What happens when blending water fails its spec? Hold, release, rework — and the corrections that do not exist

Key takeaways

• The failures split into two classes, and only one of them is fixable. A strength miss is arithmetic and can be corrected in the tank; a chemical or sensory taint carried in by the water cannot be taken back out, because there is no step after proofing that removes it.
• US regulation puts a formal hold point exactly where it is needed. A proprietor “must gauge the product after any filtering, reduction, or other treatment, and before bottling or packaging begins,” at label proof, within a tolerance that “may not exceed 0.3 percent alcohol by volume above or below the alcohol content stated on the label” (27 CFR 19.353; 27 CFR 19.356).
• Codex names four things a non-conforming batch can become — “destroyed, used for purposes other than human consumption, determined to be safe for human consumption, or reprocessed” (Codex, General Principles of Food Hygiene CXC 1-1969, §13.5). For a proofed whisky, two of those four are usually unavailable.
• The water spec is a legal spec in India, not just a purchase spec. FSSAI’s alcoholic beverages standards require that “the water used for dilution to bottling strength shall meet the requirements as specified in Indian Standards for Drinking Water, IS:10500 as amended from time to time” (FSSAI, Compendium — Alcoholic Beverages Regulations).
• Destruction is less ruinous than it sounds, and that is deliberate. In the US, “there is no tax liability on spirits … destroyed in accordance with this section” on bonded premises (27 CFR 19.459). You lose the spirit, not the duty — provided the problem is caught before tax determination.

It depends entirely on whether the water got the strength wrong or brought something with it. A strength error is recoverable: the alcohol is still there, the arithmetic is reversible, and the regulation builds in a gauge and a hold point before bottling precisely so the tank can be corrected. A contamination error is not recoverable, because the step that would remove it does not exist — blending water goes in at the end, after filtration, after maturation, after every process that could have cleaned anything up. That asymmetry is the whole of the subject, and it is why the decision that matters is made at goods-inwards, not in the bottling hall.

Where does a blending water failure actually get caught?

There are only three realistic places, and they run in increasing order of pain.

At goods-inwards, on the certificate of analysis. This is the only place where a failure costs nothing but a tanker. Codex requires that incoming materials “be procured according to specifications,” with compliance “verified, where necessary” by tests that “may be conducted by a supplier that provides a certificate of analysis, the purchaser, or both” (CXC 1-1969, §13.2.8). The “or both” is doing real work: a CoA describes the batch the supplier tested, while what reaches your tank is that batch after road transport, a hose and a receiving line. Hence the value of knowing what each number on a certificate of analysis actually measures.

At the bottling tank gauge. US law places a mandatory check here: the proprietor “must gauge the product after any filtering, reduction, or other treatment, and before bottling or packaging begins,” and that gauge “must be made at labeling or package marking proof” (27 CFR 19.353). The permitted band is narrow — variations “may not exceed 0.3 percent alcohol by volume above or below the alcohol content stated on the label” (27 CFR 19.356). This gauge catches strength, and nothing else. Water that is perfect on conductivity and carrying a chlorophenol at nanograms per litre passes it without a flicker.

In the glass, weeks later. A taint that survives to a tasting panel or a customer has already been bottled, labelled, duty-paid and shipped. By then the question is not quality control but recall.

The gap between the first and the third checkpoint is where the money is.

Which failures are correctable, and which are not?

The top two rows are routine production events. The rest are not events; they are losses.

Why can’t you simply take the water back out?

Because every available method of removing water from a spirit either destroys the product’s legal identity or destroys its character — and usually both.

Scotch is the clearest case. The Scotch Whisky Regulations 2009 define the category as a spirit “to which no substance has been added, or to which no substance has been added except” — and then list only “water; plain caramel colouring; or water and plain caramel colouring,” at “a minimum alcoholic strength by volume of 40%” (SWR 2009, reg. 3). The regulation says what may go in and nothing about taking anything out, because the category was never designed around the possibility. A matured whisky run back through a still is not a corrected whisky; it is new-make spirit with an expensive history.

Nor can you adsorb a fault away without adsorbing the whisky with it. Carbon treatment works, and the trade has used it for decades, but the constraint has been understood since at least the 1960s: “the activated carbons used in the practice of the present invention must be carefully selected in order that the color, flavor, and congeners not be removed from the whiskey being stabilized” (Eek & Sucietto, US3046138A, 1962). Carbon does not distinguish between the chlorophenol you want gone and the phenols you spent twelve years acquiring. On a peated malt, that is not a remediation; it is a different product.

