Short answer: the limit that decides whether your sesame, spice or dried fruit is accepted is the destination's legal limit for that exact product form, not the Codex figure or the origin laboratory's certificate. The result depends more on how the lot was sampled than on which laboratory ran the test. Aflatoxin contamination is patchy, so a few bags can read clean while the lot fails an official plan built from dozens of increments. Aflatoxin testing protects you only when the sampling plan, method, laboratory and governing result are named in the contract before loading.
The contaminants that cause trouble in these three categories
- Mycotoxins. Aflatoxins (B1 and the sum of B1, B2, G1 and G2) in oilseeds, spices and dried fruit; ochratoxin A (OTA) in dried vine fruit, figs and spices. Both come from moulds favoured by slow drying or damp storage.
- Pesticide residues. In the EU a default maximum residue level of 0.01 mg/kg applies where a pesticide is not specifically mentioned (European Commission, EU legislation on MRLs).
- Other contaminants and microbiology. The EU contaminants regulation also covers plant toxins, metals and processing contaminants (EUR-Lex, Regulation (EU) 2023/915); Salmonella in sesame is a separate microbiological test.
The EU's RASFF Window, a public database of notification summaries from 2020 onwards, shows what has been notified for a product and origin (European Commission, RASFF). The focus here is aflatoxin.
Why limits differ, and why Codex is a reference rather than the law
Codex publishes maximum levels and sampling plans in its General Standard for Contaminants and Toxins in Food and Feed (CXS 193-1995). It defines a Codex maximum level as the concentration "recommended by the CAC to be legally permitted" (Codex Alimentarius, CXS 193-1995). It binds nobody until a country adopts it, or its own figure, into law. National limits differ in substance measured, product definition and whether the product is ready to eat or still to be sorted.
Entries rechecked against the current EUR-Lex consolidation and Codex standard on 14 October 2026; a comparison, not a compliance list:
| Jurisdiction and text | Product | Limit (µg/kg) |
|---|---|---|
| EU, Reg. (EU) 2023/915, Annex I entry 1.1.4 | Groundnuts and other oilseeds to be sorted or physically treated before sale or use | B1 8.0; sum 15.0 |
| EU, entry 1.1.5 | Groundnuts and other oilseeds used as only ingredient, or processed products, for the final consumer or as an ingredient | B1 2.0; sum 4.0 |
| EU, entry 1.1.1 | Dried fruits to be sorted or physically treated (except dried figs) | B1 5.0; sum 10.0 |
| EU, entry 1.1.14 | Dried Capsicum, pepper (Piper), nutmeg, turmeric and mixtures containing them | B1 5.0; sum 10.0 |
| EU, entry 1.2.1.1 (ochratoxin A) | Dried vine fruits and dried figs | 8.0 |
| EU, entry 1.2.17 (ochratoxin A) | Dried spices except Capsicum | 15 |
| Codex CXS 193 (recommended, total aflatoxins) | Chili pepper, nutmeg | 20 |
| Codex CXS 193 (recommended, total aflatoxins) | Dried figs, ready-to-eat | 10 |
| US, FDA CPG Sec. 555.400 (guidance) | Human food other than Brazil nuts, peanuts, peanut products and pistachios | Total aflatoxins above 20 may be considered adulterated |
- The same spice can meet Codex and fail the EU. Codex recommends 20 µg/kg total aflatoxins for chili and nutmeg; the EU sum limit is 10.
- Product form changes the limit. In the EU, oilseeds still to be sorted carry 15 µg/kg for the sum; the same seed sold to the consumer or used as an ingredient carries 4. The EU text does not name sesame here; confirm with the competent authority which entry applies.
- Guidance is not a regulation. The FDA document states that it "is not binding on FDA or the public", and its enforcement criteria require both an original and a check analysis above 20 ppb (FDA, CPG Sec. 555.400).
In the EU, food listed in Annex I "shall not be placed on the market" where a contaminant exceeds the level set there (EUR-Lex, Regulation (EU) 2023/915, Article 2(1)). For the sum of aflatoxins, the Annex treats results below the limit of quantification as zero (lower-bound). GCC, Indian, African and other markets set their own limits; read the destination's current text for your exact product and form.
