Do The Functionalbiological Role Attributes Of The Traits As A Whole Constitute A Cohesive Story

Can this constellation of features be explained by a single selective factor or a set of selective factors arising from a way of life? The answer is possibly not. There seem to be at least two separate selective regimes influencing the primate postcranial morphotype. One is a "substrate size" component that is reflected in the grasping extremities, and the other is a "mode of displacement" component, which is reflected in the leaping features. Of course, these two components may be related (e.g., they evolved concurrently as part of a single mode of positional behavior) but there is at least the possibility that they may have independent histories (i.e., one preceded another in the "origin" of Primates).

The locomotor part of the NVP hypothesis does not recognize the role of leaping in ancestral primates. If the importance of leaping to the MRCA is established, can it be accommodated in this otherwise comprehensive model? On one hand, Fleagle and Mittermeier (1980) have associated leaping with both small body size and use of the more discontinuous understory, both of which are elements of the NVP model. Cartmill (1974a) cites Microcebus as a potential model for an ancestral primate. Although it may employ deliberate quadrupedalism when foraging for insects, its dominant locomotor mode is described as rapid scurrying and leaping (Gebo, 1987; Walker, 1974). As the example of Tarsius shows, NVP as an ecological strategy is certainly not incompatible with significant use of leaping or vertical supports.

On the other hand, leaping was certainly not stressed in the original formulation of the hypothesis, and continues to be criticized as an explanatory factor for primate origins (Cartmill, 1982; 1992; Schmitt and Lemelin, 2002, articles in this volume). Indeed, part of Cartmill's purpose was to critique hypotheses that invoke leaping to explain the presence of nails on lateral digits and the evolution of orbital convergence, and his model requires nothing more than a deliberate, slow moving quadrupedal animal (Cartmill, 1974a). If slow, deliberate quadrupedalism is meant as the predominant mode of displacement of the ancestral primate, and not just a locomotor mode employed during insect foraging, then this model and GL do seem contradictory. Are there other factors that might explain the incompatibilities between this aspect of the NVP model and GL or that may allow a reconciliation of the two models?

Are the Biological Roles of Features Misidentified?

It is of course possible that the function or biological role of the features supporting GL (or NVP) are misidentified; perhaps the features cited above are not related to leaping. As noted above, forms have numerous functions; are there other possible explanations for the primate features? For example, long hindlimbs (and elements of the limb) are useful for:

(1) Leaping, as argued in an earlier section;

(2) Increasing stride length, and therefore speed of quadrupedal locomotion. Primates do generally have longer stride lengths than other mammals (Larson, 1998). Demes et al. (1990) explain the long limbs of lorises as an adaptation to increase speed without the need to increase stride frequency, which would decrease stability on small branches. On the other hand, most cursorial mammals lengthen the distal elements of the limb, especially the metatarsus, unlike primates where the femur, tibia, or tarsus are elongated (Garland and Janis, 1993; Hildebrand, 1982), (Some lizards do lengthen proximal elements (Garland and Losos, 1994)). Alexander (1995) has shown that for leaping, the distribution of mass in the limbs is less important than total limb mass, so that unlike running, mass does not have to be concentrated in the short proximal element. Primates, of course, do not exhibit any of the other postcranial characteristics of cursors. From a mechanical and comparative perspective, the hypothesis that hindlimbs elongated to enhance speed does not seem to be a better alternative than that they elongated to enhance leaping, and in any case speedy locomotion itself is inconsistent with the NVP slow quadruped model;

(3) Mammals that glide have longer femora (and humeri) than related nongliders (Runestad and Ruff, 1995) because long limbs are biome-chanically important for maintaining aspect ratio and patagial loading. Since there are no other indications of gliding in primates, this also seems an unlikely explanation for their long femora and tibiae; and

(4) A general indication of "arboreality." Polk et al. (2000) have shown that arboreal members of the Carnivora and Rodentia have longer femora (and humeri) than terrestrial members of the same lineage. Therefore, long limbs may be an indication of some other aspect of arboreality. Three (nonmutually exclusive) possibilities here are: (a) climbing (Preuschoft and Witte, 1991; Preuschoft et al., 1995); (b) bridging (Polk et al. (2000); and (c) long stride length contributing to the compliant gait that facilitates movement on small supports (Larson et al., 2000, 2001; Schmitt, 1998; 1999).

