District crime lab officials are reviewing more than 150 firearm examinations for accuracy after the lab discovered errors by three D.C. forensic analysts who incorrectly matched bullets or shell casings recovered at crime scenes to individual weapons.

The examinations date to at least August 2015, according to federal prosecutors in Washington and the three former ballistics examiners.

The errors and retesting by the Department of Forensic Sciences (DFS) were disclosed in Feb. 17 and March 17 letters to defense attorneys by the U.S. Attorney’s Office of the District, copies of which were obtained by The Washington Post.

The department said that of 44 cases reviewed so far, two contained errors and that the errors were not in high-priority cases and had not led to an arrest, charge or conviction, as far as lab officials know.

Federal prosecutors declined to comment on whether the identified mistakes had led to arrests or prosecutions and would not say how many defendants have been notified that work done by the three examiners was part of their cases.

Prosecutors wrote in the letters that they would no longer sponsor one of the examiners as an expert witness, and the other two are no longer employed by the department.

The Firearms Examination Unit at the D.C. Consolidated Forensic Laboratory is the second-busiest unit of the lab, ranking behind only the fingerprint section, and had 10 examiners as of December 2015 — three of them full time, a department spokeswoman said. The unit conducted 559 examinations from December 2015 to December 2016 and had a backlog of 156 cases.

The review for mismatches is the latest setback for the $220 million lab that opened in 2012 and underwent a management overhaul in 2015 after prosecutors found numerous errors in the lab’s DNA testing.

The shake-up also led to questions about the independence of the lab, which suspended DNA forensic work for 10 months until March 2016.

Last month, department officials also announced that federal regulators were auditing the lab’s public health operations after testing for the Zika virus in hundreds of residents was mishandled for months. At least nine pregnant women who had been exposed to the virus were wrongly told they were negative for the infection, which can cause serious defects in developing fetuses.

The ballistics errors also come at a time of high tension — with firearms examiners and their supporters in law enforcement on one side, and many scientists, legal defense organizations and law professors on the other — over the scientific validity of ballistics tracing evidence.

In a 2014 opinion, the U.S. Supreme Court weighed in and noted the importance of defendants obtaining independent ballistics testing. The court opinion was in a case that granted a retrial to an Alabama death row inmate who was then exonerated after 30 years when new firearms experts disagreed with the original examiners’ findings that sent him to prison.

Janet E. Mitchell, special counsel to the D.C. Public Defender Service, praised the forensic department for launching the inquiry but said all firearms tracing results “should be viewed with skepticism” because of the subjectivity of comparisons and lack of research into how often marks made by different weapons may look alike. Some national studies, she noted, estimated examiner error rates as high as 1 in 20.

“While we are pleased that DFS is undertaking this review, these errors are likely not unique to these examiners or to these cases,” Mitchell said. “To the extent that this type of evidence continues to be admitted in court, judges and juries should be made fully aware of the error rates uncovered in the single appropriately designed study to date.”

U.S. attorney’s office spokesman Bill Miller said the office is working closely with the forensics department and awaiting the outcomes of reviews by the lab on accuracy, as well as reviews by prosecutors about any effects on cases. “To ensure that every criminal defendant receives a fair trial, the U.S. attorney’s office has made disclosures to defense counsel, and we will continue to share results of the findings,” Miller said.

The lab’s internal quality controls detected the problems, said LaShon Beamon, a spokeswoman for the forensic sciences department. The lab then reported the errors to prosecutors and others, she said. “This incident demonstrates some of the critical changes made” at the forensics department, which Beamon said “is undergoing a tremendous reform.”

The current review of lab matching by the three examiners was triggered when veteran firearms examiner Daniel Barrett failed an annual proficiency test in August 2016, according to details in the letter sent by prosecutors. The letter from Michael T. Ambrosino, special counsel for DNA and forensics for the U.S. attorney, went to Betty Ballester, chief of the D.C. Superior Court Trial Lawyers Association, and to the federal and D.C. public defenders.

