Journal of Zoology

Journal of Zoology. Print ISSN 0952-8369 Skull mechanics and implications for feeding behaviour in a large carnivore guild: the , and spotted-tailed M.R.G. Attard, U. Chamoli, T.L. Ferrara, T.L. Rogers & S. Wroe

Computational Biomechanics Research Group, Evolution and Ecology Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, NSW, Australia

Keywords Abstract ; Thylacinus; ; Dasyurus; finite element analysis; bite force. Extinction risk varies across species and is influenced by key ecological para- meters, such as diet specialization. For predictive conservation science to be Correspondence effective, we need to understand extinction risk factors that may have implicated Stephen Wroe, Evolution and Ecology recent species extinctions. Diet and feeding behaviour of the large extinct Research Centre, School of Biological, Earth marsupial carnivore Thylacinus cynocephalus or thylacine have long been debated. and Environmental Sciences, University of Improved understanding of the skull’s biomechanical performance and its limita- New South Wales, Sydney, NSW 2052, tions in a comparative context may yield important insights. Here, we use three- Australia. dimensional (3D) finite element analysis to assess aspects of biomechanical Email: [email protected] performance in the skull of T. cynocephalus relative to those of two extant marsupial carnivores with known diets that occurred sympatrically with T. Editor: cynocephalus: the Tasmanian devil, Sarcophilus harrisii, and spotted-tailed quoll, Dasyurus maculatus. Together, these three species comprised the large mammalian Received 5 November 2010; revised 20 June carnivore guild in at the time of European settlement. The bone- 2010; accepted 22 June 2011 cracking S. harrisii produced high bite forces for its size as expected, but the stresses induced were surprisingly high. A higher proportion of cancellous bone in doi:10.1111/j.1469-7998.2011.00844.x the skull of this osteophage may act to absorb shock but decrease rigidity and hence raise stress. A relatively high bite force and rigid skull characterized D. maculatus, which may allow them to target prey of variable sizes. Compared with S. harrisii and D. maculatus, we found that the skull of T. cynocephalus was least well adapted to withstand forces driven solely by its jaw-closing musculature, as well as to simulations of struggling prey. Our findings suggest that T. cynocephalus likely consumed smaller prey relative to its size, which may have had implications for their survival.

Introduction Dental and biogeochemical evidence suggests that T. cynocephalus was a hypercarnivore restricted to eating The idea that some species are intrinsically at a greater risk vertebrate flesh (Nedin, 1991; Wroe, Brammall & Cooke, of extinction than others is a well-established concept, and 1998; Wroe, Lowry & Anton, 2008). The total energetic recent efforts have focused on identifying species susceptible expenditure and time spent hunting by carnivorous mam- to extinction (McKinney, 1997). Species traits, such as body mals increases with body mass (Carbone et al., 1999). Given size and degree of diet and habitat specialization, are key its large body mass and carnivorous diet, any limitations in predictors used to prioritize species in conservation and the capacity of T. cynocephalus to kill large bodied species management plans (Purvis et al., 2000; Cardillo & Brom- may have made them particularly vulnerable unless small ham, 2001; Kamilar & Paciulli, 2008). Top predators, which prey was available in high abundance. have less stable food supplies than species from lower There is almost no direct evidence regarding the diet of T. trophic levels, are thought to be particularly vulnerable cynocephalus (Paddle, 2000). The range of potential prey (Davies, Margules & Lawrence, 2000). taxa in 19th-century Tasmania included relatively large, now The thylacine, Thylacinus cynocephalus, was a large apex extinct subspecies of emu and grey kangaroo that exceeded predator (average body mass 29.5 kg; Paddle, 2000) and the the marsupial carnivores’ body mass. Whether or not T. only species of marsupial to become extinct in Tasmania cynocephalus was well adapted to kill prey approaching or following European settlement. The factors involved in its exceeding its own body mass remains a contentious issue. extinction have been debated (see review Paddle, 2000), and Previous morphological studies on this species suggest that yet the underlying cause of their decline remains incomplete. they targeted prey considerably smaller (1–5 kg) than

