plane than in primates. The central position of the promontorium within the plesiadapiform bulla seems to result from a n ontogenetic medialward hypertrophy of the hypotympanic sinus. In contrast, in Eocene primates the brain is expanded and the promontoria are carried laterally during development. In this case the central 494 R.F. KAY ET AL position of the promontorium is more likely the result of this lateral displacement inside an unhypertrophied bulla. The shared presence of a Nannopithex fold is the most likely possible primate-plesiadapiform synapomorphy. However, here it should be noted that this feature is poorly developed in the more primitive members of each clade (i.e., Teilhardina and Cantius for Primates and Purgatorius for Plesiadapiformes), suggesting homoplasy in this feature. Primates and Scandentia are a clade Our analyses support the hypothesis that Primates and Scandentia are a monophyletic group. This corroborates the once widely held view that primates and tree shrews are closely related (see, e.g., Le Gros Clark, 1959). Monophyly is supported by four characters in the single most-parsimonious cladogram: presence of postorbital bar (character 13), maxillary artery pierces ectopterygoid plate (character 20), arteries of middle ear enclosed in bony canals (character 28), and jugular foramen dual (character 32). Unity of the Archonta In only one of the 31 cladograms under consideration was Archonta a monophyletic group. The uniting characters of Archonta have always been few. Novacek and Wyss 11986) cited a pendulous penis and confluent sustentacular and navicular facets of the talus a s uniting characters. The cladograms of Gregory (1910) and Novacek and Wyss (1986) support monophyly of Archonta (Fig. 9), but these are less parsimonious than our cladograms (56 vs. 54 steps for the taxa they included). The monophyly of Archonta is in need of review. Several other suggestions as to the phylogeny of Primates and archontans can be evaluated with our data. Pettigrew et al. (1989) proposed that Megachiroptera was the sister group of Primates, a view not endorsed by our analysis of cranial data. Pettigrew’s cladogram is considerably less parsimonious than our results (48 vs. 41 steps). Miyamoto and Goodman (1986) proposed that ScandentEa is the sister group to Lipotyphla; their cladogram took 46 steps based on our character matrix, whereas our most parsimonious cladogram took only 39. ACKNOWLEDGMENTS We thank Dr. Peter Houde who collected and prepared the Ignacius specimens and kindly allowed us to study them. Figure 5 was drawn by Ms. Jo-Ellen Trecartin; Figures 6 and 9 were prepared by Dr. Bradley Smith. Photographs in Figures 1-4 were made by Mr. Victor Krantz of the Smithsonian Institution. The authors profited greatly from conversations with Drs. Matt Cartmill, Ross MacPhee, John Wible and Richard W. Thorington, J r . , about cranial anatomy and primate evolution generally. Much of the research on the fossils was undertaken by R.F.K. under a Smithsonian Institution Fellowship in 1987. Original specimens were made available for our study by the kindness of Dr. Matt Cartmill, Duke University; Drs. R. Emry and R. Thorington, Jr., Smithsonian Institution; and Dr. D.E. Russell, MNHN, Paris. LITERATURE CITED Beard KC (1990) Gliding behaviour and palaeoecology of the alleged primate family Paromomyidae (Mammalia, Dermoptera). Nature 345:340-341. Bown TM, and Rose KD (1976) New early Tertiary primates and a reappraisal of some Plesiadapiformes. Folia Primatol. 26:109-138. Bugge J (1974) The cephalic arterial system in insectivores, primates, rodents and lagomorphs, with special reference to the systematic classification. Acta Anat. 871 Suppl. 