Barrett’s failed skills test prompted the lab to check a sample of his recent work in criminal cases, where officials uncovered a serious error in which he incorrectly linked shell casings to a weapon.

Lab officials then expanded the re-examination to all 120 cases Barrett has handled since August 2015 — when he last passed a skills test — and that bigger, ongoing sweep turned up a second mistake, prosecutors said in a March 17 letter.

The two cases in which Barrett reached the wrong conclusions had each been confirmed by another colleague as part of a routine verification system used in the lab — meaning three examiners are part of the errors.

Cases by all three are now being investigated, although how the review is being conducted and which cases are being examined is not clear.

Prosecutors and the District’s lab would not say exactly how many cases are being studied or whether the review of the three examiners is limited only to cases in which each was the first to report a result.

Reached at his Virginia home, Barrett, 65, confirmed prosecutors’ account of his errors and said, “I was wrong, I don’t want to embarrass the unit.” A certified firearms examiner for more than 40 years, Barrett said he joined the D.C. police in 2007 as a civilian analyst and moved to the forensic sciences department in 2012.

Most of his reviewed lab work involved homicides and police shootings, he said, and he was a senior examiner who trained others.

Once prosecutors barred him from being used as an expert witness, Barrett said, he decided to retire, which takes effect in the summer, and is no longer working on cases.

“It’s very painful for me. I had quite a reputation. And then this happened. I lost everything,” he said with a sigh. “Maybe I’m a dinosaur. I lost confidence in myself, and I feel I let my co-workers down.”

The two examiners who confirmed Barrett’s erroneous cases were Luciano Morales and Kevin Webster, longtime D.C. police officers and firearms examiners who also moved to the city’s forensic sciences lab in 2012.

Morales, 50, defended himself and Barrett, saying their error was the result of an administrative mix-up. He blamed managers for poor record-keeping and reporting, saying, “I think the issue here is management at DFS. I don’t believe that things were done right within [the] firearms” unit.

Morales worked 27 years as a D.C. officer, was certified as a firearms examiner in 1998 and later detailed to the forensic department. He worked there until last year, then became a contractor until September. Prosecutors did not specify which of his cases are being reviewed, but Morales said he conducted only about 24 examinations in recent years.

Webster, 55, spent 25 years as a D.C. police officer and was a firearms examiner for two years before returning under a contract in 2015. He rejoined the police department in February as a senior police officer teaching cadets.

Webster said he learned only in the past week that his cases for the District lab since August 2016 — the last time he passed a proficiency test — were being scrutinized. He estimated he had worked on ballistics matches in about 25 cases since then and said he was notified that his results were being investigated because one of his cases was going to trial.

“As far as my abilities, I was trained by Morales, and Morales is an excellent, excellent trainer, and he’s an excellent examiner,” Webster said.

Ballester, the trial lawyer association chief, called the lab errors “disheartening,” particularly because the unit’s process of having a second examiner verify test results proved fallible. Ballester said that because of the cost and because requests require court approval, defense attorneys often do not order their own forensic firearms testing. As a result, defendants often rely on DFS examinations.

“They are supposed to be independent. We are supposed to trust in their results,” Ballester said.

The review of the D.C. lab results comes months after a White House scientific advisory panel renewed scientists’ challenge to whether firearms testing and other forensic techniques should be admitted as scientific evidence, reinforcing criticism by the National Research Council that matched bullets, hair or tire treads to a single source are overstated.

The President’s Council of Advisors on Science and Technology said that the one valid study, funded by the Pentagon in 2014, established a likely error rate in firearms testing at no more than 1 in 46. Two less rigorous recent studies found a 1 in 20 error rate, the White House panel said.

“Expressions of consensus among practitioners … [are] no substitute for error rates,” the council said.

A leading standard-setting group of firearms analysis practitioners sponsored by the National Institute for Standards and Technology “fundamentally disagreed,” saying the council disregarded less rigorous studies that more accurately reflect casework as well as methods and training developed over hundreds of years, in favor of a “purely statistical” approach.