A full investigation of the Challenger explosion will probably take many more weeks, but enough details have emerged to piece together a relatively complete picture of what happened to the spacecraft that tragic morning one month ago.

As is typical of such major disasters, it appears that the destruction of Challenger followed a long chain of events that began hours before the 11:38 a.m. launch on Jan. 28.

To sum up what is known, it appears that cold temperatures hardened the normally supple rubbery O-rings in the solid rocket boosters, preventing them from sealing gaps between rocket segments. One set of O-rings began to burn away in the first second after ignition, but then sealed adequately until, about a minute into the flight, it burned through. White-hot gases then erupted from the booster, causing it to break at its lower attachment point and swing away, crashing into the external tank and releasing the tank's highly explosive liquid hydrogen and oxygen.

What follows is a "best guess" scenario of the events during the hours leading to the launch and the final devastating fireball, in which Challenger's seven crew members were killed. It is based on information released by the National Aeronautics and Space Administration, facts elicited by the presidential investigating commission from officials at NASA and Morton Thiokol Inc., which made the boosters, and information from others close to the investigation.

The sequence began about 3 a.m. the day of the launch, when technicians began pumping liquid hydrogen and liquid oxygen into separate chambers in the vehicle's external tank. Florida was experiencing an unusual cold snap, and air temperatures on the launch pad were in the low 20s through the night. As a result, the solid fuel boosters were unusually cold. Less than an inch of steel separated the cold air from the O-rings just inside the booster walls.

Although the morning sun would eventually warm the air to 38 degrees, the right-hand booster would be further chilled as a result of the very cold liquids being pumped into the adjacent tank. The 143,000 gallons of liquid oxygen being pumped into the upper part of the tank were 297 degrees below zero. The 383,000 gallons of liquid hydrogen in the lower tank were 423 degrees below zero. These are the normal liquid temperatures of these substances.

As the 27.5-foot-diameter tank was filled, the outer surface of its insulated aluminum wall cooled, dropping to 2 degrees. This normal phenomenon appears to have combined with the abnormally cold weather to set in motion a deadly sequence of events.

Through most of the morning that Challenger sat on the pad before launch, winds of about 10 mph were blowing from the west -- the spacecraft's left side -- becoming colder as they passed over the frigid external tank and then swirling around the right-hand booster. The effect was that of a fan blowing over a block of ice.

Although the right-hand booster was to the east and one side of it probably was warmed by the rising sun, the lower part of the booster would soon have been shaded by the orbiter's right wing. The side of the booster that would eventually rupture, closest to the external tank, was continually shaded.

By the time an ice inspection team visited the pad, three hours before liftoff, the outer surface of the part of the booster where the rupture would occur had dropped to 19 degrees. Calculations indicate that the rubber O-rings were about 29 degrees.

Like other materials, the O-rings shrink and harden when cooled. Both effects appear to have doomed mission 51L. When Challenger's two boosters ignited, something went wrong within the first second. To understand why, one must appreciate the design of booster joints.

The rocket motor of each booster consists of four cylindrical segments, each 12 feet in diameter, which must be stacked and fastened. Each joint must do two things: hold one segment to the next and prevent the hot, high-pressure gases of burning fuel from leaking out.

At the top of each segment, the edge of the steel casing widens to accommodate a 3 1/2-inch-deep vertical groove running around the 37-foot circumference of the segment. Into this groove will slide the bottom edge of another segment lowered from above. In cross-section this looks like a tongue-and-groove joint.

The outer wall of the groove is ringed with 177 holes, each one inch in diameter. The tongue, the lower lip of the segment above, has similar holes. Once the segments are joined and aligned, steel pins are pushed in from the outside, through holes in both segments and into a socket bored partway through the groove's inner wall. A steel band wrapped around the joint keeps the pins from falling out. This is all that holds the segments together.

However, because the tongue fits loosely in the groove, it is not pressure-tight. Sealing the loose fit is the job of the two synthetic rubber O-rings. Each is a closed loop of black rubber, a cylinder just over a quarter of an inch thick and long enough to encircle the booster.

