Inés Horovitz* and The quaternary Cuban platyrrhine Ross D. E. Paralouatta varonai and the origin of MacPhee† Antillean monkeys *Department of Vertebrate Paleontology, American We describe recently recovered dental and mandibular remains of the Museum of Natural History, Cuban platyrrhine Paralouatta varonai, previously known from the New York, NY 10024-5192, holotype only (a nearly complete skull with very worn teeth). We also U.S.A. E-mail: expand on the original description of the type specimen. [email protected] and Paralouatta is one of three extinct taxa of Greater Antillean †Department of Mammalogy, Quaternary monkeys known from craniodental remains. The other American Museum of Natural two, Xenothrix mcgregori and Antillothrix bernensis, occurred in Jamaica History, New York, NY and Hispaniola, respectively. It has been common practice to assume 10024-5192, U.S.A. that Antillean monkeys were more closely related to individual mainland taxa than to each other. Thus, P. varonai was thought to be Received 23 February 1998 related to Alouatta; Antillothrix bernensis to Saimiri or Cebus; and X. revision received 20 June mcgregori to Callicebus, or to callitrichines, or even to be of unknown 1998 affinity. With the discovery of well-preserved dental remains of and accepted 6 July 1998 Paralouatta, it can now be ascertained that this species was in fact very different from Alouatta. Cladistic analysis reveals a sister-group Keywords: New World relationship between Antillothrix and Paralouatta, followed on the monkeys, platyrrhines, cladogram by Xenothrix and Callicebus (last taxon being the closest Antilles, Quaternary, mainlaind relative of the Antillean clade). This conclusion has an Antillothrix, Xenothrix, important biogeographic implication: recognition of an Antillean Paralouatta, Caribbean. clade, as advocated here, assumes only one primate colonization from the South American mainland, not several as previously believed.  1999 Academic Press

Journal of Human Evolution (1999) 36, 33–68 Article No. jhev.1998.0259 Available online at http://www.idealibrary.com on

Introduction 1984; Fleagle & Bown, 1983; Fleagle et al., 1987; Fleagle, 1990). By comparison, the Platyrrhine primates have been a notable part diversity of the endemic platyrrhines of the of the mammalian fauna of South America Greater Antilles (hereafter, the Antillean since at least the late Oligocene/early monkeys) is known only in barest outline. Miocene (Hoffstetter, 1969; MacFadden, The objective of this paper is to review 1990). The living species are found from recent progress in unravelling and interpret- southern Mexico to northern Argentina, but ing the phylogeny of these monkeys, with the fossil record establishes the New World particular emphasis on the late Quaternary monkeys once ranged more widely, from species from Cuba, Paralouatta varonai. Patagonia to the Greater Antilles (i.e., the Our understanding of the fossil history of island group consisting of Cuba, Hispaniola, Antillean monkeys may be getting broader Puerto Rico, and Jamaica). The past diver- in a geographical sense, but it is not very sity of platyrrhines at the southern end of deep temporally. Cuba, Jamaica, and their distribution is becoming increasingly Hispaniola each supported one or more well known, thanks to more than a century endemic platyrrhines during the late Quater- of dedicated work in this region (Ameghino, nary, but Cuba is the only island with a 1906; Rusconi, 1934, 1935; Kraglievich, Tertiary primate record—currently consist- 1951; Hershkovitz, 1970, 1974, 1981, ing of a single fossil, a talus of early Miocene

0047–2484/99/010033+36$30.00/0  1999 Academic Press 34 .   . . .  age from the important locality of Domo de rel monkey (Rímoli, 1977) or some kind of Zaza (MacPhee & Iturralde-Vinent, 1995; cebine related to both Saimiri and Cebus Iturralde-Vinent & MacPhee, in press). (MacPhee & Woods, 1982). Similarly, it Fossils of extinct Antillean monkeys were seemed unlikely that Jamaican Xenothrix, found (although not properly recognized as variously considerd to be a callicebin, such) as early as the 1920s in Hispaniola and cebine, or a taxon of unknown affinities Jamaica (MacPhee, 1996), but Williams & (Rosenberger, 1977; MacPhee & Fleagle, Koopman (1952) were responsible for shed- 1991), could be closely related to the mon- ding the first real light on the subject. Since keys from Cuba and Hispaniola. Adding to then, a number of new discoveries have been this already complex picture, Ford (1986a) made (Rímoli, 1977; Rosenberger, 1977; and Ford & Morgan (1988) argued that MacPhee & Woods, 1982; Ford & Morgan, several of the previously mentioned isolated 1988; Ford, 1990; Rivero & Arredondo, postcranials from Hispaniola and Jamaica 1991; MacPhee et al., 1995; MacPhee, are specifically marmoset like, which implied 1996), but the Antillean record is still that yet another major platyrrhine clade was clouded by a variety of uncertainties. From once represented in the Greater Antilles. a systematic standpoint, the most serious It is self-evident that, if the primate com- of these concerns how many species are plements of Cuba, Jamaica, and Hispaniola actually represented in existing fossil assem- had multiple origins, there had to have been blages, and how they are related to one multiple colonization events. However, another. There are three named taxa: thanks to a number of new discoveries, Paralouatta varonai Rivero & Arredondo, especially ones made in the present decade, 1991, from Cuba; Xenothrix mcgregori this picture of multiple separate coloniza- Williams & Koopman, 1952, from Jamaica; tions by unrelated propagules is no longer and Antillothrix bernensis MacPhee et al., so easily maintained. The multiple origins 1995 (formerly Saimiri bernensis Rímoli, viewpoint was explicitly challenged by 1977), from Hispaniola. There are, in MacPhee et al. (1995), who demonstrated addition, a number of primate-like post- on the basis of a preliminary cladistic study cranial remains from these islands that, for that Paralouatta and Antillothrix are related one reason or another, have not yet been as sister groups, and that Callicebus may be allocated; some may represent unknown the closest living relative of this dyad. The taxa (MacPhee & Fleagle, 1991; MacPhee, position of Xenothrix was not investigated. 1996). It is a safe bet that we are far However, it is of interest that Rosenberger from having a complete roster of Antillean (1977) independently posited that Xenothrix monkeys, even for the Quaternary (Ford, might be related to Callicebus on the basis 1990). of evidence then available. New fossils of The relationships of Antillean monkeys Xenothrix (description of which is in prep- are obscure. Indeed, until very recently, the aration) provide an opportunity to determine only conclusion one could draw from the whether all Antillean monkeys (or at least scanty literature on this subject was that those represented by adequate material) may these monkeys had little to do with one derive from a single colonizing ancestor. another. Thus, it might be concluded that One objective of this paper is to put on Cuban Paralouatta, supposedly a close rela- record a description and evaluation of all tive of Alouatta (Rivero & Arredondo, craniodental specimens of P. varonai so far 1991), could not also be a near affine of recovered. Another is to present a com- Hispaniolan Antillothrix bernensis because prehensive cladistic analysis of Antillean the latter was viewed as either a giant squir- monkeys. We address the latter goal within   PARALOUATTA VARONAI 35

Figure 1. Map of Cuba showing principal localities: CA, Cueva Alta and Cueva del Mono Fósil (Quaternary); DZ, Domo de Zara (early Miocene). the larger issue of platyrrhine systematics. ated on the same south-facing slope of the To avoid ambiguity, in this paper we explic- Sierra de Galeras, one of several blocks of itly define pitheciins as including Pithecia, uplifted Jurassic limestones that make up the Chiropotes, and Cacajao, atelines as Ateles, Cordillera de Guaniguanico (Figure 1). Brachyteles, Lagothrix, and Alouatta, and cal- The first primate fossils from Cueva del litrichines as Callimico, Callithrix, Cebuella, Mono Fósil were found in 1988 and Leontopithecus, and Saguinus. included the type skull (Figure 2) and a distal humerus. Remains belonging to other Abbreviations vertebrates indicated that the temporal con- text of the site was Quaternary. The first AMNH, American Museum of Natural detailed account of this material, by Rivero History, New York & Arredondo (1991), included a formal CMC, Claremont-McKenna College, description of P. varonai and emphasized California the similarity of the type skull to that of the GPBSEC, Grupo Espeleológico ‘‘Pedro living howler monkey, Alouatta. Borrás’’ of the Sociedad Since the late 1980s, accessible portions Espeleológica de Cuba of Cueva Alta and Cueva Mono Fósil have MNHNH V, Museo Nacional de Historia been carefully combed for new fossils by Natural, La Habana, expeditions of the GPBSEC, MNHNH, and Vertebrate Collection AMNH. A number of additional platyrrhine M-D, mesiodistal remains have been recovered as a result, B-L, buccolingual including a mandible, numerous isolated teeth, and several postcranial pieces. Craniodental morphology and Although exploration continues in the relationships of Paralouatta varonai numerous caves that riddle the rest of the Fossils allocated to P. varonai have been south face of Sierra de Galeras, to date no recovered at only two localities in Cuba, new primate-bearing fossil sites have been Cueva del Mono Fósil and nearby Cueva identified. This is, therefore, an opportune Alta (Rivero & Arredondo, 1991; Jáimez time to review, describe, and interpret what Salgado et al., 1992). These sites are situ- has been collected thus far. 36 .   . . . 

Figure 2.

Cueva del Mono Fósil and Cueva Alta are It is now evident that most of the remains located on a mogote (a steep-walled lime- found in each cave did not occur there stone hill with a flat or gently rounded top). originally, but were brought down from   PARALOUATTA VARONAI 37

Figure 2. Holotype specimen of Paralouatta varonai (MNHNH V194): (a) dorsal, (b) ventral (from a cast), (c) lateral, and (d) frontal views [(a) and (c) from Rivero & Arredondo, 1991]. passages higher on the mogote through a important fossil, both to round out know- system of fissures. Evidence for this suppo- ledge of the cranial anatomy of Paralouatta sition is the discovery of small bones (not and to provide a descriptive basis for some primate) in the upper part of the major of the character analyses conducted in the vertical fissure which transects Cueva del next section. It should be noted that the Mono Fósil, and the fact that all discoveries temporary museum number cited by Rivero of consequence made at Cueva Alta came & Arrendondo (1991) for the holotype from a single locus—a small, sediment- specimen (MNHNH 90-25) has now been choked chimney (Jáimez Salgado et al., replaced by a permanent accession string 1992). Efforts to date bone from these caves (MNHNH V194). chronometrically have not been successful Because the teeth of MNHNH V194 are because the bones retain too little collagen exceptionally worn, Rivero & Arrendondo to permit 14C (radiocarbon) dating. (1991) were unable to provide an adequate Whether this means that they are actually description of the dentition of P. varonai. beyond the range of radiocarbon dating has Later collecting has resulted in the discovery not been established, but it is certainly a of a lower jaw (MNHNH V195) and some possibility. If the Galeras faunule is com- 80 isolated teeth, many in excellent condi- paratively old, it would help to explain tion. These new fossils yield a wealth of (although not fully clarify) why monkey detail unknown previously. Our description bones have not been recovered elsewhere is conducted by element and we provide in Cuba. measurements of dental elements in Table 1 [for additional measurements, see original description of type skull by Rivero & Craniodental morphology of Arrendondo (1991)]. Paralouatta Rivero & Arrendondo (1991) interpreted Rivero & Arredondo’s (1991) description of overall skull shape in Paralouatta as essen- the holotype skull of P. varonai focused on tially indicative of alouattine affinity. For systematically important features and was example, both display a substantial angle not intended to be comprehensive. In this (airorhynchy) between the facial and neural section we provide a fuller account of this regions of the skull. However, Alouatta is a 38 .   . . . 

