LONG BEACH, CALIF., JULY 25 -- Documents apparently kept secret for more than a year by a California laboratory indicate that blood samples taken from former captain Joseph Hazelwood were mishandled and that charges of his drunkenness before the disastrous Exxon Valdez oil tanker spill may have had no legal foundation.

Hazelwood's license to operate a ship was suspended today for nine months after he pleaded no contest to charges of negligently leaving the tanker bridge in dangerous waters and violating a rule against drinking less than four hours before sailing. Administrative law judge Harry J. Gardner also placed Hazelwood on probation for one year.

The Coast Guard dropped allegations of drunkenness and improper procedure against Hazelwood and agreed to recommend that his license be suspended for no more than a year. Hazelwood's attorneys said that was a response to the new evidence, but Lt. Cmdr. John Ast-ley, a Coast Guard attorney, said the documents did not enter into the decision.

Hazelwood's attorneys released two 1989 chain-of-custody forms kept by ChemWest Analytical Laboratories Inc. They said the forms suggest a coverup of vital evidence by the firm's Sacramento laboratory, and possibly also by the Coast Guard, the National Transportation Safety Board and Alaska prosecutors.

One form shows that the laboratory noted receiving three "red-stoppered" tubes of Hazelwood's blood on March 28, 1989, not the one red-stoppered and two grey-stoppered tubes prepared by a Coast Guard health technician on board the oil tanker about 10 1/2 hours after it ran aground on a well-marked Alaskan reef.

The different colors suggest a break in the chain of evidence needed to prove the samples were Hazelwood's and call into question the testing process, because red-stoppered tubes lack a chemical needed to keep blood alcohol from fermenting. The second form shows that a laboratory employee changed the notation four days later to show two grey-stoppered and one red-stoppered tube, but neither form was given to defense attorneys despite their repeated requests for such information.

"This completely invalidates the blood test, which has been the cornerstone of all the charges against captain Hazelwood," said Hazelwood attorney Thomas M. Russo.

Michael A. Peat, director of the laboratory, said "there was no tampering done with the specimens," but referred reporters to the Coast Guard for further comment.

Astley denied any Coast Guard coverup of the evidence and said he had become aware of the discrepancy between the forms only in the last few days. He said the laboratory delayed sending material in response to a subpoena and that he did not receive the altered form until July 12. He said he first saw the unaltered version of the form when Hazelwood's attorneys showed it to him sometime within the last two months.

Astley said the Coast Guard was concerned about evidence that the samples might have been mishandled in shipping by being left too long at the wrong temperature.

Alaska Attorney General Doug Baily said state prosecutors until today had no knowledge of the conflicting forms and would immediately investigate what impact they might have on Hazelwood's conviction on a minor misdemeanor, negligent discharge of oil, which is now being appealed. An Anchorage jury earlier this year cleared Hazelwood on three felony charges, including operating a vessel while intoxicated.

Baily said the new evidence might explain why laboratory officials resisted testifying or producing relevant documents during Hazelwood's trial.

Acting Anchorage District Attorney Mary Anne Henry, one of the prosecutors in the Hazelwood case, said ChemWest and its Sacramento subsidiary CompuChem Laboratories Inc. argued that federal privacy rules prohibited release of information on Hazelwood's test and said they could give information to no one but the Coast Guard, which ordered the test. The court rejected the privacy argument, and the Coast Guard waived its right to keep the tests confidential.

Peat testified at the trial but was asked nothing about the discrepancies because prosecutors were not aware of them, Henry said. She said ChemWest is "a nationally recognized company" of good repute.

The Exxon Valdez hit Bligh Reef about nine minutes after midnight March 24, 1989, and ripped holes in its bottom that unleashed nearly 11 million gallons of oil into Prince William Sound. It was the largest U.S. oil tanker spill and led to the deaths of at least 36,471 oil-fouled birds, 1,016 otters and 151 eagles.

The test results showed Hazelwood, who had drunk-driving convictions and had been treated for alcoholism, with a .061 percent blood alcohol level and a .094 percent urine alcohol level. The prosecution failed to convince the jury that Hazelwood's blood alcohol level must have been above Alaska's legal limit, 0.10 percent, at the time of the grounding and had dissipated by the time he was tested. The Coast Guard prohibits operation of a vessel by anyone with a blood alcohol level at .040 percent or above, so it moved to suspend Hazelwood's master's license.

Hazelwood was seen drinking in the port of Valdez before the ship departed March 23 and, allegedly to attend to paperwork, left a junior officer in charge of the bridge just as the ship was moving out of normal sea lanes to avoid ice in the water. The Coast Guard cited those incidents in seeking today's suspension, but after negotiations with defense attorneys dropped the drunkenness charge and the charge alleging that the junior officer was uncertified to navigate in the sound.

Hazelwood, who showed little emotion as the nine-month suspension was announced today, declined to comment on the new blood test evidence, but his attorneys greeted it as a complete reversal.

Hazelwood attorney Michael Chalos and Russo said they hope the new information will alter the public view of Hazelwood.

The National Transportation Safety Board, the first agency to release the results of Hazelwood's blood test after the mishap, is scheduled to release its final report on the Exxon Valdez grounding on Tuesday. Hazelwood's attorneys said they asked to meet with the board before then but were told to submit any materials in writing.

NTSB spokeswoman Drucella Andersen said she could not comment on whether board officials knew of the discrepancy in blood sample documents, but Hazelwood's attorneys have not filed a formal complaint.

Each change on the altered form is neatly marked "KLM," apparently the initials of laboratory employee Karen L. Metcalf, who received the samples. Russo noted that on the day of the changes, May 1, 1989, the samples were checked out to Judy Peat, whose name appears nowhere else in the handling record. A receptionist at the laboratory said she did not know if Judy Peat and Michael Peat are related.