Nocturnal Visual Predation and the Evolution of Orbit Orientation and the Postorbital

The interspecific analyses indicate that the presence of a postorbital bar is correlated with greater orbital convergence in herpestids and pteropodids, and with increased orbital frontation in felids and herpestids. Therefore, moderate support is provided for the NVPH's prediction about orbital convergence, as well as our suggestion regarding the influence of orbital frontation. As both orbital parameters increase during euprimate origins (Cartmill, 1970, 1972, 1974, 1992; Fleagle, 1999; Martin, 1990; Simons, 1962), it is reasonable to infer that both changes in orbital morphology are implicated in the development of the primate postorbital bar (Ravosa et al., 2000a; see also later section). In fact, the coevolution of these orbital parameters characterized the origin of anthropoids and the evolution of a postorbital septum is also thought to dampen ocular movements (cf., Cartmill, 1980; Ravosa, 1991b,c, unpublished; Ross, 1995; Ross and Ravosa, 1993) (Table 4).

The allometry of orbital parameters also appears important to a consideration of postorbital bar formation. For example, orbital convergence increases with size in herpestids and pteropodids (Table 3) and larger taxa in each clade evince a higher occurrence of a postorbital bar (Figure 7). Due to the negative scaling of brain size across felids—a pattern common to mammalian clades (Gould, 1975; Martin, 1990; Shea, 1987)—smaller cats are more frontated and only such diminutive species tend to exhibit postorbital bars (Figure 8A) (Ravosa et al., 2000a). Therefore, orbital frontation in felids appears to be proportional to relative brain size, with larger, more anteriorly located frontal lobes displacing the anterior cranial base and superior orbital margins rostrally so that the orbital apertures are more vertical (Cartmill, 1970, 1972, 1980; Radinsky, 1968). While herpestids with postorbital bars are more frontated, the fact that they are not as encephalized as felids may explain the lack of allometric decreases in orbital frontation in this clade (Figure 8B) (Ravosa et al., 2000a).

Levels of orbital frontation are likewise elevated in more encephalized primates (Cartmill, 1980, 1992; Ravosa, 1991b,c, unpublished; Ross, 1995; Ross and Ravosa, 1993). Contrary to a recent suggestion (Heesy, 2005), added support for this structural pattern is provided by suborder comparisons of the primate growth data. Due to greater encephalization and increased basicranial flexion (Ross and Ravosa, 1993), anthropoids typically exhibit relatively elevated levels of orbital frontation throughout ontogeny (Table 4). Furthermore, all 12 primates exhibit age-related decreases in orbital frontation (Table 4)—a pattern which strongly belies the negative scaling of brain size and basicranial flexion common to the postnatal ontogeny of a wide variety of mammals (Gould, 1975; Lieberman et al., 2000; Martin, 1990; Shea, 1987).

One long-standing controversy regarding the craniodental adaptations and behavior of basal euprimates centers on the mammal clade(s) used to elucidate the functional underpinnings of an important euprimate synapomorphy (forward-facing orbits). On one hand, the NVPH emphasizes that felid-like nocturnal visual predation is critical for understanding the adaptive significance of increased convergence levels and binocular visual acuity during euprimate origins (Allman, 1977, 1982; Cartmill, 1972, 1974, 1992). Alternative scenarios regarding the evolution of the euprimate circumorbital region posit that better functional analogs are to be found among nocturnal frugivores such as pteropodids and didelphimorh marsupials (Crompton, 1995; Pettigrew et al., 1989; Rasmussen, 1990; Sussman, 1991; Sussman and Raven, 1978). While the principal source of disagreement centers on the extant clade(s) selected to elucidate the functional underpinnings of orbital character states, this unresolved debate is further complicated by the lack of a broad-based empirical analysis of factors posited to influence variation in orbital orientation. This is especially surprising given that elevated levels of orbital convergence are also purportedly linked to the presence of relatively smaller orbital diameters (Cartmill, 1980), so that convergence increases with skull size due to the negative scaling of eye/orbit size (Kay and Cartmill, 1977; Martin, 1990).

