Recently discovered cracks in a foam "ramp" on the space shuttle's external fuel tank may present an "unacceptable safety threat" to the orbiter, raising the possibility that NASA could delay its next launch while engineers decide whether to get rid of the ramp altogether, shuttle program manager N. Wayne Hale said in an internal memo.

Discovery of the PAL ramp cracks during an inspection early last month marked a new setback for the troubled shuttle program, which has flown only once since Columbia disintegrated over Texas in February 2003. Hale had predicted earlier that making changes to eliminate the ramp would delay the next shuttle launch until next fall, at least four months later than currently planned.

Further delays could also add cost and uncertainty to President Bush's "Vision for Space Exploration," which calls for the shuttle to finish building the international space station by 2010 and then retire, stepping aside for a new-generation spaceship designed to take humans back to the moon and eventually to Mars.

In the e-mail Sunday to shuttle engineers and managers, Hale said vertical cracks in the external tank's "protuberance air load," or PAL, ramp extended deep into the foam insulation and appear to have been caused by contraction and expansion as the tank was being filled with supercooled liquid hydrogen and oxygen.

Because all shuttle fuel tanks undergo such "cryoloading," cracking "must be presumed possible in any PAL ramp on any flight vehicle," Hale wrote. As a result "this . . . represents a critical and unacceptable safety threat to the flight of the space shuttle."

Hale stressed his judgments were "preliminary" but it "appears mandatory" that the shuttle team focus on eliminating the ramp altogether for all upcoming shuttle flights, including the May launch.

Despite the dire nature of the e-mail, however, Hale said in a telephone interview yesterday that he intended to challenge his team to "explain to me how we could do anything with the cracks," including devising new foam spraying techniques.

"We don't give direction by e-mail," Hale said. "We will have a formal board meeting to review this." The e-mail, he added, "is not a decision, but a notice to my folks that tells them 'this is the way I'm thinking; this is the way I'm leaning. Talk me into it, or talk me out of it.' "

NASA grounded the shuttle last July after Discovery's external tank lost a one-pound piece of PAL ramp foam during launch. The fragment flew harmlessly away, but the mishap was embarrassing, because NASA had redesigned the external tank after a piece of foam punched a fatal hole in Columbia's heat shielding.

The foam, sprayed both mechanically and by hand, insulates the 154-foot tank to protect it when it is filled with liquid oxygen and hydrogen cooled to temperatures several hundred degrees below zero.

The PAL ramps are foam ridges that run alongside electric cables and pressurized gas lines on the exterior of the tank. They serve as a windbreak for these fixtures during the turbulence of launch.

As recently as mid-October, engineers attributed the Discovery foam loss to other factors, including "crushing" by technicians crawling on top of the tank during manufacture and a possible air-filled "void" inside the foam that had expanded and burst as Discovery climbed.

The planned fix was to improve manual foam applications and develop new automated techniques. Hale at that time described elimination of the PAL ramp as an option.

But when they examined a tank that had been twice loaded with fuel and then emptied, engineers found nine vertical cracks in the lower, hydrogen, PAL ramp. Another tank that had not been cryoloaded had no such cracks.

NASA sources, who requested anonymity because they were not authorized to speak publicly about the investigation, said engineers who knew about the cracks immediately suspected they could indicate a serious problem.

But when Hale announced the findings at a Nov. 22 news conference, he said the next "one or two flights" would use the new, improved PAL ramp, while NASA tested the feasibility of flying without it. He said the testing would not finish before next fall.

In his e-mail five days later, however, Hale's tone was markedly more pessimistic. He said yesterday that he had seen a further report from his team showing that at least one of the cracks extended all the way down through the ramp and into the "acreage foam," the initial machine-applied insulation layer to which the ramp is attached.

"Further, cracks appeared in both old and new sprayed areas of the . . . PAL ramp, and therefore the new improved spray techniques do not provide protection from this phenomenon," Hale wrote in the e-mail. "It is hard to see how automated PAL ramp spray would provide any protection from these cracks."

Because of the potential pervasiveness of the cracking, "all efforts should be made to provide a no-PAL ramp flight condition," Hale said, including wind tunnel testing and computer models capable of analyzing the effects of eliminating the ramps.

Hale said yesterday, however, that he had directed his team to move ahead in investigating new spray techniques along with ramp removal, and that the May launch date "remains on the table" while planners study the options in preparation for a formal progress review in the next few weeks.

An analysis of this photo, taken by Discovery's crew in July, showed that a large piece of foam separated from the fuel tank as the shuttle rose to orbit.

N. Wayne Hale is the shuttle program manager.