Most honey that fails an export inspection does not fail on flavour. It fails on water. Moisture content is the first number a buyer’s laboratory reports, the one number a seller can still influence after extraction, and in nearly every case it is measured with a refractometer and two drops of honey.
The instrument is simple. The judgement around it is not. Two people can test the same drum, read half a percentage point apart, and end up arguing about whether the lot can ship. This article covers what a honey refractometer actually measures, which limit applies in which market, how to take a reading that survives a buyer’s check-sample, and the mistakes that quietly cost commercial operations the most money.
What the refractometer is measuring
Honey is a supersaturated sugar solution. Light passing into it slows and bends, and the amount it bends — the refractive index — rises as dissolved solids rise and falls as water rises. A refractometer measures that angle and prints it on a scale. On a honey instrument the scale is graduated directly in percent water, typically from about 12 to 27 percent, so the number in the eyepiece is the number you write down.
Two complications sit behind that simplicity.
First, the relationship between refractive index and water content is not linear, and it is not identical across honey types because the sugar profile differs. Clover honey and honeydew honey at the same water content do not refract identically. This is why the conversion is published as a table rather than a formula. The Chataway tables, still the reference behind most instruments, were derived from measured honey samples at a fixed temperature; a scale built on them is accurate for honey and inaccurate for anything else.
Second, a refractometer graduated in Brix is a different instrument. Brix is calibrated against sucrose solutions, and sucrose is not honey. Subtracting a Brix reading from 100 is an approximation that drifts at both ends of the scale, which is exactly where the commercial decisions are made. If the instrument says °Bx rather than percent moisture, either apply the maker’s honey conversion table or replace the instrument. Estimating the conversion in your head is how a lot that reads 19 percent gets shipped as 18.
The limits that apply, market by market
Moisture is regulated in most importing markets, and the ceilings are close but not identical. The table below collects the ones that come up in export documentation.
| Standard or grade | Maximum moisture | What it governs |
|---|---|---|
| Codex Alimentarius CXS 12-1981 | 20% general, 23% heather honey | The international reference standard most national rules are built on |
| EU Directive 2001/110/EC, Annex II | 20% general, 23% heather and baker’s honey | Composition requirement for honey sold in the EU; baker’s honey must also be labelled as intended for cooking |
| USDA Grade A (7 CFR 52, Subpart E) | 18.6% (81.4% soluble solids minimum) | Voluntary US grade; once printed on a label, the lot must meet it |
| USDA Grade B | 18.6% (81.4% soluble solids minimum) | Same moisture ceiling as Grade A, lower minimum score for defects, clarity and flavour |
| USDA Grade C | 20.0% (80.0% soluble solids), no sample unit below 79.0% | Lowest graded tier before substandard |
The US grade standard is worth reading in full, because the moisture rule is applied per sample unit rather than to a lot average. Four of five sample units can sit comfortably at 18 percent and the certificate will still record the fifth at 19.2. Buyers who understand the standard read the range, not the headline number, which is why a supplier who reports only an average looks evasive even when the honey is good.
Above the legal ceilings sits the contract. Most industrial buyers set a tighter specification than the regulation allows, because a cargo that only just meets the legal maximum is a cargo with no margin for a second test at the destination port. When a purchase order states 18.0 percent maximum, that figure — not Codex — is the number the lot has to hit.
Why 18.6 percent became the working line
The figure is not arbitrary. It is the moisture content that corresponds to 81.4 percent soluble solids, the threshold the US grade standard uses for its top two grades, and it has been carried into commercial practice far beyond the United States because it sits close to the point where keeping quality changes.
The mechanism is yeast. Honey carries osmophilic yeasts, chiefly Zygosaccharomyces rouxii, that stay dormant while the available water is low enough. As moisture rises, water activity rises with it, and at some point the yeasts start converting sugars to alcohol and carbon dioxide. The threshold most often quoted in extension and industry guidance is around 18 percent, and it is a risk threshold rather than a legal one: a honey at 18.4 percent with a light yeast load and cool storage may keep for years, while a honey at 17.8 percent with a heavy load and a warm warehouse may not.
Three details make the practical picture more useful than a single number.
Crystallisation moves water, not sugar. When glucose comes out of solution the remaining liquid phase becomes more dilute than the bulk average. A drum measured at 18.2 percent before it sets can have liquid-phase moisture well above that in the months that follow, and that liquid is what ferments. This is why honey intended for long storage is usually held below the threshold with margin rather than at it.
Temperature sets the rate, not the risk. Cool storage slows yeast activity without removing it. A warm container, a container in direct sun on a dock, or a slow sea freight route through the tropics all raise the rate. Two lots with identical moisture can leave the same port and arrive in different condition.
Water activity, not water content, is the variable the yeast responds to. Sugar concentration, temperature and the presence of crystallised solids all shift it. Moisture is the proxy everyone measures because it is cheap and fast, not because it is the whole story.
Taking a reading that will survive a check-sample
A refractometer is only as good as the sample on the prism. The protocol below is the one that holds up when a buyer’s laboratory re-tests the retained sample, and each step exists because of a specific way readings go wrong.
