Standard methods for Apis mellifera honey research

To cite this article: Ligia Bicudo de Almeida-Muradian, Ortrud Monika Barth, Vincent Dietemann, Michael Eyer, Alex da Silva de Freitas, Anne-Claire Martel, Gian Luigi Marcazzan, Carla Marina Marchese, Carla Mucignat-Caretta, Ana Pascual-Maté, Wim Reybroeck, M. Teresa Sancho & José Augusto Gasparotto Sattler (2020) Standard methods for Apismellifera honey research, Journal of Apicultural Research, 59:3, 1-62, DOI: 10.1080/00218839.2020.1738135 To link to this article: https://doi.org/10.1080/00218839.2020.1738135


Introduction
The honey chapter is divided into seven parts describing the main properties of honey as well as their methods of analysis. Each method is described step-by-step. The authors come from many countries such as Belgium, Brazil, France, Italy, Spain, and Switzerland, with many of the authors being members of the International Honey Commission. Honey is the most important product from the apiary and is consumed globally, making it economically important . As the only natural sweetener, honey historically was an important food for man. Honey was mentioned in many ancient cultures (the holy books of ancient India (the Vedas), book of songs Shi Jing (written in the sixth century BC in ancient China), the Holy Bible, just to name a few), and depicted in many wall drawings in ancient Egypt, Rome (mentioned by the writers Vergil, Varro, and Plinius) and also in medieval high cultures. In ancient Greece, the honey bee, a sacred symbol of Artemis, was an important design on Ephesian coins for almost six centuries (Bogdanov, 2011). Honey is a viscous and aromatic product, widely appreciated by man and prepared by honey bees (Hymenoptera: Apidae, Apis spp.) from the nectar of flowers or honeydew, the latter being produced through an intermediary, generally an aphid (Ouchemoukh et al., 2007;Almeida-Muradian et al., 2013;Ferreira et al., 2009) . This natural product is generally composed of a complex mixture of carbohydrates and other less frequent substances, such as organic acids, amino acids, proteins, minerals, vitamins, lipids, aroma compounds, flavonoids, vitamins, pigments, waxes, pollen grains, several enzymes, and other phytochemicals Almeida-Muradian & Penteado, 2015;Gomes et al., 2010;Lazarevic et al., 2010).
All methods related to honey research cited herein are based on the specialized literature, and the authors recommend reading the primary references such as Codex Alimentarius, Association of Official Agricultural Chemists (AOAC), and International Honey Commission publications. Herein, we include the main methods related to research with honey, from the hive to the final product consumed.
2. Honey authenticity-botanical, geographical, and entomological origin Pollen analysis is the initial technique used to determine the botanical origin of the honey produced by bees. This knowledge enables one to characterize the honey producing area by means of pollen grain identification. Apis and Meliponini bees show different preferences when foraging plant products.
The management of honey harvesting must follow quality procedures, aiming not only at its efficient collection, but mainly at maintaining its original characteristics and the quality of the final product. The beekeeper, when handling the harvest, must be wear appropriate clothing for beekeeping. It is recommended not to collect honey on rainy days or with high relative humidity, as it could lead to an increase in the moisture content of the product. Harvesting should take place selectively during the opening of the beehive. The beekeeper must inspect each frame, giving priority to removing only the frames that have at least 90% of their operculated storage cells, this being indicative of the maturity of the honey in relation to the percentage of moisture. In order to guarantee the quality of the final product, during the extraction of honey, all the equipment and utensils used in the various stages of handling must be specific to this activity, and there is no room for adaptation. In the case of equipment and utensils that will have direct contact with the product, all must be stainless steel, specific for food products.
Pollen analysis is the initial technique used to determine the botanical origin of the honey produced by bees. This knowledge enables one to characterize the honey producing area by means of pollen grain identification. Apis and Meliponini bees show different preferences when foraging plant products. For palynological analysis, honey samples of Apis have to be received in glass vials and kept at room temperature. In case of stingless bees, the honey has to be kept in a refrigerator at about 8 C to avoid fermentation.
For more details in sampling, please see "Sampling" sections.

Botanical origin
Pollen grains and other structured elements that compose honey sediments are indicative of the botanical origin (Barth, 1989;Vit et al., 2018). A complete palynological analysis reveals the real provenance of the raw material. Pollen grains are introduced into honey by different means (Barth, 1989). Besides the nectariferous plants, that secrete nectar in addition to producing pollen, the polliniferous plants mainly produce pollen grains and have low concentrations of nectar (Barth, 1989).

Pollen grains
The first universal technique aiming at the recognition and evaluation of the pollen in honey was elaborated by the French, Swiss, and Dutch specialists on honey analysis. This started with Zander's work in 1935, which was published by Louveaux et al. (1970aLouveaux et al. ( , 1970b and updated by Louveaux et al. (1978). The latter includes many standard methods, including the acetolysis method. A more recent publication on pollen analysis was published in von der Ohe et al. (2004).
These methodologies are used today. For the first technique, the pollen grain content is not removed, making it difficult to observe some of the structures. However, more of the elements occurring in honey sediments are preserved. The acetolysis technique enables a better visualization of pollen grain structures when only the outer pollen grain wall (exine) is preserved, but there is a significant loss of no resistant elements such as oils, fungi hyphae, yeast, insect fragments, and organic matter (Haidamus et al. 2019). A detailed analysis of the technique established by Louveaux et al. (1978) was presented also by von der Ohe et al. (2004) considering quantitative and qualitative pollen counting.

Other structured elements
Honey sediments may contain several types of particles besides pollen grains ( Barth, 1989;Zander & Maurizio, 1975). These structured elements, not nectar indicative, show what happens during honey sampling and storage, and reflect honey quality. Some adulterants include a variety of fungal spores, yeasts, colorless molds and air dispersed spores, green algae, bacteria, insect bristles and scales, and pathogenic mites, as well as starch and synthetic fibers.
(2) Add 20 mL of distilled water and dissolve well by stirring (10-15 min). (3) If the honey is crystallized, heat the mixture for a few seconds in a microwave oven to complete the dissolution process of the sugar crystals. (4) Divide the solution eventually between two conical centrifuge tubes of 15 mL capacity. (5) Centrifuge at least 252 RCF (relative centrifugal force or G) that, according to a centrifuge model, may comprise on average 1500 rpm (revolutions per minute), for 15 min. (6) Discard the supernatant and rapidly invert the tubes for approximately 30 s before returning to normal position. (7) Add 10 mL of distilled water to the pellet. (8) Centrifuge at 1500 rpm for 15 min. (9) Discard the supernatant and rapidly invert the centrifuge tubes for approximately 30 s before returning to normal position. (10) Add 5 mL of glycerin water (solution 1: 1 with distilled water and glycerin) to the pellet. (11) Wait for 30 min and then centrifuge at least 1500 rpm for 15 min. (12) Discard the supernatant and rapidly invert the tube onto absorbent paper (never return to the normal position) to allow all the liquid to drain. (13) Prepare two microscope slides. (14) Put a little piece of glycerin jelly (0.5 mm 3 ) onto a needle tip and collect part of the sediment into the centrifuge tube. (15) Place the sediment on the center of the microscope slide, heat gently on a hot plate, only to dissolve the glycerin jelly. (16) Cover with a coverslip, and seal with paraffin.
(17) The pollen grain identification is based on the reference pollen slides library from regional/country vegetation and with the aid of reference catalogs (e.g., Roubik & Moreno, 1991). (18) After the identification and palynological analysis, the prepared slides were stored in a pollen library.

The acetolysis technique
Acetolysis (Erdtman, 1952) is an artificial fossilization method to prepare pollen grains in order to get the best transparency of the grain envelope.
(1) Weigh 10 g of honey into a beaker.
(2) Add 20 mL of distilled water and dissolve well by stirring (10-15 min). (3) If the honey is crystallized, heat the mixture for a few seconds in a microwave oven to complete the dissolution process of the sugar crystals. (4) Divide the solution evenly between two centrifuge tubes of 15 mL capacity. (5) Centrifuge at least 1500 rpm for 15 min. (6) Discard the supernatant and rapidly invert the tubes for approximately 30 s before returning to normal position. (7) Add 5 mL of acetic acid (32%), centrifuge at least 1500 rpm for 15 min and discard the supernatant (8) Add the acetolysis mixture (4.5 mL of acetic anhydride (98%) and 0.5 mL of sulfuric acid (99%)). (9) Heat in a water bath just to 80 C for 3 min. (10) Centrifuge at least 1500 rpm for 15 min. (11) Discard the supernatant and add 20 mL of distilled water. (12) Wash twice. (13) Discard the supernatant and rapidly invert the tubes for approximately 30 s before returning to normal position. (14) Add 5 mL of glycerin water (solution 1: 1 with distilled water and glycerin) to the pellet. Wait for 30 min. (15) Centrifuge at least 1500 rpm for 15 min. (16) Discard the supernatant and rapidly invert the tube onto absorbent paper (never return to the normal position) to allow all the liquid to drain. (17) Prepare two microscope slides. (18) Put a little piece of glycerin jelly (0.5 mm 3 ) onto a needle tip and collect part of the sediment in the centrifuge tube. (19) Place the sediment on the center of the microscope slide, heat gently on a hot plate, only to dissolve the glycerin jelly. (20) Cover with a coverslip, and seal with paraffin.

Additional techniques
A number of other techniques to prepare honey samples were purposed and published. In general, they are complex to execute and require a broad assortment of equipment.

Honey sample evaluation
Since the standard work of Louveaux et al. (1978), the qualitative and quantitative evaluation of honey samples (Maurizio, 1977) has been a subject of many investigations until the data assembly in von der Ohe et al. (2004). The reliability of pollen analysis in honey samples was discussed in detail by Behm et al. (1996). Pollen coefficients were discussed by the R-values of honey pollen coefficients by Bryant and Jones (2001).

Super and subrepresentation of pollen types
For each region of honey production, there are plants that are significant nectar or/and pollen producers. The pollen grains present in the honey sediments can reveal the bee foraging behavior. In general, the number of pollen grains present on one microscope slide provides information to use to assign a frequency class, indicating if the pollen types observed in the honey samples are predominant (PP), accessory (AP), important minor pollen (IMP), or minor pollen (MP) (Louveaux et al., 1970a(Louveaux et al., , 1970b. Few pollen grains found in a honey sample suggests that the bees visited plant species of a low pollen production, but that have a high nectar disponibility like Citrus spp. and Croton sp. (Table 1, Figure 1). A high quantity of pollen grains from polliniferous plants (e.g., Poaceae and Melastomataceae) can indicate a super  representation since these species produce a lot of pollen but little or no nectar (Louveaux et al., 1978).

Honeydew
Bees can use the excreta of aphids (Aphidae) that feed on plant xylem to create honeydew. This bee product can be difficult to classify and presents a high electrical conductivity and mineral salt content. Generally, pollen grain occurrence is low and a moderate number of fungal spores can be observed (Barth, 1971;Maurizio, 1959) (Figure 1).

Geographical origin
Pollen grain analysis allows one to determine the nectar origin of honey. The assembly of plant species identified by pollen analysis provides information about honey's local or regional origin. A reference pollen slide collection (pollen library) must be available in order to compare the morphology of pollen grains obtained directly from flowers with the ones obtained in the honey sediment collection. The scope of this collection depends upon the study area and the radius of action of the bees. Plant assemblies recognized in the honey sediment by means of pollen grain morphology inform about the nectariferous and non-nectariferous, local and regional potential of honey production (Aronne & de Micco, 2010;Consonni & Cagliani, 2008;Salgado-Laurenti et al., 2017). Unfortunately, some honeys do not contain much pollen. In these cases, the honey origin is difficult to determine. It may derive from strong nectar secreting plant species such as those from the Asclepiadaceae. Sometimes, certain flowers secrete abundant nectar that will drip from the flower inflorescences, making it unlikely to contain pollen. This happens with Dombeya (Sterculiaceae, a shrub) and with some plants that grow in very humid environments.

Entomological origin
For purposes of this manuscript, we are discussing honey produced by bees of the family Apidae, genus Apis, species A. mellifera. There are several subspecies/races of A. mellifera occurring in different parts of the world (Carreck et al., 2013;Meixner et al., 2013). Humans have cultivated this bee since antiquity given that it has provided a source of sugar. There are other types of honey producing bees. Collectively called stingless bees (also Apidae), these bees occur mainly in the tribes Meliponini and Trigonini. Bees from both tribes produce a different type of honey than that produced by A. mellifera.
Stingless bee honey contains less sugar and more water than the honey produced by Apis spp. These bees are very important pollinators in their native habitats, being responsible for maintenance of the local vegetation, mainly the forests, and for seed production. There are circa 500-800 species occurring exclusively in the tropical and sub-tropical Neotropical regions. Apis spp. and Meliponini spp. forage on some of the same plants, but many different plants as well. Species specific preferences for various forage resources depend upon the bee species' affinity, as well as the flowering vegetation strata and environmental conditions .
3. Standard honey physicochemical analysis for nutritional composition, chemical composition, and antioxidant activity In this section, we present the main physicochemical analyses for nutritional composition, chemical composition, and antioxidant activity of honey in stepby-step format. Review articles on the composition and properties of honey from A. mellifera colonies are present in the literature Pascual-Mat e et al., 2018).
We recommend reading of the primary reference on this topic: Codex Standard for Honey (2001).

Water content determination (moisture)
Water content of honey is related to the botanical and geographical origin of the nectar, edaphic and climatic conditions, season of harvesting, manipulation by beekeepers, processing/storage conditions, etc. . It is an important physicochemical parameter for honey shelf life (Bogdanov, 2011) and it normally ranges between 13% and 25% (Uran et al., 2017), with the average being about 17% (Fallico et al., 2009). Honeys with high moisture content (>18%) have a higher probability of fermenting upon storage (Bogdanov, 2011) while those with a low moisture content (<15%) are likely to granulate.
The standard method for honey moisture determination, indicated by Codex Standard for Honey (2001), uses the refractometric method which is simple and reproducible. It is possible to use an Abb e refractometer or digital refractometers for this purpose (AOAC, 1992;Bogdanov, 2009Bogdanov, , 2011. The refractometric method is described here as a reference method (AOAC, 1990;Bogdanov, 2009).

Principle
Water content or moisture determined by the refractometric method is based on the fact that the refractive index increases with solid content. In honey, the refraction index can be converted in moisture content using the Chataway Table (Bogdanov et al., 1997).

Apparatus and reagents
Water bath or incubator Honey sample Refractometer 3.1.3. Procedure

Samples and standards preparation and determination
(1) If honey is granulated, place it in an airtight closed flask and dissolve crystals in a water bath or incubator at 50 C. (2) Cool the solution to room temperature and stir.
(3) Ensure that the prism of the refractometer is clean and dry. (4) Cover the surface of the prism from the refractometer with the sample and read the refractive index. (5) Read the corresponding moisture content from the Chataway Table (Table 2), making the temperature correction, if necessary.
Other methods can be used to determine water content of honey including Karl Fischer method and nonstandardized method Fourier transform infrared spectroscopy (FTIR) , 2014Bogdanov, 2009).

