Old softies get a buff and grind

Ancient invertebrates are hard to study but, Geoff Watts reports, computers and recycled car parts can help palaeontologists view the structures of soft-bodied creatures

Spade, trowel, hammer and brush: a few of the traditional tools of the palaeontologist. But a group of fossil hunters in Oxford has found the need for two more: a computer with good graphics software and the parts from an old windscreen wiper.

The product of this unlikely combination is a method of viewing the structure of fossil invertebrates that would otherwise be difficult to envisage, let alone study. Most palaeontology is focused on the hard bits of long-gone animals: skeletons, shells and other structures likely to remain intact long enough for the slow process of fossilisation to preserve them. The remains of mainly soft-bodied animals are few and far between. When they are found, they are difficult to study.

Such was the problem confronting Derek Siveter and Mark Sutton of the department of earth sciences and the University Museum of Natural History in Oxford. Together with colleagues in Bristol and Leicester, they wanted to study some worm-like molluscs dating from the Silurian period 425 million years ago.

The material itself comes from a site in Herefordshire and is buried inside anonymous-looking lumps of stone: nodules, to use the jargon of the trade. Siveter was alerted to the possibility that these nodules might be worth examining when, several years ago, someone showed him the exposed surface of a broken one.

Concealed within it was a small portion of the common mineral calcite. A closer examination showed that this was the fossilised remains of an arthropod with well-preserved appendages.

The outline of the animal was visible because the voids left after the decay of its soft parts had been replaced by a precipitate of calcite. "This probably happened following a volcanic event," Siveter says. "The ash that went into the atmosphere and settled into the marine environment was important in some way. It produced just the right chemistry for preservation."

These nodules have since been found to house several groups of soft-bodied creatures. When dealing with the fossilised remains of a skeleton, the palaeontologist would hope to reveal all by carefully chipping and scraping at the material surrounding it until the object itself had been exposed.

But this approach will not work when all you have is a pattern of small, calcite-filled voids. Looking at the exposed surface of a cracked nodule and seeing a cross-section of its long-entombed inmate is not unlike breaking a stick of seaside rock and reading the words "Blackpool" or "Brighton".

The difference with a stick of rock is that most breaks will be more or less transverse to the long axis of the material forming the word, and you see the same thing at whatever point you snap it. Neither of these things happens with nodules.

In a small yard behind the museum, Sutton demonstrated the Oxford technique. The nodules, most of them a few centimetres in diameter, are first broken using a screw press that applies a cold chisel. When the nodule breaks, with an explosive crack, a glance at the exposed surface is sufficient to reveal anything of interest. If it looks promising, the surface is photographed using a digital camera.

The half nodule is then mounted in a holder and placed face downwards on a lubricated grinding surface. This is where the recycled windscreen wiper comes in; it provides the motive power. About five minutes of grinding will remove a few tens of microns. Then the surface is photographed again. "We grind and photograph, grind and photograph until the fossil is completely gone," Sutton says.

Larger examples may require several hundred such slices. But whatever the number, the fossils inevitably finish up as dust. "That's the paradox in this method," Sutton admits. "But the information we get out more than makes up for the loss."

The next step is to assemble the digitised images into a whole. Biologists have been using serial section for donkey's years to establish the three-dimensional form of objects. "Palaeontologists too have been trying to do this for a long while. But it's only in recent years that we've realised that computers are the way to do it," Sutton says.

The images of the successive surfaces are fed into the computer and aligned so that they stack perfectly. Fossil material is generally lighter in colour than the rock matrix, so the computer can be instructed that anything over a certain brightness is fossil rather than matrix. With the right software, the computer can combine the cross-sections to create an image of the entire creature.

This "virtual fossil" can be rotated, have its various structures coloured for greater clarity, or even have them removed to see what's behind them: an exercise in "virtual dissection". The original streaks and smudges on the surface of the half nodules come magically to life.

The Oxford researchers are not the first to have tried this technique. But they have succeeded when others have not. "In the past," Sutton says, "most of those who've tried to do this have been either computer people or palaeontologists. There haven't been many groups that could bring these two things together, and that's what we've done."