Table 1 Dental measurements (in mm) of new Table 1 Continued remains of P. varonai1 Specimen B-L M-D Specimen B-L M-D

M2 V195 5·6 7·4 1 I V150 3·9 5·9 M3 V124 5·1 7·8 V152 3·9 5·7 V134 4·5 6·6 V153 4·2 5·4 V135 5·0 7·0 V154 3·9 5·6 V579 5·5 7·8 V156 3·9 6·1 V195 5·0 7·2 V158 4·4 5·8 2 I V105 3·8 4·3 1Measurements for best preserved teeth only. V149 4·1 4·7 V151 3·4 4·2 V155 3·7 4·1 P2 V115 5·7 5·5 clear outlier among platyrrhines for this V160 5·4 5·3 feature, while the condition in Paralouatta is V164 4·9 5·3 best described as incipient. Thus, while it V165 6·0 5·4 P3 V163 7·2 5·1 can be said that the skulls are broadly similar V169 8·0 4·9 in this aspect, the two differ in many other V176 8·2 4·6 characters—as our data matrix brings out. V178 7·0 4·5 This point applies particularly strongly to V578 8·6 5·3 dP4 V166 6·7 5·6 the dentition, which was previously poorly P4 V106 9·1 5·4 known. With the new dental remains of V116 9·1 5·5 Paralouatta now available, it can be shown V170 10·0 5·2 V171 9·2 5·2 that the Cuban monkey differs markedly M1,2 V120 9·1 6·6 from both Alouatta and Stirtonia. V179 9·0 6·9 V180 9·7 7·0 V181 9·3 6·5 Skull V183 8·8 6·6 In most respects the neurocranium and basi- 3 M V122 7·0 6·3 cranium of MNHNH V194 are exception- V191 7·2 5·5 V192 7·1 5·3 ally well preserved; by contrast, the facial

I1 V126 3·6 2·8 region is damaged extensively (Figure 2). I2 V195 3·8 3·2 C1 V127 3·2 5·1 V195 4·1 4·7 Frontal. Except for the glabella/nasion

P2 V117 5·3 5·4 region and the orbital processes, the frontal V128 5·2 5·2 bone is quite complete. Shaped like an isos- V195 4·5 4·8 celes triangle, with its largest dimension P3 V118 5·9 4·9 V129 5·8 4·9 oriented anteroposteriorly, the frontal is V130 5·7 5·4 delimited rostrally by a low supraorbital V131 5·5 4·7 ridge surmounting the orbits [Figure 2(a)]. V132 5·5 4·9 V195 5·9 5·9 Dorsally, each supraorbital ridge is thrown

P4 V119 6·7 5·8 into relief by a slight depression in the V146 6·6 5·6 frontal, immediately behind the line of the V147 5·8 5·5 V195 5·9 6·4 ridge. In lateral view, the typical platyrrhine

M1 V195 5·6 7·4 contact pattern is displayed, with the zygo- M1,2 V123 5·3 7·1 matic and the parietal interposed between V138 5·9 7·0 V144 6·0 7·1 the frontal and alisphenoid. Within the V145 5·5 6·6 orbit, the frontal runs from dorsolateral to medioventral, intersecting the lateral   PARALOUATTA VARONAI 39 orbital fissure. Medial to its contact with the The zygomatic process of the maxilla is zygomatic, the frontal contacts successively situated low on the face. The preserved part the alisphenoid, the orbitosphenoid (above of its ventral edge lies just above the plane the orbital foramen) and, slightly more of the alveolar border (<1·0 times height of anteriorly on the medial wall of the orbit, the M2 crown), as in Callicebus and Xenothrix ethmoid (little of which is preserved). (Horovitz & MacPhee, in prep.). This is in Breakage in the glabellar region provides a contrast to Alouatta, the other atelines, and window into the midsagittal portion of the pitheciins, in which the process’s position on frontal sinuses, which appear to have been the face is notably higher (d1·5 times height extensive (see below). Dorsally, the coronal of M2 crown). Cebus, Aotus, and Saimiri suture is drawn into a long, posteriorly display intermediate positions. directed ‘‘V’’, corresponding to the two The maxillary tuberosity or postdental remaining sides of the isosceles triangle. part of the maxilla is drawn out into a Bregma is located far to the rear, at the lengthy, triangular process that articulates transverse level of the posterior carotid with the elongated pyramidal process of the formen (with skull in Frankfurt plane). palatine [Figure 2(b)]. It is not obvious why the rear of the maxilla should be conspicu- Ethmoid. Because of fusions and breakage, ously enlarged in Paralouatta, as the maxil- sutural boundaries in the ethmoid region lary sinus does not appear to extend into this cannot be followed for any distance. region. (However, matrix left within the However, it is clear that the ethmoid was nasal cavity to provide support for remaining pneumatized extensively and that, in con- lamellae may obscure an ostium into the sequence, the interorbital septum was tuberosity.) In atelines the tuberosity is quite comparatively substantial. The septum’s abbreviated and the posterior wall of the thickness suggests that it was not fenes- maxilla is much more vertical, especially trated. The anterior portion of the orbital in Alouatta. In contrast, Cacajao, Pithecia, wall and upper face, including the lacrimals Callicebus, and Xenothrix display a somewhat and the nasals, are not preserved. larger tuberosity. Comparisons show that this feature varies conspicuously across Maxilla. The maxillae are well preserved platyrrhine genera. except for their medial portions. The bony The internal architecture of the nasal orbits jut from the face to a marked degree cavity is poorly preserved. However, the (especially evident in norma verticalis), which massiveness of the facial region of the skull, suggests that the eyes were very large, evident in the photographs [Figure 2(a)], although not as large as in Aotus (for obser- is principally due to the expansion of vations on orbital size in Paralouatta, see pneumatic spaces related to the nasal appar- Rivero & Arredondo, 1991). The maxilla atus. Both the nasal chamber per se and extends backward to form much of the the paranasal sinuses are exceptionally orbital floor and the medial side of the large—indeed, proportionally much larger inferior orbital fissure (sphenomaxillary fis- than in any extant large-bodied platyrrhine. sure) up to the point of contact with the The inferior surface of the nasal aperture palatine. The facial portions of the maxillae is broad and gently sloping. Comparisons bear multiple infraorbital foramina. The reveal some similarities in this respect to area normally formed by the premaxilla has Callicebus and, to a lesser degree, Callimico. been lost through breakage, and there are One maxillary turbinal, probably the ven- no identifiable remnants of the maxillo– tralmost, is partly preserved on the left side. premaxillary suture. It begins high on the lateral wall of the nasal 40 .   . . .  cavity, at the level of the highest infraorbital anterior crus of the ectotympanic and the foramen, and is directed medioventrally. postglenoid process; this is unlike most A point of minor paleopathological inter- platyrrhines, in which these two structures est concerns the large abscess chamber per- form a continuous surface or are situated forating the wall of the maxilla above the very close together. The squamosal contacts mesial root of the left M2. the alisphenoid medially, with which it forms a suture that runs parallel to the Palatine. The palatine bones are partly pre- midsagittal plane to approximately the level served [Figure 2(b)]. Each maxillo–palatine of the lateral pterygoid process. At this point suture parallels the dental series up to the the suture turns laterally and runs along the level of M2, then runs medially to meet its ventral edge of the parietal. fellow (midsagittal contact lost through The ectotympanic is completely fused to breakage). The greater palatine foramen, at the tympanic bulla (here assumed to be the level of M3, lies on or just medial to the petrosal in origin); no sign of a suture track of this suture. remains. The aperture of the meatus is an The posteromedial palatine spine is oval ring, the long axis of which runs located posterior to the transverse level of anteroventrally/posterocaudally. The meatal M3. Its apparent degree of projection is margin is raised and highly rugose except on increased by deep scalloping of the choanal its dorso-caudal face, which is smooth and margin of the palatines. The arc of the flattened. scallops does not reach the transverse level The elongated tympanic bullae are mod- of M3. The palatine enters into the con- erately inflated. The anteromedial portions struction of the orbit along the medial wall of both are broken, thus exposing parts of of the inferior orbital fissure. Posteriorly it the middle ear cavity. As seen from the contacts the alisphenoid and anteriorly the outside, the alisphenoid forms the rostral maxilla. and lateral edges of the foramen ovale, while the bulla makes up this aperture’s medial Vomer. The bladelike vomer is visible on and caudal edges. The canal for the auditory the base of the skull beneath the pre- tube opens immediately beneath the caudal sphenoid, part of which it covers. Irregularly border of the foramen ovale. The posterior sutured into the palatines, the vomer carotid foramen is located on the medial extends to the rear edge of the palate and curvature of the bulla in the typical platyr- thereby helps to define the choanal aper- rhine position, just in advance of a line tures. The choanae are relatively enormous connecting the jugular foramina. It is (right aperture, 14·1#9·1 mm), their recessed into a small basin whose border is entrances being approximately four times as defined by a sharp line or crest on the bulla’s large in area as those of AMNHM 230805, a ventral surface. The stylomastoid foramen, specimen of Alouatta seniculus of similar which is only slightly smaller than the skull length (right aperture, 8·6#3·7 mm). carotid foramen, is located posteriorly on or near the apparent margin of the ectotym- Temporal. The temporal squamae are very panic. The hypoglossal canal is buried in a low, as in platyrrhines generally, and the deep pit above the anterior margin of the postglenoid process is notably elongated. occipital condyles. The large postglenoid foramen [Figure 2(b)] The petrosal promontorium, seen by is situated on the medial side of the latter looking through the external auditory process, rather than behind it or within it. meatus, displays two bulges or prominences. There is a pronounced gap between the The first is the margin of the aperture of   PARALOUATTA VARONAI 41 the fenestra cochleae; the second bulge, and most of the pterygoid process (the situated more anterodorsally, appears to be anterior third of each process is contributed either a bony eminence or (more likely) the by the palatine). Neither the lateral nor the next turn of the cochlea itself, seen in relief medial pterygoid plates are completely pre- as it swells the lateral aspect of the pro- served, but it is clear that the lateral plates montorium. Similar bulges are present in were much larger than the medial, as in Callicebus, Pithecia, callitrichines, Cebus, platyrrhines generally. It is not possible to Saimiri, and Aotus. In all other genera exam- determine whether the medial plates pro- ined by us the promontorium is either jected substantially. However it is possible without relief in this region or only the first to see that the pterygoid fossa between the bulge is present. medial and lateral pterygoid processes was Paralouatta has a typically anthropoid very shallow and did not excavate the base of anterior accessory (paratympanic) cavity the skull. In this regard Paralouatta is similar whose posterior wall is partly defined by the to Callicebus, pitheciins and atelines, but track of the bony carotid canal (see contrasts markedly with Cebus, Saimiri, and MacPhee & Cartmill, 1986). The dorsally Aotus, in which the fossa indents the skull placed epitympanic recess is large and filled base. with short, irregular septa. Similar relief is also present in the posterior part of the Parietal.InParalouatta the parietals meet hypotympanic region. the zygomatics along an irregular, dorso- The posterior end of the temporal is com- ventrally oriented line. A projecting portion pleted by the mastoid region. In Paralouatta of the parietal’s ventralmost edge intervenes the mastoids are considerably swollen, and between the zygomatic and the alisphenoid, they project laterally and especially posteri- and thereby manages to border on the lateral orly much more than in Alouatta. Small orbital fissure. Parietal contribution to the fenestrations created by breakage expose delimitation of the lateral orbital fissure is a bony cellules within the mastoids, indicating character present (although frequently poly- significant pneumatization. morphic) in most platyrrhines. However, in Alouatta, Brachyteles, and Cebus, the lateral Sphenoid complex. The ventral aspect of the orbital fissure is normally defined by the sphenoid complex is unremarkable. The zygomatic exclusively, or by this bone and presphenoid–basisphenoid synchondrosis is (for a small distance) the alisphenoid. The completely fused, although a faint line parieto–alisphenoid suture, about one- marks its original position. Cellules of large quarter of the length of the parieto– size (possibly evidence of pneumatization) squamosal contact, terminates close to the have been exposed by abrasion at the site of posterior (squamosal) root of the zygomatic the basisphenoid–basioccipital synchrondo- arch. sis. The body of the basisphenoid bears low As seen from above, the temporal lines ridges paralleling the tympanic bullae, prob- are remarkably sinusoidal, flaring once ably for the origin of prevertebral muscles anteriorly (at the level of the supraorbital (see description of occipital). ridge), again centrally (at bregma), and once The dorsal wing of the alisphenoid con- again posteriorly (just in advance of the tacts the zygomatic, parietal, and squam- lambdoidal suture). osal. The dorsal wing forms the lateral edge of the inferior orbital fissure. The ventral Zygomatic. This bone is not preserved com- (pterygoid) wing forms a small fraction of pletely on either side: the orbital margins the medial edge of the inferior orbital fissure are missing bilaterally, and the zygomatic 42 .   . . . 

