Salvage workers, battling Potomac River water temperatures near freezing and visibility of less than a foot, recovered 30 additional bodies from the Flight 90 crash site yesterday.

Divers were sure by late afternoon that they had recovered all the bodies from the tail section of the Air Florida Boeing 737 lodged in the river, according to Francis H. McAdams, the National Transportation Safety Board member who is overseeing the investigation into Wednesday's crash, which killed 78 people.

The workers attempted but failed to fasten a sling around the jet's tail section, believed to contain the vital flight data and cockpit voice recorders that investigators need to help determine what happened to the plane. "They were unable to get straps in the proper position to raise it," McAdams said.

Federal investigators also turned their attention yesterday to a problem that the 737, a widely used twin-engine jet, has experienced in the past when taking off under weather conditions similar to those existing Wednesday.

The Boeing Company recommended in 1979 and again last June that special procedures be followed when 737 pilots take off in near-freezing temperature with snow, sleet or rain. The company made that recommendation because it was concerned about reports it had received from operators of its 737 jetliners that they had had problems with unexpected nose-up and side-to-side movements when taking off in such conditions.

Air Florida Flight 90 apparently pitched up before falling onto the 14th Street bridge Wednesday, according to witnesses and survivors. But the airline had adopted Boeing's recommended changes in its flight manual prior to the accident, according to both the Federal Aviation Administration and Air Florida. Further, the positioning of controls on the wing recovered from the Potomac Friday suggests that settings on the Air Florida plane followed the Boeing recommendation, McAdams said. Nonetheless, federal investigators were studying the Boeing recommendations carefully.

Other aspects of the investigation continued, with investigators hearing conflicting reports on whether the plane took off with ice and snow on its wings and fuselage. The investigators also established a special committee to examine longtime safety issues surrounding National Airport. McAdams said the group will study the length of National's relatively short runways, rescue operations following the crash and the routes aircraft are required to fly when using National.

At the crash site yesterday, divers wearing cold weather gear to keep them warm in the near-freezing water literally had to feel their way around the wreckage to do their work.

D.C. Police Inspector James Shugart said that if any more bodies were discovered in the fuselage they would be removed before any attempts were made to raise the wreckage.

A total of 49 bodies have been recovered since the crash, 45 from the plane and four who died in cars on the bridge, McAdams said. It is believed that the bodies of 29 people remain in the water.

The remains recovered yesterday--15 men, 13 women and two whose sex had not yet been announced--have been taken to the office of the D.C. Medical Examiner for identification and notification of relatives. Although the frigid water has preserved facial appearances, Shugart said that identifying women has proven to be a particular problem.

"Men normally carry some form of identification in their jacket pockets," he said, "but women usually carry it in their purses and of course their purses aren't with them."

The next phase of the recovery operation, Shugart said, is removal from the river of the jet's tail section, which is attached to a 40-foot length of fuselage. Attempts to recover the tail section were suspended early last evening.

That unsuccessful attempt found divers suspending cables from a mammoth 90-ton crane atop the center span of the bridge beneath the water and under the fuselage. "We don't know if we have enough equipment," Shugart said. He added that a barge-borne crane was en route from Indian Head and is expected to arrive some time today.

Besides the problems of ice, jagged metal edges on the wreckage and tricky currents was the anticipated onslaught of subzero cold heading toward the Washington area from the Midwest. "We believe the weather will be extremely cold tonight and that the river will freeze over again so that we'll have to begin tomorrow morning in breaking the ice again," Shugart said. "But we expect to continue as long as the divers consider the situation safe."

The NTSB's McAdams said last night that divers will attempt to start again after daybreak. A helicopter, equipped with special electronic gear designed to precisely locate large pieces of metal, will be used, McAdams said.

Although the focus of the investigation continued to be taken up yesterday with the question of ice on the wings, the uncovering of an old problem involving the performance of the 737 in frigid weather added a new element.

