Antillothrix USNMM

Phee, R. D. E. M & Meldrum, Jeff, 2006, Postcranial Remains of the Extinct Monkeys of the Greater Antilles, with Evidence for Semiterrestriality in Paralouatta, American Museum Novitates 3516, pp. 1-66 : 43-57

Antillothrix USNMM 254682

Ford (1986a) described the Samaná tibia ( fig. 24) in some detail and compared it to a wide range of other primates, and in regard to basic anatomical details we are in agreement with most of her observations. However, like Hershkovitz (1988) we are not convinced that the morphology of the distal tibia supports a strong case for callitrichid affinities. Indeed, of the numerous tibial characters that Ford (1986a, 1986b) considered in her cladistic analysis, only three strongly implied a callitrichid connection, and two of these are variable within the latter group. The remaining character, strong development of a bony flange for conducting the tendon of tibialis posterior m. into the foot, is produced to about the same degree in the Samaná tibia and Saguinus (which are shown as sister taxa in Ford’s cladogram). However, this feature, which is not uncommon among mammals generally ( MacPhee, 1994), shows about the same amount of development in Callicebus . As already noted the flange is absent in Paralouatta , making it further doubtful that this is a feature of much systematic valency. Hershkovitz (1988: 383) concluded that the Samaná tibia represented ‘‘a platyrrhine distinct from known living forms’’, possibly the same species as Antillothrix bernensis . Although we favor the same conclusion, it is unlikely that there will be any consensus until there is new material to evaluate.

Both Paralouatta MNHNCu 76.1034 and the Samaná tibia ( fig. 24) are notably robust for their respective sizes, as far as this can be judged, but they differ in several details (e.g., medial malleolus in Antillothrix is pyramidal in medial aspect, as in anthropoids generally, whereas in Paralouatta it is more massive and rectangular). The cotylar facet is well developed in both.

Functional Considerations

FIBULAR APPRESSION: Appression of the distal parts of the tibia and fibula occurs in several small and medium-sized platyrrhines ( Saimiri , Aotus , Callithrix , Cebuella , and Pithecia ) but not in larger taxa ( Meldrum and Kay, 1997). The appression of the distal fibula to the tibia of Paralouatta , if it occurred, suggests together with other features that the distal tibiofibular syndesmosis may have been extensive. This arrangement, combined with the equally massive medial malleolus, would have served to strongly resist dislocation by limiting lateral and rotational movement of the talocrural joint. Behaviorally, this morphology is correlated with frequent leaping in small-bodied primates, such as Saimiri , or with arboreal or terrestrial quadrupedalism in larger-bodied primates ( Meldrum and Kay, 1997). Although the fibula of Paralouatta is unknown, one might predict from the size of the lateral surface of the talocrural joint (see below) that the fibular malleolus must have been large as well. The presence of some but not all of these features in large atelines (cf. fig. 25) may indicate that the correlation is with BM.

CHEIRIDIA

A number of distinctive mammalian hand and foot bones were collected in the same caves that yielded remains unquestionably attributable to Paralouatta . After careful comparison of these elements to their homologs in other members of the Quaternary fauna, both living and extinct, we are persuaded that they probably belong to the Cuban monkey. For some elements, like the tarsals and distal phalanges, there is not much doubt because they strongly resemble those of other primates. In other cases judgment is more difficult (e.g., proximal and middle phalanges).

No carpals have been yet been recovered for any Antillean primate, and the few tarsal elements available (for Paralouatta only) are battered and not especially informative. Accordingly, descriptions and analyses in this section will be brief except for the discussion of digital ray length, which has some functional significance. Measurements for cheiridial elements can be found in tables 12 and 14.

and 76.3059 (Talus)

Two tali have been referred to Paralouatta : one is the partial talus MNHNCu 76.1037 of Paralouatta varonai from Cueva Alta ( fig. 26A–D); the other is MNHNCu 76.3059 from the Early Miocene locality of Domo de Zaza, which MacPhee et al. (2003) recently made the holotype of a second species, P. marianae ( fig. 26E–H). These authors noted that the two specimens differ only slightly— a remarkable point, given that 17–18 Ma allegedly separate them (see Introduction)— and for the characters of interest here they are nearly identical. Because MNHNCu 76.1037 is quite incomplete, we feel that it is justified to utilize the Miocene fossil as our basis for genus-level comments in this section.

In an earlier paper, MacPhee and Iturralde- Vinent (1995) concluded that there were no strong matches for the talus of Paralouatta among living platyrrhines. In particular, they noted that large-bodied platyrrhines (living atelines and Alouatta ) differ sharply from Paralouatta in possessing a ‘‘wedged’’ trochlear articular surface with overall low trochlear relief. This morphology helps insure maximum mobility at the talocrural joint; by contrast, the talus of Paralouatta seems to have been built for stability ( MacPhee and Iturralde-Vinent, 1995).

