Reconstructing the Locomotor Repertoire of Protopithecus brasiliensis. II. Forelimb Morphology


  • Lauren B. Halenar

    Corresponding author
    1. The Graduate Center, Department of Anthropology, City University of New York, New York Consortium in Evolutionary Primatology (NYCEP), New York, New York
    • The Graduate Center, Department of Anthropology, City University of New York, 365 Fifth Avenue, New York, NY 10016.
    Search for more papers by this author


The majority of previous publications have suggested that the large-bodied subfossil Protopithecus brasiliensis was a suspensory ateline with a locomotor repertoire similar to that of extant Ateles and Brachyteles. This is unexpected, as the cranial morphology of Protopithecus is very similar to Alouatta, a genus usually classified as a deliberate quadrupedal climber. Complicating matters further, as Protopithecus is twice as large as Ateles and Brachyteles, its ability to be as suspensory as those two genera is suspect and a terrestrial component of the locomotor repertoire has also been hypothesized. The forelimbs of Protopithecus, while relatively elongated as would be expected in a suspensory animal, are also quite robust and show several adaptations for climbing. To test these hypotheses about the fossil locomotor repertoire, three-dimensional geometric morphometric techniques were used to quantify the shapes of the fossil distal humerus and proximal ulna and then compare them to a broad sample of extant primates with varying body sizes and locomotor patterns. Results indicate that Protopithecus is similar to Ateles and Brachyteles in terms of its forelimb joint surface morphology; however, the overall locomotor repertoire of the fossil is reconstructed as more flexible to include forelimb suspension, climbing, and potentially some terrestrial ground use. The combination of suspensory locomotion and quadrupedal climbing supported here indicates the beginnings of the evolutionary transition from a more acrobatic style of locomotion in the last common ancestor of alouattins and atelins to the current pattern of howler locomotion. Anat Rec, , 2011. © 2011 Wiley Periodicals, Inc.


From the time of its discovery in 1836, Protopithecus brasiliensis has been associated with the extant suspensory atelines Ateles and Brachyteles (Lund,1838; Winge,1895). The first fragmentary specimens to be recovered from the Lagoa Santa (LS) caves in Minas Gerais, Brazil, a left proximal femur and a right distal humerus, were largely ignored for over 100 years after being described as a giant Brachyteles and assumed to be nearly identical to that genus (Hartwig,1995a). It was not until the mid-1990s, when a nearly complete skeleton of Protopithecus was discovered in the Toca da Boa Vista (TBV) caves in the neighboring state of Bahia, that the scientific mystery of this taxon was fully appreciated. The full skeleton seems to combine traits found in two genera that are usually seen as belonging to opposite ends of the ateline adaptive spectrum; its skull shares several derived characters with Alouatta while its teeth and postcranial skeleton are more similar to Ateles (Hartwig and Cartelle,1996). Along with its large body size that has been estimated at 23–25 kg, twice the size of any living New World monkey (Hartwig,1995b; Hartwig and Cartelle,1996; Halenar, 2011), this mosaic of traits has made interpreting the paleobiology of Protopithecus and its impact on reconstructing the evolutionary history of the ateline primates a challenge that few have attempted to address.

Locomotor behavior is one facet of Protopithecus paleobiology that has remained enigmatic. The original suggestion of ateline-like suspensory locomotion, which was mostly based on limb proportions (Hartwig and Cartelle,1996), has been challenged by another hypothesis pointing to the large body size of Protopithecus as evidence for a “high degree” of terrestrial behavior (Heymann,1998). More recently, several quantitative traits in the Protopithecus postcranial skeleton have been shown to group it with other “brachiating” primates (Jones,2008); however, because of the fragmentary nature of the fossil joint surfaces used to calculate some of the key indices, these results should be viewed with caution. Different methodology will be used here in an attempt to test the various hypotheses that have been proposed so far for the locomotor repertoire of Protopithecus. Detailed qualitative descriptions of the humerus, radius, and ulna will be given first. Then three-dimensional geometric morphometric (3DGM) techniques are used to capture the shape of the distal humerus and proximal ulna and to quantify aspects of their morphology that are known to vary in predictable ways with locomotor patterns in extant primate taxa. Principal components analyses (PCA) are used to visualize the major axes of shape variation within the sample. The species, or group of species, to which the fossil is most similar is used as an analog for reconstructing its locomotor repertoire. Results of these analyses are discussed in relation to the overall evolutionary trajectory of the alouattin tribe and the sequence of changes leading from the ateline last common ancestor to modern Alouatta.


Table 1 was constructed as an aid in comparing the Protopithecus forelimb with forelimb morphology known to vary predictably amongst taxa of different locomotor patterns. The qualitative data compiled here present a muddled picture with few truly diagnostic characters, as the same features appear under different types of locomotion and characters found in Protopithecus fall under several different behavioral categories. However, it is still instructive to examine each element of the fossil on its own in more detail.

Table 1. Qualitative descriptions of aspects of forelimb morphology known to vary with locomotor pattern in extant primatesa
 Arboreal quadrupedalismSuspensoryClimbingClinging and leapingTerrestrial quadrupedalism
  • a

    Compiled from Gebo (1993), Larson (1993), Meldrum (1993), Rose (1993), MacPhee and Meldrum (2006), Jones (2008), and personal observation.

  • bThe descriptions in bold describe the relevant morphology in Protopithecus.

