(FORTUNE Magazine) – It all looks so good in PowerPoint. By the end of 2005, Merck will be on track to unveil four remarkable new medicines, remedies to arm doctors in the fight against some of the world's most intractable ailments: rotavirus, cervical cancer, shingles, and diabetes. Each of the new drugs will have taken decades of plodding lab work. Each will be a dazzling display of Merck's scientific acumen. Millions will be cured.

And yet, on the day last December that CEO Ray Gilmartin invited analysts to the company's secluded New Jersey campus to reveal Merck's achievements, the slide show that would have stirred fits of envy in the research community drew exasperated groans from Wall Street. These drugs won't be the big moneymakers Merck needs. Instead of plaudits, cries for the company's board to fire Gilmartin, 62--well before his planned retirement in 2006--grew louder. Two days after Gilmartin's presentation, Deutsche Bank analyst Barbara Ryan wrote, "We believe that resetting the management clock is preferable to letting it run out, and that a near-term major acquisition/merger may be required to restore a pulse." Bear Stearns analyst Joseph Riccardo noted to his clients that Merck's "current problems are a direct result of what management did not do over the last five to ten years. Management is now playing a huge game of catch-up."

If Ray Gilmartin were an NFL coach, say, or a TV weatherman, he'd already be out of a job. Gilmartin has failed the essential test for a CEO of the world's third-largest pharmaceutical company, despite nearly a decade running the place: He hasn't been good at creating new drugs. Most of the 19 that Merck introduced during Gilmartin's reign were a result of research that began under former CEO Roy Vagelos. Drug companies frequently endure hard-luck cycles, to be sure, but Merck scientists discontinued development of an unprecedented four potential blockbuster compounds last year.

Meanwhile, Gilmartin began to irk Wall Street by dialing down several of the company's far-too-optimistic quarterly earnings estimates. Since the beginning of 2001, Merck shares have lost 46% of their value, compared with the Amex pharmaceutical stock index's drop of 23%. The year 2003 was one of the worst in the company's 117-year history. Sales grew just 5%; net income fell for the second consecutive year, dropping 4.5%, to $6.8 billion. Last November, Gilmartin laid off 4,400 employees, 7% of the total, in the company's largest workforce reduction ever. What's more, Merck's top seller, cholesterol pill Zocor, loses U.S. patent protection in mid-2006. Zocor represents about 18%, or $5 billion, of Merck's annual $22.5 billion in sales. Almost all of that is expected to disappear the first year. None of Merck's coming drugs are likely to replace those revenues.

Wall Street has an easy remedy for Merck's woes: Merck should identify a large competitor with a blockbuster drug and acquire that company. Now. Gilmartin has his own answer: reenergize Merck's labs and usher in a new era of Merck-made bestsellers. "If you did a checklist of what it takes to succeed in this business, 'making breakthrough medicines' is at the top," says Gilmartin. Problem is, neither of those approaches is likely to work.

The fundamental problem with Merck is that Gilmartin has allowed it to be hamstrung by its history. Merck has always been regarded as having the best scientists and research operations in the industry. But Gilmartin has fostered a culture of scientific arrogance that is harming the business. Some of the brashness is justified. Merck consistently tops the list of companies with the highest number of new chemical-compound patents, and it does so while investing a lower percentage of revenues into research than its competitors: 13%, vs. an industry average of 17%. Between 1996 and 2001, Merck patented 1,933 new chemical entities, roughly 500 more than its nearest competitor, Pharmacia (now part of Pfizer).

But only a tiny percentage of patents results in drugs, and even fewer result in blockbusters. Consider the story behind two of the four drugs that Merck wound up axing last year. In November, Gilmartin halted development of what would have been Merck's first entrant into the lucrative antidepressant market. Days later he called off another late-stage program, this one for a diabetes treatment that was found to cause cancerous tumors in mice. In both cases, prior research had suggested the compounds were not viable. But the company went ahead anyway on the sheer will of Merck's former head of research, Ed Scolnick, who retired last year. By all accounts Gilmartin, whose background is in electrical engineering, not chemistry, deferred to Scolnick on all matters of science. That gave Scolnick and his successor, Peter Kim, an amazing amount of latitude. In the case of Merck's "Substance P" depression candidate, says Scott Reines, former head of neuroscience development at Merck, "by the time we went to Phase III, no one thought it would be a winner. But we felt, 'We're Merck,' and if this compound was active, we were going to find it."