Paralouatta tali also present a definite if shallow cotylar fossa fronting a lengthy articular surface for the medial malleolus ( fig. 26A, C), thereby providing firm seating for the latter. The limits of the fossa are difficult to make out because of damage to the neck in both specimens. Talar cotylar fossae are insignificant (when present at all) in larger platyrrhines such as Ateles , Chiropotes , and Alouatta but are well developed in Old World monkeys such as Theropithecus and Lophocebus . According to Gebo and Sargis (1994), there is a similar contrast (but on a lesser scale) between moderately terrestrial Cercopithecus lhoesti and highly arboreal C. mitis . Previously, it was assumed (e.g., MacPhee, 1994) that the cotylar fossa was limited to cercopithecoids, in which it is normally well defined. Very deep cotylar fossae, not to be compared with the modest versions in primates under discussion here, occur in graviportal terrestrial mammals as disparate as the bear Ursus , the giant rodent Amblyrhiza , and the bibymalagasy Plesiorycteropus ( MacPhee, 1994; see also Zack et al., 2005).

To better characterize the distinctions of the Paralouatta talus, we took 12 morphometric measurements on MNHNCu 76.3059 and the tali of 14 other platyrrhine genera ( fig. 27; see tabulated data presented by Meldrum [1990]). Size-adjusted variates were created by scaling to the sum of talar head width and height. Paralouatta stands somewhat apart from the main trend linking talar size to m1 mesiodistal length in platyrrhines (although it is closest to uakaris and woolly monkeys), suggesting that it has a relatively large m1 for its talar length ( fig. 28).

Principal components analysis was carried out on both the individual specimens and the genus-level means, with comparable results. For brevity, the plots of the analysis of the means are presented along with the loadings of each variable on the first three axes, accounting for 99% of the variance ( fig. 29, table 13). The scatter produced a pattern that arranged the tali in a series of clusters corresponding to atelines, pitheciines-cebines, and callitrichines, with modest intergroup overlap; these results are comparable to those previously reported by Meldrum (1990). The talus of Paralouatta is notably distinct from the tali of other platyrrhines. In particular, the lack of trochlear ‘‘wedging’’ sets Paralouatta off from all other large-bodied platyrrhines along the second component in the principal components analysis, and is reflected in the high negative loading of posterior trochlear width (PTW) in table 13.

Paralouatta MNHNCu 76.1020 (Calcaneus) and MNHNCu 76.1021 (Middle Cuneiform)

The single calcaneus ascribed to Paralouatta is highly incomplete and badly preserved ( fig. 30). Among the few features that can be readily discriminated is the indication of a very wide peroneal shelf on the lateral side of the specimen, also seen (but not uniquely) in Callicebus .

Only a fraction of the plantar surface is preserved. The area still present is relatively flat as in Lophocebus rather than sharply ridged as in Alouatta . However, Lophocebus is not markedly different from Brachyteles in this regard, indicating that this feature is not correlated with a single locomotor category. Unfortunately, the plantar tuberosity for the origin of quadratus plantae m. (flexor accessorius m.), which is very pronounced and projecting in Alouatta and Brachyteles but not as developed in Lophocebus and Theropithecus , is not preserved in this specimen.

Among the more unlikely fossils recovered from Cueva Alta is the perfectly preserved middle cuneiform illustrated in figure 31. The facet for the head of the navicular is narrow but markedly rounded for reception of the navicular. This suggests that midfoot dorsoventral mobility was well developed, as is true of most anthropoids other than humans.

Paralouatta MNHNCu 76.1022, 76.1023, and 76.1025 –76.1028 (Metapodials)

Examples of both MC1 and MT1 are available, permitting them to be clearly discriminated morphologically and metrically. The single specimen of MC1 (MNHNCu 76.1022) is about three-quarters the length of the only MC2 (MNHNCu 76.1023) in the collection (table 14). Compared to that of Alouatta seniculus and other large-bodied extant platyrrhines, the MC1 of Paralouatta ( fig. 32) has a relatively less bulbous head, a straighter shaft, and a less projecting ulnar border. In addition, there are two prominent crests on the distal end not seen in the howler monkey. The one on the radial side is presumably for insertion of abductor pollicis longus m.; the one on the ulnar side, if not for the attachment of ligamentous tissue, is dubiously for the origin of an unusually developed first palmar interosseous m. A similar morphology is seen in the thumbs of some Old World monkeys (e.g., Trachypithecus pileatus ). The joint surface for the trapezoid is uniaxial and semicylindrical, as is normally the case in platyrrhines. There are obvious facets for distal sesamoids.