Proximal humerus
 HeadOval, posteriorly directedLarge, round, medially directed  Oval, posteriorly directed
 TuberclesProject proximally above headEqual to or lower the headbGreater faces anteriorlyBelow the headProject proximally above head
 Bicipital grooveBroad, shallowNarrow, deepBroad Broad
 Deltoid tuberosityBroad, v-shaped, quarter of the way down the shaftAt least halfway down the shaft  Quarter to a third of the way down the shaft
 Humeral shaftShort, bowed slightly outward in anterior viewLong, straight, slenderHigh degree of torsion Strong medial curvature
Distal humerus
 Medial epicondyle Large, medially directedLarge, medially directedMedially directedPosteriorly directed
 CapitulumBroad, distolaterally expansiveSpherical, posteriorly extensiveSpherical, “unrolled”Inflated, capitular tailFlattened distally
 TrochleaWide, conical, prominent edgesBroad, cylindrical, low edgesCylindricalNarrow, cylindricalNarrow, strong posteriomedial edge
 Zona conoideaFlatRecessed, gutteral Recessed 
 Brachioradialis flange IntermediateEnlargedEnlarged 
 Olecranon fossaShallow Shallow Deep, lateral wall articular surface
Proximal ulna
 Olecranon processIntermediateShortIntermediateRelatively longLong and retroflexed
 Trochlear notchWide, shallowWide, shallowNarrow, long Narrow, lateral articular facets
 Coronoid processIntermediateNot projectingIntermediate High
 Radial facetFlat, faces anterolaterallyFaces laterallyLarge, faces anterolaterally Outset, subdivided
 Thumb Vestigial   
 PhalangesLongLong, curvedLong, slender, curvedLongShort

Unfortunately, the scapula of Protopithecus is very fragmentary and does not preserve any functional information. Similarly, the proximal humerus, especially parts of the head that would make clear its size, shape, and orientation, is only present on the right side and is not well preserved (Fig. 1). However, the bicipital groove and size of the tubercles can be seen; the groove is relatively narrow and deep and the tubercles sit below the level of the head, as in suspensory primates (Larson,1993). The shaft of the humerus is straight and relatively long but is also quite robust as would be expected in an animal of its body size. Distally, the specimens from LS and TBV differ from one another in their size and robusticity (Fig. 1); this is especially apparent in the width of the entire joint and prominent brachioradialis flange of the TBV specimen. These are some of the features pointed out by Hartwig and Cartelle (1996) that suggest climbing as an important part of the locomotor repertoire for this specimen. The LS specimen is similar in shape to the larger TBV specimen, and its smaller size could indicate a relatively high level of sexual dimorphism for Protopithecus; other atelines, especially Alouatta, are also sexually dimorphic (Ford,1994). If the TBV specimen represents a larger male individual, a well-developed brachioradialis flange would be expected as the flexor muscles that attach there would also be larger for hoisting its heavier body through the trees and pulling itself up on branches. The medial epicondyle on both specimens is large and projects medially as in arboreal quadrupedal primates; it is not retroflexed as in terrestrial Old World monkeys. Neither do the fossils exhibit the extreme distal projection of the medial edge of the trochlea like those terrestrial primates (Rose,1988,1993). Protopithecus also does not have an entepicondylar foramen, a feature that has been lost in atelines but is frequently seen in other platyrrhines such as Cebus, Aotus, and Pithecia (Gebo,1993).

Figure 1.

Humerus of Protopithecus. Note the narrow, deep bicipital groove and low tubercles on the proximal end and the large brachioradialis flange on the distal end of the Toca da Boa Vista specimen (left). The specimen from Lagoa Santa (right) is smaller and lacks the protruding brachioradialis flange, but otherwise is similar in joint morphology to the TBV specimen. Scale bar = 1 cm.

The radius of Protopithecus is not complete (Fig. 2); only the proximal end is present and the head is heavily worn and covered in calcite deposits. Without the entire shaft it is hard to judge important characteristics of the radial head and tuberosity related to their orientation. The radial tuberosity, however, is relatively large and the head seems to be oval in shape, but the circumference is incomplete. A more circular radial head is a characteristic of suspensory primates (Rose,1993). The proximal ulna also does not have any of the classical features indicative of terrestrial locomotion in Old World monkeys (Fig. 3). More like an arboreal climber, the radial facet is inset against the shaft of the bone and faces anterolaterally as opposed to the outset condition seen in a terrestrial primate (Gebo,1993; Rose,1993). The olecranon process is not retroflexed; the superior surface is angled slightly posteriorly but the whole process is prominent and oriented proximally, much like in Ateles. However, other aspects of the joint surface are not similar to the Ateles condition. For example, the distal facet of the trochlear notch is much smaller and less convex. Also, the coronoid process is slightly more projecting and oriented at a shallower angle. As in the humerus, the ulna exhibits a combination of traits usually seen only in either suspensory taxa like the atelins or taxa that are more generalized arboreal climbers like Alouatta.

Figure 2.

Partial proximal radius of Protopithecus. Scale bar = 1 cm.

Figure 3.

Proximal ulna of Protopithecus in anterior (left) and lateral view (right). Note the relatively prominent and straight olecranon process, the inset radial notch, and the relative overall width of the trochlear notch. Scale bar = 1 cm.

