Modified microwave-assisted extraction of ergosterol for measuring fungal biomass in grain cultures

Ergosterol is a measure for fungal biomass. The recovery rates using a previously described microwave-assisted-extraction (MAE) method for ergosterol analysis tended to be low for grain cultures (pure culture in sterilized 40% moisture content grain) inoculated with Fusarium graminearum . An improved MAE method for measuring ergosterol in grain cultures was developed and compared. Modification to the original MAE included alterations in duration of microwave exposure and extraction solvents. Four autoclaved grains (wheat, rice, barley, and corn) were inoculated with F. graminearum or spiked with ergosterol at concentrations from 0.88 to 100 microg/g and extracted with both methods. The ergosterol recovery rates were significantly different (p < 0.05) for the two methods in assaying both the spiked and grain culture samples. The modified method provided greater recovery rates than the previously reported MAE method for the spiked samples and F. graminearum grain cultures.     

Parameters affecting extraction of selected fungicides from vineyard soils

This paper describes a sensitive method for the simultaneous quantification of eight commonly used grapevine fungicides in vineyard soils: cyprodinil, fludioxonil, metalaxyl, penconazole, pyrimethanil, procymidone, tebuconazole, and vinclozolin. The fungicides are extracted from the soil sample by sonication with water followed by shaking with ethyl acetate and are quantified by gas chromatography with mass spectrometry. Average extraction efficiencies in a sample of seven spiked, previously fungicide-free soils were > or =79% for all of the analytes, method precisions were > or =17%, and quantification limits were < or =50 microg/kg. However, because recoveries varied considerably from soil to soil, there is a need to control for soil matrix differences (mainly soil pH and exchangeable calcium content); as a consequence, soil fungicide contents must be quantified by the standard additions method. When the method was applied in this way to soil samples from vineyards belonging to the specified wine-growing region of Rias Baixas (Galicia, northwestern Spain) taken at the beginning of October (1 month after the crop's final treatment), levels of fludioxonil as high as 991 microg/kg were found, but at the start of the season (9 months after the previous crop's final treatment) only fludioxonil was detected at levels higher than its limit of quantification (45 and 52 microg/kg).     

Analysis of chlorothalonil and three degradates in sediment and soil

A method has been developed for the simultaneous extraction of chlorothalonil and three of its degradates (4-hydroxy-2,5,6-trichloroisophthalonitrile, 1-carbamoyl-3-cyano-4-hydroxy-2,5,6-trichlorobenzene, and 1,3-dicarbamoyl-2,4,5,6-tetrachlorobenzene) from soils and sediments; the compounds were extracted using sonication with acetone and isolation of the parent compound and matrix interferences from the degradates by solid phase extraction (SPE). The chlorothalonil fraction underwent further coextracted matrix interference removal with Florisil. The degradates were derivatized with N, O-bis(trimethylsilyl) trifluoroacetamide (BSTFA) and chlorotrimethylsilane (TMCS). All compounds were analyzed by gas chromatography-mass spectrometry (GC-MS). Recoveries on a spiked (20 and 200 microg kg (-1)) sediment ranged from 80% to 91% with calculated limits of detection of 1-5 microg kg (-1) dry weight sediment. An additional 20 sediment samples were collected in watersheds from the Southeastern United States where chlorothalonil is used widely on peanuts and other crops. None of the target compounds were detected. Laboratory fortified recoveries of chlorothalonil and its degradates in these environmental sediment samples ranged from 75% to 89%.     

Simple determination of 40 organophosphate pesticides in raw wool using microwave-assisted extraction and GC-FPD analysis

A validated analytical method for the multiresidue analysis of 40 organophosphate pesticides (OPs) and conversion products in raw wool has been developed. The method is based on the selective microwave-assisted extraction (MAE) of raw wool with acetonitrile and analysis of extracts by gas chromatography-flame photometric detector. The optimum MAE conditions were 20 min duration at 80 °C with 30 mL of acetonitrile per gram of wool. A validation study was performed according to the European SANCO guidelines 10684/2009. Limits of detection and quantification for all pesticides tested were from 0.01 to 0.2 mg/kg and from 0.2 to 1.0 mg/kg, respectively. The average recoveries of pesticides spiked at different levels were in the range of 70-120% with relative standard deviations of ≤ 20%. The extraction performance was compared to the one obtained with a reference Soxhlet extraction. The method was also applied in the analysis of real wool (after field application) samples.     