Of the two MT1s in the collection, MNHNCu 76.1025 ( fig. 34) is relatively complete; MNHNCu 76.1026 lacks its proximal end and has not been illustrated. The MT1 resembles the equivalent element of Alouatta , with the difference that the process for peroneus longus m. is better developed in the former. There is torsion of the head, but Alouatta exhibits about the same degree of twisting as Paralouatta . There are several additional metapodials in the hypodigm (table 2), the best preserved of which are MC2 MNHNCu 76.1023 ( fig. 33) and MT3 MNHNCu 76.1028 ( fig. 35). As discussed below, in relation to the estimated BM of Paralouatta , all metapodials are robust and relatively short (cf. table 14).

MNHNCu

76.1029 –76.1033 (Phalanges)

As already noted, some of the phalangeal elements described in this section assigned to the Paralouatta hypodigm have unusual features. However, there does not appear to be any other vertebrate taxon to which they could be justifiably assigned, and they are therefore accepted here as belonging to the Cuban monkey.

There are four proximal phalanges assignable to Paralouatta . The first of these (MNHNCu 76.1029, fig. 36A) is distinct morphologically, and judging from the asymmetry of its proximal articular surface, it must be part of a DR1. Its size and robusticity suggest that it belongs to the DR1 of the pes (cf. Hamrick et al., 1995), which means that the condition of the thumb in the Cuban monkey is still unknown. However, given the large size of MC1, there is no reason to suspect that the thumb was reduced as in atelines.

The remaining proximal phalanges (MNHNCu 76.1030a–c) can be placed only in a general sense, as representatives of one or another of DR2–5; only one (MNHNCu 76.1030a) is complete ( fig. 36B). Its proximal end is notably expanded, with a single concave facet for reception of a large metapodial head. The shaft is marked by ridges for the flexor sheath, and the head presents the usual pulleyshaped articular surface for the middle phalanx. Although this specimen is utilized in the reconstruction of DR2 (manus) in figure 37, there is no morphological basis for regarding it as manual rather than pedal.

A group of 12 distinctive middle phalanges from the monkey caves were compared to, and found to be unlike, those of any known megalonychid, capromyid, echimyid, or solenodontid (i.e., the other terrestrial mammalian taxa represented in the Quaternary fauna of Cuba). All are extremely short, averaging 9.6 mm in length (cf. table 14, footnote b), and are robust for their length. Apart from size, however, they otherwise resemble typical platyrrhine middle phalanges and may reasonably be considered to represent Paralouatta ( fig. 36C). Their proximal ends are dimpled by two small, concave facets separated by an indistinct ridge, to articulate with the pulleys of the proximal phalanges. Their distal ends are likewise pulley-shaped to receive the unguals. Each bears well-defined medial and lateral ridges for insertion of flexor digitorum superficialis m. and associated flexor sheath.

The five distal phalanges in the hypodigm are notably unlike their counterparts in other Cuban mammals. Four are similar to one another and display expanded or ‘‘cauliflower’’ distal ends (tuberosities or apical tuffs), to which connective tissues of the finger-tip pulp would have attached (MNHNCu 76.1032a–d, fig. 36D). One would normally conclude without hesitation that unguals having this morphology must represent DR1, because in known platyrrhines the unguals of digits other than the thumb and big toe are conical or only slightly expanded. In the existing Paralouatta hypodigm there is only one specimen (MNHNCu 76.1033, fig. 36E) an- swering to this latter description, although because the specimen in question is somewhat damaged it is difficult to be sure of its original condition. It might be argued that the odds are against recovering four unguals of DR1 but only one representative from all of the other digital rays combined, but the alternative interpretation is that distal expansion was present on many or most distal phalanges in Paralouatta . Were that the case, distal phalanx morphology in the Cuban monkey would be unlike that of any other New World monkey, large or small—or, for that matter, any Old World monkeys except the large ground-adapted species like Theropithecus gelada and Macaca nemestrina in which the unguals are distally expanded. Erythrocebus patas displays a different pattern which may be relevant to note: the distal phalanges of most fingers and toes other than the thumb and big toe are only slightly expanded, but the axial digits (DR3) have larger and wider elements not unlike MNHNCu 76.1032a ( fig. 36D).

It is unfortunate that the issue of distal phalanx placement has to be left unresolved, as the condition of DR2–5 unguals would appear to have some diagnostic value regarding preferred substrate. However, there is an additional source of inference, digital ray length, that provides some further functional insight into the Cuban monkey’s cheiridial morphology (see next section).

Functional Considerations

DIGITAL RAY LENGTH: As a group, the phalangeal specimens discussed in the preceding section are chiefly remarkable for their short length and sturdy build. If properly assigned, and if phalanges scaled in Paralouatta as they typically do in other platyrrhines, for its projected BM the Cuban monkey must have had relatively short fingers and toes. Obviously, this notion is difficult to explore in detail given the material available: With the exception of DR1, the units of which are morphologically distinctive, there is no way of allocating specific phalanges to their proper DRs with any certainty. The best that can be done is to assemble a ‘‘typical’’ DR by piecing together appropriate specimens to achieve a first approximation of ray length.

