SKULL MECHANICS IN SPHENODON

Almost all of the cranial joints examined in Sphenodon remain patent through life, which suggests following the majority of skull growth they remain important to skull function. Using inferences similar to those employed by Herring (1972), Buckland-Wright (1978) and Taylor (1992), we combine our observations on cranial joints, with current understanding of the jaw muscles (e.g., Jones et al. 2009) and observed skull architecture to develop a series of hypotheses about skull biomechanics in Sphenodon.

Forces Acting on the Skull

Forces acting on the skull in the living animal include gravity and occasional blows during fighting (Seligman et al. 2008; Jones and Lappin 2009) and prey acquisition, but most mechanical stress will be experienced during feeding: the temporal region and skull roof will be pulled downward as the jaw adductor muscles contract; the back of the cranium will be pushed up and to some extent anteriorly as the lower jaws are pulled against the quadrate condyles by the jaw adductor muscles; and the maxillae and other facial bones will be pushed dorsally as the jaws are brought against resistant food items. Therefore, in a very general way, we may expect the skull to act like a beam subject to three-point bending with compression along the dorsal surface and tension along the ventral margin (Taylor 1992; Russell and Thomason 1993; Weishampel 1993; Preuschoft and Witzel 2002; Henderson 2002; Rafferty et al. 2003). However, this simple model is complicated by several other considerations:

• As in many other amniotes the pterygoid flanges will be pushed medially by the lower jaws as they are pulled upward and medially by the jaw muscles (Taylor 1992).

• Following jaw closure there is an anteriorly directed (prooral) shearing motion in Sphenodon (Farlow 1975; Gorniak et al. 1982; Curtis et al. 2010b).

• Sphenodon possesses a row of teeth on the lateral edge of the palatine bone parallel with the tooth row on the maxilla so that when the jaws close the lower (dentary) tooth row bites between the upper tooth rows (Gray 1831; Günther 1867; Gorniak et al. 1982; Jones et al. 2009). This type of occlusion means that during biting the palatine bones will receive a substantial component of the loading forces directly (Gorniak et al. 1982). In addition, as the jaws close, food will be forced between the upper tooth rows pushing them apart, and creating lateral forces on the posterior ends of the maxillae and medial forces on the posterior end of the palatines (Figure 79). The posteroventral surfaces of the premaxillary chisel-like teeth are also probably pushed forward at the end of the shearing action as evidenced by the wear facets found on the tips of the lower jaws (Reynoso 1996; Jones et al. 2009).

• The prominent premaxillary teeth in adult Sphenodon are likely to contact a food item well before those on the maxilla and palatine. Hence, during a bite the premaxillae will experience loading well before the rest of the skull, causing shear or bending within the rostrum.

• Sphenodon is known to bite both unilaterally and bilaterally (Gorniak et al. 1982). Unilateral biting will generate complex torsional forces in the skull rather than simple long axis bending (e.g., Herring and Teng 2000; Thomason et al. 2001).

• The forces imposed on the skull by muscles are both important and complicated (Herring and Teng 2000; Herring et al. 2001; Thomason et al. 2001). When the lower jaw meets resistance (a food item) and the muscles contract, the sites of origin and attachment will experience tensile forces. The action of some muscles may cancel out the forces from others (Buckland-Wright 1978). However, the system will be dynamic because muscle vectors and forces will alter with gape and with bilateral or unilateral biting (Gorniak et al. 1982; Curtis et al. 2010b). As in vivo work on other taxa has demonstrated, some parts of the skull are likely to experience both tensile and compressive forces at different parts of a bite cycle during unilateral biting (e.g., Herring and Teng 2000; Herring et al. 2001; Thomason et al. 2001).

• The neck muscles may generate some posterior forces on the braincase and post- temporal arch during head movements related to feeding (e.g., Rayfield et al. 2001; Preuschoft and Witzel 2002; Jones et al. 2009).

• There may be some lateral force on the quadrates during food transport and swallowing.

