Axial skeleton of Cebupithecia sarmientoi (Pitheciinae Platyrrhini) from the middle miocene of La Venta Colombia.код для вставкиСкачать
American Journal of Primatology 2569-89 (1991) RESEARCH ARTICLES Axial Skeleton of Cebupithecia sarmientoi (Pitheciinae, Platyrrhini) From the Middle Miocene of La Venta, Colombia D. JEFFREY MELDRUM' AND PIERRE LEMELIN' 'Department of Biological Anthropology and Anatomy, Duke University Medical Center, Durham, North Carolina; 'Doctoral Program in Anthropological Sciences, State University of New York, Stony Brook The axial skeleton of Cebupithecia sarmientoi is described and analyzed for its functional and phylogenetic implications. The vertebrae of the holotype of C. sarmientoi (UCMP 38762) most closely resemble those of the extant pitheciine genus Pithecia and display features associated with adaptations for clinging and leaping as in that genus. Cebupithecia has a relatively long non-prehensile tail, which is most similar in absolute dimensions and proportions to Pithecia monachus. It also shares with P. monachus a distinctive morphology of the thoracic vertebrae, specifically the presence of a bony pillar spanning the vertebral lamina and body, caudal to the pedicle, herein designated the vinculum laminum. It is proposed t h a t many of these features are shared primitive retentions from the last common ancestor of the Cebupithecia-pitheciine clade. Key words: positional behavior, vertebral morphology, vinculum laminum, vinculum transversarum, Pithecia, New World monkeys, Miocene Platyrrhini INTRODUCTION The axial skeleton and limb bones of Cebupithecia sarmientoi (UCMP 38762) constitute the only associated and relatively complete postcranial skeleton of a fossil platyrrhine known. This specimen was recovered by R.A. Stirton in 1945 [Stirton & Savage, 19511 from the monkey unit in the Baraya Member of the Villavieja Formation, Honda Group. The appendicular skeleton is well represented and has been the subject of a number of studies. The axial skeleton is equally well represented, but with the exception of the craniodental fragments, little attention has been given to it. We identify among the miscellaneous fragments associated with the holotype, a number of additional vertebral elements, as well as rib fragments that were Received for publication May 22, 1990; revision accepted December 7, 1990. Address reprint requests to D. Jeffrey Meldrum, Dept. of Cell, Molecular and Structural Biology, Northwestern University, 303 East Chicago Ave., Chicago, IL 60611-3008. 0 1991 Wiley-Liss, Inc. 70 / Meldrum a n d Lemelin previously unknown. Together, the vertebral column is now represented by a partial atlas and axis, 5 thoracic vertebrae, with 7 additional fragments, 3 nearly complete lumbar vertebrae, with 4 additional fragments, and 18 caudal vertebrae. Our conclusions from the examination of this material differ slightly from the original description by Stirton and Savage  and the more extensive comparative analysis of Stirton [ 19511. Moreover, certain vertebrae were reidentified based on our comparative observations with extant platyrrhine samples. The status of the postcranial remains of C. sarmientoi has been controversial. Some authors [i.e., Rosenberger, 1979; Ford, 19861 suggest that the craniodental and postcranial remains of the type specimen do not belong to a single individual 6 e . , the skull of a n immature individual associated with the adult postcranial skeleton of a second). However, recent studies have demonstrated the integrity of the entire skeleton based on dental eruption and wear, and basicranial morphology [Meldrum & Fleagle, 1988; Kay, 19901. Altering her previous position, Ford  has adopted this point of view concerning the unity of UCMP 38762. The craniodental material of Cebupithecia displays marked similarities with living pitheciines, such as procumbent incisors, distinctive triangular-shaped everted lower canines, premolars and molars with low cusps and shallow occlusal basins, and deep mandible with a robust symphysis [Stirton, 1951; Stirton & Savage, 1951; Rosenberger, 1979; Orlosky, 1980; Kay, 19901. The postcranial skeleton is characterized by morphological traits associated with clinging and leaping, features typical of Pithecia pithecia [Fleagle & Meldrum, 19881 a s well as a number of small-bodied platyrrhines and callitrichines [Davis, 1988; Meldrum & Fleagle, 1988; Meldrum & Kay, 1990; Meldrum et al., 19901. These craniodental and postcranial traits evince the pitheciine affinities of Cebupithecia. At the same time it displays many primitive features, and lacks a number of the synapomorphies defining the monophyletic pitheciines, such as a large diastema between I2 and C,, enlarged P4, and crenulated molar enamel [Delson & Rosenberger, 1984; Kay, 19901. This is precisely what one would expect for a species near the base of a radiation. In this study we will 1)describe and analyze the vertebrae of C. sarmientoi, 2) address the functional implications for its positional repertoire, including the question of a prehensile tail, and 3) discuss the phylogenetic implications of these observations for the relationship of Cebupithecia to the Pitheciinae. METHODS The associated miscellaneous bone fragments of the axial skelton of (UCMP 38762) were cleaned and identified. Comparisons were made with a large sample of extant platyrrhine axial skeletons from the collections of the Department of Anatomical Sciences, SUNY a t Stony Brook, and the Mammalogy Section, American Museum of Natural History. The number and dimensions of vertebrae in the thoracic and lumbar regions were noted in 43 specimens of selected taxa, a s well as the position of the “anticlinal” vertebra [c.f. Erikson, 19631. Finally, measurements were made of each caudal vertebra, including the length, proximal width, and median width. Wet specimens of platyrrhines were dissected to examine selected musculoskeletal relationships of the vertebral column, with particular attention to the lumbar and caudal region. Specimens included Pithecia pithecia, Alouatta senicuZus, and additional specimens of prehensile-tailed platyrrhines [Lemelin, 1988, 1989, in prep.]