COLUMN ONE : When the Mind’s Eye Blinks : Eyewitness testimony carries great weight with juries. But experts say memory can play tricks that distort the truth and sometimes may send the wrong person to jail.

The distraught witness recalled her hour of terror in vivid detail for the jury: Two men abducted her from a parking lot, yanked down her jeans and raped her in the front seat of her car.

Asked to identify one of the rapists, the victim pointed at Frederick R. Daye, the defendant. “There is no doubt in my mind,” she testified.

Without realizing it, the San Diego woman had identified the wrong man. A second witness corroborated her misidentification, testifying that he saw Daye help push the woman into the car and drive away.

Largely because of eyewitness testimony, the San Diego barber spent 10 agonizing years in prison before the criminal justice system realized its mistake and released him in September.

The case demonstrates what legal analysts have long known: The word of eyewitnesses holds a lot of influence over juries even though memory is unpredictable. Although relatively few innocent defendants are convicted, the No. 1 reason for such errors is misidentification by a witness, usually unintentional, researchers have found.

“I am much more likely to believe a client who says he is not guilty when the only evidence against him is an eyewitness,” said North Carolina criminal defense attorney Gerald Beaver, “than I am to believe the client who has lots of circumstantial evidence against him.”

In one study by social scientists, only half of a group of bank tellers who had witnessed a staged theft attempt could pick out the correct suspect from photographs. Most of the tellers had been trained in such identification.

Eyewitnesses err about 50% of the time in experiments, according to University of Nebraska psychology professor Kenneth Defenbacher. Although the rate of mistakes is probably much smaller in the courtroom, the consequences can be devastating for the accused.

“It is the most theatrical moment in the trial when the witness points the finger at the defendant and says, ‘He did it,’ ” said UCLA law professor John Shepard Wiley Jr., until recently a federal prosecutor.

“Everybody in the jury box looks at the witness, looks at the finger and follows the line right to the defendant,” he said. “And just about every defendant squirms.”

No one knows how often the innocent are convicted because of misidentification. Estimates vary widely, from a conservative of three in California each year to thousands nationally. But researchers who study memory contend that the surprise is that so many people can accurately recall events and faces, not that an occasional distortion occurs.

The brain is not a camera, and the nervous system cannot reproduce a memory of an event exactly as it occurred, said Dr. Marsel Mesulam, professor of neurology and psychiatry at Northwestern University.

“Everything we remember is false by definition,” he said. “. . . You just hope that your re-creation is not too outlandish.”

Even a particularly vivid memory can be distorted by the mind. The brain uses the same neurons over and over again to record new information, much like a computer re-records areas on a hard disk. People can confuse the truth with what was imagined or was suggested by another witness or police, or mix up memories.

“The simplest example is somebody who parks their car in a different place every day,” said Howard Egeth, a Johns Hopkins University psychology professor. “The trace from yesterday is going to intrude, and you are not going to be able to remember where you left it today.”

When Daye was tried for rape in 1984, his court-appointed attorney vigorously attacked the witnesses’ identifications and produced the man authorities now suspect may have committed the crime. The judge even cautioned the jury about the potential unreliability of eyewitness accounts.

But two of the jurors who convicted Daye, then 26, recall that the victim’s identification cinched the case for them. She was so certain, they said.

Daye resembled the rapist that the victim initially described: He was an African American of medium height with facial hair and a metallic front tooth. But the man now suspected of having been the real perpetrator has a gold tooth; Daye’s is silver. Daye is wiry, the other man more muscular, Daye’s lawyer said.

Some studies suggest that misidentification is compounded by racial differences. Daye, however, was identified both by the victim, who is white, and the witness in the parking lot, who is black.

Daye had served a one-year sentence for a Minnesota robbery in 1977 and had given police a false name when they questioned him during a traffic stop eight days after the rape. He did not take the stand to protest his innocence because that would have allowed the jury to learn of the prior conviction.

