Technical Reports
地理的表示保護あるいは原産地呼称保護表示のあるギリシャ産リンゴにおける残留農薬
2014 年 39 巻 1 号 p. 29-35
詳細
2014 年 39 巻 1 号 p. 29-35
Consumers’ demand for superiority in agricultural products and foodstuffs followed by quality guarantees, have led to the introduction of certification labels. Among others, the European Commission enforces two certification labels: Protected Designation of Origin (PDO) and Protected Geographical Indication (PGI) (EEC Regulation No. 2082/1992).1) A PDO/PGI label guarantees the authenticity of products and helps protect them from those selling replicate products trying to benefit from the image of PDO/PGI products (European Commission, Agriculture and Rural Development). It should be noted that PDO/PGI agricultural products are exported to numerous-mainly European, Asian and American-countries. Furthermore, a recent study showed that consumers of regional products are willing to pay a premium for regional products with PDO/PGI labels, as they consider that PDO/PGI labels guarantee their authenticity, safety and high quality.2,3) However, in spite of this belief, PDO/PGI products are not necessarily free of pesticide residues. For example, in our recent investigation concerning the determination of pesticide residues in Greek PDO/PGI olive oils, 38 of the 70 investigated samples were positive; the highest detection rates were found for penconazole, α-endosulfan, β-endosulfan and flufenoxuron.4)
The domesticated apple (malus domestica), member of the plant family Rosaceae, is one of the most widely cultivated tree fruits. There is considerable variation among cultivars for most traits, such as the yield and the characteristics of the fruit as well as their resistance to biological and environmental conditions.5) The role of apples in human health has been extensively documented. Among others, apple polyphenols, such as hydroxycinnamic acids, dihydrochalcones, flavonols and catechines, help protect against increased blood pressure and hyperlipidemia6) they may also reduce the risk of colon cancer.7) In order to highlight the value of its specific features, apples are included in PDO/PGI labeling.
As is true with every agricultural product, the beneficial effects of apples on human health can be balanced by the risk of consumers’ exposure to the residue of pesticides applied to apple trees. It is well-known that crops are attacked by a variety of insect pests, such as apple maggots (Rhagoletis pomonella), codling moths (Cydia pomonella) and plum curculio (Conotrachelus nenuphar), which can reduce yields or even destroy them.8–10) In order to protect apple crops, pesticides are widely applied. To protect consumers’ health the European Commission along with Codex Alimentarius (Codex Alimentarius Commission, International Food Standards, http://codexalimentarius.org.) of the Food and Agriculture Organization (FAO) have established MRLs for several pesticides in apples (Regulation (EC) No. 396/2005).11) Limited previous studies have shown that major contaminants in Greek apple samples were chlorpyrifos-methyl, methomyl and parathion ethyl,12) while residues of other organophosphorus insecticides (diazinon, malathion) have also been detected.13) Less investigated is the presence of pesticide residues in exportable PDO/PGI apples; data in such products are very rare.
In the light of the above considerations, we found it interesting to perform an investigation in Greek exportable PDO/PGI apple samples with these aims; (i) to explore their pesticide residue content with respect to evaluating consumers’ risk, extending our work to the monitoring contaminants in Greek PDO/PGI agricultural products of wide consumption, (ii) to study the evolution of the presence of pesticides in Greek apples and (iii) to investigate possible correlations between residues in olive oil and apple samples. For this purpose, a total of 80 samples from 4 PDO/PGI apples exported to other countries were collected and analyzed for the presence of 51 pesticide residues by means of gas chromatography coupled with mass spectrometry (GC-MS). The results were compared with the corresponding MRL values and pesticide residue levels were subjected to statistical analysis with the goal of unraveling possible correlations.
A total of 80 apple samples, 20 samples of each of the 4 PDO/PGI apples under study, were collected from local markets in Greece. In particular, three apple types with protected designation of origin status (total number of samples equal to 60) and one apple type with protected geographical indication (20 samples) were examined. These investigated apple types correspond to all exportable PDO/PGI apples available in local markets. The origination of the apple samples is summarized in Table 1 and their exact production locations are given in Fig. 1. For purposes of comparison, all apple samples collected correspond to the crop period of 2011–2012. After collection, each sample was packaged in a polyethylene bag and transported in a cooling box with ice packs to the laboratory. There, it was immediately stored in a freezer at −20°C until further analysis.
