Corpus and Ramus

AMNHM 268001 (figs. 3B, 4B, 5B) consists of the left ascending ramus and body as well as both sides of the symphyseal region. The entire inferior edge of the mandible has spalled off except for a small area close to the symphysis. Although the posterior border of the ramus is missing from condyle to angle, as a whole this part of the mandible is better preserved than it is in the type specimen (figs. 3A, 4A, 5A). The lateral surface of the ramus bears a shallow triangular fossa, presumably for the insertion of part of the superficial masseter m. (area not represented on type jaw). This fossa is continued upward onto the broken root of the coronoid process, which, judging from what remains, must have been quite large. Whether the coronoid ended as a spike or was strongly recurved cannot be determined from the material available, and the reconstruction (fig. 6B) is noncommittal. On the medial side of the ramus, behind the last molar, is a large and very excavated mandibular foramen (not “mylohyoid foramen” of Rosenberger [1977: 470], a lapsus). Above the foramen a thick pillar of bone runs from the body of the mandible backward and slightly upward. In the intact state this pillar would have terminated as the neck of the condyle.

Remarkably, a tiny piece of cortical bone defining the inferior margin of the mandibular notch is still preserved (double asterisk, fig. 4B). This is of some interest because it gives a qualitative idea of the length of the mandibular load arm, assuming that the mandibular condyle of Xenothrix resembled that of other platyrrhines (i.e., in norma lateralis, superior surface of condyle situated only slightly higher than inferiormost point on mandibular notch). If this inference is correct, the effective length of the ramus (occlusal plane to condyle) in Xenothrix would have been rather short, as in larger-bodied pitheciids (fig. 8B–D).

As is typical for platyrrhines other than Paralouatta, the symphyseal region slopes sharply away beneath the incisor sockets and there is no mental eminence. By contrast, on the oral side there is a strong retromental buttress, underneath which are two deep pits for genioglossus/geniohyoid mm. As in the type, the mental foramen is located at the level of pm3. Because the canine was unquestionably small (see Dentition), the apical end of its alveolus does not swell the symphyseal planum laterally nor interrupt the smooth line of the inferior margin of the horizontal ramus, as it does in all extant noncallitrichines except Aotus and Callicebus (which likewise have small canines).

The second mandible from Skeleton Cave, AMNH 268004 (figs. 3C, 4C, 5C), consists of a small section of the corpus preserving both molars. Although the specimen as a whole is severely damaged, the inferior edge of the mandible beneath the cheekteeth is well preserved, giving a much firmer idea of the contour of this border than previously known. This specimen shows that the inferior border of the mandible in Xenothrix was conspicuously inflected medially, as in many other platyrrhines (including Callicebus and Pithecia). Large ridges define the limit of an extensive, multicompartmental fossa in the gonial region for the insertion of fascicles of the medial pterygoid m. (area not represented on type jaw because of breakage).

Although it is clear that the jaw of Xenothrix exhibited a deep gonial region, very little of this area remains on any one specimen, making it difficult to quantify the amount of gonial flare. Rosenberger (1977: 468, fig. 3) attempted to design a mandibular profile index, but noted that because the inferior border of the Xenothrix type specimen was completely broken away, only “a minimal estimate” could be made. Thanks to the recent discovery of a small piece of the type's inferior border (single asterisk, fig. 4A), found during recuration of H.E. Anthony's 1920 faunal collection from Long Mile Cave, it is now possible to gain a better (if still imperfect) idea of gonial flare in the Jamaican monkey. Although the area of the jaw's maximum flare is probably not represented, enough remains to allow the taking of the two measurements (symphyseal height and modified gonial height) presented in table 2. These measurements are not identical to Rosenberger's (1977), as we had difficulty in ascertaining the position of “minimum depth” according to his criteria. Nevertheless, except in one crucial area our results are in broad agreement. According to Rosenberger's (1977) measurements, Xenothrix would have expressed a degree of posterior deepening comparable to, if not more extreme than, that of Callicebus and Lagothrix. By contrast, according to our results, Xenothrix places closer to midsized pitheciids than to Callicebus for this feature. From this we conclude that Xenothrix possessed posterior deepening, but not in the exaggerated form seen in the titi monkey or woolly spider monkey (or, for that matter, Aotus). Some published photographs of the type jaw (e.g., Rosenberger, 1977: pl. 2A) make it seem as though Xenothrix possessed a greatly deepened gonial region, but this is due to misorientation of the specimen (symphyseal end raised excessively, thereby overemphasizing the degree of posterior deepening).

