Monofloral Is a Guess With a Number On It

Pick up a jar that says acacia and you are holding a small, unspoken claim. The word implies a flower — black locust, Robinia pseudoacacia, the cream-colored bloom that hangs off roadside trees for about ten days in late spring — and it implies that the honey in the jar came mostly from that flower and not from the clover, rape, and bramble flowering in the same fields at the same time. It implies, in other words, a measurement. Someone, somewhere, is supposed to have counted.

The reassuring part is that a method for counting exists, and has for almost fifty years. The unsettling part is that the method disagrees with itself depending on which flower you ask it about — and that whether anyone runs it at all depends almost entirely on which side of the Atlantic you bought the jar.

This is a column about what labels say. Today's word is monofloral, and it is doing more work than it can support.

The forty-five percent rule

The science of reading a honey's floral origin under a microscope is called melissopalynology, and its governing text is older than most of the people buying honey: Louveaux, Maurizio and Vorwohl, "Methods of Melissopalynology," published in Bee World in 1978 for the International Commission for Bee Botany. Bees carry pollen incidentally as they work nectar; that pollen settles into the honey; you dissolve a measured amount, spin it down, and count grains under magnification — at least 300 for a rough look, 500 to 1,000 to trust the proportions.

What you get is a ranking. The classical convention sorts the grains into tidy bands. A predominant pollen is one that makes up more than 45 percent of everything you counted. Below that sit the secondary pollens (16 to 45 percent), then the important minor types (3 to 15), then the minor ones (under 3). To call a honey unifloral — monofloral, single-origin, the word on the jar — the convention says one pollen type should clear that 45 percent line and become predominant.

It sounds clean. A number, a threshold, a yes or no. The problem is that the number is a lie of convenience, and the people who wrote it knew it.

Where the rule falls apart

Pollen is not a neutral witness. Different flowers load their nectar with wildly different amounts of it, so the count in the jar is a distorted reflection of the field. Some flowers barely shed into their nectar at all; others bury the sample in pollen they contributed nothing else to.

Consider citrus. Orange-blossom honey is one of the most recognizable monofloral honeys on earth, and citrus is a notorious under-representer: a genuine, unmistakable citrus honey can show anywhere from 2 to 42 percent citrus pollen. By the clean 45 percent rule, almost every real orange-blossom honey on the shelf would fail to qualify as what it plainly is. The same trap catches lavender (Lavandula × intermedia, as little as 1 percent), alfalfa (1 to 10), and black locust — the acacia we started with — which the harmonized tables list anywhere from 7 to 60 percent, with the working trade rule treating roughly 20 percentRobinia pollen as "pure."

Now turn the problem over. Chestnut honey is the opposite kind of liar. Chestnut sheds pollen so aggressively that it can occupy more than 86 percent of the grains in a sample even when the bees were working other flowers heavily; eucalyptus, similarly, routinely tops 83 percent; rapeseed and phacelia clear 60. A honey can read as overwhelmingly "chestnut" under the microscope and still not be a chestnut honey in any meaningful sense.

The 2004 harmonization of the method — Von Der Ohe and colleagues, in Apidologie — handles this the only way it honestly can: by abandoning the single number. It publishes correction tables, and it instructs the analyst, in plain language, to do a second count that throws out the over-represented pollens— it names forget-me-not, chestnut, eucalyptus specifically — and recalculate. Which means the real answer to "is this monofloral?" is not a threshold at all. It is a trained human being looking at a slide and making a judgment about which flower is lying to them and by how much. The harmonized method says so itself: its single greatest weakness is "the correctness of pollen identification and the subsequent interpretation." Translation: it depends on who is holding the microscope.

So two honeys can sit side by side, each showing a 30 percent pollen reading for their headline flower, and one can be an impeccable monofloral while the other is nothing of the kind — and the difference is not in the jar but in the species, and in the analyst's experience with its habits.

There is also a cruder problem. Pollen is a physical particle, and physical particles can be filtered out. Ultrafiltration— pushing honey through a fine membrane under pressure — strips the pollen entirely, and with it any possibility of a pollen count. The single most established method for proving a honey's origin is also the one most easily erased, and erasing it is not difficult.

The chemistry that's supposed to back it up

If pollen is unreliable, surely the lab can fall back on chemistry. It can — but most of the chemistry on a honey's spec sheet is measuring the wrong thing.

The numbers that regulators actually enforce are quality and freshness gates, not identity proofs. Electrical conductivity below 0.8 millisiemens per centimeter marks a blossom honey; above it, you are likely looking at honeydew or chestnut. Hydroxymethylfurfural — HMF, a marker of heat and age — is capped at 40 milligrams per kilogram (80 for declared tropical honey). Diastase, an enzyme that degrades when honey is overheated, must come in at 8 Schade units or higher. These tell you whether a honey is fresh and unbaked. They tell you almost nothing about which flower it came from.