Sugar content determination (HPLC method)
Honey is mainly composed of sugars (60 À 85% of the total volume). Monosaccharides (fructose and glucose) represent about 75% of the sugars found in honey. Disaccharides (sucrose, maltose, turanose, isomaltose, maltulose, trehalose, nigerose, and kojibiose) compose about 10-15% of the sugars. The rest is composed of a small amount of other sugars, including trisaccharides (mainly maltotriose and melezitose). Due to the wide variety of botanical sources, the sugar profile presents great variation; however, it has been demonstrated that among the same species of plants, the profile of sugars and other physicochemical components can be used as fingerprint for the identification of the botanical origin of monofloral honeys (El Sohaimy et al., 2015;Ruoff et al., 2005Ruoff et al., , 2006. According to the publication of the Harmonized Methods of the International Honey Commission (2009), the methodologies for sugar determination are grouped as specific and nonspecific. Liquid chromatography -refractive index detector (specific methodology) and a methodology proposed by Lane-Eynon (nonspecific methodology) are among the most used methodologies for the quantification of sugars presented in honeys (Bogdanov, 2009).

Sample preparation and determination
(1) Weigh 5 g of honey into a beaker and dissolve in 40 mL distilled water.
(2) Pipette 25 mL of methanol into a 100 mL volumetric flask and transfer the honey solution quantitatively to the flask. (3) Complete the volumetric flask with water until 100 mL. (4) Filter the solution through a 0.45 lm membrane and collect the filtered solution into sample vials. (5) Store for four weeks in the refrigerator (4 C) or for six months at freezer (À18 C).

Calculation and expression of results.
Honey sugars can be identified and quantified by comparison of their retention times and the peak areas with those of the standard sugars (external standard method). The mass percentage of the sugars (W) is calculated according to the following formula: W ¼ Mass percentage of sugars (g /100 g). A 1 ¼ Peak areas or peak heights of the given sugar compound in the sample solution (expressed as units of area, length, or integration). A 2 ¼ Peak areas or peak heights of the given sugar compound in the standard solution (expressed as units of area, length, or integration). V 1 ¼ Total volume of the sample solution (mL). V 2 ¼ Total volume of the standard solution (mL). m 1 ¼ Mass amount of the sugar in grams in the total volume of the standard (V 2 ). m 0 ¼ sample weight (g)

Reducing sugars
Apparent reducing sugars and apparent sucrose content of honey are most commonly determined for honey quality control purposes. The Lane and Eynon method (Codex Alimentarius Standard, 1969) is used for the determination of sugars. Apparent reducing sugars are defined as those sugars that reduce a Fehling's reagent under the conditions specified (Bogdanov et al., 1997;Bogdanov & Martin, 2002;Bogdanov, 2009;Granato & Nunes, 2016).
(2) Weigh 0.5 g of glucose in a beaker, add distilled water, dissolve, and transfer to a 100 mL flask. The correction factor corresponds to the amount of sugar required to reduce 10 mL of the Fehling solutions.

Preparation of the sample.
(1) Dissolve 5.0 g of honey in distilled water up to 25 mL in a volumetric flask (Solution 1:5).

Apparent sucrose
Apparent sucrose content is determined indirectly by calculating difference in total reducing sugar before and after inversion of sugars in honey. The determination of apparent sucrose requires the use of acid hydrolysis to break the glycoside bonds of the disaccharides, thus releasing reducing sugars such as glucose and fructose (Bogdanov et al., 1997;Granato & Nunes, 2016 (1) Weigh 100 g of NaOH and dissolve in distilled water.
(2) Increase the volume in a volumetric flask to 500 mL with distilled water. (3) Condition the solution in a plastic container.

Sample preparation
(1) From the 1:5 diluted honey solution prepared in the reducing sugars analysis, transfer 2 mL to a 100 mL glass beaker and add 40 mL of distilled water and 1 mL of concentrated HCl. (2) This solution must be boiled, cooled, neutralized by adding 5 mol/L NaOH until a pH of 7 ± 0.2. (3) Add distilled water up to 100 mL in a volumetric flask. and make the titration according to the same procedure as "Reducing sugars" (3.3.2.4.). The result is expressed as percentage of total sugars (% AT): % AT ¼ ðdilution of honey solution ¼ 250Þ Â F Â 100 average of the volumes spent in the titration AT ¼ total sugars F ¼ correction factor obtained from the standardization of Fehling's solutions Heating to boiling under strongly acid conditions ensures that the sucrose is hydrolyzed; having as products the separate molecules of glucose and fructose. The molecular mass of sucrose (MM ¼ 342 mol/L) is 95% of the molecular mass of glucose and fructose together (MM ¼ 360 mol/L), taking into account that 5% of the mass refers to water (MM ¼ 18 mol/L) which was used in the hydrolysis. Thus, the percentage of apparent sucrose in a sample is calculated using the following formula: % apparent sucrose ¼ ð% total sugars -% reducing sugarsÞ Â 0:95

Proteins (proline)
The presence of protein in honey is low and varies depending on the species of honey bee (A. mellifera 0.2-1.6% and A. cerana 0.1-3.3%) (Lee et al., 1998) and the contact of these with flower pollen, honeydew, and nectar (Davies, 1975).
In the last decades, the use of more sensitive and robust methodologies has enabled the identification of a larger quantity of proteins and amino acids in honey samples from several countries . Among the currently used methodologies, the most preferred is HPLC (Hermosı n et al., 2003) associated with identification by mass spectrometry (Chua et al., 2013).
The classical method to determine protein uses the total nitrogen content that is usually calculated by multiplying the total nitrogen content from the Kjeldahl method by a factor of 6.25 (AOAC, 1990). However, about 40-80% of the total honey's nitrogen comes from the protein fraction and most of the remainder resides in the free amino acids (Chua et al., 2013).
According to the published update "Harmonized Methods of the International Honey Commission" in 2009, the proline content is recommended for quality control (being considered a criterion of honey ripeness) and it can be an indicator of sugar adulteration (Bogdanov, 2009). Since most of the amino acids are present in trace quantities, only proline content is included for the quality parameters of honey. No other methods for protein quantification are included in the methods referenced above or in the Codex Standard for Honey (2001).

Principle
Proline and ninhydrin form a colored complex. After adding 2-propanol, color extinction of the sample solution and a reference solution at a wavelength maximum is determined. The proline content is determined from the ratio of both and it is expressed in mg/kg. The method is based on the original method of Ough (1969).

Apparatus and reagents
Apparatus: Spectrophotometer measuring in the range of 500-520 nm Cells1 cm 20 mL tubes with screw cap or stopper 100 mL measuring flask and 100 mL beaker Water bath Reagents: Distilled Water Formic acid (H.COOH), 98-100% Solution of 3% ninhydrin in ethylene glycol monomethylether (methyl-cellosolve) Proline reference solution: Prepare an aqueous proline stock solution containing 40 mg proline to 50 mL distilled water (volumetric flask). Dilute 1 mL to 25 mL (volumetric flask) with distilled water (solution containing 0.8 mg/25 mL) 2-propanol solution, 50% by volume in water Proline reference solution.
(1) Prepare an aqueous proline stock solution containing 40 mg proline to 50 mL distilled water. (1) Weigh to the nearest milligram about 5 g honey into a beaker and dissolve in 50 mL distilled water. (2) After dissolving, transfer all the solution to a 100 mL volumetric flask. (3) Complete the volume with distilled water and shake well.

Determination.
Note that the coefficient of extinction is not constant. Therefore, for each series of measurements, the average of the extinction coefficient of the proline standard solution must be determined at least in triplicate.
(1) Pipette 0.5 mL of the sample solution in one tube, 0.5 mL of distilled water (blank test) into a second tube, and 0.5 mL of proline standard solution into each of three additional tubes. (2) Add 1 mL of formic acid and 1 mL of ninhydrin solution to each tube. (3) Cap the tubes carefully and shake vigorously using a vortex machine for 15 min.
(4) Place tubes in a boiling water bath for 15 min, immersing the tubes below the level of the solution. (5) Transfer to a water bath at 70 C for 10 min. (6) Add 5 mL of the 2-propanol-water-solution to each tube and cap immediately. (7) Leave to cool and determine the absorbance using a spectrophotometer for 45 min after removing from the 70 C water bath at 510 nm, using 1 cm cells.
Note: Adherence to the above times is critical.
3.5.3.3. Calculation and expression of results. Proline in mg/kg honey at one decimal place is calculated according to following equation: E s ¼ Absorbance of the sample solution E a ¼ Absorbance of the proline standard solution (average of two readings) E 1 ¼ mg proline taken for the standard solution E 2 ¼ Weight of honey in grams 80 ¼ Dilution factor In Germany, a honey with < 180 mg proline/kg honey is considered either unripe or adulterated (Bogdanov, 2009).

Vitamins (HPLC)
The presence of vitamins in honey is quite variable and is related to the botanical origin of the honey. Water soluble vitamins (vitamins C and B-complex) occur in higher levels in honey than do lipid-soluble vitamins (A, D, E, and K) due to the low presence of lipids in honey. In general, this foodstuff cannot be considered as an important source of vitamins due to the fact that they are in very low quantities. As mentioned by Le on- Ruiz et al. (2013), the identification of vitamins in different honeys allows their characterization to botanical type and is a valuable approach of honey quality control, but no official methods have been described .
There are few studies that identified and quantified vitamins in honey. Some authors such as Ciulu et al. (2011)  (1) Weigh 10 g of honey into a 50 mL beaker and dissolve in 10 mL of ultra-pure water. (2) Add 1 mL of NaOH 2M to complete solubilization. (3) Add 12.5 mL of phosphate buffer 1M (pH ¼ 5.5). (4) Raise the volume to 25 mL using ultra-pure water. (1) Add 40 mL of ultra-pure water.
(2) Add 4 mL of NaOH to complete dissolution. 3.6.2.3. Chromatographic analysis. The vitamins are identified and quantified by comparison of the retention times and the peak area of the pure standards. The elution program in the RP-HPLC gradient elution of water-soluble vitamins in honey is described in Table 3.

Minerals
The mineral content of honey samples has been receiving significant attention from many scientists globally. In recent years, the mineral profile of honey became an important indicator of quality control and environmental contamination by agrochemicals (Almeida- Silva et al., 2011;Kacaniov a et al., 2009). In the same way, some authors proposed that mineral profile can be used as an authentication analysis for botanical identification or designation of geographical origin (Anklam, 1998;Louppis et al., 2017).
Many factors can contribute to the mineral composition of honey, including the botanical origin of nectar, the harvesting treatment and material of storage, seasonal climatic variations, and geographical origin (Almeida- Silva et al., 2011;Anklam, 1998). Other minerals in honey may derive from those present in the air, water, and soil. Furthermore, bees can be exposed to minerals while visiting flowers, contacting branches and leaves, drinking water from pools, and while flying (Kacaniov a et al., 2009).
The mineral content in honey ranges from 0.04% in light honeys to 0.2% in dark honeys (Alqarni et al., 2014). Potassium is the most abundant mineral in honey, composing generally up to 1/3 of the total mineral content (Alqarni et al., 2014;Y€ ucel & Sultanoglu, 2013). In smaller quantities, honey also contains sodium, iron, copper, silicon, manganese, calcium, and magnesium. Macro elements (such as potassium, calcium, and sodium) and trace minerals (such as iron, copper, zinc, and manganese) perform a fundamental function in biological systems: maintaining normal physiological responses, inducing the overall metabolism, influencing the circulatory system and reproduction, and acting as catalysts in various biochemical reactions (Alqarni et al., 2014;da Silva et al., 2016).
Honey should be free from heavy metals in amounts that may represent a hazard to human health. However, maximum residue levels of these potentially toxic elements in honey have not been established. The World Health Organization (WHO, 1996), has proposed acceptable levels of 15 lg/kg for arsenic, 25 lg/kg for lead, 5 lg/kg for mercury, and 7 lg/kg for cadmium.
As described by Bogdanov et al. (2007), the characterization of trace elements by inductively coupled sector-field plasma mass spectrometry (ICP-SFMS) is significantly improved by virtue of enhanced sensitivity and separation of polyatomic interferences in a high resolution mode. In contrast, the option of inductively coupled plasma with mass spectroscopy detector (ICP-  (4) Perform a wet digestion of the solution using a microwave oven set at a digestion program consisting of three potency steps: first step at 500 W for 2.5 min, second step at 1000 W for 20 min, and the third step at 1200 W for 30 min.

Calculation and expression of results.
(1) Perform a quantitative analysis using the calibration curve method.
(2) Calibration curves should be constructed with a minimum of five concentrations of standards per element. (3) The limits of detection (LODs) should be calculated as three times the standard deviation of 10 consecutive measurements of the reagent blank, multiplied by the dilution factor used for sample preparation (Table 4).

Calories
Honey calories can be determined indirectly using calculation criteria according to the Brazilian Food Composition Table ( Food Research Center, 2017) and (WHO/FAO, 2002) which uses the Atwater general factors system. The calculation is done by multiplying the content of proteins and available carbohydrates by the factors described as follows: Observation: In the case of lipids, dietary fiber, and alcohol, it is zero for honey.

Hydroxymethylfurfural
Hydroxymethylfurfural (HMF) is a compound produced by sugar degradation (dehydration of hexoses in an acidic medium) (Bogdanov, 2011). HMF is a freshness parameter of honey; and in fresh honeys, it is absent or occurs in trace amounts. High values of HMF are naturally present in honeys from areas with warm climates (Sodr e et al., 2011). HMF is enhanced with honey processing and heat treatment. Adulteration with commercial sugars and long storage can also enhance HMF (Bogdanov, 2011).
There is more than one method found in literature to determine HMF content (HPLC and spectrophotometric method). We have chosen to summarize the spectrophotometric method of Bogdanov et al. (1997).

Principle
The spectrophotometric method is based on the principle that HMF absorbs at 284 nm.

Apparatus and reagents
Apparatus:

Procedure
The HMF spectrophotometric procedure can be found in AOAC (1990) and Bogdanov et al. (1997).

Samples, standards preparation and determination
Carrez I Solution-Dissolve 15 g of ferrocyanide in water and fill to 100 mL in volumetric flask.
Carrez II Solution-Dissolve 30 g of zinc acetate in water and fill to 100 mL in volumetric flask.
Sodium Bisulfite Solution 0.2%-Dissolve 0.20 g of sodium bisulfite in water and dilute to 100 mL in volumetric flask.
(1) Weigh approximately 5 g of honey into a 50 mL beaker.  (12) If the absorbance at 284 nm exceeds a value of about 0.6, dilute the sample solution with distilled water and the reference solution with sodium bisulphite solution (as described before) in order to obtain a sample absorbance low enough for accuracy.