Heterogeneity: why a contaminated lot can produce a clean sample
For peanuts, Codex says that since there is no way to know whether contaminated kernels are uniformly dispersed, "it is essential that the aggregate sample be the accumulation of many small portions" drawn from throughout the lot (Codex Alimentarius, CXS 193-1995, Annex III). The EU sampling regulation treats groundnuts and tree nuts as products with heterogeneous contamination (EUR-Lex, Implementing Regulation (EU) 2023/2782).
No laboratory can measure seed that never entered the sample; origin and destination samples are different material, so two honest results can disagree.
Sampling plans and increment counts, and why the contract must name one
An official plan sets the number and weight of increments, how they are combined and the maximum lot size. In the EU, Commission Implementing Regulation (EU) 2023/2782 lays down the methods of sampling and analysis for official control of mycotoxins in food. It applies from 1 April 2024 and repealed Regulation (EC) No 401/2006 (EUR-Lex, Implementing Regulation (EU) 2023/2782). Examples rechecked on 14 October 2026:
| EU 2023/2782 part | Covers | Example row |
|---|---|---|
| Part II.A | Cereals, oilseeds other than groundnuts, and their products | Lot above 10 and up to 20 t: 60 incremental samples; aggregate 6 kg, or 1.5 kg for small-particle oilseeds |
| Part II.D | Groundnuts, apricot kernels, tree nuts, large-particle dried spices | Lot 15 to 125 t: sublots of 25 t, 100 increments, 20 kg aggregate, 2 laboratory samples |
| Part II.E | Dried spices | Lot above 0.5 and up to 1 t: 30 increments, 3 kg aggregate |
Part II.A sets an incremental sample of about 100 g, but about 25 g for oilseeds or grains where 1,000 seeds weigh less than 10 g. Have your inspector confirm the part and row.
Codex has its own plans for peanuts, tree nuts and dried figs. If the contract names no plan, each party uses the one that suits it and the evidence settles nothing. Name in the contract:
- the sampling plan by title and part (for example, the destination's official rules, or a named Codex annex);
- who samples, where and when (in the store, at stuffing, at discharge);
- how the aggregate sample is ground, mixed and split, and how many sealed portions are kept for each party;
- the method, the limit of quantification and the laboratory;
- which result governs, and the route for a second analysis.
Laboratory selection, accreditation, method and measurement uncertainty
Choose a laboratory accredited to ISO/IEC 17025, which sets requirements "for the competence, impartiality, and consistent operation of laboratories" (ISO, ISO/IEC 17025:2017), and check that aflatoxin in your matrix (sesame, chilli, dried fruit) is within the accredited scope. Since 1 January 2026, Global Accreditation Cooperation Incorporated has carried forward the work of ILAC and the IAF; the IAF site is now maintained only as a legacy reference (IAF, Global Accreditation Cooperation launch). Verify the accreditation body's current recognition status rather than relying on an old certificate or directory entry. For the wider evidence chain, see what an independent inspection certificate covers.
Ask for:
- Separate results for B1 and for the sum of B1, B2, G1 and G2, because limits are written both ways.
- The limit of quantification. "Not detected" reassures only when the LOQ is well below the destination limit.
- Recovery and expanded measurement uncertainty. Under the EU rules, a lot is rejected when the result, corrected for recovery where applicable, "minus the expanded measurement uncertainty" is above the maximum level (EUR-Lex, Implementing Regulation (EU) 2023/2782). A result slightly above the limit is not automatically a rejection; one slightly below is not a safe margin.
Testing at origin, at load port and at destination
| Point | What it protects | What it cannot do |
|---|---|---|
| Origin, before contract or at processing | Lets the seller sort, re-dry or divert a lot before it is committed | Speaks only for the material sampled at that time |
| Load port, sampled during stuffing | Links the result to what went into the container; can be a payment document | Cannot see changes in transit or bind the destination authority |
| Destination, official control | Decides admission | Arrives after freight has been paid and exposes the lot to detention costs |
Match the programme to the risk: test each lot from a new supplier or origin at stuffing with an official-style plan; test an established supplier less often, with the same plan so results stay comparable. Where payment depends on the result, the certificate wording belongs in the credit; see reading an LC before you accept it and the document set behind one shipment.
