In the letter, Merck reported that patients taking Vioxx in the largest clinical trial of the drug ever, only 0.5 percent had incurred "cardiovascular events," or heart and circulation problems. That would mean only about 20 patients among the more than 4,000 who took Vioxx during the study.

But in fact, 14.6 percent of the Vioxx patients - or 590 people - had cardiovascular troubles while taking the drug, according to Merck's own report on the study to federal regulators. And 2.5 percent, or 101 people, had serious problems, like heart attacks.

Merck sent the letter to thousands of doctors, including in April 2001 to Dr. Brent Wallace, who had prescribed Vioxx to Robert Ernst. Mr. Ernst, who was 59, died suddenly in May 2001 after taking Vioxx for eight months, and his family is suing Merck, claiming the drug caused his death.

Merck lied to Dr. Wallace, said W. Mark Lanier, a lawyer for the Ernst family.

"Y'all didn't tell him the truth about the safety of Vioxx, did you?" Mr. Lanier asked his witness, Dr. Nancy Santanello, a senior Merck scientist.

The gap between Merck's internal analysis of the study and its letter to physicians may undercut a crucial aspect of the company's defense: that Merck fully disclosed Vioxx's potential heart risks to doctors and patients during the five years the drug was on the market. Merck stopped selling Vioxx last September after a clinical study disclosed heart risks that the company said were its first conclusive evidence of the drug's safety problems.

Unlike much of the other evidence that will be introduced at this trial, the discrepancy between the letter and the study results has not been previously disclosed.

In testimony Tuesday, Dr. Santanello struggled to respond to Mr. Lanier's repeated questions about the letter.

"You know the way this is written is wrong, isn't it?" Mr. Lanier asked her.

"It's not very precise," Dr. Santanello said.

But Dr. Santanello also noted that along with the letter to doctors Merck had sent a New England Journal of Medicine article about the trial. The article contained additional data about the trial and Vioxx's heart risks, Dr. Santanello said.

"The letter does say to read the study for complete information," she said.

The 0.5 percent figure that Merck mentioned in its letter to doctors was actually a number that referred to the risk of heart attacks faced by patients taking Vioxx in the trial, not the total risk of cardiovascular problems generally.

During the study, more than 4,000 patients took Vioxx for an average of nine months, while about the same number took naproxen, an older painkiller sold under the brand name Aleve. The results indicated naproxen was much safer for heart problems.

Compared with 0.5 percent of Vioxx patients who had heart attacks, only 0.1 percent of the naproxen users - or four patients - had heart attacks. Fewer than 50 naproxen users suffered other serious cardiovascular problems, according to Merck's analysis of the study.

After the day's testimony, a lawyer for Merck said that the evidence presented Tuesday had been available for years.

"This is no new information," said Jonathan B. Skidmore, a partner at Fulbright & Jaworski, which represents Merck. Mr. Lanier has "taken data which the whole scientific community has had, and he has played with it in front of a jury, and in doing so he's clouded what's important," Mr. Skidmore said.

A jury of seven men and five women is hearing the case, Ernst v. Merck, in Brazoria County Courthouse in Angleton, a town of 18,000 about 45 miles south of Houston. The case has attracted national attention as the first of thousands of Vioxx-related lawsuits to reach trial.

About 20 million people took Vioxx before Merck stopped selling the drug last fall, and some epidemiologists have estimated that the drug may have been linked to tens of thousands of heart attacks or strokes.

Merck denies that Vioxx was responsible for Mr. Ernst's death, which according to the autopsy report was the result of arrhythmia, or an irregular heartbeat. Vioxx has never been proved to cause arrhythmia, the company says.

The discrepancy between Merck's letter and its analysis of the study's data was the crucial point in Mr. Lanier's day of questioning Dr. Santanello, who first took the stand Monday afternoon.

Mr. Lanier upbraided Dr. Santanello dozens of times for not answering his questions, saying in one typical instance, "I asked you to quit dodging and focus on my questions please."

Gerry Lowry, a lawyer for Merck, complained repeatedly to Judge Ben Hardin that Mr. Lanier was making speeches to the jury instead of simply asking questions. Near the end of the day, the judge twice asked Mr. Lanier to give Dr. Santanello time to answer.

Mr. Lanier also played Merck's television advertisements for Vioxx, which did not mention the drug's possible heart risks. Later, he showed the jury e-mail messages between two top Merck executives in which they complained about the Food and Drug Administration's proposed label for Vioxx. In one message, Dr. Edward M. Scolnick, Merck's top scientist, referred to the F.D.A. officials as "bastards."

The day was not entirely without humor. At one point, Dr. Santanello joked that she might be having heart palpitations under Mr. Lanier's questions.

"Down here we call them runs," Mr. Lanier said with a smile. "If you get them, let us know and we'll take a break."