BOSTON — A human test of the American government’s experimental AIDS vaccine, similar to a failed Merck & Co. product, was canceled after a top scientist determined it was unlikely to give useful results.

The director of the National Institute of Allergy and Infectious Diseases in Bethesda, Md., Anthony Fauci, said yesterday he was unwilling to contribute the resources needed for the trial. The test, PAVE-100, was set to include about 3,000 people, and was originally planned to enroll as many as 8,000.

Mr. Fauci went against the recommendation of an advisory panel that voted May 30 in favor of conducting the test in some form. He said he would consider a smaller trial. The government’s vaccine is intended to stimulate immune cells to reduce or eliminate levels of HIV, the AIDS virus, in the blood.

“Given the fact that there are not a lot of leads on an AIDS vaccine, I’m not willing to entirely shelve the concept,” Mr. Fauci said in a telephone interview. “But it would have to be a leaner, meaner, less-expensive trial, with less people, that’s focused on the question of whether the vaccine can lower viral load.”

Merck, based in Whitehouse Station, N.J., developed a vaccine called Ad5, which was also intended to lower the amount of virus in the bloodstream. An international test of the vaccine in about 3,000 people was halted in September when 49 HIV infections occurred among those who received it, while just 33 who got placebo vaccinations caught HIV. That suggested the vaccine may have inadvertently increased HIV risk among people who were exposed to blood or semen containing the virus.

Both the Merck vaccine and the government’s experimental shot, called VRC for the Vaccine Research Center where it was developed, contain a cold virus called adenovirus-5. Because of that similarity to the Merck vaccine, the VRC shot probably wouldn’t have gained market clearance, scientists have said.