Three pesticide-free apple samples, one commercially supplied and two collected by organic farming were chosen; they were spiked with increasing concentrations of standard solutions of pesticides for calibration process. The same samples were also used as matrices in the method validation procedure.
2. ReagentsPesticide standards, in a purity range between 97.0% and 99.0%, were supplied by Riedel-de Haën (Seelze, Germany). Stock standard solutions of each pesticide were prepared in ethyl acetate in a concentration range of 200–1000 mg/L and stored in tapered bottles at −20°C, protected from the light. Working standard solutions, in a concentration range of 1–10 mg/L, were obtained daily by appropriate dilution of the corresponding stock solutions using ethyl acetate.
Ethyl acetate (≥99.8%), acetonitrile (≥99.9%), toluene (≥99.8%) and hexane (≥99.0%) were high purity grade solvents for pesticide residue analysis that were supplied by Sigma-Aldrich. Analytical grade dodecane (≥99.8%) was supplied by Fluka. Analytical grade anhydrous magnesium sulfate (≥97%) and sodium citrate dibasic sesquihydrate (≥99.0%) as well as ACS grade sodium chloride (≥99.5%) and sodium citrate tribasic dehydrate (≥99.5%) were obtained from Sigma-Aldrich. Dispersive-SPE sorbents, Supelclean primary secondary amine (PSA) and graphitized carbon black (GCB) (Discovery) were also purchased from Sigma-Aldrich. High Purity Water was obtained by means of an EASYpure II (Barnstead International, USA) water purification system.
All glassware used for the experiments had been carefully washed previously with acetone (ACS, Merck) and high purity water.
3. Sample preparation and extractionThe apple samples were first homogenized in a food processor. The extraction of pesticide residue was carried out according to a modification of the “quick, easy, cheap, effective, rugged and safe” (QuEChERS) method.14) Briefly, 10 g of homogenized apple was transferred into a 50-mL centrifugation tube and 10 mL of acetonitrile (containing 1% acetic acid), 4 g of anhydrous MgSO4, 1 g of NaCl, 1 g of trisodium citrate dehydrate and 0.5 g of disodium hydrogen citrate sesquihydrate were added. The tube was vigorously shaken for 1 min and it was centrifuged at 3000 rpm for 5 min. An aliquot of 6 mL of the supernatant was then transferred into a 15-mL centrifuge tube and cleaned up by dispersive solid-phase extraction using 0.9 g of anhydrous MgSO4, 150 mg of PSA and 15 mg of GCB. The tube was again vigorously shaken for 2 min and centrifuged at 3000 rpm for 5 min. Finally, 2 mL of the supernatant liquid was transferred into a vial, 0.040 mL of dodecane was added and the mixture was evaporated to dry at 30–40°C. The residue was redissolved with 0.50 mL of a 1 : 1 mixture of toluene–hexane and after filtration (0.20 µm), it was inserted into the autosampler vial.
4. Gas chromatography analysisThe GC system comprises a Thermo Scientific (model trace GC ultra) gas chromatograph equipped with a DSQ II mass spectrometer. A ZB 50 (50% phenyl-, 50% dimethylpolysiloxane) (30-m length, 0.25 mm i.d. and 0.25 µm film thickness) chromatographic column was employed.
Helium (99.9% purity) with a flow rate electronically adjusted at 1.2 mL/min was used as the carrier gas. An oven temperature was programmed as follows: 50°C for 1 min, 30°C/min to 180°C, 1.8°C/min to 230°C, 30°C/min to 280°C and maintained at this temperature for 6 min, 50°C/min to 300°C and maintained at this temperature for 2 min. The temperature of the ion source was maintained at 200°C.
The GC system was operated in a solvent flush post mode and the injection volume was 1 µL for each injection. Xcalibur 2.1. software (Thermo Fisher Scientific Inc., USA) was employed for chromatographic data processing. The retention times (tR) of the 51 pesticide residues under study along with their selected ions are presented in Table S1, in the supplementary material. As an internal standard for quantitative analysis using matrix-matched calibration standards mirex was selected because it was found in preliminary experiments to have quantitative recoveries and it had not been detected in apple samples in previous studies.