. Axial Skeleton of Cebupithecia sarmientoi I 71 A B Fig. 1. Stereophotographs of the partial atlas and axis, cervical vertebrae 1 and 2, of Cebupithecia sarmientoi, seen in ventral (A) and dorsal views (B). Scale bar = 1 cm. RESULTS Basicraniurn and Cervical Vertebrae A newly recognized right half of the atlas was removed from a piece of matrix. The anatomical relationship with the axis is illustrated in Figure 1. Stirton and Savage [19511 have described and reconstructed a portion of the basicranium from the known right periotic region of the temporal bone, with part of the occipital condyle, and the left periotic region with the left glenoid fossa. We also recovered from the miscellaneous bones embedded in matrix the missing left occipital condyle, the right glenoid fossa, and the basilar portion of the occipital bone. The proportions of the occipital condyles precisely match those of the cranial articular surface of the newly discovered atlas. This precise articulation, together with the fact that no bones of the basicranium (or any other skeletal region) have been duplicated, should lay to rest any lingering doubts a s to whether the craniodental and postcranial remains represent a single individual [see also Meldrum & Fleagle, 1988; Kay, 19901. Stirton and Savage  briefly described but did not figure the axis. The body, odontoid process, portions of the pedicles, and part of the right transverse process are preserved. On the ventral surface of the body, a median ridge can be distinguished, running from the dens caudally. This ridge gives attachment to the longus colli muscle. The ridge widens posteriorly to form a single tubercle, most resembling the condition in Pithecia. I n Chiropotes, Cacajao, as well as Saimiri and Saguinus, the tubercle bifurcates posteriorly. Two shallow fossae are present on either side of the median ridge of the axis in Cebupithecia. They are most similar to the configuration found in Pithecia. A centrally located depression is present on the dorsal surface of the body of 72 I Meldrum and Lemelin A B Fig. 2. A Dorsal view of the lamina of the thoracic vertebrae of Cebupithecia sarrnientoi. The lamina from thoracic vertebra seven is indicated (T7),as is the bifid spinous process of the pre-anticlinal vertebra (arrowhead). B: Lateral view of selected thoracic vertebra. The lumbar-like caudal articular surfaces of the anticlinal vertebra are indicated (arrowhead). Scale bar = 1 cm. the axis of Cebupithecia. An arch of bone that would have spanned the depression craniocaudally is eroded away, revealing a single large nutrient foramen situated in the center of the depression. In Pithecia, a very similar configuration is found. This condition is absent in both Chirpotes and Cacajao. In Saguinus, the depression is very shallow and posteriorly located. Aotus has a depression intermediate in position and depth to the conditions described for Pithecia and Saguinus. Finally, Saimiri is distinctive in that a foramen is present on each side of a low median crest. The transverse processes and the cranial articular surfaces sweep caudally more acutely in P. monachus than in P. pithecia. These features of the axis in Cebupithecia are most similar to the condition in P. monachus. Thoracic Vertebrae Stirton and Savage  reported that the third through seventh thoracic vertebrae are present in the type of Cebupithecia. They noted that these vertebrae were slightly smaller than in Pithecia. However, the body of the "third" thoracic vertebra is very shallow dorsoventrally and is more similar to the first or second thoracic vertebra of Pithecia (Fig. 2B). The seventh thoracic vertebra is repre- Axial Skeleton of Cebupithecia sarmientoi I 73 C D Fig. 3. Dorsal view of the seventh thoracic vertebrae of (A) Cebupithecia sarmtentoi, (B) Pithecia prthecza, (C) = 1 cm. Saimiri c.f sciureus, (D) Saguinus sp. Scale bar sented only by the well-preserved lamina and transverse processes. The overall morphology of this vertebra is virtually identical with Pithecia, and quite distinctive from Saimiri and Saguinus (Fig. 3). The cranial articular processes in Cebupithecia and Pithecia are more oval in outline and project farther cranially. At the cranial edge of the base of the spinous process, a inverted V-shaped depression for the attachment of a well-developed ligamentum flavum is present. In humans, this stout yellow fibroelastic ligament runs from the inner surface of the lamina a t its inferior (caudal) edge to the superior (cranial) edge of the next lower vertebra [Stern, 19881. Its elastic properties limit excessive flexion of the vertebral column. In quadrupedal leapers it is well developed and serves as a n elastic recoil mechanism, tending to extend the flexed vertebral column a t the onset of a leap. 74 / Meldrum and Lemelin TABLE I. Number of Vertebrae in Extant New World Primates and Cebupitheeia sarmientoi* Genus Cebus Saimiri Aotus Callicebus Callimico Saguinus Leontopithecus Callithrix Pithecia Ch iropotes Cacajao Alouatta Lagothrix Ateles Brachyteles C. sarmientoi Thoracic region a 14 (13-15)b 13 (13-14) 14 (13-14) 12 (12-13) 12 12 (11-13) 12 (12-13) 13 (12-13) 12 (12-13) 13 (12-13) 13 (12-13) 14 (13-16) 14 (13-14) 14 (13-15) 14 (13-14) 12 Lumbar region 6 (4-7) 7 (6-7) 7 (7-8) 7 (7-8) 7 7 7 (6-7) 6 (5-7) 7 (6-7) 6 (6-7) 6 (6-7) 5 (5-6) 4 (4-5) 4 (4-5) 4 (4-5) 7 Caudal region 24 (21-26) 28 (25-31) 25 (24-31) 25 (23-26) 28 28 32 (28-34) 27 125-30) 24 (22-25) 24 (23-26) 13 (11-15) 27 (25-28) 26 (24-29) 31 (28-35) 29 ?(>18) "Data from Erikson 119631,Schultz 119611, Schultz and Straus [19451,and personal observations. Cebuella is not included. "The no. of thoracic vertebrae is defined on the basis of a rib-bearing criterion. 