Dentition

The teeth remaining in the new mandibles are not in good condition, most having lost significant sections of their crowns (figs. 3– 5). They are also extensively worn, both occlusally and interproximally, indicating that the new jaws came from animals that were ontogenetically somewhat older than the one represented by the holotype (the molars of which show virtually no wear). Although dental measurements for these specimens are provided in tables 3 and 4, most should be regarded as minimum estimates due to breakage and loss of material.

AMNHM 268001 preserves on the left side parts of pm3–m2 and sockets for i1– pm2 (figs. 3B, 4B, 5B). On the right side, sockets (but no teeth) are partially preserved for i1–pm1. AMNHM 268004 retains only the two left molars and the distal wall of the left pm4 socket (figs. 3C, 4C, 5C). A noticeable gap, not present in the holotype, exists between the distal margin of the m2 and the edge of the ascending ramus in both AMNHM 268001 and 268004. To confirm that this gap is not due to the presence of an unerupted m3, AMNHM 268001 was radiographed. No evidence of a crypt was found (cf. Williams and Koopman, 1952), and we conclude that the difference between jaws in this respect is probably due to intraspecific variation or is a function of age.

Despite damage to the molars, it is obvious that their occlusal surfaces were comparatively large. The few linear measurements that can be taken indicate that the occlusal area (BL × MD) of each molar is more than two times the size of individual m1s and m2s of Callicebus (table 5). Furthermore, summed m1–m2 occlusal area (based on mean dimensions) in Xenothrix is 54.2 mm², which is effectively identical to the average for the entire molar row in a sample (N = 3) of Cacajao (52.7 mm²; see also table 5). This comparison is of some interest, inasmuch as Xenothrix was probably similar to uakaris in body size (see Discussion). This finding supports Rosenberger's (1977: 474) contention that “in Xenothrix the total relative length of the molars appears to have remained stable or to have increased [compared to the primitive condition], the absolute number of components notwithstanding.”

The molars of Xenothrix mcgregori and Cebus apella are also similar in size (cf. Swindler, 2002). However, despite evident bunodonty in Xenothrix, its crown enamel is not relatively thickened (L. Martin, personal commun.), in contrast to C. apella in which extremely thick enamel apparently facilitates the mastication of palm nuts and other very hard foods. It is tempting to speculate that Xenothrix may have been chiefly a “seed predator” in the manner of the large living pitheciines (Kinzey, 1992), although we cannot offer any morphological support for this argument at present. Even within Pitheciinae, a morphologically homogeneous group, there is some dietary plasticity with regard to the intake of ripe fruits, leaves, and insects, which should make one cautious about interpreting the fossil monkey's preferences too narrowly.

The middle and posterior premolar crowns are also large and stoutly built. As Rosenberger (1977) noted, BL widths of the molars of Xenothrix are broad compared to those of extant platyrrhines. This also applies to the premolars: for example, the broken left pm4 of AMNHM 268001 has a BL width of ∼5.2 mm, as compared to a mean of 4.4 mm for the homologous tooth in the Cacajao sample (tables 4, 5). The condition of the anterior premolar has to be inferred from alveolar measurements (table 6), as the tooth is not preserved in any specimen. Interestingly, the right pm2 alveolus in AMNHM 268001 seems to be unusually shallow (?anomalous retention of deciduous tooth), although so little is left of this feature that it is hard to be certain (fig. 4B). In any case, no crypt for an unerupted successor is in evidence. The left pm2 alveolus appears normal (fig. 5B). In larger pitheciines the equivalent of the curve of Spee is quite pronounced, the premolars showing a steady forward increase in crown height (fig. 7C, D). In Callicebus and probably also Xenothrix the curve is flatter, in keeping with the imputed small size of the canine (cf. figs. 6B and 7B).

