High above the Pacific Ocean last night, the Pentagon's ballistic missile interceptor failed to hit its target in a crucial test of a controversial system being developed to protect the United States against nuclear attacks by "rogue states" such as North Korea and Iran.

In the first exercise to see if the complex elements of the prototype system could work together, the small, sensor-laden "kill vehicle" appears to have missed a dummy warhead launched from Vandenberg Air Force Base in California, 4,300 miles away, said Lt. Col. Rick Lehner, spokesman for the U.S. National Missile Defense Organization, a Defense Department agency.

"There was no intercept for unknown reasons at this point in time," Lehner said.

No catastrophic event, such as an explosion on one of the missiles carrying the kill vehicle and the target warhead, were noted by the extensive array of sensors tracking the test, Lehner said. The first review of data from the intercept site, more than 120 miles above Kwajalein Atoll in the Marshall Islands test range, will take 48 hours, but a full accounting of what went wrong could take several weeks, Lehner said.

The political impact of the test's failure may be felt sooner. The prospect of fielding a national missile defense system has emerged as a key defense and foreign policy issue in the presidential campaign. The contenders for the Republican nomination generally favor proceeding full-speed with the project. The Democratic candidates have been more circumspect, calling for further testing and talks with Russia, which has indicated it would consider construction of such a system a violation of the 1972 Anti-Ballistic Missile Treaty.

The results of yesterday's test, the second so far, will weigh heavily in an upcoming Defense Department study of whether the system can be fielded by 2005, a deadline set by the Clinton administration. After the Pentagon review, which is to be completed in June, President Clinton is expected to decide by late summer whether to start building the first missile shield designed to protect the entire United States.

Under the rules of the Pentagon's development program, the system must score two kills to be judged ready for continued work toward the 2005 deadline. The first test, in October, achieved a successful intercept, and so now the Pentagon must score a kill in the next test, scheduled for April or May, which will be the last exercise before the Pentagon and the White House have to decide on the system's feasibility.

The target for yesterday's test was a dummy warhead lofted into space on a Minuteman II missile fired from California at 9:19 p.m. EST, Lehner said. An interceptor missile was fired from a test facility at Kwajalein Atoll 20 minutes later and flew a satisfactory course guided by a command and control system at the Pacific site.

The missiles appeared headed for a rendezvous in space, and the kill vehicle and the target warhead appear to have separated from their missiles normally, Lehner said.

At that point, the 55-inch kill vehicle, which is equipped with miniature thrusters, its own complex computers and an array of sensors, was supposed to start hunting the target warhead on its own and should have smashed into it, destroying the dummy warhead in the crash.

During the October test, the kill vehicle had trouble with the star map that guides its navigation and then had trouble locating the target, initially heading for a decoy. Pentagon officials said before yesterday's tests that those difficulties had been corrected.

Defense Department officials had high hopes that yesterday's test would provide the first data from a live test, as opposed to a simulation, on whether the communications and control system can take information from a network of ground radars and space-based sensors, feed it to the interceptor and guide it toward the warhead target, a senior defense official said before the launch.

All the important elements of the system, including the missile that will carry the kill vehicle into space and the ground radars that will guide it, are prototypes. The final versions will not be decided upon until after the deployment decisions are due. If Clinton decided to press forward, development and testing would continue even as the first elements of the system were under construction.

"We are on a fast, fast, fast pace," said a senior Pentagon official, reflecting a widespread view among military and civilian officials that development and testing of the missile defense system is proceeding on an unusually compressed schedule even though some of the technology dates to the Reagan administration's efforts in the mid-1980s to develop a much broader missile defense system.

That program, dubbed "Star Wars" by its critics, never went beyond the preliminary development stage. It was supposed to create a shield that would protect the United States from thousands of nuclear-armed missiles simultaneously launched by the Soviet Union from several sites. The current system has much more modest goals.

Projected to cost $12.7 billion through 2005, the system under development would initially deploy 100 missiles, most likely in Alaska, and would be expected to protect against no more than 30 fairly simple missiles, a defense official said. Even if the full shield is deployed in 2011--with more than 200 interceptors at two or more sites--it would only be able to handle "tens, not hundreds" of enemy missiles, the official said.