292 Journal of Zoology 285 (2011) 292–300 c 2011 The Authors. Journal of Zoology c 2011 The Zoological Society of London M. Attard et al. Skull mechanics in a marsupial carnivore guild themselves (Jones & Stoddart, 1998). Most data are re- restricted to smaller prey, but closely related species were stricted to Tasmania, whereas in Pleistocene and early not included in this analysis. Holocene times, the species ranged across Australia and In the present study, we apply finite element analysis, a New Guinea (Plane, 1976). At that time, still larger, now powerful engineering tool increasingly applied by biologists extinct megafauna were theoretically available and diet may (Rayfield, 2005; Strait et al., 2009; Wroe et al., 2010) to have varied in different parts of the species range. examine differences in biomechanical performance between We can infer the diet of extinct species such as the skulls of T. cynocephalus and two extant marsupial T. cynocephalus, including likely prey size, through compar- carnivore relatives: S. harrisii and D. maculatus (Cardillo isons with related species with known diets. Thylacinus et al., 2004). Our objectives are twofold. Firstly, fundamen- cynocephalus is a member of the now wholly extinct family tal structural differences between marsupial and placental Thylacinidae. Its closest living relatives are the dasyurids skulls (Wroe & Milne, 2007) make comparisons between (), a now diverse taxon comprising over 60 these taxa difficult because differences may be phylogeneti- species. Together with the monotypic (Myrmecobii- cally constrained (Wroe et al., 2007; Goswami, Milne & dae), these three families constitute the order Dasyuromor- Wroe, 2011). Additional insights into the feeding ecology of phia (Wroe & Musser, 2001). First appearing in the fossil T. cynocephalus may be derived from comparisons with record in the late Oligocene to early Miocene, dasyurid closely related, extant marsupial carnivores. Secondly, we diversity appears to have been initially low, but increased will examine whether variation in the biomechanical perfor- through the late Miocene to ultimately dominate most mance of these sympatric marsupial carnivores reflects insectivore–carnivore niches. In contrast, the Thylacinidae differences in the feeding ecology of each species. were once more diversely represented by 11 species spanning a considerable range of body sizes (Wroe, 1996; Muirhead & Materials and methods Wroe, 1998; Wroe & Musser, 2001). The two largest extant dasyurids are the Tasmanian devil, Sarcophilus harrisii (6–13 kg), and the spotted-tailed quoll, Specimens Dasyurus maculatus (1–4 kg) (Jones & Barmuta, 2000). The Dasyurus maculatus (UNSW Z20; University of New South dietary preferences of these two species have been exten- Wales), S. harrisii (AM10756; Australian Museum) and T. sively studied. Sarcophilus harrisii can generate the highest cynocephalus (AM1821; Australian Museum). The skull maximum bite force, adjusted for body mass, of any extant length of each specimen was 101, 135 and 232 mm, respec- mammalian species studied to date (Wroe, McHenry & tively. Marsupial carnivores display substantial sexual di- Thomason, 2005). This feature may assist them in catching morphism (Jones, 1997). Data on sex were unavailable; relatively large , up to 30 kg (Jones & Barmuta, however, because each is within the upper size range for 1998). Approximately 60% of the prey biomass consumed their species, it is likely that all were adult males. by adult male D. maculatus consists of medium- to large- sized mammals including bandicoots, pademelons and pos- Model assembly sums that can weigh up to three times its own body mass, although invertebrates form a larger proportion of the diet The assembly of finite element models (FEMs) largely follows of female D. maculatus (Belcher, 1995; Jones, 1997). Known the procedures applied by Wroe et al. (2007, 2010). Com- size relationships between predator and prey in S. harrisii puted tomography (CT) scanning was conducted at the and D. maculatus make them ideal candidates to compare Mater Hospital, Newcastle, NSW, using a Toshiba Aquilon with T. cynocephalus to determine prey size range. 16 scanner (Toshiba Medical Systems Corporation, Otawara, The maximum prey size that can be taken by a predator is Tachigi, Japan). Slices were 1 mm thick, with an inter-slice strongly influenced by biomechanical limitations. Skull distance of 0.8 mm and a field of view of 240 mm. Surface shape and bite force adjusted for body mass correlate meshes were generated from the CT data using Mimics 13.2 with prey size and feeding ecology in many terrestrial and converted to FEMs in STRAND7 (ver. 2.4). carnivores (Meers, 2002; Wroe et al., 2005; McHenry et al., Heterogeneous FEMs were composed of 3D four-noded 2006; Wroe & Milne, 2007). Different cranial shapes among tetrahedral ‘brick’ elements. Brick elements were assigned one extant dasyuromorphians have been correlated with diet of eight material properties based on CT density values (Rho, (Wroe & Milne, 2007), and yet little is known about how the Hobatho & Ashman, 1995; Schneider, Pedroni & Lomax, biomechanical performance of these structures may reflect 1996) ranging from low-density tissue [Young’s modulus dietary functions. (Y)=1.5 GPa; Poisson’s ratio=0.4; density=251 kg mÀ3] Computational biomechanics has been used to predict the to enamel (Y=38.6 GPa; Poisson’s ratio=0.4, density=28 capacity of mammalian carnivore species to withstand loads 61 kg mÀ3)andseeMcHenryet al. (2007). Full data for the generated by jaw adductors or struggling prey in relative eight material properties are available in Table S1. contexts. A recent study by Wroe et al. (2007) examined Jaw elevators were modelled as seven muscle subdivi- convergence with respect to biomechanical performance in sions: temporalis superficialis, temporalis profundus, masseter the skulls of T. cynocephalus and the placental grey wolf superficialis, masseter profundus, zygomatico mandibularis, subspecies Canis lupus . The results suggested that pterygoideus internus and pterygoideus externus (Wroe et al., relative to C. l. dingo, T. cynocephalus may have been 2007). Maximum contractile forces were estimated using the

