Which water in your whisky was distilled, and which was poured in? The three waters in every bottle

Key takeaways

• A 70 cl bottle at 40% ABV holds about 420 ml of water, and it has three separate origins. Water that co-distilled from the wash, water added to reduce new make to cask-filling strength, and blending water added at bottling.
• Only the first of those was purified by physics. In distillation, “any impurities in the original water, such as non-volatile or mineral components, that do not boil below or near the boiling point of water remain in the original container” (Wikipedia, Distilled water). The mash water’s calcium left the plant as pot ale, not as whisky.
• The other two are only as pure as somebody’s specification. Both are dosed from a tank, and nothing downstream removes what they carry.
• On standard Scotch assumptions, about 62% of the water in the bottle was added rather than distilled. Roughly 52% at the bottling hall and 10% at the filling store, against 38% that came over the still.
• Even cask strength is not water-free of added water. New make leaves the spirit safe at “between approx. 68 and 71%” and “is usually diluted with water to a strength of 63.5%” before filling (Whiskipedia) — so roughly a fifth of the water in a cask-strength bottling was poured in, not distilled.
• Blending water carries about five times the exposure of filling water, because it contributes about 31% of the finished bottle against the filling water’s 6%. A part per million in the blending tank lands in the glass at roughly 0.31 ppm.

Water is the largest single ingredient in a bottle of Scotch, but it did not all arrive by the same route, and the routes are not chemically equivalent. Some of the water in your glass travelled through the still as vapour, which makes it distilled water by definition — every non-volatile mineral in the distillery’s source water was left behind in the wash still. The rest was poured in from a tank, once before maturation and once at bottling, and that portion is pure only to the extent that someone wrote a specification and tested against it.

How many separate waters actually end up inside the bottle?

Three, and they enter the process years apart.

A distillery runs five or six distinct water streams — mashing, cooling, boiler feed, cleaning, reduction — each specified differently, as we set out in how many different waters a distillery uses. Most never reach the consumer. Cooling water never touches the product, cleaning water is rinsed away, and mashing water leaves the site as effluent: “for every 1 L of alcohol produced, approximately 2.5 kg of draff, 8 L of pot ale, and 10 L of spent lees (and washings) were the resulting yield” (White et al., ACS Omega, 2020).

Only three waters survive into the bottle:

1. Still water — the water that vaporised alongside ethanol in the spirit still and condensed with it. Ethanol and water cannot be separated completely by distillation, which is why Scotch must be “distilled at an alcoholic strength by volume of less than 94.8 per cent” (Scotch Whisky Regulations 2009) and why the practical ceiling is lower still, as we explain in why alcohol cannot be distilled above 96%.
2. Filling water — added to new make before it goes into oak, to bring receiver strength down to filling strength.
3. Blending water — added at the bottling hall to bring cask strength down to label strength.

Where does the still water come from, and what is in it?

It comes from the mash water, but almost nothing of the mash water comes with it.

This is the part most water marketing elides. Distillation is a separation, and non-volatile solutes do not make the journey: “any impurities in the original water, such as non-volatile or mineral components, that do not boil below or near the boiling point of water remain in the original container,” which is why distillation will “almost completely remove all dissolved minerals such as calcium, magnesium, sodium, fluoride, potassium, iron, and zinc” (Wikipedia, Distilled water).

So if a distillery mashes with water at 150 ppm calcium, the calcium does not board the vapour. It stays in the wash still and leaves as pot ale. To a first approximation the water arriving in the spirit receiver is mineral-free regardless of how hard the source was — which is the honest technical core of the soft water versus hard water regional story and of the terroir claim. Source water shapes the fermentation. It does not shape the mineral content of the bottle.

Mineral-free is not the same as pure, though. Still water arrives carrying the congeners it co-distilled with, and it arrives having been in intimate contact with hot copper — contact that is the point of the exercise, since “the copper of stills removes sulfur volatiles during distillation” and “contact with copper for both liquid in the pot and vapor in the lyne arm is the most important for reducing levels of sulfur compounds” (Wanikawa & Sugimoto, Molecules, 2022). Some copper comes across with it; pot ale itself carries dissolved copper “from 1.7 to 15.3 mg/L” (White et al., 2020).

How much water goes in before maturation?

Less than most people assume, but it is not nothing — and it is the water the industry talks about least.

New make “flow[s] through the Spirit Safe with an alcohol content of between approx. 68 and 71%,” and “before being filled into the oak casks, the New Make is usually diluted with water to a strength of 63.5%” (Whiskipedia) — a convention WSET confirms as near-universal for single malt (WSET). We cover why that number and not another in cask filling strength.

Run the arithmetic on 100 ml of spirit at filling strength. It holds 63.5 ml of alcohol, which arrived inside 63.5 ÷ 0.69 = 92.0 ml of new make at 69% ABV. That new make carried 28.5 ml of still water. The remaining 8.0 ml is filling water, poured in from a tank.

At the moment the bung goes in, about 78% of the water in the cask is distilled and about 22% was added. That ratio is then locked for the life of the cask. Evaporation cannot change it, because a water molecule that came over the lyne arm is indistinguishable from one that came out of a hose — the angels take them in proportion.

So what is the split in a finished 70 cl bottle?

Here is the full audit, on stated assumptions: new make at 69% ABV, filled at 63.5%, drawn at 58% after humid Scottish maturation, bottled at 40% in a 70 cl bottle. Volumes are nominal — the small non-additive contraction when ethanol and water mix is set aside here and handled properly in the proofing arithmetic.

