At first glance, it looks like a 1930s' documentary: Sturdy cogwheels, large-toothed gears and piston-like rods mesh in a proletarian symphony in grainy black and white.

But there are no people in the movie, and no sound in the room. What visitors are watching on the screen above their heads is not the industrial might of yesteryear, but the technological promise of tomorrow.

In fact, the gear in the film is about the size of a pollen grain. It can rotate at speeds up to 350,000 rpm, and is the principal moving part in the world's smallest engine.

The gear turns under the eye of a microscope in a small "design room" at Sandia National Laboratories here, and its image is projected where scientists can watch it as they plot what they say could be the world's next technological revolution.

Micromachines--mechanical devices so small they cannot be seen with the naked eye--have undeniable potential to inspire radical change in almost any aspect of human endeavor.

Among the anticipated innovations are tiny switches that could dramatically increase the capacity of fiber-optic telecommunications systems to deliver signals to computers and telephones.

Microgyroscopes could also keep cars from skidding and trekkers from straying in the woods. Soldiers in the field might carry a wristwatch-sized radio--beginning a new era in walkie-talkies--decades after Dick Tracy pioneered it.

There are an estimated 600 government, university and private labs doing microtechnology research worldwide. Eighty U.S. companies are working toward commercial applications.

Unlike "nanotechnology," which seeks to build devices on the atomic or molecular scale, microtechnology, which works in microns (millionths of a meter), graduated from theory long ago and has already made itself felt in a number of commercial applications.

Microscopic accelerometers trigger most of today's automobile air bags; micromachines spit the ink in ink-jet printers; microsensors monitor blood pressure in heart attack patients and regulate the fuel-air mix in automobile engines.

And enthusiasts say this is just the beginning. "We have developed some new technologies that we think are just going to revitalize the world," said Samuel L. Miller, supervisor of advanced concepts for Sandia's Microsystems Center. "It is a second silicon revolution, adding to microcircuitry complete systems that can sense, think, act and communicate."

Microdevices are small. The gear teeth in the world's smallest machine are a mere 7 to 8 microns across, about the size of a red blood cell, and at that size, gravity and inertia have little meaning. Miller tweaks a joystick at his elbow, and in an instant the cogwheel is spinning at 350,000 rpm in the opposite direction.

And microdevices are strong. Miller shows a videotape of the day a dust mite--a phenomenally ugly creature about half the size of a dandruff fleck--wandered into the micromachine.

The mite weighed about 100 times as much as the gear, but with another touch of the joystick, both gear and mite were spinning at several thousand rpm. When the gear stopped, the mite stepped down and wandered away.

And the machines are relatively easy to make, with the same procedures that engineers use to manufacture microchips. Layers of silicon are photo-engraved with the micromachine's design, then etched with acid to free the moving parts.

Scientists can make hundreds of thousands of microdevices on the surface of a silicon wafer the size of a grapefruit slice, and once the design is completed, manufacturing is easy. Like microcircuitry, everything is made in place--no assembly required.

Despite these advantages, however, the industry has been slow to develop, said San Francisco-based consultant Roger Grace, who has been tracking the progress of microtechnology for years.

"I'm a market analyst, and what it comes down to is, 'When are people going to make money from all this?' " Grace said. "When you look at the 'killer applications,' they are very limited."

Grace estimates that the industry sold $4 billion to $6 billion in microdevices in 1998, largely on the strength of a half-dozen proven technologies such as the ink jets and the air bag triggers.

But this is changing, Grace says, and he expects revenue to rise within five years to between $18 billion and $20 billion, as more firms build new devices to feed high-volume markets: "These are 'disruptive' technologies," he said. "We're not talking about a continuum. We're talking about a quantum leap."

Grace said the difficulty, in the past, has been that scientists have been too enamored of the "gee whiz" aspects of microtechnology and less focused on "consumer kinds of things."

But now, says Paul McWhorter, Sandia deputy director for microsystems, researchers are getting "past that." Without microtechnology, companies today can make dolls "that can tell when a child is holding its hand, or can follow the child around," he noted. But a microsensor could do the same thing for a tiny fraction of the cost, and a consumer "would never know it was there."

It was this combination of cheapness and compactness in 1995 that prompted Analog Devices Inc. of Norwood, Mass., to begin mass producing a fingernail-size device whose critical component is a lacy, spring-mounted microstructure that looks like a comb.

Put the device in an automobile, and fingers in the comb will vibrate gently within the fingers of a second stationary comb. But when the car is jolted, the space change between the sets of fingers sends out signals of varying strength. Jolt the combs hard enough and the car's air bag will deploy.

Before Analog, air bag triggers cost about $18 apiece, and cars needed at least three to ensure that one would fire. Analog's device costs less than $5, one to a car.

"We built an accelerometer and looked around to see where it could be used," said Analog spokesman James Fishbeck. "We were like Willie Sutton, we decided to go where the money is."

Still, even if the goal is to focus on everyday problems, micro-solutions are anything but mundane, particularly at Sandia, a Department of Energy lab that specializes in defense and nuclear weapons.

For instance, Patrick J. Eicker, Sandia's director of robotics, is overseeing a project to create armies of tiny robots capable of mapping a contaminated room: "If you have a bio-bomb, you'd rather have robots in there than humans."

Eicker calls his devices MARVs, "micro-autonomous robotic vehicles." They are about 1 1/2 inches long today and probably will be less than a quarter-inch long when they finally debut--"about the size of a pebble," Eicker says.

He has a "herd" of 35 MARVs and researchers have already developed software that gives them what he calls "a collective consciousness."

Send them on a job, Eicker explains, and they will organize themselves to map the contamination, share the data among themselves and figure out a way to transmit it. Like the Borg collective of "Star Trek," what one MARV knows, all know.

"The really cool part is that if you have a hundred cockroaches running around the room, and you stomp on a bunch of them, you still have a map," Eicker said. "The collective owns it."

MARVs, strictly speaking, are not microdevices--at least not yet--but everything they carry will be a microdevice. Sandia is developing a microscopic device known as a gas chromatograph that will sample and test for chemical and biological agents.

And while MARVs might sneak into an Iraqi factory, or eavesdrop on terrorists, they could also be just as useful mapping a nuclear accident like Chernobyl or searching a collapsed building for earthquake survivors.

Supposedly "nobody can develop a robot vacuum cleaner, because there's too much stuff on the floor, but when I saw we were going to map Chernobyl, it's the same thing," Eicker said. "People like Maytag come and visit us."