Engineers at the Morton Thiokol Inc. plant in Utah concluded yesterday that a design flaw led to the failure of a previously untested space shuttle booster part during a full-scale test firing Dec. 23.

The engineers said the effect of the flaw was probably aggravated by a severe test maneuver. More analysis is needed before a replacement design can be adopted, they added.

The most optimistic scenario would allow the National Aeronautics and Space Administration to launch a shuttle in late summer, officials said.

The failed design was one favored by NASA engineers, although there were reservations at both Thiokol and NASA. It may be replaced by one favored by contractor Thiokol that was tested successfully in September, according to NASA and Thiokol officials.

The next flight of the shuttle, scheduled for June 2, has been delayed indefinitely until the failure can be more fully studied and a recovery plan can be established. A decision on what course to pursue could come at the end of this week.

NASA officials delayed the flight schedule last week after they discovered that a large portion of the booster's nozzle assembly, which guides the vehicle, had broken apart during last month's test.

The design failed by a much greater margin than even its Thiokol and NASA critics anticipated, according to most accounts of the incident. NASA's director of propulsion, Russell Bardos, said that even if the part "hadn't broken off, there was delamination {a coming apart of bonded layers} through it that would have told us we have a problem . . . . We don't know why {the design} failed yet. But by the looks of it, no matter what, we'd not want to use it."

Another source said some of the investigators and other experts who looked at the part were "appalled."

"In layman's terms," said J.R. Thompson, director of the Marshall Space Flight Center in Huntsville, Ala., "it just kind of tended to unravel" during the test.

The part that failed was one of the few on the giant booster that had never been tested before and, in that respect, officials said, the test did its job.

Engineers yesterday finished taking the huge booster motor apart and recovered all six missing pieces of the failed nozzle part and fitted them together like a jigsaw puzzle.

Known as the outer boot ring, the failed part anchors the booster's nozzle to a flexible rubbery "boot" that allows the nozzle to swivel. One purpose of the outer boot ring is to shield the metal and rubber bearing at the core of the swivel mechanism from the intense heat of the rocket gases.

The outer boot ring was redesigned as a precaution because of problems it had in an earlier flight, when pieces of it came apart but with no effect on the vehicle's performance.

The part is one of many that have been redesigned since the Jan. 28, 1986, Challenger disaster, which killed the crew of seven. The part is not related to the flawed field joint design that caused the tragedy. The redesigned field joints performed well in last month's test, officials said.

The failed design for the outer boot ring "has apparently been used on thousands of smaller nozzles in various kinds of military rocket motors," said Thiokol spokesman Rocky Raab. "In those it has always worked."

"Some reservations were expressed" about the design, he said, "but NASA decided to go ahead with it." At both NASA and Thiokol, he added, management was "less thrilled" about the design than some engineers were.

The design "is an accepted industry construction technique" that, as is typical in such redesigns, appeared to be the best of a number of choices and was approved by both NASA and the contractor, said NASA's John Thomas, of the Marshall center, who is head of the booster redesign team.

The "Thiokol design," which NASA hopes to be able to switch to, was tested successfully in September. It is a modified version of the one used before the Challenger accident. Officials said they hope to have enough information to decide by the end of this week whether it can be substituted in both the test and flight booster nozzles that are already built. If it can, a late summer launch is possible, they said.

Engineers will probably get approval today to begin cutting up booster nozzles already built for tests, to be ready to substitute the alternative outer boot ring.

The decision to move ahead with the Thiokol design is further complicated by the fact that it has never been subjected to the maneuver that was performed during last month's test of the NASA boot ring. That maneuver -- in which the nozzle was swiveled abruptly 7 degrees to one side, then 7 degrees to the other and then raised straight up -- may have put unnecessary stress on the boot ring, officials said.

The nozzle is swiveled much less than that in normal flight and might not require that extreme a maneuver even in an actual flight emergency, Thomas and other officials said.

The maneuver was not performed in the successful September test of the Thiokol design. NASA must decide whether the maneuver should be repeated in future tests of whatever design is adopted, risking another test failure and further delays, or whether a less-severe maneuver will be a good enough test under the more stringent, post-Challenger standards.

A team of officials at the Johnson Space Center in Houston, which developed the original test requirements for an extreme swiveling of the nozzle, is studying "what drove it that way," as one official put it, and whether a somewhat less extreme and stressful maneuver would be adequate.

Another question to be resolved is whether the new design will require a third full-scale test firing, in addition to the two already scheduled before the next shuttle flight.