(FORTUNE Magazine) – One day in December 1999--near the top of what we now know as the Internet bubble--a couple of East Coast business types with a few hours to kill were cruising Silicon Valley in a chauffeured car. "Yeah, I know 'em all," announced their driver, as they meandered through the streets of Woodside, an exclusive enclave near Palo Alto. Neil Young, for instance--he lives up in the mountains, the driver said. Why, he boasted, his customers even included Mary Meeker. Would they like to see her new California digs?

Within moments the sedan had pulled up to a wooded, four-acre estate, marked with white pillars and protected by a wrought-iron gate. As his wide-eyed passengers watched, the driver stepped out and punched a security code into the keypad. The gate swung open, the driver pulled the car onto the property, and the gleeful pair scampered about the grounds, checking out the backyard pool and peeking into ground-floor windows like teenaged boys granted admission to Britney Spears' bedroom.

It was a sign of those strange times that a Wall Street analyst could excite such fascination. Then again, Morgan Stanley's Mary Meeker was no ordinary analyst. Anointed by Barron's as "Queen of the Net," lovingly profiled by The New Yorker, equated with Alan Greenspan and Warren Buffett as a market mover by the Wall Street Journal, Meeker was the unquestioned diva of the Internet Age. Tech companies begged her to cover them. Morgan Stanley paid her an eye-popping $15 million in 1999. Ordinary investors hounded her for autographs. During the dot-com craze, Mary Meeker was by far the most important voice for the Internet--and the notion that companies without earnings could transform the world and climb to the moon.

That was then. Today Meeker, 41, has become something else entirely: the single most powerful symbol of how Wall Street can lead investors astray. For the past year, as Internet stocks have crumbled and entire companies have vaporized, Meeker has maintained the same upbeat ratings on her companies that characterized her research reports in the glory days. For instance, of the 15 stocks Meeker currently covers, she has a strong buy or an outperform rating on all but two. Among the stocks she has never downgraded are Priceline, Amazon, Yahoo, and FreeMarkets--all of which have declined between 85% and 97% from their peak. For this she has been duly pummeled in the press, accused of cheerleading for Morgan Stanley's investment banking clients.

But Meeker's refusal to downgrade her stocks is only a small piece of a bigger story. This larger story is about how a smart, hard-working analyst became a big part of the world she covered. Meeker came to see herself not merely as an analyst but as a player--a power broker, a dealmaker, a force to be reckoned with. She was a true Internet insider--and other Internet insiders, most of whom were her friends, shared this exalted view of her. "I don't think of Mary as an analyst," says venture capitalist John Doerr. "I think of her as a service provider for investors, entrepreneurs, and management teams." As a result Meeker did things that utterly compromised her as a stock picker.

In responding now to criticism that she let investors down, Meeker refuses to admit--or even see--how compromised she is. She defends herself in part by saying she feels protective toward the phenomenon she helped launch--and especially toward the dozens of companies she helped Morgan Stanley take public: "There is something compelling about...playing an important role in something that will never happen again....I feel a--'stewardship' is a strong word--but I feel a keen sense of responsibility." She adds, "If you take a company public and you are really aggressive on the downside, it can be devastating." Of course, if you're not aggressive on the downside, it can be devastating for investors. But that was never a Meeker priority.

Though she was Queen of the Bubble, Meeker hardly reigned alone. The stories that follow explore the many ways investment banks now abuse the trust of their core customers--investors trying to build capital and companies trying to raise it. "Betrayal on Wall Street" explains how the IPO market became a racket in which banks consistently shortchange the startup clients they're advising in order to create quick profits for institutional traders. In "Hear No Risk, See No Risk, Speak No Risk" we learn how a chorus of telecom analysts overlooked warning signs at a company called Winstar and kept croaking "buy" right up until this broadband wannabe choked on its own debt. And in his confessional essay "Diary of a Financial Pornographer," Nelson Schwartz takes a hit for all of us at FORTUNE and in the media who got caught up in the bubble madness. But first, there's Mary Meeker.

"It is a cool thing to take a great company public," Mary Meeker is saying. "You have to go up to the plate and swing for the fences." We're sitting in a corner conference room at Morgan Stanley's Times Square headquarters, a half-hour into the first of several long interviews. I've come to ask Meeker about her record as a research analyst. But it becomes clear almost immediately that her real passion is the investment banking side of her job.