• Together, the very low position of the jaw joint and the shape of the lower jaw in Sphenodon permit the tooth rows to meet almost evenly. As the mouth closes and the posterior teeth of the lower jaw and maxilla pass each other, so do the anterior teeth (e.g., DGPC 2). This arrangement means that when the jaws are nearly closed and the food item is relatively flat, the bite force will be evenly distributed along the tooth row. The situation will be more complicated if the jaws encounter resistance when wide open.

• The marginal tooth rows of Sphenodon are relatively closer to the midline than those of many squamates (e.g., see Evans 2008) and early rhynchocephalians (e.g., Evans 1980; Whiteside 1986). The distance between the posterior ends of the tooth rows in the horizontal plane is about 60% that of the distance between the jaw joints (62%, n = 20, standard deviation = 0.33), which provides greater jaw stability during unilateral biting and makes bilateral biting more likely when dealing with large prey. Correspondingly, the skull is more likely to undergo bending rather than torsion in such instances.

• The skull of Sphenodon includes a number of bony bars and therefore, in contrast to many mammals, is an excellent example of a skull built by struts (sensu Preuschoft and Witzel 2002). Under torsional loading the junctions between these stuts may be vulnerable to high concentrations of stress.

• Under loading, sutures will respond differently from the bones they unite.

These additional considerations highlight the unique feeding apparatus of Sphenodon and make it unlikely that the skull will behave exactly as a beam. Others (e.g., Thomason et al. 2001) have suggested that a tube might be a better model, at least for mammal skulls, but they also acknowledged that forces in the skull can be highly localised, as shown by Herring and Teng (2000). Thus although bite force was recently obtained for adult Sphenodon (Jones and Lappin 2009; Herrel et al. 2010), the nature and magnitude of the stress generated during biting will vary depending on the exact location of the bite, gape and the amount of force used.

Ridges and Thickenings

Although the dermatocranial bones of vertebrates essentially develop as thin flat sheets, the adult elements are more complex in their morphology and typically show variation in bone thickness resulting from remodelling during ontogeny. This remodelling is thought to be a response to varying levels of mechanical loading within the skull, leading to the concept of 'Benninghoff trajectories' (Benninghoff 1934; Lehman 1973a, 1973b) whereby the areas of thickening form a network of reinforcing bony ridges (Benninghoff 1934; Fox 1964; Herring 1972; Lehman 1973a, 1973b; Buckland-Wright 1978; Lanyon 1980; Reisz 1981; Taylor 1992; Lieberman 1997; Thomason et al. 2001). In experiments that simulated biting action in dried skulls, Buckland-Wright (1978) found that these thickenings ("continua") were associated with the alignment of surface strains.

The skull of Sphenodon exhibits a consistent pattern of bony ridges and thickenings that is most conspicuous in larger skulls (Figure 5, Figure 50, Figure 80), which include ridges and thickenings in the margins of the orbits; the dorsal edges of the upper temporal bars; the anterior and posterior edges of the postfrontals; the anterior, posterior and ventral edges of the postorbitals; the dorsal and ventral margins of the pterygoid-quadrate wing; the ventral margins of the jugals; the boundaries of the jugal-postorbital and quadrate-squamosal joints and the secondary bone band along the alveolar margin of each maxilla. These ridges of bone increase the cross-sectional area and thus, by inference, the strength of the skull in particular regions liable to bending or twisting forces (e.g., the secondary bone band along the maxilla will provide a greater resistance to long-axis bending forces, Jones and Lappin 2009; Jones et al. 2009). Ridges of bone are not directed toward fontanelle locations (Figure 80), for example the nasofrontal junction, the central part of the nasal-prefrontal joint, and the central area between the pterygoid and frontal, or toward parts of the skull where the bone may be thin or absent, such as the central part of the quadrate wing (e.g., DGPS1, DGPS2, LDUCZ x036). Correspondingly bone is thought to be prone to removal where stress or (more specifically) compression is relatively low (e.g., Case 1924; Olson 1961; Frazzetta 1968; Oxnard et al. 1995; Preuschoft and Witzel 2002; Farke 2008).