'The nos. indicate the average no. of vertebrae and the range observed in each genus. Miscellaneous fragments of spinous processes were identified for as many a s seven additional thoracic vertebrae (Fig. 3A). However, their fragmentary nature prevents a conclusive assessment of their relative positions within the thoracic region. The primitive number of combined thoracic and lumbar vertebrae for primates is most likely 19 [Schultz & Straus, 1945; Schultz, 1961; Table I]. Our examination of over 35 pitheciine skeletons proved this to be the case in all but one instance [see also Fleagle & Meldrum, 19881. This was a specimen of Pithecia pithecia with 13 thoracic and 7 lumbar vertebrae (Hershkovitz  also reports two out of five P. monachus, and one of 11P. pithecia, each with 13 thoracic and 7 lumbar vertebrae). Therefore, since seven lumbar vertebrae have been positively identified, it seems most probable that Cebupithecia had 12 thoracic vertebra, but it possibly had 13. Two of the thoracic vertebral fragments mentioned above can be identified with a greater degree of confidence than the remainder. The first fragment is the lamina of the transitional vertebra, which marks the transition point between the thoracic and lumbar regions. The cranial articular processes are thoracic-like in shape (i.e., the articular processes are oriented in a coronal plane), whereas the caudal articular processes of the same vertebra are lumbar-like in shape (i.e., the articular processes are oriented in the sagittal plane). The base of the spinous process is preserved and is directed cranially (Fig. 2B arrowhead). The second fragment preserves the entire length of a thoracic spinous process. The configuration of the ventral surface indicates that the process was caudally directed. The fragment is distinctive because the tip of the process is bifid, as occurs in the spinous process of the thoracic vertebra immediately cranial to the transitional vertebra (Fig. 2A arrowhead). The bifurcation presumably accommodates the cranially directed spinous process of the transitional vertebra. Determining the position of the anticlinal vertebra in Cebupithecia is incon- Axial Skeleton of Cebupithecia sarmientoi I 75 clusive given the fragmentary nature of the remaining vertebrae. None of the other thoracic fragments appear to have spinous processes that were cranially directed, suggesting that the transitional vertebra was located at the level of T12. This seems unlikely. The transitional vertebra usually encroaches cranially into the thoracic region by two to four segments in living platyrrhines [Erikson, 19631. Pithecia pithecia usually has 12 thoracic vertebrae (rib-bearing) with the transitional vertebra positioned a t the level of TlO. Pithecia monachus usually has 12 or 13with the transitional vertebra at T1O or T11. Both Cacajao and Chiropotes have, on average, 13 thoracic vertebrae with the transitional vertebra at the level of T I 1. Careful removal of the matrix encrusting the sixth thoracic vertebra (in agreement with the assignment by Stirton and Savage ) revealed a bony pillar spanning between the lamina and the body, caudal to the pedicle. This creates a foramen, presumably for the passage of a spinal nerve (Fig. 4A). The presence of such foramina has been reported in the thoracic vertebra of the genera Perodicticus and Arctocebus among the Lorisidae [Mivart, 18651. Mivart [ 18651 also described a smaller foramen piercing the base of the transverse processes craniocaudally. Ankel-Simons [1972, 19831 notes the presence of similar foramina in all four genera of lorisids, and proposes that these features be considered a derived character of the group. We have observed both types of foramina in all five specimens of Pithecia monachus in the collections of the American Museum of Natural History (Fig. 4B). They were not present in any other pitheciine species, or in any of the remaining platyrrhine taxa examined. It appears that these structures arise by the ossification of a ligament. Cases of partial ossification are common, particularly in the more caudal thoracic vertebrae (e.g., Fig. 4a). It is therefore more informative to speak of the partial or complete ossification of the ligament, rather than the foramen it encloses. Since these structures have never been formally named, we propose the following terms. The perpendicular bony span between the body and the lamina is designated the vinculum laminum (Fig. 4C,D). The oblique bony span between the body and the transverse process is designated the vinculum transversum (Fig. 4C,D). The presence of the vincula in Pithecia monachus appears to be associated with the presence of very broad ribs. The ribs are so broad in some individuals of P. monachus that they are nearly in contact with the costal angle. The ribs are also oriented perpendicular to the vertebral column instead of sweeping caudally as is otherwise typical of platyrrhines. The cranial edge of the head and neck of the ribs lie in contact with the vinculum transversum of the preceding vertebra. Therefore, the vincula may serve to stabilize the rib and act as buttresses to support the transverse process and lamina. Jenkins [197O] has proposed that broad ribs in edentates and some primates, such as lorises, are somehow part of a mechanism to generate trunk stability. However, this hypothesis has yet to be substantiated. Since very little is known about the positional repertoire of P. monachus, the behavioral correlates of this peculiar morphology of the thoracic vertebrae and ribs remain unclear. The sixth thoracic vertebra of Cebupithecia has a complete vinculum laminum, but the presence of the much smaller vinculum transversum is uncertain due to the eroded state of the fossil. In the remaining thoracic vertebrae, the vincula lamina a r e broken, but there are the distinct remnants of the initial portions of the bony pillar on the body and the lamina. In addition, the associated rib fragments of Cebupithecia are broad and robust, as is typical for P. monachus. An additional proximal fragment of the first left rib was identified among the miscellany. It is particularly robust and broad, measuring 6.23 mm wide just distal to the costal tubercle. 