The recent discovery of an exceptionally well-preserved Paleocene plesi-adapiform (Carpolestes) with a unique constellation of skeletal features has rekindled debate regarding the patterning of morphological and adaptive transformations during the origin of archaic (Plesiadapiformes) and modern (Euprimates) primates (Bloch and Boyer, 2002, 2003; Kirk et al., 2003; Sargis, 2002). In positing that a series of manual and pedal features shared between Carpolestes and basal euprimates are homologous and derived (rather than simply a case of functional convergence), Bloch and Boyer (2002, 2003) argue that grasping appendages and terminal-branch feeding preceded the evolution of increased levels of orbital convergence and stereoscopic visual acuity characteristic of the earliest modern primates. With Carpolestes reconstructed as frugivorous (and having low levels of orbital convergence), grasping extremities in carpolestids and the ancestors of euprimates are inferred to be adaptations for terminal-branch foraging on fruits, flowers, and buds (Bloch and Boyer, 2002, 2003; Sargis, 2002). Accordingly, the phylogenetic independence of grasping and forward-facing orbits is incompatible with the NVPH's version of euprimate origins (Cartmill, 1972, 1974, 1992; Kay and Cartmill, 1974, 1977; Kirk et al., 2003). Instead, grasping capabilities in basal euprimates are interpreted as exaptations for nocturnal visual predation, supporting an alternative scenario regarding the sequence of acquisition and function of important primate postcranial synapomorphies (cf., Rasmussen, 1990; Sussman, 1991). Bloch and Boyer's (2002, 2003) study also suggests that grasping adaptations occurred prior to the emphasis on leaping behaviors as a component of the locomotor repertoire in basal euprimates (Dagosto, 1988; Szalay, 1973, 1981; Szalay et al., 1987).

In differentially focusing on the phylogenetic implications of the postcra-nial features, Bloch and Boyer (2002, 2003) are equivocal regarding an unresolved claim that nocturnal arboreal frugivores are the most appropriate functional analogs for understanding the evolution of euprimate visual acuity and forward-facing orbits (Rasmussen, 1990; Sussman, 1991; also Crompton, 1995; Pettigrew et al., 1989). Three sets of comparisons bear on the unresolved questions regarding the functional significance of forward-facing orbits. First, nocturnal arboreal frugivores, such as pteropodids, exhibit relatively lower amounts of orbital convergence similar to that in tupaiids, ple-siadapiforms, dermopterans, and herpestids (Figure 10). Apart from representing the plesiomorphic state for euprimate sister taxa and presumably all eutherians, more divergent orbits characterize mammals with widely disparate activity patterns, including diurnal predators, as well as nocturnal and diurnal foragers (Ravosa and Savakova, 2004; Ravosa et al., 2000a). In contrast, the earliest euprimates are similar to nocturnal primate and felid faunivores in uniquely possessing enlarged (Kay and Cartmill, 1977; Martin, 1990) and relatively convergent orbits (Figure 10). Further comparisons between similarly sized nocturnal omomyids and extant primate nocturnal faunivores indicate no significant group differences in levels of orbital convergence (Mann-Whitney U test among species means of 12-26 mm for palate length, p = 0.240). Therefore, the derived presence of relatively larger eyes and forward-facing orbits in the first modern primates suggests that consequent increases in stereoscopic acuity and image clarity at close range were functionally linked to nocturnal stalking and capturing of mobile prey (Allman, 1977, 1982; Cartmill, 1972, 1974, 1992; Ravosa and Savakova, 2004; Ravosa et al., 2000a).

More phylogenetically restricted comparisons further demonstrate the lack of an association between nocturnal frugivory and marked orbital convergence (Ravosa and Savakova, 2004). In procyonids, the more frugivorous, nocturnal kinkajou exhibits an amount of orbital convergence equivalent to its nocturnal sister taxa (Figure 11A). Interestingly, kinkajous and other nocturnal predators display relatively greater levels of convergence than diurnal predators (MannWhitney U test of residuals from the procyonid LS line, p = 0.001). This suggests that the small vertebrate and insect component of the kinkajous diet—a proclivity shared with all nocturnal procyonids (Ewer, 1973; Nowak, 1999; Zeveloff, 2002), underlies variation in orbital form within and between nocturnal and diurnal members of this clade. Likewise, the nocturnal faunivore

Ptilocercus exhibits relatively greater convergence than diurnal tupaiid sister taxa (Figure 11B; studentized residual of 5.974 from the tupaiid LS line). In didel-phimorphs of similar size, four nocturnal faunivores of the genera Philander, Chironectes, and Metachirus (mean = 53.5°; range = 50.4-57.8°)—all possess significantly more convergent orbits than three nocturnal arboreal frugivores of the genus Caluromys (mean = 42.4°; range = 42.0-43.0°) (Mann-Whitney U test among species means of 30-38 mm for the nasion-inion chord, p = 0.034). Compared to other largely diurnal herpestids, the diminutive nocturnal predator Dologale also displays an elevated level of convergence (Figure 10; studentized residual of 3.057 from the herpestid LS line). In support of the NVPH, these independent analyses clearly indicate that the presence of forward-facing orbits in extant and extinct mammals are related to an adaptive strategy of nocturnal visual predation.

