EDEN REGAINED.

Spectacularly colonizing a Cornish china clay-pit, the Eden Project is a monumental palm house for the twenty-first century, its ingeniously engineered biomes inspired by natural processes and structures.

Honeycomb, flies' eyes, frog spawn, cuckoo-spit -- choose your organic simile. Built to contain biological specimens, the biomes of the Eden Project look like giant biological specimens themselves, some kind of fungus from outer space, perhaps, fruiting weirdly in this worked out Cornish china clay-pit. The design seems to have been inspired by natural and/or science fiction images but, though some Grimshaw buildings are indeed image-inspired, in this case the impression is misleading. The inspiration was not what nature looks like but how it works, its processes and structures. The fact that the Eden Project is a ready-made set for Quatermass and the Pit has been useful in the marketing of the whole enterprise, but it was a by-product rather than the starting point of the design.

The greenhouses had to be sited in the unshaded strip at the foot of the cliffs on the north side of the pit. The first idea was for a linear, lean-to structure rather like Grimshaw's International Terminal at Waterloo station (AR September 1993). This form posed a number of problems, however. For one thing the three-dimensional profile of the site, far more complicated than the level curve of Waterloo, meant that it was difficult to use cheap, standardized components. To make matters worse, the ground profile was constantly changing during the development of the design, because the site had not yet been taken over by the client and was still being quarried. A long-span, arched structure would have been heavy, bulky and difficult to carry down into the pit. It would also have cast unwanted shadows on the plants inside. A more promising alternative was a much lighter and more economical geodesic dome, but it had the wrong plan-form and would have been impossible to divide up into different zones. The idea of a line of smaller, intersecting geodesic domes was arrived at late in the day, but it solved all the problems at once and made the project possible.

It works like this: take a row of spheres of different sizes, made like footballs out of two-dimensional hexagons and pentagons, and squash them into one another, forming perfect circles where they intersect. Then squash the whole row into the site, in the angle between the cliff and the quarry bottom. Circles become arches, and the hexagons and pentagons are removed as necessary around the perimeter to accommodate the irregular ground profile. Structural components, mainly of tubular steel joined by spherical nodes, are identical in each dome and small enough to be easily handled. These are not conventional domes in that they exhibit tensile as well as compressive structural behaviour. The outer compressive grid is linked by tetrahedrons to an inner tensile grid. The double grid is necessary because the lattice steel arches break the continuity of the structure. For the same reason, the domes were not self-supporting during erection but had to be assembled from a temporary scaffold so big that it has entered The Guinness Book of Records. This is a slight disappointment for techno-organicists raised on Buckminster Fuller (nature does not use scaffolding), but there is nothing heavy or awkward about the finished structure. The geodesic grid is scaled according to the size of each dome and except in the smallest dome, where it becomes rather dense, the effect is amazingly light for such enormous spans. At the junctions with the arches, the grid is adapted ad hoc, creating irregular geometrical shapes. Architecturally, this may seem a worrying inconsistency, but it is exactly what happens in nature when, for example, the hexagonal grid of veins in a dragonfly's wing meets a leading edge or a structural spar.

The largest hexagons are 11 m across and therefore impossible to span with a single sheet of glass, especially since it would have to be double glazed and toughened. The lightness of the structural grid is made possible by a new high tech material -- ethyltetrafluorethylene foil (ETFE). This light, transparent, flexible film forms triple-membrane cushions which are kept inflated by a constant low pressure air supply. Because they were formed and fitted on site, the ETFE cushions could adapt easily to geometrical variations without any need for complicated scheduling or production planning. The biomes are beautiful structures because they are efficient structures -- a kind of beauty common in nature but rare in architecture.

Like their humbler horticultural cousins, however, they also have a rugged practicality. The branching network of flexible air-supply pipes, for example, is clipped to the structural steel members with no attempt at concealment. The heating and ventilating system simply consists of free-standing air handlers in ordinary metal boxes placed at intervals around the perimeter, poking their twin circular ducts straight through the walls of the domes. Such artless functionalism is easy to accept, though the heavy duty adjustable glass louvres associated with the ducts are perhaps a little too clumsy, their insistent linearity stubbornly at odds with the fluidity of the geodesic grid.

But once inside the enormous bubbles of the Humid Tropics Biome, such details are insignificant. A winding gravel path climbs up through what will be a dense forest (the planting is still immature) to a big, noisy waterfall. Though we can never quite imagine that this is a real rainforest, it is nevertheless a unique spatial experience, certainly more like nature than architecture. The sheer size of the enclosure, the word 'biome' and the very name 'Eden Project' all lead you to expect a complete ecosystem, or at least an approximation of one, but it soon becomes clear that this is really just a botanical garden, the Palm House at Kew writ large. There are no animals, apart from the crowds of people. The neighbouring Warm Temperate Biome is smaller and more comfortable, not just because it is relatively cool and dry, but because the structure of the domes is close enough to give it scale. It feels more human, more like architecture, though the technology is exactly the same.

In early versions of the design, the entrance to the biomes was housed in a chain of very small domes. This proved to be too fussy and expensive, but it was hard to imagine any kind of conventional building that would look comfortable between the big domes. The answer was to bury the building in the ground, reducing it to a few simple planes - a curved, grass-covered roof, a glass curtain wall and an entrance bridge leading to a first floor concourse overlooking restaurants below. Another curved, linear, earthbound building forms an artificial crest high on the opposite ledge of the pit. Visitors arrive at the back of this building from the cascade of car parks beyond, pay their entrance fees and emerge onto a terrace, cameras at the ready for their first view of the whole site. From here they make their way down to the entrance bridge through a richly cultivated open air theatre - the 'roofless biome'. Compared with the biomes, which express a compelling engineering logic, the ancillary structures seem rathe r sketchy and artificial. The arrival building (AR August 2000), for example, which houses shops, cafes and offices, is elegant and well planned but its use of materials like shingles, rammed earth (taken from the clay-pit) and gabions, seems more like a symbol of green construction than the real thing.

But then the Eden Project is not an architectural expo: it is a theatre in which humankind's relationship with the plant world is dramatized. The specimen plants are magnificent, the garden arrangements are imaginative and the scale is breathtaking. The crowds in the biomes soon forget about the delicate net arching high over their heads. They have come to look at the plants, not the greenhouses.

EDEN PROJECT, CORNWALL, ENGLAND

ARCHITECT

NICHOLAS GRIMSHAW & PARTNERS

Architect

Nicholas Grimshaw & Partners, London

Project team

Nicolas Grimshaw, Andrew Whalley, Jolyon Brewis, Vincent Chang, David Kirkland, Michael Pawlyn, Jason Ahmed, Vanessa Bartulovic, Dean Boston, Chris Brieger, Antje Buithaup, Arnanda Davis, Florian Eckardt, Alex Haw, Perry Hooper, Bill Horgan, Oliver Konrath, Angelika Kovacic, Quintin Lake, Richard Morrell, Tim Narey, Monica Niggemeyer, Killian O'Sullivan, Debra Penn, Martin Pirnie, Juan Porral-Hermida, Mustafa Salman, Tan Su Ling

Structural engineer

Anthony Hunt Associates

Services engineer

Ove Arup & Partners

Landscaping

Land Use Consultants

Glass louvres

M&V

Photographs

All photographs were by Peter Cook/VIEW except no 7 by Chris Gascolgne/VIEW