It might have been the last straw for Lt. Col. Craig MacNab, except that the Army does not allow him to think that way.

A large artilleryman who spends his days folded uncomfortably into a tiny cubicle in the Army's media relations shop at the Pentagon, MacNab had fended off attacks on the Bradley Fighting Vehicle for weeks.

Now, a reporter was on the telephone with a new verse in the long litany of criticisms against the Army's cherished and costly troop carrier.

"Why does the Bradley need a roof?" the reporter asked.

MacNab patiently explained why a roof is not a gold-plated frill, entering the latest skirmish in a frustrating war against a persistent but frequently unseen enemy.

Critics of the Defense Department, who usually call themselves "military reformers," have mastered and improved upon many of the Pentagon's public relations techniques -- selective leaks, dark insinuations about classified data and a refusal to be held accountable subsequently. The hapless Army bureaucracy does not respond adeptly to such guerrilla attacks.

Sometimes, history proves the critics correct. For years, the Army defended its division artillery antiaircraft gun (Divad), only to admit defeat when Defense Secretary Caspar W. Weinberger this year decided that Divad would not do the job.

Sometimes, weapon systems deflect the critics' withering fire and prove themselves in the field, prompting the critics to drop the issue and move on to another weapon.

The Air Force Airborne Warning and Control System (AWACS) airplane, the Army's M1 tank and the Navy's Aegis radar were attacked mercilessly as "gold-plated lemons" that would never work. Now, they are embraced by airmen, troops and sailors. While experts still question whether the weapons are worth their cost, the most dire predictions are forgotten.

Now it appears to be the Bradley's turn. The Army hopes to buy almost 7,000 of them, to carry troops and shoot missiles and cannon, at a cost of more than $13 billion.

Plagued by delays, fantastic cost growth, exaggerated claims and the Army's refusal to conduct realistic tests until forced to do so, the Bradley long ago became a natural target for the critics.

They could raise fundamental questions about its mission that had no easy answers: Is it worth 10 times the cost of its predecessor? Is it sensible to combine combat and transport functions, forcing soldiers to share space with volatile ammunition?

On some issues, those questions become lost as zealous critics, aided by reporters less likely to challenge a "reformer's" credibility than that of a general, help mingle truth with fiction.

Thus, the Divad never shot out a latrine fan in early testing, as was reported widely, including by The Washington Post. But the image of a gun that could not tell a helicopter from an outhouse became a part of lore never shaken by Divad.

Similarly, critics lambasted the Army for hosing test dummies inside the Bradley during early tests, making them less likely to catch fire and putting a sunny spin on test results. MacNab confirmed that it happened.

But the story turned out to be not quite so simple. MacNab said those tests were aimed at discovering whether aluminum vapors from the hull could poison soldiers when the vehicle was hit, another count in the critics' indictment. If the dummies had burned, MacNab said, the sensitive toxicity monitors would have burned, too.

The Army has completed its most rigorous "vulnerability" test of the Bradley, a set of 10 live-fire shots against fully loaded and fueled vehicles designed to show how one will behave on the battlefield. Now, the battle has begun on how to interpret the tests.

The results are classified, so as not to tell the Soviets where to aim, the Army says. Therefore, congressional aides, Army officers, representatives of the Bradley's maker in California and others have been whispering their versions of the results for weeks.

Last week, the Project on Military Procurement -- a foundation-funded clearinghouse for critics inside and outside the military -- made available to reporters excerpts from an internal analysis of the results by an Air Force colonel critical of Army testing methods.

The excerpts gave an impression that the Army again had cushioned the vehicle against tough tests.

At pains to support the tests, Army officials had said they shot at the Bradley's ammunition stores four times during the 10 tests, in one case destroying the vehicle.

Yet one part of the critical report said: "Each of the 10 live-fire shots was aimed so as to avoid intentionally striking the explosive elements of internally stowed ammunition."

Which side was shading the truth? Apparently, neither. Col. James G. Burton, who wrote the critical analysis and initially had pressed the Army to conduct realistic tests, was unavailable for comment, apparently on orders of his boss.

But another critic outside the military, who asked not to be named, explained the carefully worded excerpt this way: One test shot hit a missile but not its warhead (the "explosive element"); two other shots hit stored .25-mm ammunition inside the vehicle but had not been aimed there (not "intentional"), and a fourth struck ammunition loaded on the vehicle's side (not "internally stowed").

Angered by the criticism, Army officers decided to come out swinging for the Bradley and got into trouble for that, too. A three-star general was trotted out to describe test results to reporters and show videotapes of the 10 shots.

The next day, top Defense Department officials rebuked the Army for conducting a news conference without permission, according to sources. And the chairman and ranking Republican of the House Armed Services Committee, piqued about seeing the videotapes on network television before Congress had viewed the results, wrote to the defense secretary to say they were "dismayed" and "deeply disturbed" by the Army's actions.

"This ain't no way to run a railroad!" Rep. William L. Dickinson (R-Ala.) fumed to Weinberger.

Meanwhile, the critics abandoned some accusations and raised new charges. The tests apparently showed that the Bradley was not excessively vulnerable to fuel fires, as had been suggested. Instead of welcoming that as good news, the critics said the effective fire-extinguishing system might be more dangerous than fuel fires, because its gases would force soldiers out of the Bradley for air.

"Nonsense," MacNab said with a sigh, adding that soldiers could easily survive the extinguishers. But he spoke without much apparent hope of ending the controversy.

"Anybody can make up these things," he said. "I kill 400 snakes, and then somebody calls up and says, 'I hear riding in the Bradley for three years causes skin warts.' I have to deal with that, or it will sure show up in an editorial somewhere.