Pollen analysis of honey: how melissopalynology works

Melissopalynology reads the pollen in honey to tell which plants the bees used and roughly where. Here is how the method works, the numbers behind it and what it cannot prove.

Blog · 12 min read · Published 28 September 2026 · Sources checked 28 September 2026

Pollen analysis of honey works by dissolving a weighed sample in water, spinning the pollen out in a centrifuge, and identifying and counting the grains under a light microscope. In the harmonised method published by members of the International Honey Commission, that means 10 g of honey and a count of at least 500 grains before any percentage is reported. The resulting pollen spectrum shows which plants the bees visited and hints at where they were.

Labs use it to check claims such as “lavender honey” or “chestnut honey” and to support, or challenge, a stated country or region of origin. It is not a purity test. Honey stretched with sugar syrup is caught with isotope ratio mass spectrometry, sugar profiling and NMR, and New Zealand’s official definition of mānuka honey relies on chemical markers plus a DNA test rather than a microscope count.

This guide covers the method, the numbers labs actually use, the legal texts behind them, and the limits that sales pages tend to leave out.

What is melissopalynology?

Melissopalynology is the branch of palynology that studies the pollen and other microscopic particles in honey. Its main job is to establish a honey’s botanical origin (which plants supplied the nectar or honeydew) and to help confirm its geographic origin. It sits alongside sensory and physico-chemical testing: the harmonised method itself recommends taking “sensory and physicochemical data” into account before naming a honey’s origin.

It is one of the branches of palynology, and you will find the short definition in our glossary entry on melissopalynology.

Why honey contains pollen

Bees do not add pollen to honey on purpose. It gets there as a side effect of foraging and hive life:

• From the flower. Pollen falls from the anthers into the nectar, so the nectar a bee collects already carries grains of that plant.
• From the bee. Foragers are dusted with pollen and some of it ends up in the nectar they bring home.
• From the hive. Stored pollen and wax debris can be mixed in when combs are uncapped and extracted.
• From the air. Wind-pollinated plants and fungal spores add a background that says little about the nectar source but something about the setting.

Honeydew honey, made from the sugary secretions of sap-sucking insects rather than nectar, usually contains less pollen. Analysts recognise it by “honeydew elements” such as fungal spores and hyphae and compare their number with the number of pollen grains (the HDE/PG ratio described in the harmonised method).

Why filtered honey loses its fingerprint

Because the pollen spectrum is the honey’s botanical and geographic signature, removing it removes the evidence. EU law defines “filtered honey” as honey obtained by removing foreign matter “in such a way as to result in the significant removal of pollen”, and both EU law and the Codex standard state that no pollen may be removed “except where this is unavoidable in the removal of foreign inorganic or organic matter”. A normally strained honey keeps its pollen; a finely filtered one may keep too little to analyse, and it must be labelled as filtered honey.

What pollen analysis can tell you

Botanical origin: unifloral vs multifloral

The analyst reports the relative frequency of each pollen type and sorts them into the classic frequency classes that go back to Louveaux, Maurizio and Vorwohl (1978):

The harmonised method states the rule plainly: a honey is considered to come predominantly from a given botanical origin (unifloral honey) “if the relative frequency of the pollen of that taxon exceeds 45%”. The same paper then explains why that rule cannot be applied blindly, which is the subject of the section on representation below. A multifloral (polyfloral) honey is simply one in which no single nectar source dominates.

Geographic origin

No single grain proves a country. What does the work is the association of taxa: a combination of cultivated crops, native trees and weeds that only co-occur in certain regions, often including the minor pollen types. A 2024 study of 87 honeys from French Guiana, for example, found pollen of Mimosa pudica in every sample, a plant that would be very odd to find in a honey sold as coming from northern Europe (Foods, 2024).

This is why a good lab needs a reference collection and regional reference data for the origin being claimed. It can say that a spectrum is consistent, or inconsistent, with a region; it cannot usually pinpoint one.