None of these hypotheses, alone or in concert, seems adequate to explain the extreme degree of lengthening of the primate hindlimb compared to other arboreal mammals, including Caluromys, which does not appear to have hindlimbs any longer than its terrestrial relatives (Figures 2, 3A). Although humeri are also longer in primates than in arboreal carnivores or rodents (Polk et al., 2000), the difference in femur length is more marked. Primates do differ significantly from tupaiids and plesiadapids in hindlimb length, but not in humeral or forelimb length; in fact plesiadapids appear to have humeri that are as long or longer than primates (Figure 7, Table 3).

« Primates □ Scandentia a Dermoptera

• Plesiadapidae

♦ Fossil primates x Marsupials

+ Cheirogaleidae

Tupaiids y = 0.31x + 1.85 R = 0.98

ln body mass

o Primates □ Scandentia a Dermoptera

♦ Fossil primates

• Plesiadapidae x Marsupials

+ Cheirogaleidae

ln body mass

Figure 7. (A) Regression (OLS) of ln humerus length on ln body mass. Conventions as in Figure 2. (B) Regression (OLS) of ln forelimb length (humerus + radius) on ln body mass. Conventions as in Figure 2.

If primates were shown to climb vertical supports or bridge more often or more efficiently than other arboreal mammals there could be some validity to these hypotheses. Among primates, however, the lorises, likely the most frequent climbers and bridgers, do not have longer hindlimbs than other primates, while leaping primates do (Connour, 2000; Runestad, 1997). Prosimians that use vertical supports more frequently (indriids, tarsiers) also have relatively longer femora, but these taxa are also leapers, complicating attempts to separate these factors.

Although it is most certainly naive to think that a single selective factor is involved in the origin or evolution of any trait, the combined evidence of hindlimb morphology (both limb elongation and joint shape) is strongly supportive that the mechanical requirements of leaping have driven, at least in part, the evolution of the primate hindlimb.

Are these Features Part of the Primate Morphotype?

It is also possible leaping features were not part of the primate morphotype, but evolved after the MRCA only in a clade of prosimian primates, or independently in several different groups of primates. Many of the features listed as primate apomorphies are more strongly expressed in notharctines, omomyids, and lemurs and less so in anthropoids. Anthropoids, for example, do not differ greatly from tupaiids in talar height (Table 3). If we accept the traditional phy-logeny where anthropoids evolved from some sort of prosimian, the hypothesis that leaping features were part of the euprimate morphotype is strongly supported. However, if a phylogeny in which prosimians are the sistergroup of anthropoids and in which anthropoids exhibit primitive primate postcranial morphology is accepted (Ford, 1988, 1994; Godinot, 1992; Godinot and Beard, 1991; Godinot and Jouffroy, 1984), the hypothesis is less certain (Figure 8).

These propositions are problematic on several levels. Anthropoids, although not as specialized as most prosimians, still differ significantly from tupaiids or plesiadapids in several critical features, particularly hindlimb length, suggesting that this was a feature present in the MRCA of all primates. The convergence hypothesis also requires the independent development of leaping and the same leaping features in several groups of primates, an idea made unlikely by the presence of these features in adapids, omomyids, and eosimiids (Gebo et al., 2000), the likely stem groups for the living clades. Furthermore, if the grasping complex, which differs significantly in morphological detail between anthropoids and prosimians (Szalay and Dagosto, 1988), is considered to be

Figure 8. Two versions of primate phylogenetic relationships. In (A) anthropoids are derived from an ancestor with prosimian-like ancestry. Leaping features (gray box) are characteristic of the primate MRCA. In (B) anthropoids are the sistergroup of prosimians, leaping features are not characteristic of the MRCA, but of the prosimian ancestor.

Figure 8. Two versions of primate phylogenetic relationships. In (A) anthropoids are derived from an ancestor with prosimian-like ancestry. Leaping features (gray box) are characteristic of the primate MRCA. In (B) anthropoids are the sistergroup of prosimians, leaping features are not characteristic of the MRCA, but of the prosimian ancestor.

a highly corroborated synapomorphy of primates, then isn't it inconsistent to simultaneously argue that the morphology of the knee joint, hip joint, or ankle joint, which differ less in detail, are convergences?

