Steve Farrar reports on the team of scientists striving to build the world's largest telescope and the anthropologist who is observing them

When the multifaceted eye of the California Extremely Large Telescope opens in a little over a decade, its gaze will bring the heavens closer to Earth than ever before. With a primary mirror 30m in diameter, Celt will be three times the size of any existing telescope and gather nine times as much light. It will be able to pick out detail that has up to now escaped astronomers: the most distant galaxies and most subtle phenomena, perhaps even extra-solar planets. But at the moment it exists only in the minds of a small but remarkable group of Californian scientists and engineers who are about to take an unprecedented technological leap.

Celt will be possible only with a host of pioneering features, most daunting of which is a laser-generated constellation of atmospheric "stars" to help counteract the blurring effects of the atmosphere. The Celt team is led by the University of California's Jerry Nelson, project scientist at the W. M. Keck Observatory, home of the world's largest existing telescopes, and by Caltech's Richard Ellis, director of the Palomar Observatory, home to a previous largest telescope. The team is bullishly optimistic. It needs to be.

Not only is the task going to be extraordinarily difficult, but there is competition. Two rival European plans for similarly large telescopes pre-date Celt, and the US National Optical Astronomy Observatory is considering its own plans. Celt's genesis was prompted by recent European advances that, in turn, had followed Californian innovation. The 10m Keck telescopes are about to be eclipsed by the European Southern Observatory's Very Large Telescope, and then it emerged that ESO was considering building a 100m telescope.

ÖйúAƬ

Size is the principle currency of observational astronomy and California's astronomers, long used to being the best-endowed in the world, were stung into action. Europe was putting pressure on the Americans, Nelson says. He welcomes competition, but says: "We would hate it if the Europeans built a giant telescope and we didn't."

While ESO's 100m proposal and a Swedish 50m plan are still at their earliest stages, the Californians have moved fast to catch up. Tom Tombrello, chair of Caltech's division of physics, mathematics and astronomy, set up a steering committee to explore the idea last summer. The telescope's size would be dictated to a large degree by what was technically and financially possible. They agreed on 30m. "It was as much as we could imagine doing in one jump," Tombrello says.

ÖйúAƬ

Within months, the steering committee was slimmed down to a focused design team, including several of the University of California Keck pioneers. They already had indispensable experience of working on telescopes that had been criticised as too radical when first proposed.

New blood was added to this core: brilliant young physicists and engineers, eager to make their mark, and established scientists, including several Britons, with dazzling records and still brimming with enthusiasm. Ellis quit his post as director of Cambridge University's Institute of Astronomy in 1999 to lead the scientific effort. Then Tombrello added an unexpected ingredient - anthropologist Marianne de Laet.

De Laet's presence in meetings and her interviews with team members give her access to a major scientific project at its outset. She sees her brief as studying "how this bundle of visions and notions of what might be possible fits together and what it might turn into". Her study might be used to help future groups approach similarly tough challenges.

When de Laet was first introduced to the team, Tombrello advised members:

"Think of her as the Jane Goodall of astronomers," and then turned to Nelson and told the anthropologist: "Here's your alpha male." The scientists and engineers seem pleased to have de Laet on board. They have even taken to calling themselves "Marianne's chimps".

The pictures Celt will paint of our universe will outstrip the best Hubble Space Telescope snapshot. Past experience suggests there will be many pleasant surprises. Palomar's 200-inch Hale telescope found quasars and vast intergalactic gas clouds. Since Keck assumed Palomar's mantle as the sharpest eye on Earth in 1993, it has pinpointed gamma ray bursters and revealed the accelerating expansion of the universe.

ÖйúAƬ

If Celt is to add to this list, significant hurdles have to be cleared. Not least, the need to keep the cost down. Simply scaling up Keck would give Celt an impossible $1.5 billion (?1.05 billion) price tag. To stand a chance, the innovations have to involve a cost saving of at least a factor of four. While Keck's 10m primary mirrors are composed of 36 hexagonal segments, Celt will require 1,000, each positioned with nanometre precision to work in concert as a single piece of reflective glass.

Every bit of this insect-eye mosaic has to be polished to an exact asymmetric shape and kept in place by a system of computer-controlled sensors and precision pistons or actuators to preserve the overall shape of the 30m mirror against the distortions of gravity, thermal stresses and the wind. Douglas MacMartin, a control systems engineer, quit his job tackling vibrations in helicopters to develop a way to control 3,000 actuators and 6,000 sensors on a microsecond time scale. But he feels reliability may give him his biggest headache: "If every actuator lasted on average 100 years, then you would have one fail every ten days. Quite simply, you need to design actuators that do not fail."

ÖйúAƬ

Keith Taylor, former head of instrumentation at the Anglo-Australian observatory, joined the team to work out how to extract the most interesting information from the light Celt collects. Back-of-the-envelope calculations suggest instruments will be the size of a tennis court. Each will perform a single task, from producing sharp images to studying the spectrum of the light of thousands of objects simultaneously. Every job is tough, but all agree that Richard Dekany has the toughest job of all.

Dekany has to overcome the blurring effect that the Earth's atmosphere has on celestial images through a technique called adaptive optics. This will involve monitoring the distortions of a constellation of natural and possibly "artificial stars" created around the real focus of the astronomers' attention by training a battery of lasers on free sodium ions 90km up in the atmosphere. Powerful computers use the information gathered to calculate a constantly changing model of the atmosphere's turbulence in the sky above the telescope and hence control the precise shape of high speed, deformable mirrors to compensate for every twitch and wobble. Ridding starlight of atmospheric interference is pushing into unknown territory. Dekany knows a lot rests on his success.

Testing much of Celt's new technology will give the Palomar Observatory a role at the cutting edge, postponing its inevitable retreat from the frontline of science. Meanwhile, the exhilarating research continues apace. De Laet views Nelson as the centre of the network of connections that Celt has created, though not necessarily higher up any hierarchy. "It is a collaborative effort rather than a competitive one, and the innovations and insights emerge from the interactive process," she says.

The brainstorming will have to end some time. A draft proposal verifying that there are no "show-stopping"problems should be filed by October and a project manager will pull the effort together some time later. The final design should emerge by 2006.

Negotiations with possible donors to fund the next stage are under way, with the prospect of a race against Europe to build the biggest telescope helping entice support.

Ironically, while Celt is pencilled in to join the community of observatories on the summit of the dormant Hawaiian volcano Mauna Kea - Keck among them - environmental concerns may yet force it to seek an alternative home. The fallback location would most likely be in Chile, where ESO's Very Large Telescope is sited, In which case, Californian and European giants might eventually wind up as neighbours.

ÖйúAƬ