Journal of Zoology 285 (2011) 292–300 c 2011 The Authors. Journal of Zoology c 2011 The Zoological Society of London 293 Skull mechanics in a marsupial carnivore guild M. Attard et al. dry skull method (Thomason, 1991) on the basis of esti- The general approaches applied here have been validated mated maximal cross-sectional areas with an estimated force against experimental data from other taxa (Wroe et al., of 300 kPa for vertebrate striated muscle (Weijs & Hillen, 2010; Tsafnat & Wroe, 2011), but not yet using FEMs of 1985). marsupial carnivores. In the absence of validated data for The protocols used for the 3D modelling of muscle material properties and forces generated in the despatch of architecture following Wroe et al. (2007) are described in large prey, actual mechanical limits on maximal prey size the supporting information. Muscle origin and insertion cannot be directly predicted. We apply a comparative areas and lines of action of muscle forces are provided in approach wherein the results must be interpreted in relative Figure S1 and pretension values for each muscle subdivision context and not as predictions of absolute stress magnitudes are listed in Table S2. A rigid link framework at the occipital (Wroe et al., 2010). condyle was used to restrain the cranium in all simulations. Bone fails under a ductile model of fracture and von Static loading cases were set to an optimum gape angle of Mises (VM) stress is a good predictor of failure in ductile 351, which is consistent with the optimum gape angle found materials (Nalla, Kinney & Ritchie, 2003; Tsafnat & Wroe, in C. l. dingo (Bourke et al., 2008). 2011). Therefore, VM stress was used here as a metric to compare the structural strength of FEMs under different loading conditions. Brick elements were selected at set Load cases points along the mid-sagittal plane, zygomatic arch and We applied six loading cases to each model: two intrinsic mandible to assess changes in stress magnitudes and dis- loading cases to simulate bites driven solely by jaw muscu- tributions under different loadings. At each point, values lature and four extrinsic loading cases to test skull perfor- were calculated by averaging VM stress of the bricks mance under the influence of unrestrained prey. surrounding each node, with between five and nine bricks The maximum bite force of each specimen was measured joined at any given point. Nodes were equally distributed for both intrinsic cases using unscaled models with speci- and VM stress was measured using the same node points in men-specific estimated muscle forces (Table S2). We also each simulation. For mid-sagittal sampling, stresses were solved two sets of load cases with models scaled to the same sampled from the anterior edge of the nasal bone to the total surface area of T. cynocephalus to remove the effects of posteriormost point of the sagittal crest. At the zygomatic size where (1) the same muscle forces were applied to each arch, nodes were selected midway between the top and the specimen to quantify differences in bite forces after adjust- bottom of the arch from the most anterior extreme of the ing for size differences; (2) the same bite reaction force as T. jugal to the posterior extreme of the zygomatic arch. Nodes cynocephalus (747 N at canines and 1659 N at carnassial) were also selected along the right side of the jaw from the was allocated to each model to determine how well adapted mid-point between the first incisors and ventral surface of each species was to withstand the stresses generated in the the jaw to the temporomandibular joint. production of a given bite reaction force. Scaled models The proportion of cortical to cancellous bone in the were also used to measure the muscle recruitment required mandible and cranium was estimated based on volume to produce the same bite force as T. cynocephalus. ratios (Table S3). Brick elements with material properties The intrinsic cases comprised a bilateral bite at the 1–4 were allocated as cancellous bone, while 5 (i.e., density canines and a unilateral bite at the fourth molar, or of 1855.9 kg mÀ3) and higher were considered cortical (Ta- ‘carnassial’. Constraints were applied at the tip of the upper ble S1). and lower canines on both sides of the skull for the bilateral bite and at the upper and lower fourth molar on the right Results side of the skull for the unilateral bite. Extrinsic loading cases were applied only to scaled Bite forces were higher at the carnassial than at the canines models, and consisted of (1) a lateral ‘shake’ (50 N); (2) axial for each model and were positively correlated with body size twist (5000 N mm); (3) pull back (50 N); (4) dorsoventral (Fig. 1a). In unscaled models, the actual bite performance of force (50 N). These extrinsic loading cases simulated differ- S. harrisii was similar to that of the much larger T. ent behaviours used by carnivores during prey capture and cynocephalus. However, among models scaled to a uniform killing. Axial twisting and lateral shaking in particular are surface area and muscle recruitment force, S. harrisii had the used by dasyurids to disorient unrestrained prey and assist lowest bite force for both intrinsic cases (Fig. 1b) and in canine penetration into the skull of restrained prey (Pellis required the highest muscle recruitment to produce the same & Officer, 1987). All extrinsic forces were applied using an bite force (Fig. 2). Overall, the magnitudes and distributions H-frame that linked the upper and lower jaws, with forces of VM stress in S. harrisii loading simulations are broadly applied at the centre of the frame (and see Wroe et al., 2007). similar to those generated in T. cynocephalus and D. macu- The extrinsic forces applied were based on the estimated latus. Dasyurus maculatus produced the highest bite forces in body mass of T. cynocephalus taken from Wroe et al. (2007) scaled models for both loading cases (Fig. 1), which suggests and are low relative to the size of the predators. However, as that the geometry of its lever system is the most efficient. this is a comparative analysis using scaled models, the actual The distribution of stress in all models was similar on value used is immaterial, provided that it is held constant both sides of the skull for bilateral loads (Fig. 3a–c) and was across species (Dumont, Grosse & Slater, 2009). largely concentrated on the working side for unilateral loads