Though it's hardly news anymore that the Chinese Wall once separating investment banking from Wall Street research has eroded, what you realize in talking to Meeker is the extent to which these two supposedly conflicting functions--keeping companies happy and giving investors honest stock advice--are now organically intertwined. She talks unashamedly, for instance, about how she has used her research to help land banking deals for Morgan Stanley. And she describes how upset she became when Morgan Stanley lost a hotly contested deal to archrival Goldman Sachs. "I had never lost an IPO mandate in my life for a company that I wanted to take public," she says, sounding like, well, an investment banker. "I flipped into overdrive."

A notorious workaholic, Meeker left the office at 4 A.M. the day she first spoke to FORTUNE. She seems none the worse for wear; overdrive, in fact, appears to be her perpetual state. She launches into lengthy, digressive answers that hop self-consciously--and often humorously--from one subject to the next. That's typical Meeker. "Mary will just go off on those tangents," says a friend. "It's like trying to catch a meteor. That's how her mind works. It's a complicated place." It's also part of her charm. While she can be brutally high-handed with colleagues and subordinates--she's a diva, remember--she is also quick-witted, fun, and, when the occasion calls for it, self-deprecating. Being an analyst, she says, "is like being a Gumby. You're pulled in a lot of different directions."

Meeker got her first taste of the life 15 years ago, when she entered the analyst training program at Salomon Brothers. Growing up in a small Indiana farm town, she had been fascinated with Wall Street since high school, when she entered a stock-picking contest and watched her stocks promptly double. By the time she joined Salomon Brothers, she had spent two years as a broker with Merrill Lynch, had earned her MBA from Cornell, and had a clear goal: She wanted to be a portfolio manager. She viewed the Salomon job as a useful stepping stone.

Instead, she got hooked. Playing second banana to the firm's respected computer analyst, Michele Preston, Meeker covered the likes of Pitney Bowes and Eastman Kodak as well as the computer industry. When Preston jumped to S.G. Cowen in 1990, Meeker followed. A year after that investment banker Frank Quattrone, the legendary leader of Morgan Stanley's technology-banking team--it had taken Cisco and Apple public--offered Meeker a job as a senior analyst covering PCs and computer software. Though Morgan (along with Goldman Sachs) is the most prestigious investment-banking firm, Meeker turned Quattrone down the first time. "I didn't think I was ready to be a stand-alone analyst at a firm like Morgan Stanley," she explains now. But finally she succumbed.

For the next four years, Meeker labored in the lengthy shadow of Quattrone, a swashbuckling character who ran Morgan's tech group like a personal fiefdom. Although Meeker was involved in some IPOs--no underwriting can take place unless the firm's analyst "signs off" on it--her primary responsibility was covering such established companies as Microsoft and Compaq for Morgan Stanley's institutional clients. Soon after arriving at Morgan in 1991, she issued a massive report initiating coverage on ten software and computer stocks. She rated eight of them as buys. The following year most of them plummeted, giving her, as she puts it now, one of her "two shitty years as a stock picker." (The other "shitty year," of course, was 2000.)

That first Morgan Stanley report was more than just a stock-picking document, though. It also set out what Meeker called her "Ten Commandments for Investing in Technology Stocks"--and it offered remarkably sound advice. "Technology stocks are volatile...." she warned. "Buy when no one is interested in them ...sell when everyone is interested in technology (or when attendance at technology conferences reaches record levels or when your grandmother wants to buy a hot technology IPO)." Investors should buy "when fundamentals are intact"; they should sell when stocks "begin to trade down after large rises." And the tenth commandment: "Don't fall in love with technology companies. Remember to view them as investments." Sadly, these were rules that Meeker herself would abandon in the coming years.

Two things happened in the mid-1990s that changed everything for Meeker. First, on Aug. 9, 1995, Morgan Stanley underwrote the initial public offering of a tiny software company called Netscape. To this day people in Silicon Valley mark the Netscape IPO as ground zero for the Internet era. Netscape set the pattern for all the hot dot-com IPOs to come: It had incredible buzz and no profits (it gave away its popular Web browser), and it enjoyed a huge pop in the market on its first day of trading. Perhaps most important, it alerted the investing public to the promise of this new thing called the Internet.