Although the condition of the c1 in Xenothrix is not known, because there are no osteological indications of diastemata in the available jaws we agree with Rosenberger that it seems unlikely that the c1 crown could have flared either mesially or distally to any important degree. Our reconstruction (fig. 6B) suggests that c1 may have been more bulbous than the teeth on either side of it, but it probably did not project above them to any significant extent.

None of the mandibular incisors is preserved, but the close packing and parallel walls of their highly compressed alveoli suggest relatively narrow crowns (see Discussion). In the available specimens, incisor alveoli are too incomplete to determine whether the roots of the centrals were larger than those of the laterals. In Pitheciidae generally (including Callicebus), lateral incisor crowns are appreciably larger than those of the centrals, but roots are only slightly larger. In Aotus, laterals are subequal to very slightly larger than centrals for both features. Whether lower incisor crowns were relatively “heightened” (a diagnostic feature of Pitheciinae originally noted by Kinzey [1992]) cannot be decided in the absence of the teeth themselves.

Facial Skeleton and Orbits

As preserved, AMNHM 268006 is reasonably intact (figs. 9, 10; table 6): the alveolar process of the maxilla and palate are complete except for minor abrasions, and substantial portions of the sphenoid, palatines, vomer, lacrimal, and interorbital septum remain. Other parts of the face, including the upper orbits and most of the orbital floors, zygomatics, and upper nasal cavity were evidently broken away some time ago, as no fresh breaks can be seen in the relevant areas.

Morphological evaluations of the midfacial, nasal, and palatal regions are provided in subsequent sections of this paper. It is unfortunate that so little is left of the orbital skeleton in AMNHM 268006, since the plausibility of Rosenberger's Aotus hypothesis largely turns on his interpretation of its morphology. Features connected with the eye and orbit of Xenothrix are reserved for separate treatment (see Discussion).

Dentition

As in the case of the mandibles, the cranial fossils preserve only molars and some premolars (figs. 9, 11). However, unlike the teeth in the jaws, the maxillary cheektooth crowns are beautifully preserved except in one instance (right M1 of AMNHM 268006). Alveoli of missing teeth are also largely intact in AMNHM 268006, providing some idea of the relative sizes of the roots of the anterior teeth.

Like the mandibular cheekteeth, the maxillary molars and premolars of Xenothrix are large and ovoid to quadrangular in outline. Principal cusps are swollen, with puffed-out shoulders that obscure the presence of cristae. Buccal cingula are absent, and lingual cingula are inconspicuous or absent. In size and to some degree in shape the upper cheekteeth of Xenothrix strongly resemble those of larger pitheciines (Chiropotes and Cacajao) rather than Callicebus. Indeed, in most metrical regards the cheekteeth of Xenothrix can be described as superficially uakarilike (tables 4, 5), although of course there are differences in detail. In comparison to Cacajao (fig. 12D) the following features of Xenothrix are noteworthy: (1) Principal cusps and general crown outlines are puffier in Xenothrix. (2) M1 of Xenothrix is very large and expresses unusual development of buccal moiety. (3) M2 is quite reduced, like M3 in living pitheciids (Kinzey, 1992). (4) Premolars are very large and similar in BL and MD dimensions to those of Cacajao (tables 4, 5). (5) PM4 is premolariform as in Callicebus, not distinctly molariform as in pitheciines. (6) Roots of cheekteeth are strongly divaricating, as in Cacajao and to a lesser degree in other pitheciids. (7) Arrangement of principal cusps and other small-scale features is very similar, except that whereas the cheekteeth of Chiropotes and Cacajao wear into what may be described as an aligned series of troughs, the main cusps and principal blades of Xenothrix tend to retain more definition, as in Callicebus (fig. 12B) or Pithecia (fig. 12C). (8) Xenothrix is distinctively different in lacking the intense pitting, crenulation, and slight polycuspidation which uniquely distinguish cheektooth occlusal surfaces of living pitheciines. In Callicebus crenulation is much weaker than in pitheciines. Regarding this last feature, Rosenberger (1977) came to a different conclusion, arguing that there was some evidence of “enamel papillation” on lower molars of Xenothrix. Slight crevices and ripples can be detected under magnification, but they are no more frequent or obvious than in many nonpitheciids. Similar remarks apply to the upper molars of Xenothrix.