Figure 3. Skull of adult Alouatta (lateral view) showing several characters (see Appendix 1). roots of the zygomatic arches are only par- Ridging is extensive, more so than in any tially preserved. The fronto– and parieto– living platyrrhine, including Alouatta. The zygomatic sutures and the formation of the width of the skull in norma occipitalis lateral orbital fissure have already been is considerably increased by the lateral described. The maxillo–zygomatic suture projection of the mastoids. slants posteroventrally, with the result that The skull qualifies as airoryhnchic the maxilla and the zygomatic form respect- because the basioccipital is angled (ventrally ively the ventral and dorsal halves of the root flexed) with respect to the dental series, of the zygomatic arch. somewhat more than most platyrrhines Multiple zygomaticofacial foramina are except Alouatta, in which the angle is located on the face, near the ventrolateral markedly larger. margin of each orbit (although this is not The foramen magnum is positioned so obvious in the figures because of breakage). that if faces downward as much as back- At least three foramina are preserved on the ward when the skull is placed in the right side, the central one being the largest Frankfurt plane, as in platyrrhines other [see Rivero & Arredondo (1991)]. In most than Alouatta. By contrast, in Alouatta the platyrrhines, there is a single zygomatico- foramen magnum is more posteriorly posi- facial foramen on each zygomatic (Figure 3); tioned and is oriented at a sharp angle to the however Alouatta, some pitheciins, and Ao- basioccipital. tus occasionally display multiple foramina. The supraoccipitals form an angle of approximately 45) with the Frankfurt plane. Occipital. The basioccipital widens caudally. The surface of the occipital planum is excep- Its ventral surface is marked by the con- tionally rugose, suggesting that there was a tinuation of the low ridges for muscle considerable investment in postural muscle attachment seen on the basisphenoid. mass.   PARALOUATTA VARONAI 43

Figure 4. Paralouatta varonai mandible (MNHNH V195), in (a) occlusal and (b) lateral views.

Mandible adult animal, although it is far too small to The mandible (MNHNH V195) of Para- articulate with the type skull. The specimen louatta was found in 1991 in Cueva del preserves the left ascending and horizontal

Mono Fósil, but at a considerable remove rami, with P2–M3 in situ and alveoli for other from the position of discovery of the type anterior teeth, the symphyseal portion, bear- skull. The mandible (Figure 4) is that of an ing I2 and C1 and a small part of the inferior 44 .   . . . 

Figure 5. P. varonai, stereoscopic views of isolated front teeth: (a) I1 (MNHNH V154), (b) I2 (MNHNH

V155), (c) I1 (MNHNH V126), and (d) C1 (MNHNH V127). border of the right horizontal ramus. On way that asperities (cusps and cristae) were the preserved left ascending ramus, the rapidly removed, effectively converting the coronoid process is broken at its root and occlusal surfaces of the cheek teeth into the medial aspect of the mandibular condyle large, flat milling surfaces (see Figure 4). is abraded. Otherwise, the specimen is in Paralouatta is unusual in displaying, in com- extremely good condition. However, as the bination, the following features usually teeth of both the type skull and the mandible associated with hominine dentitions: are exceedingly worn, descriptions of indi- extreme bunodonty, premolar hypertrophy, vidual loci are based largely on isolated widened maxillary incisor crowns, and, most specimens found during screening opera- interestingly, marked canine reduction tions at Cueva Alta. (crowns incisiform, projecting little or not at all beyond occlusal plane of cheek teeth, and Dentition canine root small). Although Rivero & Arrendondo (1991) con- The dental formula of Paralouatta is 2133/ cluded that P. varonai was a definite alouat- 2133, as in all known platyrrhines except tin, they noted that the teeth of the Cuban callitrichines (other than Callimico) and species appeared to differ in several ways Xenothrix. from those of howler monkeys. Thus they noted that molar pericones were present and Maxillary teeth large, and that, curiously, the maxillary Central incisors. MNHNH V150, 152–154, canine (as indicated by a preserved root) 156–158. Large size (compared to lateral would have been tiny. Additional marked incisors), substantial breadth, and low differences between Paralouatta and crown height are noteworthy features of the Alouatta were briefly documented by maxillary central incisors [Figure 5(a)]. In MacPhee et al. (1995). Alouatta, Brachyteles, Ateles, and, to a lesser To a greater degree than in any other degree, Lagothrix, incisors are higher- known platyrrhine, the entire dental series of crowned but narrower. As is frequently the Paralouatta evidently wore down in such a case in platyrrhines, the mesial margin of   PARALOUATTA VARONAI 45 the crown is higher than the distal, and the aligned. In Paralouatta, the I2 distal facet is biting edge is distinctly curved. On little- vertical, circular rather than oval in form, and worn specimens, a bean-shaped fovea is separated from facetting on the bite edge by visible on the lingual surface. a significant gap. Therefore it cannot be a C1 Incisor tooth use in Paralouatta seems to facet, but must instead be for the C1. Taken have involved considerable amounts of bite- together, these features strongly imply that pulling, because worn teeth display numer- in the Cuban monkey, maxillary incisors ous small grooves arrayed normal to the and canines were in intimate contact and edge of the occlusal surface. Wear features prone to develop interproximal facetting. suggest that maxillary incisors were drawn orthally across lowers during biting, or that Canine. MNHNH V194. The type skull food items were shredded by pulling them retains only the deepest part of the root of manually between clenched incisor teeth. As the right C1, which at that level, is smaller a result, wear on the lingual or foveal surface than the alveolus for P2. The tiny size of of I1 is often substantial (for example, all this root was confirmed by X-ray of the enamel lost from foveal surface MNHNH holotype. V158). Step-fracturing and pitting of the enamel is visible in, and restricted to, the Premolars. MNHNH V115, 160–162, 164, areas immediately adjacent to the bite sur- 165 (P2); 163, 169, 174–178, 578 (P3); face. This localization suggests that these MNHNH V106, 116, 167, 168, 170–173 defects were acquired during life, probably (P4); MNHNH V166 (dP4). as the result of small fragments of enamel P2 (MNHNH V115, 165) is double- spalling off during powerful biting. rooted in the type specimen (MNHNH V194), as reported by Rivero & Arredondo Lateral incisors. MNHNH V105, 149, 151, (1991). However, the three isolated speci- 155. Maxillary lateral incisors have shorter mens in which the roots are complete or roots, considerably smaller crowns, and nearly complete (V160, 162, 164, 165) are narrower cervical regions than do central in fact single-rooted, as is normally but not incisors [Figure 5(b)]. universally the case in other platyrrhines Characteristic of I2sofParalouatta is the (Hershkovitz, 1977). For example, in most presence of interproximal facets on both specimens of Callicebus the P2 root is single the mesial and distal margins of the crown. and undivided, but a split root is occasion- The existence of the distal interproximal ally seen. P2 bears a single cusp on its buccal facet implies that I2 and C1 were in perma- side, and its trigon widens distally [Figure nent contact, that is, they were not separated 6(a)]. P3 and P4 [Figure 6(b), (j)] are mor- by a diastema. (This point was not obvious phologically similar, although P3 is some- when only the type skull was available, since what smaller. Both of the distal premolar it lacks incisors and preserves only the stump loci bear two cusps, with the protocone of one canine root.) In this feature Paral- being situated in the mesiobuccal quarter of ouatta is unlike any living large-bodied New the trigon. The lingual cingulum is promi- World monkey, in all of which the canines nent and mesially projecting relative to the are enlarged and diastemata are present. In trigon (cingular buccolingual width at least these latter platyrrhines, the distal aspects of half that of trigon). In the type specimen, I2 crowns sometimes display an oval facet for both P3 and P4 exhibit two buccal roots in the mesial aspect of the tip of C1. However, addition to a lingual root. Among isolated 2 because the C1 meets the I at an oblique specimens, MNHNH V163 and V178 (both angle, the resulting facet is never vertically P3s) are single-rooted, whereas MNHNH 46 .   . . . 

Figure 6.

V106 (a P4) has two roots (one buccal, one Saimiri, Aotus, Cacajao, Alouatta, Callicebus, lingual). Platyrrhines display some varia- and Xenothrix,P3 normally possesses two bility in root number in P3 and P4.InCebus, roots (three occasionally in Alouatta). On   PARALOUATTA VARONAI 47

Figure 6. P. varonai, stereoscopic views of permanent premolars and molars with explanatory keys: (a) P2 4 (MNHNH V115), (b) P (V116), (c) P2 (MNHNH V117), (d) P3 (MNHNH V118), (e) P4 (MNHNH 1,2 3 V119), (f) M (MNHNH V120), (g) M (MNHNH V122), (h) M1,2 (MNHNH V123), (i) M3 2 1,2 (MNHNH V124), (j) P (MNHNH V115), (k) P3 (MNHNH V118), (l) P4 (MNHNH V119), (m) M (MNHNH V120), and (n) M1,2 (MNHNH V123).

Figure 7. Stereoscopic view of P. varonai deciduous P4 (MNHNH V166). the other hand, in Ateles the roots of P3 tend cingulum is less developed (crown more to be fused and only the tips of the roots triangular), and overall the tooth is much are separate. P4 normally has two roots in smaller than true molars. Cebus, Saimiri, Cacajao, Alouatta, Callicebus, and Ateles (as in P3, tips barely separate); in Molars. MNHNH V104, 120, 121, 179– Aotus the normal condition is three roots. 184, 186–190 (M1,2); MNHNH V122, 185, Only one deciduous premolar, a dP4 191–193 (M3). We have found no reliable (MNHNH V166), has been recognized with way to separate first and second upper mo- certainty (Figure 7). Very low crowned with lars, and therefore will discuss them together. three broadly splayed roots (two buccal, one Upper molars of Paralouatta display sev- lingual), it resembles upper molars in gen- eral diagnostically interesting features. One eral outline, but the proportions of the major such feature is the size and differentiation cusps are somewhat different. The lingual of cingular structures [Figure 6(f), (m)]. 48 .   . . . 

Except on the mesial side of molar trigons, ridges. As in other platyrrhines, M1,2 display a continuous ledge borders upper molar three roots, two buccal and one lingual. crowns; the ledge is best developed buccally M3, smaller than M1,2, displays four and lingually. In the least worn examples cusps, of which the paracone is best devel- of M1,2 (e.g., MNHNH V120, 180), the oped. The hypocone appears on the lingual buccal cingulum is continuous from the cingulum bordering the distal side of the preparacrista to the postmetacrista, although tooth and is barely cuspiform. The floor of it is scored by a tiny groove in the ectoflexus. the trigon is irregularly rugose, as in M1,2. The buccal cingulum is adorned with a This tooth may exhibit two discrete roots relatively discrete distostyle distobucally, a (as in MNHNH V185, 193) or only a very faint parastyle, and two low bumps single, fused one (MNHNH V122, 191, (=mesostyles) on either side of the groove in 192) [Figure 6(g)]. the ectoflexus. The lingual cingulum is massively devel- Mandibular teeth oped and forms a continuous band from the Central and lateral incisors. MNHNH V126 mesiolingual aspect of the protocone to the (I1); MNHNH V195 (I2). The narrow- distostyle of the buccal cingulum. There is crowned I1 and I2 are tiny compared to the a well-developed hypocone and a smaller upper incisors [Figure 4, 5(c)]. There is only but nevertheless well-developed entostyle one identifiable I1 in the existing sample (=pericone), doubled in some specimens (MNHNH V126); its flat occlusal edge (e.g., MNHNH V120, 180). The hypocone presents a large dentine lake exposed is connected to the metaloph via a well- through wear. Central incisors are missing in developed prehypocrista (hypocone crest). the jaw (MNHNH V195), but I2 is pre- The hypocone is large in howler monkeys, served on the right ramus. Compared to I1, and there is a deep incisure between the the I1 has a narrower crown and a smaller hypocone and protocone that is lacking root (as suggested by empty alveoli in the in Paralouatta. jaw). In both teeth, many sub-parallel One of the outstanding features of molar striations can be seen running normal to the construction in Paralouatta is the position of bite surface (cf. maxillary incisors). the protocone, which appears to be situated virtually in the centre of the occlusal surface. Canine. MNHNH V195, 127 [Figure 4,