Calibrate before the session, not once a season. Distilled water should read zero on the moisture scale — full water, zero solids. If it does not, adjust with the calibration screw and re-check. A prism that has drifted two tenths of a point will misreport every sample taken that day, and the error is invisible without the water check.
Bring the sample to 20 °C. The scales are referenced to 20 °C. Instruments with automatic temperature compensation correct within a limited band around that, and the correction assumes the sample itself is at the measured temperature, not the room. A sample pulled from a cool store and read immediately will not be at 20 °C even if the room is. Leave the sample to equilibrate, or use a bench instrument with a temperature-controlled prism.
Handle crystallised honey properly. A set sample cannot be read as it stands. Warm it gently until the crystals dissolve, mix thoroughly, then cool it back to 20 °C before reading. Reading warm honey gives a falsely low moisture figure, and reading partly dissolved honey gives a figure that depends on where the drop came from.
Use enough sample and no air. Two or three drops, spread across the prism with the cover closed, no bubbles. Air gaps read as dryness. Excess honey that squeezes out is fine; a partial film is not.
Read the boundary, not the middle of the shade. The dividing line between the light and dark fields is the reading. On an optical instrument, hold it toward a consistent light source and take the line at the centre of the field to avoid parallax. Take three readings from three fresh drops and record the spread; if the spread is wider than a couple of tenths, the sample is not homogeneous and needs more mixing.
Clean the prism between samples. Honey is sticky and residues build a film that changes the reading. Distilled water and a soft lint-free cloth, then dry. Do not use paper that sheds fibres, and do not scrub an optical prism with anything abrasive.
A bench Abbe refractometer with a flow-through cell removes most of the operator variability, at a price that only makes sense when readings are part of a documented quality system. For a producer taking a handful of readings a week, a well-calibrated handheld digital instrument is enough — provided the calibration and temperature steps are followed every time.
Sample the lot, not the jar
Moisture varies inside a single container. It varies more between containers in the same batch, and most of all between apiaries and between harvest dates. A reading taken from the top of a drum after settling describes the top of that drum.
For a bulk lot, take increments from more than one depth after the contents have been mixed, combine them, and read the composite. Where a lot is made up of several drums, sample each one and report the range as well as the mean, because the standard is applied per sample unit. Where a batch is going into a contract with a maximum, the highest unit is the one that decides whether the lot passes.
Retain a sealed sample of every lot you test. Moisture migrates through some plastics and through imperfect closures, so a retained sample stored in a sealed glass or rigid food-grade container, kept with the reading and the date, is the only evidence available if a dispute arrives months later.
Moisture bands and what they mean commercially
The same reading means different things depending on where it sits. The bands below are the ones worth writing into an internal procedure, because they translate a number into an action.
| Reading | Commercial position | Practical action |
|---|---|---|
| Below 17.0% | Long-keeping; suitable for any market and for extended storage | Hold in sealed tanks; protect from humidity; no drying needed |
| 17.0–18.0% | Inside the common contract band and inside Grade A/B moisture | Ship as normal; verify with a retained sample |
| 18.0–18.6% | Meets the US grade ceiling but has no margin for a re-test or a warm transit | Move quickly, keep cool, avoid blending upwards |
| 18.6–20.0% | Outside Grade A/B; inside Codex and EU general limits; fermentation risk is real | Blend down with drier stock or dry before sale; disclose the figure to the buyer |
| Above 20.0% | Outside Codex and EU limits for table honey | Dry the lot, or sell only where a specific allowance exists, such as baker’s honey within its own labelling rules |
Two of those actions need a caveat. Blending is legitimate when it produces a homogeneous lot that genuinely meets the specification, and it is fraudulent when it is used to move water past a limit without the buyer knowing the composition. Drying is legitimate when it is done gently: heat is the enemy of the other parameters a buyer will test, and the same lot that fails on moisture can fail on HMF or diastase if it is warmed hard to fix the first problem. A warm, dry, moving-air room and shallow layers take longer than a heater, and they leave the enzyme and HMF figures intact.
Where the water gets in
Extraction is the step where moisture is decided. Frames pulled before the cells are capped carry nectar that has not been through the final evaporation stage, and no amount of care afterwards puts the missing days back. The commonsense check — shake a frame horizontally and see whether anything flicks out — is worth doing before the uncapping knife comes out, even when the refractometer is in the room.
After extraction, the exposure is mostly about surfaces and air. Open-top settling tanks in a humid room gain moisture across the whole exposed surface overnight. Pumps that whip air into the honey create foam that reads inconsistently and holds bubbles in the finished jar. Strainers left half-full in a warm room are a large wetted surface with a thin film of honey on it, which is close to an ideal moisture-exchange setup.
The controls are unglamorous: lids on tanks, minimum headspace, a dehumidifier running in the extraction and filling rooms during the season, strainers emptied promptly, and transfer pumped rather than poured. Equipment choices make the discipline easier. A stainless steel storage tank with a close-fitting lid gives you a sealed volume instead of an open one, a double-layer stainless strainer clears wax fast enough that the honey is not sitting in it, and a filling machine shortens the window in which the finished product is exposed in an open hopper.