76 / Meldrum and Lemelin A B C D Fig. 4. A Stereophotographs of a thoracic vertebra of Cebupithecia sarrnzentoi, seen in craniolateral view, illustrating the foraman created by the presence of a bony column caudal to the pedicle. B: Ventral view of thoracic Vertebrae and a rib ofpithecia monachus. Note the incomplete vinculum transversum. Caudal view (C) and lateral view (D) of a schematic of a thoracic vertebrae of P. monachus illustrating the relationships of the vinculum laminum (vl) and the vinculum transversum (vt). Axial Skeleton of Cebupithecia sarmientoi I 77 We have also identified among the miscellany the isolated body of a vertebra, presumed from the lower thoracic region, due to its relatively large size, and because rib facets are visible on the caudolateral surface of each side of the body. The lateral surfaces of the body are marked by two distinct fossae. These suggest the presence of a well-developed quadratus lumborum muscle. The presence of these fossae on the lower thoracic vertebrae is characteristic of leaping platyrrhines, such as Pithecia, but not seen in more deliberate quadrupeds such as Chiropotes or Cacajao. Lumbar Vertebrae Stirton and Savage [ 19511 describe fragments of five lumbar vertebrae. With the assumption that L1 and L3 were missing, they suggest that Cebupithecia had seven lumbar vertebrae as in Pithecia. Two vertebral lamina, found in the miscellaneous material, articulate with two of the described lumbar vertebrae, making them nearly complete. Portions of the bodies of two additional lumbar vertebrae were also found. This has necessitated a slight modification of Stirton and Savage’s original description. This reassessment of the lumbar series is based on careful examination of the relative size and fit of the articular surfaces of the vertebral bodies, and development of the transverse processes. The body and neural arch of the first lumbar vertebra are complete. The base of the broken spinous process is very wide craniocaudally. The articular surfaces of the right and left prezygapophyses are eroded. The second lumbar vertebra is the most completely preserved of the series (Fig. 5) and best illustrates the distinctive morphology of these vertebrae. The prezygapophyses are only slightly eroded. The third lumbar vertebra consists of the caudal third of the body. The fourth lumbar vertebra preserves the body and portions of the neural arch. It lacks the left preand postzygapophyses. The fifth vertebra preserves the cranial half of the body, with the base of the left transverse process and pedicle. Finally, the seventh consists of the caudal half of a body, also with the base of the left transverse process and pedicle. The lumbar vertebrae are quite large and robust relative to the vertebrae of the thoracic region and the ventrolateral surface of each vertebral body of the lumbar region is deeply excavated in Cebupzthecia. This is also characteristic of Pithecia. Two fossae separated by a median sagittal keel mark the origin of the m. psoas major [Stern, 19711. The lumbar transverse processes in the first and second vertebrae are extremely small and non-projecting in Cebupithecia, as in P. monachus in contrast to P. pithecia. In the more caudad lumbar segments the transverse processes appear to be relatively larger and are more laterally directed, as is the case in Pithecia and the other pitheciines. In constrast, the transverse processes in small-bodied cebids and callitrichids are more ventrolaterally directed. The spinous processes in Cebupithecia, best preserved in L2, have a more perpendicular orientation and a blunter profile than in other platyrrhine taxa. The average number of lumbar vertebrae for various species of the Pitheciinae ranges between six, the typical number in both Chirpotes and Cacajao, and seven, the typical number in Pithecia pithecia. P. monuchus has equal numbers with six or seven. The primitive number of lumbar vertebrae for primates is believed to be six [Schultz & Straus, 1945; Table I]. The elongation of the lumbar region is associated with leaping adaptations in primates [Erikson, 1963; Fleagle, 19771. This association has been demonstrated in a contrast of skeletal anatomy of two sympatric pitheciines, P. pithecia and Chiropotes satanas [Fleagle & Meldrum, 19881. This elongation is further indicated by the greater length and robusticity of the lumbar region relative to the thoracic region in Pithecia. Therefore, the ap- A B c Figure 5 Axial Skeleton of Cebupithecia sarmientoi I 79 parent elongation of the lumbar region in Cebupithecia, both by increasing the number of segments and their proportional lengths, suggests leaping behaviors were an important part of this fossil primate’s positional repertoire. This is consistent with morphological features associated with leaping found in the appendicular skeleton of Cebupithecia [Meldrum & Fleagle, 1988; Meldrum & Kay, 19901. Sacral Vertebrae Although no remains of the sacrum have been recovered for this specimen, some inferences can be made based on the morphology of the associated pelvic fragments. The pubis of Cebupithecia is quite long relative to acetabular width (pubis length/acetabular width = 1.6). P. monachus also has a relatively long pubis, in contrast to P. pithecia [Meldrum & Kay, 1990; Meldrum, in prep.]. In a sample of 41 platyrrhines, representing 14 genera, the length of the pubis was highly correlated with the width of the sacrum (r2= .94). Therefore, the relatively long pubis of Cebupithecia implies a proportionately wide sacrum, similar to that of P. monachus. This is also suggested by the long and deep auricular surface preserved on the ilium of Cebupithecia. Hershkovitz  reports that P. monachus typically has three sacral vertebrae, while P. pithecia has two or three in nearly equal frequency. The primitive number of sacral vertebrae for primates is believed to be three [Schultz & Straus, 19451 and the modal number for platyrrhines is clearly three. Therefore, it seems reasonable to conclude that Cebupithecia preserved the primitive platyrrhine condition of three sacral vertebrae, as in P. monachus. Caudal Vertebrae Stirton and Savage [19511 identify 17 caudal vertebrae in their description of the type specimen of Cebupithecia. Presuming that the total number of caudal vertebrae in this fossil was 25 (the hypothetical primitive number), they