Case Study 2 — The Champagne Crisis at the Pommery House, 1874

In the spring of 1874, in a chalk cellar 30 meters under the city of Reims, a problem that had been creeping up for two years became impossible to ignore.

The Pommery house — Veuve Pommery, "Widow Pommery," then under the management of Louise Pommery after her husband's death in 1858 — had been steadily expanding its champagne production. The cellars beneath the Pommery property in Reims were vast: 18 kilometers of tunnels carved into the soft Cretaceous chalk that lay beneath the Champagne region of northeastern France, with a constant cellar temperature of about 10°C year-round, perfect for the slow second fermentation that gives champagne its bubbles.

The problem was that of those bubbles, between 5 and 10% of the bottles in any given batch were exploding.

A champagne bottle finishes its second fermentation under 5–6 atmospheres of internal pressure — about the pressure inside a passenger car tire, but with no flexible wall to absorb shock. A flawed bottle, or a bottle filled with wine that had over-fermented, or a bottle whose wall had a hidden weakness, could detonate with the force of a small grenade. In a working cellar, this was not just a financial loss; it was a danger to the workers walking the rows. Cellar staff at all the major Champagne houses wore iron face masks during inspection, like beekeepers, because flying glass came at speed.

Louise Pommery's chief enologist that year, a man named Jacques Bouchet, had a theory about why the explosions had been increasing. The theory was correct. Implementing it would change champagne production for the next 150 years.

The bottle's accidental physics.

Champagne is made by the méthode champenoise, a technique formalized in the 17th and 18th centuries (with Dom Pérignon's name attached to it more for marketing than for accuracy). The basic idea: produce a still wine through normal grape fermentation; before the wine has fully clarified, transfer it to bottles along with a small precise dose of additional sugar and a fresh yeast culture; cap the bottles; allow a second fermentation to occur inside the sealed bottle.

That second fermentation produces ethanol (slightly raising the alcohol content) and CO₂ gas. With no escape, the gas dissolves into the wine — and continues dissolving until the partial pressure of CO₂ above the wine inside the bottle equals the equilibrium imposed by Henry's law and the production of more gas. In a typical champagne bottle, that pressure builds to about 5–6 atmospheres at cellar temperature. If you uncork the bottle, you release the pressure suddenly, and the dissolved gas comes out as bubbles. (This part of the science was understood by 1874, though not in those exact terms.)

The trouble was the dose of sugar. The dose — liqueur de tirage — was added by hand, with a measuring ladle, by cellar workers running thousands of bottles a week. It was supposed to be a specific amount: roughly 24 grams of sugar per liter, calibrated to produce 6 atmospheres of pressure. But "roughly" was the operative word. A ladle-and-eye dose varied wildly. Bottles that received slightly more than 24 g/L would over-ferment and produce more pressure than the bottle could hold. Bottles that received slightly less would under-ferment and have weak bubbles, which the customer would notice and complain about.

Bouchet, examining the broken-bottle records over two years, made a striking observation. The explosion rate correlated almost perfectly with cellar temperature. In the cool months — December through March, cellar temperature 8–10°C — the explosion rate was about 3%. In the warmer months — July through October, when the cellar warmed to 13–15°C — the rate climbed to 12% or more.

This was Henry's law again, though Bouchet wouldn't have used the name. (Henry's law had been published in 1803, but it was a chemist's tool, not a winemaker's.) At the cooler temperature, more CO₂ stayed dissolved in the wine, lower partial pressure built up above the wine, and the bottle held. At the warmer temperature, the same total CO₂ partitioned more into the gas phase — out of solution, into pressure — and the bottle was more likely to fail.

Two factors compounded: the over-doses produced too much total gas, and the warmer cellar pushed more of that gas into the gas phase. An over-dosed bottle in a 9°C cellar might survive; the same bottle in a 14°C cellar would burst.

Bouchet's two interventions.

