First readings of the flight data recorder recovered from the Air Florida jetliner that crashed into the Potomac 13 days ago indicate that the plane took an abnormally long time to reach takeoff speed on the runway, federal safety investigators said yesterday.

Seconds after liftoff, the plane did reach a climb speed that is normally considered safe but then began to slow down, according to the recorder. The jet climbed to about 340 feet and then began to fall. After less than 30 seconds in the air, it swiped the northbound span of the 14th Street bridge and plunged into the frozen river. Seventy-four people aboard the plane and four on the bridge were killed.

Yesterday, Dr. James L. Luke, the chief D.C. medical examiner, said 73 of the victims on the plane either were killed instantly or suffered injuries so severe that they could not have lived had they been rescued. He said the large majority suffered skull fractures and many had brain hemorrhaging.

Luke said the plane's pilot, Larry Wheaton, 34, of Miami, suffered "severe head injuries that were almost immediately fatal." A deputy medical examiner, Dr. Douglas Dixon, said the copilot, Roger Pettit, 31, of Miami, also was killed virtually instantly. Both doctors said toxicological tests on both pilots showed no drug or alcohol use.

The medical examiners said one victim aboard the plane drowned. He was Arland D. Williams, a 46-year-old directing bank examiner for the Federal Reserve Bank in Atlanta.

But the examiners declined to say that Williams, who had no broken bones, was the man a Park Police helicopter crew saw selflessly passing a lifeline to others in the river before he died. Williams had a beard, and the helicopter crew says the man passing the lifeline did not.

"We can't be certain it was Williams," Luke said.

In describing the data recorder findings yesterday, Francis H. McAdams, a member of the National Transportation Safety Board, said they indicated that the Boeing 737 took 47 seconds to reach 166 miles per hour, a speed that allows a safe climb just after liftoff. It should have reached that speed in a little over 30 seconds, he said.

"It appears that the takeoff roll was longer than a normal takeoff roll," McAdams said. Survivors have said that they sensed the jet was slow in gaining momentum as it took off, while witnesses on the ground have placed the liftoff point about 1,300 feet farther down the runway than normal.

The recorder's readout could fuel controversy over the length of National Airport's runways and lack of precise runway distance markers. The runway the Air Florida jet used is just under 7,000 feet long; at many commercial airports runways extend 10,000 feet and more.

The findings also could lend credence to speculation that one or both of the two jet engines were putting out insufficient thrust as the plane took off in a snowstorm at about 4 p.m.

One theory holds that sensors inside the engines might have been frozen, giving the crew false readings on thrust. Another holds that snow or slush on the runway might have retarded acceleration.

However, McAdams stressed that it is too soon to make firm judgments on why the plane crashed. Other analysts have cautioned that the recorder, which etches out graphs of airspeed, altitude, bearing and "vertical G-forces" (the force with which a plane bounces up and down), can give distorted information.

The data will be analyzed in Seattle by computer on a flight simulator belonging to Boeing, the 737's manufacturer, then will be compared with information from the cockpit voice recorder, air traffic control communications and radar in an attempt to reconstruct the jet's brief flight.

The recorder data left unanswered why the plane reached what should have been sufficient speed to keep it in the air but still crashed. Its lowest airborne speed was about 150 mph, or about seven mph over its theoretical "stall speed," at which air ceases to move over the wings fast enough to give sufficient lift.

One line of speculation theorizes that the plane raced down the runway caked with ice that hampered the wings' ability to lift. A phenomenon known as "ground effect," a cushion of air that builds up under planes on takeoff, might have helped the 737 off the ground. But when ground effect wore out as it climbed, it might have had trouble staying airborne, according to the theory.

Yesterday, McAdams noted again that the plane's nose apparently was angled upward too sharply shortly after liftoff. This could have caused drag or changed the flow of air over the wings, adding to any control problems the pilots were already experiencing due to ice.

Pilots say that when faced with dwindling speed and altitude shortly after takeoff, they often react by increasing the plane's pitch, to try to climb. In these circumstances, however, such a reaction might have worsened the plane's predicament.

Another explanation for the pitch blames a control peculiarity of the 737. Last year, Boeing put out a bulletin to airlines warning that during takeoffs in snow, 737s had shown a tendency to pitch up at the nose without warning.

Yesterday, working amid mangled wreckage laid out on the concrete floor of a National Airport hangar, investigators continued to search for clues to the crash. Sources said examination indicated the plane's landing gear was down at impact, as many witnesses had said.

The plane was going at least 150 mph when it hit the bridge, and the severity of the impact was clear. Dixon, the deputy medical examiner, said that 57 of the 74 people on the plane who died from injuries so severe that their deaths were almost instantaneous. Of the other 16 besides drowning victim Williams, Dixon said, 11 had at least two injuries that were life-threatening and another five had one such injury.

Luke said, "It's hard to say how many were conscious or unconscious" after the impact of the crash. "Most of these people had severe injuries."

He said there was "no way to know" how long the 17 lived who were not killed instantly.

Dixon said 26 of the 74 victims had water in their lungs, but said nonetheless that, with the exception of Williams, their deaths were caused by other factors.

Luke, Dixon and another deputy medical examiner, Dr. Brian D. Blackbourne, performed the autopsies. Blackbourne said they were able to identify the crash victims through a variety of types of identification, including fingerprints, clothing, jewelry, dental records, and most importantly, pictures shown to relatives of the victims.

Luke said all the victims were identified by at least two means. He said that after the bodies were brought to the city morgue, the examiners looked at the victims' personal belongings, and later fingerprinted them and searched for any body marks, such as surgical scars, that might help in identification. Eventually, autopsies were performed and the bodies were released to relatives for burial.

Blackbourne said it took four or five days to identify one woman, whose identity was eventually learned from two scars on her neck from recent surgery. But he said the identity of most victims was determined within 24 hours after they were brought to the morgue.

Across town at the Washington Hilton, Melvin M. Belli, the flamboyant San Francisco lawyer once dubbed the "King of Torts," announced at a press conference that he plans to file lawsuits within a week on behalf of 10 victims and a survivor aboard the plane and for a person injured on the 14th Street bridge. Belli said he expected all the cases would be consolidated in the U.S. District Court in Washington.