Pesticide residues in apple samples were identified when the following criteria were met: the chromatographic peaks of both species- the unknown and the standard samples- coincided at the same relative (to internal standard) retention time (±0.5%) and the ratio between the selected ions (Table S1) were the same. The initially positive apple samples were subjected to further confirmation analysis by employing spiked apple samples, according to the SANCO/10684/2009 recommendations.
5. Method validationThe described analytical procedure was validated according to the SANCO/10684/2009 validation protocol for analytical techniques for pesticide residue analysis in food and feed. This procedure fulfills the European Decision 2002/657/EC requirements. Linearity was evaluated by employing matrix-matched calibration curves in a concentration range from the limit of quantification (LOQ) of each analyte to a maximiun concentration of 500 µg/kg.
6. Statistical analysisThe Spearman correlation coefficients for pesticides levels of the investigated apple samples were obtained using SPSS, version 14.0, for windows.
Principal component analysis (PCA) was performed using Simca-P 10.5 (Umetrics, USA).
The so-called “quick, easy, cheap, effective, rugged, and safe” (QuEChERS) approach was applied in the present study, as it has become the method of choice for the rapid extraction and cleanup of a wide range of samples for the quantification of pesticide residues.14–17) This approach proved to be effective in minimizing the coextraction of lipids, even from lipid matrices due to the low solubility of lipids in acetonitrile,16,18) while the use of anhydrous MgSO4 and NaCl contribute to obtain high recoveries for many pesticides.18)
In our previous work concerning the determination of the 51 pesticide residues in PDO/PGI olive oil samples,4) their recoveries in a range of pesticide concentrations between 10 and 250 µg/kg were investigated. In the present work, the recoveries of the 51 target pesticide residues had to be revisited prior to the application of the QuEChERS technique for their quantification to the PDO/PGI apple samples under study. For this purpose, pesticide-free apple samples were spiked with the target pesticides in four different concentrations (10, 25, 50 and 250 µg/kg), covering a wide concentration range of pesticide residue that can be found in apple samples. The mean recoveries of pesticides along with their standard deviation, obtained from 3 replicates, in relation to the 4 spiked concentrations are presented in Table S2, in the supplementary material. As shown, most recoveries were quantitative and they can be considered as persistent (RSD <15%). Some pesticides, such as hexachlorobenzene, heptachlor, aldrin and 4,4′DDE, exhibited lower recoveries, ranging from 75 to 85%. Low recoveries for these pesticides have also been reported by other authors mainly in matrices with high lipid content,15,16,18) as well as in our previous study of olive oils.4) This can be mainly attributed to their high lipophilicity, expressed mainly by the logarithm of octanol-water partition coefficient (log Pow). Details as to the role of lipophilicity in chemical and biological interactions can be found in more specialized publications.19–21) However, in the present study, recoveries of all investigated pesticides are generally higher than those obtained in olive oils. Such high pesticide recoveries in low lipid matrices have also been reported in related publications.22,23) This can be explained on the basis that the lower lipid content of the matrices implies lower retention of the analytes to the matrix lipids and, therefore, lower losses.
Limits of detection (LOD) and limits of quantification (LOQ) were calculated from a signal-to-noise ratio of 3 and 10, respectively, by spiking at low concentrations pesticide-free apple samples and submitting them to the same analytical procedure. Matrix-matched calibration standards were prepared by adding known pesticide amounts in concentration levels between LOQ and 500 µg/kg. The results were calculated by plotting the area ratios of the residues under study versus mirex, which was selected as the internal standard. As shown in Table S3, in the supplementary material, fine linearity (R2> 0.99) in the investigated concentration range was obtained for all pesticides. The accuracy of the procedure can be expressed as recoveries of the investigated residues. In Table S3, mean recoveries obtained by 6 determinations of spiked apple samples with a concentration level of 30 µg/kg were presented. As shown, recoveries of all pesticides examined ranged between the acceptable values of 70% and 120% (Regulation (EC) No. 396/2005). Precision, expressed as relative standard deviation (% RSD) was calculated by 6 analyses of a sample loaded with a concentration level of 30 µg/kg of each pesticide. The between-day precision of the technique was also tested by performing five determinations daily for a period of five consecutive days at concentration levels of 30 µg/kg. The obtained RSD values were between 3.6 and 7.8%, fulfilling the criterion of the EU (RSD<20%).
