Since 1956, an estimated 30 million pregnant women have taken -- or have had at hand -- a drug mixture for "morning sickness" named Bendectin.

Some gave birth to malformed children. But so did some who didn't take Bendectin to relieve the nausea or vomiting of early pregnancy.

Now, a jury in U.S. District Court is considering whether Bendectin -- sold abroad as Debendox -- can injure the human embryo.

The question is sharply dividing a parade of internationally renowned expert witnesses, including some who are personal friends and admirers of each other's professional skills.

But even scientists who exonerate Bendectin say a woman usually can get through early pregnancy -- the period of simultaneous great peril to the fetus and of nausea -- without the drug. Drugs of any kind "certainly" shouldn't be prescribed during pregnancy unless essential, says Dr. Godfrey P. Oakley Jr., chief of the birth defects branch of the federal Center for Disease Control (CDC) in Atlanta.

Alternatives to Bendectin -- listed in the Journal of the American Medical Association in December -- include eating dry crackers and, particularly in the early morning, drinking only liquids that are either very hot or very cold.

Whatever the outcome of the trial, the evidence may put a cloud over the manufacturer, Merrell-National Laboratories of Cincinnati, because of several charges of improper practices made and documented by scientists.

One charge is that Merrell did no premarketing animal or clinical tests of safety for the unborn, even though women in the sensitive first trimester of pregnacy were the sole sales target. "It is true that no studies were done," Merrell's former medical research director, Raymond C. Pogge, said. Pogge, who invented Bendectin, made a deposition in Casa Grande, Ariz.

The scientists also accuse the company of withholding troubling animal test data from the Food and Drug Administration and of concealing reports on human birth deformities from inquiring physicians.

The trial record also documents that the FDA has never required Merrell to disclose to physicians that adequate safety testing wasn't done. Only recently did the FDA-approved prescribing instructions tell physicians, for the first time, that Bendectin should be used in the first trimester "only when clearly needed."

The FDA approved Bendectin only for the nausea "and" vomiting of pregnancy. Nausea doesn't imperil an embroyo, while vomiting may, the jury was told by Dr. Roger F. Palmer of the University of Miami. Yet Merrell documents show that the company promoted Bendectin heavily just for nausea -- a much bigger market than combined nausea and vomiting.

In 1978, U.S. doctors wrote 3.4 million new and refill Bendectin prescriptions -- nearly one for each pregnancy. The retail value was about $14.4 million. How many pills were taken is uncertain. In some countries, including Canada, Bendectin is sold over the counter.

In the trial, two of the opposed experts were heroes of the 1960s calamity of thalidomide, the sedative-tranquilizer that caused about 8,000 children in 50 countries to be born without arms, legs, or any limbs at all, and often with other severe congenital malformations.

One is Dr. William G. McBride, an obstetrician-gynecologist and specialist in teratology -- the study of birth defects -- in Sydney, Australia. In 1961, he established that thalidomide is an extremely potent teratogen capable of causing deformities in perhaps one of five children whose mothers took it in the first 50 days of pregnancy.

Bendectin is a "very low grade" teratogen, meaning that it causes birth deformities rarely, McBride testified last Wednesday. Most of the deformities, he said, are so-called limb reductions in which part of an arm or leg is missing.

Because victims of Bendectin-caused limb deformities, if any, would be very few in number, large-scale, precise testing would be needed to isolate them from the perhaps 3 percent of all children who are born with significant malformations that have no apparent cause.

McBride appeared for David Mekdeci, 4. He was born with a shortened right forearm. Two fingers are missing from his right hand. The thumb is webbed to the forefinger. He has trouble using the whole arm, because the major pectoral muscle is missing. In addition, his breastbone (sternum) is depressed.

David's mother had taken Bendectin in the sensitive period when embryonic limbs develop -- approximately 24 to 40 days after conception, McBride calculated. In his opinion, he testified, the limb reduction was caused by the drug.

David and his parents, Michael and Elizabeth Mekdeci, are seeking $10 million in damages from Merrell, a division of Richardson-Merrell Inc., of Wilton, Conn.

Later in the trial, which began Jan. 24 and will continue for a few weeks, Merrell will call its top expert witness, Dr. Widukind Lunz, of Hamburg, West Germany, a pediatrician, geneticist, and friend of McBride.

Lenz, too, has thalidomide credentials. Learning that an epidemic of birth deformities was occurring for the first time, he did the medical detective work that made the connection to the drug. Then he led the fight to halt its sale in West Germany, where it was invented and most widely used. It was never sold here, although the American licensee, Richardson-Merrell, distributed 2.5 million "experiental" pills.

Just what Lenz will say isn't known. In the defense opening statement, however, Lawrence E. Walsh, a director of Richardson-Merrell and a former federal judge and U.S. deputy attorney general, hinted that Lenz will support Merrell's contention that David has Poland's Syndrome, an affliction first described in 1841. It has no known cause.

Walsh also said that David's limb reduction could have been caused by constriction of an artery. Terming Bendectin "safe," he pointed out that Mrs. Mekdeci had taken other drugs in pregnancy. Walsh said the medicines "had an effect on the fetus," but McBride disagreed.

Individual experts aside, Bendectin has had on its side the sister federal agencies, the CDC and the FDA. But the validity of their support has been hit by heavy fire from the plaintiffs' lawyers, mainly Arthur N. Cohen and George A. Kokus, both of Miami, and Allen T. Eaton of Washington.

