Why does a cask gain strength in Kentucky and lose it in Goa? The angel's share, humidity, and the water you add back

Key takeaways

• Both water and alcohol leave a cask; humidity decides the ratio. In humid air “relatively more ethanol is lost than water in the barrel and the ethanol content decreases,” while “in relatively hot and dry climates… this results in an increase in the ethanol concentration” (Whiskipedia).
• The rates are not comparable across climates. Scotland loses “1-2%” a year; India and Australia reach “as high as 12%,” with Taiwan’s Kavalan recorded at 15% (Whiskipedia). Paul John reports “approximately eight per cent loss each year as Angel’s Share” in Goa against “two to three per cent in Scotland” (Paul John).
• The direction reverses inside a single building. Bardstown Bourbon Company’s Dan Callaway describes upper floors with “massive temperature swings… with relatively low humidity” producing “higher proof,” while ground-floor barrels “retain more water and lose more alcohol, resulting in lower-proof juice” (VinePair).
• Out-turn strength sets the blending-water bill. A cask emerging at 70% ABV needs roughly 35% more blending water per litre of pure alcohol than one at 58.5% — arithmetic, not opinion, worked below.
• Humid maturation eats the headroom. Scotch must have “a minimum alcoholic strength by volume of 40%,” and the only permitted additions are “water… plain caramel colouring; or water and plain caramel” (Scotch Whisky Regulations 2009). A cask drifting toward 40% has almost nothing left to dilute.

A maturing cask loses both water and ethanol through the oak, and the ratio between the two — not the total loss — is what moves the strength. Where the surrounding air is humid, water leaves slowly and ethanol leaves at close to its full rate, so ABV falls. Where the air is hot and dry, water escapes freely and the remaining spirit concentrates, so ABV rises. Everything downstream — how much blending water a bottling takes, how many bottles a cask yields, whether the liquid still qualifies as whisky — follows from which of those two regimes the warehouse sits in.

What actually evaporates from a cask?

Both components, continuously, in a ratio set by the air outside.

Oak is porous. Water and ethanol both migrate through the staves and away, which is why the volume in a cask falls every year without anyone opening it. The industry’s shorthand for that loss is the angel’s share, and the figure quoted for Scotland — “the average annual volume loss is around 2% per year” (Hedonism Wines) — describes the total liquid gone, not its composition.

The composition is the interesting part, and it is not fixed. It shifts with ambient humidity, and it can invert entirely between two climates using identical casks and identical filling strength.

Why does humidity decide whether strength rises or falls?

Because each component leaves at a rate driven by the gap between its concentration inside the cask and its concentration in the air outside — and the outside air is full of water vapour and effectively empty of ethanol.

Scotland’s warehouses sit in air with “average humidity of 88%,” where “relative humidity is often 80-90 percent, especially in winter” (Whiskipedia). Air that saturated has very little capacity left to accept more water, so water’s escape route is throttled. Ethanol faces no such restriction: ambient air contains essentially none of it, so the gradient pulling ethanol outward stays wide open regardless of the weather. The result is preferential alcohol loss — “when the air humidity is high, relatively more ethanol is lost than water in the barrel and the ethanol content decreases as the whisky matures” (Whiskipedia).

Dry the air out and water’s throttle comes off. Water now has somewhere to go, it leaves quickly, and the spirit left behind concentrates: “there is a loss of ethanol in relatively hot and dry climates. This results in an increase in the ethanol concentration during barrel maturation” (Whiskipedia). The same pattern appears in the general account of cask behaviour — humid coastal warehouses see “more alcohol evaporates than water, leading to a decrease in ABV,” while in drier environments “more water evaporates than alcohol, which can sometimes lead to an increase in ABV over time” (Hedonism Wines).

One popular explanation deserves retiring. You will often read that proof rises because water molecules are smaller and escape more easily. Molecular size is not doing the work here — ethanol is the more volatile of the two at warehouse temperatures, so a size-based account predicts the wrong answer in Scotland. What changes between Speyside and a Kentucky top floor is not the molecules. It is the humidity of the air they are escaping into.