This appearance is in large measure due to 5(d)]. The C1 is present in situ in the Mono the great width of the lingual cingulum, Fósil lower jaw (MNHNH V127); this which isolates the protocone from the crown specimen demonstrates beyond any doubt margin. In Alouatta, the protocone is not that extreme canine reduction was the nor- lingually bordered by a cingulum at all, while mal condition in Paralouatta. It is also clear in Stirtonia (see Material and methods) the from this specimen that C1 was no more latter is very narrow and discontinuous. projecting than P2, from which it may be The major trigon cusps are markedly distinguished only by its greater degree of bunodont and enclose a large but shallow buccolingual compression and single, blade- trigon basin. In unworn teeth (MNHNH like, mesiodistally oriented principal cusp. A V120, 180), the deepest part of the basin is subtle lingual cingulum is present in V127. corrugated by a number of tiny cuspules and grooves. However, the type of crenulation Premolars. MNHNH V195 (mandible with is not readily comparable to that seen on P2–4); MNHNH V117, 128 (P2); MNHNH pitheciin molars; in the latter they are much V118, 129–132 (P3); MNHNH V119, finer, and have the shape of elongated 146–148 (P4). Dimensions of mandibular   PARALOUATTA VARONAI 49

1 2 premolars increase distally, making P2 the As in the case of M and M we cannot smallest member of the premolar row identify any consistent morphological or size

(Figure 4). This premolar locus–size rela- differences between M1 and M2. All man- tionship is seen in other platyrrhines having dibular molars display two roots, one mesial diminutive canines (e.g., Callicebus). In con- and the other distal. trast, in all atelines except Brachyteles,P2 is Compared to Alouatta, lower molar cusps the largest premolar. In Paralouatta,P2 and are stouter, crests are much less marked P3 are single-rooted; none of the isolated P4s even in unworn teeth, and in worn teeth has a completely preserved root. there is almost no difference in cusp height

The P2 has one buccal cusp from which profile between talonid and trigonid cusps radiate three crests: one runs mesially, [Figure 6(h), (n)]. another curves slightly as it passes disto- Alouatta and Stirtonia distinctively differ lingually, while the third and shortest crest from Lagothrix, Ateles, and Brachyteles in that originates slightly distal to the cusp, runs the talonid is transversely wider than the lingually, and ends in a small swelling trigonid. Inspection shows that breadth dif- [Figure 6(c)]. About twice as much of the ference is mainly due to the fact that, in tooth is located mesial to the last crest as lies Alouatta, the apex of the entoconid bulges distal to it. The cingulum is not continuous lingually to a noticeable degree, giving the around the lingual and distal surfaces of the talonid an almost pear-shaped outline. Par-

P2 crown (as it is in atelines) because it is alouatta critically differs in that protoconid interrupted by the lingual crest. and hypoconid apices are not displaced.

The P3 has two distinct cusps of subequal Instead, the entire buccal aspect of the lower height, a metaconid and a larger (in diam- molars is swollen, making the protoconid eter) protoconid. A crest running mesially and hypoconid seem proportionately larger from the metaconid describes a sharp angle than the metaconid and entoconid and giv- and then curves distolingually onto the ing the tooth a lop-sided look, similar to protoconid. The talonid bears no cusps, Xenothrix molars (see figure in Williams & is longer mesiodistally on the lingual side, Koopman, 1952). and is completely enclosed by a crest. The The distal wall of lower molar trigonids is a trigonid is notably larger than the talonid continuous, raised crest joining metaconid [Figure 6(d), (k)]. and protoconid. It is on a sharp oblique, as

The P4, the largest of the three premolars in Alouatta and Stirtonia. In all molars the [Figure 6(e), (l)], has a square outline and a cristid obliqua (=premetacristid) is strongly well developed protoconid and metaconid. built and intersects the distal wall of the tri- Both cusps are subequal in height, but the gonid at a position intermediate between the protoconid has a much larger volume than protoconid and metaconid. As a result the the metaconid. As in the molars, the buccal ectoflexid is very deep in all lower molars, its (protoconid) side of the tooth is swollen and apex being situated within the middle one- projecting. Trigonid and talonid basins are third of the tooth’s breadth. In little-worn originally distinct, but quickly become worn specimens (e.g., V579) the ectoflexid is down to shallow grooves. It displays hypo- frequently adorned with one or two small conid and entoconid. cuspules (ectostylids), usually positioned on the distobuccal wall of the protoconid.

Molars. MNHNH V195 (mandible with On M3 there is a hypoconulid, of moderate M1–3); MNHNH V123, 136, 138–142, 144, size, situated directly distal to the entoconid 145, 579 (M1,2); MNHN V124, 134, 135 and separated from it by a sulcus leading (M3). into a small distolingual basin [Figure 6(i)]. 50 .   . . . 

The hypoconulid is occasionally present on phology, although a few pertain to the post- the M3 of many platyrrhine species. cranium and soft tissue anatomy. Characters taken from the literature were verified be- fore utilization. With respect to continuous Cladistic analysis characters, we used only those that showed Materials and methods states separated by gaps in distribution A total of 80 characters are listed and among groups of terminal taxa. These char- defined in Appendix 1. These are scored in acters were considered additive. Missing Appendix 2 for the taxa making up the characters were scored as question marks comparative set, which include Cebupithecia, (see Appendix 2). Phylogenetic analysis was Stirtonia, Paralouatta, Antillothrix, Xenothrix, performed using the program PAUP (Phylo- the 16 extant genera of New World mon- genetic Analysis Using Parsimony), version keys, and several outgroups. Cebupithecia 3.1.1 (Swofford, 1993) applying a heuristic sarmientoi (Stirton & Savage, 1951), Stirto- search with stepwise random addition nia tatacoensis (Stirton, 1951; Herschkovitz, sequence of taxa and TBR, done over 100 1970) and S. victoriae (Kay et al., 1987) replications. All characters were weighted are fossil species from the Middle Miocene equally and variable ones were considered of Colombia. C. sarmientoi has been pre- polymorphic. Trees were rooted by desig- viously placed among pitheciines (Stirton & nating Tarsius as the outgroup. Savage, 1951; Stirton, 1951; Orlosky, 1973; We obtained Bremer support values for Rosenberger, 1979; Kay 1990). Stirtonia is each branch (Bremer, 1988; Källersjö et al., currently considered the sister group of 1992), inspecting the strict consensus of Alouatta (Rosenberger, 1979; Setoguchi trees up to four steps longer than the most et al., 1981; Kay et al., 1987, 1989). As parsimonious ones. For example, branches mentioned in the Introduction, P. varonai, that collapse in a strict consensus of trees A. bernensis, and X. mcgregori are Quaternary a step longer than the minimum, have a species from Cuba, Hispaniola and Jamaica Bremer support of ‘‘1’’. respectively. Character analysis of X. mcgregori is based on the type mandible Results plus several new fossils recently collected The heuristic search yielded three most par- by RDEM, Donald A. McFarlane (CMC), simonious trees (CI=0.51, RI=0.66, tree and co-workers in southern Jamaica (two length=269 steps) the strict consensus of partial mandibles, a maxillary fragment, and which is shown in Figure 8. In this con- a partial skull preserving the palate and most sensus, Aegyptopithecus is positioned as the of the nasal fossae and maxillary sinuses; sister group of all other anthropoids, and Horovitz et al., 1997). Outgroup taxa con- platyrrhines appear as monophyletic. The sisted of Tarsius, a wide array of living catar- three trees differ in how relationships among rhines (including both cercopithecoids and atelines are portrayed. Brachyteles appears hominoids), and the Oligocene Fayum either as the sister group of the Alouatta– anthropoid Aegytopithecus. Unlike most Stirtonia dyad (with Lagothrix–Ateles as the fossil taxa that might have been used as next branch), or as basal within atelines. In outgroups in this investigation, Aegypto- the latter case, Ateles appears either as the pithecus is represented by relatively complete next offshoot within atelines, or as the sister skeletal remains and could therefore be group of Lagothrix. Characters supporting scored for many characters. all clades in the trees represented by the Most of the characters utilized in this consensus depicted in Figure 8 are listed study are based on craniodental mor- in Table 2.   PARALOUATTA VARONAI 51

Figure 8. Strict consensus of the three most parsimonious trees obtained with 80 morphological characters (see Appendices 1 and 2); ‘‘†’’ indicates fossil taxon. Antillean clade enclosed in rectangle. Unambiguous character support for each node (labelled with circles) shown in Table 2. Numbers above branches indicate level of Bremer support.

The Bremer support values (Figure 8) longer than the most parsimonious tree. A indicate that most branches of the depic- slightly different formulation, ten steps ted topology are unstable, including the longer than the most parsimonious tree, has relationships of Paralouatta to the other Paralouatta as sister group of the Alouatta– Antillean monkeys and Callicebus. The Stirtonia pair. Though our new hypothesis topology suggested by the original descrip- may be falsified with new data, it contra- tion of Paralouatta (Rivero & Arredondo, dicts rather strongly the contention that 1991), in which Paralouatta is regarded as Paralouatta may be closely related to the sister group of Alouatta, is 11 steps Alouatta. 52 .   . . . 

Table 2 Unambiguous characters supporting tree nodes depicted in Figure 81

Branch Character Change

node_47

Table 2 Continued

Branch Character Change node_34

1Based on morphological data set shown in Appendices 1 and 2.

Three unique, unambiguously positioned guished from other primates by the fact that characters support Platyrrhini: presence of tentorial ossification extends behind the an ossified tentorium cerebelli (Character subarcuate fossa as well as above and in 13), presence of a canal which traverses the front of it (Figure 9). Neverthless, New posterior wall of the subarcuate fossa to World monkeys differ in the degree to which connect with the channel for the sigmoid the tentorium is ossified. In atelines ossifica- sinus (Character 17), and zygomatico– tion is always extensive, whereas in some parietal contact (Character 24). These callitrichines and some specimens of Saimiri characters constitute a morphology-based only the portion of the tentorium attaching argument for platyrrhine monophyly and to the petrosal apex displays any sign of deserve additional comment. bone formation. Indeed, in some specimens of Saimiri there is no indication of tentorial Character 13. Hershkovitz (1977) noted ossification at all, and for this reason we that the tentorium is extensively ossified in score this character as polymorphic in atelines and, to a lesser degree, in other squirrel monkeys (see Appendix 1). platyrrhines; tentorial ossification also Our survey of other primates showed occurs in lemuroids. However, we find that, that lemuroids characteristically present a as a group, platyrrhines may be distin- restricted, T-shaped tentorial ossification in 54 .   . . . 

Figure 9. Medial view of sagittal section of Callicebus moloch skull (after Hershkovitz, 1977), illustrating partial ossification of tentorium cerebelli (Character 13) and presence of canal connecting subarcuate fossa and sigmoid sinus (Character 17). front of the petrosal apex, quite unlike the Two anthropoid petrosals from the platyrrhine condition. Like catarrhines and Fayum Oligocene were described by Tarsius, lemurs lack a posterior extension of Cartmill et al. (1981). One of them has been tentorial ossification behind the subarcuate assigned tentatively to the fossil anthropoid fossa. The presence and nature of tentorial Apidium; the other may belong to any of ossification could not be determined in our the larger anthropoids (i.e., Aegyptopithecus, sample of fossil Old World anthropoids. Parapithecus, Propliopithecus). The canal in the subarcuate fossa is absent in sampled Character 17. The innominate canal con- extant catarrhines, Tarsius, and the isolated necting the subarcuate fossa and sigmoid petrosals of ?Apidium and ?Aegyptopithecus sinus was first identified by Cartmill et al. discussed by Cartmill et al. (1981). (1981). This canal is located behind and below the aperture of the vestibular Character 24. Another unambiguous aqueduct, and is presumably for venous character on the most parsimonious trees drainage. Although Cartmill et al. (1981) that supports platyrrhine monophyly is reported this canal to be absent in atelines, zygomatic–parietal contact on the sidewall we determined on the basis of our more of the skull as seen in side view (Figure 3). extensive sampling that this feature is regu- In Tarsius, living catarrhines, Aegypto- larly present in all New World monkeys, pithecus, and possibly Apidium (Simons, including atelines (feature not yet confirmed 1959; Fleagle & Rosenberger, 1983), these for poorly-sampled Brachyteles). The vessel two bones do not make contact on the traversing the canal does not appear to be sidewall of the skull because the alisphenoid the homolog of the sigmoido–antral vein and frontal are interposed between them. and canal described by Saban (1963) in However, in Apidium the surface of the lemurines in approximately the same loca- temporal process of the frontal has a rugose tion. In lemurines this vein issues from the surface suggestive of broken bone (Simons, mastoid antrum and drains to the sigmoid 1959; Fleagle & Kay, 1987) which could sinus, but in platyrrhines there is no evi- have contained the parietal zygomatic dence of a mastoid connection. suture, as in platyrrhines (Fleagle & Kay,   PARALOUATTA VARONAI 55

1987). This possibility needs to be evaluated on more complete cranial remains.

