AN experiment testing the ability of non-native Japanese oysters to resist diseases in the Chesapeake Bay was abruptly halted last fall after researchers found some of the bivalves - thought to have been sterilized - were capable of reproducing.

The controversial experiment had been allowed in the York River last June only after scientists offered assurances that the oysters would receive a treatment that would make any chance of natural reproduction almost nonexistent.

Critics feared that if the oysters reproduced, the foreign species could spread throughout the Chesapeake, resulting in unpredictable consequences both for the ecosystem and for the already troubled native oyster.

But the experiment was abruptly halted last fall after scientists discovered the treatment that made the oysters sterile "triploids" - a condition thought to be permanent - was wearing off.

"There was no reason to think, when the study began, that this would happen," said Stan Allen, of the Haskin Shellfish Research Laboratory at Rutgers University in New Jersey. "There is no precedence for it in the animal kingdom that I've been able to find."

The experiment was carried out by Rutgers in Delaware Bay and the Virginia Institute of Marine Science in the York River. The change in reproductive ability took place in both experiments.

The finding may have far-reaching consequences. Scientists and management agency officials say no similar experiments are likely to be permitted until it is understood why the reversion took place.

"I think that's going to have a significant impact on research with the beasts, obviously," said Jack Travelstead, chief of the fisheries management division of the Virginia Marine Resources Commission. "It was only with the triploid process - and the as surances that it was safe - that a lot of East Coast states allowed the experiment to go ahead."

Further, officials say, the finding calls into question whether other organisms that get similar treatment to prevent reproduction - including some used in aquaculture, as well as grass carp used in ponds and lakes to control weeds - may pose a greater r eproductive risk than previously thought.

The Bay Program is already planning a "Critical Issue Forum" for late this year or early 1995 to bring together scientists to discuss the matter.

"That whole issue needs to be thoroughly examined," said Dan Terlizzi of Maryland Sea Grant, who chairs the Bay Program's Exotics Species Workgroup. "Triploid technology gives us a tremendous tool if it can be shown to work safely and effectively because then exotic species might be brought in for cultivation without the worry of establishing reproducing populations. But without that certainty, there are a lot of potentially valuable organisms that many will resist introducing for aquaculture production and other projects."

Scientists had proposed the Japanese oyster experiment to determine whether that species - Crassostrea gigas - was more resistant to the diseases MSX and dermo than the native eastern oyster, Crassostrea virginica.

The two diseases have devastated the native oyster population, and harvests in Virginia and Maryland are at historic lows.

The Japanese oyster has shown resistance to dermo in laboratory experiments, but some scientists were concerned that those tests could not adequately reproduce real-world conditions. In addition, MSX resistance cannot be tested in the laboratory.

The proposal brought strong opposition from environmental groups and many East Coast states. Critics worried that that an accidental introduction of the foreign oyster in the Bay could have unknown ecological consequences, and could ultimately wipe out the native species. Some also worried that if the oyster proved to be disease resistant, pressure would build for an outright introduction of the species to bolster diminishing oyster harvests, regardless of the consequences.

The controversy lasted more than two years. Maryland, most other states, and the Chesapeake Bay Foundation finally dropped their staunch opposition after VIMS scientists, based on research done by Allen, concluded that the risk of the experiment's oyster s producing viable offspring were "vanishingly small."

At that point, the Virginia Marine Resources Commission issued a permit for the project to begin last June.

The Japanese oysters used in the experiment were chemically treated at Rutgers so they would have three sets of genes instead of the normal two. Such oysters would be unable to produce offspring that would survive. To further bolster protection, the Rutgers lab drew a blood sample from each oyster to make sure the treatment worked.

On June 29, trays containing 200 Japanese oysters and 400 native oysters were placed in the York River. The scientists periodically withdrew oysters from the trays for disease analysis. When that was done, the Japanese oysters were also tested to make su re they were triploid. When researchers pulled oysters out for inspection in October, they found one oyster with two sets of genes - a "diploid" - which could have been capable of reproduction under the right conditions. "At first, we thought it was just a mistake," said Eugene Burreson, an oyster disease specialist at VIMS. But a follow-up sample revealed more oysters with both triploid and diploid cells - or "mosaics." At that point, Burreson said, the last 85 oysters in the experiment were pulled out of the river and returned to Allen for inspection.

An examination revealed that 20 percent of those were mosaics. It appeared that those oysters were in the process of reverting to pure diploids.

"For it to have lost triploid cells in favor of diploid ones in the first place suggests that it is a progressive process," Allen said. "And we found animals of all ilks: those that had only a few diploid cells, those that had a lot of diploid cells."

Burreson said the water was too cold by the time the reversion took place for reproduction to occur.

Allen examined whether some error had been made in the original process rendering the oysters triploid, but could find none. He concluded that the reversion from triploids to diploids was "a true thing that happened."

Allen said he could not explain why the reversion took place, but that it was a source of on-going examination. "I don't have the answers right now," he said. "Going into the study, there had been no examples - and I still have found none in the literature - to indicate that this kind of event, this reversion from a triploid to a mosaic, is to be expected."

"I would never have been so assertive in our contention that these things were reproductively sterile had I not totally believed it," he said. "A triploid is, in fact reproductively sterile for all intents and purposes. But all bets are off when it starts to change."

Allen said he had previously hoped to conduct open water experiments with another oyster species this year, "but obviously, because of these results, we have to put that on hold."

Indeed, mangers in both Maryland and Virginia said they did not foresee any future open-water experiments with triploid oysters until there is an understanding of what happened with last year's experiment.

"We would want more assurances now than we did before that this risk of reverting or some other unforeseen thing might happen," said Pete Jensen, director of the Maryland Department of Natural Resources' Fisheries Division.

Travelstead said he could not flatly say the find would end in-site research with triploids in the Bay, but added, "I think we'll have to address that issue before this kind of research can be approved." VIMS Director Dennis Taylor, who is also science advisor to the VMRC, said the outcome "makes you want to be more cautious about any of those triploids." He said he planned to make a report to the commission on the issue later this summer.