Thirty-five of the 39 D.C. police recruits suspended Monday for alleged marijuana use denied the allegations yesterday and said they were willing to take lie-detector tests to prove their point, according to a police union official.

Gary Hankins, president of the labor committee of the Fraternal Order of Police, met with the suspended recruits yesterday at police headquarters and told them he would meet today with Police Chief Maurice T. Turner to determine whether the April 5 urine tests that led to the suspensions were properly administered.

Police spokesman Lt. Hiram Brewton said yesterday that Turner would not fire the recruits until he reviewed the results of a follow-up urine test administered Monday and a police internal investigation of each of the 39 suspended recruits.

"He is not just taking this test and saying 'You are terminated.' He will review it on a case-by-case basis," Brewton said.

"This investigation will probably make him aware of who has used (drugs) in the past and who is using it currently," Brewton said. He said he was unsure how Turner would react if the investigation showed that the marijuana use occurred only before the recruits enrolled in the police academy last December.

FOP lawyers said yesterday they have serious questions about the accuracy of the tests, but said they were unable to investigate further because the police department has not yet released specific information about the type of test used.

Hankins said all the accused recruits, except several he has not spoken to, have denied using drugs. However, Robert E. Deso, an FOP lawyer, said, "Some are vehemently denying it, while others have admitted it." He said he did not have specific figures.

Kevin Zeese, executive director of the National Organization for Reform of Marijuana Laws, said yesterday that a "significant" number of persons can be wrongly accused of marijuana use through the use of such tests, which he said have been developed only in the last few years.

Zeese said the more advanced testing method is regarded as only 95 percent accurate, with frequent cases of positive readings later found to be wrong in subsequent tests. In addition, he said, persons can test positively through "passive inhalation" in which they inhale marijuana smoke at parties or rock concerts.