2. Occurrence and levels of target pesticides in Greek PDO/PGI applesAfter the successful validation of the QuEChERS approach, the method was applied to the 80 PDO/PGI apple samples for analysis of the 51 target pesticides. The overall results obtained are presented in Table 2. As shown, 12 pesticide residues were detected of the 51 species investigated. Only 5 of the 80 samples (6.3%) were found to contain pesticide residues in undetectable limits. In the 75 positive samples, the number of different pesticide residues ranged from 3 to 10 (10 different pesticides were detected only in one sample) with an average number of 6.0 different pesticides per apple sample. Two of the 80 investigated samples contained pesticide residues (parathion-methyl) in levels exceeding MRLs. The presence of pesticide residues in PDO/PGI apples is not unexpected since this labeling does not necessarily mean that such products are free of contaminants. Indeed, PDO/PGI labeling guarantees only the authenticity of products and their agricultural practice does not avoid the use of pesticides, as does organic farming practice. As mentioned above, pesticide residues, mainly penconazole, α-endosulfan, β-endosulfan and flufenoxuron, were also observed in PDO/PGI olive oils in our previous investigation.4) These findings confirm the need for monitoring of pesticide residues in PDO/PGI agricultural products.
In this study, endosulfan sulfate, azinphos-methyl and parathion-methyl possessed the lowest detection rate with 2, 5 and 9 positive samples, respectively. The highest detection rates in the investigated apple samples were observed for chlorpyrifos (n=75), quinalphos (n=75) and parathion (n=73). It should be noted that the presence of parathion and parathion-methyl in Greek apple samples was also reported in a previous study,13) while fenthion, α-endosulfan and β-endosulfan have been detected in Greek agricultural products,24,25) even in much higher concentrations. For instance, fenthion was detected in 74% of olive oil samples in concentrations ranging between 4.6 and 767 µg/kg.25) An interesting point of this investigation is the existence of cis-permethrin, penconazole, flufenoxuron and quinalphos. To the best of our knowledge, this is the first time that these pesticide residues have been detected in Greek apples, as they were not included in previous studies in Greece. However, these residues have been detected in apple samples from other countries.26,27) Flufenoxuron and penconazole were also detected in PDO/PGI olive oils in our previous investigation.4) The existence of flufenoxuron in both apples and olive oil is of interest because it was banned in the European Union in 2011 due to its high potential for bioaccumulation. Therefore, the present study underlines the need for flufenoxuron monitoring, in accordance with EU regulations.
The values of penconazole, flufenoxuron, α-endosulfan, β-endosulfan, chlorpyrifos, parathion, quinalphos and permethrin in the 75 positive samples were subjected to Spearman correlation analysis. This kind of statistical treatment used in relative studies,25) is a non-parametric measure of statistical dependence between two variables that assesses their relationship. In cases of undetectable analytes in a sample, a rough value of LOD/2 was considered, while for detectable analytes with values below LOQ, an average value of LOD and LOQ was considered. In Table 3, the Spearman correlation matrix for flufenoxuron, penconazole, α-endosulfan, β-endosulfan, chlorpyrifos, parathion, quinalphos and cis-permethrin is presented. As shown, the most powerful correlations (r> 0.500), significant at the 0.01 level, were obtained between α-endosulfan and β-endosulfan (r=0.823) and between flufenoxuron and penconazole (r=0.683) in agreement with our previous investigation of olive oil samples,4) as well as an analogous study of Greek olive oils.25) Powerful correlations, significant at the 0.01 level, were also obtained between parathion and cis-permethrin (r=0.463), α- endosulfan and penconazole (r=0.376), α- endosulfan and flufenoxuron (r=0.352) and between quinalphos and chlorpyrifos (r=0.319). These correlations imply that the pesticides are often used in combination. However, to the best of our knowledge, in Greece there is no commercially available combination of two or more of the 8 pesticides with the highest detection rates (flufenoxuron, penconazole, α-endosulfan, β-endosulfan, chlorpyrifos, parathion, quinalphos and cis-permethrin) in the investigated apple samples. Therefore, this finding can be explained on the basis of intercropping (apple trees with olive or orange trees, vineyards, etc.). Another explanation for such correlations is the belief in optimum effect in controlling diseases and pests when these pesticides are applied in combination to apple trees.