Ridges and thickenings can also be found within facets (Figure 81). Herring (1972) and Buckland-Wright (personal commun., 2007) suggested that these structures were oriented parallel to the likely direction of force transmission and slippage. However, it may be more complicated than this, as work on sheep skulls (Thomason et al. 2001) found that cranial architecture corresponded to working side compression but not overall strain magnitudes. Nevertheless, the arrangement of bony ridges and internal facet ornament allows the construction of a "hypothesis of compressive stress trajectories" (HCST, Figure 82) that can be tested by comparison to cranial joint structure and Finite Element Analysis. The lines of hypothesised stress are often located along the (thickened) edges of bones rather than running through the exact centre. As a result, the HCST is not equivalent to a "trans-suture web" (sensu Thompson 1995) of Sphenodon where lines are drawn between the midpoint of external suture seams (Figure 83). The lines also do not necessarily take the shortest route and occasionally cross each other rather than merging. This arrangement is not dissimilar to that found by Buckland-Wright (1978) in the cat skull.

Overall the HCST predicts that compressive stress in Sphenodon is being resisted around the orbits, postfrontals, and quadratojugal foramina, and along the lower temporal bars, the upper margin of the upper temporal fenestrae, the margins and midline of the palate, and the margins of the quadrate-pterygoid wings. This situation matches the suggestions of previous authors with respect to the orbits (e.g., Fox 1964; Frazzetta 1968; Buckland-Wright 1978; Preuschoft and Witzel 2002) and also to the lower temporal bar which is posited to act as a brace between the postorbital bar and mandibular joint (e.g., Rieppel and Gronowski 1981; Whiteside 1986). The prime locations where stress might arise during loading (the marginal dentition, the jaw joints and temporal fenestrae) are linked in a manner similar to that described Buckland-Wright (1978) as a "structural continuum." This system potentially allows opposing stresses from different regions to be transmitted toward each other as they are reduced and absorbed by the intervening hard and soft tissues (Buckland-Wright 1978). The prefrontal in Sphenodon should, for example, be important in re-directing dorsally directed forces from the palate and maxilla to the skull roof (Figure 82). Similarly, the ectopterygoid is positioned to transmit forces between the marginal tooth rows and the centre of the pterygoid (Figure 82.3). Forces from the jaw joint radiate along five different pathways from a point on the quadrate-quadratojugal (Figure 82.1 and Figure 82.4).

Preuschoft and Witzel (2002, text figures 4b, 5b, 5g, 12b) proposed a hypothesis of stress distribution during biting in Sphenodon, using lateral, dorsal, ventral and occipital views of the skull and depicting zones of expected tension and compression. It posits that the dorsal and occipital surfaces of the skull are held mainly in compression whereas the ventral part of the skull (including the lower temporal bar) is primarily in tension, with the postorbital bar and quadrate-pterygoid wings as "distance elements" within a neutral axis. In this scheme the skull is acting like a beam held posteriorly and loaded anteriorly. The fact that the skull roof of Sphenodon is composed of thicker bone than the palate lends some support to this interpretation but, as discussed in the previous section, a simple beam model may not be appropriate. The details of the skeletal architecture described here suggest that the lower temporal bars, postorbital bars, central part of the palate and edges of the palate at some point undergo compressive forces. Tensile forces are likely to be resisted by the fibrocellular sutures, tendons, fascia, skin and other soft tissues (Preuschoft and Witzel 2002)

The Cranial Joints

Given the hypothesis of loading represented by the HCST, we would predict that key cranial joints would be those: between the premaxilla, maxilla and neighbouring bones in the rostrum; along the longitudinal axis of the skull roof; resisting the pull and torsional effects of the jaw adductor muscles; supporting the jaw joint around the suspensorium; and reinforcing the palate against bending.