Determination of bentazone, dichlorprop, and MCPA in different soils by sodium hydroxide extraction in combination with solid-phase preconcentration

A method for the extraction of bentazone, dichlorprop, and MCPA in three selected Norwegian soils of different textures is described. Initially three different extraction methods were tested on one soil type. All methods gave recoveries >80% for the pesticide mixture, but extraction with sodium hydroxide in combination with solid-phase preconcentration was used for further recovery tests with soils of different properties spiked at four herbicide concentration levels (0.001-10 microg/g of wet soil). The method was rapid and easy and required a minimum of organic solvents. The recoveries were in the range of 82-109, 80-123, and 45-91% for the soils containing 1.4 (Hole), 2.5 (Kroer), and 37.8% (Froland) organic carbon, respectively. Limits of quantification using GC-MS were 0.0003 microg/g of wet soil for bentazone and 0.0001 microg/g of wet soil for both dichlorprop and MCPA.     

Improved HPLC-MS/MS method for determination of isoxaflutole (balance) and its metabolites in soils and forage plants

A robust multi-residue procedure is needed for the analysis of the pro-herbicide isoxaflutole and its degradates in soil and plant materials at environmentally relevant (<1 microg kg-1) levels. An analytical method using turbo-spray and heat-nebulizer high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) was developed for the analysis of isoxaflutole (IXF) and its two metabolites, diketonitrile (DKN) and the benzoic acid metabolite (BA), at sub-microgram per kilogram levels in soil and plant samples. The average recoveries of the three compounds in spiked soil and plant samples ranged from 84 to 110% and 94 to 105%, respectively. The limits of quantification were validated at 0.06 microg kg-1 for soil and 0.3 microg kg-1 for plant samples. The limits of detection (LOD) for soil analysis were 0.01, 0.002, and 0.01 microg kg-1 for IXF, DKN, and BA, respectively. Corresponding LOD for the plant analysis method were 0.05, 0.01, and 0.05 microg kg-1. The developed method was validated using forage grass and soil samples collected from a field lysimeter study in which IXF was applied to each of four forage treatments. Forage plants and soils were sampled for analyses 25 days after IXF application to the soil. In soils, IXF was not detected in any treatment, and DKN was the predominant metabolite found. In forage plants, the concentrations of DKN and BA were 10-100-fold higher than that in soil samples, but IXF was not detected in any forage plants. The much higher proportion of BA to DKN in plant tissues (23-53%), as compared to soils (0-5%), suggested that these forages were capable of detoxifying DKN. The developed methods provided LODs at sub-microgram per kilogram levels to determine the fate of IXF and its metabolites in soils and forage plants, and they also represent considerable improvements in extraction recovery rates and detection sensitivity as compared to previous analytical methods for these compounds.     

Rapid time-resolved fluoroimmunoassay for the screening of narasin and salinomycin residues in poultry and eggs

Anticoccidial drugs are extensively used in the poultry industry to control the infection of the single-cell protozoa of the genus Eimeria. The most commonly used coccidiostats in poultry are the polyether ionophores such as narasin and salinomycin. This paper presents a rapid and simple method for the screening of residues of these two coccidiostatic compounds in poultry and eggs. The method is based on time-resolved fluoroimmunoassay. Sample preparation of eggs consists only of one extraction and evaporation step, and a solid phase extraction step is needed only for the muscle sample preparation. Mean recoveries were 91.0% from muscle tissue and 81.1% from eggs for both narasin and salinomycin. The performance of the assay was evaluated only for narasin because salinomycin had a cross-reactivity of 100% in the assay, and the recoveries of the compounds were not significantly different (P >0.05). The limits of detection [mean + 3 x standard deviation (SD)] of narasin were 0.56 and 0.28 microg/kg, and the limits of quantification (mean + 9 x SD) were 1.80 and 0.57 microg/kg for muscle and eggs, respectively. The coefficients of variation (CV) of the interassay precision of the method, evaluated by five replicate analyses of muscle samples spiked with 2 microg/kg of narasin and egg samples spiked with 1 microg/kg of narasin, were 4.1 and 6.4%, respectively. The CVs of intra-assay precision tests, determined by 10 replicate analyses at the above-mentioned concentration levels, were 3.8 and 4.5%, respectively.     

