When Dow Corning Corp. scientists published results of a two-year study of the safety of the company's silicone gel breast implants in four beagles in 1973, their article said the dogs remained in "normal health." In fact, according to the outside laboratory that performed the study under contract, one dog died during the experiment and another developed a large tumor next to an implant.

No hint of the adverse outcomes appeared anywhere in the article, written by two Dow Corning scientists in a journal called Medical Instrumentation, which specializes in studies of the health effects of non-biological materials placed in the body.

The tumor was not malignant but was of a type, called a granuloma, that would be reported later in women receiving breast implants and that experts say can be caused by the immune system having an adverse reaction to a foreign substance.

Growing evidence that some women with silicone gel implants suffer immune system disorders was among the reasons Food and Drug Administration Commissioner David A. Kessler gave last week for requesting a moratorium on implants. Kessler said he had received reports from medical specialists who say they are seeing increasing numbers of implanted women coming down with diseases in which the immune system attacks cells of the women's bodies.

Only in recent years have such immunological disorders been recognized as a possible side effect of the implants. According to Noel R. Rose, chairman of the immunology department at Johns Hopkins University, the tumor seen in the dog nearly 20 years ago could have been an early sign that silicone gel causes immune system disorders.

In the 11 years before the beagle study, more than 50,000 women already had received silicone gel breast implants.

The alteration of the beagle findings is said to be one of the pieces of evidence that led a federal court jury in San Francisco in December to find Dow Corning guilty of fraud in concealing evidence of risks posed by the implants. The jury awarded Marianne Hopkins, who said her breast implants caused severe autoimune disease, a total $7.34 million, of which $6.5 million was for punitive damages.

Although the evidence in that trial has not been released officially, the discrepancies in the reports are evident from a comparison of the published journal article and an unpublished report submitted to Dow by the private laboratory that the manufacturer had engaged. The unpublished version was among a large number of internal documents Dow released last year and which are available to anyone.

The contract lab report, from Food and Drug Research Laboratories of Maspeth, N.Y., says that after the four dogs were implanted with silicone, they were examined six months later and again at 24 months, when the animals were killed and autopsied. At six months, the report says, all four dogs were doing well, most of them having only "slight connective tissue reactions around implant."

The report -- stamped "Proprietary Trade Secret" -- says, however, that one animal, Dog No. 5904F, died after 11 months. The report says no cause of death could be determined. It also says that Dog No. 5903F developed "a large granulomatous mass" next to the implant in its abdomen.

In the journal article, Dow scientists Gordon Robertson and Silas Braley first explain the need for their research: "Because these devices are implanted for long periods of time during which they must remain unaltered and must not cause physiological disturbance, detailed acute and chronic toxicological studies must be made on their materials and on their ultimate design."

The article gives a detailed breakdown of findings on the dogs at six months. Then it says, "Continued weight measurements and daily observations indicated that the beagles remained in normal health. After two years, the normal health of the dogs was confirmed by close examination." Later the article says microscopic examinations after death turned up "no long-term toxic reactions; results were similar to those in Table 2," which presented the relatively benign six-month data.

Robert R. LeVier, technical director of Dow's health care businesses, minimized the omissions, saying they had no bearing on the safety of breast implants in humans. "As a scientist, I don't think this was a well written paper," he said. "If I had written it today, I would be more inclusive."

LeVier said the paper should not be held up against today's standards of scientific publishing and that, in any event, the journal was not widely read by plastic surgeons.

In 1978, however, LeVier cited the study in testimony at an FDA hearing, saying it was evidence that implants were associated with nothing more than mild inflammation.

Rose, the Hopkins immunologist who has testified on behalf of Dow in other product liability cases, said the misrepresentation of the beagle findings should not be excused. "Certainly you would want to know about that," Rose said. "It's significance is that one would want to look for other effects."

While cautioning that there is no proof yet that silicone gel causes immune disorders, he said it was possible that the granulomatous mass in the dog was the result of an immune system reaction. "We know silicone produces a chronic inflammatory response," he said. "The question is whether this causes the immune system to respond in some more specific way" that could lead to autoimmune disorders.

Granulomas are caused when cells of the immune system called macrophages eat particles of foreign matter, as is their job. When doing so, these cells release chemical signals that recruit more macrophages and fiber-making cells, called fibroblasts, to the site. Eventually the result is a fibrous mass of cells called a granuloma. The presence of almost any foreign matter can trigger the response.

The big unanswered question, Rose said, is whether the macrophages carry out another of their roles in the immune system -- processing molecules from the foreign matter and "presenting" it to other immune system cells in a way that leads those cells to make antibodies shaped specifically to attack the foreign molecules.

Some medical researchers suspect that this happens and that the antibodies somehow become shaped to attack the body's own cells. It is well established that such autoimmune attacks are a cause of rheumatoid arthritis, among other diseases.

Rose said he suspects the immune system always contains cells primed to attack the body but they are normally suppressed by other cells of the immune system. "There's a delicate balance that can be upset easily," Rose said. Age, stress, drugs and hormones are among factors that can unleash the self-attacking cells. Another factor, he said, may be the chemical signals released by granuloma cells.

"I wish we knew whether this happens," Rose said. "It would have been nice to start looking at this years ago."