Ashes
Ashes as a quality parameter is reported for Brazilian regulation, but this method will probably be replaced by the conductivity measurement. Ash content is related to honey origin (e.g., blossom honeys have lower ash content compared with honeydew honey) (Bogdanov, 2009).The procedure for ash determination is described in Bogdanov (2009).

Principle
Samples are ashed using a temperature < 660 C and the residue is weighted.

Apparatus
Crucible Electric furnace (can run a temperature of 600 C) Desiccator with dry agent Analytical balance 3.10.3. Procedure

Samples, standards preparation and determination
(1) Prepare the ash dish by heating it in the electrical furnace (ashing temperature 600 C at least 1 h). (2) Cool the ash dish in a desiccator to room temperature and weigh to 0.001g (m 2 ). (3) Weigh 5-10 g of the honey sample to the nearest 0.001 g into an ash dish that has been previously prepared in (1). (4) Cool the ash dish in the desiccator (2 h) and weigh. (5) Continue the ashing procedure (steps 4 and 5) until a constant weight is reached (m 1 ).

Calculations.
Make the calculation using the following formula:

Free acidity
Free acidity is determined using the method of recommended by the AOAC (1990). The International Honey Commission has proposed 50 milliequivalents as the maximum permitted acidity in honey (Bogdanov, 2009

Principle
A honey sample is dissolved in distilled water and free acidity is measured by titration with 0.1 M sodium hydroxide solution at pH 8.0.

Procedure
The procedure for free acidity is found in Bogdanov (2009).

Samples and standards preparation.
(1) The pH meter should be calibrated.
(2) Dissolve 10 g of the honey sample in 75 mL of carbon dioxide-free water in a 250 mL beaker. (3) Stir with the magnetic stirrer. (4) Immerse the pH electrodes in the solution and record the pH. (5) Titrate with 0.1M NaOH to a pH of 8.30 (complete the titration within 2 min). (6) Record the reading to the nearest 0.2 mL when using a 10 mL burette and to 0.01 mL if the automatic titrator has enough precision.

Insoluble solids
Honey insoluble solids include pollen, honey-comb debris, bees and filth particles, and is considered a criterion of honey cleanness. It is possible to determine insoluble solid content by filtering honey diluted with distilled water at 80 C in porous plate crucibles (Bogdanov, 2009;Bogdanov et al., 1997;Codex, 2001).

Apparatus and reagents
Drying oven at 105 C Drying oven Pump vacuum Thermometer Desiccator with silica gel Metal spatula Beakers (50 and 250 mL) Porous crucible # 3 1000 mL kitassato flask

Procedure
(1) Weigh about 20 g of honey into a 250 mL Beaker.
(2) Dilute with hot distilled water (80 C) until the sample is dissolved.

Calculation
The percentage of insoluble solids is calculated by the following ratio: Original sample mass ¼ 100% Mass of solids ðfiltered and dried crucible À dried crucibleÞ ¼ x%

Diastase activity
Diastases (a-and b-amylases) are enzymes naturally present in honey. Diastase content depends on the floral and geographical origins of the honey. Their function is to digest the starch molecule in a mixture of maltose (disaccharide) and maltotriose (trisaccharide). Diastase is sensitive to heat (thermolabile) and consequently are able to indicate overheating of the product and the degree of preservation (Ahmed, 2013;da Silva et al., 2016;Granato & Nunes, 2016). For this analysis, the method recommended by The International Honey Commission (2009) and Codex Alimentarius (2001) is based on the "diastase activity" that corresponds to the activity of the enzyme present in 1 g of honey, which can hydrolyze 0.01 g of starch in 1 h at 40 C, expressed as the diastase number in G€ othe units (Bogdanov et al., 1997). In this step-by-step method, the modifications proposed by Santos et al. (2003) were included to give improvements in the solutions preparation and spectrophotometer procedures.

Apparatus and reagents
Apparatus:

Standard preparation
Acetate buffer solution 0.1 mol/L pH 5.3: (1) Pipette 0.57 mL acetic acid into a 50 mL volumetric flask and fill with distilled water (Solution A). to the tube and mix. It is essential that the honey solution is buffered before contact with sodium chloride, because at pH below 4, diastase activity is inhibited. (7) Add 500 lL of the sodium chloride solution 0.1 mol/L to the tube. (8) Add 150 mL of the solution of 0.02 mol/L iodine and 9.6 mL distilled water to the tube and mix. (9) Keep the tube and its contents in a water bath at 40 C ± 1 C. (10) Add 250 lL of the 1% starch solution (m/v). (11) Start the stopwatch, shaking the solution until complete homogenization occurs. (12) Transfer a part of the volume of the reaction system to complete a 1 cm cell and measure the absorbance of the solution in the spectrophotometer at 660 nm against a water blank. This first reading is the initial absorbance value (Ab si ). The tube with the solution should remain in the water bath at 40 C. (13) Make periodic readings of absorbance, always returning the tube to the water bath when not being read, until a value between 0.24 and 0.20 is reached. (14) When this value is reached, stop the timer and record the elapsed time value. (15) The last recorded absorbance value is considered the final absorbance (Abs f ). (16) Calculate the diastase index using the following formula: (previously determined by honey-free assay, given by method) T (h) ¼ time (in hours) between Abs i and Abs f measurements V ¼ volume of the 10% honey solution in the test tube (mL) 0.016 ¼ total volume in liters of the solution in the test tube (16 mL)

Electrical conductivity
This method is easy, quick, and involves inexpensive instrumentation. Electrical conductivity determination depends on the ash and acid contents of honey (e.g., the higher ash and acids, the higher the result for electrical conductivity). This parameter has been used as criterion of the botanical origin of honey (Bogdanov, 2009).
According to the European regulation (European Commission, 2001), electrical conductivity of blossom honey must be < 0.8 mS/cm of EC, while electrical conductivity of honeydew honey and chestnut honey must be > 0.8 mS/cm. Exceptions are honeys from Arbutus, Banksia, Erica, Leptospermum, Melaleuca, Eucalyptus, Tilia, and blends (De-Melo et al., 2017).

Principle
Electrical conductivity uses the measurement of the electrical resistance.

Apparatus and reagents
Conductimeter lower range 10 À7 S. Conductivity cell (platinized double electrode-immersion electrode) Thermometer Water bath Volumetric flasks (100 mL and 1000 mL) Beakers Distilled water Potassium chloride 3.14.3. Procedure The standard method for measuring honey electrical conductivity was outlined by Bogdanov (2009). 3.14.3.1. Samples and standards preparation. (2) Potassium chloride solution (0.1 M) preparation: dissolve 7.4557 g of potassium chloride (KCl), dried at 130 C, in freshly distilled water and fill up to 1000 mL volumetric flask. Prepare fresh on the day of use. (3) Connect the conductivity cell to the conductivity meter. (4) Rinse the cell thoroughly with the potassium chloride solution and immerse the cell in the solution, together with a thermometer. (5) Read the electrical conductance of this solution in mS after the temperature has equilibrated to 20 C. (6) Calculate the cell constant K, using the formula: , measured with the conductivity cell; 11.691 ¼ the sum of the mean value of the electrical conductivity of freshly distilled water in mS.cm -1 and the electrical conductivity of a 0.1 M potassium chloride solution, at 20 C). (7) Rinse the electrode with distilled water after the determination of the cell constant. (8) Dissolve an amount of honey, equivalent to 20 g of dry matter of honey, in distilled water.

Determination.
(1) Transfer 20 mL of the solution created in section "Samples and standards preparation" to a 100 mL volumetric flask and fill up to 100 mL with distilled water. (2) Pour 40 mL of the sample solution into a beaker and place the beaker in a thermostated water bath at 20 C. (3) Rinse the conductivity cell thoroughly with the remaining part of the sample solution (Step 1). (4) Immerse the conductivity cell in the sample solution and read the conductance in mS after temperature equilibrium has been reached. (5) If the determination is carried out at a different temperature, because of a lack of thermostated cell, a correction factor can be used for calculating the value at 20 C. For temperatures above 20 C, subtract 3.2% of the value per C. For temperatures below 20 C, add 3.2% of the value per C.
3.14.3.3. Calculations. Calculate the electrical conductivity of the honey solution, using the following formula: Where: S H ¼ electrical conductivity of the honey solution in mS.cm À1 ; K ¼ cell constant in cm À1 ; G ¼ conductance in mS. Express the result to the nearest 0.01 mS.cm -1 .
Also, FTIR could be used in order to determine honey electrical conductivity , 2013.

Trolox equivalent antioxidant capacity (TEAC)
Determining the scavenging ability of honey to the radical cation of ABTS [2,2 0 -azinobis-(3-ethyl-benzothiazoline-6sulfonic acid)], using trolox, a water-soluble vitamin E analog, as the standard for the calibration curve (Miller et al., 1993;Re et al., 1999;Sancho et al., 2016) is the method we propose for measuring antioxidant capacity. The described method is a modification of Re et al. (1999) procedure, optimized by Sancho et al. (2016) (2) Working ABTS þ solution: Dilute the undiluted ABTS þ solution with distilled water to obtain an absorbance between 0.70 and 0.80 at 734 nm. The working ABTS þ solution must be prepared fresh daily.

Samples and standards preparation.
Sample preparation (H-sample): (1) Weigh 10 g honey into a beaker.
(4) Further dilute to 100 mL with distilled water.

Determination.
(1) Using a pipette, transfer 990 mL of the diluted ABTS þ working solution (see "Generation and dilution of ABTS radical cation (ABTS þ )" section) into each cell. One cell corresponds to distilled water (no antioxidant), six cells correspond to the calibration curve (with trolox standard solutions of section "Samples and standards preparation"), and another cell corresponds to the sample (of section "Samples and standards preparation"). This procedure is performed in triplicate.

Calculations.
Percentage of inhibition: A 0 is the absorbance at initial time of distilled water (DW), standard solution (STANDARD), or sample solution (H-Sample), referred to section "Determination" (3).
A 6 is the absorbance measured at 6 min of distilled water (DW), standard solution (STANDARD), or sample solution (H-Sample), referred to section "Determination" (4).
Linear regression: a standard trolox calibration curve is drawn by representing trolox concentrations (mM) on the x-axis, and % inhibition on the y-axis.
value of the honey sample (mmol trolox/g honey): TEAC antioxidant activity of honey is calculated as follows:

Color of honey by the CIELab system
The colorimeter measures the reflected color of the honey samples, both liquid and crystallized, of the whole visible spectrum in a wavelength interval between 380 and 740 nm. Other methods for measuring a honey sample's color require honey dilution before measuring its color by spectrophotometry at a given wavelength. Or, it is necessary to dissolve any crystals in the honey to obtain a completely liquid honey that is transparent. Both procedures modify the honey color. The advantage of the colorimeter method is that it is the only instrument that gives the samples' color the way the human eye does, whether the sample is liquid or crystallized. Moreover, honey remains suitable for use in other analysis (Sancho et al., 2016).  (Sancho et al., 2016). (4) Other interesting values for honey comparison or characterization, calculated from a Ã and b Ã , are the coordinates C Ã (chroma, saturation, vividness, or purity of a color) and h Ã (hue angle or tone) (Tuberoso et al., 2014).
Investigating honey production with diagnostic radioentomology Diagnostic radioentomology is a technique based on the measurement of density from component materials using X-rays. The density measured by computer tomography (CT) scanning reflects the degree to which the energy of the X-ray beam is reduced when penetrating a certain material. This technique allows for the quantification and visualization of differences in density with minimal disturbance of the colony since the hive does not need to be opened to introduce light into the dark nest and since no sample need be collected. Thus, it becomes possible to investigate, noninvasively, phenomena that are affected by other common research methods that require destructive sampling. Honey originates from nectar and honeydew and undergoes a process of concentration (i.e., dehydration). The increase in sugar concentration affects honey density. This can be tracked using radioentomology.
Despite the importance of honey for beekeeping, the processes that lead to its production within the honey bee nest have rarely been investigated (Eyer, Greco, et al., 2016;Eyer, Neumann, et al., 2016). Diagnostic radioentomology was recently used to shed light on nectar storage strategies of bees by allowing for the measure of sugar concentration of the content of a large number of storage cells (Eyer et al., 2016;Greco et al., 2013). By following the evolution of the sugar concentration in individual cells over time, it is also possible to monitor several stages of honey production by the workers, from initial deposition of nectar or honeydew in cells to the capping of mature honey (Eyer et al., 2016).
This technique could be used further to determine the influence of environmental (e.g., high or low nectar flow, diverse or monotonous nectar flow, varying weather conditions) and internal factors (e.g., hive design, air circulation, colony demography) on honey ripening and storage. A better knowledge of the factors affecting honey production processes by honey bee workers could not only lead to a better knowledge of how the honey bee superorganism manages its food stock but could also lead to the improvement of beekeeping management practices by favoring a rapid ripening of honey or by increasing productivity. A further application of this technique is to track the deposition and spread of substances applied into hives (Rademacher et al., 2013) that might contaminate honey stores. The concentration of crop content of live, yet motionless honey bees can also be measured with this technique. Rapidly moving individuals will appear blurred in the scan and thus hinder precise measurements. In the following sections, we describe how to perform measurements of sugar concentration of the content of carbohydrate reserves stored in wax combs with a CT scanner.

Experimental conditions and material required
to conduct diagnostic radioentomology trials 4.1.1. Hive to scanner size ratio The most commonly available CT-scanners are those used for human or veterinary medicine. Their dimensions allow for scanning hives of most standard sizes without honey supers. Larger scanners exist for veterinary applications and can be used to study hives with supers. When high resolution is required, it can be obtained by reducing the size of the experimental hive (e.g., using Miniplus (R) hives, 30 Â 30 Â 34 cm). The focalization of the X-ray beam on a smaller area increases scan resolution. Micro-CT scanners are also available to study small items (e.g., individual motionless workers, a few cells) with high resolution.

Material to be scanned
Combs built from foundation sheets in moveable frames have a regular and planar shape, which facilitates scanning and data analysis. However, diagnostic radioentomology can be used to study honey production in combs of any shape (e.g., in wild nests in logs) thanks to 3D reconstruction tools. Depending on the research question, combs already containing stores can be used as starting material. When the starting point of the study is an empty cell, a drawn empty comb can be inserted in the test colony. Single combs without workers, workers alone or entire hives with workers can be scanned. Single combs can be gathered in an empty hive or super box and all scanned at once to decrease the number of scans required ( Figure 3).
When hives containing honey bees are to be scanned, safety of scanning facility staff and of bystanders needs to be considered. Hive entrances should be closed early on the day of scanning before foragers start their activity and tight hives should be used to prevent honey bees from escaping. Ventilation (e.g., through a screened bottom) should nevertheless be possible for the workers to be able to adjust hive temperature. Combs and hives have to be labeled in a way that they can be recognized on the scan images. Notches cut in the wood or thin metallic letters or numbers can be used. Metallic parts like nails and wires deflect the Xray beam and produce glares that make data analysis difficult. It is thus necessary to remove all metallic parts from the material to be scanned or to use plastic or polystyrol hive material.