294 Journal of Zoology 285 (2011) 292–300 c 2011 The Authors. Journal of Zoology c 2011 The Zoological Society of London M. Attard et al. Skull mechanics in a marsupial carnivore guild

(a)

Figure 2 Muscle recruitment of different dasyuromorphians to pro- duce the same bite force as Thylacinus cynocephalus in models (b) scaled to a uniform surface area.

species undergoing extrinsic loads were greatest in the cranium (Fig. 5). Lateral shaking caused the highest stresses along the zygomatic arches and mid-sagittal planes of each species. Stresses were higher for T. cynocephalus than the other two species in all extrinsic loading cases. The highest VM stresses accumulated at the parietal of T. cynocephalus during a pullback and dorsoventral loading and at the anterior point of the rostrum and parietal during a lateral shake. The proportion of cortical and cancellous bone in the skull varied considerably between species (Fig. 6). Dasyurus maculatus had the highest proportions of cortical bone in the cranium and mandible and S. harrisii had the highest Figure 1 Bite force reaction for (a) unscaled models and (b) models proportion of cancellous bone. scaled to the same surface area and muscle force as Thylacinus cynocephalus. Discussion (Fig. 3d–f). Peak VM stresses in T. cynocephalus during a Comparatively high levels of stress in T. cynocephalus may canine bite were the highest at the nasal and temporal line, be related to its relatively long rostrum (Werdelin, 1986). A which may be attributed to its relatively long rostrum (Fig. longer snout may enable higher velocities at the canine, a 3c). Unilateral carnassial loading on T. cynocephalus re- characteristic of predators specializing in relatively small, sulted in localized stress at the temporal line, with higher more agile prey and use a pounce-pursuit or ambush hunt- peak stresses than other species (Fig. 3h). In the cranium, ing style. Dasyurus maculatus and S. harrisii can catch prey VM stresses were greatest at the frontal line of S. harrisii and considerably larger than themselves (Jones, 1995; Jones & at the temporal line for D. maculatus in both intrinsic loads Barmuta, 1998). Anecdotal accounts suggest that T. cynoce- (Fig. 3g, h). The marsupial carnivores studied here all phalus may have consumed only soft tissue (Paddle, 2000). showed a similar distribution of stress along the zygomatic Certainly, its dentition shows none of the features present in arch that peaked posteriorly. specialized osteophages, such as S. harrisii, which can The highest stresses were located along the mandible in consume whole carcasses, including teeth, bone and fur scaled models during intrinsic loads (Fig. 4), a finding (Owen & Pemberton, 2005), or species consuming signifi- consistent with studies in other taxa (Wroe, 2008). The cant proportions of invertebrate food. Given the evidence mandible may be a more reliable indicator of feeding for a diet restricted with respect to prey size, type and the behaviour than the cranium because selective pressure on proportion of the prey item that it could effectively con- its morphology is less likely to be influenced by competing sume, we suggest that T. cynocephalus was particularly demands such as olfaction, hearing and vision (Figueirido, sensitive to environmental disturbance. Palmqvist & Pe´rez-Claros, 2009; Wroe et al., 2010). Our results also suggest a potential overlap between T. The ability to resist extrinsic forces may be a better cynocephalus and the two dasyurids regarding prey size. indicator of a predator’s ability to subdue relatively large Niche overlap has been found between sub-adult S. harrisii prey (Preuschoft & Witzel, 2005). Differences between and male D. maculatus (Jones & Barmuta, 1998), although

Journal of Zoology 285 (2011) 292–300 c 2011 The Authors. Journal of Zoology c 2011 The Zoological Society of London 295 Skull mechanics in a marsupial carnivore guild M. Attard et al.

(a) (b) (c) (g)

(d) (e) (f) (h)

Figure 3 Stress distribution in scaled heterogeneous models with a uniform bite force during a bilateral canine bite (a–c) and a unilateral carnassial bite (e, f). Dorsal views of models are displayed for Dasyurus maculatus (a, d), Sarcophilus harrisii (b, e) and Thylacinus cynocephalus (c, f). von Mises (VM) stress was measured from the anterior to posterior along the mid-sagittal plane for a bilateral canine bite (g) and a carnassial bite (h). Symbols for each species are: Thylacinus cynocephalus (square), Sarcophilus harrisii (triangle) and Dasyurus maculatus (circle).