There is one correction that works cleanly, and it only works on strength: blending up. If a batch has come out at 39.6% ABV against a 40% label, the fix is to add spirit of the same make at higher strength until the tank gauges within tolerance. This is ordinary practice and the reason bottling halls hold back strength stock. The arithmetic is the same arithmetic as the original reduction, run in the other direction — and it carries the same trap, since ethanol and water contract when mixed and the dose cannot be derived by subtraction. Note the direction of risk: over-dilution is cheap to fix, under-dilution is cheaper still, and a taint is not on the table.

What can a non-conforming batch legally become?

Codex sets out the disposition options for product removed or held back. Such products “should be held under secure conditions until they are destroyed, used for purposes other than human consumption, determined to be safe for human consumption, or reprocessed” (CXC 1-1969, §13.5). Map those four onto a proofed spirit:

1. Destroyed. Available, and less catastrophic than it sounds if the timing is right: in the US, “there is no tax liability on spirits, denatured spirits, articles, or wines destroyed in accordance with this section” on bonded premises (27 CFR 19.459). You lose the liquid and the labour, not the excise — provided the batch never left bond.
2. Used for purposes other than human consumption. Available. Tainted spirit has industrial outlets. It has no premium ones.
3. Determined to be safe for human consumption. This is the one buyers reach for, and it is a trap. Safety and saleability are different tests. A whisky proofed with water carrying a chlorophenol at single-digit nanograms per litre is entirely safe and entirely unsellable.
4. Reprocessed. Available for strength. Effectively unavailable for taint, for the reasons above.

Two of four, in practice. That is a thin set of options for a batch that may represent a six-figure inventory value, and it is the reason the industry’s real control is upstream.

Is out-of-spec water a safety problem or a quality problem?

Usually a quality problem — and Indian law has a precise word for that. Under the Food Safety and Standards Act 2006, an article of food is “sub-standard” if “it does not meet the specified standards but not so as to render the article of food unsafe” (FSS Act 2006, s. 3(1)(zx)). Most blending water failures land squarely in that category: the product is harmless and non-compliant at the same time.

What makes this more than semantics in India is that the water specification is written into the product standard, not left to the producer. FSSAI’s alcoholic beverages standards state that “the water used for dilution to bottling strength shall meet the requirements as specified in Indian Standards for Drinking Water, IS:10500 as amended from time to time” (FSSAI compendium). If the reduction water breached IS 10500, the finished spirit is non-conforming as a matter of regulation, whether or not anyone can taste it. A producer working to a demineralised specification sits far inside that envelope — which is the point of specifying blending water tighter than drinking water.

The genuine safety exceptions are few but real: trace metals above the drinking-water limits, and anything that arrived through a cross-connection — which is why Codex requires that water “not fit for use in contact with food … should have a separate system that does not connect with or allow reflux into the system for water that will contact food” (CXC 1-1969, §13.3).

Why does the certificate of analysis carry so much weight?

Because of one clause, which reads like boilerplate and is not. Codex: “no incoming material should be accepted by an establishment if it is known to contain chemical, physical, or microbiological contaminants which would not be reduced to an acceptable level by controls applied during sorting and/or processing” (CXC 1-1969, §13.2.8, emphasis added).

The clause assumes a downstream reduction step exists. For blending water in a spirits plant, it does not. Water is the last significant thing added; after it the operations are gauge, filter if the house style allows, fill, cap, label — none of which reduces a dissolved compound. The hurdle the clause expects processing to provide has to come entirely from the incoming specification.

This is also why a CoA alone is an incomplete control. Purified water does not hold its own specification unattended — storage and pipework will drift it — so the number that matters is the one measured at the point of use, on the day of use.

What should a hold-and-release procedure actually say?