attribute the position of each vertebra according to its length and diameter. They conclude that caudal vertebrae 1,17,and 20-25 were missing. Also, vertebrae 4 , 5 , 6 and 11 preserve only a portion of their original lengths and are reconstructed. Once again, examination of the miscellaneous material produced an additional proximal caudal vertebra. The proximal region, or initial segment, of the tail of primates is characterized by vertebrae possessing one pair of transverse processes, a vertebral canal, a spinous process, and lumbar-like articular facets [Ankel, 1962, 1963, 1972; German, 19821. These vertebrae are much shorter than the more caudal elements. The newly discovered caudal vertebra has all the features characterizing the initial segment (Fig. 6A,B). It is longer than the vertebra identified as the second caudal element by Stirton and Savage , but is more likely to be the first caudal vertebra based on our comparisons (Fig. 6 0 . The initial segment of the tail of Cebupithecia comprises a total of four caudal elements. In Pithecia and Chiropotes the initial segment of the tail also has four elements. The small-bodied cebids and callitrichids typically have five elements, while the prehensile-tailed monkeys vary from six in Cebus, to eight in the atelines [Ankel, 1962,1963,1972; personal observation]. The first caudal vertebra is nearly Fig. 5. Stereophotographs of the second lumbar vertebra of Cebupithecia sarmientoi seen in (A) lateral, (B) dorsal, and (C) ventral views. Scale bar = 1 cm. A B C Figure 6. Axial Skeleton of Cebupithecia sarmientoi I 81 complete, lacking only the spinous process and the right pre- and postzygapophyses. The transverse processes, although incomplete, were obviously large and oriented perpendicular to the long axis of the body. They bear anterior tubercles. The second caudal vertebra lacks only part of the left prezygapophysis. Its well-developed spinous process is quite perpendicularly directed, most similar to the condition in some of the small-bodied cebids, such a s Cebus or Aotus. The prezygapophyses are rather unusual in their degree of dorsal orientation. The third caudal vertebra lacks the spinous process, left prezygapophysis, and left transverse process. The transverse process is narrower and oriented more caudally. The entire process is situated on the caudal end of the vertebral body. There is a well-developed root of the spinous process. I n overall morphology, this vertebra is most similar to that of Cebus. The degree of development of the haemopophyses and the presence of a ridge on the ventral aspect of the vertebra is also similar to the condition observed in Cebus. The fourth caudal vertebra preserves the body without transverse or spinous process. Two well-defined haemopophyses are present, as well a s transverse ridges that are prolonged distally to the roots of the transverse processes. The neural canal is also well-defined and filled with matrix. This vertebra is most similar to Pithecia. The second more distal segment of caudal vertebrae in primates is characterized by the presence of two pairs of transverse processes, no spinous processes or vertebral canal, and articulations by means of the body only [Ankel, 1962, 1963, 1972; German, 19821. This distal region of the tail in Cebupithecia is represented by 14 caudal vertebrae (Fig. 7). The fifth caudal vertebra preserves the proximal half, with broken transverse processes and eroded prezygapophyses and haemopophyses. Very small and atrophied neural arches, typical of the transitional vertebra (i.e., first vertebra distal to the initial segment), are present on the body. The sixth and seventh vertebrae preserve the distal third of the bodies, with postzygapophyses and posterior transverse processes. The eighth caudal is an almost complete vertebra, lacking only the left anterior transverse process. Caudal nine and ten are complete, with only slightly eroded transverse process on ten. The 11th caudal vertebra lacks its distal half and its right transverse process. Caudal 12 to 18 are complete, with only occasional minor erosion to transverse processes. The longest vertebra of the tail is generally encountered in the second or third element of the distal region of the tail [Ankel, 1972; personal observation]. Beyond this vertebra, the length of each successive vertebra becomes progressively shorter. The absolute lengths of the caudal vertebrae of any primate tail decrease similarly [German, 19821. However, when ratios of median width or proximal width relative to the length of the vertebrae are examined, the change in vertebral dimensions occurs in two distinct patterns, which separate primates with prehensile tails (e.g., Cebus, Alouatta, and Ateles) from those with nonprehensile tails (e.g., Pithecia, Chiropotes, and Presbytis). These differences are most evident in the distal region of the tail. These ratios express the relative breadths of the vertebral body and development of the proximal transverse processes. Lemelin [ 1988, 19891 has shown that the degree of development of these structures is related to the insertion of specific caudal muscles. I n New World monkeys, the proximal transverse process is the combined insertion site for each long tendon of the lateral Fig. 6. A Stereophotographs of the initial segment of the tail of Cebupithecia sarrnientoi seen in dorsal view, Scale bar = 1 cm. B Lateral view of the initial segment. C: Stereophotographs of the first caudal vertebra seen in lateral view. 82 / Meldrum and Lemelin Fig. 7. Dorsal view of the preserved elements of the tail of Cebupithecia sarrnientoi. Scale bar = 1 cm network of flexor caudae lateralis and each bundle of intertransversarii caudae. In prehensile-tailed monkeys, especially atelines, the strong development of intertransversarii caudae in the distal part of the tail, as well a s the medial and lateral flexors, contrasts dramatically with the condition of these muscles in nonprehensile-tailed monkeys. As a result, the degree of development of the transverse processes in the caudal vertebrae from the distal region (expressed as the ratio