One can nonetheless identify apparent support for the relationship between elevated orbital convergence and frugivory. In nocturnal pteropo-dids, larger-bodied species are more frugivorous and exhibit greater convergence than smaller, more insectivorous sister taxa (Nowak, 1999; Table 3). This pattern also characterizes orbital and dietary variation between adapids and the smaller-bodied omomyids (Figure 10). However, there are several reasons why these correlations are spurious and due rather to the independent scaling of dietary preference and orbit orientation. Controlling for activity cycle, and given the preponderance of faunivory (Ewer, 1973; Nowak, 1999), orbital convergence nevertheless increases with size in diurnal her-pestids, nocturnal procyonids, diurnal tupaiids, and felids (Table 3; Figures 10 and 11). Postnatal development in three lorisiformes and three lemuri-formes, taxa arguably most similar to basal euprimates in skull form, is uniformly characterized by size-related increases in convergence (Table 4). Moreover, increases in orbital convergence among these six strepsirrhines continue long after the early postnatal shift to weaning (inferred from the eruption of the first permanent molar - Smith et al., 1994) and relatively invariant adult feeding and chewing behaviors (see Watts, 1985 and review in Ravosa and Hogue, 2004). Such findings, particularly the ontogenetic evidence, offer strong empirical support for the argument that the allomet-ric patterning of orbital convergence is simply a structural consequence of the negative scaling of orbital aperture diameter both during ontogeny and across a size series of close relatives (Cartmill, 1980; Ravosa and Savakova, 2004).

Contrary to previous assertions (Crompton, 1995; Pettigrew et al., 1989; Rasmussen, 1990; Sussman, 1991; Sussman and Raven, 1978), phylogeneti-cally and allometrically controlled analyses highlight the unmistakable imprints of nocturnality and visual predation on the evolution of the skull and sensory system in the first primates of modern aspect (Allman, 1977, 1982; Cartmill, 1972, 1974, 1992; Kay and Cartmill, 1974, 1977; Kirk et al., 2003; Ravosa and Savakova, 2004; Ravosa et al., 2000a). Since basal euprimates perhaps were no larger than the 100-g Carpolestes (Bloch and Boyer, 2002, 2003; Fleagle, 1999; Martin, 1990, 1993; Sargis, 2002), allometry cannot be invoked to explain the derived presence of relatively higher levels of orbital convergence in this clade (as would be the case if a descendant were larger in body size than its ancestor). Thus, while the evidence of Carpolestes supports earlier studies regarding the more mosaic sequence of acquisition of eupri-mate synapomorphies (Rasmussen, 1990; Sussman, 1991), accompanying ecomorphological models positing the importance of nocturnal arboreal fru-givory as a basis for elevated orbital convergence are unfounded (Ravosa and Savakova, 2004). Conversely, although the NVPH's explanation for the functional significance of increased visual acuity is well supported, because certain grasping features arguably predate the origin of Euprimates (Bloch and Boyer, 2002, 2003; Sargis, 2002), the shift to predatory and leaping behaviors in this clade occurred in an ancestor already frequenting a terminalbranch milieu (Dagosto, 1988; Szalay, 1973, 1981; Szalay et al., 1987). This is at odds with the NVPH's adaptive scenario for the coevolution of a wider range of euprimate cranial and postcranial synapomorphies from an ancestor with minimal grasping capabilities (Cartmill, 1972, 1974, 1992; Kay and Cartmill, 1974, 1977; Kirk et al., 2003; Soligo and Martin, 2006).

Phylogenetic and functional similarities between the pteropodid and primate skull have been further disputed. For instance, putative neuroanatomi-cal synapomorphies of the visual system (Pettigrew et al., 1989) have been refuted by more comprehensive analyses (Johnson and Kirsch, 1993; Thiele et al., 1991). Our data also point to different influences on postorbital bar formation in primates and megabats: frontation and convergence in the former clade and only convergence in the latter. Moreover, whereas primates first evolved a postorbital bar at very small body sizes, only larger megabats exhibit this derived condition. Contrary to Pettigrew et al. (1989), it is unlikely that the postorbital bar of megachiropterans and primates is a synapomorphy (Ravosa et al., 2000a).