The digital ray chosen for illustrative purposes is DR2 of the manus, which we reconstruct in figure 37 using MC2 MNHNCu 76.1023 (33.9 mm) + proximal phalanx MNHNCu 76.1030a (17.0 mm) + an ‘‘average’’ middle phalanx (9.6 mm) + distal phalanx MNHNCu 76.1033 (7.6 mm) (table 14; fig. 37A). The middle phalanx depicted in figure 37A, MNHNCu 76.1031a, is actually the longest in the available sample at 11.2 mm, but it is scaled to the average. As noted in the preceding section, MNHNCu 76.1033 is the only distal phalanx in the collection with the conical extremity typical of DR2–5 unguals in other platyrrhines. The ‘‘cauliflower’’ unguals are only a millimeter or two longer, however, so for this purpose it matters little which kind is chosen.

Adding these values together provides the basis for computing a ‘‘metacarpal contribution’’ ratio (MPCR); in table 15 values for this ratio in a variety of anthropoids are set out in descending order of magnitude. The range is fairly substantial, 0.35 to 0.57, although most values lie within the range 0.38–0.41 regardless of BM. The ratio for the composite DR2 of Paralouatta (0.50) is notably high in comparison to other platyrrhines in the table, and is consistent with the interpretation that the fingers would have been rather short in relation to the palmar skeleton. In the comparative set the Cuban monkey’s value is exceeded only by that of the highly terrestrial cercopithecids Theropithecus gelada and Erythrocebus patas . (Estimates using the longest and shortest middle phalanges in the available Paralouatta sample yield ratios of 0.49–0.51. Using a short metapodial like MC2 actually favors a lower metapodial contribution ratio; utilization of the same phalangeal elements with the other complete metapodials would result in higher ratios.)

Although among measured platyrrhines Chiropotes satanus (BM, 2.6–2.9 kg; Fleagle, 1999) makes the closest approach to Paralouatta for this ratio, comparison needs to be made to a species of comparable BM. In figure 37B (see also table 15), a ligamentous preparation of the DR2 of a relatively small male Alouatta seniculus (AMNHM 70087) is depicted at the same scale as the Paralouatta reconstruction and positioned so that their metacarpophalangeal joints are in tandem (arrows). Morphologically, the elements making up the howler DR2 appear to be slightly more gracile than their counterparts in Paralouatta , but actual differences are minor. However, the blunt appearance of the fingertip in the Paralouatta reconstruction would have been even more pronounced if one of the distal phalanges with an expanded distal end had been used instead. Although MC length is similar in the howler specimen and the Paralouatta reconstruction, there is a sharp contrast in ray length because A. seniculus possesses relatively much longer PPs and MPs. Indeed, the red howler’s index finger is about a third longer, its PP alone being as long as the combined length of the Cuban monkey’s PP and MP (cf. fig. 37).

Having fingers and toes that were short relative to metapodial length need not mean that the former were short in an absolute sense. However, in the case of Paralouatta this appears to be true, as can be appreciated by examining the ‘‘middle phalanx contribution’’ ratio in various taxa (table 15). Once again there is a considerable spread in values (0.18–0.40), although most species cluster in the range 0.31–0.35, which indicates that middle phalanges do not vary much in proportion to total finger length. The outstanding contrast in this column is between Theropithecus , which has an extremely short MP (ratio, 0.18), and all other taxa. This is similar to Schultz’ (1963) finding that ground-living monkeys like Theropithecus and Erythrocebus (and humans) have dramatically shorter toes relative to MT length than do other anthropoids. Paralouatta is next, with 28% of total phalangeal length being contributed by the MP (mean length, 9.6 mm). However, the Cuban monkey cannot be sharply separated from the mass of other taxa in the table unless the absolute size of DR units is also taken into account. Thus in the case of the measured specimen of A. seniculus , the ratio is 0.31, which is only a little higher than that of Paralouatta . However, in absolute terms the howler’s middle phalanx is almost twice as long as that of Paralouatta . In the case of Brachyteles , the MPCR is also about the same, but the woolly monkey’s MP is more than two and a half times longer than the mean value for Paralouatta .

We will not repeat this exercise for the DR skeleton of the foot, as it is obvious that results would be similar because the same made-up set of phalanges would have to be used for Paralouatta . Nevertheless, it seems reasonable to conclude that, just as with many other parts of the skeleton of Paralouatta , cheiridial metrics point strongly away from the Cuban monkey’s having had a locomotor repertoire like those of living atelines.