Moving down to the distal end of the forelimb, degree of phalangeal curvature is also strongly correlated with substrate preference: arboreal taxa show much more curved phalanges than terrestrial taxa (Susman et al.,1984; Hamrick et al.,1995; Jungers et al.,1997). Qualitative inspection of the Protopithecus phalanges suggests that they are curved as in arboreal primates (Fig. 4). In fact, the included angle of curvature (Stern et al.,1995) for one proximal phalanx of Protopithecus is approximately 60 degrees; this is in the high end of the range of values reported for Ateles and Hylobates but below the range of Pongo (Jungers et al.,1997). The phalanges also have relatively strong and distally placed flexor sheath ridges, indicating strong grasping abilities (Almécija et al.,2007,2009). Unfortunately, it is unclear whether the phalanges in the sample are from the hands or the feet. Also, there are no relevant metacarpals preserved to determine whether Protopithecus had a vestigial pollex, a hallmark of both Ateles and Brachyteles (Biegert,1963; Erikson,1963; Jouffroy et al.,1991; Tague,1997).

Figure 4.

Composite ray of Protopithecus. Lines on the proximal phalanx show the dimensions necessary for calculating the included angle of curvature, which on this specimen was approximately 60 degrees. Scale bar = 1 cm.

Although some single aspects of the joints might appear similar to nonplatyrrhine taxa, there is no suite of characters seen in the Protopithecus forelimb that would suggest extreme locomotor specializations like those seen in leaping or slow climbing strepsirhines or terrestrial Old World monkeys. As has been observed in other parts of the skeleton, the forelimb combines traits possessed by extant primates that practice both below-branch forelimb suspension and arboreal climbing. These qualitative aspects of the Protopithecus forelimb will be quantified using landmark-based 3DGM techniques.


A comparative sample of living primates was used consisting of adult male and female wild-shot individuals from collections at the American Museum of Natural History (AMNH) in New York, the National Museum of Natural History (NMNH) in Washington, DC, and the Museu Nacional (MN), Rio de Janeiro, Brazil (Table 2). This set of taxa was chosen to represent a diverse array of body sizes, locomotor patterns, and phylogenetic affinities, extending the comparative framework used in previous studies of the Protopithecus postcranial remains. The fossil material includes the original Protopithecus distal humerus discovered in the LS caves, which is now housed in the Universitets Zoologisk Museum in Copenhagen, Denmark; the nearly complete Protopithecus skeleton from TBV, curated in the Museu de Ciências Naturais at the Pontificia Universidade Católica de Minas Gerais in Belo Horizonte, Brazil; the nearly complete skeleton of Caipora bambuiorum, discovered at the same time and in the same cave as the TBV Protopithecus material (Cartelle and Hartwig,1996); and several subfossil lemur specimens curated in the AMNH (Archaeolemur sp.: AMNH30042-B-10; Paleopropithecus ingens: AMNH30042-B-1, 30042-A-3, 30042-A-5; Megaladapis edwardsi: AMNH30042-A-1; plus several uncatalogued specimens assigned to those genera). The Caipora skeleton is of a subadult individual, but epiphyses of the distal humerus and proximal ulna are fully fused.

Table 2. Taxa and number of individuals included in the comparative samplea
 Male body weightb (kg)Locomotor repertoireDistal humerusProximal ulna
  • Abbreviations: VCL = vertical clinging and leaping; AQ = arboreal quadrupedalism; PHT = prehensile tail. The “?” indicate that the locomotor repertoire for those two fossil taxa is unknown/under investigation.

  • a

    Data from Rosenberger and Strier (1989), Cartelle and Hartwig (1996), Fleagle (1999), Jungers et al. (2002), Di Fiore and Campbell (2007), and Halenar (2011).

  • b

    Only male body weights are listed for comparative purposes as both TBV Protopithecus and Caipora have been suggested to be males.

Archaeolemur 14–24AQ, terrestrial, climbing  1 
Megaladapis 38–75Arboreal climbing and clinging3 3 
Palaeopropithecus 45–52Upside-down below branch suspension2 4 
Aotus0.90.7–1.2AQ, leaping14151415
Pithecia2.41.4–3.1AQ, leaping3232
Chiropotes3.22.9–3.5AQ, leaping2 2 
Alouatta7.24.0–12.5AQ, climbing, PHT33283226
Lagothrix8.57.1–10.2AQ, climbing, PHT2424
Ateles8.35.5–9.8AQ, suspensory, PHT51049
Brachyteles11.09.2–13.8AQ, suspensory, PHT4343
Protopithecus 20–25?2 1 
Caipora 20.5?1 1 
Presbytis7.55.5–12.0AQ, leaping19161716
Macaca10.55.4–18.3AQ, terrestrial157158
Colobus10.89.6–13.5AQ, leaping2113118
Nasalis20.0 AQ, leaping, swimming8181
Papio23.516.0–31.0AQ, climbing, terrestrial117117
Symphalangus12.0 Brachiation, climbing2121
Pan46.039.0–59.7AQ, suspensory, terrestrial106116

A Microscribe 3DX digitizer was used to collect the three-dimensional coordinates (x, y, z) of a set of landmarks designed to capture the shapes that are expected to vary among living taxa of differing locomotor profiles (Fig. 5; Table 3). Data were collected on the distal humerus and proximal ulna; these elements are well preserved in Protopithecus on at least one side of the body. As there are few biologically homologous Type I landmarks, such as the meeting point of two sutures, on postcranial elements, sets of Type II landmarks, those that are defined by the geometry of the specimens (Bookstein,1991), were designed specifically for this study. During data collection, both the distal humerus and the proximal ulna were positioned in such a way so that landmarks on all sides of the bone could be digitized without changing its orientation.

Figure 5.

Three-dimensional landmarks collected to describe the shape of the distal humerus (A) and proximal ulna (B). Landmarks are shown on the TBV Protopithecus and are connected as a reference for the wireframes in the following PCA figures. Scale bars = 1 cm.

