The postorbital bar and septum are circumorbital structures that are important to adaptive hypotheses for the origins of primates and haplorhines, respectively. All primates possess complete postorbital bars, bony arches formed by processes of the frontal and zygomatic bones that encompass the lateral aspect of the eye. Postorbital septa, bony walls formed by the frontal, zygomatic and alisphenoid bones, walling off the orbit from the anterior temporal fossa, are limited to tarsiers and anthropoids.

Numerous functional hypotheses have been advanced for postorbital bars and septa. Many of these hypotheses can easily be rejected (Cartmill, 1970, 1972, 1980; Ravosa, 1991a,b; Ravosa et al., 2000a,b; Ross, 1994, 1995a,b, 1996, 2000, 2001; Ross and Hylander, 1996; see Heesy, 2003). Cartmill (1970, 1972, 1980; see also Collins, 1921) suggested that in therian mammals with large eyes, relatively small temporal fossae, and derived orbit

Christopher P. Heesy • Department of Anatomy, New York College of Osteopathic Medicine, Old Westbury, NY 11568 Callum F. Ross • Organismal Biology & Anatomy, University of Chicago, 1027 East 57th Street, Chicago, IL 60637 Brigitte Demes • Department of Anatomical Sciences, Health Sciences Center, Stony Brook University, Stony Brook, NY 11794-8081

convergence (orbits facing in the same direction), the plane of the bony orbit would deviate from the "plane" of the temporal fossa. Cartmill proposed that increasing orbital convergence "drags" the anterior temporalis muscle and temporalis fascia from a posterior position to a lateral position (Figure 1). In such taxa, including Primates, Cartmill suggested that contractions of the masticatory musculature, particularly the anterior temporalis muscle, would be likely to distort the lateral orbital margin, potentially disrupting oculomotor precision. Replacement of the postorbital ligament with an osseous postorbital bar should stiffen the lateral bony orbit and prevent oculomotor disruption.

Figure 1. Cartmill's hypothesized effects of orbit orientation on postorbital ossification.

In many mammals, the postorbital ligament and temporalis fascia sit posterior to the eye (A). Cartmill hypothesized that increasing orbital convergence drags the ligament anterolaterally (B), where it can be deformed during anterior temporalis contraction and temporalis fascia tension (indicated by arrows). Preventing disruptive eye movements during mastication requires the evolution of postorbital processes or a bar to prevent the fascia from encroaching on the eye (C).

Cartmill's hypothesis predicts that increasing the angular or planar deviation between the orbit and the temporalis fascia should lead to increased replacement of the postorbital ligament with postorbital processes and other stiffening structures in order to prevent disruption of oculomotor coordination. Taxa such as megachiropterans and small-bodied herpestid carnivorans with slightly lesser degrees of orbital convergence than most strepsirrhines (Cartmill, 1970) have well-developed postorbital processes, corroborating Cartmill's hypothesis (Noble et al., 2000; Ravosa et al., 2000a). Ravosa and colleagues (Noble et al., 2000; Ravosa et al., 2000a), in an extensive analysis of megachiropterans, herpestid, and felid carnivorans, found that frontation (relative vertical orbit orientation) as well as relative encephalization and relative orbit size (a factor identified by Cartmill) are contributing factors to the evolution of bars among these taxa. The association between bone strain patterns and anterior temporalis contraction has not been directly evaluated in strepsirrhines. However, in vivo bone strain data collected on the lateral aspect of the postorbital bar of Otolemur indicate that it experiences nontrivial levels of strain during mastication (Ravosa et al., 2000a,b). If the strepsir-rhine skull is twisting (e.g., Greaves, 1985), or the palate is "rocking" on the interorbital region (Ross, 2001; Ross and Hylander, 1996) during mastication, the bar would be deformed. These results suggest that if a postorbital ligament were in place of the bar, the lateral orbit would experience substantial deformation.

The evolution of postorbital septa in haplorhines is related to Cartmill's hypothesis for postorbital bars in the following ways. Anthropoids have higher orbit convergence and frontation than strepsirrhines (Cartmill, 1974; Ross, 1995a, 2000). These high degrees of orbit convergence and frontation in anthropoids are also unique among mammals (Cartmill, 1974). The postorbital septum separates the anterior temporal fossa contents, which are postero-lateral to the orbit, from the eye and orbital cone (Cartmill, 1994; Ross, 1995b). Just as increasing orbital convergence "drags" the anterior tempo-ralis muscle and temporalis fascia from a posterior position to a lateral position in other mammals, the continuation of this to the uniquely high convergence and frontation in anthropoids probably accounts for a further lateral placement of these tissues (Cartmill, 1980, 1994; Ross, 1995b). In this new anatomical configuration, the anterior temporalis muscle would be in a position to directly impinge on the eye and orbital contents if the postorbital septum were not in place to prevent this (Ross, 1995a,b, 1996, 2000).