Origin of Primates as a Process Not an Event

The most likely explanation for the apparent discrepancy between the locomotor component of NVP and GL is that any effort to relate all derived primate characters to a single way of life is oversimplified. Surely the "Origin of Primates," like that of any other group, was not a single event (i.e., a single speciation), but a series of events occurring over a long period of time (Figure 9; see also Rasmussen and Sussman, this volume, and the distinction between "origin" of a group and the "last common ancestor" in Soligo et al. (this volume)). The presence of a long list of synapomorphies for the MRCA is an illusion attributable to the lack of a good fossil record or the continued existence of intermediate forms (Gebo and Dagosto, 2004). The MRCA had a way of life, but all of its features may not be related to that; some were acquired by ancestors along the stem as a result of their ways of life. The real problem is to identify the sequence of "ways of life," to determine how earlier ones may have channeled later ones or constrained the morphological solutions available to descendants (Cartmill, 1982; Rasmussen and Sussman, this volume; Szalay and Sargis, 2001).

Figure 9. The "Origin of Primates" can be interpreted as either the MRCA with N synapomorphies, or as the stem lineage (dashed arrow) with numerous stages, the earliest of which has only one or few of the "primate" synapomorphies. Each of the stages along the stem may have very different ways of life which constrain paths of evolution of future stages, or constrain the morphological response to new behaviors. Modified from Gebo and Dagosto (2004).

Figure 9. The "Origin of Primates" can be interpreted as either the MRCA with N synapomorphies, or as the stem lineage (dashed arrow) with numerous stages, the earliest of which has only one or few of the "primate" synapomorphies. Each of the stages along the stem may have very different ways of life which constrain paths of evolution of future stages, or constrain the morphological response to new behaviors. Modified from Gebo and Dagosto (2004).

Given the current evidence, the most likely scenario is that grasping either preceded or evolved concurrently with leaping (Figure 10). Both tree shrews, especially Ptilocercus, and plesiadapiforms exhibit features of the tarsus and hallucal-metatarsal joint which indicate some independence of the hallux, incipient grasping abilities, and ability to invert the foot greater than in the primitive eutherian state (Beard, 1991; Bloch and Boyer, 2002; Sargis, 2001; Szalay and Dagosto, 1988; Szalay and Drawhorn, 1980). Some rudimentary form of grasping was already present in the ancestral archontan. The elaboration of the primate grasping appendage with its increased hallucal independence, habitual abducted stance, and replacement of hallucal claw by nail indicates more frequent or critical use of relatively small branches (Cartmill, 1974a). Didelphids and phalangerids provide interesting parallels for this phase of postcranial evolution (Rasmussen and Sussman, this volume). The elongated hindlimbs and the stabilized knee, ankle, and hallucal-metatarsal joints evolved as leaping became an integral and frequent part of the positional behavior of the ancestral primate.

Figure 10. Phases in the evolution of the primate postcranium. A primitive form of grasping, present in the archontan ancestor is replaced by more derived hallucal grasping (opposable hallux, hallucal nail). Leaping either follows this, or evolves concurrently with it.

The primary difference between the Szalay-Dagosto point of view and that of Cartmill, Lemelin, and Schmitt (Cartmill, 1974a; Schmitt and Lemelin, 2002, articles in this volume), is that we believe the evidence clearly shows that the adoption of leaping and its morphological correlates is a shared-derived feature of all crown-group primates and therefore must have occurred before or in the MRCA, while the NVPers believe their evidence indicates that some primate locomotor apomorphies are most likely to have evolved in a more slowly moving quadruped and therefore that leaping and morphological changes associated with it must evolve convergently after the appearance of the MRCA. The two models could be reconciled if the NVP hypothesis does not apply to the MRCA, but is emphasizing events earlier in the transition from archontan to Primate; the "leaping" part of GL, later events.

The primary point to take away from this essay is that there are derived features of the primate skeleton that are most parsimoniously interpreted as being present in the MRCA, and are not easily explained simply by grasping small supports, by deliberate quadrupedalism on horizontal branches, or even by rapid scurrying and bounding by clawless animals on horizontal supports. Leaping remains the best explanation for the biological role of these features.

Marsupials such as Caluromys mimic some aspects of early stages of primate evolution, but do not provide a comprehensive model for the postcranium of the MRCA of Primates, since they lack all of the key leaping-related features (see also Szalay, this volume). Models for the Origin of Primates are incomplete if they do not account for or incorrect if they cannot accommodate the leaping component of the primate morphotype.

Was this article helpful?

0 0

Post a comment