There are four cusps present on the M1 of Xenothrix; in our examples, the protocone and hypocone are less well defined than the paracone and metacone. The trigon and talon are indistinctly separated by a vaguely defined postprotocrista running from metacone to protocone. The hypocone is worn flat in both specimens and is associated buccodistally with a small fossette (fig. 11C). The result of this configuration is a large distal shelf or wear surface, as seen in pitheciids generally (especially in larger species, cf. fig. 12C, D). The paracone bulges outward (buccally) to a greater extent than does the metacone (figs. 9A, 11C), with the result that the tooth is markedly wider proximally than distally. This feature is present but less marked in Paralouatta and Antillothrix (cf. illustrations in Horovitz and MacPhee [1999] and MacPhee and Horovitz [2002]). In other pitheciids the two principal buccal cusps are mesiodistally aligned and similar in size (fig. 12B–D).

M2 is reduced distally, with the hypocone and metacone being barely distinguishable as separate cusps. The paracone is the best defined feature on the tooth, and like the M1 protocone, presents a prominent buccal bulge. The protocone area is slightly raised and worn flat. Each molar has three roots, two buccal and one lingual (fig. 11A, B).

PM3 and PM4 are bicuspid and almost identical morphologically, although PM3 is slightly smaller overall. In both specimens (figs. 9A, C; 11B, C), paracone and protocone are heavily worn; with difficulty mesial and distal parastyles can be discriminated on either side of the paracone. PM3 and PM4 possess two roots; PM2, represented by its alveolus only, seems to have had only one root.

Although canine crowns of Xenothrix remain unknown, root sizes can be accurately gauged from alveolar dimensions. As may be seen in table 6, BL and MD measurements of alveolar orifices of the pm2 and PM2 (= maximum inside dimensions of opening) are similar to those of the canine alveoli in the same jaws. (This character is to be distinguished from ch. 37, appendix 1, which compares the size of the fourth premolar's alveolar orifice to that of the canine.)

Greater discrimination can be gained by multiplying BL and MD measurements by AD (root length) in order to generate a volume modulus as a proxy for PM2/pm2 and C1/c1 root size. As may be seen in table 6, at least for the two specimens of Xenothrix that could be measured adequately, the root volume modulus for C1 is about 30% larger than that for c1. This is to be expected, as the usual condition in platyrrhines is one in which the maxillary canine is appreciably larger than the mandibular. However, the size relationship of the premolars is the reverse: pm2 root volume is about 20% larger than that of PM2. The same situation obtains in the comparative set, indicating that this relationship of premolar to root size is probably primitive.

Alveolar data cannot be used directly to infer whether the crowns of the canines projected past the occlusal level of adjacent teeth in Xenothrix, although on the basis of conditions in platyrrhines generally and in pitheciids specifically we consider this probable (figs. 6, 8, 14). At the same time, the degree of projection was probably quite small, along the lines seen in extant Callicebus and Aotus rather than Pithecia or any of the larger pitheciines (or other ceboids and ateloids generally). Inspection of a range of skulls of extant pitheciines lacking anterior teeth indicates that the C1 socket is typically about two to three times deeper than that of PM2. By contrast, if AMNHM 268006 is representative, in Xenothrix (table 6) the upper canine socket is only about 60% longer than that of the anterior premolar. In AMNHM 268001, c1 and pm2 roots are subequal.