Another example of Merck's arrogance is its refusal to pursue "me-too" drugs--second-or third-to-market copies of drugs that another company has already introduced. GSK's Levitra and Lilly's Cialis, which now compete against Viagra in the erectile-dysfunction market, are good examples. They won't win any awards for ingenuity, but me-toos can be very lucrative. Merck would do well to add at least some of them to its portfolio. Gilmartin pooh-poohs the idea. "We go for novel medicines that are true advances," he says. Adds research head Peter Kim: "You can't attract and retain good scientists if you're telling them to go make a copy of someone else's drug."

Merck has reason to worry about retaining top scientists. Kim, who joined Merck in 2000 after a stellar academic research career and took over as research head last year, has confronted a mass exodus of talent. It began with Roger Perlmutter, a former executive vice president of research, who left the company to head Amgen's R&D efforts in early 2001. Four of Amgen's 11 executive officers, in fact, are former Merck scientists who have exited in recent years. Recently Merck's top AIDS researcher, Emilio Emini, announced he would leave the company to become vice president of vaccine development at the International AIDS Vaccine Initiative. Notes one former Merck research executive who is now at Amgen: "One strength Merck always had was its balance of people who were academics and people who knew drug development from end to end. I fear the company is losing that."

Merck is more efficient with research spending than its competitors, but some argue that is no longer a positive and maintain the company must invest more money to compete effectively. Compare it with Eli Lilly. In 1997--three years before its blockbuster Prozac lost patent protections--Lilly began investing a hefty 20% of revenues (compared with Merck's 13%) in research each year. It now has the strongest pipeline in the industry: three product introductions in 2003 and four expected in 2004.

Kim insists that the percentage of revenues paradigm is the wrong way to view the company's research spending. "I'm not given a number. If we need to spend a large amount of money, we spend it," he says. "What we spend is $3 billion, and it's not as if I'm banging on Ray's door asking for more." Gilmartin concurs, insisting that Merck is spending wisely. "It's not the ratio that matters so much but growing it with sales," he says. "It's the absolute dollars that you spend and how productive you are."

What about a merger as a cure for what ails Merck? Gilmartin has been a vocal detractor of what he calls "disruptive" mega-mergers such as the ones in the late '90s that beefed up Pfizer and created GlaxoSmithKline. "I've been clear about this," he says. "If you subject Merck to a large-scale merger, does it really contribute to long-term growth and long-term pipeline? No. It's good for short-term cost cutting." He's probably right. At Pfizer, for instance, mergers with Warner Lambert and Pharmacia have given the company blockbuster drugs Lipitor and Celebrex and allowed it to eke out some efficiencies--and may have contributed to a modest rise in the stock price--but they haven't made Pfizer's own labs any more effective at discovering new drugs.

As Merck's troubles have grown, rumors that Merck's board will force Gilmartin into early retirement are swirling around Wall Street and even within Merck. But that isn't likely. Gilmartin is determined to ride out his last two years and help the board choose a successor. He says he believes he can help the company. To his credit, he has made some smart moves. He is forging more joint venture alliances: Merck licensed chemical compounds from 47 small biotech firms in 2003, up from ten such deals in 1999. And he instituted a long-overdue distribution program that prevents wholesalers from stocking up on Merck products before a price increase. More to the point, Merck's board unanimously supports Gilmartin. "Ray is a very clear-thinking, straight person. He really understands the business and the industry and the needs that Merck has," says William Bowen, president of the Andrew Mellon Foundation and a Merck board member since 1986. "The board has never questioned that."

What the board should be questioning is whether Merck is doing enough--however long Gilmartin stays--to address its main problem: its culture of scientific arrogance. The world has changed, and Merck hasn't changed with it. The time to start is now.