A judge has ruled for the first time that fingerprint evidence, a virtually unassailable prosecutorial tool for 90 years, does not meet the standards set for scientific testimony and that experts in the field cannot testify that a suspect's prints definitely match those found at a crime scene.

The decision, by a senior federal judge in Philadelphia, comes after two years of efforts by defense lawyers to hold fingerprint analysis to standards set by the Supreme Court in 1993.

The judge, Louis H. Pollak, found that fingerprint analysis had not been subjected to the rigorous testing required under those standards.

Judge Pollak ruled that fingerprint experts could still point out the similarities between prints from a crime scene and those of a defendant. They may also still point out that no two people have the same prints. But, the judge wrote, "what such expert witnesses will not be permitted to do is to present `evaluation' testimony as to their `opinion' that a particular latent print is in fact the print of a particular person."

The decision, issued on Monday, is not binding on any other court and is unlikely to affect past convictions, legal experts said. But it is expected to prompt challenges in other jurisdictions, not just to fingerprints but to other techniques like ballistics and handwriting and hair analysis.

Several legal experts said the careful reasoning of Judge Pollak's decision would make it hard for other judges to disregard it.

"I think it's going to be a blockbuster opinion, because fingerprints are always held up as the most reliable type of evidence," said Paul C. Giannelli, a forensic expert at Case Western Reserve University Law School who was the co-author of the book "Scientific Evidence."

The executive director of the California District Attorneys Association, Lawrence Brown, said that a single ruling from a federal judge did not mean the demise of fingerprinting. But Mr. Brown said that if Judge Pollak's ruling became the legal standard it would make prosecutors' jobs more difficult.

"To be able to talk only about how many `points of similarity' there are in a fingerprint — a jury might be left somewhat confused as to what all that means," Mr. Brown said. "Juries rely on experts to assist them in making important decisions, and if the expert wouldn't be able to give that final punch line that this is in fact the accused's fingerprints, it would undermine cases to some extent."

Law enforcement officials have been concerned that a decision like Judge Pollak's would cast doubt on other forensic techniques. David L. Grieve, the editor of fingerprinting's main trade journal, the Journal of Forensic Identification, warned in 1999 that "if the court determines that fingerprint analysis is not a science, other types of forensic examination will surely follow."

Handwriting and hair analysis have already been weakened by other rulings, and Professor Giannelli said that ballistics — the practice of identifying guns by the marks they leave on bullets — was now particularly vulnerable because guns, unlike fingerprints, change the marks they leave over time and with use.

DNA analysis, usually presented in terms of statistical probability of a match rather than as certainty, has been found to meet the Supreme Court's standard for a science.

Judge Pollak, a former dean of the Yale and University of Pennsylvania law schools, issued his decision as a pretrial ruling in a murder case in Philadelphia in which prosecutors used fingerprints to link two defendants, Carlos Ivan Llera Plaza and Wilfredo Martinez Acosta, to four drug-related murders in 1998. The men's lawyers had asked that the fingerprint evidence be declared inadmissible.

A spokesman for the United States attorney's office in Philadelphia, Richard Manieri, said of the decision, "We're evaluating it and considering our options."

The decision was first reported on Wednesday in the Legal Intelligencer, a newspaper in Philadelphia.

The reliability of fingerprint analysis went unchallenged for decades. But two Supreme Court decisions on expert testimony made the practice vulnerable. In two product liability suits the court declared that federal judges must determine the reliability of expert testimony before admitting it. The court held that an expert could not testify about the scientific certainty of a technique unless its accuracy had been tested and its rate of error was known.

The first legal challenge to fingerprinting under this standard, filed by a federal public defender, Robert Epstein, in 1999, was rejected. But many defense lawyers adopted Mr. Epstein's arguments that standards of what constituted a match and of training varied widely and that the accuracy of crime-scene print identification had never been scientifically tested.

Earlier in 1999, the Justice Department began the process of putting together a study of fingerprint identification's accuracy.

Both sides in the current case agreed to let Judge Pollak base his ruling on the court record in Mr. Epstein's case.

Fingerprint analysts work by comparing crime-scene prints, which are often smudged, distorted or fragmentary, to "clean" prints taken from a suspect.

Judge Pollak dismissed as irrelevant the government's position that while fingerprint experts made subjective decisions and could occasionally make mistakes, their methods, correctly applied, had an error rate of zero.

For Mr. Epstein's case, the Federal Bureau of Investigation sent the defendant's official prints to 53 law enforcement agencies. Eight of the 34 laboratories that responded were unable to match at least one of the two crime-scene prints with the defendant's.

"While the survey results fall far short of establishing a `scientific' rate of error," Judge Pollak wrote, "they are (modestly) suggestive of a discernible level of practitioner error."

Mr. Epstein called the ruling, "the decision we've all been waiting for," adding, "He didn't go all the way, but he went 98 percent of the way."

Mr. Brown of the district attorneys association noted that fingerprint evidence has long been so persuasive that when prosecutors do not have it, defense lawyers routinely attack the case as weak.

"It's somewhat amusing," Mr. Brown said. "Now they want to also cast aspersions when there is a fingerprint."

Copyright © 2002 ABCNEWS Internet Ventures.