Gas chromatographic determination of N-nitrosamines, aromatic amines, and melamine in milk and dairy products using an automatic solid-phase extraction system

A reliable analytical method was presented for the simultaneous determination of six N-nitrosamines, nine aromatic amines, and melamine in milk and dairy products using gas chromatography coupled with mass spectrometry. The sample treatment includes the precipitation of proteins with acetonitrile, centrifugation, solvent changeover by evaporation, and continuous solid-phase extraction for cleanup and preconcentration purposes. Samples (5 g) containing 0.15-500 ng of each amine were analyzed, and low detection limits (15-130 ng/kg) were achieved. Recoveries for milk and dairy products samples spiked with 1, 10, and 50 μg/kg ranged from 92% to 101%, with intraday and interday relative standard deviation values below 7.5%. The method was successfully applied to determine amine residues in several milk types (human breast, cow, and goat) and dairy products.     

Application of gas chromatography coupled to triple quadrupole mass spectrometry for the multiresidue analysis of pesticides in olive oil

A new multiresidue method has been developed and validated for the simultaneous determination of 100 pesticide residues in olive oil. The determination of pesticide residues was carried out in only 19 min by gas chromatography coupled to tandem mass spectrometry using a triple quadrupole mass analyzer. The mass spectrometer was operated in electron ionization and the selection reaction monitoring mode was used, acquiring two or three fragmentation reactions per compound. Two extraction processes were studied, and an evaluation of the stability and sensitivity of the chromatographic system has been performed for the tested extraction procedures. The final proposed methodology was based on a liquid-liquid partition with an n-hexane/acetonitrile mixture followed by a gel permeation chromatography cleanup step. An adequate lineal relation was obtained in the studied concentration range (10-200 microg kg (-1)); the recovery values were in the range 70-110% for the two levels of concentration studied: 12 and 50 microg kg (-1). Precision values, expressed as relative standard deviation, were lower than 18% at the aforementioned spiking levels; detection limits, confirmation limits, and quantitation limits were below or equal to 1.9, 2.6, and 3.6 microg kg (-1), respectively. The developed methodology was applied to the analysis of pesticide residues in real samples of olive oil from the south of Spain.     

Occurrence of ochratoxin A in cocoa products and chocolate

In this work, the occurrence of ochratoxin A (OTA) in 170 samples of cocoa products of different geographical origins was studied. An immunoaffinity column with HPLC separation was developed to quantify low levels of OTA in cocoa bean, cocoa cake, cocoa mass, cocoa nib, cocoa powder, cocoa shell, cocoa butter, chocolate, and chocolate cream with >80% recoveries. The method was validated by performing replicate analyses of uncontaminated cocoa material spiked at three different levels of OTA (1, 2, and 5 microg/kg). The data obtained were related on the acceptable safe daily exposure for OTA. The highest levels of OTA were detected in roasted cocoa shell and cocoa cake (0.1-23.1 microg/kg) and only at minor levels in the other cocoa products. Twenty-six cocoa and chocolate samples were free from detectable OTA (<0.10 microg/kg). In roasted cocoa powder 38.7% of the samples analyzed contained OTA at levels ranging from 0.1 to 2 microg/kg, and 54.8% was contaminated at >2 microg/kg (and 12 samples at >3 microg/kg). Ochratoxin A was detected in cocoa bean at levels from 0.1 to 3.5 microg/kg, the mean concentration being 0.45 microg/kg; only one sample exceeded 2 microg/kg (4.7%). In contrast, 51.2% of cocoa cake samples contained OTA at levels > or =2 microg/kg, among which 16 exceeded 5 microg/kg (range of 5-9 microg/kg). These results indicate that roasted cocoa powder is not a major source of OTA in the diet.     

Gas chromatographic determination of N-nitrosamines in beverages following automatic solid-phase extraction

A semiautomatic method for the determination of seven N-nitrosamines in beverages by gas chromatography with nitrogen-phosphorus detection is proposed. Beverage samples are aspirated into a solid-phase extraction module for preconcentration and cleanup. The influence of the experimental conditions was examined by using various sorbents among which LiChrolut EN was found to provide quantitative elution and the highest preconcentration factors of all. The proposed method is sensitive, with limits of detection between 7 and 33 ng/kg, and precise, with relative standard deviations from 4.3% to 6.0%. The recoveries of N-nitrosamines from beverage samples spiked with 0.5 or 1 microg/kg concentrations of these compounds ranged from 95% to 102%. The method was successfully applied to the determination of residues of the studied N-nitrosamines in beverages including beer, wine, liquor, whisky, cognac, rum, vodka, grape juice, cider, tonic water, and soft drinks. The analytes were only detected in beer samples, positives being confirmed by gas chromatography coupled with impact ionization mass spectrometry.     