In addition to highly convergent orbits linked to nocturnal visual predation, basal primates and felids share the following features/trends: neural specializations of the visual system (Allman, 1977, 1982; Cartmill, 1970, 1972, 1974, 1992); relatively larger brains (Figure 9); enlarged orbits related to nocturnality (Ravosa et al., 2000a); postorbital bar development at small skull sizes (Figure 8A); and somewhat reduced olfactory bulbs presumably associated with a diminished emphasis on olfaction (Ravosa et al., 2000a using data from Gittleman, 1991). Such evidence supports claims of the NVPH that many early primate cranial adaptations approximate those of felids (Allman, 1977, 1982; Cartmill, 1970, 1972, 1974, 1992; Kay and Cartmill, 1974, 1977; Ravosa et al., 2000a).

As alluded to the earlier section, several of these features appear to be uniquely critical for explaining shared patterns of circumorbital covariation in the felid and basal euprimate skull (i.e., allometric decreases in orbital frontation and the evolution of a postorbital bar at small sizes). Felids are similar to the first (omomyid-like) euprimates in exhibiting greater encephalization likely related to a nocturnal visual predation strategy (i.e., a relatively larger visual cortex) (Barton, 1998; Cartmill, 1992). Felids and omomyids are also alike in exhibiting relatively larger, more convergent orbits associated with a nocturnal predatory lifestyle (Figure 10) (Cartmill, 1970, 1972, 1974, 1992; Covert and Hamrick, 1993; Heesy and Ross, 2001; Kay and Cartmill, 1977; Kirk and Kay, 2000; Martin, 1990; Noble et al., 2000; Ravosa et al., 2000a). Increases in the relative size of these adjacent structures appear to have created a spatial packing problem in which the orientation of the orbits (and thus supraorbital rims) are more highly affected by the position of the anterior cranial fossae and, in particular, relative brain size and shape (cf., Cartmill, 1970, 1972, 1980; Lieberman et al., 2000; Radinsky, 1968; Ravosa, 1991b,c, unpublished; Ross, 1995; Ross and Ravosa, 1993). Thus, due to the negative allometry of neural and orbital size, this structural constraint is especially pronounced in smaller taxa—exactly the range of skull sizes in which felids and basal euprimates exhibit greater orbital frontation and develop postorbital bars (Ravosa et al., 2000a).

In sum, our reformulation of the NVPH uniquely emphasizes the role of encephalization on patterns of covariation in circumorbital form and function during the origin of euprimates—due to increased relative brain size, greater orbital frontation (and in turn postorbital bar formation)—is a structural consequence of a nocturnal visual predation adaptive strategy. This explanation does not preclude the possibility that increased relative brain size among basal euprimates is also related to their unique combination of arboreality, preco-ciality, and small body size (Shea, 1987, this volume).

Although the comparative data indirectly support the argument that more convergent and/or more frontated visual predators or foragers may require rigid lateral orbital margins, obviously, the most direct test of this prediction is to investigate experimentally the functional relations among orbital orientation, ocular oscillations, and postorbital bar formation (see also Heesy et al., this volume). In this regard, we briefly discuss the implications of the galago in vivo data vis-à-vis the NVPH and the evolution of the primate postorbital bar.

Perhaps the best insight offered by the galago strain data centers on the significant disparity between loading patterns along the working and balancing sides of the facial skull. As predicted by the NVPH, the galago WS postorbital bar encounters posteroinferiorly directed tension during mastication. However, opposite the NVPH, tensile strains at the BS postorbital bar are oriented posterosuperiorly (Table 2). This indicates that only the WS lateral orbital margin in basal primates was likely to have been compressed more along the optical axis instead of along the plane of the orbital aperture as inferred for putative sister taxa such as plesiadapiforms.

Other differences exist in galago WS/BS loading patterns. Plesiadapiforms and basal primates had unfused mandibular symphyses (Beecher, 1977, 1979, 1983; Ravosa, 1991a, 1996, 1999; Ravosa and Hylander, 1994; Ravosa et al., 2000c), and it is now well documented that primates with this character state recruit less BS jaw-adductor force (especially the transverse component) and exhibit correspondingly low BS corpus strains during mastication (Hylander, 1979a,b; Hylander et al., 1998, 2000, 2004; Ravosa and Hogue, 2004; Vinyard et al., this volume). This explains why galago peak strains are much higher along the WS postorbital bar and corpus (Table 1; Figures 2 and 3) (Ravosa et al., 2000a,b). In taxa with unfused symphyses and only a postorbital ligament, a galago-like recruitment pattern would result in an asymmetrical loading pattern in which elevated levels of orbital/ocular deformation occur along the chewing side of the face. As BS ocular acuity would be minimally affected, the corresponding asymmetry in visual disruption presumably poses deleterious consequences for effective binocular stereoscopic acuity, and this constitutes another reason why a rigid postorbital bar is important in the evolution of a nocturnal visual predation strategy in basal primates (Ravosa et al., 2000a).