Some destinations add origin-specific controls: the EU applies increased official controls to certain goods from certain third countries under Implementing Regulation (EU) 2019/1793, amended many times (EUR-Lex, Implementing Regulation (EU) 2019/1793). Check its current annexes each season.
Mitigation at origin: what it achieves and its limits
The Codex code of practice for spices says drying to a water activity below 0.65 "is adequate to prevent mycotoxin production", that drying should start immediately after harvest and not on bare ground, that mouldy material should be sorted out, and that storage should be dry and ventilated (Codex Alimentarius, CXC 78-2017).
Its limits:
- Prevention stops new mould growth. It cannot undo existing contamination.
- Sorting lowers levels but does not certify a lot. The EU writes separate, higher limits for oilseeds and dried fruit still "to be subjected to sorting or other physical treatment".
- Storage and transit count. A lot shipped dry can pick up moisture in a warm container; condensation is a logistics problem as much as a quality one.
What to do with a detained consignment
- Get the official report: sampling record, plan, method, result, recovery and measurement uncertainty.
- Check the plan against the rules: increments for the lot size, and how the lot was defined.
- Consider a second opinion. In the EU, Article 35 of Regulation (EU) 2017/625 gives operators whose goods are sampled in official controls the right, at their own expense, to a second expert opinion: a documentary review of the sampling and analysis by another recognised, qualified expert. Where feasible, ask at sampling for enough material to be taken for it. A further analysis by another official laboratory exists only where the member state provides for it, and the request does not stop the authority acting on the risk (EUR-Lex, Regulation (EU) 2017/625, Article 35). Ask your customs broker at once.
- Understand re-testing. A new sample is a new draw from a patchy lot; a clean re-test does not prove the first result wrong.
- Count the clock. Storage and container charges accrue while the lot is held; see demurrage and detention explained.
- Decide within the options allowed. Ask which treatment or re-export routes the regime permits, and who pays under the contract.
Worked example: a sesame lot that passed at origin and failed at destination
Hypothetical teaching example. Names, figures, results and costs are invented. The 4.0 µg/kg figure is the EU entry 1.1.5 sum limit, used only for the arithmetic.
A trader sells 20 t of hulled sesame, 400 bags of 50 kg, at 1,450 CU per tonne (29,000 CU), for use as a food ingredient. Unknown to anyone, 20 bags (5%) come from a batch dried slowly and contain clusters of contaminated seed.
Origin. The supplier's staff take grab samples from 3 bags at the front of the stack. An accredited laboratory reports the sum of aflatoxins as 1.2 µg/kg.
Destination. The official inspector applies a plan like EU Part II.A for 10 to 20 t: 60 increments. The accredited destination laboratory reports 6.2 µg/kg with an expanded uncertainty of 1.6.
| Step | Calculation | Result |
|---|---|---|
| Chance that 3 bags miss all 20 affected bags | C(380,3) / C(400,3) | 85.7% |
| Chance that 60 bags miss all 20 affected bags | C(380,60) / C(400,60) | 3.6% |
| EU-style decision test | 6.2 − 1.6 = 4.6, above 4.0 | Exceeds beyond reasonable doubt |
| Cost of rejection | 1,100 storage and container charges + 420 second expert opinion + 2,600 re-export freight + 7,250 resale discount (25% of 29,000) | 11,370 CU |
| Cost of an official-style plan at stuffing | 450 sampling + 380 analysis | 830 CU |
The model is simplified: it treats each increment as coming from a different bag and ignores dilution in the aggregate. Neither laboratory erred: the small sample very likely missed the affected bags, the larger plan very likely found them. A matching plan at origin would have cost 830 CU and allowed the 20 bags to be sorted or diverted before loading.
Common mistakes
- Quoting a Codex figure, or the origin country's limit, in the contract instead of the destination's legal limit.
- Accepting "aflatoxin: complies" without B1, the sum, the LOQ and the limit applied.
- Leaving the sampling plan, the governing result and the second-analysis route out of the contract.
- Re-testing a detained lot until a clean result appears, as if the latest result cancelled the first.