proximal width/length) is a n excellent indicator of prehensile function in the tails of platyrrhines. This method can be applied to the caudal remains of fossil platyrrhines such as Cebupithecia. Median widthllength and proximal widthllength ratios are computed for extant genera of both prehensile- and non-prehensile-tailed New World monkeys and Cebupithecia. Measurements were not taken on reconstructed or partial vertebrae. In some cases the incomplete preservation of some of the vertebrae made it necessary to take the width of the distal transverse processes instead of the proximal transverse processes for caudal vertebrae 8, 12, and 14. Our observations on numerous primate caudal vertebrae indicate that the distal transverse process width is slightly smaller than the width of the proximal transverse processes. Consequently, the substituted values for the proximal widthllength ratio of vertebrae 8, 12, and 14 should be considered slightly underestimated. Axial Skeleton of Cebupithecia sarmientoi I 83 For the first ratio, middle widthllength, the values for Cebupithecia are very similar to those for Pithecia, and in particular P. monachus (Fig. 8). They are also similar to Chiropotes. The Cebupithecia values lie well below those for the prehensile-tailed species, and above those for Callicebus. Turning to the second ratio, proximal widthllength, the values for Cebupithecia are again particularly similar to P. monachus (Fig. 9). They are also well below the values for prehensile-tailed species, and surprisingly, well below the values for Chiropotes. The relatively high values for Chiropotes may be related to its habit of feeding suspended from its hindlimbs. In this posture the tail is frequently draped over the branch as a supporting brace. Finally, the values for Cebupithecia are well above those for Callicebus and P. pithecia. We conclude therefore, that the Cebupithecia possessed a nonprehensile tail very similar in development and robusticity to P. monachus. Also, the absolute dimensions of the caudal vertebrae more closely approximate those of P. monachus than any other extant platyrrhine. DISCUSSION The study of the axial skeleton of Cebupithecia sarmientoi has revealed a number of characteristics of the vertebrae that this fossil primate has in common with Pithecia, and several that appear to be shared uniquely with Pithecia monachus. These include: 1. Presence of a ventral median ridge on the body of the axis, ending caudally in a single tubercle; 2. A rather deep centrally located depression surrounding a single nutrient foraman on the dorsal surface of the axis; 3. Similarity in the overall morphology of the lamina and transverse processes of the thoracic vertebrae; morphology of the prezygapophyses (i.e., oval and cranially projecting); 4. A vinculum laminum present on the thoracic vertebrae; 5. Broad ribs; 6. Elongation of the lumbar segment (i.e., seven lumbar vertebrae); increased relative size of the individual lumbar segments relative to the thoracic segments; 7. Initial lumbar vertebrae with small non-projecting transverse processes; 8. Ventrolateral surface of the lumbar vertebral bodies with two deep fossae separated by a median keel; transverse processes laterally oriented; 9. Initial segment of the tail consisting of four vertebrae; 10. Similarity in absolute and relative caudal dimensions and development of the transverse processes. The implications of these vertebral traits for elucidating the phyletic relationships of Cebupithecia are difficult to evaluate because so little is known about the evolution of the vertebral column in New World monkeys and primates generally. Therefore, there is very little upon which to base a hypothetical ancestral morphotype for the axial skeleton of platyrrhines or, more specifically, the Pitheciinae. Most recent analyses of dental features indicate that pitheciines are a n early offshoot within the platyrrhine clade, with Chiropotes and Cacajao sharing a more recent common ancestor than Pithecia [Kay, 19901. Kay  also concluded that Cebupithecia is the sister taxon to the pitheciine clade. It has also been suggested that the pitheciines may have evolved from a n ancestor structurally resembling Pithecia [Hershkovitz, 19791. Among the species of Pithecia, i t has been suggested that P. monachus has the most primitive pelage coloration and is believed to occupy the original home range of the genus [Hersh- 84 / Meldrum and Lemelin A A . paniscus (n=2) L a g o t h r i x sp (n=2) 0S O 0.45 0.40 0.3s 0.30 M W / L e R a t i o 0.25 0.20 0.15 0.10 0.05 0.00 1 3 5 7 9 1 1 1 3 1 5 1 7 1 9 2 1 2 3 2 5 2 7 2 9 3 1 V e r t e b r a It 6 0 50 T 0 45 0 40 .. .. 0 35 .* 0 30 M W / L e Ratio 0 25 0 20 .. C satanas (n=2) *. .’ C rnoloch ( n = 2 ) 0 15.. 010.0 05 .* 0004::::: : : : : : : : : : : : : : : : : : : : I C MW /Le Ratio 0.05 0.00 3 1 3 5 7 9 1 1 1 3 1 5 1 7 1 9 2 1 2 3 2 5 V e r t e b r a It Fig. 8. Ratio of middle widtwlength of individual caudal vertebrae plotted for Cebupithecia sarmzentoi and selected extant platyrrhine taxa. Axial Skeleton of Cebupithecia sarmientoi I 85 A 1 .o A . paniscus ( n S ) 0 9 agothrix s p . (n=2) 08 seniculus (n=2) 07 06 PW/Le Ratio 0 5 04 03 02 C. sarrnientoi 0.1 00 1 3 5 7 9 1 1 1 3 1 5 1 7 Vertebra B 0 50 1 9 2 1 2 3 2 5 2 7 2 9 3 1 * - 0 45 * ' 0 40 C satanas (n=2) .. 0 35 .. 0 3 0 .. PW/Le Ratio 0 25 .. C rnoloch (n=2) C sarmientoi 0 20 0 15.. 0 10.. 