(a) (b) (c) (g)

(d) (e) (f) (h)

Figure 4 Lateral view of the stress distribution of heterogeneous models with a uniform bite force in scaled models during a bilateral canine bite (a–c) and a unilateral carnassial bite (d–f). von Mises (VM) stress was measured from the anterior to posterior along the jaw for a bilateral canine bite (g) and a carnassial bite (h). Symbols for each species are: Thylacinus cynocephalus (square), Sarcophilus harrisii (triangle) and Dasyurus maculatus (circle). male D. maculatus consume a significantly higher propor- scavenged from kills made by T. cynocephalus. Until late tion of arboreal prey than do sub-adult S. harrisii. Current Pleistocene times, this guild also included the still larger hypotheses suggest that T. cynocephalus preyed on juvenile Thylacoleo carnifex or marsupial lion (Wroe et al., 1999) S. harrisii, and S. harrisii ate T. cynocephalus young and that likely preyed upon these smaller predators. A shared

296 Journal of Zoology 285 (2011) 292–300 c 2011 The Authors. Journal of Zoology c 2011 The Zoological Society of London M. Attard et al. Skull mechanics in a marsupial carnivore guild

(a) (b) (c) (m)

(d) (e) (f) (n)

(g) (h) (i) (o)

(j) (k) (l) (p)

Figure 5 Comparison of the distribution of von Mises (VM) stresses in MPa in models scaled to the same surface area during four extrinsic loads: an axial twist (a–c), pullback (d–f), lateral shake (g–i) and dorsoventral (j–l). Species from left to right are Dasyurus maculatus (a,d,g,j),Sarcophilus harrisii (b, e, h, k) and Thylacinus cynocephalus (c, f, i, l). VM stress was measured from the anterior to the posterior along the mid-sagittal plane for each species during an axial twist (m), pullback (n), lateral shake (o) and dorsoventral (p). Symbols for each species are: Thylacinus cynocephalus (square), Sarcophilus harrisii (triangle) and Dasyurus maculatus (circle). White areas indicate VM exceeds the scale maximum (2.5 MPa) in those areas. preference towards mammalian prey may have exposed threat display, as does S. harrisii (Pemberton & Renouf, these species to increased periods of resource scarcity 2003). Several instances of agonistic gaping behaviour were (Wiens, 1993). described in both wild and captive T. cynocephalus in Our finding that T. cynocephalus was not well adapted to response to humans (Paddle, 2000). Alternatively, the wide take relatively large prey compared with other gape angle of T. cynocephalus may have been a retained brings into question the putative role of its large gape if not plesiomorphy. The contention that T. cynocephalus would as a means to assist in the capture and killing of large prey. It have been able to open its jaws to angles of 1201 (Mittlebach is possible that T. cynocephalus used their wide gape as a & Crewdson, 2005) is unrealistic and would almost certainly

Journal of Zoology 285 (2011) 292–300 c 2011 The Authors. Journal of Zoology c 2011 The Zoological Society of London 297 Skull mechanics in a marsupial carnivore guild M. Attard et al.

harrisii showed the highest proportion of less stiff cancellous bone. However, it may be that a more compliant skull may enable S. harrisii to absorb the shock induced by biting down on relatively hard bone. Allometric factors may further complicate interpretation here. Recent work on primates suggests that cranial rigidity decreases with increasing size (Strait et al., 2010). Our findings are partially consistent with this suggestion insofar as we found that the skull of the smallest species, D. maculatus, had the highest proportion of stiff cortical bone. Sarcophilus harrisii is currently critically endangered by both facial cancer and the introduction of the red fox Vulpes vulpes (Hawkins et al., 2006). It has been argued that interference competition with S. harrisii has been an impor- tant factor in thwarting the success of previous attempts to Figure 6 Volume ratio of cortical to cancellous bone in the mandible