Cartmill hypothesized that the function of the postorbital septum was to prevent impingement on the eye and orbital contents in order to improve visual acuity (Cartmill, 1980). Cartmill related the importance of visual acuity to the evolution of the haplorhine retinal fovea (a depression in the retina with a concentration of cones that is associated with high visual acuity), suggesting that the septum was required to insulate the foveate eye (Cartmill, 1980; Ross, 2004; Walls, 1942). Ross' hypothesis for the function of the septum differs from Cartmill's in that he emphasized the shift in orbit orientation and the possible disruption of normal oculomotor function as the predominant factors related to the evolution of the septum, regardless of whether the lineage that first evolved it had high acuity (Ross, 1995a,b, 1996; Ross and Kay, 2004).

From the preceding, it is clear that the postorbital bar and septum have been suggested to perform two interrelated functions: (a) maintain the shape of the orbital margin, and (b) prevent disruptive movement of the eye. Two questions remain: (a) Would deformation of the orbital margin disrupt binocular vision and, if so, how? (b) Does a bony orbital margin prevent the masticatory muscles from displacing the eye and disrupting vision during mastication? In the remainder of the introduction, we will relate these hypotheses to what is currently known about the neurological and morphological systems responsible for oculomotor coordination.

Neurological and Morphological Maintenance of Eye Position

The hypotheses discussed above for the functions of the postorbital bar and septum relate to the function and maintenance of oculomotor control. In this section we provide a brief overview of the mammalian oculomotor system, describe how minor disruptive eye movements are corrected, and relate these to the anatomy of the orbit in taxa with and without postorbital bars.

The crucial function of oculomotor control is to prevent disruptive image blur across the retina during visual targeting and locomotion (Land and Nilsson, 2002; Walls, 1962). To achieve this, the oculomotor system generates precise, coordinated movements of the eyes to maintain image position on the retinae during movement of either the object or the observer (Goldberg et al., 1991). In addition to primary oculomotor commands (i.e., brainstem and cortically-based commands, reviewed in Robinson, 1975), oculomotor control is maintained by a collateral discharge system under which collateral axons from extraocular interneurons provide information on the timing and magnitude of contractions to other extraocular motoneurons (e.g., Delgado-Garcia et al., 1977; Evinger et al., 1981; Guthrie et al., 1983; Highstein and Baker, 1978; Highstein et al., 1982; Matin et al., 1982). Since significant forces other than those imparted by the extraocular muscles typically do not act on the eye, set contractions of extraocular muscles should predictably position the eye (Goldberg et al., 1991; Guthrie et al., 1983; Ruskell, 1999). These contraction and eye position data are "stored" such that for subsequent eye movements, extraocular motoneurons are recruited based upon prior eye position (Kandel et al., 2000).

Several mechanisms have been shown to effectively correct for small movements at low frequencies in humans and may be present in other taxa as well. For example, Ilg et al. (1989) found that small magnitude movements, less than 1 degree, and frequencies less than 1 Hz are corrected for, probably by compensatory extraocular muscle firing. Additional studies have found that movements due to bone transduction of the forces generated during chewing, or simply those associated with eye movements or fixation (i.e., retinal slip, the image blur traveling across the retina proportional to the rotation speed of the eyes) are compensated for by cortical calculation of perceived relative movement of images across the retina (Murakami and Cavanagh, 1998, 2001; Sasaki et al., 2002). However, these corrective adjustments of eye position seem to operate optimally at low magnitude eye displacements, and fail for large magnitude and high frequency eye movements (Ilg et al., 1989; Rashbass and Westheimer, 1961; Velay et al., 1997). What these studies do demonstrate is that neurological corrective mechanisms exist to correct for displacement of the eye in humans. If these same mechanisms exist in other primate and mammalian taxa, then small magnitude and low frequency movements of the eye due to bone transduction of masticatory forces, or disruptive movements generated by the anterior tem-poralis tissues may be corrected, possibly with compensatory firing of the extraocular muscles.

Several aspects of orbital anatomy are crucial for maintaining oculomotor stability. Principal among these are the ligamentous attachments that suspend the eye within the orbit and maintain the normal or rest position. In humans, connective tissue extends from Tenon's capsule (the fibrous membrane that envelops the eye from the cornea to the optic nerve) to the periorbita (Koornneef, 1992; Wolff, 1948). The check ligaments of the medial and lateral recti, which attach to the medial and lateral palpebral ligaments as well as the lacrimal and lateral orbital rim respectively in humans, have long been thought to be collagenous extensions of the recti (Bannister et al., 1995; Lockwood, 1886). It has been demonstrated that not only do they contain innervated muscular components but that these so-called orbital heads function as specialized pulleys of the extraocular muscles, and have been found in humans, macaques, rabbits, and rats (e.g., Briggs and Schachat, 2002; Clark et al., 1997, 1998; Demer et al., 1995, 1996; Khanna and Porter, 2001; Oh et al., 2001). The medial and lateral extraocular muscle pulleys are believed to function to maintain linear position of the eye and influence the position of its rotational axis (Demer, 2002; Demer et al., 2000; Haslwanter, 2002). The orbital heads may also contribute to eye movement during saccadic (foveating) movements (Briggs and Schachat, 2002; Demer, 2002).