3. Comparison of occurrence of target pesticides between PDO/PGI apples with different geographical originationsA comparison of the occurrence and levels of the 12 pesticide residues detected in PDO/PGI apples with different geographical originations is shown in Table 4. No considerable differences among datasets coming from the 3 investigated regions were observed. The two apple samples that exceeded the MRL value for parathion-methyl are originated from Kastoria (one sample) and Tripoli (one sample). Apple samples from Pelion exhibited the lowest mean number of different pesticides detected per sample (5.5). However, they possessed the highest levels of some pesticides, such as parathion, chlorpyrifos and quinalphos. Apple samples from Kastoria showed the minimum detection rate (15% of samples with no detectable residues). The average number of different pesticide residues detected per positive sample (5.9) was comparable to those originating from Pelion. Apple samples originating from Tripoli (Peloponnese) gave the highest average number of pesticides per positive sample (7.2). Their profile is characterized by lower levels of pesticides, such as parathion, chlorpyrifos, quinalphos, and permethrin, but higher levels of penconazole, flufenoxuron, α-endosulfan and β-endosulfan were found. It should be noted that the last 4 pesticides were detected in higher levels in olive oil samples coming from two neighboring regional units of Peloponnese, Lakonia and Messinia. This fact may imply geographical origination’s role in the pesticide levels of agricultural products.
4. Principal component analysisIn an effort to investigate the parameters affecting the incidence and levels of pesticide residues in Greek apples and olive oil (investigated in our previous work4)) as well as to unravel similarities/dissimilarities in the underlying distribution mechanism, principal component analysis was employed. PCA is a way of identifying patterns in data by using an orthogonal transformation to convert a set of observations of possibly correlated variables into a set of linearly uncorrelated variables, called principle components. As the number of principal components is usually less than the number of original variables, PCA unravels data similarity by reducing the number of dimensions without loss of information.28) PCA analysis was conducted by using the mean values found in the 11 PDO/PGI agricultural products for the 10 pesticides: penconazole, flufenoxuron, α-endosulfan, β-endosulfan and fenthion (highest detection rates in olive oil samples) as well as parathion, parathion-methyl, quinalphos, chlorpyrifos and cis-permethrin (highest detection rates in apple samples). These mean values come from 10 samples analyzed in the case of each PDO/PGI olive oil and from 20 samples in the case of each PDO/PGI apple; they are presented in Table S4 in the supplementary material. As stated above, in cases of undetectable analytes in a sample, a rough value of LOD/2 was considered, while in cases of a positive signal below the LOQ, an average value of the LOD and LOQ was taken into account.
Three principal components were extracted with a cumulative correlation coefficient of R2=0.873. The score plot of the two principal components served to provide an overview of the behavior of the 10 pesticides. As shown in Fig. S1, in the supplementary material, the pesticides are classified in three pairs with similar behavior in the investigated agricultural products: the two pesticide pairs, penconazole-α-endosulfan and fenthion-β-endosulfan, are very closely located, implying similarities in their existence in agricultural products. The third pair of pesticides, quinalphos-chlorpyrifos, is located in the lower right quarter. The loadings on the principal components were then considered in order to unravel relationships between the variables (pesticide residue concentrations). As shown in the histogram depicted in Fig. S2(a), in the supplementary material, the loading column plot of the first principal component, accounting for 40.7% of the variance, did not show significant variation among the different PDO/PGI products. This fact implies the general use of the above-mentioned pesticides for agricultural purposes. On the contrary, the loading column plot of the second principal component (Fig. S2(b) in the supplementary material), accounting for 35.9% of the variance, revealed differences according to the kind of product (olive oil or apple). Thus, data sets measured in olive oil and in apples possess positive and negative p-values, respectively. The loading column plot of the third principal component (Fig. S2(c) in the supplementary material), which accounts for 10.7% of the variance, showed that we can observe differences according to other parameters, mainly attributed to the geographical origination of the products. For example, the two PDO apples from Pelion exhibited slight negative p-values. However, as olive oil originating from the 4 regional units of Crete showed differences, this principal component is likely to be driven by parameters separate from the geographical origination of the agricultural products.
