Along the lanes of Burgundy, A. J. Liebling once wrote, the very road signs read like wine labels. The place names speak of the flinty or chalky earth, of soils that have for centuries sacrificed themselves to yield up tantalizing notes of apple or leather. But most Americans, having mastered a simple five- or six-part vocabulary of varietals (Cabernet, Chardonnay...), never learned to tell a Chassagne Montrachet from a Puligny-Montrachet. California wines have mostly been defined by the type of grape and the wine maker's style, not the vineyard or the year. That suits the typical wine drinker seeking consistent enjoyment rather than an intellectual challenge, says Linda F. Bisson, an oenologist at the University of California, Davis. The soil and weather, being Californian, were assumed to be perfect.

But precisely because their products are so routinely excellent, California wine makers have begun to look for ways to distinguish themselves. "A few years ago, customers never asked about specific vineyards," says Dennis Cakebread, proprietor of Napa's Cakebread Cellars. "Now they seek them out." Even Robert Mondavi's inexpensive Woodbridge label, whose typical offerings are produced from grapes trucked in from all over California, has introduced a Select Vineyard series of wines (suggested retail: $10.99) grown exclusively in Lodi, a region near Sacramento. In the $175 bracket, the ultracult wine maker H. William Harlan has just released the third vintage of his single-vineyard Bond wines, for connoisseurs desiring to "taste the different expressions of the land." Now selling for about $300,000 an acre, the fields of Napa have finally earned the honorific long bestowed on their counterparts in France: they are terroir.

The concept is still new to America, and often misunderstood. Olivier Krug, of the French champagne family, defines it as "the combined effect of all that you find in nature: topography, soil, subsoil and climate." That these things should affect the taste of wine seems intuitively obvious, but it's never been rigorously studied or proved, says David G. Howell, a U.S. Geologic Survey researcher and coauthor of last year's definitive study of Napa Valley terroir, "The Winemaker's Dance." In particular, Howell dismisses out of hand the notion that complex flavors like blackberry or chocolate are somehow inherent in the ground. Even if they were in the soil, the necessary organic molecules wouldn't be absorbed by the roots. Vines take up nutrients as individual elements (calcium, silicon, nitrogen and so on) and recombine them into the only thing they make, which is grapes. "In wine stores, they tell you there used to be an apricot orchard in the vineyard, so you get apricot notes in the wine," he complains, "and it's just hogwash." If people are tasting those things, they were produced elsewhere: in the fermentation, the aging--or the imagination.

Even the greatest wines--maybe, especially those--have qualities that can't be measured in a mass spectrometer. Terroir is best expressed by metaphor, not chemical analysis. Cakebread's Three Sisters Cabernet is grown on a steep, sun-drenched hillside of rocky volcanic soil, a terrain that produces what he calls "the Arnold Schwarzenegger of wines... a grrr wine that grabs your tongue and shakes it around." By contrast, the Benchland vineyard is well shaded, gently sloping, with a brown, loamy soil. That innocuous countryside, sure enough, produces "the Isabella Rossellini of wines, refined and elegant." The man clearly knows his own wine, but as a serious scientific proposition, this is on a par with the belief that lions are courageous.

Still, Howell believes science can shed light on the workings of terroir. Wine makers know the importance of a well-drained plot; too much water makes for an insipid and grassy-tasting wine. Excessive sun, by contrast, can produce overwhelming fruit and inordinate alcohol. At the last meeting of the American Geophysical Union, Howell led a group of research-minded journalists on a tour of the fantastic geologic diversity of the Napa Valley, illustrating his points with tasting stops at Opus One Winery and Stag's Leap Wine Cellars. The group reached consensus on one thing: more research is needed.