Table 3. Anatomical landmark definitions
  • a

    The anterior and posterior landmarks on the distal humerus were digitized together without changing the orientation of the specimen.

Distal humerusa 
  A1Most lateral point
  A2Most medial point
  A3Most lateral point on the capitulum
  A4Most medial point on the capitulum
  A5Most superiolateral point on the trochlea (excluding the capitulum)
  A6Most inferiomedial point on the trochlea (excluding the capitulum)
  A7Most superior point on the medial epicondyle
  A8Most inferior point on the medial epicondyle
  A9Most superior point on the lateral epicondyle
  A10Most inferior point on the lateral epicondyle
  P1Most superior point of the olecranon fossa
  P2Most inferiomedial point of the olecranon fossa
  P3Most inferiolateral point of the olecranon fossa
  P4Deepest point of the olecranon fossa
  P5Most medial point on the trochlea
  P6Most lateral point on the trochlea
  P7Tip of the medial epicondyle
 Proximal ulna 
  U1Most proximal point of the olecranon process
  U2Most medial point on the maximum constriction of the olecranon process
  U3Most lateral point on the maximum constriction of the olecranon process
  U4Most posterior point on the olecranon process
  U5Most anteriomedial point on the olecranon process
  U6Most aneriolateral point on the olecranon process
  U7Most medial point on the “wing” of the proximal articular facet
  U8Most lateral point on the “wing” of the proximal articular facet
  U9Most anterior point on the proximal border of the proximal articular facet
  U10Most distomedial point of the proximal articular facet
  U11Most distolateral point of the proximal articular facet
  U12Deepest point in the midline of the trochlear notch
  U13Most posteriomedial point of the distal articular facet
  U14Most anterior point of the distal articular facet
  U15Most anterior point of the radial facet
  U16Most posterior point of the radial facet
  U17Most proximal point of the radial facet
  U18Most distal point of the radial facet
  U19Deepest point in the radial facet

So far, few 3DGM analyses involving postcrania have been published (e.g., Drapeau,2008; Harcourt-Smith et al.,2008) and none of them are focused on platyrrhines. 3DGM techniques, especially generalized Procrustes analysis (GPA; e.g., O'Higgins and Jones,1998), are expected to be of particular help in answering questions about the large-bodied Protopithecus because it is a size-independent method that scales specimens to unit centroid size as a way of normalizing body mass differences within a sample. PCA were conducted on the results of a GPA using the program morphologika2 v2.5 (O'Higgins and Jones,2006) to visualize the morphological variation within the sample for each skeletal element. Clouds of points representing groups of taxa with similar locomotor behaviors were thus produced and the locomotor behavior of the fossil was inferred based on its position with respect to the comparative sample; for example, if Protopithecus were to fall in the middle of a cloud of acrobatic suspensory atelines, this would support the hypothesis that it was using acrobatic suspensory locomotion. This method was chosen over discriminant function analysis, which was used previously in a quantitative analysis involving Protopithecus (Jones,2008), because it does not require the user to specify a particular locomotor category for each specimen in the comparative sample. As discussed below, living primates use a variety of locomotor behaviors that can leave their mark on the postcranial skeleton and reducing taxa to a single categorical definition might not be as useful as allowing their variation to fall out naturally in multidimensional shape space. The inclusion of a fossil in a PCA lets the fossil speak for itself with regard to its similarities to extant taxa.

PCA were run on several different iterations of the sample for each element; one including the entire comparative sample, one including only the platyrrhine taxa, and one on the male species mean landmark configuration for those platyrrhines which was then overlaid with a minimum spanning tree to connect the most similar shapes. Both the mean landmark configurations and the minimum spanning trees were calculated using the PAST software package (Hammer et al.,2001). As the comparative sample spans a relatively wide range of body sizes, for each PCA, the first two principal component axes scores were regressed against ln centroid size of the specimens to test for correlation with body size; although the size of an animal is certainly related to how it can move through its habitat and is therefore relevant to the variation in the sample, it is not desirable for size alone to be driving the grouping patterns on the axes that represent the majority of that variation.


A PCA of the entire primate-wide sample for the distal humerus results in a clear separation of New World monkeys, Old World monkeys, hominoids, and strepsirhines (Fig. 6). The Old World monkeys are arranged across PC1 from more arboreal taxa like Colobus on the left to the more terrestrial baboons on the right, with the variable macaques spanning both groups. In both the full sample and platyrrhine-only analyses (Figs. 6 and 7), the suspensory atelins form their own group separate from the more generalized arboreal taxa (Alouatta and Cebus) and those that add more leaping to their repertoire (Aotus, Pithecia, and Chiropotes). The three fossils, both Protopithecus individuals and Caipora, are consistently part of this atelin cluster because of their mediolaterally wide distal humerus and large medially projecting medial epicondyle. In the full sample PCA, the TBV humerus is situated closer to the strepsirhine cluster on PC2 due to the enlarged brachioradialis flange that it shares with some members of that vertical clinging and leaping group (Fig. 6). When a minimum spanning tree is overlaid connecting the three-dimensional shapes representing the male mean landmark configuration for each species of platyrrhine in the sample, the TBV humerus is joined to Brachyteles as is the LS specimen, which is also linked to Lagothrix (Fig. 8). The two Protopithecus individuals are not linked to each other, emphasizing the differences between the two specimens noted in the qualitative description above.

Figure 6.