Although within the comparative set only Aotus (fig. 13A) has truly spatulate upper incisor crowns, in owl monkeys as well as pitheciids I1 crowns are larger than those of the I2s (fig. 13B–D). In AMNHM 268006 there is no question that I1 roots were more substantial than those of I2s, but there is no independent evidence that the crowns were spatulate (see Discussion).

Living ceboids (which include Aotus in our view) differ from ateloids in that the former rely to a much greater degree on insects (Horovitz and Meyer, 1997). The known morphology of Xenothrix is consistent with the view that this monkey was mainly herbivorous, like the other members of its clade. In this connection it is interesting that Xenothrix evidently lacked substantial midfacial prognathism, large canines/diastemata, and highly procumbent lower incisors—all features of larger extant Pitheciinae. Without these specializations, the feeding maneuvers of Xenothrix doubtless differed from those of pitheciine “sclerocarp foragers” as described in detail by Kinzey (1992) and Rosenberger (1992). In sum, the dental evidence suggests that Xenothrix was a rather unspecialized frugivore, probably not unlike its closest living relative, Callicebus.

Construction of the Inferior Orbital Fissure

As a first approximation it could be argued that a widely dehiscent IOF is characteristic of primates with the largest eyes, but within- group variation in strepsirhines is actually rather inconsistent for this feature (Cartmill, 1980). Among extant haplorhines, Tarsius and (to a lesser extent) Aotus have large IOFs, which of course leads to the supposition that this feature may in fact be correlated with ocular hypertrophy in this group.

The IOF is usually treated as a single anatomical entity, but because it is a gap rather than a structure this approach may obscure more than it reveals. For morphological purposes we find it convenient to divide the inferior orbital fissure into anterolateral (AIOF) and posteromedial (PIOF) sections (figs. 15, 16). In platyrrhines the AIOF is defined by the orbital surfaces of the maxilla and zygomatic which grow together in various complex ways, ranging from widely open (as in Aotus; figs. 15A, 16A; table 7) to almost completely obliterated due to the approximation of bone territories (as in most Callicebus; figs. 15B, 16B; table 7).

In pitheciids generally (fig. 15B–D), zygomatic/maxillary contact tends to be broad: these two bone territories often grow together in such a way that the anteriormost part of the AIOF is pinched off as a separate foramen or slit, here identified as foramen innominatum.6 In Aotus, the orbital surfaces of the zygomatic and maxillary are oriented differently than in pitheciids because of the laterad bulging of the orbits (figs. 13A, 15A). There is a significant gap between these bone territories not only medially but also laterally, where the lateral part of the AIOF yawns especially widely because the zygomatic's orbital surface is pushed out, as it were, onto the root of the zygomatic process. There is no question of a narrowly confined foramen innominatum being formed under these circumstances.

The PIOF is ventrally defined by the orbital surface of the maxilla and the terminal part of the maxillary tuberosity and dorsally by the lower margins of the the sphenoid wings, where the gap grades insensibly into the superior orbital fissure. In contrast to the AIOF, in most platyrrhines the PIOF tends to remain widely open, and opposing bone territories either never grow together or do so only locally. For example, in large-bodied pitheciines a small tongue of bone, usually derived from the zygomatic or sphenoid, is often seen to roof over the proximal part of the channel for the infraorbital neurovascular bundle in carefully cleaned specimens (asterisk, fig. 15D). In other specimens the tongue seems not to have formed or to have been broken off during preparation.

In a general way IOF size in adult pitheciids seems to be correlated with body size: Callicebus has relatively and absolutely the narrowest IOF, while Pithecia, Chiropotes, and Cacajao display progressively wider gaps (ignoring the tongue just described). This correlation obviously does not hold for Aotus (fig. 15A): despite this monkey's comparatively small body size, both AIOF and PIOF are very wide and the contribution of the zygomatic to the lateral wall of the orbit is relatively slim. This suggests that conditions in the owl monkey are directed by factors other than those controlling degree of closure in pitheciids.