Relationships of Greater Antillean monkeys In the most parsimonious trees, the clade consisting of Callicebus and the Greater Antillean monkeys is supported by four unambiguously placed characters. They share the derived condition of Character 15, possession of two prominences on the lateral wall of the promontorium. The primitive condition for this character is the presence of a flat surface or a single prominence. The other three unambiguously placed charac- ters are: ventral border of zygomatic arch extending below plane of alveolar border (Character 23, derived from a condition in which zygomatic arch is situated higher on face); mandibular canine root highly compressed (Character 34, derived from a more rounded condition); and alveolus of maxillary canine smaller than that of P4 (Character 62, derived from reverse condition). The Antillean clade itself is supported by three unambiguous characters: nasal fossa wider than palate at level of M1 (Character 25, derived from narrower condition depicted in Figure 10); alveolus of mandibu- lar canine buccolingually smaller than that of P4 (Character 39, derived from reverse Figure 10. Width of nasal fossa (Character 25) illus- condition), and mandibular M1 protoconid trated in (a) Cebus nigrivittatus at the level of M1,in having bulging buccal surface (Character frontal section (from Hershkovitz, 1977); (b) P. varonai 53, derived from condition in which feature at similar level, in imaginary frontal section. In Cebus, is absent). nasal fossa is narrower than palate between lingual edges of left and right M1;inParalouatta,itiswider The dyad consisting of Paralouatta and (condition shared only with X. mcgregori). Features Antillothrix is supported by six unambiguous nf and ms are nasal fossa and maxillary sinus, respectively. characters: M1 oblique cristid intersects pro- tolophid lingual to protoconid (Character 52, derived from a position directly distal to derived from absence of pericone); M1 post- protoconid); P4 lingual cingulum projects metacrista slopes distobuccally (Character mesially (Character 68, derived from condi- 73, derived from slope directed distally tion in which cingulum projects directly or distolingually); and M1 hypocone is lingually); P4 subequal to M1 in bucco- located lingually with respect to proto- lingual dimension (Character 70, derived cone (Character 74, derived from condi- from condition in which P4 is smaller); M1 tion in which this pair of cusps is aligned possesses distinct pericone (Character 75, mesiodistally). 56 .   . . . 

Discussion nuclear data set by Horovitz et al. (1998) resulted in Aotus being positioned as the At present there is no broadly accepted basal member of a clade otherwise formed cladogram of platyrrhine intergeneric rela- by the set ((callitrichines) (Cebus, Saimiri)), tionships. Although there are many reasons although the support for this topology is for this, it is nevertheless the case that in- weak. The other major clade consisted of formative morphological characters appear atelines and pitheciines, with the latter to be relatively hard to find in this group: including Callicebus and the pitheciins. individual taxa are highly autapomorphic, Most branches of the molecular and and published trees contain large amounts morphological trees mentioned above are of homoplasy. short, except for those supporting the three Rosenberger (1981, 1984), Ford (1986b), major groups, atelines (Ateles, Lagothrix, Kay (1990), and MacPhee et al. (1995) have Alouatta, and Brachyteles), callitrichines presented trees that have some features in (Callithrix, Cebuella, Saguinus, and Leonto- common. Thus all of these workers recog- pithecus) and pitheciins (Pithecia, Cacajao, nize 16 living genera, and apportion the and Chiropotes). This suggests that a rapid majority of them in the same way among basal radiation of this group of primates three major phylogenetic groupings: Atelinae, occurred, and provides a plausible reason including Ateles, Brachyteles, Lagothrix, and why the topology of their phylogenetic tree is Alouatta; Pitheciini, including Pithecia, highly unstable. The trees we obtained for Cacajao, and Chiropotes; and Callitrichinae, New World monkeys in the present study including Callimico, Callithrix, Cebuella, suggest a similar conclusion. Saguinus, and Leontopithecus. It has proven In the present study, retrieved relation- much more problematic to come to a con- ships are almost identical to those obtained sensus regarding how the specified groups by Horovitz & Meyer (1997) and Horovitz are related to each other and to the several et al. (1998) for morphological characters, genera—Cebus, Saimiri, Callicebus, and despite the addition of three fossil taxa from Aotus—that are especially difficult to fit into the Caribbean. Compared to results of other existing schemes. workers, our findings correspond closely to Molecular data amenable to cladistic Rosenberger’s original analyses, as well as analysis have only been collected since to those obtained with nuclear sequences 1993. In the first data set to be published (Schneider et al., 1993, 1996; Harada et al., (Schneider et al., 1993), sequences of 1995) and combined data sets [nuclear and nucleus-encoded å-globin genes produced mitochondrial sequences plus morphologi- highly consistent trees and a highly resolved cal data (Horovitz & Meyer, 1997; Horovitz consensus diagram. However, in that paper et al., 1998)]. the relationships among Saimiri, Cebus, Our results also confirm our preliminary Aotus, and callitrichines remained unsettled, ideas (MacPhee et al., 1995) regarding the although there were strong indications cladogeny of Greater Antillean monkeys. that these taxa form a clade. Addition Thus, in this study, as in the preceding one, of sequences of another gene, the inter- Callicebus emerges as the sister group of stitial retinol-binding protein gene (IRBP) Greater Antillean monkeys. But whereas intron 1 orthologues (Harada et al., 1995; previously we considered only the relation- Schneider et al., 1996) did not resolve this ships of Paralouatta and Antillothrix (which uncertainty. Addition of mitochondrial data remain sister taxa, as before), we now (part of the 16 S ribosomal gene and the find that parsimony favours the inclusion complete 12 S ribosomal gene) to the of Xenothrix within this group. We   PARALOUATTA VARONAI 57 conclude from this that the monophyly of Meeker (photography and Figures 1, 6, and the Antillean radiation can be considered to 10) and Clare Flemming (editorial assist- be further confirmed, with Xenothrix as the ance and Figure 3). Finally, we thank John basal taxon and Callicebus as the mainland Fleagle and three anonymous reviewers for sister group. helpful and perceptive comments on this The chief biogeographical implication of paper. the discovery that Greater Antillean mon- References keys form a monophyletic group is obvious: it is now parsimonious to assume that only Ameghino, F. (1906). Les formacions sédimentaires du Crétacé supérieur et du tertiare de Patagonie avec un one primate colonization took place from parallèle entre leurs faunes mammalogiques et celles the South American mainland, as opposed de l’ancien continent. Ann. Mus. Nac. Hist. Nat. ser. to several unrelated events invoked or III 15, 1–568. Ashley-Montagu, M. F. (1933). The anthropological implied by previous studies. The minimum significance of the pterion in the primates. Am. J. date for this colonization event is early phys. Anthrop. 18, 158–336. Miocene, since this is the age of the earliest Bremer, K. (1988). The limits of amino-acid sequence data in angiosperm phylogenetic reconstruction. primate fossil recovered from the Greater Evolution 42, 795–803. Antilles, the Zaza talus. However, the colo- Buffon, G. L. L. (1767). Histoire naturelle générale et nization could have taken place well before particulière avec description du cabinet du roi [with supplement by M. Daubenton]. Vol. 15, p. 327. this, since Paleogene land-mammal fossils Cartmill, M. (1978). The orbital mosaic in prosimians have been recovered from Puerto Rico and the use of variable traits in systematics. Folia (MacPhee & Iturralde-Vinent, 1995) and primatol. 30, 89–114. Cartmill, M., MacPhee, R. D. E. & Simons, E. L. Jamaica (Domning et al., 1997). (1981). Anatomy of the temporal bone in early anthropoids, with remarks on the problem of anthro- poid origins. Am. J. phys. Anthropol. 56, 3–21. Acknowledgements Domning, D. P., Emry, R. J., Portell, R. W., Donovan, S. K. & Schindler, K. S. (1997). Oldest West Indian The specimens of Paralouatta varonai dis- land mammal: rhinocerotoid ungulate from the cussed in this paper were collected between Eocene of Jamaica. J. Vert. Paleont. 17, 638–641. Erikson, G. E. (1963). Brachiation in New World 1988 and 1996 (see Jáimez Salgado et al., monkeys and in anthropoid apes. Symp. Zool. Soc. 1992; MacPhee et al., 1995), and have been Lond. 10, 135–163. incorporated into the permanent collections Fick, R. (1911). Handbuch der Anatomie und Mechanik der Gelenke III. Jena: Fischer. of the MNHNH. For various forms of Fleagle, J. G. (1988). Primate Adaptation and Evolution. assistance over the years, we are happy to New York: Academic Press. record our grateful thanks to Grupo Fleagle, J. G. (1990). New fossil platyrrhines from the Pinturas Formation, southern Argentina. J. hum. Espeleológico Pedro Borrás of the Sociedad Evol. 19, 61–85. Espeleológica de Cuba, especially Efrén Fleagle, J. G. & Bown, T. M. (1983). New primate Jáimez Salgado, Divaldo Gutiérrez Cal- fossils from Late Oligocene (Colhuehaupian) locali- ties of Chubut Province, Argentina. Folia primatol. vache, Rolando Crespo Díaz, and Osvaldo 41, 240–266. Jiménez Vásquez, and the staff of the Fleagle, J. G. & Kay, R. F. (1987). The phyletic posi- MNHNH, especially Manuel Iturralde- tion of the Parapithecidae. J. hum. Evol. 16, 483–531. Fleagle, J. G., Kay, R. F. & Anthony, M. R. L. Vinent, Stephen Díaz Franco, and Reinaldo (1997). Fossil New World monkeys. In (R. F. Kay, Rojas Consuegra. Our thanks also go to R. H. Madden, R. L. Cifelli & J. J. Flynn, Eds) Donald A. McFarlane for permitting us to Vertebrate Paleontology in the Neotropics, pp. 473–495. Washington D.C.: Smithsonian Institution Press. include in our investigation the specimens of Fleagle, J. G., Powers, D. W., Conroy, G. C. & Xenothrix mcgregori collected by him and the Watters, J. P. (1987). New fossil platyrrhines from junior author. Clare Flemming also partici- Santa Cruz Province, Argentina. Folia primatol. 48, 65–77. pated in the fieldwork in Cuba and Jamaica. Fleagle, J. G. & Rosenberger, A. L. (1983). Cranial We are additionally grateful to Lorraine morphology of the earliest anthropoids. In (M. 58 .   . . . 