establish V. vulpes in Tasmania. This may have been and cranium of Dasyurus maculatus, Sarcophilus harrisii and Thylaci- through either aggressive exclusion or predation on denned nus cynocephalus. juveniles. Dasyurus maculatus also experience competition from feral cats for food and den sites (Jones & Barmuta, cause dislocation due to the physical constraints of the pivot 1998). Future comparisons of mechanical performance in joint. It is more probable that they were capable of gape marsupial carnivores with a wider range of placental carni- angles similar to S. harrisii, which can open their jaws vores, including potential exotic competitors, may yield 75–801 (Pemberton & Renouf, 2003), which is still a very useful insights into the likely outcomes of such competition. wide gape for a . Possession of a relatively high bite force in S. harrisii (Wroe et al., 2005) was expected and supported by our results. We anticipated that the short, broad skull shape of Acknowledgements S. harrisii was primarily responsible for the production of extremely high bite forces in this species by reducing the We thank the University of New South Wales and Sandy length of the jaw out-lever, thereby increasing the leverage Ingleby at the Australian Museum for providing specimens of the jaw musculature. However, contra our expectations, used in this study. This work was funded by the University the skull and muscle geometry of S. harrisii did not transmit of New South Wales Internal Strategic Initiatives Grant to jaw muscle forces into bite forces more efficiently, with this S.W. and the Australian Research Council (DP0666374 and species producing the lowest bite forces in models scaled to DP0987985). the same surface area and muscle force. Rather, these high bite forces are the product of a relatively greater muscle force. As a bone-crunching specialist, we anticipated that S. References harrisii would generate less stress in the skull than flesh- Belcher, C.A. (1995). Diet of the (Dasyurus eating carnivores. In contrast to our expectations, the maculatus). Wildl. Res. 22, 341–357. magnitude and distribution of stress in the mandible of Bourke, J., Wroe, S., Moreno, K., McHenry, C. & Clausen, P. scaled models with uniform bite reaction forces were similar (2008). Effects of gape and tooth position on bite force and for all species studied here. Differences in the mechanical skull stress in the Dingo (Canis lupus dingo) using a 3- performance of these dasyuromorphians may be related to PLoS One their different foraging strategies. Sarcophilus harrisii pri- dimensional finite element approach. 3, e2200. marily scavenge whereas D. maculatus are predominantly Buchmann, O.L.K. & Guiler, E.R. (1977). Behavior and active hunters. They are both able to crush the skull of live ecology of the Tasmanian devil, Sarcophilus harrisii. In: prey (Ewer, 1969; Buchmann & Guiler, 1977), although the The biology of marsupials: 155–168. Stonehouse, B. & enlarged cheek-tooth cusps of S. harrisii are better adapted Gilmore, D. (Eds). Baltimore: Macmillan. for breaking open and consuming bone. Dasyurus maculatus Carbone, C., Mace, G.M., Roberts, S.C. & MacDonald, feed on both fast- and slow-moving prey of various sizes, D.W. (1999). Energetic constraints on the diet of terrestrial utilize patchy resources and will take advantage of short- carnivores. Nature 402, 286–288. term fluctuations in prey abundance (Glen & Dickman, Cardillo, M., Bininda-Emons, O., Boakes, E. & Purvis, A. 2006; Dawson et al., 2007). (2004). A species-level phylogenetic supertree of marsu- Distinct differences between the relative proportions and pials. J. Zool. (Lond.) 264, 11–31. distributions of cancellous and cortical bone in the skull, as Cardillo, M. & Bromham, L. (2001). Body size and risk of revealed in our study, may influence the distribution and extinction in Australian mammals. Conserv. Biol. 15, magnitude of stress (Thomason, 1995). Surprisingly, S. 1435–1440.