Turning now to the partial skull AMNHM 268006, it is evident that, while no detailed reconstruction of the IOF is possible, it is clear nevertheless that conditions are not at all Aotus-like. First, although most of the orbital wing of the zygomatic is not represented, the remaining part is in sutural union with the maxilla in the orbital floor. Zygomaticomaxillary contact does not occur in the floors of extant Aotus in the region of the IOF. Secondly, in the fossil the trailing margin of the zygomatic wing defines a notch (figs. 9D, 16E), which by virtue of its position must be the anterior end of the AIOF, in agreement with conditions not only in pitheciids like Cacajao (fig. 16D) but also in all other platyrrhines we have examined. The importance of this observation may be appreciated by imagining what the orbital skeleton of an owl monkey would look like if the zygomatic wing were broken away along the margin of the AIOF. Since there is no contact with the maxilla in this region, it follows that no part of the zygomatic would be represented in the orbital floor except anteriorly, where by necessity the zygomatic meets the maxilla to complete the postorbital plate and orbital sill. Since zygomatic material is unquestionably present in the fossil, pressed up against the lateral part of the maxillary tuberosity, this may be considered unequivocal evidence that the AIOF could not have been widely dehiscent as in modern Aotus.

A minor point that cannot be elucidated further without more complete remains is whether in Xenothrix the orbital wing of the zygomatic was typically reflected back onto the maxillary tuberosity, thereby creating a foramen innominatum. What is left of the orbital skeleton of Xenothrix does not preclude the possibility that the foramen was present, but nothing makes it particularly likely (or any more likely than the moderately open fissure seen in callitrichines, cebines, and large- bodied pitheciids). More positively, the morphology seen in Alouatta, Brachyteles, Lagothrix, Ateles, and Paralouatta, in which the anterior end of the AIOF is reduced to a very narrow slit, cannot have been present in Xenothrix.

In sum, the only thing that can be conclusively inferred about the AIOF in Xenothrix is that it must have been much narrower and differently shaped than in Aotus. Further, IOF size and shape are not necessarily coupled with orbit size: in Paralouatta, where orbit size is comparable to that of Aotus, the IOF nevertheless displays the narrow condition. We conclude from these considerations that Rosenberger's (2002) view that Xenothrix and Aotus are similar in IOF construction garners no support from comparative anatomy.

Estimating the Size of the Bony Orbit

Rosenberger (2002) asserted that enough remains of the orbital rim in AMNHM 268006 to permit a reasonable estimation of eyeball size, although he did not attempt to quantify this. In principle an empirically meaningful estimate of eyeball size in a fossil primate can be achieved by using a substitute, such as an appropriately shaped sphere of modeling clay. A model that is just large enough to pass through the orbital opening should be approximately equivalent to the size of the original eyeball (cf. MacPhee, 1987), although this makes no allowance for the intrinsic orbital muscles, fat, and other tissues that also fill the bony orbit (Schultz, 1940). Experience shows, however, that this method works well only if the bony orbit is complete or nearly so. If too little of the orbital rim and walls remain, then the diameter of the model cannot be adequately constrained and an erroneous (or, at any rate, an insecure) estimate of eyeball size may result. This is the case with AMNHM 268006: primate orbital rims are rarely perfectly circular, and a plausible arc of curvature cannot be derived from the few millimeters of rim that are still preserved on this specimen. In fact, the medioventral section of the orbital rim that is preserved in Xenothrix is typically almost straight in many platyrrhines (e.g., Cebus, Callicebus, Saimiri, Pithecia) and therefore independent of orbit size (cf. fig. 13B).