Sakka, Ed.) Morphologie Evolutive, Morphogenèse Horovitz, I. (1997). Platyrrhine systematics and the du Crâne et Origine de L’Homme, 141–153. Paris: origin of Greater Antilles monkeys. Ph.D. Disser- CNRS. tation. State University of New York at Stony Ford, S. M. (1986a). Subfossil platyrrhine tibia Brook. (Primates: Callitrichidae) from Hispaniola: a possible Horovitz, I. & Meyer, A. (1995). Systematics of the further example of island gigantism. Am. J. phys. New World monkeys (Platyrrhini, Primates) based Anthropol. 70, 47–62. on 16S mitochondrial DNA sequences: A com- Ford, S. M. (1986b). Systematics of the New World parative analysis of different weighting methods in monkeys. In (D. R. Swindler & J. Erwin, Eds) cladistic analysis. Mol. Phylogenet. Evol. 4, 448–456. Comparative Primate Biology, 1: Systematis, Evolution Horovitz, I. & Meyer, A. (1997). Evolutionary trends in and Anatomy, pp. 73–135. New York: Alan R. Liss. the ecology of New World monkeys inferred from a Ford, S. M. (1990). Platyrrhine evolution in the West combined phylogenetic analysis of nuclear, mito- Indies. J. hum. Evol. 19, 237–254. chondrial, and morphological data. In (T. J. Givnish Ford, S. M. & Morgan, G. S. (1988). Earliest primate & K. J. Sytsma, Eds) Molecular Evolution and Adap- fossil from the West Indies. Am. J. phys. Anthropol. tive Radiation, pp. 189–224. Cambridge: Cambridge 75, 209. University Press. Goodman, M. (1973). The chronicle of primate Horovitz, I., Zardoya, R. & Meyer, A. (1998). Platyr- phylogeny contained in proteins. Symp. Zool. Soc. rhine systematics: A simultaneous analysis of London 33, 339–375. molecular and morphological data. Am. J. phys. Goodman, M. (1976). Toward a genealogical descrip- Anthropol. 106, 261–281. tion of the primates. In (M. Goodman & E. Richard, Horovitz, I., MacPhee, R. D. E., Flemming, C. & Eds) Molecular anthropology: Genes and proteins in the McFarlane, D. A. (1997). Cranial remains of evolutionary ascent of the primates, pp. 321–353. New Xenothrix and their bearing on the question of York: Plenum Press. Antillean monkeys origins. J. Vert. Paleont. (Suppl. Goodman, M., Czelusniak, J. & Beeker, J. E. (1985). 3), 54A. Phylogeny of the primates and other eutherian Iturralde-Vinent, M. A. & MacPhee, R. D. E. (in orders: a cladistic analysis using amino acid and press). Cenozoic paleogeography of the Caribbean nucleotide sequence data. Cladistics 1, 171–185. Region: Implications for historical biogeography. Harada, M. L., Schneider, H., Cruz Schneider, M. P., Bull. Amer. Mus. Nat. Hist. Sampaio, I., Czeluzniak, J. & Goodman, M. (1995). Jáimez Salgado, E. J., Gutiérrez Calvache, D. G., DNA evidence on the phylogenetic systematics of MacPhee, R. D. E. & Gould, G. C. (1992). The New World monkeys: support for the sister-grouping monkey caves of Cuba. Cave Science 19, 25–28. of Cebus and Saimiri from two unlinked nuclear Källersjö, M., Farris, J. S., Kluge, A. & Bult, C. (1992). genes. Mol. Phylogen. Evol. 4, 331–349. Skewness and permutation. Cladistics 8, 275–287. Hershkovitz, P. (1970). Notes of Tertiary platyrrhine monkeys and description of a new genus from the late Kay, R. F. (1990). The phyletic relationships of extant Miocene of Colombia. Folia primatol. 12, 1–37. and fossil Pitheciinae. J. hum. Evol. 19, 175–208. Hershkovitz, P. (1974). A new genus of late Oligocene Kay, R. F. & Meldrum, D. J. (1997). A new small monkey (Cebidae, Platyrrhini) with notes on post- platyrrhine and the phyletic position of Callitrichinae. orbital closure and platyrrhine evolution. Folia In (R. F. Kay, R. H. Madden, R. L. Cifelli & J. J. primatol. 21, 1–35. Flynn, Eds) Vertebrate Paleontology in the Neotropics, Hershkovitz, P. (1977). Living New World Monkeys pp. 435–458. Washington D.C.: Smithsonian Insti- (Platyrrhini) with an Introduction to Primates. Chicago: tution Press. Chicago Univ. Press. Kay, R. F. & Williams, B. A. (1994). Dental evidence Hershkovitz, P. (1981). Comparative anatomy of for anthropoid origins. In (J. G. Fleagle & R. F. Kay, Eds) Anthropoid Origins, pp. 361–445. New York: platyrrhine mandibular cheekteeth dpm4,pm4,m1, with particular reference to those of Homunculus Plenum. (Cebidae) and comments on platyrrhine origins. Kay, R. F., Madden, R. H. & Guerrero Díaz, J. (1989). Folia primatol. 35, 179–217. Nuevos hallazgos de monos en el Mioceno de Colombia. Ameghiniana 25, 203–212. Hershkovitz, P. (1984). More on Homunculus dpm4 and M1 and comparisons with Alouatta and Stirtonia. Kay, R. F., Madden, R. H., Plavcan, J. M., Cifelli, Am. J. Primatol. 7, 261–283. R. L. & Guerrero Díaz, J. (1987). Stirtonia victoriae,a Hill, J. P. (1926). Demonstration of the embryologia new species of Miocene Colombian primate. J. hum. varia (Development of Hapale jacchus). J. Anat. 60, Evol. 16, 173–196. 486–487. Kinzey, W. G. (1973). Reduction of the cingulum Hoffstetter, R. (1969). Un primate de l’oligocène in the Ceboidea. In (W. Montagna, Ed.) Symp. inférieur sud-américain: Branisella boliviana gen. et Fourth Int. Congr. Primatol., 3, pp. 101–127. Basel: sp. nov. C. r. hebd. Séanc. Acad. Sci. sér. D 269, Karger. 434–437. Kraglievich, J. L. (1951). Contribuciones al cono- Horovitz, I. (1995). A phylogenetic analysis of the cimiento de los primates fósiles de la Patagonia. basicranial morphology of New World monkeys. Am. I. Diagnosis previa de un nuevo primate fósil del J. phys. Anthropol. (Suppl. 20), 113. Oligoceno superior (Colheuhuapiano) de Gaiman,   PARALOUATTA VARONAI 59

Chubut. Comm. Inst. Nac. Invest. Cien. Nat. II, Rosenberger, A. L. (1984). Fossil New World monkeys 57–82. dispute the molecular clock. J. Hum. Evol. 13, 737– Lewis, O. J. (1974). The wrist articulation of the 742. Anthropoidea. In (F. A. Jenkins, Ed.) Primate Loco- Rusconi, C. (1934). Nuevos restos de monos del motion, pp. 143–169. New York: Academic Press. terciario antiguo de la Patagonia. An. Soc. Cient. MacFadden, B. J. (1990). Chronology of Cenozoic Argent. 116, 286–289. primate localities in South America. J. hum. Evol. 19, Rusconi, C. (1935). Las especies de primates del 7–21. Oligoceno de Patagonia. Ameghiniana 1, 39–125. MacPhee, R. D. E. (1996). The Greater Antillean Saban, R. (1963). Contribution à l’étude de l’os tem- monkeys. Rev. Cièn. (IEB) 18, 13–32. poral des Primates. Description chez l’Homme et les MacPhee, R. D. E. & Cartmill, M. (1986). Basicranial Prosimiens. Anatomie comparée et phylogénie. Mém. structure and primate systematics. In (D. R. Mus. nat. Hist. Nat., sér. A, Zoologie 29, 1–378. Swindler & J. Erwin, Eds) Comparative Primate Schneider, H., Sampaio, I., Harada, M. L., Barroso, Biology, 1: Systematics, Evolution and Anatomy, C. M. L., Schneider, M. P. C., Czelusniak, J. & pp. 73–135. New York: Alan R. Liss. Goodman, M. (1996). Molecular phylogeny of the MacPhee, R. D. E. & Fleagle, J. G. (1991). Postcranial New World monkeys (Platyrrhini, Primates) based remains of Xenothrix mcgregori (Primates, Xenotrichidae) on two unlinked nuclear genes: IRBP Intron 1 and and other Late Quaternary mammals from Long Mile å-globin sequences. Am. J. phys. Anthropol. 100, Cave, Jamaica. In (T. A. Griffiths & D. Klingener, 153–179. Eds) Contributions to mammalogy in honor of Karl F. Schneider, H., Schneider, M. P. C., Sampaio, I., Koopman. Bull. Am. Mus. Nat. Hist. 20, 287–321. Harada, M. L., Barroso, C. M. L., Czelusniak, J. & MacPhee, R. D. E., Horovitz, I., Arredondo, O. & Goodman, M. (1995). DNA evidence on platyrrhine Jiménez Vázquez, O. (1995). A new genus for the phylogeny from two unlinked nuclear genes. Am. J. extinct Hispaniolan monkey Saimiri bernensis Rímoli, phys. Anthropol. (Suppl. 20), 191. 1977, with notes on its systematic position. Am. Mus. Schneider, H., Schneider, M. P. C., Sampaio, I., Novitat. 3134, 1–21. Harada, M. L., Stanhope, M., Czelusniak, J. & MacPhee, R. D. E. & Iturralde-Vinent, M. A. (1995). Goodman, M. (1993). Molecular phylogeny of the Earliest monkey from Greater Antilles. J. hum. Evol. New World monkeys (Platyrrhini, Primates). Mol. 28, 197–200. Phylogenet. Evol. 2, 225–242. MacPhee, R. D. E. & Woods, C. A. (1982). A new Schultz, A. H. (1930). The skeleton of the trunk and fossil cebine from Hispaniola. Am. J. phys. Anthropol. limbs of higher primates. Hum. Biol. 2, 303–438. 58, 419–436. Schultz, A. H. (1961). Vertebral column and thorax. Martin, R. D. (1992). Goeldi and the dwarfs: the Primatologia 4, 1–66. evolutionary biology of the small New World Setoguchi, T., Watanabe, T. & Mouri, T. (1981). The monkeys. J. hum. Evol. 22, 367–393. upper dentition of Stirtonia (Ceboidea, Primates) Napier, J. R. (1961). Prehensility and opposability in from the Miocene of Colombia, South America, the hands of Primates. Symp. Zool. Soc. London 10, and the origin of the postero-internal cusp of 115–132. upper molars of howler monkeys Alouatta. Kyoto Orlosky, F. J. (1973). Comparative dental morphology Univ. Overseas Res. Rep. of New World Monkeys, of the extant and extinct Cebidae. Ph.D. Disser- 51–60. tation. University of Washington. Simons, E. L. (1959). An anthropoid frontal bone from Pocock, R. I. (1925). Additional notes on the external the Oligocene of Egypt: the oldest skull fragment of a characters of some platyrrhine monkeys. Proc. Zool. higher primate. Am. Mus. Novitat. 1976, 1–16. Soc. London, 27–42. Stirton, R. A. (1951). Ceboid monkeys from the Rímoli, R. (1977). Una nueva especie de monos Miocene of Colombia. Bull. Univ. Calif. Pub. Geol. (Cebidae: Saimirinae: Saimiri) de la Hispaniola. Sci. 28, 315–356. Cuadernos de CENDIA, Universidad Autónoma de Stirton, R. A. & Savage, D. E. (1951). A New Monkey Santo Domingo 242, 1–14. from the La Venta Late Miocene of Colombia. Bogotá: Rivero, M. & Arredondo, O. (1991). Paralouatta Servicio Geol. Nac. varonai, a new Quaternary platyrrhine from Cuba. Williams, E. E. & Koopman, K. F. (1952). West Indian J. hum. Evol. 21, 1–11. fossil monkeys. Am. Mus. Novitat. 1546, 1–16. Wislocki, G. B. (1939). Observations on twinning in Rosenberger, A. L. (1977). Xenothrix and ceboid phy- marmosets. Amer. J. Anat. 64, 445–483. logeny. J. hum. Evol. 6, 461–481. Rosenberger, A. L. (1979). Phylogeny, evolution and classification of New World monkeys (Platyrrhini, Primates). Ph.D. Dissertation. City University of Appendix 1 New York. Rosenberger, A. L. (1981). Systematics: The higher Character list taxa. In (A. F. Coimbra-Filho & R. A. Mittermeir, NOTE: Characters that are multistate and Eds) Ecology and Behavior of Neotropical Primates, 1, pp. 111–168. Rio de Janeiro: Academia Brasileira de non additive are noted. All others are Ciencias. additive. 60 .   . . . 

