When China tested its new DF-31 mobile missile last month, CIA and Pentagon experts already knew a great deal about the rocket and its proposed warhead, thanks in part to the same exchanges of nuclear scientists and Chinese launches of U.S.-built satellites that have come under recent sharp congressional criticism.

While a furor has arisen over Chinese spying on the United States, the federal government has been silent about the other side of the coin--what U.S. intelligence agencies have learned during visits by Chinese scientists to U.S. weapons laboratories and trips by U.S. scientists to China's nuclear research facilities.

"We got more out of those Chinese visits than they got," said Richard Kerr, a former deputy director of the CIA who served this year on an intelligence community panel that reviewed allegations of Chinese espionage at America's nuclear labs.

At the beginning of the 1980s, the United States knew little about China's nuclear weapons and missiles. Years of exchanges involving hundreds of scientists from both countries have gradually filled in some of the blanks, according to Kerr and other experts.

One key document illustrates both sides of the intelligence coin. In 1995, a Chinese intelligence officer ostensibly seeking to become a U.S. agent gave the CIA a Chinese military paper that contained classified data on the W-88, America's newest nuclear warhead, as well as publicly available information on five other U.S. nuclear warheads. Known as the "walk-in" document, it was interpreted by Notra Trulock, then chief of intelligence at the Energy Department, as confirmation that China had stolen U.S. nuclear secrets.

But the "walk-in" document also contained valuable intelligence about China's own nuclear arsenal, a fact that has not been previously disclosed. The information included details of the DF-31, China's first solid-fueled, truck-mounted missile, which has an expected range of 5,000 miles, enough to reach Hawaii and Alaska.

It "described the internal configuration and dimensions of their road-mobile ICBM," said one government official with access to the document. "It had lots of precise details" on Beijing's existing weapons and planning for future ones, including the Chinese military's desire to develop a warhead similar to the W-88 by the year 2002, the official said.

"It talked about their own limits on ability to miniaturize" warheads, questioned why the proposed DF-31 warhead "is longer and heavier" than the W-88, and provided the dimensions of key components, the official said. On the whole, he added, "It contained more details [on Chinese weaponry] than on the W-88."

When the CIA later concluded that the person who delivered the document was still working for Beijing, the agency warned government officials that the source of the document "was controlled" and, therefore, suspect. The CIA has been working to corroborate the information, realizing that some may be inaccurate. "We know something about their designs, but not the kind of detail in the document," one CIA official said.

Kerr said the data should be analyzed carefully, adding that the CIA must ask itself, "Why would they tell us that?" When feeding material to an opposition intelligence agency to establish a double agent's credentials--as the Chinese may have been doing in this case--Kerr said his rule while at the CIA was never to provide anything "unique and specific."

The launching of U.S.-built satellites on Chinese missiles also has been advantageous to both sides.

After the failure of a Chinese Long March missile in 1996, China provided an accident report to Loral Space & Communications, the satellite manufacturer. Loral and some of its officials face a grand jury investigation for allegedly responding with sensitive information that may have helped China to improve its rockets. But CIA and Pentagon analysts also studied the accident report, which detailed not only how the rocket functioned but also how China monitors launches.

It was "certainly useful information," said one intelligence official, adding that it "confirmed a lot of what we had suspected."

The Chinese accident report included a complete description of the telemetry, the electronic signals that allow measurement of the performance of a rocket. "That's a level of detail that is important," said one former intelligence official, because knowing all the channels and sensors used by the Chinese "let's us see how good our own monitoring system is."

Previously, U.S. intelligence gathering was based on external listening and imaging devices. But as China and other countries increasingly encrypt their signals and use reduced power transmissions that are harder to intercept during test shots, "monitoring has become a black art," he said.

China's use of U.S.-made satellites also has important benefits for U.S. intelligence agencies, the former official said.

Following allegations that the rocket launches might help Beijing to improve its ballistic missiles, Congress put the satellites on a controlled munitions list. But China's use of U.S.-made equipment provides the United States with a lot of technical data about Chinese communications.

"It's an advantage for eavesdropping," said a former intelligence officer, noting that China is not allowed to open up the U.S.-made satellites to see what is inside them prior to a launch.

Kerr said the scientific exchange programs should continue, although he said that there has been no formal "net assessment" of the intelligence gains and losses. The Energy Department is planning to undertake such an analysis.

Kerr also said the "feedback system" in which U.S. scientists reported on their conversations with Chinese counterparts has been uneven. Often, "our scientists considered it tiresome, inconvenient, and a chore," Kerr said, adding that the data improved only after U.S. delegations began to include scientists with intelligence training. The Energy Department has formalized the foreign contact reporting, a step that some scientists dislike but that guarantees even more information for intelligence analysts.

One retired U.S. scientist recently recalled being engaged by Chinese hosts in the early 1980s in a discussion of neutron weapons, which emit more radiation than normal thermonuclear devices. "They went into great detail about their program," he said, adding that he reported the conversation when he returned home.

Another scientist recalled a Chinese counterpart talking in the mid-1980s about the search for new, smaller warheads. "Everyone was discussing boosting the power of smaller amounts of nuclear material," he said.

A once-classified, 1984 study by the Defense Intelligence Agency of Chinese nuclear weapons systems reflects such information from scientists. The report, released by the National Security Archive, also predicted 15 years ago that Chinese weapons programs would benefit from "overt contact with U.S. scientists" as well as from "covert acquisition of U.S. technology."