During the Cold War, scientists at two labs, one in California and one in New Mexico, eyed each other warily. They ridiculed and often back-stabbed each other at the Pentagon in a race to provide the U.S. Navy, Air Force and Army with nuclear weaponry.

The two labs, both run by the University of California, evolved into different cultures: Los Alamos in New Mexico, more conservative and academic; Livermore, flashier, more prone to taking risks. They clashed often -- over nuclear testing, over the feasibility of X-ray lasers for the Reagan administration, over laser fusion.

They were, in the words of a senior weaponeer, "scorpions in a bottle."

Yet the lesser-known story is that these two warring houses of defense scientists propped each other up, setting their enmity aside to collaborate and trade secrets concerning some of humankind's most lethal explosives.

Now, after 61 years, the U.S. Department of Energy is considering a split between these siblings, inviting bids from other contractors to run one or both of the labs. More than a dozen defense contractors, engineering firms and universities are forging alliances to challenge UC's do-

minion over two heavyweights in the world of defense science.

In fending off these aspiring nuclear weaponeers, the university and its nuclear scientists argued last week that awarding one or both labs to private, corporate contractors could subordinate U.S. nuclear policy to profits. Competing contractors also could plunge the labs into a fight-to-the-finish race for weapons and national security work.

Lab scientists told a panel of the National Academy of Sciences that only a single contractor can make the two labs work together even as they compete for weapons designs and an array of other national-security missions, ranging from intelligence analysis to energy and anti-terrorism research.

And they insist that only the University of California or a contractor untainted by desire for profits and market share can be trusted to advise the United States on the reliability of its nuclear weapons and whether to restart explosive nuclear testing after a 12-year hiatus.

A U.S. return to nuclear testing could reverberate around the world and spark rounds of nuclear tests in Russia, China and perhaps South Asia. It's a decision of great importance for U.S. and worldwide nuclear proliferation. Livermore director Michael Anastasio suggests that no profit-driven corporation should have a hand in it.

"The potential is there for the contractor to be influenced by the (financial) interests of the contractor in extension of the contract or other business," he cautioned.

The National Academy panel faces huge questions in advising the Energy Department on how to run a competition for the labs, and high on the list is a deceptively simple one: One manager or two?

John Sommerer, chief technology officer at Johns Hopkins University's Applied Physics Laboratory, says preservation of the labs' "constructive, competitive relationship" in a period without nuclear testing is a central issue, especially since it appears the university may lose management of at least one lab.

"There's a lot of comfort in leaving it the way it is. But frankly I think the probability of that is low from a game standpoint and from the track record on lab contract competitions," Sommerer said. "The question is, what can we do to make sure the competition (between the labs) will still be healthy?"

Hugh Gusterson, an MIT anthropologist who has studied scientists at both labs, said it's unclear why the University of California and its scientists think their interests are purer than a defense contractor's.

"It's a fallacy that universities are somehow immune from material pressure. There is no ivory tower," Gusterson said. "What's true here is that over a period of 50 years, an old boys' network has grown up to manage the labs. So when we're talking about deregulation, if you're in that network, it can be a little worrying."

In pressing their case for keeping UC in charge, lab executives are offering insight into the 50-year clash of wills between two major Cold War powers -- not the Soviet Union and the United States, but between Livermore and Los Alamos.

"It's been tooth and nail at times," said Bruce Goodwin, a former Los Alamos physicist who heads Livermore's nuclear-weapons program.

Yet Los Alamos and Livermore also shared H-bomb secrets, nuclear components and top managers so neither would fail and leave the other alone to stagnate intellectually in the closed world of U.S. nuclear-explosives design.

The grudging aid from one sibling lab to the other began immediately.

Then-Los Alamos director Norris Bradbury turned from his losing fight against Livermore's creation and offered help to its upstart weaponeers on bomb components and nuclear tests.

The new lab famously failed in its first two tries at hydrogen bombs, both Edward Teller ideas. Its third bomb finally delivered nuclear yield, all of it from a plutonium trigger supplied by Bradbury's Los Alamos.

Cooperation between the rivals, in public and classified channels, persisted through and beyond the Cold War. In the late 1960s -- a decade after Livermore revolutionized hydrogen-bomb design with egg-shaped nuclear components and other advances that shrank thermonuclear warheads down to a fraction of their original size -- Livermore sent full blueprints, design software and nuclear test data for its then-most modern warhead to Los Alamos, along with a top designer to instruct the older lab in its finer points. Los Alamos later tapped the ideas in the W68 for the strategic warheads now in the U.S. arsenal.

"Much of the major, primary contributions over the last 40 years are based on pioneering work that Livermore did. Los Alamos was slow to adopt that," said veteran Los Alamos weapons and testing manager John Hopkins. Since the lab did, "I think the whole stockpile is better off for it."

After the Sept. 11 attacks, despite an early Bush administration proposal to turn Livermore into the nation's chief homeland-security lab, executives of Los Alamos and Sandia joined Livermore in a pact to share the work.

In the nuclear-explosives arena, it became obvious that unlike the conventional defense and aerospace industries, the two labs would remain the only places in the United States to teach weapons physics. Both were needed to critique each other's work.

"It would have been much harder to develop this collaborative environment with two separate contractors," said Los Alamos' Hopkins.

In the last 25 years, U.S. civilian government lost its nuclear weapons expertise, and the weapons labs assumed a stronger hand in devising U.S. nuclear weapons research.

If labs were split and run by different contractors, the federal government would have to become a bigger player, said George Miller, a former weapons designer and head of the National Ignition Facility directorate at Livermore.

Otherwise, the two labs might devour each other.

"You could imagine the type of competition there would be, like between Oracle and PeopleSoft, where they try to put each other out of business. We think it's important for national security that there be two voices (on nuclear weapons)," Miller said. "You would hate for the nuclear-weapons program to have a Microsoft."