0 05 -. Vertebra C 0'50 0.45 -. .0 35 0.40 0.30*. PW/Le Ratio 0.25 0.20 0 .. monachus (n=2) .. P pithecia ( ~ 2 ) 1s.. 0 05 O.'O t 0.004: : : : : : : : : : : : : : : : + : 1 3 5 7 9 1 1 1 3 1 5 1 7 1 9 : : : : : 2 1 2 3 I 2 5 Vertebra Fig. 9. Ratio of proximal widthilength of individual caudal vertebrae for Cebupitheciu sarmientoi and selected extant platyrrhine taxa. 86 / Meldrum and Lemelin kovitz, 19871. The present distribution of P. monachus is also geographically nearest to the area where fossils of Cebupithecia are found. Pithecia monachus also has shorter limbs relative to trunk length than P. pithecia [Hershkovitz, 19871 and in P. monachus a number of features of the appendicular skeleton associated with clinging and leaping are less pronounced than in P. pithecia (Meldrum, unpubl. data). These skeletal differences would seem to suggest that P. monachus is less specialized for leaping than P. pithecia, which agrees with the limited field observations of the ecology of this species [Happel, 19821. Therefore, many of the similarities between Cebupithecia and P. monachus are most likely shared primitive retentions of the last common ancestor of the Cebupithecia-pitheciine clade. More problematic is the distribution of particular features, such as the vincula lamina and transversa in the thoracic vertebrae. Superficially, this character would appear to unite Cebupithecia and P. monachus, to the exclusion of P. pithecia. Thus, if we assume this is a shared derived character uniquely linking Cebupithecia with P. monachus, we are forced to conclude that some species of Pithecia are more closely related to Chiropotes and Cacajao than to the other species of Pithecia. This is certainly not the case. Alternatively, this character could be primitive for the pitheciine clade (including Cebupithecia). In this case, the character has been lost in Chiropotes and Cacajao, and independently in some Pithecia. Finally, this character may simply be a result of homoplasy between Cebupithecia and P. monachus. Given the likelihood that the vincula result from the ossification of existing ligaments, i t is conceivable, even likely, that they could easily arise in parallel. Until the distribution and function of this character are more fully understood, we feel it is impossible to determine which of the latter two hypotheses is more probable, for this particular feature. The same line of reasoning can be followed for most of the other features listed at the beginning of this discussion. The common features of the axis, the particulars of the thoracic vertebral lamina and transverse processes, and the similarities in the absolute and proportional dimensions of the tail may be primitive for pitheciines generally. Elongation of the lumbar region and deep ventral fossae may be primitive, or may have arisen in parallel a s common adaptations to leaping behavior. From analysis of the postcranium, it is evident that leaping was a significant component of the positional repertoire of Cebupithecia. This has been demonstrated conclusively by analyses of the appendicular skeleton, and now has received further corroboration from the analysis of the axial skeleton. However, some of the skeletal correlates of leaping behavior in Cebupithecia appear to be convergent, rather than homologous with those of Pithecia [Meldrum & Fleagle, 1988; Davis, 1988; Ford, 1990; Meldrum & Kay, 19901. Many aspects of the pitheciine postcranium are convergent on a n ateline-like condition. These are most marked in the hindlimb, such as the relatively wide and shallow femoral condyles and broad patellar groove, and relatively low broad talar trochlea, characteristic of the extant pitheciines and atelines. These are generally the very features in which Cebupithecia differs from the extant pitheciines. For example, the distal femur of Cebupithecia is relatively narrow and deep and more similar to the condition in Callicebus, Aotus, or Saguinus, than i t is to Pithecia [Ciochon & Corruccini, 1975; Davis, 1988; Meldrum & Kay, 19901. It seems clear that the femoral and talar morphology of Cebupithecia, as well a s other selected features of the postcranium, more closely approach the ancestral condition for platyrrhines and thus provide a reasonable model for a basal pitheciine morphotype. Therefore, the specializations for leaping in the genus Pithecia might be interpreted as primitive retentions, modified by the derivation of ateline-like morphologies, or as parallelisms, inde- Axial Skeleton of Cebupithecia sarmientoi 1 87 pendently derived in Cebupithecia and Pithecia (especially P. pithecia) from a less saltatory last common ancestor of the Cebupithecia-pitheciine clade. This issue will be examined in more detail elsewhere [Meldrum, in prep.]. CONCLUSIONS 1. A re-examination of the holotype of Cebupithecia sarmientoi has produced a number of additional, previously undescribed, elements of the axial skeleton. Among these were additional fragments of the basicranium and the atlas, or first cervical vertebra, which unquestionably establish the unity of craniodental and postcranial remains of this fossil specimen. 2. Numerous characters of the vertebrae, including aspects of the cervical, thoracic, lumbar, and caudal regions were identified that bear the greatest similarity to the extant pitheciine genus Pithecia. Some of these particularly resemble Pithecia monachus. Of particular interest is the presence of a bony pillar (vinculum laminum) spanning between the thoracic vertebral bodies and the laminae of Cebupithecia. This feature appears to be uniquely present in P. monachus among the extant platyrrhines. Due to our limited knowledge of the evolution of platyrrhine vertebral morphology, the phylogenetic implications of these characters are uncertain. It has been suggested that P. monachus most nearly approximates the ancestral pitheciine morphotype. In that case, some of the similarities between Cebupithecia and P. monachus may represent primitive retentions from the last commmon ancestor of the Cebupithecia-pitheciine clade. Alternately, others may represent homoplasies between Cebupithecia and P. monachus. 3. The lumbar region (non-ribbearing vertebrae) of Cebupithecia consists of seven vertebral elements. The absolute and relative proportions of the lumbar vertebrae relative to the thoracic region indicate leaping behavior. This condition is likely derived in parallel in Cebupithecia and Pithecia (especially P. pithecia) from a less saltatory last common ancestor of the Cebupithecia-pitheciine clade. 