62 ):1-160. Butler PM (1948) On the evolution of the skull and teeth in the Erinaceidae, with special reference to fossil material in the British Museum. Proc. Zool. Soc. London I 18:446-50 1. Butler PM (1956) The skull of Ictops and the classification of the Insectivora. Proc. 2001. Soc. London 126t453-481. Butler PM (1988) Phylogeny of the insectivores. In M J Benton (ed.):The Phylogeny and Classification of the Tetrapods, Vol. 2: Mammals. Oxford: Clarendon Press, pp. 117-141. Cartmill M (1975) Strepsirhine basicranial structures and the affinities of the Cheirogaleidae. In WP Luckett and FS Szalay (eds.): Phylogeny of the Primates. New York: Plenum Publ. Corp., pp. 313-354. Cartmill M (1978) The orbital mosaic in prosimians and the use of variable traits in systematics. Folia Primatol. 30:89-114. Cartmill M, and MacPhee RDE (1980) Tupaiid affinities: The evidence of the carotid arteries and cranial skeleton. In PW Luckett (ed.): Comparative Biology and Evolutionary Relationships of Tree Shrews. New York: Plenum Pub. Corp., pp. 95-132. IGNACIUS AND PRIMATE ORIGINS Conroy GC, and Wible JR (1978)Middle ear morphology of Lemur uariegatus: Some implications for primate paleontology. Folia Primatol. 29:81-85. Evans HE, and Christensen GC (1979)Miller’s Anatomy of the Dog. Philadelphia: WB Saunders. Forsyth Major CI (1901)On some characters of the skull in the lemurs and monkeys. Proc. Zool. SOC.London Zt129-156. Gingerich PD (1976) Cranial anatomy and evolution of early Tertiary Plesiadapidae (Mammalia, Primates). Papers Paleontol. Univ. Michigan 15:l-140. Gingerich PD (1989)New earliest Wasatchian mammalian fauna from the Eocene of northwestern Wyoming: Composition and diversity in a rarely sampled highfloodplain assemblage. Papers Paleontol Univ. Michigan 28.1-97. Gingerich PD (1991) Systematics and evolution of early Eocene Perissodactyla (Mammalia)in the Clarks Fork Basin, Wyoming. Contrib. Paleontol. Univ. Michigan 28t181-213. Gingerich PD, and Martin RD (1981) Cranial morphology and adaptations in Eocene Adapidae. 11. The Cambridge skull of Adapis parisiensis. Am. J. Phys. Anthropol. 561235-257. Gregory WK (1910) The order of mammals. Bull. Am. Mus. Nat. Hist. 27:l-362. Gregory WK (1920) On the structure and relations of Nothai-ctus, an American Eocene primate. Mem. Am. Mus. Nat. Hist. 3t51-243. Hershkovitz P (1977) Living New World Monkeys (Platyrrhini): Vol. 1.Chicago: University of Chicago Press. Hunt RM J r , and Korth WW (1980) The auditory region of Dermoptera: Morphology and function relative to other living mammals. J . Morphol. 164r167-211. Kay RF, and Cartmill M (1977) Cranial morphology and adaptations of Palaechthon nacimienti and other Paromomyidae (Plesiadapoidea, ?Primates), with a description of a new genus and species. J . Hum. Evol. 6:19-53. Kay RF, Thorington RW Jr. and Houde P (1990) Eocene plesiadapifol-m shows affinities with flying lemurs not primates. Sature 345t342-344. Kielan-Jaworowska Z (1981) Evolution of the therian mammals in the Late Cretaceous of Asia. Part IV. Skull structure of Kennalestes and Asioryctes. Palaeontol. Polon. 42:25-78. Kielan-Jaworowska Z (1984) Evolution of the therian mammals in the late Cretaceous of Asia. Part VII. Synopsis. Palaeontol. Polon. 46:173-183. Krause DW (1991) Were paromomyids gliders? Maybe. Maybe not. J . Hum. Evol. 21:177-188. Le Gros Clark WE (1959) The Antecedents of Man. Edinburgh: Edinburgh University Press. MacPhee RDE (1981) Auditory regions of primates and eutherian insectivores: Morphology, ontogeny, and character analysis. Contrib. Primatol. 19:l-282. MacPhee RDE, and Cartmill M (1986)Basicranial structures and primate systematics. In DP Swindler and J Erwin (eds.): Comparative Primate Biology, Vol. 1: Systematics, Evolution, and Anatomy. Plenum Press New York pp. 219-275. MacPhee RDE, Cartmill M, and Gingerich PD (1983) New Palaeogene primate basicrania and the definition of the order Primates. Nature 301t509-511. 