There is exactly one varietal on earth with a real, codified, government-enforced chemical definition, and it is instructive precisely because it is so lonely. In December 2017, New Zealand's Ministry for Primary Industries made it illegal to export honey labeled mānuka unless it passed a five-part test: four chemical markers and one DNA marker. To qualify as monofloral mānuka, a honey must show 3-phenyllactic acid at 400 milligrams per kilogram or more, plus 2′-methoxyacetophenone, 2-methoxybenzoic acid and 4-hydroxyphenyllactic acid above their own floors, plus mānuka pollen DNA detected by qPCR. Drop the 3-phenyllactic acid between 20 and 400 and you fall into multifloral mānuka; drop below 20 and you cannot use the word at all.

The detail worth pausing on is the 2′-methoxyacetophenone threshold, which the regulator deliberately raised from 1 milligram per kilogram to 5. The change had nothing to do with biology. It was set there to make it arithmetically harder to stretch a small amount of real mānuka with cheaper honey and still clear the line. That is what a genuine authentication standard looks like: not a single number, but a panel of them, chosen partly to anticipate the specific ways people cheat.

Now notice that almost no other honey has anything like it. Tupelo — the swamp honey from the Apalachicola, prized because it barely crystallizes — is "authenticated" mostly by its sugar ratio (an unusually high fructose-to-glucose balance, around 1.5 to 1 against a more typical 1.09) and by a pollen count, both of which are suggestive rather than dispositive. Chestnut has a promising chemical fingerprint in kynurenic acid, which runs from trace levels in most honeys up into the hundreds or low thousands of milligrams per kilogram in real chestnut — but that is a research marker, not a legal one. For the overwhelming majority of varietal claims, there is no panel, no DNA check, no enforced threshold. There is a pollen count, an analyst's judgment, and the producer's word.

The test that can't see the modern fraud

Set varietal identity aside for a moment and ask the blunter question: is there even honey in the jar, or has it been cut with syrup? Here the headline method is AOAC 998.12, the C4 sugar test, and it rests on an elegant piece of botany. Most nectar plants use what's called C3 photosynthesis; corn and sugarcane use C4, and the two pathways lay down measurably different ratios of carbon isotopes. Compare the carbon signature of the whole honey against its own protein fraction, and if they diverge by one part per thousand or more, you can infer the honey holds more than 7 percent added C4 sugar. For decades this was the workhorse.

The trouble is that fraudsters read the same literature. The cheap syrups of the moment are made from rice, wheat and sugar beet — and rice, wheat and beet are all C3 plants. Their carbon signatures sit right on top of genuine C3-nectar honey. The C4 test does not catch them because there is, isotopically, almost nothing to catch. The dominant purity test in global trade is structurally blind to the dominant adulterants in global trade, and has been for years.

This is the backdrop for the number you may have seen in headlines: that 46 percent of imported honey is fake. It deserves a closer reading than it usually gets.

The forty-six percent, read carefully

In 2021 and 2022, under a coordinated action the EU called "From the Hives," investigators pulled 320 consignmentsat the bloc's borders and sent them to the Joint Research Centre in Geel, Belgium. When the results came out in March 2023, the finding was stark: 147 of the 320 — 46 percent — were "suspected" of adulteration with sugar syrups. Broken out by origin, the proportions were worse: 74 percent of the consignments from China, 93 percent of those from Turkey, and, eyebrow-raisingly, 100 percent of the ten samples that arrived via the United Kingdom — the last almost certainly because Britain had become a re-blending and re-export waypoint rather than a source. The economics make the temptation legible. Imported honey averaged 2.32 euros a kilo; rice syrup ran 40 to 60 cents.

It is a damning figure, and it is also a more fragile one than the headline admits — a fragility flagged not by the honey industry's lobbyists but by the EU's own scientists. The 46 percent did not come from the accredited isotope methods. It came from a newer, qualitative screen on a high-resolution mass spectrometer, looking for the presence of marker compounds — mannose, a thing called difructose anhydride III, an oligosaccharide signature — rather than measuring how much syrup was there. A sample was flagged if at least one markerturned up. The Commission's own Q&A concedes the methods sat "outside the scope of accreditation." And the JRC's own follow-up paper, Paiano and colleagues in Separations in 2025, admits the inconvenient fact that one of those markers, mannose, can occur naturally in linden, eucalyptus, jujube and chestnut honeys.