Monitoring the storage of artificial diets
Nectar and honeydew collected by foragers are highly diverse in their composition. Consequently, it can be useful to use artificial diets to simplify the study of honey production. The storing of artificial diets (e.g., commercial feeding solutions) can also be the purpose of the experiment. Commercial feeding solutions or a solution of controlled sugar concentration can be provided to the colonies in feeders within the hive or outside of the hive at a close distance, which reflects a more natural foraging situation.
To prevent the foragers from collecting other sources of carbohydrates than those experimentally provided, test colonies can be placed in tunnel cages (see Medrzycki et al. (2013), section "Scanning settings"). When tunnel cages are not available, field colonies can be used during a period without natural nectar flow. To confirm absence of incoming nectar, the concentration of returning foragers' crop content can be measured at regular interval using the following method: (1) Collect returning foragers that have no pollen baskets on their legs at the hive entrance with the help of forceps.
(2) Gently squeeze their abdomens between your fingers to force them to regurgitate a small droplet of crop content. (3) Place this droplet onto a refractometer adapted to the range of expected concentrations (see Human et al. (2013), section 4.7.3.1). (4) Read the sugar concentration through the refractometer. (5) Repeat this procedure daily for 10-50 foragers per test-colony.
If the concentration of crop content does not correspond to that of the diet provided, the foragers are collecting nectar from other sources. If this input cannot be accounted for in the analysis, the experiment should be performed when naturally available resources do not interfere with the aim of the study.

Studying the storage of natural diets
When the aim of the experiment is to study natural diets, field colonies with freely flying honey bees can be used. Choose a period with the nectar or honeydew flow of interest for your research question. The range of sugar concentrations collected by foragers can be verified by squeezing the abdomen of returning foragers (see section "Monitoring the storage of artificial diets").

Labeling of the diet to increase contrast
When the storage of qualitatively different solutions of similar densities needs to be studied, it is possible to discriminate between these solutions by labeling one of them with a contrast agent (Eyer et al., 2016). One example of a suitable contrast agent is Visipaque, which is harmless to honey bee workers at a concentration of 10%. Visipaque (iodixanol Injection) is a contrast media solution, with multiple indications in human medicine (aortography, venography, urography, etc.) for use in Xray scans. Other contrast agents may be used but their toxicity for honey bees need to be determined (Medrzycki et al. 2013).
If a contrast agent is used, whether it mixes with the solution or sediments in the cell needs to be determined prior to the experiment (Eyer et al., 2016). This verification is required to exclude possible bias from the contrast agent when monitoring density patterns or measuring density of cell or crop content.

Calibration
The density of cell content is measured in Hounsfield units. This unit is used as a proxy for sugar concentration. A regression of sugar concentrations on density of the artificial carbohydrate solutions or natural sources is thus required to convert Hounsfield units in % sugar concentration (Eyer et al., 2016). To (1) Prepare vials (20-50 mL) with carbohydrate solutions (mixing granulated sugar with water) covering the range of concentrations required for the experiment (e.g., 30%, 50%, 60%, 70%, and 80% sucrose).  Ideally, a portable CT-scanner should be used to avoid interfering with the biological processes in the colony. However, until the use of portable devices in the field becomes more practical and affordable and since the most frequently accessible devices are fixed, it is recommended to minimize transport of the test hives or combs by conducting the field component of the experiments near the CT-facility. In cases when transportation cannot be avoided, its effects can be minimized by careful handling, avoidance of shaking and shocking the hives or combs to be tested. If combs are scanned outside of the hive, not using smoke during comb collection will limit methodological bias by not provoking cell content take up by workers. Workers can be removed from the combs by gently brushing them off, taking care that the bristles do not contact the cell content. Shaking the combs to get rid of workers is not recommended due to possible effects on cell content, especially if storage of nectar of low sugar concentration is investigated. In the absence of honey bees, further biases such as cell content evaporation can be minimized by scanning the comb immediately after collection.

Scanning settings
The scans can be performed with a Philips Brilliance CT 16-slice apparatus (e.g., Philips Healthcare, 5680 DA Best, The Netherlands) using 120.0 kVp and 183.0 mA as settings. For other models, detailed device setting should be discussed with scanner operator according to the output required.

Performing a scan
Constant conditions during scanning should be maintained, holding the temperature in the scanning room at 18-20 C. If required by the study question, scans can be repeated at defined time intervals. Adhere to safety measures as instructed by scanner operator to avoid the danger of exposure to X-rays.
(1) Place the hive or combs to be scanned on the scanner bed. (2) Adjust bed movement range and position of material to be scanned using the device's positioning laser beams ( Figure 4). (3) Place yourself in the designated, protected area; personal other than operators leave the room. (4) Press scan and survey the scan-procedure on the computer. (5) Wait until scan is complete (usually indicated by the extinction of an acoustic signal) before leaving the designated area. (6) Save the scan output to a specific location on a server or a hard drive.

Analysis of density patterns in individual cell
For the visualization of cell content density patterns, CT-images can be analyzed with specific 3D rendering software that permits the visualization of differences in density with a color gradient ( Figure 5). This feature enables visual monitoring of density measured following this procedure: (1) Load file.
(2) Choose parameter settings. Detailed device settings for the windowing feature should be discussed with the scanner operator according to the output required. A dark color represents low density content, whereas light color represents high density content. Image analysis software also has several coloration presets (e.g., Figure  5) that can be tried to render the best image output for the analysis. There is no empirical manner to determine the best color settings. The appearance of the images generated will determine which is the most informative. (1) Load image file.
(2) Navigate/Scroll vision plane to the cells of interest. (3) Use the ellipse tool to draw a circle over the content of the cell, three scanning frames away from its bottom end. (4) Repeat the measurement of the same cell, three frames below the surface of the cell content.
Steps 3 and 4 are necessary to take into account within cells variation of density (see Eyer et al., 2016). Scanning a few frames away from cell bottoms and openings helps avoiding the inclusion of air volumes or wax in the circles, which would strongly bias the measurement. (5) Average the values of these measurements.

Measuring cell filling status
The filling status of the cells can be estimated as the number of scanning frames for which the content fills the whole cell diameter.
(3) Scroll through each frame with full content while counting.

Analyzing content of individual cells over time
(1) Take a picture of each test comb or load the image file with eFilm. In case the whole process of honey production is to be monitored but not all cells are capped at the end of the experiment, the selection of cells measured can be done a posteriori, once cells have been capped.

Spatial analyses at comb level
The following method can be used to investigate spatial patterns in nectar processing (Eyer et al., 2016).
(1) Select combs that contain a sufficient number of filled cells (e.g., Figure 7).

Pros and cons
Pros: The use of CT-scanning permits nondestructive observations and measurements within the dark hive.
Repeating the observations and measurements in time is thus possible without disrupting the phenomenon under scrutiny. Three-dimension imaging is possible. Cons: Different materials of similar density cannot be distinguished based on their Hounsfield value. Unless a portable device is available, the material to be scanned has to be transported to the scanning facility. Only snapshots and hence low frequency time lapse images can be captured, filming is not possible to observe the behavior of workers for example. Metallic parts create glare and need to be removed from the scanned material.

Perspectives
Diagnostic radioentomology is a powerful tool to monitor otherwise difficult to observe processes. However, this recently developed method could be further developed to increase its usefulness. Scans performed immediately after behavioral observations in hives with transparent sides can help relate density patterns with the behavior that generated them. Better labeling techniques (e.g., avoiding sedimentation of contrast agent) of cell content would provide further opportunities to monitor processes involved in nectar storage and honey production with an even higher resolution. Scanning of storage combs should also be performed more Figure 7. A coronal scan through a partly filled comb. Colors reflect differences in density, with black for low density to yellow for high density. Cell content appear in green-yellow shades, wax and wood in blue, air in black.
frequently as done to date, aiming at investigating the processes of honey production with higher time resolution. Further, generating a database of densities of nectar and honey of specific origins (with various sugar compositions) or at different ripening stages could improve the sets of tools available and might relieve the need for calibration before each experiment. Scanning honey bee workers, to investigate crop content, for example, is limited by the relatively low scanning speed. This is especially the case with older devices and when workers are moving. However, the motors of new generations of scanners are spinning the X-ray and detection units at high speeds and will allow for the freezing of a greater proportion of the worker's movement. Other imaging techniques (e.g., radioactive labeling and laser scanning microscopy) could also be employed for studying honey production by workers.

Pesticide residues in honey
The term "pesticides" represents many different substances used in various crop protection products to treat plants against pests and in veterinary drugs used against animal pests/parasites (including those used in honey bee colonies to control bee pests/parasites), in/ around structures to protect against structural pests, etc.
Pesticides applied on crops can contaminate plants, soil, water, and air, and honey bees may be exposed to them via contact with these matrices. The bees collect and transport contaminated products (nectar, pollen, and water) into the hive. There is also a risk of finding pesticide residues in bee products following treatment of the hive.
Honey samples can be screened for pesticide residues for food safety purposes ("Commission Regulation (EU) No 37/2010", European Union Commission, 2009;"Regulation (EU) No 396/2005. The European Union (EU) requires that honey be tested for pesticide residues within the framework of the monitoring program covered by the Council Directive 96/23/EC (European Communities, 1996). Maximum residue limits (MRLs) are defined for pesticides in honey (EU Pesticides Database) which mostly are between 0.01 and 0.05 mg/kg. For acaricides used in beekeeping, two MRLs were defined in the Commission Regulation (EU) No 37/2010 (2009) for amitraz (and its metabolite, 2,4dimethylaniline) and coumaphos: 0.2 and 0.1 mg/kg, respectively.

Chemical families of pesticides
The most common use of pesticides is as plant protection products in agriculture. Pesticides are classified in three main groups according to the nature of the "pest" to be controlled: herbicides, fungicides, and insecticides (including acaricides). In each group, there are different chemical families of pesticides.

Herbicides
Herbicides are used to eliminate weeds which disturb the growth of the crops. They are widely used in agriculture to kill plants or to inhibit their growth or development. Residues are found in air, water, and on various plant parts. There are multiple chemical classes of herbicides: carbamates, triazines, triazoles, and ureas to name a few.

Fungicides
Fungicides are used to kill fungi in plants, stored products, or soil, or to inhibit their development. Fungicides can either be contact, translaminar, or systemic. The main chemical families of fungicides used are benzimidazoles, dicarboximides, triazoles, chloronitriles, and carbamates.

Insecticides and acaricides
Insecticides and acaricides are used to kill or disrupt the growth/development of insects or mites. Insecticides are applied on crops to protect them against pests. They can be classified into two groups. (i) Systemic insecticides are incorporated into the tissues of treated plants. Insects ingest the insecticide while feeding on the plants. (ii) Contact insecticides are toxic by contact to insects. Acaricides are also used against honey bee parasites such as Varroa in hives . The main chemical insecticide families are organochlorines, organophosphorus, pyrethroids, neonicotinoids, carbamates, and phenylpyrazoles.

Analytical methods
For protecting the health of consumers, the analytical challenge is to achieve limits of quantification at or below the MRL specified for pesticides under EU or other similar legislation. The laboratories conducting the residue analyses usually develop and validate a multiresidue method. Within the framework of official controls on pesticide residues, laboratories follow the requirements specified in the guidance document on analytical quality control and validation procedures for pesticide residues analysis in food and feed ("Document SANTE/11813/2017", European Commission, 2018).
The development of residue analysis methods depends on the properties of both the matrices (honey, wax, pollen, etc.) and pesticides. Residue analysis involves several steps. There are many extraction and clean-up procedures used by different authors to determine the amount of pesticides in honey (Barga nska & Namie snik, 2010). Solvent extraction and solid-phase extraction (SPE) are the techniques most commonly used for the extraction of pesticides from honey.
There are two general analytical approaches used to determine residues in food and environmental samples: (i) specific methods where a single pesticide and its metabolites are quantitatively determined in the sample, and (ii) multiresidue methods which are analytical methodologies for the simultaneous analysis of trace amounts of a large number of analytes. The number of pesticides tested in the sample can be limited in order to get reliable results, higher recoveries, and lower quantification limits.
Several analytical methods have been used to separate and detect pesticides in honey. Gas chromatography (GC) and liquid chromatograpy (LC) are used for the detection and quantification of pesticide residues (Souza Tette et al., 2016). The choice of the separation technique depends mostly on the characteristics of the pesticides of interest. The volatile, semi-volatile, and thermally stable compounds can be determined by GC, whereas nonvolatile and/or thermally unstable ones should be determined by LC.
At least three steps are required for the analysis of pesticides, among them extraction, separation, and detection. Each one of these steps will be described in the sections that follow.

Sampling
The samples to be analyzed must be representative of the entire honey batch in question. The different steps for the sampling of honey are as follow: (1) If the honey sample contains impurities (e.g., wax), the sample should be filtered through a stainlesssteel sieve. If needed, the honey can be gentle pressed through the sieve with a spatula. (2) If a honeycomb is sampled, the honeycomb is drained through a 0.5 mm sieve without heating in order to separate honey from the comb. (3) For crystallized honey, the sample is homogenized with a spatula and an analytical test portion is collected by coring of the honey. (4) Honey must be homogenized before analysis. (5) According to the protocol of sample preparation applied for extraction of pesticide residues in honey, collect 1-20 g honey into a centrifuge tube or a beaker for sample preparation.

