Researchers at the National Institutes of Health violated federal rules by engaging in lucrative collaborations with pharmaceutical and biotechnology companies and not reporting those arrangements to ethics officials as required, according to documents released yesterday as part of an escalating congressional investigation into conflicts of interest at the agency.

The House oversight subcommittee had already identified several instances in which scientists engaged in outside activities that posed at least the appearance of a conflict of interest. But in those cases the arrangements had been approved by top legal and ethics officials. Now, NIH officials said, disciplinary actions may be needed.

Testimony yesterday also provided evidence that Lance A. Liotta, a researcher at the National Cancer Institute, continued to receive thousands of dollars in compensation from one such business arrangement through May, despite his testimony under oath last month that he had suspended the collaboration months before.

The subcommittee said Liotta and others used their government computer systems to exchange e-mails relating to their private consultancies, supporting some lawmakers' contention that some government scientists have been illegally using federal resources for personal gain.

It remains to be seen how many of the subcommittee's allegations will stand up to closer scrutiny as bona fide breaches of ethics rules. Many details of the cases were still missing as of yesterday, and key individuals could not be reached to comment after the hard-hitting six-hour hearing came to a close.

But having learned of some of the new findings late last week, NIH Director Elias A. Zerhouni came before the subcommittee yesterday with proposed revisions to NIH ethics rules more severe than those he had recommended a month earlier.

"I have reached the conclusion that drastic changes are needed," he said.

The new allegations emerged as part of the panel's expanding investigation into government employees' consulting deals with private companies. Although House members began by focusing on NIH, where top scientists' spare time is in great demand by drug companies wishing to capitalize on their expertise, they widened their probe last week to include 15 other federal agencies. In letters sent to agency heads, the subcommittee chairman, James C. Greenwood (R-Pa.), and House Energy and Commerce Committee Chairman Joe Barton (R-Tex.) asked that records of all such collaborations be provided to them by July 2.

Yesterday's surprise disclosure that many NIH scientists may be engaging in outside deals without the required agency reviews and approvals grew from inquiries Greenwood made to 20 pharmaceutical companies.

Given the lack of a centralized NIH database of all agency scientists' outside collaborations, Greenwood went directly to the companies, asking them to reveal all the arrangements they had.

Of the 264 arrangements the companies reported, Greenwood said, "about 100" were apparently unknown to NIH officials.

That sampling has Congress wondering, "What else is out there?," said Rep. John D. Dingell (D-Mich.).

In one example detailed by Greenwood, drug giant Pfizer Inc. reported that Trey Sunderland, a researcher at the National Institute of Mental Health, was paid $517,000 in fees, honoraria and expense reimbursements related to consulting arrangements with the company over the past five years. Greenwood said the information was not on Sunderland's financial disclosure reports as required by federal ethics rules.

An NIH official said Sunderland was traveling abroad and could not be contacted.

In another highlighted arrangement, Alan Moshell of the National Institute of Arthritis and Musculoskeletal and Skin Diseases was retained as an expert witness in several private product-liability lawsuits involving the drug Accutane at a rate of $600 per hour -- and did so without required agency permission -- Greenwood said. Those arrangements were described by Health and Human Services general counsel Alex M. Azar II as particularly worrisome as Moshell allegedly testified in those trials to the inadequacy of the government's own warning label on the drug.

Moshell did not respond to calls and an e-mail late yesterday.

The subcommittee also provided new details regarding an alleged conflict of interest outlined in a May hearing, in which cancer researcher Liotta and an FDA scientist became paid consultants for a California biotechnology company that is in competition with a Bethesda company with which the two scientists and the Cancer Institute were already collaborating.

Liotta testified last month that by March of this year he had suspended the California arrangement, pending a fresh ethics review by agency officials who initially approved the deal but later expressed regret at having done so.

Yesterday, Greenwood flashed on a giant screen copies of several canceled checks from the company -- Biospect Inc. of South San Francisco, recently renamed Predicant Biosciences -- made out to Liotta. The latest check, for $3,125, was dated May 1.

Greenwood also showed evidence supplied by the company that it had paid Liotta a total of $70,000, significantly more than the approximately $49,000 that Liotta reported to ethics officials.

A cancer institute spokesman said yesterday that Liotta had an appointment and would not be able to respond to media queries.

Zerhouni has already imposed new tiers of ethics review for all proposed outside consulting arrangements by NIH employees and greater public disclosure of approved arrangements. Yesterday, he proposed additional restrictions, including some that could be accomplished internally and others that may require new legislation.

Among them: a ban on ownership of drug company or biotech stocks by some key employees, and restricted stock ownership for all other employees; no membership on corporate boards; creation of a centralized registry of all outside arrangements and a public list of the awards that employees may receive; and prohibition of all paid consulting or speaking engagements at institutions that receive NIH funding.

"I have reached the conclusion that drastic changes are needed," said NIH Director Elias A. Zerhouni, who recommended more severe revisions to ethics rules.James C. Greenwood, chairman of the House oversight panel, made inquiries to 20 pharmaceutical companies.