Twelve pesticide residues were detected in 80 PDO/PGI Greek apple samples. Five of the 80 samples (6.3%) were pesticide free; the average number of different pesticides detected in the 75 positive samples was 6.0. The highest detection rates were observed for chlorpyrifos (n=75), quinalphos (n=75) and parathion (n=73). Only 2 of the 80 investigated samples contained pesticide residues (parathion-methyl) in levels exceeding MRLs. According to the Spearman correlation analysis, the most powerful correlations were obtained between α-endosulfan and β-endosulfan (r=0.823) and between flufenoxuron and penconazole (r=0.683). These correlations imply that the pesticides may be used in intercropping or they may be applied in combination for optimum effect in controlling diseases and pests of apple trees. Principal component analysis revealed similarities in pesticides levels in PDO/PGI samples, while the type of product also impacts its residue content.
| Location | Region | PDO/PGI | Number of samples |
|---|---|---|---|
| Pelion | Thessalia | PDO | 20 |
| Pelion | Thessalia | PDO | 20 |
| Tripoli | Peloponnese | PDO | 20 |
| Kastoria | West Macedonia | PGI | 20 |
| Total samples under investigation | 80 | ||
| Pesticide | Number of Positive samples a | Mean value (µg/kg) | Concentration range (µg/kg) | MRL (µg/kg) | Samples exceed MRL No (%) |
|---|---|---|---|---|---|
| Azinphos-methyl | (5) | < 9.0 | < 9.0 | 50 | 0 (0%) |
| Chlorpyrifos | 62 (13) | 24.6 | 9.4–45.1 | 500 | 0 (0%) |
| α-Endosulfan | 16 (15) | 4.5 | 3.9–5.0 | 50b | 0 (0%) |
| β-Endosulfan | 12 (15) | 4.5 | 3.9–5.6 | 50b | 0 (0%) |
| Endosulfan sulfate | (2) | < 8.7 | < 8.7 | 50b | 0 (0%) |
| Fenthion | 2 (10) | 5.1 | 4.9–5.2 | 10 | 0 (0%) |
| Flufenoxuron | 11 (16) | 6.7 | 6.0–8.1 | 500 | 0 (0%) |
| Parathion | 60 (13) | 24.5 | 10.1–41.8 | 50 | 0 (0%) |
| Parathion-methyl | 3 (6) | 11.0 | 9.6–11.9 | 10c | 2 (2.5%) |
| Penconazole | 28 (16) | 4.1 | 3.0–6.7 | 200 | 0 (0%) |
| Permethrin cis | 45 (25) | 13.9 | 4.1–35.4 | 50d | 0 (0%) |
| Quinalphos | 60 (15) | 21.0 | 8.5–43.4 | 50 | 0 (0%) |
a The number denotes the samples with residues above the limit of quantification. In parenthesis, the number of further samples in which the pesticide exists in levels below the limit of quantification is indicated. b Sum of α- and β-endosulfan and endosulfan sulfate. c Sum of parathion-methyl and paraoxon-methyl expressed as parathion-methyl. d Sum of cis-and trans-isomers.