4. While some of the individual elements of the initial segment bear some affinities to those of Cebus, the tail of Cebupithecia was nonprehensile, and in absolute dimensions and relative proportions was most like P. monachus. ACKNOWLEDGMENTS We are grateful to Donald Savage of the Museum of Paleontology, University of California at Berkeley, for allowing us to study the holotype of Cebupithecia sarmientoz (UCMP 38762); to Guy Musser and Wolfgang Fuchs of the American Museum of Natural History; and to John G. Fleagle, Richard F. Kay, David W. Krause, and Liza Shapiro for comments and discussion on earlier drafts of this manuscript. This research was supported by a grant from the National Science Foundation (BNS 8719448) to D.J.M. and by a predoctoral fellowship from the Natural Sciences & Engineering Research Council of Canada and a grant from the Doctoral Program in Anthropological Sciences, SUNY a t Stony Brook, to P.L. REFERENCES Ankel, F. Vergleichende Untersuchungen iiber die Skelettmorphologie des Greifschwanzes siidamericanischer Affen. ZEITSCHRIFT FUR MORPHOLOGIE UND OKOLOGIE DER TIERE 5131170, 1962. Ankel, F. Zur Morphologie des Greifschwanzes bei siidamerikanischen affen. ZEITSCHRIFT FUR MORPHOLOGIE UND ANTHROPOLOGIE 53:12-18, 1963. Ankel, F. Vertebral morphology of fossil and extant primates. Pp. 223-240 in THE FUNCTIONAL AND EVOLUTIONARY BIOLOGY OF THE PRIMATES. R.H. Tuttle, ed. Chicago, Aldine Press, 1972. Ankel-Simons. F. A SURVEY OF LIVING 88 / Meldrum and Lemelin PRIMATES AND THEIR ANATOMY. New York, Macmillan, 1983. Ciochon, R.L.; Corruccini, R.S. Morphometric analysis of platyrrhine femora with taxonomic implications and notes on two fossil forms. JOURNAL OF HUMAN EVOLUTION 4:193-217, 1975. Davis, L. Morphological evidence of locomotor behavior in a fossil platyrrhine. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 75:202, 1988. Delson, E.; Rosenberger, A.L. Are there any anthropoid primate living fossils? Pp. 5061 in LIVING FOSSILS. N. Eldredge, S.M. Stanley, eds. New York, Springer-Verlag, 1984. Erikson, G.E. Brachiation in New World monkeys and in anthropoid apes. SYMPOSIUM OF THE ZOOLOGICAL SOCIETY OF LONDON 10:135-163, 1963. Fleagle, J.G. Locomotor behavior and muscular anatomy of sympatric Malaysian leaf-monkeys (Presbytis obscura and Presbytis melalophos). AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 46: 297-308, 1977. Fleagle, J.G.; Meldrum, D.J. Locomotor behavior and skeletal morphology of two sympatric pitheciine monkeys, Pithecia pithecia and Chiropotes satanus. AMERICAN JOURNAL OF PRIMATOLOGY 16: 1-23,1988. Ford, S.M. Systematics of the New World monkeys. Pp. 73-135 in COMPARATIVE PRIMATE BIOLOGY, VOLUME 1: SYSTEMATICS, EVOLUTION, AND ANATOMY. D.R. Swindler, J . Erwin, eds., New York, Alan R. Liss Inc., 1986. Ford, S.M. Locomotor adaptations of fossil platyrrhines. JOURNAL OF HUMAN EVOLUTION 19:141-173,1990. German, R.Z. The functional morphology of caudal vertebrae in New World monkeys. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 58:453 -459, 1982. Happel, R.E. Ecology of Pithecia hzrsuta in Peru. JOURNAL OF HUMAN EVOLUTION 11:581-590, 1982. Hershkovitz, P. The species of sakis, genus Pithecia (Cebidae, Primates) with notes on sexual dichromatism. FOLIA PRIMATOLOGICA 31:l-22, 1979. Hershkovitz, P. The taxonomy of South American sakis, genus Pithecia (Cebidae, Platyrrhinin): a preliminary report and critical review with the description of a new species and new subspecies. AMERICAN JOURNAL OF PRIMATOLOGY 12: 387-468, 1987. Jenkins, F.A. Anatomy and function of expanded ribs in certain edentates and primates. JOURNAL OF MAMMALOGY 51: 288-301, 1970. Kay, R.F. The phyletic relationships of extant and fossil Pitheciinae (Platyrrhini, Anthropoidea). JOURNAL OF HUMAN EVOLUTION 19:175-208, 1990. Lemelin, P. Myologie functionnelle de la queue prehensile des singes du Nouveau Monde (Platyrrhini, Primates). Unpublished M.Sc. Thesis, Universite de Montreal, 1988. Lemelin, P. Functional myology of the platyrrhine prehensile tail. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 78:260, 1989. Lemelin, P.; Meldrum, D.J. Vertebral morphology of Cebupithecia sarmientoi and its implications for pitheciine relationships. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 81:257,1990. Meldrum, D.J. Appendicular Skeleton of Cebupitheciu sarmientoi (Pitheciinae, Platyrrhini) from the Middle Miocene of Colombia. In prep. Meldrum, D.J.; Fleagle, J.G. Morphological affinities of the postcranial skeleton of Cebupithecia sarmientoi. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 75:249, 1988. Meldrum, D.J.; Kay, R.F. A new partial skeleton of Cebupithecia sarmientoi from the Miocene of Colombia. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 81:267, 1990. Meldrum, D.J.; Fleagle, J.G.; Kay, R.F. Partial humeri of two Miocene Colombian primates. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 81:413422, 1990. Mivart, G. Contributions toward a more complete knowledge of the axial skeleton in primates. PROCEEDINGS OF THE ZOOLOGICAL SOCIETY OF LONDON 545-592, 1865. Orlosky, F.J. Dental evolutionary trends of relevance to the origin and dispersion of the platyrrhine primates. Pp. 189-200 in EVOLUTIONARY BIOLOGY OF THE NEW WORLD MONKEYS AND CONTINENTAL DRIFT. R.L. Ciochon, A.B. Chiarelli, eds., New York, Plenum Press, 1980. Rosenberger, A.L. Phylogeny, evolution and classification of New World monkeys (Platyrrhini, Primates), Unpublished Ph.D. Dissertation, CUNY, New York, 1979. Schultz, A.H. Vertebral column and thorax. Pp. 1-66 in PRIMATOLOGIA VOL. 4. A.H. Schultz; D. Starck, eds. Basel, Karger, 1961. Schultz, A.H.; Straus, W.L. The number of vertebrae in primates. PROCEEDINGS OF THE AMERICAN PHILOSOPHICAL SOCIETY 89:601-626, 1945. Axial Skeleton of Cebupithecia sarmientoi I 89 Stern, J.T., J r . Functional myology of the hip and thigh of the cebid monkeys and its implications for the evolution of erect posture. BIBLIOTHECA PRIMATOLOGICA 14:l-318, 1971. Stern, J.T., J r . ESSENTIALS OF GROSS ANATOMY. Philadelphia, F.A. Davis, 1988. Stirton, R.A. Ceboid monkeys from the Miocene of Colombia. BULLETIN UNIVER- SITY OF CALIFORNIA PUBLICATIONS IN GEOLOGICAL SCIENCES 28:315356,1951. Stirton, R.A.; Savage, D.E. A new monkey from the La Venta Miocene of Colombia. COMPILACION DE LOS ESTUDIOS GEOL. OFICIALES EN COLOMBIA, SERVICIO GEOL6GICO NACIONAL, BOGOTA 7:345-356,1951.