GMP+ 2026 Protocol: Aflatoxin B1 Monitoring in Maize

Exporting Ukrainian feed maize (corn) and its wet-milling or dry-grind processing co-products (distillers dried grains with solubles - DDGS, corn gluten feed, corn gluten meal, and maize germ cake) into the European Union feed supply chain requires strict adherence to the GMP+ framework (GMP+ Feed Certification scheme 2020, GMP+ FSA module - Feed Safety Assurance). The core technical standard governing mycotoxin surveillance in bulk grain corridors is Technical Specification TS 1.7 Monitoring (specifically the section Aflatoxin B1 protocol for maize and maize by-products), operating alongside TS 1.5 Specific Feed Safety Limits and TS 1.2 Purchase.

On 23 January 2026, a revised geographical risk classification for Aflatoxin B1 in maize took effect, published by GMP+ International in coordination with mutually recognized European feed safety scheme holders (including QS, OVOCOM, and AIC). Following an evaluation of thousands of verified analytical records uploaded by certified operators to the GMP+ Monitoring Database across the 2024-2025 harvest cycles, country-of-cultivation risk profiles were updated as follows:

• Ukraine: reclassified downward from Medium risk to Low risk;
• Romania: reclassified downward from High risk to Medium risk;
• France (Departments 71 Saône-et-Loire and 39 Jura): removed from the Southern France classification and grouped with the remaining French departments;
• Prior 2025 baseline revisions: United States (Low -> Medium effective 24 April 2025), Slovakia (Low -> Medium effective 4 September 2025), and Brazil alongside Southern France (Medium -> High and Low -> Medium adjustments effective 12 November 2025).

For quality assurance engineers, inland silo managers, port terminal operators, and commodity traders in the agriculture sector, Ukraine's reclassification into the Low risk tier allows a documented reduction in mandatory batch-testing frequency under TS 1.7. However, reducing sampling and analytical frequency without formally updating the site HACCP hazard analysis, revising Gatekeeper intake protocols, and maintaining strict chromatographic verification is treated by certification auditors as a Major Nonconformity.

Regulatory limits for Aflatoxin B1 ($\text{C}{17}\text{H}{12}\text{O}6$) are driven by its potent hepatotoxicity, teratogenicity, and genotoxic carcinogenicity (classified as a Group 1 human carcinogen by IARC). Unlike trichothecene mycotoxins such as deoxynivalenol (DON) or zearalenone (ZEA), which undergo partial microbial de-epoxidation in the rumen of cattle, Aflatoxin B1 ingested by lactating dairy ruminants is rapidly absorbed in the gastrointestinal tract and metabolized in the liver by cytochrome P450 monooxygenases (primarily CYP1A2 and CYP3A4) into its 4-hydroxy derivative: Aflatoxin M1 ($\text{C}{17}\text{H}_{12}\text{O}_7$).

The metabolic carry-over rate of Aflatoxin B1 from daily feed intake into raw milk ranges from $1%$ to $2%$ in mid-lactation cows, but climbs to $5% - 6%$ in high-yielding dairy cattle during peak lactation (producing over 35-40 liters of milk per day) due to increased blood-milk barrier permeability. Because European Union food safety legislation (Commission Regulation (EU) 2023/915) enforces a strict Maximum Permitted Level (MPL) for Aflatoxin M1 of $0.05\ \mu\text{g/kg}$ (50 ppt) in raw milk and $0.025\ \mu\text{g/kg}$ (25 ppt) in infant formulae, compound feed mills must maintain a wide safety margin.

Both EU Directive 2002/32/EC on undesirable substances in animal feed and GMP+ TS 1.5 Specific Feed Safety Limits establish tiered statutory thresholds referenced to a standardized feed moisture content of $12%$:

• Feed materials (including raw maize grain and maize processing by-products such as DDGS and gluten): the statutory Rejection Limit is $0.02\ \text{mg/kg}$ (20 ppb);
• Complete and complementary compound feed for dairy cattle (lactating cows, dairy sheep, dairy goats) and young animals of all species (calves, lambs, piglets, young poultry): the statutory Rejection Limit is $0.005\ \text{mg/kg}$ (5 ppb);
• Complete and complementary feed for other livestock categories (finishing pigs, adult poultry): $0.01\ \text{mg/kg}$ (10 ppb), or $0.02\ \text{mg/kg}$ for adult beef cattle and non-dairy sheep/goats as categorized in TS 1.5.