Rostrum

The primary forces experienced by the rostrum will be those generated by loading on the anterior premaxillary chisel teeth and/or on the anterior maxillary teeth. The former may occur when the premaxillary teeth are used to impale relatively large prey (Gorniak et al. 1982, p. 337). This behaviour will generate dorsally directed forces within the premaxilla relative to the rest of the skull, resulting in dorsal shear in the premaxilla-nasal joints and ventral tension in the premaxilla-vomerine and premaxilla-maxilla joints (Taylor 1992; Preuschoft and Witzel 2002; Rafferty et al. 2003). The premaxilla-nasal joint will resist excessive posterodorsal slippage of the premaxilla (Figure 12, Figure 13) whereas the extensive soft tissue between the base of the premaxilla and its neighbouring bones (vomer, maxilla) may permit some limited separation. This location of flexibility may also be important during prooral jaw movement when the premaxillary chisel tooth is contacted by the lower jaw.

The anterior maxilla will be loaded when Sphenodon uses its caniniform teeth (Figure 5.1), with compressive forces being directed up the anterior edge of the facial process (Figure 82.1) into the nasal and prefrontal. The rather box-like or tubular cross-section of the rostrum at this level is shaped to resist both bending and torsional stress (Preuschoft and Witzel 2002; Rafferty et al. 2003), aided by soft tissue in the large overlaps between the prefrontal, nasal and maxilla. The maxilla-nasal slot joint (Figure 17) may serve to support the maxilla as the lower jaw pushes gripped food forward against the maxillary teeth during prooral shearing.

The Skull Roof

The skull roof may be subject to long axis bending because of upward force from the teeth and downward force from the adductor muscles (Taylor 1992; Russell and Thomason 1993; Rayfield 2004, 2005a, 2005b), aggravated by upward forces from the jaw joints and a posterior pull from the neck muscles. There will also be torsion across the orbital region, especially during unilateral biting, and mediolateral tension due to the pull of the adductor muscles.

Longitudinal bending in the antorbital skull is resisted by the long overlapping nasal-frontal joint that resists significant separation while allowing small anteroposterior adjustive movements (as postulated for Allosaurus, Rayfield 2005b). The longitudinal ridges on the frontal and nasal facets (Figure 41, Figure 43) may help translate any small mediolateral movements (caused by torsion) into anteroposterior movement. Longitudinal bending of the parietal will be resisted by its dorsoventral expansion. The HCST suggests that compressive forces in the skull roof will converge between the orbits and in front of the upper temporal fenestrae (Figure 82). Significantly, this is where some of the most heavily interlocked joints are located (Figure 77, e.g., frontal-prefrontal, frontal-postfrontal, postorbital-postfrontal). The frontal-parietal joint is relatively narrow in Sphenodon but the interlocking structure maintains a rigid connection, resisting both dorsoventral bending forces and mediolateral torsional forces. It is reinforced laterally by the large spanning postfrontals (Figure 77.2). The interfrontal joint may act as a 'keystone' (Figure 79) to the arches of the orbits with forces travelling posteriorly up the postorbital bars (Figure 82) and being directed transversely against each other in another arch meeting around the postfrontal-parietal-frontal suture (Figure 82).

As previously suggested (e.g., Arnold 1998), the complex joints between the parietal, frontal and postfrontals would prevent fronto-parietal hingeing ('mesokinesis'). Even in the hatchling Sphenodon where there is a large midline fontanelle, overlaps lateral to this make bending at this joint unlikely (Rieppel 1992).

The Temporal Region and Adductor Muscles

The jaw adductor muscles originate from the sides of the parietal, from the proximal parts of the postorbital and post-temporal bars, and from fascia over the lower temporal fenestra (Gorniak et al. 1982; Jones et al. 2009). Contraction of these muscles during biting imposes powerful anteroventral, anterolateral and ventrolateral forces on the posterior skull roof and temporal region (e.g., Beherents et al. 1978; Herring and Teng 2000; Sun et al. 2004; Byron et al. 2004). At the same time, the region will be subject to a posterodorsal force from the maxilla anteriorly, and a strong upward force from the jaw joint posteriorly. Hence, there may be shear between the component parts, as well as torsion in the upper and lower temporal bars during unilateral biting.