Records that hold up
Every serious buyer eventually asks for evidence, and the ask arrives long after the batch has gone. What a food-safety auditor or a customer’s technical team wants to see is a lot record that ties a moisture figure to a specific container, a specific instrument and a specific date.
The minimum that satisfies that request: lot or batch identifier, apiary and harvest date, container or drum numbers, instrument identifier and its last calibration date, the reading, the sample temperature, the operator, and the retained sample reference. Where a specification exists, the record should also show the limit and pass or fail against it, because an auditor reading a bare number cannot tell whether it was judged against 18.0 or 20.0.
Digital instruments that log readings remove most of the transcription risk, and where they are not available, a paper log filled in at the tank is still far better than numbers reconstructed from memory at shipping time.
The parameters that move with it
Moisture is rarely tested alone. A buyer’s certificate of analysis usually carries four or five figures, and the way you fix moisture affects the others. Knowing which ones are linked prevents the common outcome where a lot is dried into compliance on one parameter and pushed out of compliance on another.
HMF. Hydroxymethylfurfural forms as honey is heated or stored warm, and the Codex ceiling is 40 mg/kg, with an allowance to 80 mg/kg for honey of tropical origin. Because drying is usually done with warm air, every hour at temperature spends part of the HMF budget. A lot that arrives at the dryer at 14 mg/kg has room; a lot that has already been warmed for filtering does not.
Diastase activity. The enzyme figure is the other heat marker, with a Codex minimum of 8 Schade units and a lower allowance where HMF is itself low. Diastase falls as temperature and time increase, so it tracks HMF closely. Gentle drying protects both; aggressive drying breaks both at once, and the pairing is what makes the failure obvious to a laboratory.
Sucrose and reducing sugars. These describe maturity rather than handling. They are not moved by drying, which is why a lot that reads high on sucrose cannot be rescued downstream. If the sugar profile is wrong, the problem is in what went into the supers or when the crop was taken, not in the tank.
Water activity. The parameter that actually governs yeast behaviour is rarely on a routine certificate, because it needs a different instrument. Where a buyer specifies it, the figure that matters is measured on the finished product, not on the bulk sample, and crystallisation during storage can move it after the batch has been signed off.
The practical consequence is a sequencing rule: decide the moisture target with the other parameters already in hand, and set the drying method from the tightest of them. Where a lot is close to the HMF ceiling, time and air do the work and heat does as little as possible.
Mistakes that cost the most
The failures below account for most of the disputes and downgrades in bulk honey trade. None of them is exotic, and all of them are avoidable.
| Mistake | What it does to the result | Fix |
|---|---|---|
| Reading one jar and calling it the lot | Hides the worst unit, which is the one the standard and the contract measure | Sample each container, report the range, judge on the maximum |
| Reading honey straight out of a cool store | Sample is below the 20 °C reference; compensation is only valid near it | Equilibrate the sample before reading |
| Reading set honey without dissolving the crystals | Reading depends on where the drop landed; liquid phase is wetter than the bulk | Warm gently, mix, cool to 20 °C, then read |
| Using a Brix refractometer and subtracting from 100 | Drift that gets worse exactly where the decisions are made | Use a honey-scaled instrument or the maker’s conversion table |
| Skipping the distilled-water check | A drifted prism misreports every sample that day | Zero the instrument at the start of every session |
| Cutting corners on capping to hit a schedule | Locks in high moisture that no later step can remove cheaply | Test the frame, not the calendar |
| Leaving honey in open tanks overnight | Gains moisture across a large exposed surface in a humid room | Lid, minimum headspace, dehumidify |
| Drying with high heat | Fixes moisture and breaks HMF and diastase | Warm air and shallow layers, not hot surfaces |
The first and fourth are the expensive ones. A supplier who tests thoroughly but reads a Brix scale is worse off than one who tests roughly with the right instrument, because the first has documentation that will not survive re-testing. Consistency between your figure and the buyer’s figure is the whole point of the exercise; a number that cannot be reproduced at the other end of the shipment is not a measurement, it is an opinion.
It is also worth writing the test method into the specification, not just the limit. “Maximum 18.0 percent moisture, determined refractometrically at 20 °C on a composite sample per container” removes the argument about method before it happens, and it is the sort of detail that makes a small supplier look like a safe one to a technical buyer.
For the underlying limits, the primary sources are the Codex Alimentarius standards list, the European Commission’s honey market and composition pages, and the US grade standard itself in 7 CFR Part 52, Subpart E. On the measurement side, NIST’s Office of Weights and Measures publishes the calibration and traceability material that sits behind any instrument worth trusting.
On the equipment side, the honey pail and filter range covers the handling stages where moisture is won or lost, from settling and straining through to draw-off into smaller containers. Buyers who need a sealed bulk volume before shipment usually pair a stainless honey pail with a storage tank rather than decanting from an open vessel.