Limits and what to recheck
Limits and sampling rules are amended. Before each contract, recheck the destination's current consolidated rules, any origin-specific controls and your laboratory's accredited scope. EU, Codex and US figures were rechecked on 14 October 2026; other markets differ. This is general information: consult an inspection provider, an accredited laboratory, the competent authority or a customs broker, and your lawyer on contract wording.
FAQ
What is the aflatoxin limit for sesame? It depends on the destination and product form. The EU sets different limits for "groundnuts and other oilseeds" still to be sorted and for those sold to consumers or used as an ingredient. Other markets set their own.
Why did my lot pass at origin and fail at destination? Usually because the samples were different material from a patchy lot, drawn under different plans.
Is a Codex maximum level legally binding? No. Codex describes its maximum levels as recommended to be legally permitted. A limit applies only once a country adopts it or its own figure.
How many samples are needed for aflatoxin testing? It depends on the plan and lot size. Under the EU rules, for example, a 10 to 20 t lot of oilseeds takes 60 incremental samples.
Next step
Before accepting the next lot, write a one-page testing instruction that names the destination limit, product form, sampling plan, laboratory scope, method, limit of quantification, sealed-sample custody and governing result. Give the same signed instruction to the supplier, inspector and laboratory before loading.
Resources
- EUR-Lex — Commission Regulation (EU) 2023/915 on maximum levels for certain contaminants in food, consolidated 19 August 2026 — https://eur-lex.europa.eu/eli/reg/2023/915/2026-08-19/eng (accessed 14 October 2026)
- EUR-Lex — Commission Implementing Regulation (EU) 2023/2782, methods of sampling and analysis for the control of mycotoxins in food, consolidated 24 March 2024 — https://eur-lex.europa.eu/eli/reg_impl/2023/2782/2024-03-24/eng (accessed 14 October 2026)
- EUR-Lex — Commission Implementing Regulation (EU) 2019/1793 on the temporary increase of official controls — https://eur-lex.europa.eu/eli/reg_impl/2019/1793/oj/eng (accessed 23 September 2026)
- EUR-Lex — Regulation (EU) 2017/625 on official controls — https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32017R0625 (accessed 23 September 2026)
- FAO/WHO Codex Alimentarius — General Standard for Contaminants and Toxins in Food and Feed, CXS 193-1995, last amended 2025 — https://www.fao.org/fao-who-codexalimentarius/sh-proxy/en/?lnk=1&url=https%253A%252F%252Fworkspace.fao.org%252Fsites%252Fcodex%252Fstandards%252FCXS+193-1995%252FCXS_193e.pdf (accessed 14 October 2026)
- FAO/WHO Codex Alimentarius — Code of Practice for the Prevention and Reduction of Mycotoxins in Spices, CXC 78-2017 — https://www.fao.org/fao-who-codexalimentarius/sh-proxy/en/?lnk=1&url=https%253A%252F%252Fworkspace.fao.org%252Fsites%252Fcodex%252FStandards%252FCXC%2B78-2017%252FCXC_078e.pdf (accessed 23 September 2026)
- US FDA — Compliance Policy Guide Sec. 555.400, Aflatoxins in Human Food — https://www.fda.gov/media/149666/download (accessed 14 October 2026)
- European Commission — Rapid Alert System for Food and Feed (RASFF) — https://food.ec.europa.eu/food-safety/rasff_en (accessed 23 September 2026)
- European Commission — EU legislation on MRLs — https://food.ec.europa.eu/plants/pesticides/maximum-residue-levels/eu-legislation-mrls_en (accessed 23 September 2026)
- ISO — ISO/IEC 17025:2017 General requirements for the competence of testing and calibration laboratories — https://www.iso.org/standard/66912.html (accessed 23 September 2026)
- ILAC — ILAC MRA and signatories — https://ilac.org/ilac-mra-and-signatories/ (accessed 23 September 2026)
- IAF — Global Accreditation Cooperation Incorporated launch unifies international accreditation organisations — https://iaf.nu/en/news/global-accreditation-cooperation-incorporated-launch-unifies-international-accreditation-organisations-and-strengthens-worldwide-trust/ (accessed 23 September 2026)