Later Meeker would claim to have brought Netscape to Quattrone's attention, telling him, "This is a really big idea." Though Quattrone loyalists scoff at her account, citing his close ties to Netscape co-founder Jim Clark, what's clear is that Netscape was huge for Meeker. For one, she was associated with a big winner in an exciting new industry, something every analyst craves. For another, as the analyst who brought Netscape public, she was suddenly in a position to become the authority on the Internet.

She quickly set out to make the most of this opportunity. Four months after the Netscape IPO, Meeker published a 300-page research report called simply "The Internet Report." Crammed with charts, primers, tables, and stock recommendations, it quickly became the bible for investors interested in the Web. No less a personage than Intel's Andy Grove read it while on vacation in Hawaii--and suddenly realized that Intel needed to embrace the Net. Demand for the report was so high that Morgan Stanley even arranged for it to be commercially published--a first for a piece of Street research. Meeker was becoming a star.

There's one other thing about Meeker's role in the Netscape deal that seems important in retrospect. In covering the newly public company, Meeker displayed the protectiveness that would come to characterize her work. Referring privately to Netscape as "my baby," she aggressively defended the stock, even after Microsoft took dead aim at the company with its competing browser. In January 1997, Deutsche Morgan Grenfell analyst Bill Gurley, citing execution risks, downgraded Netscape from buy to accumulate. The stock dropped nearly 20%. Meeker responded two days later by upgrading her own rating to a strong buy--and angrily told another analyst, "If this company dies, it's Bill Gurley's fault." The shares continued their tumble, recovering only after AOL announced it was buying the beleaguered Netscape--almost two years later.

And the second critical mid-1990s event? That was the sudden departure of Frank Quattrone and his top deputies from Morgan Stanley on Easter Sunday, 1996. In what became known inside the firm as "the Easter massacre," Quattrone decamped for Deutsche Morgan Grenfell, which had agreed to give him unheard-of autonomy and pay. In an earlier age a move like that would have been unthinkable. But the world had changed. Though an old-line firm like Morgan still had plenty of prestige, it could no longer expect bankers to stay for life--or business to fall in its lap. Quattrone's departure was just another signal that Morgan had to get in the trenches and work for deals like everyone else.

Though none of Morgan's tech analysts had known of Quattrone's plans, he belatedly made a bid to steal the top three: Meeker, George Kelly, who covered data-networking, and Chuck Phillips, who covered business software. But after Morgan Stanley made them generous counter-offers, they all stayed put.

For Meeker, at least, there was another calculation at play. "Morgan had just lost the people who had built their tech practice," explains a friend. "Mary opted to stay, knowing Morgan needed a new star." With Quattrone gone--replaced by blander souls who lacked his stature--Mary Meeker soon became the new Quattrone.

We need to stop a moment to absorb the implications of this. Plenty of publications, including this one, have pointed out that analysts have become far more involved in the process of landing banking business than they once were. The modern analyst helps the banking team smoke out promising companies, sits in on strategy sessions, and promises--implicitly at least--to "support" the company once it has gone public with favorable research. That this makes tough-minded, independent stock research difficult, if not impossible, is no longer even an issue at most firms; investment banking brings in far more money than, say, brokerage commissions from grateful investors, thankful for unbiased research. Indeed, these days most analysts' pay is directly linked to the number of banking deals they're involved in.

What Meeker did once Quattrone left, however, went far beyond the usual analyst's accommodation. Rather than supporting Morgan's Internet banking effort, she began driving it. She became the firm's Internet rainmaker and its key dealmaker. "When I was at Deutsche Morgan Grenfell," says Gurley, now a venture capitalist (and FORTUNE columnist), "we talked about Mary being one of the best investment bankers on the planet." She developed close ties to VCs like John Doerr, who gave her early peeks at promising companies. She visited all the top startups to decide which were Morgan Stanley material. "Mary started to be the primary relationship person," says a high-profile banker at a competing firm (who, like most people FORTUNE spoke to about Meeker, would talk only if promised anonymity). "She'd almost bring her bankers in as a sort of execution team." Replies Meeker: "Our bankers were great. I was part of a team."

