What's the difference between a four-karat emerald-cut diamond and a lump of coal? One doesn't throw any heat, and the other makes a funny- looking ring, but the critical distinction is one you can't see at a glance. Because all that really distinguishes anything from everything else -- healthy tissue from a cancerous lump, sand from a silicon computer chip, or a tabby cat from a Bengal tiger -- is an invisible arrangement of atoms.

What if we could alter that matrix, shifting atoms the way we move bricks and mortar, restructuring the commonplace or creating new materials? Technology does that now in a relatively primitive way: Chemists produce new drugs and compounds by mixing reactive ingredients to coax atoms into different molecular structures.

But this is baby stuff compared with the imminent possibilities of molecular engineering and its promise (some would say threat) of a startlingly new world. Imagine, for instance, being able to move atomic particles about with sub- microscopic machines -- molecule-size cranes and cell-size robots scooping up atoms from one place and shoving them into a custom design. Atoms, after all, are objects with mass and size. Manipulating them would let us create original materials: diamond-hard carbon-based metals as cheap as firewood; tasty, nutritious food produced right in our homes; powerful computers smaller than living cells.

This is speculation but it's not fantasy, says K. Eric Drexler, who warns it's time we started thinking big thoughts about small things. Drexler is a 31-year-old MIT scientist and the author of Engines of Creation (Anchor/ Doubleday), the first book on the mind-boggling field of molecular engineering known as nanotechnology. Drexler is also the founder of MIT's Nanotechnology Study Group, a gathering of biologists, physicists, chemists and computer scientists who share hopes and fears about what might happen once we wander across this technological frontier.

The word comes from the Greek "nanos," meaning dwarf; "nano" is a prefix indicating one billionth. Working with individual atoms, which is what nanotechnologists plan to do, would put industry and science on the scale of a billionth of a meter. Nature already operates at nanolevels, of course -- living cells are tiny biological machines programmed to take direction from chemical signals.

Nanotechnologists like Drexler believe it will be possible one day to build robots with molecular-scale parts that, like cells, can be directed to perform specific duties. Called "nanoassemblers," they would enable us to construct matter atom by atom.

Scientists are already using incredible tools, such as the scanning tunneling microscope and the atomic-force microscope, to see molecules and atoms in action. As more powerful tools are developed, that capability will grow more sophisticated, leading to the construction of the first assemblers, Drexler predicts.

These will probably be built of protein, and will produce a more advanced, second generation of micro-manipulators made of plastic or ceramic. Drexler says once nanotechnology enters that phase, we will be able to make almost anything. Our creations will have to follow the laws of nature, but that hardly limits the prospects.

Cells multiply to create blue whales and bacteria, redwoods and radishes -- everything that exists. With nanoassemblers and molecular blueprints, we could rapidly do the same. A bacterium can replicate itself in a little more than 15 minutes. Replicating assemblers operating at roughly the same speed, each reproducing itself, would multiply from one assembler to one trillion in 10 hours. In less than a day, says Drexler, we'd have a ton of new material.

We could feed the world with inexpensive food, growing it quickly in tanks and vats -- producing meat without slaughter, and vegetables, fruits and grains without farmland. We could inject computer- controlled molecular machines smaller than cells into a sick person's bloodstream, directing them to seek and destroy diseased cells or stimulate the growth of new organs and tissue.

In nanotech factories (which would produce no toxic wastes), we could grow not only sheets of super-strength, lightweight building materials, but also complete structures -- including subminiature computers a trillion times more compact and a million times more efficient than anything we have today.

From chemical stews, we could produce rocket engines that would make space travel as accessible as an ocean crossing. We could even begin healing our mismanaged biosphere, converting excess levels of carbon dioxide back into coal and oil.

Applied properly, says Drexler, nanotechnology could end poverty, prolong human life to near immortality and create a decentralized industrial base that would transform the planet. But nanotechnology could also create overwhelming problems, beyond our wildest predictions.

Some might imagine that the greatest risk would be the possibility of a nanotech industrial accident, in which some bionic bozo escapes from a vat of chemicals and ravages the populace. In fact, the real threat lies not in the potential for inadvertent accidents with nanotechnology, but in the potential for malice and deliberate abuse.

The military, for example, will no doubt take a microscopic look at nanotechnology's potential for growing engines of destruction. "Smart" viruses could be programmed to make germ warfare cheap and strategically effective. Just think (with an appropriate shudder) what terrorists and politicians -- and sometimes it's hard to tell them apart -- could do with that.

Even if only some of the predictions come true, nanotechnology's possible socio-economic effects could severely disrupt traditional social systems, making the agricultural and industrial revolutions look like historical flyspecks. So our worst mistake would be to dismiss this as science fiction.

Drexler and the Nanotechnology Study Group say now is the time to plan for the age of molecular engineering, lest we be caught unprepared for the decisions we will have to make. We may not have much time. One of Drexler's friends says an optimistic estimate for the arrival of nanotechnology is 30 years. The pessimistic estimate, he says, is 10 years. ::