PCA results for the distal humerus of the entire comparative sample. The taxa are arrayed across PC1 (26% total variance) based on the width of the joint, which is affected by the length and orientation of the medial epicondyle and height and depth of olecranon fossa. PC2 (14% total variance) shows the variation based on the height of the epicondyles. The wireframes show the morphology in the nearest cluster of a right humerus in anterior view (see Fig. 5A for reference). Bottom left Protopithecus = TBV; top right = LS.

Figure 7.

PCA results for the distal humerus of the platyrrhine taxa only. PC1 (23% total variance) is being driven by the height of the olecranon fossa and the length of the medial epicondyle. PC2 (14% total variance) shows the variation in the height of lateral epicondyle and the orientation of medial epicondyle. The wireframes represent the morphology in the nearest cluster of a right humerus in anterior view (see Fig. 5A for reference). Top left Protopithecus = TBV; bottom right = LS.

Figure 8.

PCA results for the average male distal humerus shape for all platyrrhine species in the sample overlaid with a minimum spanning tree connecting the most similar shapes in three-dimensional space. Both Protopithecus specimens are connected to Brachyteles, while not being connected to each other, emphasizing the differences between the two fossil specimens as described in the text.

The full sample PCA for the proximal ulna produces clustering of New World monkeys, Old World monkeys, and hominoids similar to that seen for the distal humerus (Fig. 9). The strepsirhine group is different in this case, as the subfossil lemurs are separate from their extant relatives, Propithecus and Indri; the larger size of the subfossils and their extremely reduced olecranon process make them more similar to the living hominoids in the sample. When the platyrrhines are analyzed separately, the distinctions between the various taxa are clearer for the proximal ulna than they were for the distal humerus (Fig. 10). Again, Protopithecus and Caipora group with the more suspensory atelins, not with the less agile Alouatta, because of their shorter olecranon process and wider, more anteriorly facing trochlear notch. A minimum spanning tree produces the same results as for the distal humerus, linking Protopithecus with Brachyteles as the extant taxon sharing the most similar joint surface morphology (Fig. 11). For neither element is the much larger size of the fossils driving any of the grouping patterns on either PC1 or PC2; when ln centroid size of each individual in the sample is regressed against its PC score, R2 values are all less than 0.5.

Figure 9.

PCA results for the proximal ulna using the entire comparative sample. PC1 (30% total variance) represents the variation in the height of the olecranon process and the width of the distal facet of the trochlear notch. Variation along PC2 (13% total variance) is driven by the orientation of the proximal portion of the trochlear notch and the radial facet. The wireframes represent the left ulna in anterior view showing the morphology represented by the nearest cluster (see Fig. 5B for reference).

Figure 10.

PCA results for the proximal ulna of the platyrrhine taxa only. Morphological variation along PC1 (20% total variance) is driven by the length and orientation of the distal facet of the trochlear notch. PC2 (13% total variance) is driven by the height of the olecranon process and the overall width of the trochlear notch. The wireframes represent the left ulna in anterior view showing the morphology in the nearest cluster (see Fig. 5B for reference).

Figure 11.

PCA results for the average shape of the male proximal ulna shape in the platyrrhine portion of the comparative sample overlaid with a minimum spanning tree connecting the most similar shapes in three-dimensional space. Protopithecus is linked with Caipora, most likely due to the larger body size of the two fossils, as well as with Brachyteles as in the results for the distal humerus.


Broad categories of locomotor behavior, that is, “quadrupedal,” “arboreal,” “terrestrial,” or “leaper,” are not necessarily useful when it comes to describing the complex behavior of living primates (e.g., Prost,1965; Stern and Oxnard,1973; Hunt et al.,1996). The entire locomotor repertoire of an individual or a species might be different from the locomotor behavior it uses most often on a daily basis. This creates a situation akin to that surrounding “fallback foods” (e.g., Marshall and Wrangham,2007) and the “critical function” hypothesis (Rosenberger and Kinzey,1976); the morphology preserved in a fossil of the postcranial skeleton could be reflecting adaptations to the way the animal moves most often while traveling or, instead, to those parts of the behavioral profile that could be most important to survival, such as fleeing predators or feeding. This is especially important for a fossil like Protopithecus, which presents a mosaic of traits that could fall under several different, commonly recognized locomotor categories. Another relevant issue involves work done on comparing ateline “brachiation” to hylobatid “brachiation.” Historically, the term “semibrachiation” has been used for the atelines (Ashton and Oxnard,1963,1964; Napier,1963; Oxnard,1963; Ashton et al.,1965; Napier and Napier,1967; Rose,1973); however, there are several problems with this conceptualization, most notably the fact that further detailed behavioral studies have shown that the animals that have been included in this category all move in very different ways (Stern and Oxnard,1973; Mittermeier and Fleagle,1976).

Protopithecus has fallen victim to this terminological chaos as well and has been lumped together with the Ateles-style “brachiators” based on possession of a suite of traits associated with this genus, including long forelimbs relative to hindlimbs, long and straight diaphyses, mobile shoulder joints, curved phalanges, and (inferentially) a prehensile tail (Hartwig and Cartelle,1996; Jones,2008). However, Hartwig and Cartelle (1996) also pointed out unique features of the Protopithecus postcranium, such as the well-developed brachioradialis flange on the distal humerus and the large attachment site for the gluteus medius muscle on the ilium, that suggest climbing as an adaptively important part of the locomotor repertoire in addition to suspension. Alouatta has a large gluteus medius as well as a large gluteus maximus, both of which are useful in “antipronograde” postures (Stern,1975). The gluteus maximus and tensor fasciae femoris of howler monkeys form a muscle mass in the hip, similar to the deltoid muscle in the shoulder, that is useful in suspension by the hindlimbs (Stern and Oxnard,1973). This positional behavior is used often by Alouatta during feeding and is seen much more frequently than is the type of suspension by the forelimbs characteristic of Ateles (Stern,1971; Gebo,1992). This combination of suspensory locomotion and deliberate quadrupedal climbing in Protopithecus is confirmed here, both qualitatively and quantitatively.