Is the angel’s share mostly alcohol or mostly water?

It depends on the climate, and the arithmetic is worth doing because the answer flips.

Take a humid-warehouse cask filled with 200 litres at 63.5% ABV. Applying the ten-year figures of “a 5% drop in ABV and 20% loss in volume” (Hedonism Wines), it comes out at 160 litres and 58.5% ABV:

• Pure alcohol: 127.0 litres in, 93.6 litres out — 33.4 litres of alcohol gone
• Water: 73.0 litres in, 66.4 litres out — 6.6 litres of water gone

Roughly 84% of what the angels took was alcohol. Now take a hot, dry rickhouse, where “the legal maximum barrel entry proof for United States whiskey, including bourbon, is 125 proof, or 62.5 percent alcohol by volume” (Spirit Sight) and upper-floor barrels climb from there (VinePair). Model 200 litres in at 62.5%, out after eight years at 150 litres and 70% ABV — an illustrative out-turn, not a measured one:

• Pure alcohol: 125.0 litres in, 105.0 litres out — 20 litres of alcohol gone
• Water: 75.0 litres in, 45.0 litres out — 30 litres of water gone

Here 60% of the loss was water. Same oak, same physics, opposite composition. (These are volume-of-component figures; the small non-additive contraction when ethanol and water mix is set aside here and handled in full in the proofing arithmetic.)

Does this really differ inside one warehouse?

Yes — vertically, by several floors, which is why placement is treated as a production variable rather than logistics.

Dan Callaway, master blender at Bardstown Bourbon Company, puts the internal spread bluntly: “It is amazing to see both the temperature and humidity differences between the first floor and the seventh floor.” During a Kentucky heat wave the top floor sees “massive temperature swings… with relatively low humidity,” while “the bottom floor will consistently have a higher humidity than the outdoor temperature” (VinePair).

The consequences track the humidity exactly. Upper barrels undergo a “higher level of water evaporation,” yielding “higher proof and bolder, more concentrated flavors”; ground-floor barrels “retain more water and lose more alcohol, resulting in lower-proof juice featuring softer, sweeter notes with less burn” (VinePair). Warehouse design amplifies the split: tall modern buildings “introduce a considerable height difference between the top and bottom of the warehouse and the associated temperature stratification” (Spirits & Distilling).

How fast does the loss run in different climates?

Two cautions on reading that table. First, the early years are not representative — there is “about 10 percent reduction in spirit volume from inside the cask in the first year” before the rate settles (Spirits & Distilling), because fresh oak absorbs liquid it never gives back. Second, high loss and falling ABV are separate consequences of the same humid conditions, not the same fact: Kentucky loses more liquid than Scotland while moving its strength in the opposite direction.

How does cask drift change the blending water you need?

It changes it substantially, and in the direction most people would not guess: the cask that lost the most water needs the most water added back.

Reduce both worked casks to 40% ABV. Bottling volume is pure alcohol divided by 0.40:

The dry-matured cask takes about 35% more blending water per litre of alcohol — and that added water is nearly 43% of everything in the finished bottle. This is the practical reason out-turn strength belongs on a purchasing forecast rather than only a tasting note, and why the demineralised water specification matters more, not less, for high-strength casks: a larger share of the product is water, so any calcium or off-flavour it carries is present in greater quantity.

It also explains the yield incentive. As one cask-brokerage guide notes, “if you’re flexible on bottling strength then you can legally add Scottish water to reduce the whisky’s strength down to anything above 40% to get more bottles from the same volume of liquid” (Mark Littler) — headroom that a dry-matured cask has in abundance and a humid-matured one may not have at all.

What happens if a cask drifts below 40% ABV?

It stops being whisky in the legal sense, and no amount of water can fix it — water is the only thing you are permitted to add, and adding it makes the problem worse.

The Scotch Whisky Regulations 2009 require “a minimum alcoholic strength by volume of 40%” and permit no addition except “water; plain caramel colouring; or water and plain caramel” (legislation.gov.uk). That 40% floor is a century old — it dates to a 1917 wartime ruling “that whisky had to be sold at no more than 30 degrees under proof (40% abv),” entrenched when 1920 duty rises made stronger bottling uneconomic (Scotch Whisky).