General requirements for pesticide residue analyses
(1) All analyses should include negative control honey (honey free from pesticide residues-the "blank") and matrix calibration standards prepared by adding pesticides to blank honey before pretreatment of the samples. (2) The sample spiked with pesticides at the level corresponding to the limit of quantification (LOQ) for Table 5. Analysis of pesticide residues in honey using a liquid-liquid extraction.
Pesticide analyzed Extraction protocol References Organochlorines (a-, b-, and Ç-hexachlorocyclohexane (HCH), hexachlorobenzene (HCB), aldrin, p,p'-DDE, p,p'-DDD, o,p'-DDT and p,p'-DDT) 1. Dissolve 5 g of honey with 50 mL 4% aqueous solution of sodium sulfate in a centrifuge tube. 2. Add 20 mL of ethyl acetate to the sample. Repeat the liquidliquid extraction two times with 15 mL of ethyl acetate. 3. Centrifuge at 3000 rpm for 10 min if emulsion is formed. 4. Filtrate the organic phase through anhydrous sodium sulfate. 5. Evaporate the organic phase under a stream of nitrogen to 2.5 mL for analysis in graduated centrifuge tube. 6. Analyze by GC-ECD, and Gas Chromatography/Mass Spectrometry (GC-MS) for confirmatory analysis following authors' protocols. Blasco, Lino et al. (2004) Coumaphos, bromopropylate, amitraz and tau-fluvalinate 1. Mix 20 g of honey in an Ultra-Turrax blender with a mixture of n-hexane (60 mL), propanol-2 (30 mL) and 0.28% of ammonia. The pH of this mixture is 8. 2. Filter the solution through a filter paper. 3. Repeat steps 1 and 2 with the mixture of n-hexane (60 mL), propanol-2 (30 mL) and 0.28% of ammonia. 4. Rinse the Ultra-Turrax with 40 mL of n-hexane and filter this washing solution on the same filter paper as in step 2. 5. Transfer the combined extracts from steps 2, 3 and 4 to a separating funnel (500 mL). 6. Add 50 mL of distilled water and 0.28% of ammonia (pH 10). 7. Shake the separating funnel vigorously. 8. Allow the filtrate to separate into two phases. 9. Discard the aqueous phase (lower). 10. Repeat steps 6, 7, 8 and 9 twice. 11. Filter the n-hexane phase through a layer of anhydrous sodium sulfate placed in a funnel plugged with a filter paper. 12. Concentrate the extract by evaporation to dryness under reduced pressure in rotary evaporator using a 35-40 C water bath. 13. Recover the residue obtained with 1 mL of acetone before analysis. 14. Analyze by HPLC-DAD following authors' protocol.

Martel and
Zeggane (2002) Table 6. Analysis of pesticide residues in honey using a solid-phase extraction (SPE).
Pesticide analyzed Extraction protocol References 450 pesticides 1. Dilute 15 g of the test sample in a 250 mL glass jar with 30 mL of water. 2. Shake for 15 min at 40 C in a shaking water bath. 3. Add 10 mL of acetone to the jar. 4. Transfer the jar contents to a 250 mL separating funnel. 5. Rinse the jar with 40 mL of dichloromethane and transfer this rinse to the separating funnel for partitioning. 6. Shake the funnel eight times and pass the bottom layer through a funnel containing anhydrous sodium sulfate into a 200 mL pear-shaped flask. 7. Add 5 mL of acetone and 40 mL of dichloromethane into the separating funnel. 8. Repeat steps 6 and 7 twice. 9. Evaporate the organic phase to about 1 mL with a rotary evaporator at 40 C for clean-up. 10. Add sodium sulfate into a graphitized carbon black cartridge to about 2 cm. 11. Connect the cartridge to the top of the aminopropyl cartridge in series. 12. Condition the cartridges with 4 mL acetonitrile/toluene 3:1 (v/v). 13. Add 1 mL of the sample. 14. Rinse the pear-shaped flask with 3 Â 2 mL acetonitrile/toluene 3:1 (v/v) and decant it into the cartridges. 1. Dissolve 10 g of honey in a 60 mL glass tube with 10 mL of water and 10 mL of acetone. 2. Homogenize the mixture with Ultra Turrax for 2 min. 3. Wash the Ultra Turrax with about 2 mL of acetone and collect the washings into the glass tube. 4. Load the solution obtained in step 3 into an EXtrelutV R NT 20 column. 5. Allow to drain for 10 min to obtain an even distribution into the filling material. 6. Elute pesticides with 5 Â 20 mL of dichloromethane using the first aliquot to wash the glass tube. 7. Collect the eluate into a 150 mL Erlenmeyer flask. 8. Concentrate to nearly 1 mL with a rotary evaporator at 40 C (reduced pressure) and by drying manually by rotating the flask. 9. Dissolve the residue in 2 mL of n-hexane. 10. Divide the sample solution obtained in step 9 in two portions. 11. Transfer one portion of the sample (1 mL) into a 25 mL Erlenmeyer flask and dry by manually rotating the flask. 12. Dissolve the residue in 0.5 mL water/acetonitrile 50:50 (v/v). 13. Add the appropriate amount of 4-Br-NMC as an internal standard. 14. Inject the sample into the LC/DD/Fl (liquid chromatography-double derivatization coupled with spectrofluorimetric detector) for the determination of NMC pesticides following author's protocol. 15. Add the appropriate internal standards (PCB 209 and TPP) in the second portion of the original sample left in the Erlenmeyer flask (1 mL).

Inject this portion in GC-MSD and in GC-FPD following authors'
protocols in order to analyze bromopropylate, OC, PYR and OP residues respectively. Amendola et al. (2010) each target analyte must be prepared with other samples to control the sensitivity of the equipment. (3) A supplementary test (named "test sample") is conducted to measure the recovery of all target analytes. (4) For each analytical sequence, the extracts are injected into the analytical instrument in the following order: blank solvent, negative control honey (blank sample), samples spiked with pesticides for calibration (the "matrix calibration standards"), blank sample, unknown samples (samples to quantify), test sample (to calculate the recovery for each pesticide) and, again, a spiked sample from the calibration to control any variation during the sequence and blank sample.

Sample preparation (extraction and clean-up)
The purpose of the extraction is to extract the substance from the sample with minimum co-extractives matrix interferences. The choice of the solvent is important. Acetone, acetonitrile, ethyl acetate, and hexane are commonly used in this step. The analytes are extracted from the matrix and then, through a clean-up process, the co-extractives are removed. The most frequent technique used for sample extraction is via the homogenization of the honey sample with an organic solvent or water followed by clean-up with a liquid-liquid extraction (LLE) or a solid phase extraction (SPE) on column. Relatively new techniques like the dispersive solid phase extraction (dSPE) and solid phase microextraction (SPME) are used. Main parameters of methodologies mentioned in the literature for the extraction of pesticides from honey are reported by Rial-Otero et al. (2007) and Souza Tette et al. (2016). Different techniques are presented below with corresponding protocols. The main criterion is to find one method that gives acceptable recoveries for all analytes with one protocol.

Liquid-liquid extraction (LLE).
In multiresidue methods, an important step is the extraction procedure, especially for complex matrices such as honey, which contains high sugar content. LLE is the most common extraction and purification technique used in the determination of pesticides in honey (Souza Tette et al., 2016). However, LLE usually employs large sample sizes and toxic organic solvents. It is also characterized by the use of multiple sample handling steps, which makes it susceptible to error and contamination. Furthermore, it usually enables the extraction of analytes belonging to only one chemical class. Despite the disadvantages described above, LLE continues to be used in the analysis of pesticides in honey. Barga nska and Namie snik (2010) reviewed in their paper the techniques already used to extract pesticides from samples of honey. The general protocol involves dissolving honey in water and applying a LLE with various solvents (such as petroleum ether, hexane, dichloromethane, ethyl acetate or mixtures of petroleum ether/ethyl acetate (Tahboub et al., 2006), acetonitrile/acetone/ethyl acetate/dichloromethane (Rissato et al., 2007), acetone/dichloromethane (Pang et al., 2006), hexane/propanol-2 (Martel & Zeggane, 2002)) to extract pesticides according to the pesticide's polarity. The experimental procedures are given in Table 5.
The extracts obtained before analysis by chromatography are clean but these methods use a considerable amount of solvent. Furthermore, these methods are limited to the extraction of a few pesticides due to their different solubility.

Solid phase extraction (SPE) and dispersive solid phase extraction (dSPE)
. Sometimes, after a LLE, a clean-up step on a SPE cartridge with different adsorbents may be necessary before quantification of pesticide residues in honey. Protocols using SPE are given in Table 6 for the main classes of pesticides.
The SPE technique (Barga nska & Namie snik, 2010; Rial-Otero et al., 2007;Singh et al., 2014;Souza Tette et al., 2016) combines in a single step the extraction and the clean-up procedures based on the separation of LC, where the solubility and functional group interactions of sample, solvent, and sorbent are optimized to affect retention and elution. In SPE, the sample is passed through a cartridge or a packed column filled with a solid sorbent where the pesticides are adsorbed and then eluted with an organic solvent. Moderately polar to polar analytes are extracted from nonpolar solvents on polar sorbents. Nonpolar to moderately polar analytes are extracted from polar solvents on nonpolar sorbents. Most multiresidue methods include a clean-up step using adsorption columns on polar sorbents (florisil, alumina, silica gel) for normal-phase SPE and on nonpolar sorbent like C18 for reversedphase SPE. The reversed-phase C18 cartridge is by far the most common choice used by researchers for the extraction of insecticides, acaricides, fungicides, herbicides, organochlorines, and organophosphorus pesticides from honey (Rial-Otero et al., 2007). As florisil retains some lipids preferentially (25 g florisil with 3% water retains 1 g of fat), it is particularly well suited for the clean-up of fatty foods. When a florisil column is eluted with solvent mixtures of low polarity, nonpolar residues are recovered almost quantitatively (Singh et al., 2014). Florisil sorbent has been used for pyrethroids, organochlorines and organophosphorus pesticides. Lores et al. (1987) used a silica gel clean-up method for organophosphorus pesticide analysis.
Other SPE clean-up approaches include the combination of GCB (graphitized carbon black) and PSA (primary secondary amine) columns, the combination of C18, GCB and aminopropyl, and the combination of GCB, PSA, and SAX (strong anion-exchange sorbent) columns. GCB is such a sorbent, being nonspecific and generally of hydrophobic nature. Contrary to sorbents based on SiO 2 , these may be used without the pH of the treated solutions being considered. Because of difficulties with elution of certain planar or aromatic pesticides from GCB, only PSA is used for very efficient clean-up of acetonitrile extracts. The dSPE with PSA is effectively used to remove many polar matrix components, such as organic acids, certain polar pigments, and sugars. Thus, the PSA clean-up method is selected as the most efficient for cleaning honey samples.
The choice of one sorbent or another depends on the analyte polarity and on the possible co-extracted interferences. Sample pH can be critical to obtain high yields of pesticide retention in the sorbent. Thus, in some cases, sample pH modification can be necessary in order to stabilize the pesticides and increase their absorption in the solid phase. Once the pesticides have been retained in the SPE cartridges, they are then eluted with an organic solvent such as acetone, dichloromethane, ethyl acetate, hexane, methanol, tetrahydrofurane or mixtures of hexane/ethyl acetate, hexane/dichloromethane, and methanol/water or methanol/ ethyl acetate/dichloromethane.
There is also a method based on SPE, the on-column liquid-liquid extraction (OCLLE) or liquid-liquid extraction on a solid support (SLE), a technique based on classical LLE principle, but assisted by inert solid support (Pirard et al., 2007). This inert matrix consists of diatomaceous earth, well-known for its high porosity, its high dispersing capacities and its high capacity for aqueous adsorption. Pesticides in honey were studied by Amendola et al. (2010) using EXtrelutV R NT 20 columns packed with a specially processed wide pore kieselguhr with a high pore volume as support for the repartition process. Honey is dissolved in a mixture of water and acetone and is loaded into an EXtrelutV R NT 20. The majority of the co-extractive compounds are retained on the adsorbent material of the column, while the pesticides are eluted by dichloromethane (Amendola et al., 2010). Other columns can be used such as ChemElut 5 mL cartridges (Kujawski et al., 2014).

Quick Easy Cheap Effective Rugged and Safe
(QuEChERS). The QuEChERS method consists in salting-out LLE using acetonitrile, MgSO 4 , and NaCl salts and a dSPE step based on primary and secondary amine bonded silica (PSA) to remove co-extractive impurities (U.S. EPA, 2013; Barga nska & Namie snik, 2010). The QuEChERS method is particularly applied for the determination of polar, middle polar, and nonpolar pesticide residues in various matrices. This method is combined with sensitive analytical techniques such as LC-MS/MS and GC-MS/MS. Tomasini et al. (2012) demonstrated that the matrix effect depends on the floral origin of honey samples and that quantification by the standard addition method in blank matrix is needed.
With this approach, the sample should be >75% water. Then, an initial dissolution of the honey sample is required. Acetonitrile is used as the water-miscible extraction solvent and the separation phase is achieved by the addition of MgSO 4 . The heat produced by the water binding process promotes extraction to acetonitrile. The addition of NaCl also increases the extraction efficiency (Kujawski et al., 2014). The supernatant is further extracted and cleaned using a dSPE technique. The dSPE centrifuge tube format (available in 2 mL and 15 mL sizes) contains magnesium sulfate (to remove residual water) and PSA sorbent (to remove sugars and fatty acids). These tubes are available with or without GCB (to remove pigments and sterols) and/or C18-EC (endcapped) packing (to remove nonpolar interferences such as lipids). Blasco et al. (2011) showed that the QuEChERS method presented the highest recoveries (mean recovery 91.67%) followed by the SPE (mean recovery 90.25%) whereas the solid-phase microextraction (SPME) showed the lowest recovery (mean recovery of 49.75%) for the pesticides studied. Thus, the QuEChERS method was the most adapted method with around 58% of recoveries >90%.
The QuEChERS multiresidue procedure replaces many complicated analytical steps commonly employed in traditional methods with easier ones. The QuEChERS method has been the most commonly used method for the analysis of pesticides in honey. However, one limitation is that the sample should be >75% water; thus, an initial dissolution of the honey sample is required, which leads to lower concentration of the sample compared to other sample preparation techniques. In order to overcome this limitation, Wiest et al. (2011) added a sample concentration step by evaporation which was satisfactory for extraction of organohalogens, organophosphorus, pyrethroids, and insect growth regulators in honey. According to these authors, evaporation may be necessary when the MRL is lower than the method LOQ. A general protocol is given in the Table 7.

Solid phase microextraction (SPME)
. SPME is a rapid and simple procedure of extraction that can be easily automated and does not need an organic solvent. This technique consists of two separate steps: an extraction step and a desorption step. Both steps must be optimized for the procedure to be successful (Singh et al., 2014). A fused silica fiber coated with a polymeric film is immersed into an aqueous sample for a given amount of time. The analyte enrichment is by partitioning between the polymer and the aqueous phase according to their distribution constant. Factors influencing the extraction step include fiber type, extraction time, ionic strength, sample pH, extraction temperature, and sample agitation. The pesticides are adsorbed into the stationary phase and later thermically desorbed into the injection port of a gas chromatograph (Rial-Otero et al., 2007). Variables affecting the desorption step include temperature, desorption time, focusing oven temperature, the solvent used, and its volume. Most pesticides have been extracted with polydimethylsiloxane (PDMS, 100 mm) fiber.
The PDMS has the following advantages: (i) enhanced reproducibility, (ii) lower detection limits, (iii) extended linearity, (iv) improved correlation coefficients, (v) low extraction time, and (vi) better chromatograms. The use of a sol-gel CROWN ETHERV R fiber (40 mm) was also proposed by Yu et al. (2004) to remove 11 organophosphorus pesticides from honey with good relative recoveries (74-105%) and low detection limits (<0.001 mg/kg). Blasco et al. (2011) employed silica fibers coating with 50 mm carbowax/template resins (CW/TPR) to analyze fenoxycarb, penthoate, temephos, fonofos, diazinon, pyrazophos, phosalone, profenofos, pirimiphos-ethyl, bromophos-ethyl, chlorpyrifos-methyl, and chlorpyrifos-ethyl in honey. The SPME is accurate as a monitoring method for the extraction of the selected pesticides from honey but cannot be implemented as currently applied as a quantification method due to its low recovery for pyrazophos, chlorpyrifosmethyl, temephos, and bromophos-ethyl. The application of the internal standard should be considered.

