Like all companies, Forest pays each of its board members a small "honorarium" to participate in meetings, covers their travel expenses, and, more important, puts up money to pay for their research as well. Green says the company follows the American Psychiatric Association guidelines -- paying doctors about $500 to speak at a local symposium, $1,500 if it is national. Such payments from drug companies have caused controversy, most notably in the case of Dr. Martin Keller, a Brown University professor of psychiatry and member of Forest's advisory board. The Boston Globe reported that in 1998, out of $842,000 in income, he made nearly $556,000 from consulting to drug companies -- none of which he disclosed when he published research about their drugs.

Dr. Steven Roose, a researcher at the New York State Psychiatric Institute, is another member of Forest's advisory board. "It's nice to get an honorarium of a couple thousand dollars, or to go give a talk, especially if it is in a nice place -- don't let anybody tell you that the money means nothing," he told me, sitting underneath a Jasper Johns poster in his small office. "But the honorarium is gone in two pair of Prada shoes and a bag. What really matters is the research -- that's where the deep issues are."

An academic medical researcher's career often depends on being able to conduct important large-scale studies, and funding from drug companies can be a big help. Some fret that drug-company funding inevitably alters the course of research. "Sometimes, when research funds are limited, you will go along with drug-company studies that may be very good, but they aren't what you would do otherwise," says Roose. "There really is a pharmaceutical-academic complex." But Forest impressed its experts by going the extra mile. The company agreed to fund a project that its researchers originally proposed -- a landmark study of depression in people over 75, known as the "old-old study." Forest liked the idea because Celexa may have real advantages with very old patients, because of its limited interactions with other drugs. The company is even paying for expensive brain-imaging tests and patient interviews.

"The real way to influence researchers is to let them do a study like this -- that is the quid pro quo, the benefit you get from helping them," Roose says. Of course, every researcher knows that a single lapse in integrity could ruin his credentials, he says. Investigating the effects of Prozac on the heart, Roose happened to discover that it is much less effective than earlier drugs in treating severe melancholia. "I didn't expect it. It just hit me in the ass. Was I a popular person with Lilly? No. So maybe I wasn't invited to speak for a while -- who cares?" He told Forest that Celexa isn't so effective against severe depression, either.

Drug makers, of course, don't tend to publicize views like that. In a Forest press release, its advisory-board members sound like critics in a movie ad. "Celexa will be a friendly drug for the physician to use, and certainly a friendly drug for the patient," raves Dr. George Alexopoulos of the Cornell Medical School, another leader of the old-old study.

"Did he say that? I would be shocked," Roose says. "A thoughtful opinion of a drug can't be reduced to a quip."

To warm up the market before Celexa's debut, Forest began with an advance team of nine "medical-science liaisons" hired in August 1998. The liaisons, many with degrees in pharmacy, set out "to form relationships" with the top 200 or so "thought leaders" in psychiatry, another tier of trendsetting researchers and clinicians. Each medical liaison worked on a list of 22 doctors at institutions like Columbia-Presbyterian Hospital in New York and Massachusetts General in Boston.

In the summer of 1998, still months before the official launch, Forest turned to a tried-and-true technique to kick off the Celexa campaign: paying doctors, essentially, to prescribe it. Before the drug's approval, the "psychiatric thought leaders" on the advisory board had each put about twenty patients on the drug and answered questions about dosage and results. Next, the company added a nationwide test called Early Access to Symptom Evaluation, or ease. About 600 doctors tested Celexa on about five patients apiece. Doctors received $250 for their help, and another $50 a patient for reporting results -- a total of about $500. (Participating patients start the drug free of charge.) The tests helped gather extra information about the drug's effects. Just as important, the tests acquaint doctors with the pill and start some patients on it.

Then Forest's reps sprang into action. The company picked about 22,000 of the biggest prescribers of antidepressants and hit them with a soft sell. "We went out before the launch period, with just the FDA-approved package insert, no Madison Avenue glossy promotional material," says John MacPhee, the other Forest marketing director in charge of Celexa. "We said, 'We're here because you are so expert in treating depression. Let's talk about all the pluses and minuses. Let's talk about the beagle dogs. It's new. You can get an early experience, if you want to get a jump on a lot of physicians.' "

Some doctors were wary -- medical ethicists caution against trying new drugs without demonstrable advantages, since less is known about their risks. But in their role-playing sessions, Celexa's reps had rehearsed a ready answer to that objection: They told doctors that millions of Europeans had already taken it. And some doctors appreciate the chance to offer patients the very latest thing. Ronald Ruden, the Upper East Side doctor, says he jumps at such opportunities. "I am an early adopter," he explains. "I always want to try the newest drug to see how it works." By the day of Celexa's U.S. launch, nearly 15,000 U.S. doctors had prescribed Celexa, too.

As Celexa gained momentum, its rivals struck back. Almost immediately, drug reps from rival companies began what is known in the business as counter-detailing. I even experienced some counter-detailing myself. A Paxil rep, speaking on condition of anonymity, cut into Celexa right away: "It's not a good drug in the first place, and there are more side effects," she said, adding that she warns doctors that Celexa could cause kidney problems in overdose, alluding to the beagles. (There was a report of lethal overdoses in Sweden, but it was later debunked; a Celexa overdose isn't dangerous in humans.) A Zoloft rep took another angle: "You know, there have been some data saying Celexa is no better than a placebo." (This is true but not necessarily relevant, since placebos are often very powerful in relieving depression.) She called back to call Celexa a "me-too" drug (so are Zoloft and Paxil), quote from the Medical Letter, and describe the "tremendous pressure" Forest put on its salespeople. "They wouldn't make a bonus if they didn't sell Celexa -- it was do or die for them," she said.