GAITHERSBURG - Locked in an underground, climate-controlled government laboratory is one of the most pampered and protected scientific treasures in the country.

K20, as it's known, doesn't look like much - a shiny cylinder of platinum and iridium about the size of a large pill bottle. But since 1890, this object has served as the official U.S. kilogram, the standard by which mass and weight are measured under the metric system.

So powerful is K20 that every scale across the land must ultimately measure up to it. Even the value of the U.S. pound is derived from it.

But if a small number of scientists have their way, K20's reign may soon come to an ignominious end. Because the object, they know, also harbors a dirty secret: The kilogram has a weight problem.

At last check, K20 was off official weight by a few billionths of a gram. It might not sound like much, but over time "the drift becomes a bigger and bigger problem," says physicist Richard Steiner of the National Institute of Standards and Technology, where K20 is kept.

For years, Steiner and several competing scientific teams have been struggling to devise a more rock-steady replacement for K20 and its kin - about 80 look-alikes that serve other nations around the world. Now, after decades of frustration and failure, he thinks success may be in sight.

What also bugs scientists is that K20 is a thing- the sole survivor of an era when weights and measures were tied to objects. And that means it could be lost, damaged or stolen.

If any of those were to happen to K20 or the international master kilogram in suburban Paris (to which K20 and all other copies must ultimately measure up), "we're in deep trouble," says physicist Zeina Jabbour, who serves as K20's caretaker.

Stored in a safe

The U.S. kilogram is kept in a laboratory below NIST's sprawling campus, the country's Mecca of precision measurement. To see it, one must pass through two bolted doors requiring an agency I.D. badge and key code to open.

K20 rests in a large gray safe. Only Jabbour and one other person know the combination.

"Maybe I shouldn't be talking about security?" Jabbour says, suddenly concerned when a rare visitor inquires.

But a far greater fear, she acknowledges, is contamination. So strategically placed sticky mats pluck dirt from scientist's shoes, while rumbling air ducts constantly filter dust and other particles - keeping the lab cleaner than a typical hospital operating room.

Upon entering the kilogram's inner sanctum, Jabbour slips on a protective gown, hat and gloves to prevent errant hairs or skin flakes from mucking up the metal cylinder.

Signs of wear

Despite these precautions, K20 is already showing microscopic signs of wear: scuff marks that Jabbour speculates stem from weighings or cleanings, though both are performed with exquisite care.

These tiny scars could be behind the object's current weight troubles, because dirt can collect there. The kilogram may also be absorbing pollution molecules from the air, or oozing gas trapped during manufacturing more than a century ago.

"We really don't know for sure," Jabbour says.

Whatever the cause, scientists ultimately hope to do with the kilogram what they've already done with the meter, the second, and the other basic units of measure: link their values not to objects but to immutable natural phenomena.

The second, for example, was long defined as 1/86,400 of a solar day - until astronomers discovered the planet's spin is slowing. So in 1967, they redefined it as the time it takes a cesium-133 atom to vibrate precisely 9,192,631,770 times. Today, the best cesium-based clocks lose one second every 20 million years.

Likewise, the meter was the distance between two hash marks on a platinum-iridium bar. But this object proved unstable, expanding and contracting when the temperature changed. So in 1983 scientists voted to make the meter equal to the distance light travels in a vacuum in 1/299,792,458 of a second.

The kilogram, however, has stubbornly resisted all scientific efforts to send it to the scrap heap. "Mass is tricky," Jabbour agrees.

Few understand that better than Richard Steiner.

Seeking precision

The NIST physicist has just spent four years painstakingly building and fine-tuning a device that he thinks could ultimately be K20's undoing.

The two-story-tall apparatus, known as a watt balance, is essentially a giant scale-and the most persnickety experiment Steiner says he's ever worked on.

"This is not something you put together, push the button, and it works," says the physicist with a knowing smile. Most days he's trying to understand why it doesn't work.

First conceived of by British scientists in the 1970s, the idea behind the balance is deceptively simple: A 1-kilogram weight goes on one end, an electromagnet on the other. Then scientists add power until the push of the magnet exactly cancels the pull of gravity on the kilogram.

Whatever combination of voltage and current does the trick determines the new definition of the kilogram.

Complications

But achieving the precision needed to replace K20 and its kin has taken decades.

"Most of us would just like to get this done," says physicist Ian Robinson of the National Physical Laboratory in England, who is heading up a similar watt balance project.

Steiner has learned that the list of things that can screw up the experiment is like the universe - large, mysterious, and constantly expanding. "We're always finding something new," he says.

His main enemy is stray vibrations. To dampen their effect, the watt balance is housed in an unusual vinyl-sided structure that's really a building within a building, says Steiner.

Sitting on a concrete foundation at the center of the building, the device is surrounded by walls that are detached from the rest the structure's wooden frame. "So the wind can't shake them," he says. The walls are lined with copper to block stray electromagnetic waves, giving the room a spooky orange glow.

Steiner and his team have encased the most sensitive sections of the instrument in a vacuum chamber and moved air conditioners to a neighboring building, so its mechanical rumblings wouldn't rattle the balance.

Gravity is another complication. Steiner's team, for example, has to compensate for the subtle tug of the moon. They must even correct for the fact that the top of the instrument is one story farther from the Earth's center - and thus slightly lighter - than the bottom.

Another crack

In 1998, using an earlier version of the watt balance, Steiner's group measured the kilogram to within seven decimal places of precision. But most metrologists believe that replacing the kilogram would require at least eight.

In the next few months Steiner says his new instrument will be ready to take another crack at K20 - unless some new problem crops up.

It's enough to give any physicist a headache. But not Steiner.

"I like to play complicated games," he says. "And this is the ultimate complicated game."