For humid-climate producers this is a live constraint, not a curiosity. Under Scotland’s ~0.5% ABV loss a year (Hedonism Wines) the floor is decades away. Under tropical conditions with 8–15% annual volume loss and ABV moving down, the window between out-turn strength and the legal minimum narrows fast. A cask arriving at 45% has just 0.28 litres of blending water per litre of pure alcohol available before it hits 40% — under a quarter of what the dry-matured cask absorbed. The strength has to be watched as a maturation endpoint, not discovered at the bottling hall.

None of which changes what the water itself has to be. Whether a bottling takes 0.28 litres per litre of alcohol or 1.07, the specification is the same: near-zero minerals, no flavour of its own, and nothing that will drop the whisky out of solution once it is in the bottle. What the angel’s share changes is the quantity — and how much of the finished product that quantity represents.

Glossary

• Angel’s share — the annual evaporative loss of liquid from a maturing cask through the oak. Around 1–2% a year in Scotland, 8–15% in tropical warehouses.
• Out-turn strength — the ABV of the whisky when the cask is emptied at the end of maturation. Not the same as the filling strength it went in at.
• Filling strength (entry proof) — the ABV at which spirit is filled into a cask, conventionally 63.5% in Scotch and capped at 125 US proof (62.5%) for straight American whiskeys.
• Relative humidity (RH) — how saturated air is with water vapour, as a percentage of what it could hold at that temperature. High RH suppresses water loss from casks; it does not suppress ethanol loss.
• Stratification — the vertical temperature and humidity gradient inside a tall warehouse, which makes upper and lower barrels mature differently.
• LPA (litres of pure alcohol) — the alcohol content of a volume of spirit, independent of its strength. The quantity that survives dilution and is used for duty and yield.
• Reduction (proofing) — adding purified water to lower spirit from out-turn strength to bottling strength.
• Bottling strength — the final ABV in the bottle; legally at least 40% for Scotch (SWR 2009).

Frequently asked questions

Does whisky get stronger or weaker in the cask? Either, depending on the humidity of the warehouse. Humid conditions favour ethanol loss and the ABV falls; hot, dry conditions favour water loss and the ABV rises (Whiskipedia). The direction can differ between the top and bottom floors of the same building (VinePair).

Is the angel’s share mostly water or mostly alcohol? Neither, universally. In a humid Scottish warehouse the loss works out overwhelmingly alcohol — around 84% by the arithmetic above. In a hot, dry rickhouse where proof climbs, water is the larger share. Total volume lost and the composition of that loss are two separate questions.

Why is India’s angel’s share so much higher than Scotland’s? Heat drives the total rate of loss up. Paul John reports “approximately eight per cent loss each year” in Goa against “two to three per cent in Scotland” (Paul John), and figures up to 15% a year are reported for Indian and Taiwanese warehousing (Spirits & Distilling, Whiskipedia). Humidity then determines the direction the strength moves — down, in most tropical coastal sites.

Does a stronger cask need more blending water or less? More. Bottling volume is fixed by the alcohol present, so a cask at 70% ABV reduced to 40% takes about 1.07 litres of water per litre of pure alcohol, against 0.79 litres for one at 58.5% — roughly 35% more, and nearly 43% of the finished bottle.

Can a cask lose so much strength that it can no longer be sold as whisky? Yes. Scotch requires “a minimum alcoholic strength by volume of 40%” (SWR 2009), and because water is one of only three permitted additions, a cask that drifts below the floor cannot be corrected by dilution. It is a reason humid-climate producers track out-turn strength through maturation rather than measuring it at the end.

Does the first year lose more than later years? Yes, considerably — “about 10 percent reduction in spirit volume from inside the cask in the first year” (Spirits & Distilling), because fresh oak absorbs spirit into the staves before the steady-state evaporative rate takes over. Annual averages quoted for a long maturation smooth this out.