There was another dynamic at play: Because the Internet was so new, investors wanted a credible source to explain it to them--and tell them which companies to invest in. And who was more credible than Meeker? She became the gold standard, the person who gave a company instant stature merely by her association with it. Thus, even as she was chasing companies for Morgan Stanley, companies were chasing her. "The bankers were superfluous," says Todd Wagner, former CEO of Broadcast.com, recalling his company's decision to go with Morgan when it went public in 1998. "Our rationale was, if we went with Morgan Stanley, we'd get Mary Meeker and we'd get a lot of attention."

"We were not competing with Morgan at all," adds a rival banker. "We were competing with Mary Meeker. The clients made that very clear." Over time, competing firms cooked up strategies to combat the Meeker factor. One favorite was to spread rumors that Meeker was about to leave Morgan Stanley--which the Morgan people would have to quell.

Which is not to say that Meeker chased every deal. On the contrary: In the early years of Internet mania, she was highly selective, choosing only those companies she thought would be huge winners--and that she could support with a clear conscience. Meeker points out today that Morgan Stanley did only about 8% of all Internet IPOs--and that she turned down a tremendous number of deals. During due diligence, recalls former Morgan tech banker Andre de Baubigny, Meeker would "torture" companies as she probed management teams. Sometimes she passed on companies she should have grabbed--most notably Yahoo, which she blew off when it wanted to talk to Morgan Stanley about going public.

What almost never happened was that Meeker lost an Internet deal she really wanted. By Meeker's count, it only happened three times. As she describes the first, she starkly illustrates just how caught up she had become in the competition to win banking business. And the lengths to which she would go to win.

It was May 1998, and Morgan Stanley was in a "bake-off" to land eBay's upcoming IPO. During Morgan's presentation, Meeker and her bankers acted as if the business were already theirs, according to an eBay director. "She was on her pager a lot and getting calls from the CEO of Hewlett-Packard," he says. "The fact that she was a rock star showed itself a little at the meeting. Our concern was that we might not get enough of her attention." The eBay board chose Goldman instead.

Shocked by the rejection, Meeker "flipped into overdrive," as she later put it to FORTUNE. She offered her personal apology to eBay CEO Meg Whitman for failing to convey her full appreciation of eBay's business. Then she held out a carrot no other firm could proffer: a Mary Meeker research report. In July she met Whitman at Boston's Logan Airport for dinner and presented her with a draft of a glowing report on the company. Whitman told Meeker that she felt honor bound to stick with Goldman. Then, when eBay went public a few months later--and with the Goldman analyst forbidden from issuing any eBay research during the post-IPO quiet period--Meeker took the unprecedented step of publishing her completed eBay report on the very first day of trading. She instantly became the most visible analyst on the stock. "eBay's market opportunity is huge," she proclaimed--and gave the stock an outperform rating. Seven months later, when the company did a $1.1 billion secondary stock offering, eBay forced Goldman to split the business with Morgan.

In 1996, Institutional Investor magazine, which has been ranking Wall Street analysts for some 30 years, created a new category: Internet analysts. Mary Meeker, of course, grabbed the top spot.

And deservedly so. Without question, she was way ahead of most analysts in seeing how big the Internet could be--and how profoundly it would touch every aspect of American life and commerce. Major institutional investors devoured her "big picture" reports, which she rolled out regularly on such subjects as Internet advertising, retailing, and infrastructure.

Meeker was also important in helping popularize the notion that the Internet was a giant "land grab," in which companies had to sacrifice profits for rapid growth. And she was a leader in using new metrics to assess Internet companies. How could it be otherwise? These were companies, after all, that were going public with great hopes but no clear path to profits--and yet the analysts needed to be able to justify the stocks to investors. With the original Internet IPO--Netscape--Meeker used a metric she called "discounted terminal valuation," a novel calculation based on anticipated margins and growth rates five years down the road.

As the Internet exploded, Meeker became bolder about relying on nonfinancial metrics such as "eyeballs" and "page views." Here she is, for instance, in a July 1998 report on Yahoo (entitled "Yahoo, Yippee, Cowabunga ..."): "Forty million unique sets of eyeballs and growing in time should be worth nicely more than Yahoo's current market value of $10 billion." Four months later, when she revisited the company, which had just reported its third quarter, she wrote that there were "five key financial highlights." First on her list--even before revenues or operating margins--was the fact that Yahoo's page views had risen 25%.