Stir bar sorptive extraction (SBSE)
. SBSE is a technique theoretically similar to SPME. It has been used with success for the extraction of organic compounds from aqueous food, biological, and environmental samples (Rial-Otero et al., 2007).
For SBSE, a stir bar is coated with a sorbent and immersed in the sample to extract the analyte from solution. The sample is stirred for a given time until the analyte reaches equilibrium between the polymer and the aqueous phase according to their distribution constant. Then, the analytes are desorbed by high temperatures into the injector port of the GC or by liquid removal for liquid chromatography-tandem mass spectrometry (LC-MS) analysis.  have 3. Vortex and incubate in a water bath at 40 C until complete homogeneity is obtained. 4. Add 10 mL of acetonitrile into the tube. 5. Shake the content for 1 min using a mechanical shaker. 6. Add the QuEChERS salt kit (containing 4 g of anhydrous magnesium sulfate, 1 g of sodium chloride, 1 g of sodium citrate and 0.50 g of sodium hydrogen citrate sesquihydrate) and immediately shake for further 1 min. 7. Centrifuge at 15,000 rpm at 4-8 C for 5 min. 8. Transfer the whole acetonitrile fraction into a 15 mL dSPE polypropylene tube (containing 150 mg of anhydrous magnesium sulfate and 25 mg of PSA). 9. Shake the tube for 1 min and centrifuge for 2 min at 15,000 rpm using a cooling centrifuge. 10. Transfer 2 mL of the supernatant into 50 mL round bottom glass flask. 11. Evaporate under vacuum at 40 C till complete dryness. 12. Reconstitute the residue into 2 mL of hexane/ acetone 9:1 (v/v). 13. Ultra-sonicate and filter the sample through a disposable 0.45 mm PTFE membrane filter into an amber glass vial. 14. Analyze by GC-MS/MS following the authors' protocol. Shendy et al. (2016) applied this technique for the extraction of six organophosphorus pesticides from honey. These authors also compared the use of SBSE with SPME and concluded that although linearity and precision obtained by both techniques are similar, SBSE is more accurate and sensitive and the effect of honey matrix in the quantification is lower than that of SPME.
The most important advantages of SBSE are the same as those for SPME; however, higher recoveries are obtained with SBSE  because of the thicker polydimethylsiloxane (PDMS) coating. The procedure is presented in Table 8. Nevertheless, recoveries are less than those obtained using other techniques (LLE, SPE, QuEChERS).

Dispersive liquid-liquid microextraction (DLLME).
In 2006, the use of DLLME was developed in the field of separation science for the preconcentration of organic and inorganic analytes from aqueous matrices. The basic principle of DLLME is the dispersion of an extraction solvent (immiscible with water) and a disperser solvent (miscible in water and extraction solvent) in an aqueous solution that provides a large contact area between the aqueous phase and the extraction solvent. In DLLME, extraction and dispersive solvents are simultaneously and rapidly injected into the aqueous sample using a syringe. The main advantages of DLLME over conventional techniques are simplicity of operation, rapidity, low cost, easy handling, low consumption of organic solvents, high recovery, high factor enrichment, and compatibility with chromatographic techniques such as liquid chromatography (LC) and gas chromatography (GC). However, the QuEChERS method is demonstrated to be more robust and more suitable for the determination of pesticides in complex samples (Tomasini et al., 2011). A simple DLLME protocol for the determination of 15 organochlorine pesticides residues in honey is proposed by Zacharis et al. (2012) and is described in Table 9. The final DLLME protocol involves the addition of 750 lL acetonitrile (disperser) and 50 lL chloroform (extraction solvent) into a 5 mL aqueous honey solution followed by centrifugation. The sedimented organic phase (chloroform) is analyzed directly by GC-IT/MS (gas chromatography-ion trap mass spectrometry) or evaporated and reconstituted in acetonitrile prior to the GC-ECD analysis.
Traditionally nonpolar and middle polar pesticides are analyzed with GC. This is the case for organochlorines, pyrethroids, and organophosphorus pesticides. With the emergence of new pesticides (e.g., neonicotinoids) and due to their physicochemical properties, the LC is used.
For the multiresidue analysis of pesticides, the most convenient detector would be a mass spectrometer (MS/ MS, MS/TOF) coupled with either GC or LC, depending on the type of pesticides of interest. Volatile, semi-volatile, and thermally stable ones can be determined by GC, whereas nonvolatile and/or thermally unstable ones should be determined by LC. When there is a positive result, it has to be confirmed using mass spectrometry coupled with chromatography (GC-MS/MS and LC-MS/ MS). Mass spectrometry coupled to a chromatographic separation method is a very powerful combination for identification of an analyte in the sample extract.
The choice of the GC column is a very important task in pesticide analysis. The stationary phase should be selected as a function of the polarity of the pesticides. Nonpolar columns (5% phenyl, 95% dimethylpolysiloxane) are the most commonly used for pesticide analysis in honey (e.g., DB-5, HP-5MS, and DB-XLB). Furthermore, other column parameters such as length, inner diameter, or film thickness can be optimized as a function of the number of pesticides to be determined simultaneously. Usually, the parameters of the column are: 30 m Â 0.25 mm Â 0.25 mm (length Â inner diameter Â film thickness). In LC, a C18 column (e.g., 4.6 and 2.1 mm i.d.) is almost consensus for the separation of pesticides. Table 8. Analysis of pesticide residues in honey using the SBSE protocol .

Pesticide analyzed
Extraction protocol 6 organophosphorus insecticides (chlorpyrifos-methyl, diazinon, fonofos, phenthoate, phosalone, and pirimiphos-ethyl) 1. Place 2.5 g of honey into a 50 mL glass beaker. 2. Dilute 1/10 ratio with water and homogenize over 15 min using a magnetic stirring bar coated with PDMS. 3. Carry out the sorption for 120 min while stirring at 900 rpm. 4. Remove the stir bar from the aqueous sample with tweezers. 5. Perform the desorption of the analytes into 2 mL vial filled with 1 mL of methanol. 6. Perform the desorption of the pesticides by agitating for 15 min. 7. Inject 5 mL of this extract into the LC-MS system following the authors' protocol.

Antibiotics and chemotherapeutics in honey
Honey is generally considered as a natural, healthy, and residue-free product. However, in the early 2000s, some imported honeys were often contaminated with antimicrobial residues, even with residues of chloramphenicol, a forbidden substance (Reybroeck, 2018). Importers of honey started with regular screening of honey for residues and food authorities increased the number of honey samples at border inspection posts and in national monitoring plans. In the beginning, the monitoring was focused on streptomycin, chloramphenicol, tetracyclines, and sulfonamides. Later, the scope was enlarged to include other compounds. Presently, honey is monitored for a large list of antibiotics and chemotherapeutics of interest in apiculture: tetracyclines (oxytetracycline), aminoglycosides (streptomycin), sulfonamides (sulfamethazine, sulfathiazole, sulfadiazine, sulfamethoxazole, sulfamerazine, sulfadimethoxine), macrolides (tylosin, erythromycin), lincosamides (lincomycin), amphenicols (chloramphenicol), nitrofurans (furazolidone, furaltadone, nitrofurazone), nitroimidazoles (metronidazole), fluoroquinolones (enrofloxacin (ciprofloxacin), norfloxacin), and fumagillin .
In general, residues of antimicrobials in honey originate from apicultural use since bee diseases caused by microorganisms such as American (Paenibacillus larvae, de Graaf et al. 2013) and European foulbrood (Melissococcus plutonius, Forsgren et al. 2013) can be cured by anti-infectious agents. Also nosemosis, caused by spores of the fungi or fungi-related Nosema apis or N. ceranae (Fischer & Palmer, 2005;Fries et al. 2013) is sometimes treated with antimicrobials like sulfonamides. Table 9. Analysis of pesticide residues in honey using the DLLME protocol.

Pesticide analyzed
Extraction protocol References 15 pesticides (etridiazole,chloroneb, propachlor, trifluralin, hexachlorobenzene, chlorothalonil, cyanazine, chlorpyrifos, DCPA, transchlordane, cis-chlordane, transnonachlor, chlorobenzilate, cispermethrin, trans-permethrin) 1. Prepare a solution of honey at 50 g/L (dissolve 10 g of honey in a flask with 200 mL of water). 2. Leave the sample to equilibrate for at least for 15 min prior to performing the DLLME extraction. 3. Transfer an aliquot of 5 mL of the diluted sample into a 10 mL screw cap glass tube with conical bottom. 4. Inject rapidly into the sample solution a mixture of 750 mL acetonitrile (disperser) and 50 mL chloroform. 5. Gently shake the mixture by hand for 1 min. 6. Extract the pesticides and the ISTD (1-bromo decahexane) from the aqueous matrix/phase into the fine chloroform microdroplets. 7. Centrifuge the mixture for 3 min at 2500 rpm for phase separation. 8. Remove the sedimented chloroform volume using a microsyringe. 9. Transfer the extract into an autosampler vial with 50 mL insert and 2 mL of the organic solvent; directly inject this into the GC-IT/MS (GC with ion-trap mass spectrometer detector) following the authors' protocol. 10. For GC-ECD analysis, evaporate 20 mL of chloroform extract (obtained in step 8) to dryness by a gentle stream of nitrogen. 11. Reconstitute the sample by the same volume (20 mL) of acetonitrile. During the extraction, a precipitate formed between chloroform and aqueous phase. 5. Collect the chloroform phase at the bottom of the conical vial with a microlitre syringe. 6. Analyze by GC-MS (2 mL of sample are injected) following the authors' protocol. Kujawski et al. (2012) High levels (mg/kg or ppm) of antimicrobial residues of veterinary drugs applied to hives can be found in honey, especially the first week after dosing . Afterward, the concentration of residues in honey decreases via a dilution effect by incoming nectar and consumption of contaminated honey by the bees. Some compounds (oxytetracycline, tylosin, furazolidone) are also degraded by metabolization. In contrast to other food producing animals, honey bees do not metabolize the drugs so residues could remain in the honey for more than a year , see Table 10.
There are some cases reported of residues of antimicrobials in the honey due to agricultural practices, beekeeping practices, and environmental or fraud issues (Reybroeck, 2014): contaminated nectar from fruit trees treated with streptomycin against fireblight, natural production of streptomycin by certain Streptomyces bacteria, robbery by bees of contaminated honey, feeding bees contaminated honey, mixing clean honey with honey containing residues, the collection by bees of medicated drinking water from farms or surface water from fields where antibiotic-containing manure has been spread, contamination of nectar with sulfanilamide as a degradation product of the herbicide asulam, migration of residues from polluted wax foundation, semicarbazide formed from azodicarbonamide (ADC: a blowing agent used in the manufacturing of plastic gaskets in metal lids), and finally semicarbazide in heather honey formed from elevated arginine levels.
For residue analysis in honey, it is worth noting that some pharmacologically active compounds metabolize or degrade in honey. Thus, it is important to look for the suitable marker residue. For example, honey should be screened for both tylosin A and desmycosin (tylosin B) (Thompson et al., 2007). An overview of the most suitable marker residues for some antimicrobials of interest in beekeeping is given in Table 11. Reybroeck (2018) published a review about residues of antibiotics and chemotherapeutics in honey. This review can be a basis to decide which compounds to analyze in honey.

Legislation regarding residues of veterinary drugs in honey
MRLs were established for residues of veterinary medicinal products in foodstuffs of animal origin in order to protect public health. No MRLs have been established so far for antibiotics and sulfonamides in honey ("Commission Regulation (EU) No 37/2010 of 22 December 2009December ," 2009). This leads to the interpretation that the use of antibiotics in beekeeping is not permitted in the EU. However, based on the "cascade" system which is open to all animal species (including honey bees), antibiotics can be used for the treatment of bee diseases (Anonymous, 2007b) on condition that the active substance concerned is registered as an allowed substance in Table 1 in the Annex of Commission Regulation (EU) No 37/2010, and a withholding period has been specified by the prescribing veterinarian. In this period, no honey could be harvested for human consumption. In reality, a very long withdrawal period needs to be considered.
Some EU Member States (Belgium, France) applied action limits, recommended target concentrations, nonconformity, or tolerance levels for antimicrobial residues in honey . Presently, in application of the Commission Implementing Regulation (EU) 2018/470 for honey produced within the EU, the MRL to be considered for control purpose shall be the lowest of all the MRLs established for other target tissues in any animal species. However, some Member States do not allow treatment with antibiotics and chemotherapeutics under "cascade" in beekeeping which is making international honey trade more complex.
For certain prohibited or unauthorized analytes in food of animal origin, the regulatory limit is the Minimum Required Performance Limit (MRPL) or the Reference Point for Action (RPA). MRPLs were fixed to harmonize the level of control of those substances and to ensure the same level of consumer protection in the Community. So far in honey, a MRPL of 0.3 mg/kg was set for chloramphenicol (European Communities Commission, 2003), while the MRPL of 1 mg/kg for nitrofuran (furazolidone, furaltadone, nitrofurantoin, and nitrofurazone) metabolites in poultry meat and aquaculture products generally is considered as also applicable in honey (SANCO, 2004).
For the importation of products of animal origin from third countries, the MRPLs should be employed where they exist as RPAs to ensure a harmonized implementation of Council Directive 97/78/EC (Commission, 2005).
In a guidance paper of the Community Reference Laboratories (CRLs), recommended concentrations for testing were suggested for the harmonization of the performance of analytical methods for national residue control plans for substances without MRLs (Anonymous, 2007a). An overview of European regulatory limits and recommended concentrations for testing in honey for residues of antibiotics and chemotherapeutics of interest for use in beekeeping is given in Table 11.
In the USA, there are no authorized residue limits for antibiotics in honey despite the authorized use of certain antibiotic drugs (oxytetracycline, tylosin, and lincomycin) in beekeeping (Administration, 2017).
In Europe, for national residue monitoring plans, groups of substances that need to be monitored in honey are indicated ("Council Directive 96/23/EC of 29 April 1996April ," 1996. For honey, it concerns group B1 (antibacterial substances, including sulfonamides and quinolones), B2c (other veterinary drugs -carbamates and pyrethroids), B3a and B3b (organochlorine and organophosphorus compounds), and B3c (chemical elements).
Sampling rules are given in Annex IV of the Council Directive 96/23/EC. The same Directive specifies the frequencies and level of sampling while the Commission Decision 97/747/EC (European Communities Commission, 1997) provides levels and frequencies of sampling. The number of honey samples to be taken each year must equal 10 per 300 t of the annual production for the first 3,000 t of production, and one sample for each additional 300 t. Rules for official sampling procedures and the official treatment of samples until they reach the laboratory responsible for analysis are given in Commission Decision 98/179/EC (Commission, 1998).

