In Thursday's editions, the National Transportation Safety Board's position on mandatory drug testing for railroad workers was misstated. The board supports mandatory testing only after accidents or in cases where employers have "reasonable cause" to suspect drug use. (Published 1/23/88)

The National Transportation Safety Board concluded yesterday that the engineer of a Conrail locomotive was impaired by marijuana when his freight train ran through stop signals and a closed switch in front of a speeding Amtrak passenger train near Chase, Md., a year ago.

The crash, on Jan. 4, 1987, killed 16 persons and injured 175.

The safety board, ending its investigation into an accident that triggered a nationwide review of railroad safety and drug-testing practices, also called inadequate the Federal Railroad Administration's supervision of Amtrak and Conrail operations in the busy Northeast rail corridor, which runs between Washington and Boston.

The board chided the FRA for failing to require that locomotives in the corridor be equipped with devices that automatically slow or stop trains that fail to respond properly to signals.

"We would like to see, immediately, every train in the Northeast corridor equipped with automatic train control," said board member Joseph Nall, who supervised the year-long investigation. "We are very concerned with the possibility of another accident like this occurring."

The board's recommendations prompted a critical response from FRA chief John Riley, who said they fell far short of the steps needed to prevent railroad workers from behaving irresponsibly.

"While I'm generally in agreement with their findings" on the cause of the accident, "I'm very disappointed in the recommendations," Riley said. He said the board sidestepped the need for federal authority to sanction railroad workers who commit safety-related offenses, and the need for mandatory random drug testing.

Railroad workers, unlike airline pilots or bus drivers, operate on private property and traditionally have been beyond the reach of federal regulation, except through sanctions imposed on their employ-ers. As currently written, FRA rules require drug testing after any significant accident and allow companies to test their employees if they have "reasonable cause" to suspect drug abuse.

"If we cannot get the authority to add a random testing program . . . we will be no more able to prevent recurrence of another Chase accident than we were able to prevent its occurrence a year ago," he said. "That would be a tragedy because it would mean that those people died in vain."

Riley said he agreed with the need for automatic train control, but he defended the FRA's failure to require the devices, noting that the safety board had itself rescinded a similar recommendation in 1982. "The truth of the matter is that every agency that dealt with this issue mishandled it, and I think the board should be candid enough to add its name to the list," he said.

A safety board spokesman said the board has acknowledged its failure to pursue the automatic train control recommendation.

The spokesman declined to comment on the need for expanded regulatory authority, but said the board has not supported mandatory drug testing because it feels that existing drug testing procedures "should be tried first."

In its statement of the accident's probable cause, the board lay the blame on Conrail engineer Ricky Gates and, to a lesser extent, brakeman Edward Cromwell, both of whom showed evidence of marijuana byproducts in blood and urine samples taken after the accident.

Safety board investigators said that the two men failed to adequately check warning signals inside the cab before their string of three locomotives departed the Bay View yard in Baltimore; that a shrill whistle designed to alert the crew to signal changes had been muted with duct tape, and that the brakeman failed to alert Gates to a warning signal that could be viewed from a distance of 5,000 feet.

Gates, who was indicted on 16 counts of manslaughter in May, goes on trial in Baltimore next month.