She also became amazingly flippant about valuation. On launching coverage, in the fall of 1997, of Amazon--which quickly became one of her flagship stocks (and later used Morgan for three big bond deals)--Meeker wrote that the company's valuation "gives us heartburn of a gargantuan proportion." But she quickly dismissed the concern: "We have one general response to the word 'valuation' these days: 'Bull market.'...We believe we have entered a new valuation zone." "She moved out her horizon on fundamentals," says CSFB tech analyst Mark Wolfenberger. "It's like some of those parties you went to in college--man, until the police showed up, it was wonderful."

Did Meeker ever show concern that it might all blow up someday? In fact, she did--repeatedly. In her "overview" reports, she often expressed nervousness at how fast dot-com stocks were rising. One of her mantras throughout the Internet boom was that only 30% of Internet companies would wind up being long-term winners, while the other 70% would ultimately fall below their IPO price. But which companies were included in that 70%? That she never said--except that they weren't any of the companies she covered. With rare exceptions, she kept an outperform rating on all her stocks.

To be sure, Morgan Stanley clients who bought Internet stocks that Meeker recommended were making a boatload of money--crazy valuations be damned. Yahoo, Amazon, eBay--not to mention Priceline, FreeMarkets, and CNET--it seemed like all the Net stocks she covered were doubling and tripling. And whenever one of them hit a bump in the road, she was quick to reassure one and all that it would turn out all right in the end.

Was Meeker's unwillingness to downgrade stocks a function of her desire to protect banking business? She insists that was not the case. She acknowledges, however, that she was protective of the companies she helped take public: "I am hard-pressed to downgrade a stock of a company I fundamentally believe in over the long term," she says.

But consider the case of America Online (now AOL Time Warner, the owner of FORTUNE's publisher), a company Meeker began covering way back in 1993--shortly after Morgan Stanley did a follow-on stock offering for the company. She awarded AOL a strong buy, and has kept either that rating or an outperform on the company ever since--including the dark days of 1996, when the company was faced with huge losses, service screw-ups, and questionable accounting.

Today Meeker describes her willingness to stick by AOL during the tough times as one of her best calls ever. But at least one Morgan Stanley client has a dimmer view of her stance during that period. In the fall of 1996, a Denver fund manager heard from a Morgan Stanley salesperson that Meeker had declared during the morning meeting that maintaining her strong buy on AOL was "the worst mistake of her career."

Troubled by the call, the fund manager, who held millions of dollars worth of AOL stock, spent two weeks trying to reach Meeker. When she finally called back, he recalls, "she waved me off the stock. She basically said, 'Look, I don't think I would mess around with AOL here. There are too many big issues.' "He adds, "She was talking out of both sides of her mouth: 'Look, I've got a strong buy on the stock, but I don't have any conviction.' And when the stock recovered, she took a victory lap!"

Meeker insists she wouldn't have made the "worst mistake" remark at the morning meeting. She acknowledges, however, that her confidence in AOL had evaporated. "The number of people who truly believed in that company at that time was very small. That was pretty lonely....I was basically ready to throw in the towel on the company. The pressure got to me." Meeker also admits that these sentiments were known to members of the Morgan sales force. In fact, she says that it was a couple of institutional salesmen who talked her out of dropping her strong buy rating on the stock.

By 1999, Internet mania was in full swing--and Meeker was its reigning celebrity. Just before the year began, Barron's pasted that "Queen of the Net" label on her; in April, The New Yorker published its glowing profile, making her, in effect, a crossover hit. By midyear, fielding Meeker's media calls--two dozen or more a day--had become a full-time job for a Morgan PR staffer. When Meeker arrived at a SoHo loft for a Vanity Fair photo shoot, the photo team grew hushed--"It's Mary!"--and rushed to offer her a cappuccino.

Everyone, it seemed, wanted a piece of Meeker. Dot-com entrepreneurs tried to figure out what redeye she was taking so they could buttonhole her during the flight. Barbra Streisand, Reggie Jackson, and Saudi Prince Alwaleed tapped her personal investing counsel. Morgan's nontech clients wanted her advice on how to bring out their inner Internet. Viacom was so insistent on having Meeker at a meeting of division heads that the company dispatched a corporate jet to ferry her to Bermuda and back for a 90-minute presentation. "I'd go to China," says a former Morgan executive, "and the bankers there would say, 'You gotta get Mary Meeker out here.'"