Displacement of these extraocular muscle pulleys in humans due to abnormal development, surgery or orbital "blow out" fractures causes misalignment and improper rotation, and is associated with strabismus, which is a deviation of the visual axis of the affected eye during normal vision (Abramoff et al., 2002; Clark et al., 1998; Koornneef, 1992; Miller and Demer, 1992). The lateral pulley in particular is important when considering the functional implications of Cartmill's model. In rats and rabbits, the pulley of the lateral rectus attaches to the postorbital ligament, whereas in macaques and humans it attaches to the lateral orbital rim (Demer et al., 2000; Eglitis, 1964; Khanna and Porter, 2001). In the type of orbit deformation described by Cartmill, contraction of the anterior temporalis muscle and tension in the temporalis fascia (of which the postorbital ligament is the anteriormost portion) would deform the postorbital ligament, and also displace the pulleys of the lateral and superior recti. Displacement and the associated change in tension of the pulleys during mastication would probably cause symptoms similar to those experienced by humans with congenital dysfunction or injuries. Therefore, whereas there are neural circuits that probably maintain eye position during unavoidable displacement, these depend on maintaining the integrity of the orbit because corrective eye movements are made by the extraocular muscles, which require predictable orbit position. If the orbit were deformed during such corrective contractions of the extraocular muscles, then the subsequent eye movements would cause misalignment.

What Are the Consequences of Disruption of Oculomotor Coordination?

The disruption of oculomotor precision would have several effects on visual perception, which in turn would affect an animal's ability to interact with its environment. In taxa like primates with convergent orbits, and significant binocular field overlap and stereopsis (the cortically driven perception of depth and solidity based on binocular cues), there is a substantial field of points in space that projects images to corresponding points on each retina based on the intersection of the visual axes (point F in Figure 2). This field in visual space is called the horopter (Figure 2; see Howard and Rogers, 1995). Objects situated within the horopter can be fused into single images with associated depth and solidity visual information. Displacement of one eye leads to a shift in retinal horizontal disparity and the relative position of the horopter. If the object of interest did not shift position, the shift in retinal horizontal disparity would result in misalignment of the visual axes, thereby causing images to fall on noncorresponding areas of the two retinae (Leigh and Zee, 1991). This is schematically illustrated in Figure 2B in which the two visual axes no longer intersect at the object of interest, point F, but beyond it. This would lead to the perception of two versions of the same object, a phenomenon known as diplopia.. Mastication is a cyclical behavior occurring at 1-3 Hz in primates (Ross, unpublished data), and the impingement on the eye would also be cyclically correlated with each power stroke. This oscillating misalignment of the optic axes during mastication would result in oscillopsia (the apparent movement of objects and the environment) (Duke-Elder and Wybar, 1973). Oscillopsia and diplopia lead to visual confusion, vertigo, and nausea in humans (Duke-Elder and Wybar, 1973). These symptoms are experienced in humans with tumors deep to the anterior tem-poralis muscle that pierce the postorbital septum and impart a medially directed force on the lateral rectus muscle during temporalis contraction (Crone, 1973; Emerick et al., 1997; Knight et al., 1984). For nonconjugate eye movements to result in diplopia, the visual axis of one eye must move beyond the bounds of the disparity threshold, the largest retinal disparity between the two images presented to each eye that can be fused into a single image (Howard and Rogers, 1995). The required shift in horizontal disparity for diplopia to occur differs among taxa. For example, in Felis, lateral eye movements leading to horizontal disparities beyond approximately 1° of arc result



allows binocular fusion


Images do not fall on corresponding regions of the retina =DIPLOPIA or double vision

Images fall on corresponding regions of the retina allows binocular fusion


Fmp rtemp

Fmp rtemp

Figure 2. Schematic of disruptive movement of the right eye during either right temporalis or medial pterygoid contraction. The horopter is a substantial field of points in space that project images to corresponding points on each retina based on the intersection of the visual axes. The visual axis is defined as a line joining the object or fixation point, the center of the pupil, and the fovea centralis in anthropoids. An equivalent definition is used here for taxa without a fovea. (A) The object of interest, point F, lies within the horopter and is normal fused into a single image. (B) When displacement of the right eye occurs due to muscle contraction (the arrows indicate the presumed direction of anterior temporalis (Ft ) and medial pterygoid (F ) forces acting on the orbital contents), the visual axis of the right eye falls beyond the horopter F. Point F cannot be fused and is presented in different positions in space to each eye and retina. This would cause diplopia.

