Video: ‘Crash-tested’ skulls throw light on extinctions
STEPHEN WROE’S lab is no place to be alone in the dark. Propped against a wall is a massive set of jaws from a great white shark. On shelves above the desk are skulls that once belonged to other fearsome animals – a sabre-toothed cat, a clouded leopard, a “Tasmanian tiger”.
Despite their magnificence, what has drawn me to at the University of New South Wales in Sydney, Australia, is not the skulls themselves but the highly realistic 3D virtual models he has based on them. Unlike the real thing, these digital crash test dummies can be placed in various positions and then stressed, strained and generally pummelled in order to deduce the ways in which the real live predators killed their prey.
Advertisement
For extinct species, the knowledge gained from such tests is helping palaeontologists build up a picture of how various species may have competed against each other for food, leading to new theories for why some died out while others survived. For existing species, the same reasoning could help predict whether a predator will be able to survive changes to its environment due to global warming or loss of habitat. If a predator’s main source of prey is dwindling, for example, the models could help predict whether it would be able to switch to feed on a different, available animal instead, and successfully compete with rival predators. “In communities under stress, it’s typically the big predators that go extinct first,” Wroe points out.
The researchers’ secret is a technique called , which allows them to model the transmission of stress and strain through a material. FEA has been used for more than 30 years by architects and engineers to model the wear and tear on buildings and aircraft, but a great deal of processing power is needed to model complex biological systems and the technique has only recently become affordable for modelling skulls.
The first step in FEA is to create a digital representation of an object. For animal skulls, a computed tomography (CT) scanner images the entire skull in three dimensions. Commercially available imaging software then turns this exquisitely detailed map into a surface mesh – a “skin” of tiny triangles depicting the surface topology of the object. Next, a “finite element” program converts that mesh into a 3D object made up of millions of tetrahedral blocks, or “elements”. Each element is programmed to react realistically to the movement of adjacent elements, allowing researchers to apply virtual stresses – such as those produced by teeth tearing at flesh – and view how the skull as a whole responds. “FEA is the only modelling technique that allows us to deduce the stresses and strains experienced in structures with complex shape and complex material properties,” says Emily Rayfield at the University of Bristol in the UK.
In 2001, Rayfield published the first FEA of a palaeontological specimen – a Jurassic-era carnivorous dinosaur called Allosaurus fragilis (Nature, ). Three years later, her FEA of a Tyrannosaurus rex skull revealed that a series of gaps between the facial bones probably functioned as shock absorbers, allowing tyrannosaurs to take powerful, bone-crushing bites without overstraining their craniums.
Bone breakdown
Since then Wroe has refined the FEA technique. Both of Rayfield’s models consisted of 2D meshes and did not include the jaw. Wroe and his team now work with 3D tetrahedra and have developed a method for modelling a hinged jaw. They have also created software that calculates bone density from the CT images instead of treating the skull as if it had a uniform firmness, which was the previous standard. Wroe’s team scores each bone element on a scale from 1 – for the densest, hardest cortical bone – to 8, for the softest, spongiest bone. This provides a more realistic transmission of force, says Wroe. Meanwhile, a method for rapidly turning the CT images into FEA models, also devised by Wroe’s team, means that if a skull is in relatively good condition it will take just a few hours to create a model, rather than weeks or months.
These models are already increasing our understanding of some extinct species. In September 2007, an FEA model of the skull of the extinct “Tasmanian tiger”, or thylacine (Thylacinus cynocephalus), revealed that its skull was not adapted for biting large prey. Humans have been blamed for the thylacine’s disappearance from the Australian mainland 3000 years ago. But Wroe’s team’s analysis suggests that the dingo, whose relatively stronger head and neck could better handle the stresses of tackling big prey, may have outcompeted the thylacine when prey became scarce (Proceedings of the Royal Society B, ).
More controversially, a study led by Wroe’s colleague Colin McHenry shows that, despite its fearsome canines, the extinct sabre-toothed cat Smilodon fatalis had a surprisingly weak bite (Proceedings of the National Academy of Sciences, ). While the animal’s skull and neck were well adapted for administering a single killer bite to large prey, he found that the smilodon could only apply a modest bite force for its size – weaker than that of an African lion. “A lot of people had a problem with that,” says Wroe. “I think a lot of Americans weren’t impressed that their iconic super-predator had a bite a bit more like a pussy cat.”
“Many Americans weren’t impressed that their iconic sabre-tooth had a bite like a pussy cat”
In future, FEA might do a similar thing for the history of humans. Wroe’s team is preparing to study the skulls of modern and ancient hominids to investigate theories that different feeding habits might help to explain why Homo sapiens was successful while other lines died out.
Wroe is also working on living animals. His models of the great white shark (see Picture) suggest its jaws could tolerate a bite force equivalent to about 1.8 tonnes. “That’s a heck of a lot, but for an animal that size it’s not humongous,” he says. With its extremely sharp teeth, the great white – like the smilodon – might not need a supremely powerful bite. Its skull seems adapted to causing swift catastrophic trauma, allowing it to back off until its prey dies. The team is also looking at the tiger shark and the bull shark. Dissections have already revealed large muscles, so they expect both to pack a bigger bite than the great white.
The team has also built a virtual model of a Komodo dragon’s skull, with intriguing results. They found that it is superbly adapted to perform one killer motion: to simultaneously bite into prey using its jaw muscles, while pulling using its neck and body muscles. The results, which will appear in the , suggest the Komodo’s skull is under less stress when it both bites and pulls than when it just bites. “The combination of very sharp, wickedly serrated teeth and very clever engineering allows the big lizard to inflict serious damage despite the fact that its jaws are weak and its skull very gracile,” says Wroe.
FEAs of skulls also have practical applications. A knowledge of bite forces of various sharks could help engineers improve the design of underwater equipment, such as submarine sonar arrays, to stop them from becoming chewed up, says at the University of South Florida in Tampa, who is working with Wroe on shark modelling. Non-biting applications for FEA skull models are also in the works (see “Stressed head-bangers”). Researching the mechanics of killing in big scary animals is childishly fascinating in part, Wroe confesses, but there are justifiable scientific reasons for doing so.
Evolution – Learn more about the struggle to survive in our comprehensive special report.
Stressed head-bangers
As well as helping to tease out the feeding habits of some of the world’s scariest predators, virtual skulls should have some non-biting applications.
, an oral and maxillofacial surgeon based at Westmead Hospital in Sydney, Australia, is poised to use the computer modelling technique called finite element analysis (FEA) to improve the repair of facial fractures, together with Stephen Wroe of the University of New South Wales.Currently, the selection of plates and screws to fix a specific type of facial fracture is determined by a polymer model of a jaw with that fracture. Surgeons place a plate across the fracture and apply various forces to test its strength. But Aquilina suspects that sometimes this results in plates being selected that are either too thick or too thin. A computer model would more closely mimic the way a jaw reacts to force and so might lead to better plate selection, he says.
The technique might also help save whales. About one-quarter of the deaths of critically endangered North Atlantic right whales are caused by collisions with ships. at the Woods Hole Oceanographic Institution in Massachusetts is using FEA of the animals’ skulls to determine at what speeds ships become a deadly hazard to the whales, in order to recommend speed restrictions that should cut the death toll. “Our work is the first to approach speed restrictions from a biomechanical perspective,” she says. Ocean-going vessels typically travel at up to 22 knots. Preliminary evidence suggests that restricting speeds to below 14 knots in certain situations would greatly reduce the probability of a fatal fracture being caused to a whale as a result of a collision, she says.