Boeing's concern about bad weather and the 737 surfaced in a February 1979 recommendation to client airlines that followed a series of at least four incidents in Europe where 737s pitched upward or rolled unexpectedly. All the incidents were handled without significant incident, according to officials, but resulted in study of the problem.

Boeing's first recommendation concentrated on problems that had occurred when ice formed on the front edge of the wings after being blown there by the engines. That happened if baffles behind the engines--called reverse thrusters--had been used. The reversers diverted light snow to the wings, where it formed again as clear ice.

Boeing said the front of the wings should be examined and ice removed before takeoff.

Last June, the Boeing directive to airlines said the "roll off and/or pitch up" incidents usually involved several factors, including the near-freezing weather conditions; certain wing-control settings on the flaps (panels on the rear of the wing); the use of thrust reversers, and an excessive angle of climb after the plane left the ground.

Boeing recommended specific flight procedures to counter those conditions, including close monitoring of the angle of climb, somewhat higher power settings and special flap settings. McAdams said yesterday that the left-wing flaps, recovered Friday, were set in a recommended position.

The Federal Aviation Administration instructed U.S. airlines to adopt those changes. The British Civil Aviation Authority, the equivalent of the FAA, did not adopt them until Friday.

However, Alexander MacKellar, civil air attache to the British Embassy here, said that the date was "purely fortuitous and in no way was it related to the incident and crash that occurred here."

McAdams said investigators had determined that the Air Florida plane had received a de-icing treatment "very close to 43 minutes" before takeoff, during which time ice and snow may have formed again on the wings and fuselage.

If an airplane pulls away from the ramp and gets delayed, McAdams said, it's the pilot's decision about whether he should go back again for another de-icing treatment. If the pilot can't see the outside of the plane, he should call an airline representative to check it, McAdams said.

Friday, a Braniff pilot told investigators that he had seen snow and ice on the fuselage of Flight 90 as it was taking off. Yesterday, another pilot who had been on the National taxiways at the time Air Florida took off told investigators he looked at Flight 90 specifically to see if it had ice, because he was using it as a guide for his own plane. He could not see the wings, he said, but saw no ice on the fuselage, McAdams said.

Robin Cohn, a spokesman for Air Florida, said that airline's policy is that "If the delay is extended after de-icing, the captain or first officer should make a visual inspection by going through the cabin, looking out windows and checking. . . . We do not know if it was done in this case."

When asked the definition of extended delay, she said, "I don't have anything; I would imagine it's judgment."

A decision on whether to take off in bad weather is reached jointly by the pilot and the dispatcher, according to Dave Mulligan, Air Florida's vice president for operations. The dispatcher sits in Miami, but monitors weather conditions elsewhere. The pilot accepts the dispatch, but has the authority to decline to fly, Mulligan said.

The pilot, Larry Wheaton, came to Air Florida in 1978 when it acquired Air Sunshine, a local Florida operation for whom Wheaton had flown DC3 propeller planes. Wheaton was a 737 pilot for 18 months, Air Florida officials said.

First officer Roger Pettit joined Air Florida in September 1980 and had been a copilot on 737s since then. He had extensive Air Force experience as an F15 Eagle jet pilot.

Investigators were checking the records and training of the two cockpit crew members yesterday as a matter of routine.

McAdams said last night that the board has located a "good" witness--one with substantial aviation experience--who was on the 14th Street bridge at the time the plane crashed. When he first saw Flight 90 approaching the bridge, the witness said, he thought it was going to land, then realized that was not likely. He thought the plane turned right, then came to a wings-level position, "then mushed into the bridge in a nose-high attitude," McAdams said. That witness did not think the landing gear was extended, McAdams said.

While investigators have been most anxious to examine the two recorders, they are also interested in the two engines. McAdams said both had settled on the bottom of the river and have been marked with buoys. The engines can be checked to determine if they were functioning properly at the time of impact. A failure of engine power is another of the many possibilities investigators want to check.

FBI agents also continued to be part of the investigating team. That is normal procedure until sabotage is ruled out as a possible cause. There is no evidence of sabotage, officials said.