Signs of filtering or blending

Two other readings help flag problems:

• Absolute pollen content. Labs can count the grains per 10 g of honey. The harmonised method uses five classes, from Class I (up to 20,000 grains per 10 g) to Class V (over 1,000,000). An unusually low count can suggest filtration, but some unifloral honeys are naturally poor in pollen, so the figure has to be read against the honey type.
• Incoherent spectra. Pollen from plants that do not flower together, or that belong to regions far apart, points to blending of honeys from different origins.

How the analysis is done, step by step

The following summarises the harmonised qualitative method of von der Ohe et al. (2004), the reference most labs cite. Individual labs adapt it; ask yours which version it follows.

1. Homogenise the sample. Pollen is not evenly spread in a jar, so the honey is mixed before weighing.
2. Weigh and dissolve. 10 g of honey goes into a conical centrifuge tube with 20 mL of distilled water at 20–40 °C.
3. Centrifuge. 10 minutes at 1000 g; the supernatant is poured off, the sediment is rinsed with water and spun again (5 minutes at 1000 g).
4. Mount the sediment. The whole sediment is spread on a slide and mounted in glycerine jelly, clear or lightly stained with basic fuchsine.
5. Identify and count. Under a light microscope at 400–1000×, the analyst identifies grains against reference slides and atlases.
6. Reach a stable count. The paper asks for at least 300 grains for an estimate of relative frequencies and 500 to 1000 grains to determine them; only counts based on at least 500 grains should be expressed as percentages.
7. Interpret. The spectrum is compared with reference data for the claimed honey type and region, together with sensory and physico-chemical results.

Identification is the skilled part: many honey plants can only be separated to genus or “pollen type” level, not species. Our guide on how analysts identify pollen types explains why. For a quantitative analysis (grains per 10 g), the harmonised method dissolves the honey in more water and passes it through a membrane filter with 3 µm pores, then counts the elements on the filter surface.

Over- and under-represented pollen

A pollen percentage is not a nectar percentage. Some plants produce a lot of pollen per unit of nectar, others very little, and flower structure affects how much pollen reaches the nectar. Table I of the harmonised method lists the pollen shares typically found in well-characterised European unifloral honeys:

Two practical consequences follow:

• An under-represented honey can be genuinely unifloral below 45%. A citrus honey with, say, 20% Citrus pollen may be authentic, because citrus honeys typically contain 2–42%.
• An over-represented pollen can dominate without dominating the nectar. Chestnut honeys typically contain more than 86% chestnut pollen. A honey with 60% chestnut pollen is therefore not automatically a chestnut honey; it may be a blend in which chestnut contributed a minority of the nectar.

That is why labs interpret the count against reference data for each honey type rather than applying the 45% line mechanically, and why the same paper recommends confirming botanical origin with sensory and physico-chemical data.

Pollen analysis and the law

Pollen matters legally because it is how a label’s botanical and geographic claims can be checked. The key texts:

Note what these texts do not do. Neither sets a legal minimum pollen percentage for a unifloral name. The phrase “microscopic characteristics” is what brings pollen analysis in, while the thresholds for each honey type come from scientific reference data and, in some countries, national guidelines. If you sell honey, check whether your country has such guidelines before you make a floral claim.

The EU framework is also in motion: under the 2024 amendment, the Commission is due to adopt adulteration detection methods by 14 June 2028 and criteria on pollen removal and Union-wide traceability rules by 14 June 2029. Until then, Member States are pointed to internationally recognised validated methods, such as those approved by the Codex Alimentarius.

What pollen analysis cannot prove on its own

This is where honey marketing most often overreaches. Pollen answers questions about plants and places; it is weak on questions about added sugars.