1. Define release criteria before the tanker arrives, not after. Conductivity, TDS, pH, hardness, trace metals, TOC, microbiological count — with numbers and methods, and a named person who may release.
2. Verify, do not merely file. Run at least conductivity and a sensory check on every delivery against the CoA. Codex contemplates testing by supplier, purchaser, “or both”; for a material with no downstream control, choose both — and smell the water, which costs nothing and catches the one failure class no routine instrument panel is looking for.
3. Treat the proofing tank as a hold point, not a transfer. The US bottling tank gauge is a good model even where it is not the law: gauge after reduction, before bottling, release on that result.
4. Hold back strength stock. The correctable failure mode should always be correctable. Keep higher-strength spirit of the same make available to the bottling hall.
5. Write the taint scenario down in advance. Decide in daylight what happens if a panel rejects a proofed batch — who is told, what is quarantined, what the water-system investigation looks like. Nobody decides that well with a full tank waiting.
6. Retain a sample of every water delivery for the shelf life of the product it went into. Codex requires records kept “for a period that exceeds the shelf-life of the product”; a retained sample is what turns a later complaint into a closed investigation.

A note on scale, from an adjacent industry. In 1990 Perrier recalled 160 million bottles across 120 countries after benzene was found in its water, traced to a gas-line filter “that was supposed to be changed every six weeks” being “left on for up to three months” (UPI, 14 February 1990). A missed maintenance interval. Water systems fail quietly and invisibly, until the failure is in the product.

Glossary

• Blending water (proofing or reduction water) — the purified water added to spirit to bring it from cask strength to bottling strength.
• Certificate of analysis (CoA) — the supplier’s test report for a specific batch of material, stating the parameters measured and the results obtained.
• Corrective action — in Codex terms, “any action taken when a deviation occurs in order to re-establish control, segregate and determine the disposition of the affected product, if any, and prevent or minimize reoccurrence of the deviation.”
• Deviation — “failure to meet a critical limit or to follow a GHP procedure” (CXC 1-1969).
• Disposition — the decision about what a held or non-conforming batch becomes: destroyed, diverted from human consumption, released, or reprocessed.
• Bottling tank gauge — the mandatory US measurement of a product’s proof after reduction and before bottling (27 CFR 19.353).
• Blending up — correcting an under-strength batch by adding higher-strength spirit of the same make, rather than by removing water.
• Sub-standard — under India’s FSS Act 2006, food that “does not meet the specified standards but not so as to render the article of food unsafe.”
• Bonded premises — the duty-suspended area of a spirits plant; spirits destroyed there do not attract excise liability.

Frequently asked questions

Can a whisky proofed with the wrong water be fixed? Only if “wrong” means the wrong amount. A strength error is corrected by adding water, or by blending up with higher-strength spirit of the same make, then re-gauging. Water that carried a taint, a trace metal or an off-odour into the tank cannot be corrected: nothing downstream of proofing removes a dissolved compound from spirit.

Why not just filter the fault out? Carbon will adsorb an off-flavour, but it is not selective. The constraint is old and well documented: activated carbons “must be carefully selected in order that the color, flavor, and congeners not be removed from the whiskey being stabilized” (US3046138A). On an aged or peated spirit, treatment heavy enough to remove a taint removes the product.

Is a batch proofed with out-of-spec water unsafe to drink? Usually not. Most blending water failures are compositional, not toxicological — in Indian terms “sub-standard” rather than unsafe (FSS Act 2006, s. 3(1)(zx)). That makes it a commercial and compliance problem rather than a public health one, which is the more common and more expensive case.

At what point does it become too late to correct? Practically, at the bottling tank gauge, which US regulation places “after any filtering, reduction, or other treatment, and before bottling or packaging begins” (27 CFR 19.353). Fiscally, at tax determination: destroy a batch in bond and “there is no tax liability” (27 CFR 19.459); destroy it after duty is paid and the duty is part of the loss.

Does the supplier’s certificate of analysis discharge the buyer’s obligation? No. Codex contemplates verification by the supplier, the purchaser, “or both,” and bars acceptance of material carrying contaminants “which would not be reduced to an acceptable level by controls applied during sorting and/or processing” (CXC 1-1969, §13.2.8). Since no such reduction step exists after proofing, the buyer’s own check at the point of use is the last control there is.

Is the water specification legally binding in India? Yes, for the reduction water. FSSAI’s alcoholic beverages standards require that “the water used for dilution to bottling strength shall meet the requirements as specified in Indian Standards for Drinking Water, IS:10500” (FSSAI compendium). A breach makes the finished product non-conforming regardless of sensory outcome. The wider framework is set out in FSSAI, IS 10500 and the industry floor.

Blenders International · Water Science series