In commercial export practice, Dutch, German, and Belgian compound feed manufacturers purchasing Ukrainian maize for dairy rations routinely insert a tighter contractual ceiling into GAFTA No. 48, No. 64, or No. 78 contracts: Aflatoxin B1 in delivered raw maize must not exceed $0.0025 - 0.005\ \text{mg/kg}$ (2.5 - 5 ppb). If a maize barge arrives in Rotterdam containing $0.018\ \text{mg/kg}$ of Aflatoxin B1 (which is legally compliant as a raw feed material under the $0.020\ \text{mg/kg}$ ceiling), a dairy feed mill cannot include that maize at a $40%$ ration share because the maize alone would contribute $0.018 \times 0.40 = 0.0072\ \text{mg/kg}$, breaching the $0.005\ \text{mg/kg}$ legal limit for finished dairy compound feed.

Every numerical limit in GMP+ TS 1.5 and EU Directive 2002/32/EC is standardized to a reference moisture fraction of $12.0%$ (equivalent to $88.0%$ dry matter). Analytical laboratories typically report mycotoxin concentrations either on an as-is moisture basis ($W_{\text{actual}}$) or on a $100%$ dry matter basis. Before comparing a certificate of analysis against GMP+ limits or uploading results to the GMP+ Monitoring Database, quality engineers must apply the moisture normalization equation:

$$C_{12%} = C_{\text{measured}} \times \frac{100 - 12}{100 - W_{\text{actual}}} = C_{\text{measured}} \times \frac{88}{100 - W_{\text{actual}}}$$

For example, if wet newly harvested maize is delivered to an inland grain elevator at $W_{\text{actual}} = 21.0%$ moisture and the laboratory reports $C_{\text{measured}} = 0.0185\ \text{mg/kg}$ on the wet sample, normalizing to $12%$ moisture yields:

$$C_{12%} = 0.0185 \times \frac{88}{100 - 21} = 0.0185 \times 1.1139 = 0.0206\ \text{mg/kg}$$

Once dried in a shaft dryer to $14.0%$ commercial moisture and normalized to $12.0%$, the batch actually exceeds the $0.020\ \text{mg/kg}$ statutory Rejection Limit and must be quarantined, even though the unadjusted wet reading appeared below $0.020\ \text{mg/kg}$.

Furthermore, the GMP+ control system requires operators to enforce two distinct operational thresholds:

1. Action Limit: an early-warning intervention threshold (for instance, $0.015\ \text{mg/kg}$ for maize feed materials or $0.003\ \text{mg/kg}$ for dairy compound feed). Exceeding the Action Limit does not render the lot illegal, but obligates the operator to launch an internal root-cause investigation, intensify supplier monitoring frequency, and restrict the batch from entering dairy or young-animal feed formulations.
2. Rejection Limit (Maximum Permitted Level): the statutory ceiling ($0.020\ \text{mg/kg}$ for feed materials, $0.005\ \text{mg/kg}$ for dairy and young-animal feed). Any confirmed exceedance (accounting for expanded measurement uncertainty $U$ where applicable under compliance assessment rules) classifies the batch as non-compliant feed, triggering mandatory physical quarantine and the 12-hour Early Warning System (EWS) protocol.

A strict data-entry rule applies in the GMP+ Monitoring Database: all Aflatoxin B1 results must be entered exclusively in mg/kg. When a laboratory reports in ppb (µg/kg), the value must be divided by $1000$ ($5\ \text{ppb} = 0.005\ \text{mg/kg}$; $20\ \text{ppb} = 0.020\ \text{mg/kg}$). Uploading raw ppb values as mg/kg introduces a 1000-fold distortion into national risk analytics and triggers immediate compliance inquiries during audits.

Unlike field-borne Fusarium mycotoxins (Fusarium graminearum and F. verticillioides, which synthesize DON, T-2/HT-2, and fumonisins B1/B2 under cool, wet flowering conditions), aflatoxins B1, B2, G1, and G2 are secondary metabolites produced by thermotolerant xerophilic fungi: Aspergillus flavus and Aspergillus parasiticus. Understanding the ecophysiology of these fungi explains why TS 1.7 Monitoring ties mandatory testing frequency to both the Country of Cultivation and post-harvest storage control.