| Flufenoxuron | Penconazole | α-Endosulfan | β-endosulfan | Chlorpyrifos | Parathion | Quinalphos | Permethirn-cis | |
|---|---|---|---|---|---|---|---|---|
| Flufenoxuron | 1.000 | |||||||
| Penconazole | 0.683** (0.000) | 1.000 | ||||||
| α-Endosulfan | 0.352** (0.001) | 0.376** (0.001) | 1.000 | |||||
| β-endosulfan | 0.274* (0.014) | 0.284* (0.011) | 0.823** (0.000) | 1.000 | ||||
| Chlorpyrifos | 0.016 (0.891) | −0.052 (0.650) | 0.060 (0.598) | 0.062 (0.585) | 1.000 | |||
| Parathion | 0.048 (0.671) | 0.154 (0.173) | −0.085 (0.453) | −0.041 (0.719) | 0.129 (0.253) | 1.000 | ||
| Quinalphos | −0.010 (0.927) | −0.087 (0.444) | −0.133 (0.239) | −0.053 (0.643) | 0.319** (0.004) | 0.232* (0.038) | 1.000 | |
| Permethrin cis | −0.146 (0.196) | −0.081 (0.477) | −0.082 (0.469) | −0.023 (0.837) | 0.250* (0.025) | 0.463** (0.000) | 0.262* (0.019) | 1.000 |
*Correlation is significant at the 0.05 level (2-tailed). **Correlation is significant at the 0.01 level (2-tailed).
| Pesticide | Pelion (n=40) | Kastoria (n=20) | Tripoli (n=20) | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Positive samples a | Concentration range (µg/kg) | Samples exceed MRL No (%) | Positive samples a | Concentration range (µg/kg) | Samples exceed MRL No (%) | Positive samples a | Concentration range (µg/kg) | Samples exceed MRL No (%) | |
| Azinphos-methyl | (3) | < 9.0 | 0 (0%) | (1) | < 9.0 | 0 (0%) | (1) | < 9.0 | 0 (0%) |
| Chlorpyrifos | 35 (5) | 9.8–41.1 | 0 (0%) | 14 (3) | 9.4–45.1 | 0 (0%) | 13 (5) | 10.1–34.6 | 0 (0%) |
| α-Endosulfan | 4 (6) | 3.9–5.0 | 0 (0%) | 3 (6) | 4.0–4.4 | 0 (0%) | 9 (3) | 4.0–5.0 | 0 (0%) |
| β-Endosulfan | 4 (7) | 4.0–5.2 | 0 (0%) | 2 (3) | 3.9–4.5 | 0 (0%) | 6 (5) | 4.0–5.6 | 0 (0%) |
| Endosulfan sulfate | — | — | 0 (0%) | — | — | 0 (0%) | (2) | < 8.7 | 0 (0%) |
| Fenthion | 1 (4) | 4.9 | 0 (0%) | (3) | < 4.7 | 0 (0%) | 2 (3) | 5.0–5.2 | 0 (0%) |
| Flufenoxuron | 3 (7) | 6.0–6.1 | 0 (0%) | 1 (3) | 6.2 | 0 (0%) | 7 (6) | 6.3–8.1 | 0 (0%) |
| Parathion | 32 (7) | 10.5–41.8 | 0 (0%) | 14 (3) | 11.8–35.1 | 0 (0%) | 14 (3) | 10.1–23.1 | 0 (0%) |
| Parathion-methyl | 1 (3) | 9.6 | 0 (0%) | 1 (1) | 11.9 | 1 (5%) | 1 (2) | 11.5 | 1 (5%) |
| Penconazole | 11 (8) | 3.2–4.1 | 0 (0%) | 5 (4) | 3.0–3.3 | 0 (0%) | 12 (4) | 3.8–6.7 | 0 (0%) |
| Permethrin cis | 27 (12) | 6.8–35.4 | 0 (0%) | 12 (5) | 5.1–31.6 | 0 (0%) | 6 (8) | 4.5–12.6 | 0 (0%) |
| Quinalphos | 33 (7) | 8.5–43.4 | 0 (0%) | 14 (3) | 8.5–27.2 | 0 (0%) | 13 (5) | 9.5–29.2 | 0 (0%) |
| Samples with undetectable pesticide residues | 0 | 3 | 2 | ||||||
| Average pesticides per olive oil sampleb | 5.5 | 5.9 | 7.2 | ||||||
a The number denotes the samples with residues above the limit of quantification. In parenthesis, the number of further samples in which the pesticide exists in levels below the limit of quantification is indicated. b For pesticides’ positive samples.