There is also limited evidence from both extant primate suborders indicating allometric increases in the cranial strain gradient. During powerful mastication, a 2-kg Otolemur crassicaudatus exhibits lower peak-strain magnitudes at the dorsal interorbit and WS postorbital bar than a 1-kg O.garnettii (Ravosa et al., 2000a,b). In papionins, 14-kg baboons possess lower interorbital strain levels during chewing than 4-kg macaques (Hylander et al., 1991a). This positive allometry of the cranial strain gradient highlights a possible design criterion requiring the circumorbital region of larger forms to be increasingly overbuilt for masticatory loads—an interpretation consistent with the positive scaling of primate postorbital bar and supraorbital torus proportions (Hylander and Ravosa, 1992; Ravosa, 1988, 1991a-c; Vinyard and Smith, 1997, 2001). Such size-related decreases in circumorbital strains are also in contrast to an apparent pattern of strain similarity along the WS mandibular corpus across an interspecific primate size series (Hylander, 1979a; Hylander et al., 1998; see also Hylander, 1985; Vinyard and Ravosa, 1998, regarding stress similarity).

Given that the cranial strain gradient appears to scale positively, smaller taxa and individuals experience relatively higher strains along the WS lateral orbital margins that are variably lower than WS mandibular levels. This scaling pattern suggests that the first basal primates, which were quite small in body size, would have experienced relatively higher levels of WS ocular deformation coupled with a pronounced asymmetry in the amount of deformation between WS and BS lateral orbital margins.

Finally, as a close correspondence between orbital and ocular size characterizes only small-bodied species (Kay and Cartmill, 1977; Schultz, 1940), deformation of the lateral orbital margins is more likely to compromise ocular acuity in the earliest primates as they were quite small (100-300 g: Fleagle, 1999). Therefore, the development of a rigid postorbital bar may have been especially critical for maintaining high-effective levels of nocturnal visual acuity in small animals.

As anthropoids are further derived in having an orbital cavity mostly walled-off from the temporal fossa by a postorbital septum, data on strep-sirhine masticatory and circumorbital function is important for understanding anthropoid evolution. For instance, the first anthropoid with a bony postorbital septum was small (700-900 g) and had only partial symphyseal fusion (Simons, 1989, 1992) and presumably a jaw-adductor activity pattern like galagos (Ravosa, 1999; Ravosa et al., 2000a-c). Thus, our argument regarding circumorbital-loading asymmetry and negative scaling of ocular size also may apply to the origin of the postorbital septum in stem anthropoids, especially since WS/BS asymmetry characterizes the temporalis and masseter of primates with unfused joints (Vinyard et al., this volume). This in turn suggests that functional and adaptive investigations of the anthropoid postorbital septum should account for the fact that the circumorbital region of extant anthropoids is loaded differently than in basal forms. Moreover, such studies should integrate data on orbital orientation in stem taxa so as to better estimate the extent to which the eye was disrupted along the orbital axis, as well as information on the position of the anterior temporalis versus the orbital contents so as to better estimate the proximity and direction of disruptive adductor forces.

In terms of the NVPH and postorbital bar function, there is recent evidence that in cats with high levels of orbital convergence and frontation, a complete bony bar does not prevent ocular movements during bilateral tetanic stimulation of the jaw adductors (Heesy et al., this volume). While this experimental information suggests that visual acuity may be compromised in mammals with postorbital bars, such ocular movements are documented under conditions hardly mirroring the biological role of alert cats during biting and chewing (cf., Gorniak and Gans, 1980). In fact, to evaluate if a postorbital bar does or does not serve to stiffen the lateral orbital margins, and thereby influence ocular movements, it is critical to determine if the magnitude of ocular movement in alert animals with and without bony bars is sufficient enough to inhibit effective stereoscopic vision. Furthermore, perhaps during the normal recruitment of the jaw-closing muscles, a neuromuscular mechanism may exist to counter ocular oscillations during mastication (thus obviating the need for a rigid postorbital bar). Finally, behavioral data on the extent to which arboreal taxa actively forage and/or predate while simultaneously chewing would directly test a fundamental premise of the NVPH regarding food procurement, activity level, visual acuity, and postorbital bar function.

0 0

Post a comment