KINGS BAY, GA. -- The nuclear-tipped Trident missiles move in a slow truck convoy to the Explosives Handling Wharf on St. Marys channel here and are carefully lowered inside the 24 tubes on a submarine preparing for underwater patrol.

Several times a year, emergency teams practice their response should a loading accident ever set off a missile explosion that disperses a fine dust of highly dangerous plutonium from the core of a thermonuclear warhead.

But some nuclear-weapons experts worry about an accident so grim that emergency-rescue teams would be irrelevant. In this scenario, a mishap sets off blasts of conventional explosives in several warheads, and their shock waves in turn compress the plutonium of one bomb, starting a chain reaction and thermonuclear detonation.

Experts at the three U.S. nuclear-weapons laboratories have been working secretly since last year to determine if such a calamity is possible. They have been feeding data representing missile mishandling and explosion scenarios into supercomputers looking for a clear answer. So far, the government says, the computers have indicated no firm basis for such an accident.

But the scientists also say more study is needed to resolve lingering uncertainties, perhaps even a special underground test of the W-88 Trident nuclear warhead. A small group of U.S. legislators has arranged for an independent inquiry into these and other nuclear weapons issues by three physicists with nuclear-weapons experience, to begin this summer.

And Energy Secretary James D. Watkins has ordered a special scientific study on the cost and feasibility of redesigning the warhead to enhance its safety.

In a recent interview, he said he believes the W-88 Trident warhead is too close to "the margin" of federal safety standards and should not have been designed without features to limit the risk of accidental plutonium combustion or dispersal.

But some scientists say the risks lie partly in the design of the Trident missile system itself, posing a far more complex safety dilemma than the government has stated openly. The three-stage missile contains the most explosive rocket fuel in the U.S. strategic arsenal, some of it in a cylinder at the center of the uppermost stage, and surrounded by eight thermonuclear warheads.

By bringing the warheads, each containing conventional explosives, into such close proximity with this explosive propellant, the Navy deviated from routine missile design, and, some experts say, substantially heightened risks that a serious accident -- no matter how unlikely -- could become a catastrophe.

The Trident missile and warhead design decisions were made after a heated, secret technical debate in the early 1980s in which the weapon's military features were given much higher priority than its safety, according to senior U.S. military and civilian officials.

The Navy sought the combination of long-range missile and high-yield warhead to fulfill its longstanding goal of matching the military capability of Air Force land-based strategic missiles with its own submarine force. But that ruled out using heavy, safer materials or a less "energetic" propellant. The W-88 is said by independent experts to have a potential yield of 475 kilotons, equivalent to exploding 475,000 tons of TNT. The World War II nuclear blast at Hiroshima was 15 kilotons.

The result of this configuration is that rebuilding the warhead with less-explosive components may not change "anything," said a senior Defense Department official last week. "They may have to come up with something even more dramatic" to increase the Trident missiles' safety significantly.

So far, only one Trident submarine, the USS Tennessee, has been loaded here with strategic D-5 missiles carrying W-88 warheads. At least eight more Trident submarines are to be loaded here before beginning Atlantic Ocean patrols and eight other submarines now carrying different missiles are eventually to be equipped with the D-5 at Bangor, Wash.

The greatest risk in handling the missiles is said to be when the W-88 warheads are fitted around the column of propellant in the third stage. This is done in two buildings tightly guarded by Marines, where any open flames are prohibited.

The missiles then are transported to the wharf or stored in special earth-covered bunkers of reinforced concrete, so an accidental explosion of one missile will not detonate others nearby.

Bob Dietz, a former research engineer for Lockheed Missiles and Space Co. who played a key role in the design of the D-5 missile, said the accompanying warhead initially was to use so-called "Insensitive High Explosives" (IHE) to start the nuclear explosion.

The material, widely used in nuclear weapons in the early 1980s, is considered relatively impervious to accidental detonation from gunshots, fire or nearby explosions. But these plans were dropped, according to John H. Birely, a senior official at Los Alamos National Laboratory, where the warhead was designed.

"Had we used . . . {IHE}, the weapon would have weighed more and, therefore, for a given missile, we would either have had less range . . . or less warheads per missile," Birely told a Senate panel in 1987. "So, the tradeoff was done. It was felt that, for overall military effectiveness, the use of the conventional, more sensitive high explosive was the better way to go."

Robert Barker, assistant to the secretary of defense for atomic energy matters, confirmed to the committee that "it really was a defense requirement which drove the decision." But he added, "The other factor that was relevant is the fact that the propellant on the Trident is a very high-energy propellant" in extremely close proximity to the warheads.

"In fact, some studies indicated that, even had we put insensitive high explosives on there, in some kinds of accident situations, the explosive force of that third-stage motor would be sufficient to scatter plutonium around anyway," Barker said.

Dietz, noting that the D-5 missile uses a "detonatable" propellant, said he nevertheless doubts the possibility of a serious Trident accident. He added that no such missile has ever exploded as a result of mishandling. However, the D-5 in early underwater launch tests veered off course several times and was deliberately destroyed, and several land-based missiles have exploded accidentally.

Unlike earlier submarine missiles, which ignite when released from their launch tubes underwater, the engines of the D-5 and a similar missile, the C-4, ignite only after they breach the surface, Dietz said. The change was made because of the danger -- however remote -- that the C-4 or D-5 engines might blow up on launch and destroy the sub.

"The Navy wanted to avoid any direct coupling" of force from a potential missile explosion to the submarine, Dietz said.

The D-5's third-stage design helps extend its range 1,500 miles farther than older submarine missiles, but prevented use of insulating material between the third stage's propellant and the surrounding eight nuclear warheads.

In 1987, Roger Batzel, then director of the Lawrence Livermore National Laboratory, told a congressional hearing that an accidental dispersal of plutonium from any nuclear warhead accident "could be worse" than the 1986 Soviet nuclear reactor accident at Chernobyl.

Another weapon scientist said secret studies indicated that "if the wind was blowing the right way," a dispersal accident at Bangor, Wash., could contaminate Seattle, force its evacuation and cost millions to clean up.

Concerns about this design were revived when Sylvester R. Foley Jr., a director of defense programs at the Energy Department, sought a review of nuclear warhead designs in the aftermath of the 1986 Challenger space shuttle disaster, caused by a defective solid-fuel rocket booster.

A new study of Trident warheads failed to prove that retrofitting IHE and other features would significantly improve their safety, according to several officials. That study urged instead that future missiles be redesigned and use less-energetic propellants.

The concerns of some weapons scientists were heightened when preliminary computer calculations indicated in late 1988 that such an accident could produce a nuclear yield, not just the dispersal of plutonium. "Huge acreage was purchased for those {Trident loading} facilities in anticipation of a potential plutonium dispersal," one source said, but many deaths would obviously ensue from a full nuclear detonation.

"There was some ambiguity -- not enough to halt {the W-88's} deployment, but enough to warrant further study," another source said. One idea under discussion is withdrawal of one or more of the warheads atop each missile to permit installation of some shielding material.

Capt. Gerald Nelson, commander of the Strategic Weapons Facility here, said that he believes that a handling accident is highly improbable. "Safety is the most important thing we do here, followed closely by security," he said.

But experts say there have been several Navy missile mishandling incidents, including a 1981 accident in which a Poseidon missile armed with 10 nuclear warheads fell 14 feet as it was being transferred from a submarine to a tender at Holy Loch, Scotland. The missile did not explode, but some workers were evacuated and measures were taken to protect against possible radioactive contamination.