For the second time in as many months, the Bush administration's new missile defense system failed to complete a key test yesterday, automatically shutting down a few seconds before an interceptor missile was to launch toward a mock enemy warhead.

Defense officials said initial data pointed to a malfunction in support equipment at the launch site in the Marshall Islands in the central Pacific Ocean, rather than a failure on the missile itself.

The failure could fuel debate in Congress over the costs and benefits of the multibillion-dollar system, which some lawmakers and scientists say the Pentagon is rushing to deploy before it is proved effective. President Bush first pledged to build the system in the 2000 campaign.

Defense Secretary Donald H. Rumsfeld has so far refrained from putting the system on alert -- a move that had been expected last fall when the first six interceptors were installed at a launch facility near Fairbanks, Alaska. The system, intended to protect the United States against a long-range missile attack, envisions the creation of a multilayered network of land- and sea-based interceptors and space-based weapons.

Yesterday's miscarried test was created to duplicate a Dec. 15 trial that also failed. Both tests were to have marked the first flights of the advanced interceptor missile. Earlier tests had used a slower, less sophisticated interceptor.

The latest tests called for the interceptor's booster rocket to shoot into space and release a "kill vehicle" that closes in on a mock enemy warhead and destroys it in a high-speed collision.

In both recent tests, the mock enemy warhead was launched successfully from Kodiak, Alaska, but the interceptor failed to get off the ground. Defense officials said each test appears to have failed for a different reason.

In the Dec. 15 test, a software failure on the interceptor itself led to a communications breakdown between the flight computer and the component that steers the missile, which caused the test to automatically abort 23 seconds before launch.

Yesterday, the countdown went well beyond that stage, suggesting that problem, which officials had described as a very minor software glitch, had been solved. Instead, initial reports suggest yesterday's trouble lay with the "command launch equipment" on the ground, which can automatically halt a launch if a problem is detected.

"The problem last night had nothing to do with the interceptor. It could be the silo," said Rick Lehner, spokesman for the Missile Defense Agency here. The 85-foot-deep concrete-and-steel silo has sensors that measure temperature, fumes and other safety criteria, Lehner said.

The target missile carrying a mock warhead launched from Kodiak at 1:22 a.m. Eastern time yesterday and flew toward the southwest. The interceptor was scheduled for takeoff about 15 minutes later from a test site on Meck Island in the Marshall Islands. When the intercept failed, the mock warhead crashed north of Wake Island, about 4,000 miles from Kodiak.

Investigators are sifting through all the launch data to determine the exact cause of the failure, Lehner said.

Since 1999, the Pentagon has conducted 10 tests of the missile defense system, five of which have resulted in hits. But only the last two tests have used the actual interceptor designed for real-world missions; earlier tests employed surrogates.

David Wright, co-director in the global security program at the Union of Concerned Scientists, said in a statement that yesterday's failure showed "the program is being pushed ahead for political reasons regardless of its capability."

Defense officials stressed that the last two tests were far more realistic, and thus more technologically challenging than the previous ones.

"This was a much more robust and difficult test," said Chris Taylor, another spokesman for the Missile Defense Agency. He said the average cost of the tests is $85 million, although in this case it could be less because the interceptor was preserved.

Original URL