Britain's food-safety reviewers went further. In February 2025, a Defra and FSA working group recommended retiring the word "suspicious" in favor of "flag for further investigation," on the grounds that "the term suspicious is generally not considered as proof." They assigned the mannose and oligosaccharide tests low evidential weight. And they made one observation that quietly reframes the whole panic: a syrup level below 5 to 10 percentis unlikely to reflect deliberate fraud at all, and "may in fact reflect carryover of sugar and associated markers from bee feed" — that is, from the sugar water beekeepers legitimately feed their colonies in lean months.

None of this means the honey supply is clean. The same EU exercise, run in 2015–2017 with the older accredited methods, still found roughly 14 percentnon-compliant across some 2,300 samples, and the Commission is careful to say the jump to 46 reflects a more sensitive net as much as a dirtier ocean. The point is narrower and more useful: "46 percent suspected" is a screening result under a contested method, not a verdict. The honest version of the sentence is that almost half of sampled imports tripped a sensitive, imperfect alarm — and that the people who built the alarm are the first to say so.

The fraud underneath the statistics is real enough when it is proven the hard way. The American case still cited as the template is Project Honeygate, which unwound a German trading house, Alfred L. Wolff, that disguised Chinese honey's origin to dodge antidumping duties running "as high as 221 percent." Seven defendants were charged; investigators put the evaded duties at more than 180 million dollars. Two of the implicated packers, Groeb Farms and Honey Holding, paid fines of two million and one million dollars and signed deferred-prosecution agreements in 2013. That is what adulteration looks like when it is established by documents and forensics rather than by a single tripped marker — slower, narrower, and far more certain.

The gap you can see from space

Which brings us back to the jar, and to the single fact that should change how you read every varietal label you own: whether any of this machinery applies to you depends almost entirely on where you are standing.

In the European Union, honey has a legal definition, enforced compositional limits, and — as of a directive that applies from 14 June 2026 — a new requirement that blends name their countries of origin in descending order with the percentage from each, rather than hiding behind the old "blend of EU and non-EU honeys." The bloc has stood up a Honey Platform of ninety technical experts to harmonize exactly the authentication methods this column has been picking apart. The system is flawed, contested, and slow. But it exists, and it is getting sharper.

In the United States, there is no federal standard of identity for honey at all. The American Beekeeping Federation petitioned the FDA for one in 2006; the agency denied the petition in 2011, calling a standard "unnecessary," and has since offered only non-binding labeling guidance. USDA grades exist, but they measure quality — clarity, moisture, absence of defects — not identity or purity, and by the government's own admission they "do not provide the means to enforce against intentional circumvention or adulteration." A few states, Utah among them, have written their own standards into the vacuum. Federally, a varietal claim on an American jar is enforced mainly by the general prohibition on outright lying, which is a low bar and a hard one to clear a case under.

So when an American jar says "monofloral wildflower" or "pure clover" or "raw acacia," the word is, in the legal sense, mostly on its honor. No agency is running the pollen count. No panel of four markers and a DNA probe is standing behind it. The claim rests on the producer's integrity and the importer's paperwork — which, as Honeygate demonstrated, is precisely the thing that can be forged.

What would actually close the gap

The emerging hope is DNA. Metabarcoding — sequencing the plant DNA in a honey rather than eyeballing its pollen — can detect flowers that microscopy misses and resolve them to species; one study pulled 926 plant species out of honey samples and used a subset to pin geographic origin with better than 90 percent accuracy. It is genuinely powerful for the question of what is present.

It is much weaker at the question that matters for a varietal label: how much. DNA metabarcoding is not quantitative— the share of genetic reads does not track the share of nectar. A 2015 study found no significant correlation between the microscope's pollen percentages and the sequencer's read percentages, with relative deviations running from 39 to 84 percent. It misses whole plant families whose DNA doesn't bind the standard primers. And it can only recognize what's already in its reference database. It is a better witness to presence and a worse one to proportion, which is the opposite of what "monofloral" needs.

So watch for three things, because any of them would move this from honor system toward enforcement: a single validated reference method for syrup detection that the C3 fraudsters can't walk around; the EU publishing actual per-varietal pollen thresholds as binding law under its new directive, replacing the analyst's judgment with a number that finally accounts for citrus and chestnut; and — least likely, most consequential — an American standard of identity that gives the word on the jar something behind it.

Until then, "monofloral" remains what it has always quietly been: a guess with a number on it. The guess is often a good one, made by serious people using a real method. But it is a guess that changes with the flower, contradicts itself across the correction tables, vanishes under a filter, and, in most of the United States, answers to no one. The word on the jar is a measurement the jar is hoping you won't ask to see.

Ask anyway. ◆