From prison he wrote letters to talk-show hosts and asked them to investigate. He combed the telephone book, sending appeals to anyone who might be able to help. No one responded. His hair turned gray; at times he wished another inmate would kill him.

The man who was convicted of being the other rapist in a separate trial swore from prison that Daye was not involved. A lawyer appointed to look into the matter for two years failed to take action. Daye said the lawyer did not even return calls.

Daye was one of an estimated 77,000 suspects across the nation who are arrested each year primarily based on eyewitness identification, according to a survey of prosecutors published in 1988.

Another survey of participants in the criminal justice system led Ohio State University professor Ronald Huff to estimate conservatively that several thousand defendants each year are wrongly convicted of the eight most serious felonies, including rape and murder. Mistaken identification, the criminal justice researcher found, was the leading cause.

Some criminal defense lawyers cite FBI statistics showing that one-third of 2,500 rape suspects each year are exonerated by DNA testing.

But UC San Diego professor Ebbe B. Ebbesen said police and prosecutors usually weed out unreliable witnesses before trial. Ebbesen, who estimates that misidentification produces three wrongful convictions in California each year, said witnesses also excuse themselves by admitting they cannot remember clearly.

That is not much comfort to the wrongly convicted, who are forced to prove their innocence from behind bars. During the last few years, more than a dozen prisoners in the United States have been released because advocates arranged for DNA tests of evidence at the crime scene.

Daye owes his release in part to luck. It began when Carmela Simoncini, an appellate criminal defense lawyer, took over his case. The tiny, bespectacled woman quickly grew to like Daye. He was so considerate, she said, always apologizing for taking up her time.

One day she happened to learn that the evidence in the case was about to be destroyed. Without any plan in mind, she decided to go to the San Diego courthouse to look at the crime’s remnants.

She found a large cardboard box that contained the victim’s clothes, blood vials and stain swatches. Looking at the victim’s semen-stained jeans, she wondered: Could a DNA test be done? Daniel T. Williams, who prosecuted Daye, was in his San Diego law office a few months later when he got a call from the San Diego district attorney’s office.

“Sit down,” a former colleague told him. “I have some news for you.”

Over the objections of the district attorney’s office, Daye’s lawyer had obtained $2,000 to do a DNA test. The results excluded Daye as either of the rapists.

Williams was horrified. Other defendants he had prosecuted flashed through his mind. Could there be another innocent man he had helped send to prison, he asked himself. The answer, he concluded, was no. Daye, he said, had been “the ultimate nightmare.”

Aside from their legal and ethical responsibilities, prosecutors want to ensure that a witness is reliable because a conviction may depend upon it. Former prosecutor Wiley recalled his shock a year ago when his lead witness in a bank robbery pointed to the wrong man in the courtroom.

The witness, a bank teller, had correctly identified the robber from a photo spread months earlier. But during the trial she pointed to his co-defendant, whom she had never seen before.

“I couldn’t believe this happened,” Wiley said. “It opened new vistas of possibilities for me--that someone could be that wrong on something that important.”

Brian Piszczek was behind bars for four years because a rape victim had been that wrong. She identified him as the man who attacked her in her darkened Cleveland living room.

Even though the description she initially gave police did not match Piszczek’s, she later became convinced that he was the rapist. The two had met once before.

During the trial four years ago, the victim wept as she recounted the rape in brutal detail.

“I looked at the jury,” Piszczek, 27, recalled, “and a couple of the ladies’ mouths dropped. I could read their minds. They were thinking, ‘You horrible, filthy man you.’ ”

Piszczek was released in September after a group of New York lawyers took an interest in his case and arranged for a DNA test.

“I just want to settle down now and start a family,” said the factory worker, who lives with his parents in Ohio.

Timothy Hennis, 36, suffered a more horrifying ordeal. He was sentenced to die in 1986 for savagely slashing to death a mother and her two little girls in their home in a small North Carolina town. A witness testified that he had seen Hennis outside the victims’ house on the night of the murders, and other circumstances indicated a possible connection.

