The Pentagon's last hope of flight-testing critical new elements of an antimissile system, before activating the system this autumn, appeared to vanish yesterday with the disclosure that the next flight test has been postponed until late this year, well past the November election.

The Air Force general in charge of the program said the setback will not affect plans to begin operating the system in the next month or two. But the delay leaves the Pentagon pressing ahead with a system that will not have been flight-tested in nearly two years -- and never with the actual interceptor that will be deployed.

The postponement also comes against the backdrop of a wide disparity in estimates about the system's likely effectiveness that has emerged among key Pentagon officials.

The Pentagon's chief weapons evaluator has calculated that the system may be capable of hitting its targets only about 20 percent of the time. The Missile Defense Agency (MDA), which is responsible for developing the system, offers estimates of greater than 80 percent, according to several officials familiar with the classified figures.

The missile defense system, a top Bush administration priority, is designed to send interceptors into space to knock down enemy warheads. The first two interceptors have already been lowered into silos at a newly constructed launch facility at Fort Greely, Alaska, and more are to follow.

Since the last flight test in December 2002, a number of critical hardware and software changes have been incorporated into the system, and officials have counted on the next test to gather critical data about the system's accuracy and reliability.

Democratic lawmakers and other critics of the system accused the administration yesterday of playing politics with the test schedule, seeking to avoid the risk of an embarrassing flop during the presidential campaign.

"This has been a program so fraught with political calculation, rather than strategic and scientific thought, that I would assume there's some political aspect to the delay," said Sen. Jack Reed (D-R.I.), a member of the Armed Services Committee.

But Lt. Gen. Henry A. "Trey" Obering III, MDA's director, attributed the delay solely to technical considerations.

"I have not been asked, influenced or pressured one way or another with respect to putting this system through its paces, through its tests," he said in an interview, disclosing the postponement.

The flight test, already delayed several times, had most recently been slated to occur at the end of September. Obering said he decided to delay it until the end of November after learning last week of a number of modifications to the test interceptor that were not checked out fully in ground tests.

The modifications were made after the interceptor had been moved to a U.S. launch site in the Marshall Islands, Obering said. The interceptor will now be shipped back to a U.S. assembly facility for reexamination.

Another factor contributing to the delay, Obering said, was the inability so far to find the root cause of a software glitch in the flight computer of the interceptor's booster rocket. That glitch led to an earlier flight test delay -- from mid-August to late September -- as the interceptor was removed from its silo to put in a new computer.

Obering said the problems with the test interceptor should have no bearing on the deployment at Fort Greely. Although the interceptors being installed there underwent the same modifications as the test interceptor, they were thoroughly checked at assembly plants, he said. Further, the flight computer glitch seems to involve only test telemetry data, which is not an issue for the Fort Greely interceptors.

Nonetheless, Reed and others said they are baffled at how the administration could be proceeding with deployment while hesitating to test the system.

"If you're not confident enough to take a chance on a test, how can you say that this can engage successfully in a real operational mission?" the senator said in an interview.

Since the start of intercept tests in 1999, the Pentagon has had difficulty keeping up a consistent pace of flights. A couple of early test failures led to months of delay as the causes of the failures were probed.

Several successful intercepts after President Bush was elected gave the test program momentum. But after a failure in December 2002, MDA officials ordered a halt to more intercept tests until a newly designed booster could be completed.

The booster carries a "kill vehicle" into space. That vehicle -- a 120-pound package of sensors, computers and thrusters -- then separates and homes in on an enemy warhead, destroying it with a collision.

Development of the booster has proven more problematic than expected. As a result, several intercept tests that the Pentagon had planned to run last year and this year have been postponed, cancelled or recast as component tests.

The administration thus finds itself proceeding with deployment after only eight intercept tests -- all held before Bush's decision 21 months ago to start fielding a system in 2004. Five were hits, but all occurred under heavily scripted conditions.

All also involved a surrogate booster that flew only half as fast as the booster that will be used in the system. That booster has been launched successfully several times, but it has never flown attached to an actual kill vehicle. The maiden flight of the booster and kill vehicle is slated to be a central feature of the next test.

The absence of realistic flight testing has prompted Thomas Christie, the Pentagon's chief weapons evaluator, to say he will not be able to provide a confident assessment of the system's viability ahead of the planned deployment. In recent weeks, his Operational Test and Evaluation office has argued with MDA over widely different estimates of the system's likely effectiveness.

The differences, Obering said, reflect disagreement over which test data to include in computing the estimates.

Sen. Jack Reed (D-R.I.) said: "This has been a program so fraught with political calculation . . . that I would assume there's some political aspect to the delay."