The bottle holds 280 ml of pure alcohol and 420 ml of water. To supply 280 ml of alcohol you need 280 ÷ 0.58 = 483 ml of cask spirit, which brings 203 ml of water with it. The remaining 217 ml is blending water. Split that 203 ml by the 78:22 ratio fixed at filling, and the picture is:

Nearly two-thirds of the water in a standard bottle of Scotch — 62% — never went through a still. It was measured out of a tank by someone whose only quality control was a specification and a certificate of analysis. The single largest component of the bottle after alcohol itself is the blending water, at 31% of everything in the glass.

Does the split change with strength and climate?

Substantially, and in both directions.

Two things fall out of that table. First, a higher bottling strength genuinely does mean less added water — a 46% bottling carries about a third less blending water than a 40% one, which is part of why the minimum bottling strength of 40% is a commercial decision as much as a legal floor.

Second, warehouse climate moves the number more than bottling strength does. Where humidity is high, “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). A cask that gained strength in a dry warehouse needs far more water at bottling — 64% of the finished bottle’s water, against 52% from a humid Scottish cask. Indian, Australian and Taiwanese producers live at one end of that range and Speyside at the other; we work through the consequences in the angels’ share and cask strength drift.

The fourth row is the one worth pinning up. A cask-strength bottling is not free of added water. About a fifth of its water was dosed at the filling store, which is why that water deserves a specification even at a distillery that never proofs below 50%.

Are the three waters interchangeable?

No, and conflating them is the most common error in how distilleries think about water risk.

Still water required no specification, no testing and no supplier. The still did the work, and it cannot fail to do it: whatever the source water’s hardness, nitrate or TDS, none of it boarded the vapour.

Filling water and blending water are the opposite. Both are dosed, both can carry whatever the treatment plant let through, and crucially nothing downstream removes what they bring. Spirit is not filtered to strip calcium, boiled again, or carbon-treated the way sake is. The tolerance is unforgiving — “as little as two ppm of calcium in the final product can cause precipitation” (Artisan Spirit), a level most drinking-water standards would not bother to record. The law is equally blunt: Scotch is spirit to which “no substance has been added… except—(i) water; (ii) plain caramel colouring; or (iii) water and plain caramel colouring” (Scotch Whisky Regulations 2009). Water is one of only two permitted additions. It is an ingredient, in law.

The two added waters are not equal risks either, and the arithmetic says which to worry about. Blending water is 31.0% of the bottle; filling water is 6.3%. A contaminant present at the same concentration in both tanks arrives in the glass at roughly five times the dose from the blending water — and the filling water has had a decade in oak, with an active, extracting, oxidising system around it, to blunt whatever it brought. Blending water gets no such grace period. It goes in, and then the bottle is sealed.

What should a buyer take from this?

Three practical conclusions.

1. Rank your water specifications by share of the bottle, not by order of use. Mash water is the largest volume and the smallest product risk, because the still filters it absolutely. Blending water is the smallest volume and the largest product risk. Spending on the first while under-specifying the second is a common and expensive inversion — the logic behind what blending water is.
2. Specify your filling water too. It is roughly a tenth of the water in a 40% bottle and about a fifth of the water in a cask-strength one, and it is routinely drawn from whatever is nearest in the filling store.
3. Know your out-turn strengths before forecasting water volumes. A dry-matured portfolio needs materially more blending water per litre of alcohol than a humid-matured one, and that water is a larger proportion of what the customer drinks. The purity argument gets stronger, not weaker, as casks come out hotter — see the demineralised water specification.

Glossary

• Still water — our term for the water that co-distilled with ethanol and condensed into the spirit receiver. Distilled by definition; carries congeners and trace copper, but no non-volatile minerals.
• Filling strength — the strength at which spirit enters the cask, conventionally 63.5% ABV for single malt Scotch, reached by adding water to new make.
• Filling water — the water added to new make to take it from receiver strength to filling strength.
• Blending water / reduction water — demineralised water added after maturation to bring cask strength down to bottling strength. The last water in and the only one with nothing downstream of it.
• Out-turn strength — the ABV at which a cask is actually emptied, after years of evaporative drift in one direction or the other.
• Pot ale — the residue left in the wash still after the first distillation, which carries away the non-volatile minerals from the mashing water.

Frequently asked questions

How much of a bottle of whisky is water? At 40% ABV, 60% of it by nominal volume — about 420 ml in a 70 cl bottle. At 46% it is 54%, and at cask strength around 58% ABV it is 42%. In every case water is the largest ingredient other than ethanol.

Does the distillery’s famous spring water end up in the bottle? Its minerals do not. Distillation leaves “non-volatile or mineral components” behind in the still (Wikipedia, Distilled water), so the source water’s calcium, magnesium and TDS leave the plant as pot ale. The source water shapes mashing and fermentation. What the drinker swallows is distilled water plus whatever water was added afterwards.

Is cask-strength whisky undiluted? Not entirely. It has no water added at bottling, but Scotch new make is “usually diluted with water to a strength of 63.5%” before filling (Whiskipedia), so roughly 22% of the water in a cask-strength bottle was still added from a tank. A handful of distilleries fill at receiver strength and genuinely avoid this.

Why is blending water specified more tightly than mashing water? Because mashing water passes through two filters the blending water never sees — a fermentation and a still — while blending water is added last, to a finished product, with no removal step after it. Calcium at 2 ppm can precipitate in the bottle (Artisan Spirit), a level that is unremarkable in drinking water.

Why does warehouse climate change how much blending water is needed? Because it decides whether the cask gains or loses strength. In humid air ethanol escapes faster and ABV falls; in hot, dry air water escapes faster and ABV rises (Whiskipedia). A cask drawn at 65% needs far more water to reach 40% than one drawn at 58%, so the added water is a bigger share of the finished bottle.

Blenders International · Water Science series