In the debate ignited by the Soviet accident at Chernobyl, there is at least one point of agreement between advocates and opponents of nuclear technology: In theory, at least, a disaster of similar magnitude could happen in the United States.

"What we've always said is 'yes,' " said Brian Sheron, head of the Nuclear Regulatory Commission's reactor systems division. "You can always postulate the conditions."

"All the plants we now have in this country can have an accident comparable to what happened in the Soviet Union," said Robert Pollard, a physicist with the Union of Concerned Scientists. "The only rational basis for discussion is the probability."

There the agreement ends and the argument begins.

Industry representatives and federal regulators contend that Chernobyl's misfortune points up the strength of American reactors, with their multiple safety devices and thick concrete shells.

With equal vehemence, nuclear foes argue that the accident betrays the inherent weakness of all reactors.

"Our problem is not with nuclear physics," Pollard said. "Our problem is plumbing."

The statement goes to the heart of a controversy that hinges less on the exotic technology of nuclear power production than on the more mundane science of industrial engineering -- the pipes, pumps, wires, gauges and walls that defend against an accident that could damage the reactor's core and unleash a lethal plume of radiation.

To guard against the fallibility of machinery, the NRC relies on a policy known as "defense in depth." In general, that means a reactor must have independently operated backup systems that would automatically take over the function of damaged primary systems, such as the cooling apparatus that keeps water flowing through the reactor core.

Yet, like the war that was lost for want of a nail, the 1979 accident at Three Mile Island in Middletown, Pa., began with a stuck valve and escalated into a partial meltdown. Less than a year ago, the Davis-Besse plant near Toledo, Ohio, narrowly escaped a meltdown when 14 separate components malfunctioned or were mistakenly deactivated, including the main cooling system and its emergency backup.

The NRC's critics say the accidents illustrate their point as well, if not as dramatically, as the cloud of radiation from Chernobyl.

"You can lose sight of the principal lesson to be learned here if you immerse yourself in the differences in the technical minutiae," said Thomas B. Cochran of the Natural Resources Defense Council.

"You can hang all sorts of whistles and bells on them and make all sorts of claims on their safety. But any of these reactors, of diverse designs, all have one common problem in that they can all experience a full core melt."

Neither the TMI nor the Davis-Besse accidents, however, resulted in life-threatening releases of radiation. At Davis-Besse, operators were able to jury-rig a cooling system for the core. At TMI, most radiation from the damaged core and its contaminated cooling water was held inside the reactor's massive containment dome, the last line of defense against a potentially catastrophic radiation release.

In the days following Chernobyl, the 7-year-old nightmare of TMI became a source of comfort for U.S. regulators and industry representatives -- pointed to as proof that the safety design of U.S. commercial reactors made a similar disaster all but impossible in this country.

But as more details emerged about the design of the Soviet reactor and the magnitude of the accident there, the assurances began to waver.

"I'm not convinced personally that even if Chernobyl had a containment it would have been able to contain that event," said the NRC's Sheron.

NRC officials emphasize that, once the accident was beyond control, the Soviet reactor's graphite design created problems that might not be confronted in a typical U.S. commercial reactor.

"That plume went up something like a mile," Sheron said. "That's a different release mechanism from U.S. reactors."

But, like U.S. reactors, the Chernobyl plant was equipped with some features designed to prevent an accident from progressing to that point. There were, for example, duplicate power sources and protective steel and concrete barriers around some critical reactor areas.

Of particular concern to nuclear critics is the discovery that the Soviet reactor, while not equipped with a prestressed concrete dome like those that shield more than half of U.S. reactors, did have a pressure-suppression system similar to the kind that protects 49 civilian reactors in this country.

A containment dome, essentially a huge vault surrounding the reactor, is designed to withstand pressure because of sheer strength and size. Reactors equipped with pressure-suppression systems use a different approach: The system is supposed to prevent pressure from building to the point where a super-strong dome would be needed.

The Chernobyl plant's system consisted of two water pools sandwiched between thick concrete directly under the reactor. Should the reactor's core begin to overheat and pressurized steam begin to build up, the steam would be forced down into the pools, where it would be cooled, condensed back into water and thus depressurized.

Similar systems are in use in 39 General Electric reactors that are operating or under construction in the United States, and 10 Westinghouse reactors use ice rather than water to fill the same function. The reactors in those plants are surrounded by steel and concrete, but the stations are not covered by a massive dome like the one that is credited with preventing greater radiation releases from Three Mile Island.

The design was questioned more than a decade ago by Stephen Hanauer, a top Atomic Energy Commission safety adviser, who warned in a 1972 memo that pressure-suppression systems were far more vulnerable to valve failures than the larger -- but more expensive -- concrete domes.

"Valves do not have a very good reliability record," Hanauer wrote.

Hanauer's memo went to Joseph M. Hendrie, a colleague who later became NRC chairman. Hendrie responded that Hanauer's idea to ban pressure-suppression containments "is an attractive one in some ways," but he rejected it.

"Reversal of this hallowed policy . . . could well be the end of nuclear power," he wrote. "It would throw into question the continued operation of licensed plants, would make unlicensable the GE and Westinghouse ice condenser plants now in review, and would generally create more turmoil than I can stand thinking about."

NRC official Jack Kudrick, who heads the division that oversees reactors of the GE and Westinghouse design, said in a recent interview that the agency evaluated Hanauer's concerns at the time and decided that the suppression systems were adequate.

"We responded to those concerns," he said. "The design would allow at least one valve failure. Probably you could get by with just one functioning, and there are usually four to six."

Nevertheless, he said, additional safety devices were added to the pressure-suppression systems after the accident at TMI to avert the possibility of a hydrogen burn. At TMI, hydrogen created by reactions between water and zirconium-clad fuel rods exploded inside the containment dome.

In the General Electric Mark I and Mark II designs, the NRC required the reactor to be surrounded by inert gases -- helium and nitrogen -- that will not support combustion. In the later Mark III reactor and the Westinghouse ice-condenser design, a hydrogen igniter was incorporated so that the hydrogen could be burned in a controlled fashion.

The agency's critics question the value of those improvements, noting that the NRC allows operators of GE reactors to remove the inert gases when the reactor is being shut down so that work crews can enter the reactor area. The Chernobyl reactor, which also was surrounded with an inert gas chamber, was partly shut down when a massive hydrogen explosion apparently triggered the fire there. It is not known if the chamber contained inert gas at the time.

U.S. experts think that the suppression pool at Chernobyl helped prevent molten fuel from burning through the plant's concrete base, but the system clearly did not prevent the explosive escape of radiation. Without more data from the Soviets, NRC officials said it is impossible to know whether that means the system failed to function or whether it was simply overwhelmed by an accident beyond its capacity to control.

According to Sheron, the pools apparently were designed to accommodate a major plumbing accident, not an overheating of the reactor core that would result in a meltdown of its fuel rods.

"The Chernobyl plant did not appear to have any design features that could accommodate a core-melt accident," he said. "There are a lot of unanswered questions about their protection systems."

In a briefing for NRC commissioners two weeks after the Chernobyl accident, however, chief staff official Victor Stello said agency officials were satisfied that the U.S. safety system was sound. "We see nothing coming out of this accident to suggest we need to change," he said.