Tension across the interparietal joint is resisted by soft tissue spanning the deep contact surface. Anteriorly, as in other amniotes, the convex "arch" formed by the postorbital bars would resist the downward pull of the adductor muscles (Frazzetta 1968), aided by the tight fit of the postorbital-postfrontal joint and, further ventrally, by the tall medial process of the jugal which is positioned to brace the postorbital (Figure 79). The long overlapping postorbital-squamosal joints in the upper temporal bars (Figure 62) should allow the small adjustive movements necessary to reduce torsion and shearing in this part of the skull. Similarly, as the plane of the jugal-postorbital joint is almost parallel to the orbital margin, the intrasutural collagen fibres should be orientated perpendicular to resist posterodorsally directed forces from the upper jaw (Figure 82). The parietal-squamosal joint alternatively resists the anteroventral pull of the m. adductor mandibulae externus and posterior pull of the m. depressor mandibulae and neck muscles (Gorniak et al. 1982; Al-Hassawi, 2007; Curtis et al. 2009; Jones et al. 2009). The slotted joint allows fibres to be arranged to resist both these movements.

The Jaw Joint and Suspensorium

The forces generated at the jaw joint may be greater than those from the upper tooth row as the jaw joint is closer to most of the muscles (Crompton and Parker 1978; Jones et al. 2009). The exact direction of these forces will also vary as gape and muscle activity changes over the course of a bite and swallowing cycle. Fusion of the quadrate and quadratojugal may reflect the need to further strengthen this part of the skull (Herring 2000). The jugal-squamosal and jugal-quadratojugal joints appear well-structured to prevent the posterior process of the jugal from rotating medially and/or the posteroventral corner of the skull from twisting laterally. The near vertical facet surfaces permit intersutural soft tissue to be orientated so as to prevent excessive shear or torsion between the lower temporal bar and squamosal-quadratojugal. The joint between the squamosal and quadrate-quadratojugal is roughly parallel to the jaw joint (Figure 77) and is probably important for transmitting forces from the jaw joint to neighbouring parts of the skull (Figure 82). The lateral part of the interlocking quadrate-squamosal joint appears already well established in hatchlings (Rieppel 1992) but the medial lappet that supports the quadrate in adults does not. The presence of a synovial cavity within the quadrate-squamosal joint (Rayfield 2005a; Jones 2006; Holliday and Witmer 2008) needs to be confirmed by direct observation or histological sections, but it could be important for dissipating compressive forces from both the jaw joint and jaw muscles.

Medially, the quadrate-pterygoid joint may be strained as the pterygoid muscles pull the two bones down but the quadrate is pushed upward by the lower jaw. However, excessive movements of the quadrate on the pterygoid would be resisted by the strong overlap (Figure 77), present even in hatchlings, whereas the deep pterygoid-quadrate wings will resist dorsoventral bending while probably allowing for some lateral deflection during swallowing.

The Palate

During biting the palatal bones and joints in Sphenodon are subject to forces from the marginal and palatal tooth rows; from the lower jaws pushing against the pterygoid flanges (Taylor 1992); from the dentary teeth wedging between the maxillary and palatine tooth rows during prooral shear; and from the actions of the deeper adductor muscles pulling the braincase (e.g., m. adductor mandibular profundus) and jaws (e.g., m. pterygoideus) against the pterygoids (Haas 1973; Gorniak et al. 1982; Wu 2003; Curtis et al. 2009; Jones et al. 2009).