Bouchet proposed two changes to Louise Pommery in March of 1874, in a meeting documented (with somewhat dramatic language) in the Pommery house records that are now in the regional archives at Reims.

First, the dose itself had to be measured by weight, not by ladle. Bouchet sourced precision balances from a chemical-supply house in Paris and trained the cellar staff to weigh out 24 g of sugar per bottle, ±0.5 g, before adding it to the bottle. This sounds obvious to a modern reader. In 1874, it was an unusual investment. The balances cost more than the workers' annual wages and required care that the cellar staff had to be trained to provide. Within six months of the change, the variance in the liqueur de tirage dose had fallen by an order of magnitude. The mean dose was the same; the distribution was tight. Quality control through measurement.

Second, the cellar temperature itself had to be controlled. Pommery's cellars were already among the deepest in Reims (the deeper, the more thermally stable), but they varied by 5–6°C across the year. Bouchet proposed segregating the cellar into temperature zones — using the deepest, coolest galleries for the prise de mousse (the second-fermentation phase, where pressure built) and the warmer galleries (closer to the surface) only for the long aging that followed, after the bottles had been disgorged and re-corked at lower pressure.

This was a logistical revolution. The Pommery cellars had to be re-mapped. Bottles in different stages of production were moved to different depths, with the youngest, most-pressure-sensitive bottles deepest. The thermal gradient of the chalk itself was exploited as a free temperature-control system.

Within a year, the explosion rate had fallen to under 1%. Within two years, it was under 0.5%. The cellar staff stopped wearing the iron masks for routine inspection. Production rose substantially because fewer bottles were being lost.

The lesson and the legacy.

Bouchet's interventions were applied chemistry — Henry's law, mass balance, temperature control — and they came from a man who had probably never heard the name Henry. He had observed, measured, and corrected. The chemistry he was applying was real; the framework he was using was experiential.

This is part of what theme #4 of the book points at. Across centuries, in fermentation traditions that range from village bread to champagne to pu-erh tea, the people doing the work have arrived at correct chemical insights through observation and trial. The chemistry we name now is what their work discovered. Bouchet would have been comfortable in any modern laboratory; the laboratory would have to be comfortable acknowledging him.

The system of precise dosing and zoned-cellar production he developed at Pommery in 1874 became the standard across the Champagne region within a decade. Other houses — Krug, Roederer, Bollinger — adopted it. The annual rate of bottle explosions in Champagne, which had been a known industrial hazard for a century, fell to nearly zero. The méthode champenoise you read about today, with its precise dose and its monitored cellar temperature, traces directly back to that 1874 reform.

There is a small ritual still observed in some Champagne cellars. When the harvest comes in, and the first bottles are being prepared for the second fermentation, the new dose is weighed — not ladled, weighed — on a small precision scale that lives on a shelf near the liqueur de tirage tank. The scale is symbolic now; modern automated dosing equipment has long since replaced hand-weighing in most large operations. But in a small house, a careful enologist will still do it once a season, to start the year right, in honor of an old habit whose origin few people remember.

📊 Analysis prompt:

1. Bouchet's analysis of bottle-explosion records found a strong correlation between cellar temperature and explosion rate. Translate this into the language of Henry's law and gas-phase partitioning: why does a warmer cellar give more failures from the same total fermentation?

2. The hand-ladle dose of sugar produced wildly variable bottle pressures, while the precision-weighed dose was tight. Estimate (with reasoning) what the variance in bottle pressure would have been at each precision level, given that pressure scales roughly linearly with the dose. What does this tell you about the sensitivity of the failure rate to dose precision?

3. Modern champagne is bottled with a degree of precision Bouchet would have envied. But the ratio of liqueur de tirage to total wine is still chosen with care. If a winemaker wanted a champagne with a lower internal pressure (4 atm instead of 6 atm — say, for a "softer" style of sparkling wine), what change to the recipe would they make? If they wanted a higher pressure (some German Sekt traditions push to 6.5 atm), what would they change? Use mass-balance thinking, not an empirical recipe.