One of the most mysterious aspects of the evolutionary history of New World monkeys is the lack of terrestrial species. To date, no fossil or living South American primate has been described as spending the majority of its daily activity budget on the ground, as do many Old World monkeys and lemurs of Madagascar. This is not for lack of open habitat; there are, and have been for approximately 25 million years (MacFadden,1997), plenty of grasslands in South America that are home to many diverse mammal species, but no primates. Several suggestions have been made for why the platyrrhines are restricted to an arboreal lifestyle: for example, some argue that the South American forest structure is more conducive to suspensory behavior with the aid of a prehensile tail (Emmons and Gentry,1983; Lockwood,1999), whereas others point to the high predator pressure on the ground for the relatively small-bodied taxa that are found in the New World (Di Fiore,2002; Campbell et al.,2005).

Recently, several fossil platyrrhines have been suggested to have been either terrestrial or “semiterrestrial” (Heymann,1998; Kay et al.,2002; MacPhee and Meldrum,2006; Kay,2010). Semiterrestrial has been defined as a separate locomotor category for Old World monkeys that have adaptations for transitioning between the trees and the ground by climbing and leaping as well as adaptations for quadrupedal running (Gebo and Sargis,1994; Anapol et al.,2005). A combination of terrestrial and arboreal traits such as relatively long distal limb segments, a long tail, and smaller body size would indicate membership in this group; an example in the Old World is Cercopithecus aethiops, the vervet monkey (Anapol et al.,2005). In their study of the Paralouatta postcranium, MacPhee and Meldrum (2006) compared several morphological features of the elbow joint, ankle joint, and digits to the same elements in living New World and Old World primates of various locomotor patterns. Because of its unique combination of skeletal features, which includes a retroflexed medial epicondyle and short straight phalanges, Paralouatta's postcranial skeleton was suggested to function more like that of a cercopithecine than any other platyrrhine. Whether this means that Paralouatta is directly analogous to the vervet monkey as another semiterrestrial species is not exactly clear. However, it is worth noting that Paralouatta, like Protopithecus, does not seem to be moving around in its environment in the same was as any extant ateline.

Spending at least some time on the ground has also been suggested for the oldest platyrrhine, Branisella boliviana (Kay et al.,2002). However, this was based on its high-crowned and heavily worn molars (the assumption being that grit in the terrestrial diet would wear teeth faster, necessitating higher crowns) combined with paleoenvironmental reconstruction of a more open habitat, as no postcranial remains are yet known for this taxon. Similar reasoning has been used to infer “scansorial” behavior from the high-crowned molars of the newly discovered Argentinian primate Mazzonicebus almendrae (Kay,2010). Protopithecus and its Pleistocene subfossil relative Caipora (Cartelle and Hartwig,1996; Hartwig and Cartelle,1996) provide new morphological information relevant to this topic. For these taxa, which are both represented by nearly complete skeletons, hypotheses about locomotor adaptations can be tested by looking directly at the joint surfaces themselves, instead of distant functional systems such as teeth.

Heymann (1998) has proposed a counterhypothesis stating that Protopithecus and Caipora were not as suspensory as Hartwig and Cartelle (1996) believed and instead would have practiced a “high degree” of terrestriality. Most of the evidence presented is intended to counter the pendulum model of brachiation as practiced by gibbons and siamangs as the main mode of locomotion for the fossils. For example, the intermembral indices of the fossils, when taking into account their extremely large body size for New World primates, make them look more similar to chimps or bonobos than to spider monkeys (see Fig. 1 in Heymann,1998). In fact, if the ateline portion of that figure is isolated, Protopithecus and Caipora seem to fall on a line with Alouatta and Lagothrix, the less-acrobatic atelines, whereas Ateles and Brachyteles are the more specialized suspensory outliers (Fig. 12). Although the intermembral index of Protopithecus, at 104, is within the range of values for Ateles (Erikson,1963), this value is more correctly seen as being in line with expectations for its body size based on a regression model restricted to the nonspecialized extant ateline genera. However, this is simply evidence against suspensory locomotion for the fossil, not for terrestriality. The intermembral index is a gross indicator of limb proportions and, hence, locomotor capabilities. However, conclusions about locomotion in fossil taxa should not be based on this value alone as it can mask subtler differences in limb structure and behavior among related taxa.

Figure 12.

Ateline portion of Fig. 1 from Heymann (1998). The regression line added through the nonsuspensory taxa, although driven by the marked contrast between species of two different size classes, still suggests that the intermembral index of Protopithecus and Caipora is on trend for their larger body size. Despite having high intermembral indices in the range of the more suspensory taxa, when their body size is taken into account, the fossils appear more similar to the slower, more quadrupedal atelines.

Just as limb proportions cannot provide a smoking gun for locomotor capabilities, neither can body size alone. Many other variables, both intrinsic to the animal-like joint surface morphology and extrinsic-like habitat type and social behavior, can influence how an animal moves (Remis,1995). However, in regards to suspensory locomotion, there is some evidence that siamangs, at 10–12 kg, are at the body size limit for biomechanically efficient brachiation (Preuschoft and Demes,1985). Even the largest New World suspensory taxa, Ateles and Brachyteles, do not exceed more than about 12 kg (Di Fiore and Campbell,2007) despite possessing a prehensile tail that provides more support and weight distribution throughout the canopy. Protopithecus, at an estimated 20–25 kg (Hartwig,1995b; Hartwig and Cartelle,1996; Halenar, 2011), is well above this threshold.