Sampling
The analyzed honey samples should be representative of the honey lot.
(1) Crystallized honey should be homogenized before starting the analysis. This can be done by placing the sample in a water bath at maximum 40 C until the honey is fully liquified and all of the sugar crystals are dissolved.
(2) The honey should be homogenized before starting the analysis by stirring thoroughly (at least 2 min). (3) In the event the honey contains extraneous matter, the honey should be strained through a stainlesssteel sieve with a mesh diameter of 0.5 mm. (4) If needed, the honey could be gently pressed through the sieve with a spatula. (5) Comb honey needs to be uncapped and drained through a 0.5 mm sieve without heating in order to separate honey from the comb.

Sample pretreatment
Some honeys require a special sample pretreatment. In honey, sulfonamides tend to bind sugars via the formation of N-glycosidic bonds through their aniline group (Sheth et al., 1990). Therefore, the sulfonamides need to be released from the sugar concentrates by hydrolysis using strong acids (Schwaiger & Schuch, 2000) prior to analysis. Otherwise, it is possible that the sulfonamides are missed or underestimated.

General remarks for honey testing on antimicrobial residues
It is necessary to integrate at least one negative and one positive control sample in each test run. The negative honey is honey free from antimicrobial residues and, by preference, a mixture of the different types of honey (origin, color, and texture) all tested as negative in prior analysis. The positive control sample is prepared by contaminating blank honey with standard material. If commercial kits are used, follow the kit instructions set by the kit manufacturer and validate the method internally before using the method in routine testing. In Europe, validation should be performed according to Commission Decision 2002/657/EC (European Union Commission, 2002. For screening methods, the guidelines published by the Community Reference Laboratories for Residues could also be taken into account (Anonymous, 2010).

Microbiological screening tests
Microbiological screening is not often used for screening honey for antibiotic residues, this due to the very low detection capabilities (CCb) that need to be reached because of the issue of zero-tolerance for antimicrobial residues in honey applied in many countries. The high sugar content in this special matrix makes the use of microbial inhibitor tests also less evident. A broad-spectrum detection of antimicrobials in honey by Eclipse 50 (ZEULAB, Zaragosa, Spain) and PremiTest (R-Biopharm AG, Darmstadt, Germany) was suggested by . In their validation, a high false positive rate of 5 and 14%, respectively, was observed. This study also showed that the detection capabilities are not in line with the action or reporting limits or recommended concentrations for testing for different compounds (e.g., streptomycin) applied in some European countries. The extraction and test procedure for the Eclipse 50 and PremiTest for honey testing are given in Tables 12 and 13, respectively. Noncommercial microbiological methods, two for the detection of tetracyclines and one for the detection of tylosin, were published for honey based on the use of Bacillus cereus (Gordon, 1989), Bacillus subtilis    ATCC6633 (Khismatoullin et al., 2003), and Micrococcus luteus ATCC 9341 (Khismatoullin et al., 2004), respectively. However, the preparation of the test medium and the testing procedures themselves are not fully and clearly described. The tests are utilizing gold conjugated antibodies as signal reagents and a drug protein conjugate as a solid phase capture reagent. As the sample flows through the absorbent sample pad, the liquid reconstitutes the dried monoclonal gold conjugate. The drug in the sample will bind to the conjugate antibody and will migrate further up the membrane to the test line. If there is no drug in the sample, the free antibody conjugate will bind to the test line giving a negative result. In case the sample contains drug residues, the antibody conjugate will not bind to the test line giving a positive result. There is a short sample pretreatment for honey but without the need of instrumentation for most of the tests. With all these test devices, a result is obtained within 3-5 min and visual interpretation of the result is possible. As an example, the test protocol of the Chloramphenicol Residue Rapid Test Device (Nankai Biotech Co., Ltd.) is given in Table 14.
The TetraSensor Honey KIT008 (25 tests)/KIT009 (100 tests) (Unisensor s.a., Li ege, Belgium) sensitively (<10 mg/kg) detects the four most important tetracyclines in honey in 30 min, without any special equipment, making analysis at the production site possible (Alfredsson et al., 2005;Reybroeck et al., 2007; . The test procedure of the TetraSensor Honey is given in Table 15. During this first incubation period, tetracyclines possibly present in the honey bind with the specific receptor forming a stable complex. Afterward, during the second incubation, the liquid is absorbed by the dipstick; and while flowing over the dipstick, the liquid passes through the green capture lines. In case the honey is free from tetracycline residues, the first line captures the remaining active receptor and a strong red line will appear. The second line, serving as a control line, takes a certain amount of the excess of reagent that passed through the first line. A red control line should always become visible; otherwise, the test is invalid. Results can be read visually or by means of a ReadSensor, comparing the color intensity of both capture lines. During this first incubation period, tetracyclines possibly present in the honey bind with the specific receptor. Afterward, the dipstick is dipped into the vial and a second incubation at room temperature occurs for 15 min. When the liquid passes through the green capture lines, a red color appears. The first line captures the remaining active receptor and the second line takes a certain amount of the excess reagent that passed through the first line. The second line serves as a control line and always has to become visible; otherwise the test is invalid. Results can be read visually or by means of a ReadSensor, comparing the color intensity of both capture lines. Only limited laboratory equipment is required to run the Sulfasensor Honey KIT033 (Unisensor s.a.), a generic monoclonal antibody test, for the detection of sulfonamides in honey in 20 min. A sample pretreatment (acid hydrolysis) by heating a mixture of honey sample and buffer at 95 C for 5 min is needed to release the sulfonamides that are chemically bound to the sugars (Chabottaux et al., 2010;Gaudin et al., 2012;Reybroeck & Ooghe, 2010). The kit manufacturer claims a detection of several sulfa drugs at 25 mg/kg in honey. The extraction procedure and test protocol are shown in Table 16.
With the competitive multiplex dipstick Bee4Sensor KIT059 (Unisensor s.a.), the screening for tylosin, (fluoro)quinolones, sulfonamides, and chloramphenicol in honey is possible (Heinrich et al., 2013). The test could be used in two different ways: as a field test or as a lab test with better detection capabilities. The sample extraction and test protocol of the field test is described in Table 17. To run the lab test, one aliquot (A) is dissolved using acid hydrolysis, whereas the other aliquot (B) is dissolved in water. After liquid/liquid partitioning of both aliquots with ethyl-acetate, the organic layers are evaporated until dry under nitrogen. After reconstitution in a buffer, aliquots A and B are combined and applied to the well of a Bee4Sensor test kit for 5 min at 40 C, as one sample extract. Afterward, a dipstick is then incubated in this prepared well for 15 min at 40 C. The dipstick can be assessed visually or instrumentally via the ReadSensor.

Enzyme-linked immunosorbent assays (ELISA)
There are many commercial and noncommercial ELISA's developed for the detection of antibiotic residues in different matrices, including honey. Table 18 summarizes existing ELISA kits (the list is nonexhaustive). The sample preparation for honey ranges from very simple, just a dilution followed by a filtration, to more complex, including acidic hydrolysis, derivatization, solvent extraction, or purification on a SPE column. The technical brochures of the different kit manufacturers should be followed. Validation studies of ELISA kits have been published regarding the detection of chloramphenicol (Scortichini et al., 2005), tylosin and tilmicosin (Peng et al., 2012), nitrofuran metabolites (Elizabeta et al., 2012), and tylosin en streptomycin   (Figure 8).
Competitive enzyme immunoassays are used in most cases for the detection of small molecular weight components such as antimicrobial residues. Such assays use either an enzyme-linked antibody or an enzymelinked analyte to detect a particular antigen (drug). As an example, the sample preparation and test procedure for the detection of chloramphenicol in honey by means of the Chloramphenicol ELISA of EuroProxima (Arnhem, The Netherlands) is shown in Table 19. In this kit, the walls of the wells of the microtiterplate are precoated with sheep antibodies to rabbit IgG. A specific antibody (rabbit anti-CAP), enzyme labeled CAP (enzyme conjugate) and CAP standard or sample are added to the precoated wells followed by a single incubation step. The specific antibodies are bound by the immobilized antibodies. At the same time, free CAP (present in the standard solution or sample) and enzyme conjugated CAP compete for the CAP antibody binding sites. After an incubation time of 1 h, the nonbound (enzyme labeled) reagents are removed in a washing step. The amount of CAP enzyme conjugate is visualized by the addition of a chromogen substrate. The bound enzyme conjugate elicits a chromogenic signal. The substrate reaction is stopped to prevent eventual saturation of the signal. The color intensity is measured photometrically at 450 nm. The optical density is inversely proportional to the CAP concentration in the sample.
In some laboratories, noncommercial ELISA methods are used to monitor honey for the presence of antimicrobial residues (Heering et al., 1998;Jeon & Rhee Paeng, 2008). A microplate and magneto iELISA was developed for the detection of sulfonamides in honey using magnetic beads to reduce nonspecific matrix interferences (Muriano et al., 2015).

Enzyme-linked aptamer assays (ELAA)
Assays were developed using DNA or RNA aptamers to get specific binding to streptomycin (Zhou et al., 2013) or tetracyclines (Wang et al., 2014) in honey.

Radio-labeled receptor/antibody techniques (Charm II tests)
The Charm II (Charm Sciences Inc., Lawrence, MA) is a scintillation-based detection system for chemical families of drug residues utilizing class-specific receptors or an antibody in immune-binding assay format. The sample preparation is mostly just a dilution of the honey with an extraction buffer supplied as part of   (McAleer et al., 2010). The Biacore biosensor system is based on surface plasmon resonance. Analytical Qflex Kits are offered for use with the Biacore Q-instrument for screening for the antibiotics chloramphenicol, streptomycin, sulfonamides, and tylosin (Caldow et al., 2005). The sample preparation includes dissolving the honey in an aqueous buffer and filtering. Analysis for chloramphenicol requires an extra LLE. For sulfonamides, a hydrolysis step is needed to release the sulfonamides bound to sugars. With this system, a high throughput and a rapid (around 5 min) multi-analyte screening in honey is possible (Weigel et al., 2005). However, the instrument costs are high.
Randox Laboratories Limited offers a multi-analyte quantitative testing platform, the Evidence Investigator using Biochip Array Technology for the monitoring of honey for antimicrobials (Daniela et al., 2012;Gaudin et al., 2014;O'Mahony et al., 2010;Popa et al., 2012). An overview of available reagents for the monitoring of honey on antimicrobials is given in Table 21. The sample preparation typically takes about 20 min for dilution. This does not include the Antimicrobial Array III which requires 4 h due to the derivation step or 50 min if CAP only due to extraction step. The incubation and assay times take 2 h for all arrays. Arrays for the detection of nitroimidazoles in honey are in development.

Methods using MIPs and other immunotechniques
Molecularly imprinted polymers (MIP) with specific recognitions sites are sometimes used as a preconcentration step as for the determination of chloramphenicol in honey (Thongchai et al., 2010). MIP sensors based on electropolymerized oligothioaniline-functionalized gold nanoparticles are prepared and applied to the detection of tetracycline in honey (Bougrini et al., 2016). Furthermore, there are publications on the development of a immunosensor for the detection of residues of sulfonamides in honey Valera et al., 2013).

Physico-chemical methods (chromatographic techniques)
Methods for the detection of residues of veterinary drugs in honey based on chromatographic techniques have been described in several publications (Benetti et al., 2004(Benetti et al., , 2006Carrasco-Pancorbo et al., 2008;Dubreil-Ch eneau et al., 2014;Edder et al., 1998;Kaufmann et al., 2002;Kivrak et al., 2016;Maudens et al., 2004;Nozal et al., 2006;Sporns et al., 1986;Van Bruijnsvoort et al., 2004;Verzegnassi et al., 2002) and in a review article (Barga nska et al., 2011). In the 1990s, the use of HPLC was popular; but today, confirmation of antibiotic residues in honey is performed by LC-MS, mainly LC-MS 2 (tandem mass spectrometry) (Blasco et al., 2007). When mass fragments are measured using techniques other than full scan, the system of identification points (IP) is applied. Confirmatory methods must fulfil the criteria listed in Commission Decision 2002/ 657/EC (European Union Commission, 2002) and must be based on molecular spectrometry providing direct information concerning the molecular structure of the Some compounds such as nitroimidazoles (Polzer et al., 2010), fumagillin Kanda et al., 2011;Tarbin et al., 2010), and nitrofuran residues (Khong et al., 2004;Lopez et al., 2007;Tribalat et al., 2006) are mostly directly screened in honey using LC-MS/MS detection, since no or only few immunochemical methods for the detection in honey have been developed. By the development of multiresidue methods, LC-MS is used more and more for multiclass screening for antimicrobial residues in honey (Azzouz & Ballesteros, 2015;Galarini et al., 2015;Hammel et al., 2008;Lopez et al., 2008). Table 19. Extraction and test protocol (laboratory test) for the Chloramphenicol ELISA (EuroProxima B.V.). Extraction protocol 1. Bring 10 ± 0.1 g of honey into a centrifugation tube of 50 mL; fill also a centrifugation tube with the same amount of blank honey, honey doped with 0.3 mg/kg of chloramphenicol, and honey doped with 0.1 mg/kg of chloramphenicol, respectively. 2. Add to each sample 5 mL of distilled water. 3. Vortex the samples until all honey is dissolved. 4. Add 5 mL of ethyl acetate. 5. Mix (head over head) for 15 min. 6. After centrifugation (5 min at 2,600 g), pipette 2 mL of the upper layer (ethyl acetate) into a glass test tube. 7. Evaporate at 45 ± 5 C under a mild stream of nitrogen. 8. Dissolve the residue in 2 mL of n-hexane for defatting purposes and vortex. 9. Add 1 mL of reconstitution buffer and vortex for 20 s. 10. After centrifugation (10 min at 2600 g), pipette the layer underneath (reconstitution buffer, ± 800 mL) into a short glass test tube. 11. Add again 1 mL of n-hexane and vortex for 20 s. 12. Add 1 mL of reconstitution buffer and vortex for 20 s. 13. After centrifugation (10 min at 2,600 g), pipette the layer underneath (reconstitution buffer, ± 800 mL) into a short glass test tube. 14. Use the extract in the ELISA (extract could be stored for one day at 4 ± 2 C).

