Government officials created "the most serious breach of security" in the history of the nation's nuclear-weapon development program by erroneously declassifying a secret document on the hydrogen bomb in 1975 and allowing unrestricted access to it for nearly four years, a Senate hearing was told yesterday.

The testimony came from Theodore B. Taylor, a designer of nuclear weapons from 1949 to 1956 and a Princeton University professor since 1976. He had read the document - "UCRL 4725"-at the request of Sen. John Glenn (D-Ohio), chairman of a subcommittee on nuclear proliferation.

In addition to the erroneous declassification of UCRL 4725, which didn't come to light until two weeks ago, Glenn revealed that the old Atomic Energy Commission also put a second document-said by subcommittee sources to be of equally grave security importance-in the public domain 5 1/2 years ago.

The nature of the second decument, UCRL 5280, is clearly denoted by the title, listed on the page eight of a 16-page Dec. 31, 1973, AEC declassification listing: "Weapons Development During June 1958."

The subcommittee wasn't able to obtain the other 15 pages of the listing and so could not check whether there have been other accidental serious security breaches.

The Department of Energy, which absorbed the AEC's functions, discovered last summer-but did not disclose-the erroneous declassification of UCRL 5280, in a review triggered by a March 1978 hearing of the subcommittee.

A subcommittee source told a reporter that the directive to declassify the second document was "absolutely unambiguous," thus raising the possibility not of mere clerical error, but of "negligence up and down the system starting in Washington."

In the case of UCRL 4725, dated June 1956, the intention was to declassify only a six-page section on propulsion of nuclear rockets. Because the 1973 AEC publication listed the subject of the entire report as propulsion rather than nuclear weapons, the entire report was declassified and twice escaped reviews.

DOE's Duane C. Swell, assistant secretary for defense programs, admitted to Glenn, "The assumption that all weapons documents could be identified by title was incorrect. This is the major mistake that was made."

Taylor testified that his reading of UCRL 4725 left him "thoroughly shocked" and "extremely concerned." He gave three principal reasons:

UCRL 4725 has been "directly accessible to foreign agents and others" in the public library of DOE's Los Alamos Scientific Laboratory in New Mexico since the AEC declassified it on July 30, 1975.

That library is believed to house "the world's most extensive collection of information related to nuclear weapons," Taylor said. It's where he'd start if he wanted to bring a foreign government "up to speed," he said.

"I'd be surprised if any of those countries that have undertaken nuclear-weapon development in the last four years" didn't go to the Los Alamos library and didn't find UCRL 4725, Taylor said in rresponse to questions from Glenn.

In combination with other widely accessible unclassified information, UCRL 4725-even though prepared 23 years ago-"could provide foreign governments with points of departure for development of types of fission [atomic] and fusion [H-bomb] weapons it took this country more than 10 years to develop," Taylor said.

"I am alarmed at the possibility of nations that may now be in the process of acquiring nuclear weapons standing that high on our shoulders as they pursue their quest," he added.

Finally, "The document contains information that would be of value to other nations, such as the Sovent Union, that may have monitored U.S. nuclear tests in the atmosphere before the Limited Test Ban Treaty [of 1963] went into effect," Taylor said.

"This information could be used to revise conclusions they may have drawm concerning the nature of nuclear explosives the United States has tested in the atmosphere," he said.

As summarized by Taylor, UCRL 4725 describes "major progress...on each of a number of nuclear weapon development programs," "contains design and performance data related to several past and planned tests of fission [atomic] and thermonuclear [H-bomb] explosives," and "presents key numbers and tables of numbers representing the results of highly detailed calculations of times, dimensions and densities, and explosive yields of the test explosives."

Taylor said the report also reveals particularly sensitive data: the exact mass of each of various types of fission and fusion materials used in U.S. H-bomb explosives.

The erroneous declassification of UCRL 4725 was discovered at Los Alamos after researcher Dimitri A. Rotow made copies for the American Civil Liberties Union, which is trying to overturn a court order preventing publication of an H-bomb article by Progressive magazine. The washington Post reported last Friday that Rotow had mailed several copies to unidentified persons.

But DOE lawyer Erick J. Fygi testified he was unaware until Wednesday night that copies may be in Glenn's words, "going 'round the Bloomin' world." Rotow was willing to name the recipients, but DOE never asked, "Incredible," said Glenn.