R McNeill Alexander, an expert in biomechanics, explains how palaeontologists are using engineering ideas to look back at the past.
You can argue over whether it's a good or a bad thing that we can't observe dinosaurs moving - would you really want to meet a T rex? But since we definitely can't, we have to use the next best thing: fossil evidence.
The most direct form comes from fossil footprints formed in mud that has since turned to stone. We can use these to calculate speeds of movement, because the faster an animal runs, the farther each footprint is from the one before.
Of course a long-legged animal such as a horse takes longer strides than a short-legged one such as a dog, even when both go at the same speed, so we have to make allowance for the very long legs of the large dinosaurs. Unexpectedly, this can be done by using Froude numbers, an idea taken from nautical engineering that was devised to take account of size differences when measurements on models are used in designing ships.
Froude number calculations for all known tracks of large dinosaurs indicate low speeds - about as fast as human walking. But there are some remarkable tracks of a medium-sized (half-tonne) dinosaur showing strides so long that it must have been running fast enough to win the 100-metre sprint at the Olympics.
The low speeds calculated from the tracks of really large dinosaurs do not prove that they never ran, only that they usually walked. Fast running needs strong legs: because each foot is on the ground for only a small fraction of the time, the forces while they are on the ground must be large. You might think we could learn more by measuring the strength of dinosaur bones, but the strength of a bone is changed in the process of fossilisation, as the organic matter decays and is replaced by minerals. However, we can calculate the strengths that the leg bones would have had in the living dinosaur from the bones' dimensions, in the way that engineers calculate the strengths of girders.
We find that some of the largest dinosaurs had legs strong enough to move quite briskly. For example, the 35-tonne herbivore apatosaurus could have been as athletic as a three-to-five-tonne elephant, able to run at moderate speeds.
James Farlow of Indiana University has suggested that T rex could not run fast because its tiny arms could not save it if it tripped. A rough calculation, based on research on traffic accidents, confirms that a fall would have hurt it badly, but in any case the dimensions of its leg bones show that they were too weak for high speeds. Despite anything you may have seen in Jurassic Park, T rex would have been no good at chasing jeeps.
Sauropod dinosaurs, such as diplodocus, had very long necks and tails. A suggestion that they reared up on their hind legs to feed on the leaves of tall trees seemed hard to believe, until experiments showed that because their tails were so long, their centres of gravity were close to the hip joints. That implies that most of the animal's weight would have been supported by the hind legs, even when it was standing on all fours. All it would have had to do to rear up is to lean back a little, to bring its centre of gravity directly over the hind feet. The long, heavy tail would have enabled the animal to balance on two feet.
This seems to explain the heavy base of the tail, but the tail then tapers down to a long and thin, whip-like end. Some years ago I suggested (tongue in cheek) that the tail could have been cracked like a whip. But Nathan Myhrvold of Microsoft Corporation has recently shown by computer modelling that the idea is feasible for diplodocus. A sudden side-to-side movement of a whip or any other long, flexible structure will send a wave travelling along it. If the whip or tail tapers from a thick base to a thin tip, the motion will get faster and faster as the wave travels down it (because of conservation of momentum). If the tip is thin enough in comparison to the base, it may exceed the speed of sound and a crack - in fact a miniature sonic boom - will be heard.
Diplodocus had a tail that tapered enough for quite a gentle twitch at the base to have caused supersonic movement at the tip. Perhaps male diplodocus competed for females by tail-cracking matches. I imagine something like the roaring matches of stags, which demonstrate their stamina by roaring repeatedly before engaging in a real fight with their antlers. (If you can roar louder and longer than I can, at a faster rate, I might be wise to sneak off without fighting.)
There is good evidence for one other dinosaur sound. parasaurolophus was a 10-metre bipedal herbivore with an amazing, metre-long crest. Its nasal cavity extended all the way up the crest and down again, like the looped tube of a trombone. It would have resonated at a particular sound frequency that can be calculated from the laws of acoustics, just like the pitch of an organ pipe. It seems likely that it used this resonance to enhance its voice, and the calculations tell us that the pitch would have been at the lower end of the range of a trombone.
Recently, scientists at the New Mexico Museum of Natural History have made a more detailed analysis of the acoustics of parasaurolophus's crest that enables them to reproduce the quality of the sound as well as the pitch.
The crest was too fragile to have been used as a weapon, so the voice was probably used in a different way from diplodocus's tail crack - not to scare off rival males, but to attract females. It seems likely that female parasaurolophus thought long crests and deep voices were sexy, just as peahens prefer peacocks with long tails. It is well known that preferences can evolve for extreme and irrational male fashions. (You can hear the sound on the Internet at http:// www. nmmnh-abq.mus.nm.us/ nmmnh/ parasound.html - the hyphen is part of the address.)
In most of these examples, biologists have done the research, taking their methods from engineering, but a current project is led by engineers using biologists' advice. The Palaiomation Consortium is building a sheep- sized robotic iguanodon for museum display.
It is fully autonomous, carrying its own computer and batteries, much more sophisticated than existing animatronic exhibits. It walks around choosing its own path and avoiding obstacles. So far it moves very slowly, but we plan to speed it up in order to make it do more things.
We are taking robotics to its current limits in our efforts to build a dinosaur that seems to be alive. Sadly, a robot is as near as we are likely to get to a living dinosaur. Molecular biologists told us last week at a meeting at the Natural History Museum that they have no hope of recovering enough dinosaur DNA to make dinosaurs as imagined in Jurassic Park.
This is a shortened version of a public talk at 6.30pm tonight at the Zoological Society of London. Call 0171-449 6261 for tickets and details.
Professor Alexander is professor of Zoology at the University of Leeds and secretary of the Zoological Society of London.