Did this take a toll? Of course it did. Meeker's hours, always brutal, became death defying. She had assistants working around the clock. She also became stretched incredibly thin, with junior analysts doing much of her research, and fund managers and companies alike complaining that it was impossible to get through to her. Although the firm had steadily expanded its Internet research team--ultimately attaching the word "Internet" to the titles of 26 analysts--there was only one Morgan Stanley Internet analyst anyone wanted to hear from: Meeker. For her part, she was intent on keeping her finger in everything. Says de Baubigny: "It became hard to scale the business because there was such a reliance on one person."

In time Meeker began signing her name to research reports that were largely written by others. For instance, the business-to-business software company Ariba fell naturally to Meeker's colleague Chuck Phillips, a top-ranked analyst in his own right. But Ariba told Morgan Stanley that it would not get its IPO business unless Meeker's name was on the research reports. Though Phillips did most of the work, Meeker co-authored Morgan's Ariba research. Ditto Martha Stewart Living Omnimedia, a traditional media company that wanted to be perceived--and valued--as a dot-com. That meant it needed Meeker. To get its IPO business, Meeker signed on as co-analyst.

At the peak, Meeker was following 30 companies--twice the normal research load. "My name ended up on things where I was not the point person," she says now--now that many of those stocks have tanked. "It's the way corporations work. You gotta be a team player." She adds, "I did what I was asked to do."

Trying to cope with the craziness of 1999, Meeker began to feel new pressures both as an analyst and as an investment banker. And, she concedes today, she succumbed.

Let's take investment banking first. Meeker had long prided herself on Morgan Stanley's taking public only the very best prospects--the premier company in every "space," as they used to say. But in 1999 the number of Internet IPOs exploded from 42 the year before to 294, making it virtually impossible for Morgan Stanley to stay on the high road and not fall in the closely watched market-share rankings. With Credit Suisse First Boston--whose tech group was then being run by none other than Frank Quattrone--and Goldman Sachs aggressively pursuing Internet IPOs, Meeker was suddenly under intense pressure to maintain Morgan's share of the business.

For instance: In 1999, Meeker had decided not to pursue the IPO of the Internet grocer Webvan because she didn't have faith in its business model. Similarly, she passed on the women's content site iVillage. But both companies had spectacular IPOs and were soon sporting monster market caps--$8 billion in the case of Webvan, and $2.6 billion for iVillage. With iVillage now at $1.42 a share and Webvan at 13 cents, she was clearly right to stay away--from an analyst's standpoint. But from a banking perspective, the initial success of those IPOs meant that Meeker had egg on her face.

How did she react? By lowering her standards. Abandoning her "only the best" dictum, she took public a raft of second-tier companies that have now become embarrassments to Morgan Stanley: HomeGrocer, Women.com, AskJeeves, tickets.com, Lastminute.com.

Meeker could see that deal quality was deteriorating. But, she says, "we had to take risks....We couldn't not be in that game." Thus, in 1999, Morgan Stanley did 27 Internet IPOs. Though that was still less than 10% of all Web IPOs, it was more Internet underwritings than the firm had done the previous four years combined. One result: Meeker's salary jumped from a reported $6 million to $15 million.

As an analyst, she was also feeling pressure. Forced to support companies that weren't all that good, at valuations that were increasingly difficult to justify, she was boxed in. In May 1999, Morgan Stanley took a company called Scient public at $20. By the time the quiet period had ended--and Morgan, with Meeker as co-analyst, could initiate coverage--the stock had jumped to almost $40 a share. Because of the run-up, the Morgan analysts gave the stock a rare "neutral" rating, citing valuation concerns. Three months later they upgraded to "outperform"--even though the stock had climbed to $63. "The market was saying, 'You're an idiot, you're an idiot, you're an idiot,' " Meeker says now. "This is the reality of the world we lived in."

"The hardest call to make is to be negative and wrong," says CSFB analyst Wolfenberger. "You had analysts going negative on Amazon and having the stock triple from that point. Those analysts aren't in business anymore."

And then it all fell apart. NASDAQ collapsed, dot-coms imploded, and the valuation bubble burst. Mary Meeker went from hero to goat.