298 Journal of Zoology 285 (2011) 292–300 c 2011 The Authors. Journal of Zoology c 2011 The Zoological Society of London M. Attard et al. Skull mechanics in a marsupial carnivore guild

Davies, K.F., Margules, C.R. & Lawrence, J.F. (2000). Which McHenry, C.R., Wroe, S., Clausen, P.D., Moreno, K. & traits of species predict population declines in experimental Cunningham, E. (2007). Supermodeled sabercat, predatory forest fragments? Ecology 81, 1450–1461. behavior in Smilodon fatalis revealed by high-resolution 3D Dawson, J.P., Claridge, A.W., Triggs, B. & Paull, D.J. (2007). computer simulation. Proc. Natl. Acad. Sci. USA 104, Diet of a native carnivore, the spotted-tailed quoll (Day- 16010–16015. surus maculatus), before and after an intense wildfire. Wildl. McKinney, M.L. (1997). Extinction vulnerability and selec- Res. 34, 342–351. tivity: combining ecological and paleontological views. Dumont, E.R., Grosse, I.R. & Slater, G.J. (2009). Require- Annu. Rev. Ecol. Syst. 28, 495–516. ments for comparing the performance of finite element Meers, M.B. (2002). Maximum bite force and prey size of models of biological structures. J. Theor. Biol. 256, Tyrannosaurus rex and their relationships to the 96–103. inference of feeding behavior. Hist. Biol.: J. Paleobiol. 16, Ewer, R.F. (1969). Some observations on the killing and 1–12. eating of prey by two dasyurid marsupials: the , Mittlebach, M. & Crewdson, M. (2005). Carnivorous nights: Dasycercus cristicauda, and the Tasmanian devil, Sarco- on the trail of the Tasmanian tiger. New York: Villard. philus harrisii. Z. Tierpsychol. 26, 23–38. Muirhead, J. & Wroe, S. (1998). A new genus and species, Figueirido, B., Palmqvist, P. & Pe´rez-Claros, J.A. (2009). Badjcinus turnbulli (Thylacinidae: Marsupialia), from the Ecomorphological correlates of craniodental variation in late Oligocene of Riversleigh, northern Australia, and an bears and paleobiological implications for extinct taxa: an investigation of thylacinid phylogeny. J. Vertebr. Paleon- approach based on geometric morphometrics. J. Zool. tol. 18, 612–626. (Lond.) 277, 70–80. Nalla, R.K., Kinney, J.H. & Ritchie, R.O. (2003). Mechan- Glen, A.S. & Dickman, C.R. (2006). Diet of the spotted-tailed istic fracture criteria for the failure of human cortical bone. quoll (Dasyurus maculatus) in eastern Australia: effects of Nat. Mater. 2, 164–168. season, sex and size. J. Zool. (Lond.) 269, 241–248. Nedin, C. (1991). The dietary niche of the extinct Australian Goswami, A., Milne, N. & Wroe, S. (2011). Biting through marsupial lion: Thylacoleo carnifex Owen. Lethaia 24, constraints: cranial morphology, disparity and conver- 115–118. gence across living and fossil carnivorous mammals. Proc. Owen, D. & Pemberton, D. (2005). Tasmanian devil: a unique Roy. Soc. Lond. Ser. B: Biol. Sci. 278, 1831–1839. and threatened . Australia: Allen & Unwin. Hawkins, C.E., Baars, C., Hesterman, H., Hocking, G.J., Paddle, R. (2000). The last Tasmanian tiger: the history and Jones, M.E., Lazenby, B., Mann, D., Mooney, N., Pem- extinction of the thylacine. Cambridge: Cambridge Univer- berton, D., Pyecroft, S., Restani, M. & Wiersma, J. (2006). sity Press. Emerging disease and population decline of an island Pellis, S.M. & Officer, R.C.E. (1987). An analysis of some endemic, the Tasmanian devil, Sarcophilus harrisii. Biol. predatory behaviour patterns in four species of carnivorous Conserv. 131, 307–324. marsupials (Dasyuridae), with comparative notes on the Jones, M.E. (1995) Guild structure of the large marsupial Eutherian cat Felis catus. Ethology 75, 177–196. carnivores in Tasmania. PhD thesis, University of Tasma- Pemberton, D. & Renouf, D. (2003). A field study of com- nia at Hobart. munication and social behaviour of the Tasmanian devil at Jones, M.E. (1997). Character displacement in Australian feeding sites. Aust. J. Zool. 41, 507–526. dasyurid carnivores: size relationships and prey size pat- Plane, M.D. (1976). The occurrence of Thylacinus in Tertiary terns. Ecology 78, 2569–2587. rocks from Papua New Guinea. J. Aust. Geol. Geophys. 1, Jones, M.E. & Barmuta, L.A. (1998). Diet overlap and 78–79. abundance of sympatric dasyurid carnivores: a hypothesis Preuschoft, H. & Witzel, U. (2005). Functional shape of the of competition? J. Anim. Ecol. 67, 410–421. skull in vertebrates: which forces determine skull morphol- Jones, M.E. & Barmuta, L.A. (2000). Niche differentiation ogy in lower primates and ancestral synapsids? Anat. Rec. among sympatric Australian dasyurid carnivores. J. Mam- A. Discov. Mol. Cell. Evol. Biol. 283A, 402–413. mal. 81, 434–447. Purvis, A., Gittleman, J.L., Cowlishaw, G. & Mace, G.M. Jones, M.E. & Stoddart, D.M. (1998). Reconstruction of the (2000). Predicting extinction risk in declining species. Proc. predatory behaviour of the extinct marsupial thylacine Roy. Soc. Lond. Ser. B: Biol. Sci. 267, 1947–1952. (Thylacinus cynocephalus). J. Zool. (Lond.) 246, 239–246. Rayfield, E.J. (2005). Aspects of comparative cranial me- Kamilar, J.M. & Paciulli, L.M. (2008). Examining the extinc- chanics in the theropod dinosaurs Coelophysis, Allosaurus tion risk of specialized folivores: a comparative study of and Tyrannosaurus. Zool. J. Linn. Soc. 144, 309–316. Colobine monkeys. Am. J. Primatol. 70, 816–827. Rho, J.Y., Hobatho, M.C. & Ashman, R.B. (1995). Relations McHenry, C.R., Clausen, P.D., Daniel, W.J., Meers, M.B. & of mechanical properties to density and CT numbers in Pendharkar, A. (2006). Biomechanics of the rostrum in human bone. Med. Eng. Phys. 17, 347–355. crocodilians: a comparative analysis using finite-element Schneider, U., Pedroni, E. & Lomax, A. (1996). The calibra- modeling. Anat. Rec. A. Discov. Mol. Cell. Evol. Biol. 288A, tion of CT Houndsfield units for radiotherapy treatment 827–849. planning. Phys. Med. Biol. 41, 111–124.

Journal of Zoology 285 (2011) 292–300 c 2011 The Authors. Journal of Zoology c 2011 The Zoological Society of London 299 Skull mechanics in a marsupial carnivore guild M. Attard et al.