495 McKenna MC, Xue X and Zhou M (1984) Prosarcodon lonunensis, a new Paleocene micropternodontid palaeoryctoid insectivore from Asia. Am. Mus. Nov. 1278O:l-17. Miyamoto MM, and Goodman M (1986) Biomolecular systematics of Eutherian mammals: Phylogenetic patterns and classification. Syst. 2001.35:230-240. Muller J (1934)The orbitotemporal region of the skull of the Mammalia. Arch. Neerl. Zool. 1.118-259. Novacek MJ (1986) The skull of leptictid insectivorans and the higher-level classification of eutherian mammals. Bull. Am. Mus. Nat. Hist. 183:3-111. Novacek MJ, and Wyss AR (1986) Higher-level relationships of the recent eutherian orders: morphological evidence. Cladistic? 2257-287. Novacek MJ, Wyss AR, and McKenna MC (1988) The major groups of eutherian mammals. In MJ Benton (ed.): The Phylogeny and Classification of the Tetrapods, Vol. 2. Oxford: Clarendon Press, pp. 31-71. Pettigrew JD, Jamieson BGM, Robson SK, Hall LS, McAnnally KI, and Cooper HM (1989) Phylogenetic relations between microbats, megabats and primates (Mammalia: Chiroptera and Primates). Phil. Trans. R. SOC.London [Series Bl325t489-559. Rose KD, and Gingerich PD (1976) Partial skull of the plesiadapiform primate Ignacius from the early Eocene of Wyoming. Contrib. Mus. Paleontol. Univ. Michigan 24:181-189. Russell DE (1964) Les Mammiferes Paleocenes #Europe. Mem. Mus. Nat. Hist. Nat. Series C 13:l324. Stehlin HG (1912, Die Saugetiere des schweizerischen Eocanes. Kritischer Catalog der Materialien. Siebenter Teil, erste Halfte: Adapis. Abhandl. schweizerischen Palaontol. Ges. 38:1165-1209. Straney DO (1984) The nasal bones of Chzroderma (Phyllostomidae).J . Mammal. 65:163-165. Swofford DL (1989)PAUP: Phylogenetic Analysis Using Parsimony, Version 3.0. Champaign, IL: Privately published. Szalay FS (1972) Cranial morphologl of the early Tertiary Phenacolemur and its bearing on primate phylogeny. Am. J. Phys. Anthropol. 36:59-76. Szalay FS (1975) Phylogeny of primate higher taxa. The basicranial evidence. In WP Luckett and FS Szalay (eds.): Phylogeny of the Primates. A Multidisciplinary Approach. New York: Plenum Publ. Corp., pp. 91-125. Szalay FS, and Delson E (1979) Evolutionary History of the Primates. New York: Academic Press. Szalay FS, Rosenberger AL, and Dagosto M (1987) Diagnosis and differentiation of the order primates. Yrbk. Phys. Anthropol. 30:75-105. Thewissen JGM (1985) Cephalic evidence for the affinities of Tubulidentata. Mammalia 49:257-284. Thewissen JGM and Gingerich PD (1989) Skull and endocranial cast ofEoryctes melanus, a new palaeoryctid (Mammalia, Insectivora) from the early Eocene of western North America. J. Vert. Paleontol. 9,459470. Wible JR (1984) The Ontogeny and Phylogeny of the Mammalian Cranial Arterial Pattern. PhD Thesis, Duke University. Wible JR (1986) Transformations in the extracranial course of the internal carotid artery in mammalian phylogeny. J. Vert. Paleontol. 6,313-325 496 R.F. KAY ET AL Wible J R (1987) The eutherian stapedial artery: Character analysis and implications for superordinal relationships. 2001. J. Linnean Soc. 91:107-135. APPENDIX A: CHARACTERS USED IN THE CLADISTIC ANALYSES DESCRIBED IN THE TEXT. Many of the characters described below are based on Novacek (1986), Wible (1984, 19871, Novacek and Wyss (19861, and Butler (1948,1956). These authors are often not the original source for the characters they use. We have included only informative characters (i.e., those that may represent synapomorphies of some of the investigated taxa.) Characters 5, 7,10,12,13, and 15 are scored as ordered multistate. Other multistate characters are considered to be unordered. with a n r2 of 0.624. For each taxon the “expected” IOF size was calculated from this regression equation. The observed (measured) size of IOF for each species was compared with the expected and expressed a s a residual: 100 . (observed - expected)/(expected). This residual ranges from -22.1 (Loris)to + 17.6 (Erinaceus).Residuals were grouped into three categories: large (0)for taxa with residuals greater than 7.0; moderate (1)for taxa with residuals smaller than 7.0 but greater than -8.0; small (2) for taxa with residuals smaller than -8.0. Ignacius has a value of 12.1, essentially identical to the value for Echinosorex (12.2). From observation of the figures and descriptions of &e1. Primitively the caudal part of the na- lan-Jaworowska (19811, Asioryctes has a sal flares laterally (0); in the derived condi- large IOF. Plesiadapis was scored a s modertion the nasals are of uniform width and ate based on Kay and Cartmill (1977). 6. Contact between lacrimal and palahave a narrow suture with the frontal (1). The nasals are wider than long in certain tine in the orbit (0; Butler, 1956; Novacek Microchiroptera (e.g., Rhinolophus) due to and Wyss, 1986). Derived conditions include extreme shortening of the rostrum, but this a contact between maxilla and frontal (1)or does not bear on the scoring of this charac- large ethmoid exposure (2), which reduce ter. This and other sutural characters of the the size ofthe orbital process ofthe palatine. microbats are difficult to score because the The condition of this character in Ignacius is cranial bones fuse very early in ontogeny unknown. The condition in Tarsius was determined by Cartmill (1978). Tupaia and (Straney, 1984). 2. Left and right incisive foramina sepa- Ptilocercus were scored according to Le Gros rated by a bony process in the median plane Clark (1959). 7. Caudal midsagittal margin of palate (0) or fused in the midsagittal plane (1 NOCrostra1 to M3 (O), at M’ (11, caudal to M3 ( 2 ) . tilio lacks a n incisive foramen iscored as 2 j . 8. Posterior palatine torus present (0) or 3. Rostra1 edge of palate complete (0) or interrupted a t the suture between premax- absent (1).Unlike Novacek (19861, we conilla and maxilla (1).The incisive foramina sider the presence of the postpalatine torus are open rostrally a t the suture between pre- primitive because it is present in Asioryctes. 9. Lacrimal foramen surrounded by the maxilla and maxilla in certain Microchilacrimal bone (0) or lies on the lacrimal/ roptera (2). 4. Infraorbital canal long (0) or short, de- maxillary suture (1).Noctilio lacks a lacrimal foramen and consequently could not be veloped as just a foramen in the maxilla (I). 5. Infraorbital canal large. The relative scored. 10. Lacrimal foramen opens into orbit (0), size ofthe infraorbital foramen was based on measurement of its area, the product of onto orbital rim (l),or onto rostrum (2). 11. Absence (0) or presence (1) of a lagreatest diameter times diameter a t right angles to that measurement. Log IOF area crimal tubercle. The tubercle is absent was regressed against log prosthion-inion in Cretaceous proteutherians (Kielan-Jalength. The resulting least-squares regres- worowska, 1981). 12. The zygomatic is primitively large sion is and forms a significant portion of the zygolog IOF = 1.583(log skull length) - 2.673, matic arch (0). We recognize three states: IGNACIUS AND PRIMATE ORIGINS 497 zygomatic separates squamosal and maxilla cuspidens (MNHN CR 965) but was labelled (01, squamosal and maxilla share a suture a s “t.d.a.” by Russell (1964: Fig. 19). 17. Foramen rotundum absent (0) or on the zygomatic arch (implying that the zygomatic is small, scored a s (11, and zygo- present (1).Muller (1934) established that matic absent (2). In Echinosorex, maxilla the fusion of foramen rotundum with the and squamosal are in contact on the medial sphenorbital fissure is primitive for placentals. side of the zygomatic arch. 18. Lateral pterygoid plate present (0) or 13. Postorbital process of frontal absent (O), present but not touching zygomatic arch absent (1). 