The essential charge that has been hurled at her this past year centers on her refusal to downgrade her stocks, even as they dropped 70%, 80%, more than 90% in some cases. In effect she's being accused of selling out investors to keep Morgan Stanley's banking clients happy. In the New York Times last December, Gretchen Morgenson noted that Meeker had an outperform rating on all of her Internet stocks--down an average of 83%--and pointedly asked a Morgan Stanley PR official whether "her nonstop optimism had anything to do with the fact that most of the companies had engaged Morgan Stanley as an investment bank."

Meeker, for her part, feels unfairly maligned. "I'm tired of the witch-hunt punching-bag stuff," she says in frustration. "The other side of the story never seems to get told." The "other side," as Meeker sees it, goes like this. Last year, admittedly, was an awful year for her stock picks--but only the second bad year of her career. ("Do you judge Ted Williams on one bad year?") If you look at the totality of her record, she says, you'll see that in the time she's been at Morgan Stanley, her picks have created nearly $700 billion in wealth. But 88% of that figure comes from three computer stocks--Microsoft, Dell, and Compaq, where she hasn't been an important voice in years--plus AOL Time Warner. The dot-com stocks she is most closely associated with--such as Priceline, Yahoo, and Amazon--have been disasters in the past year.

Yet even now Meeker is not ready to concede that these stocks are disasters, even with Priceline at $4 (from $162), Yahoo at $19.50 (from $237), and Amazon at $15 (from $106). On the contrary, she insists that they still deserve their "outperform" rating. After the "nuclear winter," she says, we'll see "the spring bloom" for her favorite stocks. Sometime within the next two to three years, the "aggregate" valuation of the "leading names" that make up her recommended list will exceed even the heights they reached during the dot-com bubble. Says Meeker: "Our bet is that the winners that come out of this, the market value of the leaders, are going to make all the things that came before them look like chump change."

As for disgruntled individual investors who feel misled by her recommendations, she offers little sympathy. As she told the Wall Street Journal, "Every individual has got to be accountable for how they're allocating their investments." Her "real" constituency, she added, is professional money managers and other institutional investors.

To FORTUNE she said, "Did we do some deals we shouldn't have done? Yes. Did we recommend some stocks we shouldn't have? Yes." But, she added, "it's difficult to get hit and hit and hit when we did a better job than any other firm."

She's right, of course, that individuals who have lost money in this market need to look themselves in the mirror before they begin blaming Wall Street analysts. But the plain implication of her statement is that institutions are sophisticated enough to see through her ratings. And, indeed, some of them are. Former hedge-fund manager Jim Cramer says that it's been clear since last summer that "no one [on Morgan's sales staff] was pushing her stuff, because she wasn't enthusiastic. She has a buy because she's gotta have a buy. I thought she distinguished herself by not pushing the stuff. You might call that duplicitous. I was happy because it made me money."

The most critical point is this: Mary Meeker got so caught up in the allure of the Internet--the celebrity, the money, the thrill of dealmaking--that she forgot that she was supposed to be analyzing companies. "She was flying at 50,000 feet, talking about trends," says a rival banker. "She had no idea what the fundamentals were." Thus, when a company like Priceline turned out to have serious problems, including its disastrous foray into gasoline and groceries, Meeker didn't have a clue. Instead, here's what she says about Priceline, a company (it almost goes without saying) that Morgan took public in 1999: "It wasn't troubled until it was troubled. It was fine on Wednesday, bad on Thursday....You can say, 'Why didn't the idiot analyst figure it out?' [But] you have to have some degree of trust in the concepts and the management team. With Priceline, that was a mistake." Even so, she maintains her outperform rating on the stock.

Then again, why wouldn't she? "It's one thing if you've got a disciplined valuation criteria," says a banker who worked with Meeker. "But if you don't and you're outperform for no good reason, what's the reason for turning it to an underperform?"

In that first interview with FORTUNE, Meeker said that her approach has always been to find the big idea and latch onto it. "If the approach that worked for me in the past doesn't work in the future, I'll make the wrong call," she said. Well, that's one way of looking at it. Here's another: "If you're going to say 'Buy, buy, buy,' " says a former Morgan Stanley executive, "at some point, you gotta say 'Sell.'"

REPORTER ASSOCIATES Mary Danehy, Jessica Sung, Julie Schlosser

FEEDBACK: pelkind@fortunemail.com