(1) Offspring per birth, number (Wislocki, ( 7) Carpo-metacarpal joint of thumb 1939; Hill, 1926): 0=one, 1=two. (Napier, 1961; Fick, 1911): 0=non- Callithrix, Cebuella, Saguinus, and saddle, 1=saddle. Catarrhines possess Leontopithecus always have dizygotic a saddle-shaped joint; all other haplo- twins. Singletons occur rarely and are rhines possess conformations that are believed to be survivors of dizygotic not saddle-shaped. Napier (1961) pregnancies (Hershkovitz, 1977). described saddle-type articulating sur- ( 2) Lumbar vertebrae, number (Erikson, faces as reciprocally concavo–convex 1963): 0=more than five, 1=five or in section and incongruent in at least fewer. one plane in joint’s mid-position. ( 3) External thumb (Pocock, 1925): ( 8) Rib cage, shape (Schultz, 1961): 0=absent or reduced, 1=present. 0=larger dorso-ventrally, 1=larger Thumb is nearly or completely absent laterally. Platyrrhines and Old World as an external feature in Ateles and monkeys differ from hominoids in the Brachyteles. In the latter, it is reduced transverse shape of their rib-cages: in to a very small metacarpal and a single monkeys, cages are taller than they phalanx. are wide, in hominoids the reverse ( 4) External tail: 0=absent (not project- obtains. ing), 1=present. External tail universal ( 9) Ulnar participation in wrist articu- in New World monkeys; not project- lations (Lewis, 1974): 0=present, ing (as coccyx) in hominoids. 1=absent. In most primates except ( 5) Tail, ventral glabrous surface (Pocock, hominoids, ulna articulates with tri- 1925): 0=absent, 1=present. Several quetral and pisiform. In hominoids, genera of New World monkeys display ulna does not articulate with any wrist different degrees of prehensility in bones. their tails, which they use either as an (10) Sternebral proportions (Schultz, additional point of support, to suspend 1930): 0=manubrium shorter than themselves, and/or as a fifth member 36% of corpus length, 1=manubrium in locomotion. However, only four longer than 46% of corpus length. genera (Ateles, Lagothrix, Brachyteles, (11) Orbit size (Character 4, MacPhee and Alouatta) display a patch of naked et al., 1995): 0=smaller than 1.9, 1= skin on the distal end of the ventral larger than 2.1. Size measured as orbit side of the tail. This feature is readily height divided by foramen magnum recognizable on specimens and is pre- width; definition of states takes advan- sumably related to use in suspension. tage of gap in distribution of ratios. ( 6) Claws on all manual and pedal digits (12) Postglenoid foramen (Horovitz, 1997): except hallux (Buffon, 1767): 0= 0=absent, 1=reduced, 2=large. Cod- absent, 1=present. Of taxa considered ing of this character distinguishes in this study, only five genera possess between large foramina, through this feature as defined. Claws and which interior of braincase is easily nails are both horny coverings of the visible, and conditions in which terminal phalanges; they differ in that foramina are reduced or absent (in- claws are long, sharp, laterally com- terior of braincase not visible). pressed, transversely convex and pro- (13) Tentorium cerebelli, ossification truding, whereas nails are broad, (Hershkovitz, 1977; Horovitz, 1995): blunt, and protrude little or not at all 0=absent, 1=present. Tentorium cer- (Hershkovitz, 1977). ebelli presents some degree of ossifica-   PARALOUATTA VARONAI 61

tion in all New World monkey genera processes is shallow, and does not (coded as present); it is unossified in excavate base of skull. all outgroups in which this region (17) Canal connecting sigmoid sinus could be explored (all extant outgroup and subarcuate fossa (Character taxa used in this study, coded as 6, MacPhee et al., 1995)(Cartmill, absent). Degree of ossification is vari- 1981; Horovitz, 1995): 0=absent, able, with atelines showing greatest 1=present. Present in all examined degree and callitrichines and Saimiri platyrrhines: Brachyteles scored as the least. Saimiri is polymorphic (some unknown because status could not be individuals show slight ossification, determined in material available (see others none) (see Figure 9). Figure 9). (14) Middle ear, pneumatization of antero- (18) Vomer, exposure in orbit (Cartmill, ventral region (Horovitz, 1997): 1978; Rosenberger, 1979): 0=absent, 0=absent, 1=present. Anterior floor of 1=present. Exposed only in Cebus and middle ear in Callithrix, Cebuella, and Saimiri. Leontopithecus displays typical macro- (19) Ectotympanic, shape (nonadditive): scopic consequences of pneumatic 0=tube I, 1=ring, 2=tube II. Catar- activity (Cartmill et al., 1981). These rhines and Tarsius display a tube- include formation of septa, small shaped ectotympanic. However, they vacuities or cellules, and repositioning differ in detail, suggesting nonhomol- of entire sheets of bone such that shape ogy (scored as ‘‘I’’ and ‘‘II’’). For of internal carotid canal is revealed on discussion, see MacPhee & Cartmill floor. (1986). (15) Middle ear, paired prominences on (20) Temporal emissary foramen cochlear housing (Horovitz, 1997): (Character 7, MacPhee et al., 1995): 0=absent, 1=present. Lateral wall of 0=present and large, 1=small or cochlear housing, exposed in medial absent (see Figure 3). wall of middle ear cavity, shows vari- (21) Eyeball physically enclosed (Martin, able relief in New World monkeys. In 1992): 0=absent, 1=present. Diam- some taxa it is flat or displays a single eter of external orbital aperture is prominence. In others (callitrichines, smaller than greatest internal diameter Cebus, Aotus, Saimiri, Callicebus, of eyeball in Callithrix, Cebuella, Pithecia) there is an additional promi- Leontopithecus, and Saguinus. This is nence, situated higher and more medi- so because the eyeball is physically ally than the first. Also presumably enclosed by the external orbital margin present in Apidium. (Martin, 1992). (16) Pterygoid fossa, depth (Horovitz, (22) Cranial capacity (Horovitz, 1997): 1997): 0=deep, 1=shallow. Pterygoid 0=less than 15 cm3, 1=more than fossa is defined by medial and lateral 15 cm3. Cranial capacities of New pterygoid processs. In Paralouatta, World monkey genera overlap over atelines, living pitheciins, Callicebus, a wide range of values; however, a and callitrichines, medial pterygoid gap (reflected in character definition) process is greatly reduced. Instead of separates Callithrix, Cebuella, Callimico, originating from base of skull, it is Leontopithecus, and Saguinus (as a restricted to medial side of lateral group) from other genera. pterygoid process. As a consequence, We use this character instead of the pterygoid fossa defined by these more traditionally used body size 62 .   . . . 

because the former displays an actual present, 2=styliform, lingual heel

gap in its distribution, whereas there is absent. Deciduous I2sofCallithrix and a slight overlap in body size between Cebuella are unique among examined Saimiri and the callitrichines. haplorhines: they are bladelike, verti- (23) Zygomatic arch, ventral extent cally implanted in mandible, with flat (Horovitz, 1997): 0=below plane of lingual and buccal sides (no lingual alveolar border, 1=above plane of heel). Deciduous incisors of Pithecia, border. In most terminal taxa, entire Cacajao, and Chiropotes also lack a zygomatic arch is located above tooth lingual heel, but they are styliform and row. Some degree of ventral projection closely resemble permanent ones. All may be present, but rarely exceeds other taxa have spatulate deciduous level of plane of alveolar border (Cal- and permanent mandibular incisors licebus, Aotus, and Xenothrix only). with lingual heels.

(24) Pterion region, contacts (Ashley- (29) Relative height of I1–I2 (Rosenberger, Montagu, 1933): 0=zygomatic- 1979): 0=I1 absent, 1=I1 lower than parietal, 1=frontal-alisphenoid. In I2, 2=I1–I2 subequal. Tarsius is only platyrrhines, zygomatic and parietal taxon in this analysis having a single bones are almost always in contact incisor on each side. In Callithrix, on external surface of skull in norma Cebuella, and Alouatta, central incisors lateralis. In all outgroups in which are lower than lateral ones; subequal in condition is known, frontal contacts all other taxa. AMNHM Brachyteles alisphenoid (see Figure 3). specimens not evaluated because teeth (25) Nasal fossa width (Horovitz, 1997): too worn to interpret. 1 0=narrower than palate at level of M , (30) Alignment of I1–I2 (Hershkovitz, 1=wider. In anthropoids, nasal fossa is 1970, 1977; Rosenberger, 1979): generally narrower than width of palate 0=transversely arcuate, 1=staggered. measured between lingual edges of Hershkovitz (1970, 1977) noted that, alveoli of M1s. By contrast, Paralouatta uniquely in Callithrix and Cebuella, and Xenothrix possess unusually wide mandibular incisors are staggered in nasal fossae that (at their maximum such a way ‘‘that their mesiodistal axes width) are broader than palate (see are not linearly aligned but parallel Figure 10). [to] one another in a ranked, en echelon (26) Infraorbital foramen, vertical posi- spacing’’ (Rosenberger, 1979: 107). tion relative to maxillary cheekteeth Incisors arrayed in transverse or in Frankfurt plane (Character 5, arcuate arrangement in all other taxa.

MacPhee et al., 1995): 0=above inter- (31) Permanent I1–I2, shape (Rosenberger, val between (or caudal to) M1 and P4, 1979): 0=spatulate, 1=styliform. Per- 1=above interval between P4 and P3, manent mandibular incisors of Tarsius, 2=above (or rostral to) anteriormost Callithrix, Cebuella, Pithecia, Cacajao premolar (see Figure 3). and Chiropotes are mesiodistally com- (27) Zygomaticofacial foramen, size relative pressed and display no lingual heel; in to maxillary M1 breadth (Character 1, all other taxa they are spatulate, with a MacPhee et al., 1995): 0=small, lingual heel.

1=large (see Figure 3). (32) Mesostyles and distostyles of I1–I2 (28) Deciduous I2, shape (nonadditive) (Hershkovitz, 1977): 0=absent, 1= (Horovitz, 1997): 0=bladelike, lingual present. Mandibular incisors in heel absent, 1=bladelike, lingual heel Callithrix and Cebuella display small   PARALOUATTA VARONAI 63

mesial and distal cusps flanking major (43) Deciduous P3, metaconid (Kay & cone. Meldrum, 1997): 0=absent, 1=

(33) Diastema between C1 and I2 present. (Rosenberger, 1979): 0=absent, 1= (44) Protoconid of P3, size relative to P4 present. Pithecia, Cacajao, Chiropotes, protoconid (Horovitz, 1997): 0=P3 and Cebupithecia all display man- and P4 protoconids subequal, 1=P3 dibular diastemata large enough to protoconid largest.

accommodate maxillary canines when (45) Talonid of P3 (Horovitz, 1997): tooth rows are in complete occlusion. 0=larger than P2 talonid, 1=subequal (34) Root of C1 shape (Character 11, to P2 talonid. MacPhee et al., 1995): 0=rounded/ (46) Metaconid of P3, height relative to suboval, 1=highly compressed. protoconid height (Rosenberger,

(35) Lingual cingulum on C1, complete- 1979): 0=metaconid absent, 1=meta- ness (Kinzey, 1973): 0=complete, conid lower than protoconid, 2= 1=incomplete or absent. metaconid and protoconid subequal,

(36) Lingual crest on C1, sharpness (Kay, 3=metaconid taller than protoconid. 1990): 0=rounded, 1=sharp. (47) Metaconid of P4, height relative to (37) Lingual cingulum on C1, mesial elev- protoconid height (Rosenberger, ation of (Horovitz, 1997): 0=not 1979): 0=metaconid lower than proto- elevated, 1=elevated. In most taxa, conid, 1=metaconid and protoconid lingual cingulum of mandibular subequal, 2=metaconid taller than canines is mesially elevated (i.e., cin- protoconid.

gular band is tilted relative to the long (48) Hypoconid of P4 (Kay & Williams, axis of the tooth such that it intersects 1994): 0=absent, 1=present.

tooth’s mesial edge at higher level than (49) Entoconid of P4,(Kay & Williams, its distal edge, usually the lowest point 1994): 0=absent, 1=present. the cingulum reaches). (50) Number of premolars: 0=two,

(38) Lingual cingulum on C1, forming spike 1=three. on mesial edge of tooth (Horovitz, (51) M1 projection of distobuccal quad- 1997): 0=absent, 1=present. Lingual rant (DB complex) (Character 14, cingulum of mandibular canine termi- MacPhee et al., 1995): 0=not project- nates in a spikelike feature on mesial ing, 1=projecting (crown sidewall edge of tooth. In other taxa, this region hidden in occlusal view).

is flat, not spikelike. (52) M1 intersection of oblique cristid and (39) Buccolingual breadth of alveolus of C1 protolophid (Character 15, MacPhee compared to P4 (Horovitz, 1997): 0=C1 et al., 1995): 0=intersects protolophid larger than P4, 1=C1 smaller than P4. buccally, directly distal to apex of (40) Deciduous P2, angle subtended by dis- protoconid, 1=intersects protolophid tal portion of mesiodistal axis and more lingually, distolingual to apex of postprotocristid (Horovitz, 1997): 0= protoconid.

