Warren Buffett, who was probably the greatest investor of the 20th century, is fond of quoting the salacious actress Mae West as saying, Too much of a good thing can be wonderful. In the market, such a motto would lead you to avoid diversification and instead concentrate your portfolio in stocks you really, really like.

Peter Lynch, who was probably the best mutual fund manager of the 20th century, calls spreading yourself too thin diworseification.

Smart, witty and brilliant at picking stocks, Buffett and Lynch may not need diversification, but the rest of us do. When you own one stock, youre out on a limb. For example, very few analysts with or without a conflict of interest predicted that shares of Enron, the energy and trading company, would tumble by 90 percent in a year. Put all your eggs in a basket like that and you end up with a gooey mess. The more stocks you own as long as they are in different industries the more the overall riskiness of your portfolio is modulated.

The reason you dont want a super-risky portfolio is simple: While Warren Buffett may be calm and prescient enough to ride out severe dips in the value of his holdings, most investors are not. A portfolio that increases in price by 10 percent each and every year is worth exactly the same at the end of three years as a portfolio that falls by half the first year, rises by three-quarters the second and rises by 52 percent the third. But reasonable investors prefer the consistent ride. It prevents them from doing something stupid, such as selling all their stocks after losing half their money during that first disastrous year.

Consider the sad case of James D. McCall, who earlier this month resigned as manager of the Merrill Lynch Focus Twenty mutual fund. Two years ago, Merrill wanted McCalls services so desperately that the firm went to court to pry McCall away from his previous employer, Pilgrim Baxter, where he rang up impressive gains in the late 1990s. (His big success was called PBHG Large Cap 20.) And when they got McCall, Merrills brokers raised more than $1.5 billion from their clients for him to invest. While the average growth-stock mutual fund owns about 100 stocks, with the top 10 holdings representing about one-fourth of the portfolios total value, McCall specialized in what are called concentrated portfolios. In the case of Merrill Lynch Focus Twenty, he owned, as the name implies, just 20 stocks. At last report, his top 10 holdings accounted for a whopping two-thirds of the funds assets.

If McCall had spread his 20 stocks among, say, a dozen different industries, he might have smoothed his ride. Instead, 69 percent of his assets went to technology firms. The Focus fund and a smaller one that McCall ran called Premier Growth were launched in March 2000. Within just 17 months, all but $650 million of the clients original $1.5 billion had vanished.

It is hard to imagine losing as much as Focus Twenty did even if you tried. As of Nov. 9, the week McCall resigned, the fund was down 72 percent for the year, compared with a loss of 14 percent for the Standard & Poors 500-stock index, the benchmark for fund managers. According to the latest report from Morningstar Mutual Funds, 19 of McCalls 20 stocks had declined during 2001, the only exception being Harley-Davidson. More amazing, 16 of the 19 losers had fallen by at least half. (By the way, Enron was McCalls seventh-largest holding.)

This fund has had a wretched existence, wrote Morningstar analyst Kunal Kapoor, who did admit a grudging admiration for McCalls perseverance. McCalls faith may turn out to be well placed over time, Kapoor said. Unfortunately, time ran out.

My point here is not to pick on McCall but to reveal the perils of concentration. Buying Focus Twenty as a technology fund, and consigning it to no more than one-fifth of your holdings (with the rest of your assets in diversified, conventional stocks or funds) might have made sense, but Focus Twenty was touted as a long-term capital appreciation fund, not a sector fund. Here, it failed, but maybe it didnt have to.

The manager who made the concentrated fund popular, Tom Marsico, who ran Janus Twenty, took care to spread his holdings around. His successor, Scott Schoelzel, has suffered losses lately (he is down 28 percent year-to-date, but thats after a total gain of 546 percent in the preceding five years), but they have not been nearly so catastrophic and for good reason. Schoelzels last report lists among his top 10 holdings three tech stocks, two financials, one drug company, one energy firm (whoops, Enron again), one industrial, one consumer-durables company and one services firm.

For investors in individual stocks, the important question is this: How much diversification is enough? Some risk is inherent in even the broadest portfolio. This is called market, or systematic, risk. Over the past 75 years, market risk, as measured in standard deviation, has been about 20 percent. In other words, in two-thirds of the years the annual return of the S&P has fallen into a band ranging from 20 points lower to 20 points higher than its average return of 11 percent; that is, between a loss of 9 percent and a gain of 31 percent. Thats still volatile, but if you invest in stocks you have to live with it.

What you dont have to live with is anything more volatile. So your objective in building a portfolio is to try to approximate systematic risk and avoid what is called idiosyncratic, or extra, risk. A portfolio with just a few stocks, or one like McCalls, that is overloaded in a single sector, has lots of idiosyncratic risk. In 1977, an influential study found that investors could nearly eliminate that extra risk by owning just 20 stocks in a wide variety of sectors; in fact, owning eight or 10 stocks depressed risk sharply.

Recently, however, the market has appeared to be far more volatile, and a new study by a group of economists headed by John Campbell of Harvard found that many more stocks were needed around 50 to bring a portfolio down to the same level of riskiness as the broad market. What Campbells group found was that neither the market itself nor individual sectors had become more volatile in the 1990s, but that stocks within those sectors had, so you need to own more of them.

But owning 50 stocks is a pain in the neck and it brings up the Buffett-Lynch admonitions about too much diversification. It is hard just to take the time to make the selections, but even buy-and-hold investors need to keep track of the companies they own to spot adverse changes in management, product failures or new competition (not to mention Enron-style accounting shenanigans) signs that its time to sell.

One good answer is to achieve balance by owning a combination of mutual funds and stocks. For example, you might want to put 50 percent of the money you have allotted for stocks into a fund that mimics the S&P itself, like Vanguard Index 500, which charges rock-bottom expenses and guarantees that risk wont exceed systematic levels. You could also consider a broad fund thats managed by human beings, such as Meridian Value or Baron Growth, which are recommended by Sheldon Jacobs, editor of the No-Load Fund Investor newsletter. Then another 25 percent of your holdings can go into a few sector funds that specialize in technology, real estate, energy and small-caps, and the final 25 percent into a portfolio of 10 to 20 individual stocks. (I own 16, at last count.)

There are many valid variations. Just dont emulate Mark Twain.

In a letter to clients recently, Anthony M. Maramarco of David L. Babson & Co. the Cambridge, Mass. investment firm, recalled the aphorism of Twains Puddnhead Wilson: Put all your eggs in the one basket and watch that basket! Unfortunately, such a philosophy emphatically does not work in stock investing as Twain himself learned when he sank nearly all his fortune into the Paige Linotype, a machine that flopped.

We all make mistakes. (It was Twain, after all, who pointed out that human beings are the only animals that blush or need to.) But smart diversification helps investors avoid some of the worst of them.