19, 20. Alar canal and maxillary artery. (I), or touching zygomatic arch to form a postorbital bar ( 2 ) . Intermediates between The course of the maxillary artery from the these states occur. The size ofthe postorbital basicranium to the orbit may lead through a process is variable in Ptero-pus.where i t may bony foramen or canal in the ahsphenoid. or may not reach the zygomatic process This canal is often developed a s a bridge on the lateral side of braincase and lateral (scored a s ‘‘?,,). 14. Ethmoid foramen present (0) or ab- pterygoid plate (e.g., Tenrec). In other mammals (e.g., Tarsius) the maxillary artery sent (1). 15. Optic foramen (OF) small (01, moder- may pierce the lateral pterygoid plate. Both ate (l),or large (2). The relative size of the canals or foramina have been described a s optic foramen was based on measurement of alar or alisphenoid canal, and they are its area, the product of greatest diameter equivalent in that they carry the maxillary times diameter a t right angles to that mea- artery in a bony structure from which the surement. A line was assigned to a bivariate lateral pterygoid muscle originates. Howcluster of log skull length vs. log OF area. ever, both structures can be present in the The line was assigned slope 2.0 (slope of same taxon (e.g., Rattus), so we distinguish isometry) and passed through mean log two characters. We retain the name alar canal for the caskull length and mean log OF area. The nal on the lateral wall of the braincase equation expressing this line is (character 19). Primitively this canal is Log (optic foramen diameter) = 2Uog skull present (01, but it can also be absent (1).The alar canal is present but lacks a medial wall length) - 3.167. in Megachiroptera and thus lies within the For each taxon the “expected” OF size was cranial cavity (Wible, 1984). This condition calculated from this equation. The observed is scored as 2. The alar canal is commonly (measured) size of OF for each species was prePent in Echinosorex but is sunietimes abcompared with the expected and expressed sent (Thewissen, 1985). The foramen in the lateral pterygoid plate as a residual: 100 . (observed - expected)/ (expected). This residual ranges from -93.5 differs from the alar canal in that it trans(Tenrec)to +283.1 (Tarsius).Residuals were mits the maxillary artery and/or other strucgrouped into three categories: small (0) for tures from medial to lateral side of the lattaxa with residuals less than -75.0, moder- eral pterygoid plate (character 20), whereas ate (1)for taxa with residuals greater than the alar canal does not pierce the plate. This -75.1 but less than -25.0, large (2) for taxa foramen is apparently absent in Asioryctes. with residuals greater than -10.1. Ignacius Absence is considered primitive (0). The foramen of the lateral pterygoid plate has a value of 76.6, essentially identical to is usually a simple window, but in Ptilocerthe value for Erinaceus (-76.1). 16. Suboptic foramen absent (0) or cus it is elongated to a canal. In addition to present (1).The suboptic foramen carries a the artery, this foramen may also carry the vein that interconnects the orbits (Butler, branch of the trigeminal nerve to the medial 1948). This character was scored a s 1 if the pterygoid muscle, or this nerve may travel venous canal is completely separated from through a second foramen in the lateral the optic foramen by a bony process. A sub- pterygoid plate, a s in Tupaia. Dissections optic foramen is present in Plesiadapis tri- confirm that the foramen does not carry the ~ 498 R.F. KAY ET AL. maxillary artery in Lemur, and instead carries only a branch of the mandibular branch of the trigeminal nerve. Because we cannot distinguish the soft structures transmitted on the basis of osteological evidence, we score the presence of one or more foramina in the lateral pterygoid plate a s (1).Other perforations of the lateral pterygoid plate include the foramen crotaphiticum and foramen Civini described by Hershkovitz (1977). 