The most bitter and longest running "Star Wars" debate involves a seemingly straightforward matter: How many orbiting battle stations would be needed to provide enough firepower over Soviet missile fields at all times.

Star Wars critics -- led by Richard L. Garwin, a physicist who worked on the hydrogen bomb and is now a defense consultant, and the Union of Concerned Scientists -- said early last year that the number would be prohibitive. A system based on the most promising weapon, an orbiting chemical laser, would require 2,400 battle stations, and each would need nearly a full space shuttle load of fuel to fire the laser, he calculated then.

Star Wars proponents in military-oriented national laboratories such as Lawrence Livermore and Los Alamos quickly objected that the true number was around 90 and that Garwin and the UCS were either bad mathematicians or trying to kill Star Wars.

Then UCS discovered errors in its estimates. It had calculated the number of satellites needed to keep the entire world in range, not just the Soviet Union, and it had not allowed for the fact that closer targets can be destroyed faster than distant ones. Its estimates dropped first to 800 and then to 300, but only temporarily.

Robert Jastrow, a Star Wars advocate and National Aeronautics and Space Administration astrophysicist, pounced on the errors, citing them as evidence that Garwin and the UCS were misleading the public.

"The work of the UCS on the question of the satellite fleet," Jastrow wrote in Commentary magazine in December, "is the poorest that has appeared in print, to my knowledge." Jastrow also called the UCS work "shoddy" and contrasted its estimates with "some fine work by the theoretical physicists at Los Alamos."

Garwin responded with a 23-page letter to Commentary -- not yet published -- denouncing Jastrow's article as a "screed . . . rehashing the same demolished criticisms." Garwin added that Jastrow, who is not an expert on strategic defense technologies and who used other people's data, has "made a career of hyena-like behavior."

Not all of the clashes between Jastrow and Garwin are so emotionally worded. Some are more factually verifiable.

Jastrow, for example, asserted that the UCS report always chooses the most pessimistic end of a range of performance possibilities. "All the errors and omissions," Jastrow wrote, "go in one direction only -- toward making the president's plan seem impractical, costly, and ineffective."

In fact, Garwin responded, the UCS evaluation presumes perfection: that Star Wars will work flawlessly at every step of the way even though the system can never be tested under battle conditions. It assumes that laser-aiming, 30-foot-diameter mirrors will have perfect reflecting surfaces, that the mirrors can swivel instantly from target to target, that the beams will never miss their fast-moving quarries thousands of miles away, and that the battle management computers work flawlessly in assigning targets to lasers.

Last month Garwin calculated that if some of these improbably perfect allowances are adjusted, the number of needed battle stations goes back up again.

For example, Garwin figured that if it takes the giant mirror three seconds to swing toward a new target and become steady enough to aim -- the same time it takes NASA's much smaller space telescope -- Star Wars could require up to 2,263 satellite battle stations. Even if it achieves its goal of cutting the settling time to 0.1 second, the system would still need 1,110 satellites, Garwin calculated.

These new figures also assume that the Soviets would try to counteract the U.S. missile defense by increasing their missiles from the current 1,400 to 3,000 and that they would fly faster, reaching space and releasing their warheads in less than half the current time.

Although the Strategic Defense Initiative (SDI), as Stars Wars is formally known, says its weapons must be effective against the Soviet force of the 21st century, its advocates complain that Garwin is introducing these factors to recoup the position lost when the UCS errors were found.

"Garwin isn't playing by the rules," said Gregory Canaven, a pro-Star Wars physicist from Los Alamos. Canaven said that without Garwin's newly introduced factors, they would be close to agreement on the number of satellites needed to counter the current Soviet missile force.