But none of the evidence in the house implicated Hennis: not the semen (the woman had been raped), or the fingerprints, bloody shoe print, and the pubic hair that police found.

*

A jury convicted him anyway. He spent two years and four months on Death Row before he was acquitted in a second trial. The defense discovered that a man who resembled Hennis had regularly walked the victims’ neighborhood at night.

Hennis returned to the Army and resumed his life with his wife and baby. He served in the Gulf War, had a second child and was promoted.

Now he worries about how he will save enough money to send his children to college and repay his parents more than $100,000 they spent on his legal fees.

He wants to put the nightmare behind him: the hearse he saw from his cell window as the body of a fellow inmate was removed after his execution; the certainty of the community that Hennis was a triple murderer, and the fact that the gruesome crimes remain unsolved.

“Some of it fades,” Hennis said in a recent interview from Florida. “Some of it I just let go, otherwise I would be too angry about it.”

Such cases have made the science of memory a legal battleground. Many states, including California, allow experts to testify about eyewitness reliability.

About five years ago, concern about eyewitness reliability prompted England to bar trials when the only evidence against a suspect is an eyewitness, said Kathy Pezdek, a psychology professor at the Claremont Graduate School.

“When witnesses make a misidentification,” she said, “it is most often an unconscious process and it doesn’t mean that they are not doing their best.”

But potential problems could be alleviated long before a case gets to court if police took more care with such witnesses, said Gary L. Wells, professor of psychology at Iowa State University in Ames.

When witnesses are asked to look at a lineup, Wells said, they should be told that the perpetrator might not be present. The foils should also closely resemble the suspect, he said, and the officer conducting the lineup should be unaware of the suspect’s identity. Even a careful detective, he said, can give unconscious clues that direct the witness.

Memory could be improved if witnesses wrote down what they saw as soon as possible after a crime. “A lot of information is lost very quickly,” Wells said. Witnesses are also advised to look for a peculiar feature, like a scar or the location of a mole, that will help authorities later. Hypnosis is rarely helpful, experts say, because it can be suggestive and taint the witness’s testimony.

Although experts do not know why memory fails in certain situations, studies offer some possible explanations. Some suggest that it is more difficult to remember a face when under extreme stress.

Risk of misidentification is compounded if the upper part of a suspect’s face is obscured, according to Pezdek, because that is where people tend to focus. “The best disguise is a baseball cap and sunglasses,” she said.

People also categorize an experience in terms familiar to them and mold their recollection of events accordingly, said Johns Hopkins’ Egeth. Expectations and attitudes about an event can shape memory.

Participants in one study were told an Eskimo tale, then asked questions about it. A character died and something black came out of his mouth.

“People end up saying, ‘The person died and he frothed at the mouth,’ ” Egeth said. “They take a more familiar concept and insert it.”

*

Just because a memory feels authentic does not mean it is accurate. “Your brain may be re-creating something very vivid,” said Northwestern’s Mesulam, “but that doesn’t prove that what was being re-created was true. Hallucinogens are the best example of that.”

The woman who had been so certain that Daye raped her clung to her convictions even after she was told that DNA had exonerated him.

“It has upset her quite a bit,” former prosecutor Williams said.

San Diego authorities are investigating the possibility of bringing charges against the man whom the other rapist identified from prison as his accomplice. It was the man Daye’s trial lawyer had presented as a suspect many years before.

But the initial misidentifications could make prosecution difficult, particularly if the victim remains convinced that Daye did it.

After Daye was released in September, the talk-show hosts who had ignored his letters invited him to appear on their shows.

He returned to his hometown of Des Moines, Iowa, got a job as a postal worker and, shortly before Christmas, married a childhood sweetheart. Now 37, he dreams of becoming a comedian.

But the past haunts him. He wakes up in the middle of the night weeping. He feels tears running down his face while he is eating or listening to the radio.

“Even though I know I am happy,” he said, “I just start crying for no reason whatsoever.”