Pre-harvest colonization of maize ears in the field occurs during silking and kernel milk-dough stages when two stress factors converge:

1. Thermo-hydrological drought stress: daytime ambient temperatures exceeding $+30^\circ\text{C}$ to $+35^\circ\text{C}$ combined with soil moisture deficit induce pericarp micro-fissures and suppress plant phytoalexin defenses;
2. Entomological vectoring: ear feeding by larvae of the European corn borer (Ostrinia nubilalis) and cotton bollworm (Helicoverpa armigera), which carry Aspergillus flavus conidia directly into damaged starchy endosperm.

Even when maize is harvested in a Low risk country of origin (the tier assigned to Ukraine from 23 January 2026), explosive post-harvest Aflatoxin B1 synthesis can occur inside corrugated steel silos or ocean vessel holds during transit. The metabolic optimum for A. flavus toxin production in stored maize is defined by:

• Grain bed temperature: $+25^\circ\text{C}$ to $+37^\circ\text{C}$ (peak aflatoxin biosynthesis occurs between $+28^\circ\text{C}$ and $+32^\circ\text{C}$);
• Water activity ($a_w$): above $0.82 - 0.85$, which corresponds to an equilibrium maize moisture content of $14.5% - 15.5%$ at $+25^\circ\text{C}$.

When freshly dried maize is loaded into a large metal silo (24-32 meters in diameter) without sufficient cooling in the dryer recuperation zone (entering storage at $+22^\circ\text{C}$ to $+26^\circ\text{C}$), nighttime radiative cooling of the external galvanized steel walls drives convective air currents: warm interstitial air rises through the core of the grain pile and condenses against the cold headspace roof and northern wall. In this condensation zone, local grain moisture spikes from $13.8%$ to $17.5% - 19.0%$, creating a self-heating microbiological hotspot where Aflatoxin B1 concentrations can reach $500 - 2000\ \mu\text{g/kg}$ within 7 to 10 days. Consequently, during feed safety system implementation and integrated FSSC 22000 audits, automated multi-point silo thermometry and dew-point-controlled aeration are audited as critical operational controls.

The largest source of error in Aflatoxin B1 surveillance is not the mass spectrometer in the laboratory, but primary sampling at the grain elevator or port terminal. Because Aspergillus flavus colonizes grain in discrete localized pockets, mycotoxin distribution within a bulk maize lot is highly skewed (following a negative binomial distribution). A single heavily colonized maize kernel weighing $0.3\ \text{g}$ can contain up to $40,000\ \mu\text{g/kg}$ of Aflatoxin B1. If that single kernel is present among $10,000$ completely clean kernels (a $3\ \text{kg}$ sample), the true mean concentration of the entire $3\ \text{kg}$ unit jumps to $4\ \mu\text{g/kg}$ ($0.004\ \text{mg/kg}$).

Metrological studies by FAO and the European Commission Joint Research Centre (JRC) demonstrate that total measurement variance ($S_{\text{total}}^2$) partitions into three additive components:

$$S_{\text{total}}^2 = S_{\text{sampling}}^2 + S_{\text{subsampling}}^2 + S_{\text{analysis}}^2$$

Here, primary sampling variance ($S_{\text{sampling}}^2$) accounts for $75%$ to $85%$ of total uncertainty, sample comminution and mass division ($S_{\text{subsampling}}^2$) accounts for $10%$ to $15%$, and analytical instrument variance ($S_{\text{analysis}}^2$) represents only $2%$ to $5%$.

To overcome hot-spot heterogeneity, GMP+ TS 1.7 Monitoring prohibits grab sampling from truck tailgates or hatch covers and mandates representative sampling in strict accordance with Commission Regulation (EC) No 152/2009 (Annex I) and ISO 24333 (Cereals and cereal products - Sampling):

• Maximum Sub-Lot Size: large silo or vessel shipments must be divided into physical or virtual sub-lots (for example, $500\ \text{tonnes}$ per sub-lot for high-risk origins, or according to the TS 1.7 frequency matrix for medium and low-risk tiers);
• Incremental Samples: at least $100$ incremental samples (for sub-lots of $100 - 500\ \text{tonnes}$, minimum $100\ \text{g}$ each) must be collected systematically across the moving grain stream using an automated mechanical cross-cut sampler installed on the elevator leg discharge or conveyor belt, or via multi-aperture pneumatic probes penetrating the full depth of stationary railcars/trucks;
• Aggregate Sample Mass: all incremental samples from a sub-lot are combined into an aggregate sample weighing at least $10.0\ \text{kg}$.

