Lawrence B. Mulloy, head of NASA's solid rocket booster program, said today he is now convinced that "all the evidence" from the Challenger explosion points to a faulty booster joint as the cause and that engineers also believe that a failure of the putty used in the joints may have contributed to the rupture.

It was the first time a ranking National Aeronautics and Space Administration official has explicitly identified a failure in the joint as the probable culprit, though others, including members of the presidential investigating commission, reached that tentative conclusion weeks ago.

"I think all the evidence is now pointing to a leak in the right-hand booster's aft center joint," said Mulloy, who had just examined the most recent tests of the joints by the booster manufacturer, Morton Thiokol Inc., in Brigham City, Utah.

The boosters are made of four segments stacked vertically and sealed at the joints to prevent exhaust gases from escaping from the sides of the booster during launch.

Inside each joint are two rubbery O-rings, running the circumference of the booster and designed to make the seals airtight. To seal properly, pressure from the rocket must force the O-rings tightly into metal grooves. Putty is used to protect the O-rings from the hot gases inside the boosters.

Mulloy, in a telephone interview from Marshall Space Flight Center in Huntsville, Ala., said it is now believed that the putty may have momentarily shielded the O-rings from pressure building up inside the booster. If the O-rings are so shielded even briefly, Mulloy said, the joints may not seal properly.

He said the possibility is being investigated that cold temperatures before launch may have stiffened the putty. Problems with putty have long been suspected as a cause for O-ring erosion problems, in part because the putty is believed to be excessively "susceptible to environmental effects," including moisture and cold, according to NASA documents and commission testimony. The putty is stored at 32 degrees Fahrenheit before use to keep it dry.

Air temperatures were in the 20s the night before the Jan. 28 launch, forming icicles on the launch tower. Officials from Morton Thiokol, as well as Rockwell International, another major NASA contractor, had cautioned against a launch because of these conditions.

Mulloy said that in six days at Thiokol he saw test data showing that the joints would seal properly in launch conditions of 10 degrees.

A second set of tests, however, showed that cold reduces the ability of the O-rings to seal a gap created briefly by metal expansion after ignition.

The second tests did not account for pressure, however, and Mulloy said that if the putty works properly, pressure from ignition of the rockets forces the O-rings to seal, regardless of the cold.

"We only have 20 to 30 percent of the test data in on the joints , so it is entirely premature to draw a conclusion, but if I was forced to draw a conclusion from this test data, I would say it shows that the joint will seal properly at 10 degrees," he said.

The temperature when Challenger was launched was 38 degrees, and had been as low as 24 degrees the night before. Officials have said temperatures inside the seals may have been colder because of the proximity of supercold gases used to power the shuttle.

Also today, NASA for the first time confirmed reports that Navy divers have recovered four of Challenger's five main on-board computers, along with three data and voice recorders.

The computers may provide information about the last milliseconds of the flight, and the recorders may contain crew conversations that would reveal whether the seven crew members realized the craft was breaking up before it exploded. NASA officials have said no information about the cause of the accident is likely to be contained in the devices.

The NASA announcement indicated it would take "several months" to retrieve data from the computers and more than two weeks to clean and process tapes from the voice and data recorders.