The Pentagon said yesterday that nearly one-third of the new recruits taken into the armed services last year were in the lowest acceptable mental category, a report certain to refuel the debate about the all-volunteer forces.

The figure for the Army alone was 46 percent, much higher than previously reported.

These new disclosures came in a report to Congress in which the Pentagon acknowledged that it had taken into the military vastly larger numbers of people from the lower mental categories than it thought it had. The report tried to explain how it all happened.

Overall, the Pentagon thought -- and had publicly reported -- that this lowest acceptable category comprised only 5 percent of recruits in all the services combined. But the correct figure, yesterday's report said, is 30 percent -- six times as high.

Rather than the 9 percent figure for so-called Category IV recruits publicly reported by the Army, the actual figure is 46 percent, the report said.

The Marines have 26 percent Category IV recruits, rather than the 4 percent they thought.

In the Navy, it is 18 percent, rather than 4 percent.

The Air Force, which thought it had not taken any Category IV recruits, actually had 9 percent.

The report does not cite any estimates for previous years, so the overall service totals could be even higher. Sources say the Army, for example, wound up taking in about 40 percent Category IV recruits in previous years, when it thought it was taking far fewer.

All of these new figures are the result of a massive recalculation carried out by the Pentagon, the services and independent testing agencies.

The problem has come to light gradually since 1976, the Pentagon explained in its report. The report said the situation resulted from the technical complexities of trying to modernize entry qualification tests and match up results of the new tests with the old ones, which date back to World War II.

In 1976, the Pentagon put into use a new, standard qualification test. Problems arose when attempts were made to convert results from that test to yield the same balanced assessments as earlier tests, some of which differed from service to service.

When it became apparent that scores seemed inflated in some areas, the Pentagon brought in outside experts to restudy the results. Earlier this year, it alerted Congress that something might be amiss.

Defense officials flatly reject the idea that the test results were tailored to make the all-volunteer force look good initially.

Military officials are extremely sensitive to the prospect that this report could hurt morale within the armed forces or raise further public doubts about whether the military is able to handle its job.

The report says that "millions of low-scoring soldiers, sailors, airmen and Marines served their nation well in the past . . . and they continue to do so today." The Pentagon added, "The first findings from this analysis suggest that most of the low-scoring people are performing adequately."

Category IV recruits score between the 10th and 30th percentile in the qualification test. The report points out that 21 percent of those who served in World War II were in this category, while 9 percent were in an even lower mental grouping that is not accepted today.

The Pentagon said that new tests are being prepared for introduction in October, and are also being introduced to reduce the potential for cheating by recruits or coaching by recruiters.