Although the limb proportions and large body size of Protopithecus may indicate a nonsuspensory mode of locomotion, this does not automatically mean that Protopithecus was terrestrial. The picture is most certainly more complicated than a strict dichotomy between arboreality and terrestriality. A primate probably cannot be an efficient brachiator at 20–25 kg; however, this does not make terrestriality its only other option. Other large-bodied primates, like orangutans and the truly giant subfossil “sloth” lemurs, have found other ways to move around successfully in the trees (e.g., Godfrey and Jungers,2003; Thorpe and Crompton,2006). Heymann (1998) pointed out that the great apes, while being designed for brachiating, do not actually brachiate and that most primates have a flexible locomotor repertoire, the full range of which might not be represented in their postcranial morphology. Analogy suggests that, perhaps, Protopithecus was not completely arboreal or terrestrial but could be better compared to a versatile great ape, such as the chimpanzee. This would just be one more example of convergence between atelines and hominoids (for a review, see Di Fiore and Campbell,2007).

Other evidence presented for a high degree of terrestriality in the fossil repertoire is largely reliant on circular logic: for example, if Protopithecus was terrestrial, its large body size would be a good defense against predators such as jaguars and other large cats. Heymann (1998) also brings up contradictory statements made by Cartelle and Hartwig (1996) about the paleoenvironmental reconstructions of the area. They appear to vacillate between an interpretation of the area as being more open and similar to the cerrado vegetation that exists in parts of Bahia and Minas Gerais today while also inferring that the existence of the large-bodied suspensory primates is evidence for a more closed forest habitat type. The paleoenvironment in the TBV and LS areas relevant to Protopithecus is hard to pin down, especially as the fossils have not been dated directly. Various lines of evidence suggest mixed habitat types in alternating wet and dry climates for the region through the late Pleistocene (Cartelle,1994; Auler et al.,2004,2006; MacFadden,2005). If the climate was more dry and habitat more open at the time the fossils were extant, the primates would have had an opportunity to spend more time on the ground. Even if the habitat was more closed, Heymann (1998) suggested that the New World forest structure, with its lack of lianas and more fragile branches (Emmons and Gentry,1983), cannot support a large-bodied suspensory taxon, even one with a prehensile tail. Both of these points, potential for predator defense and the possible existence of a more open habitat, are valid considerations but do not prove that Protopithecus was terrestrial.

As mentioned above in regards to “brachiation,” the terminology being used here, and elsewhere in the literature, can benefit from clarification. Even though results presented here do not ally Protopithecus with terrestrial Old World monkeys like some macaques and baboons, none have suggested that these extinct animals did not use the ground at all. Several other ateline species, while not “terrestrial” in their dominant locomotor pattern, use the ground in many ways during their daily activities. In the Atlantic Coastal forest, both Brachyteles and Alouatta guariba come to the ground, especially in more disturbed parts of the habitat, to cross cleared patches (Mourthe et al.,2007). Even Ateles, the most suspensory and highly reliant on an arboreal habitat, does come to the ground, most often for reasons related to specialized feeding behaviors such as drinking water during the dry season and geophagy at mineral licks (Campbell et al.,2005); some of these mineral licks are even in small caves (Link et al.,2011), an interesting point to consider regarding the taphonomy of the nearly complete Protopithecus skeleton from TBV. The majority of these behaviors are described in the literature as opportunistic and more properly called “ground use,” not “terrestrial locomotion,” and as such they may not leave an adaptive signal on a fossilized postcranial skeleton. There is a difference between being labeled a terrestrial primate in the manner of an Old World monkey, which is what previous authors have been suggesting for various fossil platyrrhines, and allowing for ground use as part of a more flexible locomotor repertoire that could be more correctly given a different categorization. Perhaps “semiterrestrial” could be used, but as with “semibrachiation,” lumping taxa into a catchall intermediate category can obscure distinctions between them and caution should be exercised; a description of a range of possible behaviors may be preferable to a single categorical label.


Despite the postcranial similarities creating functional links to Ateles and Brachyteles, Protopithecus is a member of the alouattin tribe, more closely related to extant Alouatta and fossils such as Stirtonia and Paralouatta (Hartwig and Cartelle,1996; Cooke et al.,2007; Rosenberger et al., in review). This designation is mostly based on cranial synapomorphies seen in the TBV specimen such as the strong temporal lines, posteriorly directed nuchal plane, large airorynchous face, and a relatively small brain and foramen magnum. Based on the results presented here, and considering their similar Atlantic Coastal forest habitats, an alternate hypothesis originally proposed by Hartwig (1995b) based on the LS postcrania could be revived suggesting that Protopithecus is actually more closely related to Brachyteles. There would be a size decrease and transition to a semifolivorous diet in the subsequent evolution of Protopithecus toward its closest modern relative, whether that is Alouatta or Brachyteles. However, if Protopithecus is more closely related to Brachyteles, the cranial morphology shared by Protopithecus and the alouattins would be convergent, an unlikely scenario given how very derived the alouattin condition is when compared with the atelins. The inclusion of Protopithecus in the alouattin tribe as a basal form that retains primitive aspects of its postcranial skeleton from the more suspensory common ancestor of both alouattins and atelins is a more parsimonious hypothesis that does not necessitate convergence in any part of the skeleton. Positing a medium-sized ateline last common ancestor with a prehensile tail used during feeding postures and a moderately suspensory mode of locomotion can help explain why the Protopithecus cranium is so similar to Alouatta but the postcranial skeleton examined here is not. The results presented here corroborate this description of the ateline last common ancestor, which was also one of the suggestions made by Jones (2008) in her study of the evolution of “brachiation” in atelines.