Each ring fits inside a groove, each held in its own channel so it can close the space -- about an eighth of an inch -- between the inner wall of the groove and the tongue from the segment above. To ease installation of the rings, technicians first slather both them and the metal surfaces with heavy grease. Like the O-rings, the grease would have been chilled to 29 degrees that morning, becoming unusually thick and viscous.

In a normal launch, as soon as the booster fuel ignites, the pressure of hot gases is supposed to force the first O-ring to move a fraction of an inch until it fits tightly into the space, blocking any opportunity for leakage. A second O-ring is in place as a backup if the first fails.

The failure in Challenger's right-hand booster occurred less than one second after its fuel was ignited. The cold O-rings at one joint in the right-hand booster -- too stiff, too small and too sluggish to move because they were stuck in cold-hardened grease -- apparently failed to move immediately into the gap.

Less than 0.4 seconds after ignition, the resultant hot gases inside the booster were pressing in all directions with a normal pressure of about 800 pounds per square inch. If the O-rings had sealed, the gases could escape only through the nozzle at the bottom of the rocket, which produces the desired upward thrust.

Because Challenger's O-rings did not immediately seal the gap, some of the pressure leaked out, blowing a hole through the thick putty that is supposed to protect the O-rings from the heat of burning rocket fuel.

Although the inside of each booster segment is lined with a layer of solid fuel varying from three to five feet thick, there is a space between fuel segments at the joints. Hot, high-pressure gases from the burning fuel normally press against the putty but the pressure usually stops once the putty transfers the pressure to the O-rings and they move into sealing position, blocking further movement.

As hot gases rushed past Challenger's frozen O-rings, they apparently burned the putty, the rubber and the grease. Before the vehicle was even 20 feet off the pad, billows of thick black smoke 10 feet wide and 25 feet high had slipped through the joint gap, at a point near the external tank, and were spewing upward, aimed by the design of the joint.

For the next 12 seconds, as Challenger gained speed and completed its roll maneuver, tracking cameras recorded the smoke continuing to pour out. Then the smoke appears to have stopped. One possibility is that heat from the burning fuel eventually warmed and softened the O-rings, or what was left of them, enough to let them seal.

For the next 47 seconds everything seemed fine.

Then, 58.7 seconds into the flight, as Challenger was flying through what is normally the most turbulent part of its launch -- a period that subjects the vehicle to severe, rapid shaking -- smoke again broke through the same joint. The eroded O-ring was burning again.

Half a second later the burnthrough was complete and a small but intense plume of hot gases, about 5,900 degrees, erupted from the joint and began spreading, cutting its way along the O-ring, like a welder's torch around the booster, toward the external tank.

At just over 72 seconds Challenger began to break up. Data transmitted from the booster indicate that it had somehow torn loose the struts attaching its base to the external tank and was beginning to pivot on the strut from near its nose to the tank.

In less than one second, thrust from the errant plume pushed the booster's tail away from the tank and may have crashed it into the orbiter's right wing. Debris recovered at sea suggests that the plume scorched the orbiter's right aft side.

As the booster swiveled, its nose crashed into the tank, crumpling it between the hydrogen and oxygen chambers. This appears to have been what caused the fireball.

At 73.175 seconds, cameras show, a huge cloud of gas -- leaking hydrogen or oxygen or both -- began streaming out of the tank just where the booster's nose hit. Near the bottom of the tank, the leaking gases flared and flickered momentarily before a flash of fire burst forth under the orbiter's nose.

Less than three-hundredths of a second after the flash, at 73.226 seconds, the enormous fireball burst forth, releasing both boosters to fly on as the explosion disintegrated the tank and the orbiter. Tracking cameras recorded the right booster rocketing away, a huge plume still spewing from its ruptured side, until it and the left booster were deliberately blown apart 37 seconds later by radio command from the ground.

Meanwhile, data transmission from the orbiter continued automatically for almost another half-second after the explosion started. At 73.534 seconds, onboard computers, sensing a loss of fuel, did what they were supposed to do and shut down one of the main engines in the orbiter.

Challenger's last spark of life, its automatically transmitted stream of computer data updating Mission Control on the chaotically changing status of the spaceship, flickered out at 73.605 seconds after liftoff