What a Bottle Does in the Dark

A bottle in a cellar is not resting. It is running a slow set of reactions, and what people call ageing is the sum of them.

The tannins and pigments (phenolic compounds) link into progressively longer chains. As the chains grow they become less soluble, and eventually they precipitate out as the sediment in an old bottle. This is why an old red is paler and softer than a young one, and why it needs decanting off its deposit.

Meanwhile the primary aromas of young fruit fade, and a tertiary register emerges: undergrowth, mushroom, truffle, game, dried flowers, tea. In red Burgundy this transition is the whole point, and it is why the wines are cellared.

The crucial thing to be clear about is that the wine is not improving indefinitely. It is changing, continuously, in one direction, and the window during which that change is an improvement is finite and specific to the bottle. After it, the fruit is gone and what remains is structure and decay.

How long that takes is far less predictable than the culture of cellaring suggests. A village wine from a light vintage may be at its best at four years and tired at ten. A grand cru from a structured vintage may need fifteen. Vintage, appellation, producer style and storage all matter more than price, and a great deal of expensive Burgundy is drunk well past its point by people who were told to wait.

The closure turns out to matter more than most drinkers assume.

Natural cork admits a very small and variable amount of oxygen. That variability is a substantial part of why two bottles from the same case can differ noticeably after twenty years: same wine, same cellar, different corks, different trajectories. Screwcaps with engineered liners give far more consistent transmission, which is an advantage if you value consistency and a subject of genuine argument about whether the variation is a fault or part of what makes old bottles interesting.

Then there is the episode that shook the region's confidence in its own white wines.

From the mid-1990s, white Burgundies that should have had a decade ahead of them began turning brown and flat at five or six years. Not occasional bottles: a widespread, cross-producer, cross-vintage failure that destroyed cellars and reputations.

Premature oxidation (premox) was never conclusively diagnosed. The candidate causes include changes to cork preparation after chlorine bleaching was abandoned, over-vigorous lees stirring introducing oxygen, lower sulphur additions as a matter of fashion, riper grapes with less protective acidity, and gentler pressing removing phenolics that would have protected the wine.

Practices changed on all of these fronts, sometimes for unrelated reasons, and the problem has largely receded. It remains a genuinely unsolved case, and it is worth remembering whenever anybody describes winemaking as a solved technology.

Two more things worth knowing if you keep bottles.

Wines go through a closed phase. A wine that showed beautifully on release can be mute and hard a few years later, and open again after several more. The mechanism is imperfectly understood (aromatic compounds appear to bind into non-volatile forms for a period) and the phenomenon is consistently reported by people who taste the same wines repeatedly over decades. It is a good argument for buying more than one bottle, since a single sample tells you only about that day.

And old wine is fragile in a way young wine is not. A twenty-year-old Burgundy can be sublime for half an hour and collapse in the glass, because the tannins and sulphur dioxide that protect a wine against oxygen have both diminished. The oxygen that first lifts its aromatics then destroys them.

Which is why old bottles are opened at the table rather than decanted hours in advance, and why the standard advice about letting a wine breathe runs in precisely the opposite direction for something young and closed.