A frequent critical finding uncovered during diagnostic audits at grain elevators and export terminals occurs during sample reduction: the sampler collects $100$ incremental samples into a $10\ \text{kg}$ sack, pours the whole unground maize kernels through a Boerner or Riffle divider to reduce it down to a $500\ \text{g}$ laboratory dispatch bag, and sends those $500\ \text{g}$ of whole grain to the laboratory.

Under Regulation (EC) No 152/2009 and GMP+ TS 1.7, reducing an aggregate sample in whole-kernel form invalidates the mycotoxin test report. Dividing $10\ \text{kg}$ of whole maize down to $500\ \text{g}$ carries a $95%$ statistical probability that the 2 or 3 contaminated kernels in the aggregate sample will be discarded into the rejected $9.5\ \text{kg}$ fraction.

Only two sample preparation workflows meet engineering compliance:

1. Complete Dry Grinding of the Entire $\ge 10\ \text{kg}$ Aggregate Sample: the entire $10\ \text{kg}$ aggregate sample is ground without prior mass reduction in a high-capacity cutting/hammer mill (such as a Retsch SM 300 or a combined grinding-dividing mill like the Romer RAS Mill) to pass a $< 1.0\ \text{mm}$ sieve (optimally $\le 0.85\ \text{mm}$), mechanically homogenized, and only then sub-sampled into a $1.0\ \text{kg}$ laboratory sample and a $25 - 50\ \text{g}$ analytical test portion.
2. Wet Slurry Homogenization: the entire $10\ \text{kg}$ maize sample is blended with water ($1:1$ or $1:1.5$ ratio) in a high-shear industrial slurry mixer until a uniform paste is achieved, reducing the sub-sampling coefficient of variation ($CV_{\text{subsampling}}$) below $5%$ while eliminating airborne spore and toxin dust.

Under GMP+ BA 10 Minimum Requirements for Inspection and Analysis and TS 1.2 Purchase, analytical certificates used to demonstrate compliance with TS 1.5 limits and populate the GMP+ Monitoring Database must be issued exclusively by laboratories that hold active GMP+ Registered Laboratory status and maintain accreditation under ISO/IEC 17025 from an ILAC/EA-recognized accreditation body (such as the National Accreditation Agency of Ukraine - NAAU) specifically covering Aflatoxin B1 in cereal/feed matrices.

Quality managers must distinguish between rapid intake screening and confirmatory reference chromatography:

1. Rapid ELISA and Lateral Flow Immunochromatographic Strips (LFD):

   • Operational role: fast 15-minute screening at elevator truck intake to segregate incoming farm deliveries.
   • Technical limitations: antibody-antigen binding is vulnerable to matrix interference from maize carotenoids (zeaxanthin, lutein), germ lipids, and ambient temperature shifts. Near the $0.005\ \text{mg/kg}$ (5 ppb) dairy threshold, rapid strips exhibit $\pm 30% - 40%$ variance and cross-reactivity with Aflatoxins B2, G1, and G2.
   • GMP+ compliance status: internal elevator strip results without ISO/IEC 17025 accreditation and GMP+ B11 Registered Laboratory registration cannot fulfill mandatory TS 1.7 monitoring requirements.

2. High-Performance Liquid Chromatography with Fluorescence Detection (HPLC-FLD per EN ISO 17375 / EN 15851):

   • Engineering mechanism: extraction with methanol/water or acetonitrile/water, cleanup through a monoclonal Immunoaffinity Column (IAC) to isolate Aflatoxin B1 from co-extracted matrix interferents, reversed-phase C18 separation, and post-column derivatization (electrochemical bromination via a Kobra Cell or UV photolysis in a PHRED reactor) prior to fluorescence detection.
   • Metrological performance: Limit of Quantification ($LOQ$) of $0.0005 - 0.001\ \text{mg/kg}$ (0.5 - 1.0 ppb), recovery between $80%$ and $110%$, and high baseline cleanliness.

3. Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS):

   • Engineering mechanism: positive electrospray ionization (ESI+) monitoring specific Multiple Reaction Monitoring (MRM) ion transitions (precursor ion $m/z\ 313.1 \rightarrow$ product ions $m/z\ 285.1$ and $241.0$), enabling simultaneous multi-mycotoxin quantification (Aflatoxins B1/B2/G1/G2, DON, ZEA, Fumonisins B1/B2, Ochratoxin A, T-2/HT-2).
   • Mandatory validation safeguard: when using Dilute-and-Shoot extraction without IAC cleanup, co-eluting phospholipids and peptides in maize, DDGS, and corn gluten cause severe electrospray ion suppression (up to $-40%$ signal loss), artificially under-reporting the true toxin level. Therefore, the ISO/IEC 17025 laboratory must apply either Matrix-Matched Calibration or spike an isotope-labeled internal standard ($^{13}\text{C}_{17}\text{-Aflatoxin B1}$) prior to extraction.

Primary agricultural producers (arable farms and agro-holdings in the agriculture sector) typically operate outside the scope of GMP+ certification, as the GMP+ chain begins at grain collection, drying, elevator storage, and trading. Instead, primary farms may hold sustainability certifications under ISCC (ISCC EU / ISCC PLUS) or good agricultural practice certification under GLOBALG.A.P..

When a GMP+ certified grain elevator, feed mill, or commodity trader purchases raw maize directly from an uncertified farm, the certified buyer acts as a Gatekeeper under Section 4 of GMP+ TS 1.2 Purchase. To legally bring farm-origin maize into the assured GMP+ FSA supply chain, the Gatekeeper must execute five engineering controls:

1. Contractual Supplier Qualification & Grower Declaration: execute a purchase contract accompanied by a signed Grower Declaration confirming compliance with statutory plant protection product (PPP) pre-harvest intervals and verifying that harvested grain has never come into contact with treated seed corn;
2. Grain Drying Hazard Assessment (HACCP): if the farm dried the maize on-site, the Gatekeeper must verify the dryer design (indirect heat exchanger vs direct-fired burner) and fuel type. Direct flue-gas drying using diesel, heavy fuel oil, wood chips, or sunflower husk pellets introduces a severe hazard of contaminating maize with polycyclic aromatic hydrocarbons (PAH4), dioxins, and dioxin-like PCBs;
3. Representative Batch Sampling per TS 1.7: collect a representative aggregate sample ($\ge 10.0\ \text{kg}$) per Regulation (EC) No 152/2009 from each incoming lot or consolidated silo bin prior to release;
4. Confirmatory Testing in a GMP+ Registered Laboratory: analyze the lot for Aflatoxin B1 under TS 1.7 alongside other HACCP-identified hazards (pesticide residues, heavy metals Pb/Cd/As/Hg, Fusarium mycotoxins DON/ZEA/Fumonisins, and PAH4/dioxins where direct drying was used);
5. Positive Release Quarantine: hold the purchased grain in Blocked / Pending QA status within the warehouse management system (WMS/ERP) until the verified certificate of analysis is reviewed and normalized to $12%$ moisture.

What happens when a GMP+ Registered Laboratory issues a certificate of analysis showing an Aflatoxin B1 concentration of $0.028\ \text{mg/kg}$ (28 ppb) at $12%$ moisture, exceeding the TS 1.5 Rejection Limit ($0.020\ \text{mg/kg}$) beyond expanded measurement uncertainty $U$?

This triggers the mandatory GMP+ TS 1.4 Early Warning System (EWS) protocol. From the moment the exceedance is confirmed, a strict 12-hour clock starts, requiring the certified operator to execute the following incident response workflow:

• Step 1 (Hours 0 - 2): Immediate ERP and Physical Quarantine. The affected silo bin, barge hold, railcar block, or flat-store bay is locked in the ERP system and physically sealed with "Non-Conforming Feed - Quarantine" signage. All loading, blending, or dispatch operations are halted immediately.
• Step 2 (Hours 2 - 4): Metrological Verification and Re-Testing Rules. Operators frequently attempt to draw a second sample from another section of the silo hoping for a lower reading. Under Regulation (EC) No 152/2009 and GMP+ rules, re-sampling a lot to overturn a positive mycotoxin result is prohibited: because aflatoxin forms discrete hotspots, a clean second sample does not disprove the contaminated hotspot captured in the first sample. The operator's sole legitimate verification step is to instruct the laboratory to perform an immediate duplicate re-analysis from the original ground and homogenized laboratory sample to rule out instrument drift or calculation error.
• Step 3 (Hours 4 - 8): Forward and Backward Traceability. The HACCP team reconstructs the full mass balance of the contaminated lot: identifying all farm deliveries that entered the silo, shared conveyor paths, and any partial shipments already dispatched to buyers.
• Step 4 (Within 12 Hours): Official EWS Notification Filing. Once the exceedance is confirmed, the company submits the mandatory EWS Notification Form simultaneously to:
  1. GMP+ International (ews@gmpplus.org);
  2. Its Certification Body (CB);
  3. The national competent authority: in Ukraine, the State Service of Ukraine on Food Safety and Consumer Protection (SSUFSCP / DPSP) pursuant to Article 20 of the Law of Ukraine No. 2042-VIII "On State Control over Compliance with Legislation on Food, Feed, Animal By-Products, Animal Health and Welfare" and the Law of Ukraine "On Feed Safety and Hygiene" (and, for shipments reaching EU ports, the relevant EU national authority such as NVWA in the Netherlands or BVL in Germany for RASFF alerting);
  4. Downstream Customers, if any portion of the lot has already left the facility.

Statutory Prohibition of Dilution (Blending Ban): both Article 5 of EU Directive 2002/32/EC and the GMP+ scheme strictly prohibit mixing a batch that exceeds a Maximum Permitted Level (e.g., maize at $0.030\ \text{mg/kg}$) with clean grain ($0.001\ \text{mg/kg}$) to bring the blended average below $0.020\ \text{mg/kg}$. Dilution of contaminated feed is an illegal adulteration offense that leads to immediate certificate withdrawal and regulatory prosecution. A contaminated lot may only be diverted outside the feed/food chain (such as industrial biogas anaerobic digestion or combustion) or routed to a detoxification facility officially authorized by an EU competent authority under Commission Regulation (EU) 2015/786.

Field experience from certification and surveillance audits supported by our consultants during audit support and pre-audit diagnostic assessments shows that even after Ukraine's transition to Low risk on 23 January 2026, feed operators frequently receive Major Nonconformities in five areas:

1. Unsynchronized HACCP Monitoring Plans: reducing testing frequency in 2026 without recording a formal HACCP team validation minute referencing the 23 January 2026 TS 1.7 revision;
2. Deficient Aggregate Sample Mass and Whole-Grain Division: sampling logs showing $2.0\ \text{kg}$ samples instead of the mandatory $\ge 10.0\ \text{kg}$ aggregate sample, or splitting whole maize kernels through a Riffle divider prior to grinding;
3. Using Non-Registered Testing Laboratories: submitting maize samples to a laboratory that holds national ISO/IEC 17025 accreditation but lacks active registration in the GMP+ Registered Laboratories database under BA 10;
4. Overlooking 3x Mycotoxin Concentration in Maize Co-Products (DDGS and Gluten): Aflatoxin B1 is thermally stable up to $+250^\circ\text{C}$ and is not degraded by Saccharomyces cerevisiae during ethanol fermentation. When starch (which makes up $65% - 70%$ of the maize kernel) is fermented into ethanol and distilled off, Aflatoxin B1 concentrates in the remaining protein, fiber, and fat fraction (DDGS and corn gluten) by a factor of $2.8\times$ to $3.3\times$. Consequently, incoming maize containing a compliant $0.008\ \text{mg/kg}$ of Aflatoxin B1 will yield DDGS at roughly $0.025\ \text{mg/kg}$, breaching the $0.020\ \text{mg/kg}$ statutory Rejection Limit;
5. Unit Errors in the GMP+ Monitoring Database: uploading ppb figures ($5.0$) instead of mg/kg ($0.005$) or omitting the Country of Cultivation field.

To safeguard export contracts and pass your upcoming surveillance or recertification audit without findings, schedule a technical verification of your sampling infrastructure, laboratory service agreements, and Gatekeeper procedures. Reach out to our engineering team via the contact page for a consultation, pre-audit diagnostic audit, or full GMP+ implementation.