smaller than 45), 1=larger than 45). (53) M1 buccal bulging of protoconid (41) Deciduous P2, cross-sectional shape (Horovitz, 1997): 0=absent, 1= (Horovitz, 1997): 0=rounded, 1= present. In both Paralouatta and

mesiodistally elongated. Xenothrix,M1 displays a very convex, (42) Size of P2, relative to P3 and P4 bulging buccal surface, especially in (Horovitz, 1997): 0=P2 smallest in the region buccal to the protoconid. premolar series, 1=P2 not smallest. When seen from occlusal view, this 64 .   . . . 

feature makes it seem that the mass of except in Callithrix and Cebuella, which the protoconid and, to a lesser degree, lack a differentiated trigon. that of the hypoconid, are placed more (64) Deciduous P3, hypocone (Horovitz, lingually (‘‘centrally’’) than in other 1997): 0=absent, 1=present. platyrrhines. (65) P3 preparacrista (Horovitz, 1997):

(54) M1 entoconid position (Rosenberger, 0=absent or vestigial, 1=present. 1977): 0=on talonid corner, 1=dis- (66) P4 protocone position (Character 23, tally separated from talonid corner by MacPhee et al., 1995): 0=mesial to sulcus. Rosenberger (1977) noted that widest point of trigon, 1=on widest Xenothrix resembles Callicebus and point. Pithecia in displaying a sulcus which (67) P4 lingual cingulum (Kinsey, 1973): separates the entoconid from the post- 0=absent, 1=present. cristid in such a way that the cusp (68) P4 lingual cingulum mesial projection does not appear to be located on (Character 22, MacPhee et al., 1995): tooth’s distolingual corner. Alouatta 0=absent, 1=present. and Paralouatta also display such a (69) P4 hypocone (Kay, 1990; MacPhee sulcus. et al., 1995): 0=absent, 1=present. 4 1 (55) M1/M2 buccal cingulum (Kinzey, (70) P and M , relative buccolingual 1973): 0=absent, 1=present. breadth (MacPhee et al., 1995): 0=P4 1 4 (56) M2 trigonid/talonid relative height narrower than M , 1=P subequal to (Kay, 1990): 0=trigonid taller than or wider than M1. talonid, 1=subequal. (71) M1 mesostyle/mesoloph (nonadditive)

(57) M2 mesoconid (Horovitz, 1997): (Kinzey, 1973): 0=absent, 1=meso- 0=absent, 1=present. style present, 2=mesoloph present. 1 (58) M3/P4 relative length (Horovitz, (72) M hypocone/prehypocrista presence 1997): 0=M3 absent, 1=M3 shorter, (Rosenberger, 1979; Character 30, 2=subequal, 3=M3 longer. MacPhee et al., 1995): 0=hypocone (59) Molar enamel surface (Rosenberger, and prehypocrista present, 1= 1977): 0=smooth, 1=crenulated. hypocone present and prehypocrista (60) I1 lingual heel (Rosenberger, 1979): absent, 2=hypocone and prehypo- 0=absent, 1=present. crista absent. (61) I2 orientation (Rosenberger, 1979): (73) M1 postmetacrista slope (Character 0=vertical, 1=proclivious. 26, MacPhee et al., 1995): 0=disto- (62) C1 alveolus size relative to P4 equiva- buccal slope, 1=distal or distolingual lent (Character 21, MacPhee et al., slope. 1995): 0=C1 larger than P4, 1=C1 (74) M1 mesiodistal alignment of proto- smaller or equal to P4. Size is cone and hypocone (Character 27, measured as length times width of each MacPhee et al., 1995): 0=parallel, alveolus. The two characters reflecting 1=hypocone lingual. canine size (39 and 62) show different (75) M1 pericone/lingual cingulum distributions across taxa, and therefore (Character 29, MacPhee et al., act in ensemble as additive multistate 1995): 0=absent, 1=lingual cingulum character describing canine size. only, 2=distinct pericone on lingual (63) Deciduous P2, trigon (Horovitz, cingulum. 1997): 0=absent, 1=present. Decidu- (76) M2 hypocone (Rosenberger, 1979; ous P2 is morphologically similar to Character 32, MacPhee et al., 1995): permanent P2 in most platyrrhines, 0=absent, 1=present.   PARALOUATTA VARONAI 65

(77) M2 cristae on distal margin of trigon (79) Maxillary M’s parastyles (Horovitz, (nonadditive) (Character 31, MacPhee 1997): 0=absent, 1=present. et al., 1995): 0=cristae form distinct, (80) Ventral flexion of the skull (airo- continuous wall between protocone rhynchy): 0=absent, 1=present. The and metacone, 1=cristae interrupted angle that the basioccipital forms with by fossette or do not form distinct the tooth row varies among platyr- wall, 2=cristae absent or differently rhines. The largest angle occurs in organized. Alouatta; the next largest in Paral- (78) M3/P4 relative mesiodistal length ouatta, although this species does not (Rosenberger, 1979; Horovitz, 1997): actually greatly differ from the average 0=M3 absent, 1=M3 shorter than P4, for Platyrrhini. Nevertheless, both 2=M3 and P4 subequal, 3=M3 longer were scored as having ventrally flexed than P4. Mandibular (Character 53) skulls. and maxillary third molars sizes rela- tive to P /P4 show different distribu- 4 Appendix 2 tions in their character states among taxa scored. Since they provide differ- Matrix of morphological characters ent information they were both employed to obtain most parsimonious trees included. whose consensus is shown in Figure 8. 66

Appendix 2 1

123456789101112131415161718192021222324252627

1 Tarsius 001100000 0 1 2 0 0 0 0 0 0200011010 2 Leontopithecus 101101000 0 0 2 1 1 1 1 1 0111010020 3 Saguinus 101101000 0 0 2 1 0 1 1 1 0111010020  4 Callimico 001101000 0 0 0&11 0 1 0 1 0111010021 . 5 Callithrix 101101000 0 0 2 1 1 1 1 1 0111010020    6 Cebuella 101101000 0 0 2 1 1 1 1 1 0111010020 7 Aotus 001100000 0 1 2 1 0 1 0 1 0111100020 8 Cebus 001100000 0 0 2 1 0 1 0 1 1110110020 9 Cacajao 001100000 0 0 0&11 0 0 1 1 0110110020 10 Pithecia 001100000 0 0 1 1 0 1 1 1 0110110020 11 Chiropotes 001100000 0 0 0&11 0 0 0&11 0110110020

12 Saimiri 001100000 0 0 2 0&10 1 0 1 1100110020 . 13 Alouatta 011110000 0 0 1 1 0 0 1 1 0100110001  .

14 Lagothrix 011110000 0 0 0&11 0 0 1 1 0100110011  15 Brachyteles 010110000 0 0 1 1 0 0 1 1 0100110011 .  16 Callicebus 001100000 0 0 0 1 0 1 1 1 0111100021 17 Ateles 010110000 0 0 0 1 0 0 1 0&1 0100110011 18 Homo 0010?0111 1 0 0 0 1 0 0 0 0010111?00 19 Hylobates 0100?0111 1 0 0 0 0 0 0 0 0010111010 20Cercopithecoids0011001000 0 0 0 0 0 0 0 0010111000 21 Aegyptopithecus ???1??0???0 2 ? 00 0 0 ?1?0111?00 22 Cebupithecia ????????? ? ? 0 1 0 0 ? 1 ? 1 ??????20 23 Paralouatta ????????? ? 1 2 ? 0 1 1 ? 01101?0101 24 Antillothrix ????????? ? ? ? ? ? ? ? ? ?????????? 25 Xenothrix ????????? ? ? ? ? ? ? ? ? ?????0?11? 26 Stirtonia ????????? ? ? ? ? ? ? ? ? ?????????? Appendix 2 2

28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

1 Tarsius 10?100010 1 0 10 1 0 001 1 0 0 010 0 00 2 Leontopithecus 120000010 1 1 00 1 1 ?10 1 1 0 010 0 00   3 Saguinus 120000010 1 1 00 1 1 0101&21 0 010 1 00 4 Callimico 120000010 1 1 00 1 1 ?00 2 1 0 011 1 00 5 Callithrix 011110100 1 0 00 1 1 011 1 0 0 010 0 00 6 Cebuella 011110100 1 0 00 1 1 011 1 0 0 010 0 00 7 Aotus 1200000100&10&101 1 1 100 21&20&1110 0 00 8 Cebus 120000000 1 0 01 1 1 100 3 2 1 110 0 00 9 Cacajao 220101011 0 0 01 0 0 100 1 1 1 110 0 01 10 Pithecia 220101011 0 0 01 00&1100 11&21 110 0 01 11 Chiropotes 220101011 0 0 01 00&1100 2 1 1 110 0 01 12 Saimiri 1200000100&11 000&11 100 2 1 1 1100&100 VARONAI PARALOUATTA 13 Alouatta 110000010 1 0 01 0 1 100 1 0 1 111 1 01 14 Lagothrix 1200000100&10 01 0 1 100 2 10&1110 0 01 15 Brachyteles ? ?0000010 1 0 0? ? 0 ?00 1 10&1111 0 0? 16 Callicebus 120000110 1 0 01 0 0 100 1 1 1 110 0 01 17 Ateles 120000010 1 0 01 0 1 100 2 10&1010 0 01 18 Homo 120000100 ? ? 1? ? ? ? ? ? 1 00&1100 0 01 19 Hylobates 120000010 1 0 0? ? ? 1? ? 0 1 1 100 0 01 20Cercopithecoids120000010 1 1 0? ? ? 1? ? 0 1 1 101 0 01 21 Aegyptopithecus ?20000110 1 0 0? ? ? ? ? ? 0 0 0 000 0 01 22 Cebupithecia ? ?0? ?1011 0 ? 0? ? ? ? ? ? ? ? 1 010 0 01 23 Paralouatta ? ?0000110 1 0 1? ? 0 ?00 2 1 1 110 1 11 24 Antillothrix ????????? ? ? ?? ? 0 ??? ? ? ? ?101 ?? 25 Xenothrix ??0??01?? ? ? 1? ? 0 ??? ? ? ? ?100 11 26 Stirtonia ?????0010 1 1 0? ? 1 ? ? ? 1 ? 0 111 1 01 67 68

Appendix 2 3

55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80

1 Tarsius 10030 0 01000? 1000020?100310 2 Leontopithecus 10000 0 001?0010000&121?100010 3 Saguinus 10000 0 00100110000&120?100010  4 Callimico 00010 0 001?0110001011110110 . 5 Callithrix 10000 0 00000110000&120?100010    6 Cebuella 10000 0 0000010&1000121?100010 7 Aotus 00020 0 001001100000100&110100 8 Cebus 01010 0 00101110010110011100 9 Cacajao 01121 1 101011 0 ?0&11 0 110111200 10 Pithecia 01111 1 101011 0 ?0&10 0 010111300 11 Chiropotes 01121 1 101111 0 ?0&11 0 110111200

12 Saimiri 10010 0 00110110010&1100210110 . 13 Alouatta 00030 1 001000 0 ?0&10 2 1000&11 1 301  .

14 Lagothrix 00030 1 001000 0 ? 100110010300  15 Brachyteles 00030 ? 00? ?00 0 ? 00211001?2?0 .  16 Callicebus 01030 1 01100110100010111200 17 Ateles 00030 1 001000 0 ? 100110010200 18 Homo 010300&100? ?01 0 ? 0 0 0 1100&11 0 300 19 Hylobates 00030 1 00?001 0 ? 0 0 0 1100&11 0 300 20Cercopithecoids00030 1 10?101 0 ? 0 0 0 1100&11 2 310 21 Aegyptopithecus 10030 1 00? ?0110100&1100110&1310 22 Cebupithecia 0??20?10??1110?0001011?11? 23 Paralouatta 00030 0 ?1? ?0011011001210201 24 Antillothrix ????? ? ?0???1 1 1 ? 1 0 101211?1? 25 Xenothrix 01100 ? ?1? ?11100000100? ?00? 26 Stirtonia 000?0? ???0000?102110111?1?