21. Postglenoid process present (0) o r absent (1).The process is well-developed in Asioryctes . 22. Suprameatal foramen present (0) or absent (1).This foramen dorsal to the mandibular fossa in the lateral aspect of the mandibular process of the squamosal transmits veins. It is present in Asioryctes. 23. External auditory meatus membranous (0) or bony (1).Primitively the external auditory meatus is unossified. In several taxa there is a distinct bony tube lateral to the tympanic membrane. In primates, this tube is often formed by the ectotympanic. 24. Auditory bulla unossified (O), parfully tially ossified by the ectotympanic (l), ossified by the petrosal (2), fully ossified by the entotympanic (3), or fully ossified by the alisphenoid (4).Cynocephalus was scored according to Hunt and Korth (1980), Tupaia and Ptilocercus according to MacPhee (1981). 25. Promontory artery present in adult (0) or absent in adult (1).For those taxa in which the promontory artery has involuted in the adult, Ignacius and Loris (Cartmill, 1975), Cynocephalus (Wible, 1986), and Lemur (Bugge, 1974; Conroy and Wible, 1978), we could nonetheless determine its course in fetal life by the location of the sympathetic plexus of the internal carotid artery. The location of this plexus is a reliable indicator of the course of the promontory artery. All of the studied taxa exhibit a transpromontorial course (sensu Wible, 1984) of the internal carotid artery or its sympathetic plexus. 26. Stapedial artery present in adult (0) or absent in adult (1). A groove on the caudal promontory for the stapedial artery is primitive for placentals (Wible, 1987). Grooves are not always visible for taxa with stape- dial arteries. We were able to score three taxa because their morphology relating to this character has been previously published: Myotis received a score of 0 (Wible, 19841, Pteropus a score of 1 (Wible, 19871, and Ptilocercus a score of 0 (Cartmill and MacPhee, 1980). The Lipotyphla were scored according to MacPhee (1981). 27. Vascular foramina in parietal or squamosal present (0) or absent (1).Primitively these foramina, which may be venous or arterial, are present. 28. Arteries of middle ear cavity exposed (0) or enclosed in bony canals (1). The arteries of the middle ear are primitively in ventrally open grooves. 29. Mastoid tubercle weak or absent (0) or strong (1).A weak mastoid process if present in Asioryctes. Loris and Cynocephalus both received a score of 0 even though the scoring is somewhat ambiguous because of the extreme mastoid inflation in both taxa. Pteropus and Macroglossus were scored a s 1. However, the tubercle is formed by the squamosal rather than the mastoid. 30. Mastoid foramen present in mastoid bone (O), on suture of mastoid and occipital (11, in occipital bone (21, or absent (3). Mastoid foramen is used to describe any venous foramen that drains a dural sinus and pierces the occipital part of the mastoid or its suture with the occipital bone. Such a foramen is present in the mastoid bone of Asioryctes. 31. Paroccipital process weak or absent (0) or strong (1).Asioryctes lacks a ventral process of the occipital caudal to the external auditory meatus. 32. Jugular foramen single (0) or dual (1). The jugular foramen transmits venous drainage from the dural sinuses a s well a s cranial nerves IX, X, and XI. Although scored a s having the derived state, Loris and Tupaia are potential intermediates. In both taxa, two foramina are visible within a shallow fossa. It is not clear, however, whether the two foramina are separating venous drainage from nerves o r if they are both channels for venous drainage. 33. Supraorbital foramen absent (0) or present (1).