All extant genera in the subfamily have diverged away from that moderately suspensory common ancestor, with Ateles and Brachyteles becoming more specialized acrobats and Alouatta and Lagothrix becoming less specialized. For the Alouatta lineage, this evolutionary transition began with a size increase toward Protopithecus, a large-bodied animal that would need to decrease its reliance on acrobatic behaviors to preserve energetic efficiency. Next came a size decrease during which climbing and slow quadrupedalism became even more prevalent as the newly enlarged hyoid bone and an increase in percent of leaves included in the diet continued to create obstacles to acrobatic movement and enhance the energy-minimizing strategy that Alouatta follows today (Rosenberger and Strier,1989; Fig. 13). The fossil evidence presented here allows a fleshing out of the hypotheses of Schön Ybarra (1976,1984) regarding the influence of the enlarged hyoid on possible locomotor patterns. As sub-basal space decreased with decreasing body size over evolutionary time, the cranial base was forced to flatten and elongate to maintain the volume necessary for the enlarged hyoid. The smaller-bodied Alouatta ancestor with the newly unbalanced mass in its throat would have altered its locomotor repertoire accordingly toward the more slow, cautious quadrupedalism practiced by the extant species.

Figure 13.

Ateline cladogram modified from Jones (2008). The branching pattern and divergence dates for the extant taxa reflect results from molecular studies (Meireles et al.,1999; Collins and Dubach,2000a, b; Collins,2001; Cortés-Ortiz et al.,2003). Protopithecus provides support for these evolutionary transitions. Several steps not included in the original figure along the alouattin branch can be added, such as a size decrease and craniodental changes associated with an increase in folivory and use of howling behaviors. The exact order and timing of those changes are unknown; however, they must have occurred between the more primitive alouattin Protopithecus and the appearance of Paralouatta and Stirtonia, fossil taxa from the Miocene which are more similar to extant Alouatta cranial and dental morphology, respectively.

Protopithecus can thus be described as a taxon that includes both types of locomotion in its repertoire and is therefore positioned as transitional between the more suspensory ateline last common ancestor and the more quadrupedal alouattins. This means that the actual subfossil material that exists with its Pleistocene date is the remains of a platyrrhine “living fossil” (Delson and Rosenberger,1984). The divergence date for the alouattin lineage obtained from molecular clock studies is approximately 15.5 Ma, and the extant species of Alouatta begin their divergence from one another at approximately 7 Ma (Meireles et al.,1999; Collins and Dubach,2000a, b; Collins,2001; Cortés-Ortiz et al.,2003). It has been suggested that Protopithecus is part of a more primitive Miocene radiation (Rosenberger et al.,2009) and the same could be said of Paralouatta that is represented by Pleistocene material as well as the Miocene talus referred to Protopithecus marianae (Rivero and Arredondo,1991; MacPhee et al.,2003). Just because we have not yet uncovered older Protopithecus fossils, it does not mean that they were not there. The lack of specifically Alouatta-like traits in the forelimb reported here further supports a basal position for Protopithecus; it is unlikely that a late-evolving Pleistocene giant Alouatta would become more suspensory and not retain Alouatta-like aspects of the postcranium or dentition.


The general qualitative observations made above are backed up by the quantitative analyses performed here as the fossils are never grouped with extant primate leapers or terrestrial quadrupeds. The hypothesis presented by Heymann (1998) suggesting a high degree of terrestrial locomotion for Protopithecus is refuted by these results, although using the ground for specialized feeding behaviors or to cross short open patches in the habitat is still likely. Nothing in the morphology of these large monkeys would prohibit ground use. The multiple links to Brachyteles seen throughout the analyses vindicate the original nineteenth century functional interpretations of Protopithecus as a large member of this genus. The fossils are never grouped with Alouatta, reinforcing their mosaic nature and emphasizing the differences between the Protopithecus cranial and postcranial skeleton (also see Rosenberger et al., in review). However, there is no need to invoke homoplasy or convergent evolution, as was suggested when the TBV skeleton was discovered (Hartwig and Cartelle,1996); newer analyses suggest that the last common ancestor of atelins and alouattins was a more suspensory animal than previously thought (Jones,2008). This supports the hypothesis that Protopithecus is a primitive alouattin, moving toward the extant condition in regards to cranial adaptations, but lagging behind in retaining a forelimb more adapted for suspensory locomotion.


The author thanks Dr. Castor Cartelle of the Museu de Ciências Naturais at the Pontificia Universidade Católica de Minas Gerais in Belo Horizonte, Brazil, and Dr. Kim Aaris at the Universitets Zoologisk Museum in Copenhagen, Denmark, for access to the fossil material, and Eileen Westwig (AMNH), Linda Gordon (NMNH), and Dr. Leandro Salles (MN) for access to the extant primate collections. The author also thanks Drs. Jeffrey Laitman, Alfred Rosenberger, and Walter Hartwig for sponsoring this special issue and allowing her to participate. The author especially thanks Drs. Eric Delson, Michelle Singleton, Melissa Tallman, and Sergio Almécija for their support and very helpful comments during the revisions of the manuscript.