Overall, the palate is relatively thin but vaulted (Figure 84, Figure 85, Figure 86), the peak of the vault occuring where the vomer, palatine, and pterygoid meet just anterior to the palatal tooth rows. Palatal vaulting has also been described in an extinct crocodylomorph (Pristichampsus vorax, Busbey 1995) and non-mammalian synapsids (Jenkins et al. 2002). A disadvantage is that the palate is moved closer to the neutral axis reducing the skull's overall resistance to torsion and dorsoventral bending (Busbey 1995, p. 190). However, vaulting may help to dissipate dorsally directed forces from the teeth in a manner analogous to the effects of a 'flying buttress' (Busbey 1995). Perhaps correspondingly the palatal bone thickenings in Sphenodon (and by inference compressive stresses) converge toward the midline of the palate (Figure 82.3). The vomer, for example, although apparently thin, is reinforced by medial and lateral thickenings. The large overlapping palatal joints in Sphenodon (Figure 77.3) potentially allow slight adjustive movements (to dissipate or redistribute torsional stress) while resisting excessive displacement. Similarly, torsional stress during unilateral biting might be reduced by long axis slippage along the midline joints, but posteriorly the interdigitated interpterygoid joint would resist medial forces from the pterygoid flanges and palatal tooth rows (via the palatine) (Rafferty and Herring 1999; Herring and Teng 2000; Markey and Marshall 2007a).

During biting and prooral shear, the complex interlocking prefrontal-palatine joint is positioned to buffer dorsally directed forces from the palatine tooth rows (Jaslow 1990; Jaslow and Biewener 1995), posterior forces generated by pterygoideus atypicus and, potentially, anterodorsal and dorsolateral (lower jaw wedging) forces during prooral shearing. Prooral shearing also imposes lateral forces on the maxilla, particularly at the posterior end where the jaws first occlude. These will strain both the maxilla-palatine joint and the extensive maxilla-jugal joint.

Together, the maxilla and jugal form much of the lateral wall of the skull as well as the margins of the palate. The maxilla-jugal joint is thus crucial to the overall stability of the skull, linking its anterior and posterior halves. The posterior end of the maxilla is dorsally and medially expanded, giving it an 'L' shape in cross-section (Figure 86). This shape strengthens the long axis of the bone against both dorsoventral bending and torsion, and braces the jugal medially. Medial movement of the jugal (caused by m. adductor mandibulae externus superficialis sensu stricto pulling on the lower temporal fascia and postorbital, Jones et al. [2009]) would also be resisted by the palatine and ectopterygoid.

The ectopterygoid is a relatively small bone positioned to transfer compressive loading between the jugal and pterygoid that might occur during biting. Within the ectopterygoid-jugal joint, the dorsally located interdigitation suggests this is an area of compression, whereas the smoother ventral part may be indicative of tension (Herring and Mucci 1991; Rafferty and Herring 1999). This pattern of stress may occur during biting as the maxillary tooth row is pushed laterally (Figure 79), and the upper temporal bars are pulled down by the muscles. The large pterygoid-ectopterygoid joint will resist compression from the overlying pterygoid muscle, medial forces from the lower jaw, maxilla and jugal (Taylor 1992) and lateral forces from contraction of the m. pterygoideus.

The Metakinetic Axis

Metakinesis is a movement of the whole braincase against the rest of the skull. It occurs in at least some lizards (e.g., Herrel et al. 1999, Evans 2008) but distribution within other amniotes remains uncertain. In Sphenodon, the joints between the braincase and the dermal skull do not appear to allow metakinetic movements comparable to those found in some lizards (contra Gardiner 1983, p. 50). Thus, the constrictor dorsalis muscles (Johnston 2010), important in lizard metakinesis (e.g., Evans 2008), are more likely to be important for proprioception and support in Sphenodon (Evans 2008; Johnston 2010). Nevertheless, the small movements that doubtless exist between the skull and braincase, particularly at the synovial basipterygoid joint, would help to transmit or dissipate stress between the braincase and the rest of the skull (Evans 2008; Johnston 2010).