Utilization of Kay and Kirk's (2000) osteological method to derive information concerning activity pattern and visual acuity in extinct primates is also precluded, because none of the required variables (skull length, orbit diameter, optic foramen [canal] size) can be measured on AMNHM 268006. However, there is another method that can be applied in the present case. Although the foramina defined by the lesser wing of the sphenoid are no longer present in AMNHM 268006, the position of the medial wall of the optic canal can be reliably identified (see Description of New Specimens, Cranial Remains). Using this landmark as one terminus and the point at which the zygomaticomaxillary suture intersects the inferior margin of the orbit (internal wall) as the other, a proxy measurement for orbital depth (POD) can be taken (fig. 15C, inset). This measurement is related to, but is obviously not the same as, true eyeball diameter. POD makes no allowance for eyeball sphericity, projection of the globe beyond the confines of the bony orbital margin, or distance between the orifice of the optic canal and the anatomical origin of the optic nerve on the eyeball. Our interest is simply to compare taxa (specimens) of varying body size for unit measurements that bear on the size of the bony orbit, in order to test whether or not Xenothrix had orbits of a size larger than expected.

Table 8 compares PODs and average body size (ABM) estimates for several platyrrhine taxa, including Xenothrix. POD is slightly smaller in Xenothrix than Aotus, but this is not especially meaningful in view of their substantial difference in body size. Rough adjustment for size can be achieved by dividing POD by ABM to generate an index, as is done in table 8 using natural log (ln) transformations. Results indicate that the orbits of Xenothrix are relatively small compared to those of other members of the comparative set, especially in the case of Aotus.

The same conclusion obtains when suitable allowance is made for the effect of allometry. As Schultz (1940) showed, there is a negative allometric relationship between eye size and body size in primates. This also applies to orbit size. Thus we should expect that, in proportion to body size, large primates will inevitably have smaller orbits than small primates do (cf. Kay and Kirk, 2000).

In the present case, performing an allometric correction is hindered by the fact that we lack body size estimates for specimens actually measured for POD (body size varies intraspecifically to a substantial degree within Platyrrhini [Ford and Corruccini, 1985]). Since appropriate data are unlikely to be gathered in the foreseeable future, any correction will have to be approximate. We utilized Schultz' (1940) data (which cover a much broader range of primate species than the set utilized here) to develop a linear regression expression relating orbital depth to body size. Derivation of the expression, in the form ln POD = 0.19(ln ABM) − 0.21, is explained in footnote e, table 8. It should be noted that all values in the last column of table 8 are much higher than 1 (expected ratio of left and right sides of the allometry equation), which probably reflects the introduction of various errors (original expression based on much wider sampling of primates, linear measurement indirectly derived from volumetric data, orbit depth and body weight not obtained from the same individuals/species). Although the dataset is therefore less clean than might be hoped, two general observations are warranted.

(1) With or without correction for allometry, Xenothrix falls close to taxa that have relatively small orbits as compared to Aotus. Indeed, Aotus stands apart from all other platyrrhines considered here: this is surely to be expected for a primate whose orbital index as defined by Martin (1990) places it in the top end of the range occupied by nocturnal lorises and lemurs. No matter how one wishes to express relative orbital size, Xenothrix emerges as different from Aotus—indeed, so different that they lie at opposite ends of the spectrum presented here.

(2) Accordingly, the bony orbits of the Jamaican monkey are not larger than expected, either in absolute terms or when compared with taxa in the same body size range. Instead, orbital depth appears to scale with measurements of other mid- to large-sized pitheciines, and not with Aotus (or Callicebus, for that matter). Even if one uses the lowest body weight (2 kg) that MacPhee and Fleagle (1991) considered reasonable for Xenothrix, the resulting index value is still only 2.5, comparable to Pithecia of average body size but quite unlike Aotus.

The case for Aotus-like orbital construction in Xenothrix thus founders. At least on the evidence considered here, Rosenberger's (2002) contention—that the Jamaican monkey followed the same “adaptive trend” that conditioned the evolution of orbital size in the owl monkey—appears to be without strong foundation.