Standard method for honey sensory analysis
Sensory analysis of honey describes the organoleptic profile and may be used to assess the quality, as well as the botanical and geographical origin of a honey sample. Furthermore, it may be used to assess the pleasantness to the end-consumers and to monitor the quality of honey during harvesting, packaging, and storage. In order to evaluate uniflorality, "honey may be designated according to floral or plant source if it comes wholly or mainly from that particular source and has the organoleptic, physicochemical and microscopic properties corresponding with that origin" (Codex 2001). It should be noted that some pollen is either under or over-represented in a honey sample and the percentage of one specific pollen does not necessarily correspond to the amount of nectar from that specific species. Moreover, the pollen entering the nectar during bee collection (primary pollen spectrum, strictly related to botanical origin) can be significantly modified from secondary pollen contamination, that is, pollen grains entering in honey in the hive, during transformation from nectar to honey by bees, or as a consequence of beekeeper operations. Using only pollen analysis for determining uniflorality may be misleading and a unique generalized threshold may be too high for under-represented and too low for over-represented pollens (see, e.g., the descriptive sheets Person Oddo & Piro, 2004). Thus, sensory evaluation, in addition to physico-chemical and melissopalinological analyses, is essential for determining uniflorality, also because organoleptic characteristics are the only ones that consumers can identify and evaluate.
Traditional sensory analysis was first applied to honey by Michel Gonnet who trained specialists to evaluate honey on the basis of their sensory experience (Gonnet et al., 1985;Gonnet & Vache, 1979, 1992. Modern techniques relying on a panel of assessors and controlled protocols are applied more regularly now . One selected technique for the sensory analysis of honey is presented here, while a wider guide is published in Marcazzan et al. (2018).
The following proposed standard method will be developed by referring to specific articles (Marcazzan et al., 2014;Mucignat-Caretta et al., 2012;Piana et al., 2004;Sabatini et al., 2007) or to ISO standards (ISO 5492, 2008;ISO 6658, 2017). For detailed information regarding testing rooms and the selection and training of assessors or technicians, one can refer to the specific ISO standards (ISO 8586, 2012;ISO 8589, 2007). For a complete list of ISO standards on sensory analysis, see the ISO catalog (http://www.iso.org).

Methods
Many methods for the sensory analysis of honey are available. The general requirements and tasting procedures are common to all methods. The descriptive semiquantitative method will be described in detail, as it requires less training and has a wider use; the other methods are described in Marcazzan et al. (2018), and references herein.

Test room
Evaluations should be conducted under known and controlled conditions with minimal distractions. Individual testing booths should be used ( Figure 9) and located in rooms with controlled lighting and temperature, without interfering noises and odors. The general guidance for the design of test rooms is reported in ISO 8589 (2007).

Selection and training of assessors and choice of panel
The guidelines for the selection, training, and monitoring of assessors are reported in ISO 8586 (2012). Panel sessions involve at least five specialized expert assessors working as part of a panel managed by a panel leader (ISO 13300, 2006a(ISO 13300, , 2006b) who oversees the selection, training, monitoring, and performance of the assessors. The panel leader is responsible for organizing and coordinating the tests as well as collecting and processing the results.

Product storage
Recommended temperature for honey sample storage should be <20 C to maintain the stability of honey. Avoid high temperatures (>35 C) and exposure to sunlight.

Preparation and presentation of the samples
The sensory analysis is performed in containers as detailed below. It is useful to separate the visual from the olfactory-gustatory assessment: general aspect and certain defects should be assessed in the original container if possible. In any case, the container should not have any sort of identification except a code. Otherwise, honey samples must be presented in the containers mentioned below.

Sampling glasses.
The same type of glasses or containers should be used within each tasting session. The glasses/containers should meet the following requirements: Containers should be odorless. Present samples in a homogeneous and anonymous way (identical containers with no distinguishing marks apart from the identification code). Cover samples to prevent the contamination and dispersion of the honey odors.
Use a 150/200 mL capacity balloon stemmed wine glass for samples ( Figure 10) or other types of glasses if they satisfy the requirements mentioned above, maintaining the sample/volume ratio near 1/5. Use colored or opaque containers if it is necessary to mask the color of the honey. For example, oil sampling beakers defined in the COI/T.20/Doc. No. 5 norm (International Olive Oil Council, 2007) and ISO standard (ISO 16657, 2006) meet these requirements. The color of honey may be masked by using a red light.

Preparation of the samples.
(1) Assign each sample a random three-digit code.
(2) Pour at least 30 g of the sample (suggested for a 150 mL glass, one for each assessor) into the container. (3) Cover samples with a suitable lid (e.g., petri dish, aluminum foil, cling film). (4) Use minimal manipulation during transferring of samples to ensure that the subsamples look the same. (5) To guarantee anonymity, the technician preparing the samples should not perform the testing. (6) Maintain the sample at room temperature, preferably between 20 and 25 C. (7) The evaluation should ideally occur within 24 h after sample preparation. Figure 9. Assessors in testing booths. Figure 10. Balloon stemmed glasses proposed for sensory honey evaluation. Table 22. Guidelines for assessors. 1. In the day of assessment, the assessors should: avoid perfumed toiletries, and avoid anything else that may introduce odors into the tasting room.

Tasting procedures 7.3.1. Guidelines
Each assessor will be given instructions and materials, including evaluation forms and computers, necessary to conduct the sensory analysis. The assessors must have at their disposal neutral spoons (without any odor and/or taste, metallic material should not be used) to stir the honey during the olfactory assessment and to take the product to be tasted. Water and a juicy, slightly acidic apple are suggested to cleanse the mouth between tastings. Also rose-hip tea or low-salt bread may be used. The assessors must follow simple but important rules (ISO 6658, 2017;Piana et al., 2004), see Table 22.
(1) Limit the number of samples according to the type of honey to prevent oversaturation and fatigue of the senses. (2) The evaluation should occur when sensory sensitivity is at its optimum, at least 2 h after a meal. (3) Rest at least 30 min between sessions. (4) Randomly present the order of the samples to each assessor.

Evaluation of the visual characteristics
(1) The evaluation is carried out under day light in a well illuminated room or using an artificial illumination as lights with a correlated color temperature of 6 500 K. (2) Evaluate the visual cleanliness and quality of the crystallization of a honey in its original container. (3) Evaluate all other characteristics by assessing the sample in a separate glass.

Evaluation of the olfactory characteristics
Evaluate olfactory characteristics before tasting. Avoid receptor desensitization by taking short sniffs of the sample.
(1) Spread the honey around the inside of the glass using the spoon. This increases the surface area and releases the volatiles. (2) Take first sniff to evaluate odor immediately.
(3) Wait 15-20 s before taking a second sniff, so that the full intensity of the odor is perceived. (4) Leave part of crystallized honey untouched for tactile evaluation.

Evaluation of the olfactory-gustatory and tactile characteristics
(1) Bring to the mouth 1/2 g of honey with a spoon.
(5) Evaluate the retro-nasal odor: aroma intensity and quality, persistence, after-taste. (6) Evaluate other mouth sensations (viscosity, texture, type of crystals). (7) Concentrate on the chemical aspects of the mouth sensations apart from the tactile characteristics. (8) Wait 1 or 2 min before taking a second taste, longer in the case of persistent honeys. Eat a slice of apple or bread or cleanse the mouth with water or rose-hip tea if necessary. (9) Evaluate tactile characteristics simultaneously. We encourage doing so in a second taste. Press the sample between tongue and palate in order to perceive and assess the viscosity, texture, and crystals characteristics. (10) Wait a few minutes between samples, eat a slice of apple or bread or cleanse the mouth with water or rose-hip tea.

Descriptive semiquantitative analysis
All numbered lists correspond to a scale, while bulleted lists correspond to a nonexhaustive list of descriptors. An example of descriptive card is presented in Figure 11.

Appearance and visual descriptors
The principal visual descriptors for honey are: (1) Physical state: liquid or crystallized.
(2) Color intensity: 1. very light; 2. light; 3. medium; 4. dark; 5. very dark. (3) Color tone: normal honey color; bright yellow; grayish; reddish; with orange tone; fluorescent green; dull; bright. (4) Presentation: depending on the aim of the evaluation (e.g., competition or checking botanical declaration etc.), the importance of the assessment of some or all of these elements may vary. The following characteristics can be assessed: clarity (for liquid honey); impurities; air bubbles or foam. (5) Defects can be assessed as marbling or white spots, incomplete crystallization, phase separation.
Viscosity, cohesion and type of crystals are observed in both visual and tactile evaluation but assessed only in the latter (see Section "Texture and other tactile descriptors" for description). (1) The intensity of the odor may be defined as: 0absent; 1-weak; 2-medium; 3-strong. (2) The description of the odors may be classified in families and subfamilies. These are shown in Table  23. The examples used in the table to describe the families should be considered useful but not exhaustive. Figure 11. Example of descriptive card.

Flavor descriptors
Flavor is the "complex combination of the olfactory, gustatory and trigeminal sensations perceived during tasting. It may be influenced by tactile, thermal, painful and/or kinaesthesic effects" (ISO 5492, 2008).
(1) The intensity of the taste sensation sweetness, sourness, salty, and bitterness (Table 24) is quantified according to the scale: 0. absent; 1. weak; 2. medium; 3. strong. The taste of umami is not considered; if perceived, it is reported as present. (2) The odor perceived via retro-nasal pathways (aroma) is described as reported in olfactory description. (3) Other sensations we can perceive in the mouth include: (i) piquant: an aggressive and burning sensation perceived by the mucous membranes and/or in the throat; (ii) metallic: a feeling similar to a slight tingling sensation due to the presence of traces of iron (or perhaps other metals) in ionic form in some honeys. In honey, it indicates a defect. (4) Persistence can be evaluated. Persistence refers to the duration of the aroma sensations after swallowing. Less than 30 s is considered "short" persistence, and longer than 5 min, "long" persistence. Persistence is evaluated according to the scale: 0. absent; 1. short; 2. medium; 3. long. (5) Aftertaste indicates the presence of aroma sensations in the mouth that remain after swallowing the product and are different from those initially perceived. The quality of the sensations can be described with the same terminology used to explain the olfactory sensations (Table 23). Floral Odor reminiscent of flowers orange flower (distilled) water, b-ionone (subfamily "fine") rose, ylang ylang, phenylacetaldehyde (subfamily "heady") 2 Fruity Odor reminiscent of fruit fruit aroma (strawberry, pear, apple, raspberry) (subfamily "fresh fruit") isoamyl acetate, c-decalactone, d-decalactone (subfamily "tropical fruit") grape sugar (concentrated rectified must) (subfamily "syrup") apricot fruit juice, peeled tomatoes (subfamily "processed fruit") white wine (subfamily "fermented fruit") 3 Warm Odor reminiscent of food characterized by sweet taste vanillin, benzoin, beeswax (subfamily"fine") concentrated sugared milk, diacetyl (subfamily "lactic") cane sugar, furanone (subfamily "candied") toasted hazel nuts, toasted almonds, toasted peanuts (subfamily "toasted") barley malt syrup, rice syrup (subfamily "malted") toasted barley, toasted bread (charred part) (subfamily "burnt") 4 Aromatic Odor reminiscent of something fresh, balsamic, cosmetic, not necessarily related to food products cloves, nutmeg (subfamily "spicy") mugo pine, incense (subfamily "resinous") oak musk (subfamily "woody") peppermint, anise, eucalyptus (subfamily "balsamic") lemon, sweet orange (subfamily "citrus fruit") coumarin (subfamily "bitter almond") 5 Chemical Odor generally not related to food products, characterized by a certain degree of aggressiveness phenol, vegetal tar (subfamily "phenolic") castile soap (not perfumed) (subfamily "soap") jute smoke, cigarette butt extract (subfamily "smoky") white wine vinegar (subfamily "vinegar") ammonia (subfamily "ammonia") 6 Vegetal Odor reminiscent of nonaromatic plants fresh raw beans, cut grass, fresh ivy leaves (subfamily "green") raw champignon mushrooms, lightly boiled spinach (subfamily "damp") green tea, hay (subfamily "dry") 7 Animal Odor reminiscent of animal world or decomposition hardboiled egg, dimethyl sulphide, methionine (subfamily "sulfurized") dried ceps mushrooms, stock cube (subfamily "proteic") isovalerianic acid, butyric acid, hexanoic acid (subfamily "valerianic") cassis (subfamily "cassis") A further tactile descriptor for honey is astringency, a complex sensation due to substances such as the tannins of persimmon or unripe fruit, which precipitate the proteins in the saliva responsible for lubricating the mucous membranes in the mouth. The astringency is perceived as a lack of lubrication and dryness of mucous membranes; it is often associated with a bitter taste.
In finely crystallized honey, it is common to experience a refreshing sensation. This is not due to the temperature of the product, but it is linked to the solubility of some type of crystals (glucose).

Acknowledgments
The following authors were responsible for the respective manuscript sections: Ligia Bicudo de Almeida-Muradian, (Introduction); Ortrud Monika Barth and Alex da Silva de Freitas (Honey authenticity-botanical, geographical, and entomological origin); Ligia Bicudo de Almeida-Muradian, Maria Teresa Sancho, Ana Pascual-Mat e, Jos e Augusto Gasparotto Sattler (Honey physicochemical analysis for nutritional, chemical composition, and antioxidant activity); Michael Eyer and Vincent Dietemann (Investigating honey production with diagnostic radioentomology); Anne-Claire Martel (Pesticide residues in honey); Wim Reybroeck (Antibiotics and chemotherapeutics in honey); Gian Luigi Marcazzan, Carla Mucignat, Carla Marina Marchese (Sensory analysis). The authors thanks Lucia Piana for participation in the Sensory Analysis section. Ana Pascual-Mat e and Maria Teresa Sancho-Ortiz are very grateful to Cipriano Ramos and M onica Pascual for their valuable help. Ligia Bicudo de Almeida-Muradian thanks Stefan Bogdanov and Peter Gallmann for their technical support and also Adriane Alexandre Machado de Melo. We further thank Leila Aparecida Bonadio for helping with the reference management.
The COLOSS (Prevention of honey bee COlony LOSSes) Association aims to explain and prevent massive honey bee colony losses. It was funded through the COST Action FA0803. COST (European Cooperation in Science and Technology) is a unique means for European researchers to jointly develop their own ideas and new initiatives across all scientific disciplines through trans-European networking of nationally funded research activities. Based on a pan-European intergovernmental framework for cooperation in science and technology, COST has contributed since its creation more than 40 years ago to closing the gap between science, policy makers and society throughout Europe and beyond. COST is supported by the EU Seventh Framework Program for research, technological development, and demonstration activities (Official Journal L 412, 30 December 2006). The European Science Foundation as implementing agent of COST provides the COST Office through an EC Grant Agreement.