Strait, D.S., Grosse, I.R., Dechow, P.C., Smith, A.L., Wang, and niche overlap. Proc. Roy. Soc. Lond. Ser. B: Biol. Sci. Q., Weber, G.W., Neubauer, S., Slice, D.E., Chalk, J., 274, 2819–2828. Richmond, B.G., Lucas, P.W., Spencer, M.A., Schrein, C., Wroe, S., Ferrara, T.L., McHenry, C.R., Curnoe, D. & Wright, B.W., Byron, C. & Ross, C.F. (2010). The struc- Chamoli, U. (2010). The craniomandibular mechanics of tural rigidity of the cranium of Australopithecus africanus: being human. Proc. Roy. Soc. Lond. Ser. B: Biol. Sci. 277, implications for diet, dietary adaptations, and the allome- 3579–3586. try of feeding biomechanics. Anat. Rec. (Hoboken) 293, Wroe, S., Lowry, M.B. & Anton, M. (2008). How to build a 583–593. mammalian super-predator. Zoology 111, 196–203. Strait, D.S., Weber, G.W., Neubauer, S., Chalk, J., Rich- Wroe, S., McHenry, C. & Thomason, J. (2005). Bite club: mond, B.G., Lucas, P.W., Spencer, M.A., Schrein, C., comparative bite force in big biting mammals and the Dechow, P.C., Ross, C.F., Grosse, I.R., Wright, B.W., prediction of predatory behaviour in fossil taxa. Proc. Roy. Constantino, P., Wood, B.A., Lawn, B., Hylander, W.L., Soc. Lond. Ser. B: Biol. Sci. 272, 619–625. Wang, Q., Byron, C., Slice, D.E. & Smith, A.L. (2009). The Wroe, S. & Milne, N. (2007). Convergence and remarkable feeding biomechanics and dietary ecology of Australo- constraint in the evolution of mammalian carnivore skull pithecus africanus. Proc. Natl. Acad. Sci. USA 106, shape. Evolution 61, 1251–1260. 2124–2129. Wroe, S. & Musser, A. (2001). The skull of Nimbacinus Thomason, J.J. (1991). Cranial strength in relation to esti- dicksoni (Thylacinidae: Marsupialia). Aust. J. Zool. 49, mated biting forces in some mammals. Can. J. Zool. 69, 487–514. 2326–2333. Wroe, S., Myers, T.J., Wells, R.T. & Gillespie, A. (1999). Thomason, J.J. (1995). To what extent can the mechanical Estimating the weight of the Pleistocene marsupial lion, environment of a bone be inferred from its internal archi- Thylacoleo carnifex (Thylacoleonidae: Marsupialia): im- tecture. In: Functional morphology in vertebrate paleontol- plications for the ecomorphology of a marsupial super- ogy: 249–263. Thomason, J.J. (Ed.). Cambridge: predator and hypotheses of impoverishment of Australian Cambridge University Press. marsupial carnivore faunas. Aust. J. Zool. 47, 489–498. Tsafnat, N. & Wroe, S. (2011). An experimentally validated micromechanical model of a rat vertebra under compres- sive loading. J. Anat. 218, 40–60. Supporting information Weijs, W.A. & Hillen, B. (1985). Cross-sectional areas and Additional Supporting Information may be found in the estimated intrinsic strength of the human jaw muscles. Acta online version of this article: Morphol. Neerl. Scand. 23, 267–274. Werdelin, L. (1986). Comparison of skull shape in marsupial Figure S1. Muscle origin and insertion areas and lines and placental carnivores. Aust. J. Zool. 34, 109–117. of action of muscle forces for skulls of (a) Dasyurus Wiens, J.A. (1993). Fat times, lean times and competition maculatus, (b) Sarcophilus harrisii, and (c) Thylacinus cyno- among predators. Trends Ecol. Evol. 8, 348–349. cephalus constructed using finite element software (Strand 7, Wroe, S. (1996). Muribacinus gadiyuli (thylacinidae: marsu- Vers 2.4). pialia), a very plesiomorphic thylacinid from the Miocene Table S1. Values used to assign eight material proper- of Riversleigh, northwestern Queensland, and the problem ties to brick elements in heterogeneous models. Table S2. Force used for each jaw muscle division in of paraphyly for the Dasyuridae (marsupialia). J. Paleon- unscaled models of Dasyurus maculatus, Sarcophilus harrisii tol. 70, 1032–1044. and Thylacinus cynocephalus. These were calculated using Wroe, S. (2008). Cranial mechanics compared in extinct muscle mass proportions from dissected D. Virginiana marsupial and extant African lions using a finite-element (Turnbull 1970). approach. J. Zool. (Lond.) 274, 332–339. Table S3. Volume (mm3) and percentage volume (%) of Wroe, S., Brammall, J. & Cooke, B. (1998). The skull of each material property in the skull of each model. Ekaltadeta ima (Marsupialia, Hypsiprymnodontidae): an analysis of some marsupial cranial features and a re- As a service to our authors and readers, this journal investigation of propleopine phylogeny, with notes on the provides supporting information supplied by the authors. inference of carnivory in mammals. J. Paleontol. 72, Such materials are peer-reviewed and may be re-organized 738–751. for online delivery, but are not copy-edited or typeset. Wroe, S., Clausen, P., McHenry, C., Moreno, K. & Cunning- Technical support issues arising from supporting informa- ham, E. (2007). Computer simulation of feeding behaviour tion (other than missing files) should be addressed to the in the thylacine and dingo as a novel test for convergence authors.

300 Journal of Zoology 285 (2011) 292–300 c 2011 The Authors. Journal of Zoology c 2011 The Zoological Society of London