The CDC, as part of a broad birth-defects survey in Atlanta, gathered some date on Bendectin, among other drugs. A special computer run on Bendectin yielded data that seemed to clear it, although no effort had been made to find out if mothers of children with limb reductions who had taken Bendectin had done so in the critical period for limb development.

The Journal of the American Medical Association (JAMA, learning of the Bendectin data in a phone call to Dr. Jose Cordero of the CDC's birth-defects branch, published the results last Dec. 7. There was "no indication that use of (Bendectin) leads to limb abnormalities," JAMA concluded.

The article was intended partly to overcome fears generated by the National Enquirer, which is seen by millions. On Oct. 9, it carried a Bendectin story under this headline: "Experts Reveal . . . Common Drug Causing Deformed Babies." One of the fears was that large numbers of women would seek needless abortions.

But plaintiffs' lawyers obtained a Dec. 20 letter in which Cordero, upset by the JAMA article, told the American Academy of Pediatrics that the Bendectin data were "very preliminary and should not be considered definite result." At this time, he wrote, "a statement that our study exonerates Bendectin as a cause of birth defects is not a completely accurate statement."

The FDA's associate director for new drug evaluation, Dr. Marion J. Finkel, prepared a statement on Bendectin in 1977 and affirmed it last week. "There is no evidence of any risk due to Bendectin," the statement says.

Her sweeping exoneration was based in part on monitoring of birth-defect reports. Only 190, including 88 limb reductions, were known to the FDA. The numbers are so small, in view of the enormous use of Bendectin, as to be virtually meaningless. The company -- and many scientists -- believe, in addition, that limb deformties are the most likely ones to be reported.

But the plaintiffs' laywers say that the FDA and the company know of only a fraction of the true number of deformities, partly because both have virtually excluded the possibility of a cause-effect link, partly because of vast underreporting of adverse drug reactions generally, and partly because Merrell had misled physicians who did report defects in children of Bendectin mothers. They also said that the National Enquirer article, which mentioned Dr. Palmer, brought him 30 additional reports of limb deformities in children of mothers who say they took the drug.

Perhaps most importantly, the lawyers say, an FDA computer printout on drug-associated birth defects shows a proportion of limb reductions for Bendectin greatly exceeding that for any other drug. The FDA sells the printout for $50, but Merrell didn't buy one, the lawyers say.

The FDA's Finkel also relied on the medical literature, the CDC and studies of animals and of groups of women who had taken Bendectin.

Take the Merrell studies -- triggered by the thalidomide disaster -- in rats and in rabbits, which are closer to man.

The rat tests gave no indication that Bendectin is a teratogen. As for the rabbits, they, for lack of accommodations, were left outside in the cold, and all but two died. The litters of the survivors were free of drug-related deformities.

Did a test in two rabbits "have any value in establishing whether . . . Bendectin was teratogenic?" Merrell science executive Dorsey E. Holtkamp was asked last June "No," he admitted. But a Merrell vice president, Robert H. Woodward, cited the happy result to the "entire Merrell sales force" in a Jan. 10, 1963, status report on teratologic testing.

Dr. Holtkamp, aware that a new try had to be made, assigned the task to Dr. Robert E. D. Staples, a teratologist. The results raised a strong enough possibility that Bendectin caused birth deformities -- in rabbits -- that Staples recommended further tests with higher doses.

Merrell should have told the FDA about Staples' data, according to plaintiffs' scientific experts who analyzed his report and the workbooks on which it was based. And retired FDA pharmacologist Frances Da Costa, who cleared Bendectin in 1968, testified that had she known of the report she would have urged Bendectin's removal from the market or addition of a strong warning to the official label.

The University of Miami's Palmer testified that the Staples report obligated Merrell to alert physicians and the public and to seek to curb Bendectin usage.

Wayne State University's Dr. Alan K. Done testified that the report and other evidence persuaded him that Bendectin is a teratogen in humans. Agreeing, Dr. Beverly Paigen of Roswell Park Memorial Institute calculated that "the drug is a suspected teratogen" that, taken in the first 40 to 50 days of pregnancy, could cause "approximately five malformations per thousand women."

In marked contrast to all of this was Merrell's actual conduct. Staples' boss, Holtkamp, in an internal memo, minimized -- inaccurately, Palmer says -- the study findings. Holtkamp, who unlike Staples, isn't a teratologist, said in a sworn deposition that he doesn't believe to this day that thalidomide caused birth defects.

No report was made to the FDA, no alert was given to doctors or the public.And vice president Woodward, who had told Merrell salesmen of the test in two rabbits, did not tell them of Staples' test in 82 rabbits.

Three or possibly five years later -- it isn't clear which -- Holtkamp and others in the company sent to the FDA what plaintiffs' lawyer Kokus described to the jury as a "cleaned up" version of the Staples study. Its conclusion was the Bendectin doesn't cause birth defects in rabbits.

That was a misrepresentation of the Staples data, Drs. Palmer and Done told the jury.

As to the clinical studies relied upon by the FDA and the company, the plaintiffs' witnesses found serious flaws in all of them. One common criticism was that they included disproportionate numbers of women who had taken Bendectin after the sensitive period for limb development had passed. Another criticism was poor records.

But the company will bring out contrary views when it puts on the witness stand some of the eminent scientists who did the studies, including Finland's O. P. Heinonen and Britain's Richard Smithells.