Grow one plot of grapevines on the top of a hill in a really sunny, exposed, windy spot. Grow another plot at the bottom of the hill in a shady waterlogged spot. It is really not rocket science to understand that the grapes are likely to be quite different, as will wine made from them. That’s terroir, stupid.

Pushing the argument to these extremes doesn’t really resolve the issue of whether terroir effects exist that produce wines with subtle but consistent differences from adjacent vineyards where there is a little perceptible difference in growing conditions. Burgundians will argue, for example, that the adjacent vineyards of Cazetiers and Combes aux Moines in Gevrey Chambertin produce different wines. Faiveley have adjacent vineyards with vines of the same age, tended the same way. “The tractor doesn’t stop,” says Jérôme Flous of Maison Faiveley. But as I describe in my book, In Search of Pinot Noir, the wines are consistently different.

Well, yes, but maybe in such cases, producers, consciously or unconsciously, are treating the wines just a little bit differently to produce results that conform with their expectations for each vineyard. You might almost call it the Holy Grail of Terroir to demonstrate that the wines from different vineyards are intrinsically different. A group of scientists have just reported an attempt to prove this.

They took Pinot Noir berries from two vineyards managed by the same producer in Burgundy, one in Flagey-Echézeaux and one in Vosne Romanée. They don’t give exact locations, but the vineyards appear to be within a mile of one another. They report that soils types are similar in each vineyard: they don’t comment on the age of the vines or planting density, but let’s give the benefit of the doubt and assume that all extraneous parameters are the same.

They took two samples of 100 berries from each vineyard in 2010, 2011, 2012 and compared the berries and wine made from them by microvinification. Well, here’s the first problem. ONE HUNDRED BERRIES! How many berries do you suppose there are in a vineyard? Let’s try to estimate this. Suppose the vineyard is 1 ha and has 8,000 vines, each with 6 bunches, and each bunch has 50 grapes. I make that 2.5 million berries. I do not think you need to be an expert in statistics to see that 100 berries is not going to be representative of the vineyard. (For the more geekishly minded, you’d need a few hundred to achieve a 5% statistical significance level.)

I won’t go into the details of the analysis, which uses formidably complicated equipment, but comes down to the ability to measure small amounts of phenolic compounds. The authors conclude that  differences between vintages are more significant than differences between vineyards. This is not a surprise. But they go on to say they can find  differences between the vineyards, for both grapes and wine. This I do not believe. It’s definitely fair to say that when you compare the samples in one vintage, all four are different. But I’m not at all convinced that the two samples from each vineyard identify any distinct character. The results look  higgledy-piggledy, and the two samples from each vineyard seem just as different from one another as from the other vineyard. Misquoting Monty Python, Every berry is sacred, every berry is great.

I believe in terroir, but I don’t believe this study proves any more than that each sample of one hundred berries is a different from every other sample of one hundred berries. (Gives a whole new meaning to micro-cuvées.) Anyway, do we want to see a scientific basis established for terroir, wouldn’t that spoil the fun?