Amtrak's Night Owl passenger train reached a speed of about 110 mph -- 10 miles an hour above the speed limit -- in the last 1 1/4 miles before it derailed on a sharp curve and slammed into a crowded commuter train in Boston's Back Bay station last Wednesday, the National Transportation Safety Board said yesterday.

The wreck injured 265 people but caused no deaths.

Preliminary data from a locomotive "event recorder" showed that the brakes were applied at some point during the final distance, and the train's last recorded speed was 77 mph, according to a statement by Susan M. Coughlin, the board's vice chairman. She said investigators had calculated the "overturn speed" on the curve as 60 mph.

Coughlin also reported that drug and alcohol tests were negative for operating crew members of both trains. This included apprentice engineer Richard Abramson, who was operating the train, and engineer Willis Copeland, who was supervising Abramson on his third trip over the line since completing Amtrak's engineer training course.

Abramson and Copeland told similar stories in two interviews each with investigators: that they began braking the train at a signal at Ruggles Street, 6,380 feet from the point of derailment, but they did not get the braking action they expected. Abramson said he applied emergency brakes just before the derailment.

The event recorder does show that brakes were applied, but Coughlin said it would take weeks to determine where, how much they were applied and how effective the braking was.

Board investigators also determined that the brakes on all the Night Owl cars were in the applied position after the accident, but brakes go on automatically in a derailment, so this would have no bearing on the question of the engineer's actions. It does mean that all brakes on the train were in working order, investigators said.

The board said last week that preliminary tests indicated that the brakes might not have been working properly on the locomotives, even though they were clearly functioning on the rest of the train and those would normally be enough to achieve a stop.

The event recorder showed for the first time that the train was exceeding the speed limit, which for several miles leading into the station is 100 mph. The event recorder showed a speed of 110 mph at some point after Ruggles Street.

The board did not comment on the speed, but a railroad operating specialist said that under normal circumstances a train should be able to slow sufficiently if it starts braking at the Ruggles Street signal, even at the higher speed.

The board at first thought information from the event recorder was lost when Amtrak personnel placed a tone on the tape that was intended to be a brief marker tone, but which instead spread across data from the last five miles of travel.

Working through the weekend, safety board engineers were able to recapture the data, except for one key element -- the time of each event.

"Our experience working with this damaged tape underscores our belief that event recorders should be preserved after an accident for the initial use of federal investigators, as is the case with aircraft recorders," Coughlin said.