Table 1. Binocular disparity thresholds


Binocular disparity threshold

Eye radius"

Diplopia arc length4



15 min arc'

20 mm

0.087 mm

Timney and Keil



10 min arcc

12 mm

0.035 mm

Mitchell (1966)*


6-18 min arc'

11 mm

0.019 mm

Poggio and Poggio



30-50 min arcc

10.8 mm

0.094 mm

Packwood and Gordon



15-20 min arcc

5.25 mm

0.031 mm

Moran et al. (1983)


7.9 mm

0.069 mmf

Ordy and Samorajski



3 degreese

15 mm

0.785 mm

Clarke et al. (1976)

Pettigrew et al. (1984) Ramachandran et al.

Pettigrew et al. (1984) Ramachandran et al.

Relationship between binocular disparity, eye size and diplopia in mammals. The binocular disparity threshold is the largest retinal disparity between the images presented to each eye that can be fused into a single image. Differential eye movements beyond this result in double vision. The minimum values for binocular disparity thresholds can be combined with mean eye radii to estimate the minimum distance one visual axis would need to move to generate diplopia.

aEye diameter data for Equus, Felis, Macaca, Suricata, and Ovis are taken from Ritland (1982). Eye diameter data for Homo are from Williams et al. (1980). Data on Otolemur crassicaudatus were provided by C. F. Ross (unpublished data).

^The arc lengths were calculated by the formula S = R8, where R is the radius of the eye, 8 (in radians) is the angular excursion. The minimum disparity threshold values were used for these estimates. c Data derived from psychophysical or behavioral studies.

d Two studies have made the claim that images with disparities as high as 2 can still be fused. Howard and Rogers (1995) point out several flaws with these two studies, such as the fact that the criteria for perceiving fusion without depth are not described. Nor have these results been replicated by numerous other studies. For this reason, we use the Mitchell (1966) data. e Data derived from electrophysiological studies.

-^The diplopia arc length for Otolemur was computed using the Felis binocular disparity threshold value (see text).

in diplopia (Packwood and Gordon, 1975), whereas in humans disparities beyond 5-10° of arc result in diplopia (Mitchell, 1966; also see Table 1). Protruding movements of the eye would also lead to disparity and blur in a way similar to mediolateral movements. This point is listed in Figure 2B in which displacement of the right visual axis shifts the point of fixation, F, beyond the horopter and object of fixation. Protrusion in this study is defined as relative to the orbital plane and as such is anterolateral relative to the skull. This would shift the position of the image on the right retina to one that does not correspond to the position on the left retina. Depending on the magnitude of the movement and the size of the eye, protrusion could also lead to diplopia.

There are several potential behavioral and ecological consequences of oscil-lopsia and diplopia. While the animal is chewing, accurate location of potential predators or prey would be impossible, and likely the time both during and slightly after chewing would be one of visual confusion, as it is in humans that experience oscillopsia and diplopia. An arboreal animal would be unable to effectively navigate in its environment. Data on Loris tardigradus and the didelphid marsupial Caluromys derbianus (Nekaris and Rasmussen, 2001, 2003; Rasmussen, 1990), both nocturnal visually directed predators in the terminal branches, indicate that these taxa opportunistically capture flying and nonflying insect prey while feeding on flowers and fruits in the terminal branches. Diplopia and oscillopsia would prevent this feeding strategy.

For taxa like perissodactyls and artiodactyls with panoramic visual fields that spend a large component of the day masticating, oscillopsia would make the location of predators difficult. Presumably the loss of stereopsis for these animals would not be critical because the stereoscopic field is not large in these taxa (e.g., Hughes, 1977; Hughes and Whitteridge, 1973; Pettigrew et al., 1984), although the stereoscopic portion of the visual field may be critical to taxa that locomote on cliffs and other precarious substrates (Ramachandran et al., 1977).

Focus of This Study

Whereas the biomechanics of the bony orbit has received extensive attention (Hylander et al, 1991; Ravosa et al., 2000a,b; Ross, 2001; Ross and Hylander, 1996), one question common to hypotheses for the functions of postorbital bars and septa has not been directly evaluated, and that is whether the masticatory muscles displace and disrupt the eye during mastication. In this study we use ocular kinematic methods in anesthetized subjects to test specifically whether the masticatory muscles can disrupt eye position. We further distinguish between eye movement caused by anterior temporalis and medial pterygoid muscles as well as compare the magnitudes of the eye movements they cause. The goal is to characterize the eye movements, if any, caused by masticatory muscle contraction, compare these with known or estimated binocular disparity thresholds, and then reevaluate hypotheses for primate lateral orbital wall function.

0 0

Post a comment