Syrup adulteration

The EU’s 2021–2022 coordinated action tested 320 imported honey consignments and found 147 (46%) suspicious of breaching the Honey Directive because at least one marker of extraneous sugars was detected. The markers came from isotope, sugar-profiling and NMR methods, not from pollen. The report also notes that the older EA-IRMS test (AOAC 991.41) alone was not effective, which it links to a shift from maize and cane syrups towards rice, wheat or sugar-beet syrups (European Commission).

Mānuka honey

New Zealand’s Ministry for Primary Industries defines mānuka honey with “a combination of 5 attributes (4 chemicals from nectar and 1 DNA marker from mānuka pollen)”, tested by LC-MS/MS and by qPCR for Leptospermum scoparium DNA. Pollen still plays a role, but through its DNA, not through a microscope count.

How to get your honey tested

If you are a beekeeper, packer or buyer, a few points will save you time and money:

• Know your question. Botanical claim, origin check or suspected syrup? The last one needs a chemistry lab, not only a pollen lab.
• Send a homogeneous, sealed sample in a clean jar, labelled with the claimed floral source and region, harvest date and whether it was filtered or heated. The harmonised method uses 10 g per analysis; your lab will tell you how much it wants for repeats.
• Ask how many grains are counted, whether results are relative, absolute or both, and what reference data the interpretation uses for your honey type and region.
• Ask about accreditation and whether the lab takes part in proficiency tests.

We cover the full checklist, including red flags and what a report should contain, in how to commission a pollen analysis.

If you want to go deeper into identification yourself, the standard visual reference for the terms used in reports is Pollen Terminology: An Illustrated Handbook.

Frequently asked questions

Can pollen analysis prove honey is pure?

No. Pollen analysis shows which plants and regions a honey is consistent with. It cannot show whether sugar syrup was added: that is detected with isotope ratio mass spectrometry, sugar profiling and NMR, as used in the EU’s 2021–2022 honey control campaign.

How much honey does a lab need?

The harmonised method uses 10 g of honey per analysis. Labs usually ask for more so they can homogenise the sample and repeat the test if needed; check the amount with your lab before sending.

Does filtered honey have pollen?

Usually very little. EU law and the Codex standard define filtered honey as honey from which pollen has been significantly removed, and it must be sold under the name “filtered honey”. Ordinary straining to remove wax and debris is meant to leave the pollen in.

What is unifloral honey?

A honey that comes wholly or mainly from one plant source and has that source’s sensory, chemical and microscopic characteristics. As a rule of thumb, its pollen exceeds 45% of the count, but under-represented types such as citrus or lavender can qualify with less.

Can pollen analysis tell which country a honey comes from?

It can show whether the pollen spectrum is consistent with a claimed region, based on combinations of plants that grow together. It rarely pinpoints a country on its own, so labs combine it with other evidence and traceability records.

Sources

• von der Ohe W., Persano Oddo L., Piana M.L., Morlot M., Martin P. (2004). Harmonized methods of melissopalynology. Apidologie 35 (Suppl. 1): S18–S25. doi:10.1051/apido:2004050
• Louveaux J., Maurizio A., Vorwohl G. (1978). Methods of melissopalynology. Bee World 59(4): 139–157. doi:10.1080/0005772X.1978.11097714
• Melissopalynological analysis of honey from French Guiana (2024). Foods 13(7): 1073. doi:10.3390/foods13071073
• Council Directive 2001/110/EC relating to honey. EUR-Lex
• Directive (EU) 2024/1438 of 14 May 2024. EUR-Lex
• Codex Alimentarius, Standard for Honey CXS 12-1981. FAO/WHO (PDF)
• European Commission, EU coordinated action “From the hives” (honey 2021–2022). food.ec.europa.eu and JRC technical report JRC130227 publications.jrc.ec.europa.eu
• New Zealand Ministry for Primary Industries, “Ensuring mānuka honey is authentic”. mpi.govt.nz
• International Honey Commission. ihc-platform.net

This article explains scientific and regulatory background. It is not legal advice on labelling a specific product.