Simultaneous determination of ethidimuron, methabenzthiazuron, and their two major degradation products in soil

An analytical method has been developed for the quantification of two herbicides (ethidimuron and methabenzthiazuron) and their two main soil derivatives. This method involves fluidized-bed extraction (FBE) prior to cleanup and analysis by reverse-phase liquid chromatography with UV detection at 282 nm. FBE conditions were established to provide efficient extraction without degradation of the four analytes. (14)C-labeled compounds were used for the optimization of extraction and purification steps and for the determination of related efficiencies. Extraction was optimal using a fexIKA extractor operating at 110 degrees C for three cycles (total time = 95 min) with 75 g of soil and 150 mL of a 60:40 v/v acetone/water mixture. Extracts were further purified on a 500 mg silica SPE cartridge. Separation was performed on a C18 Purosphere column (250 mm x 4 mm i.d.), at 0.8 mL min(-1) and 30 degrees C with an elution gradient made up of phosphoric acid aqueous solution (pH 2.2) and acetonitrile. Calibration curves were found to be linear in the 0.5-50 mg L(-1) concentration range. Besides freshly spiked soil samples, method validation included the analysis of samples with aged residues. Recovery values, determined from spiked samples, were close to 100%. Limits of detection ranged between 2 and 3 microg kg(-1) of dry soil and limits of quantification between 8 and 10 microg kg(-1) of dry soil. An attempt to improve these performances by using fluorescence detection following postcolumn derivatization by orthophthalaldehyde-mercaptoethanol reagent was unsuccessful.     

Development of a solid-phase extraction method for phenoxy acids and bentazone in water and comparison to a liquid-liquid extraction method

A rapid solid-phase extraction (SPE) method was developed for the determination of bentazone and the phenoxy acids 2,4-D, dichlorprop, MCPA, and mecoprop in Norwegian environmental water samples. Cartridges with a high-capacity cross-linked polystyrene-based polymer were used for off-line preconcentration. The effects of elution solvent, elution volume, sample volume, sorbent mass, pH, and flow rate on the recoveries of the pesticides were investigated using HPLC. Average recovery of >90% was achieved with 500 mg sorbents using 2 mL of methanol with 5% NH3 as elution solvent. The recoveries were independent of sample pH in the tested range of pH 1-7. Using a sample volume of 200 mL, the limits of determination for the phenoxy acids and bentazone are 0.02 microg/L. Sample volumes up to 2000 mL at a flow rate of 60 mL/min could be handled without any loss of analytes, which makes it possible to lower the limits of determination. The SPE method was compared to a routinely used liquid-liquid extraction method. Three different water matrices spiked at 1.0 and 0.05 microg/L were extracted, and the quantification was performed by GC-MS. Both methods permitted the determination of phenoxy acids and bentazone in distilled water, creek water, and well water down to a level of 0.05 microg/L with recoveries >80% for 200 mL samples. Important advantages of the SPE method compared to the liquid-liquid extraction method were the short extraction times, lack of emulsions, use of disposable equipment, and reduced consumption of organic solvents.     

Screening of tetracycline residues in fish muscles by CCD camera-based solid-surface fluorescence

A methodology for the screening of tetracyclines (TCs), including tetracycline (TC), oxytetracycline (OTC), and chlorotetracycline (CTC), in different fish muscle matrices has been proposed. This method was based on in situ fluorescent derivation of TCs, transferring weakly fluorescing TCs to highly fluorescent species, on alkaline-activated solid silica gel G plates (SGGPs). By coupling solid-surface fluorescence (SSF) with charge-coupled device (CCD) camera imaging, a CCD camera-based SSF (CCD-SSF) methodology has been developed. Calibration curve, repeatability, selectivity, limit of detection (LOD), and limit of quantification (LOQ) have been explored for evaluating the performance of the method itself. Linear calibration curves were obtained over a range of 0.20-1.0 ng/spot for all three TCs. The LODs, defined as 3sigma, for TC, OTC, and CTC were 0.14, 0.15, and 0.16 ng/spot, respectively. The trueness of method was validated by HPLC, and no significant difference between CCD-SSF and HPLC was found, on a basis of 95% confidence level. By spiked recovery studies, a linear calibration curve ranging from 20 to 300 microg/kg of TC in fish muscle samples with a correlation coefficient (R 2) equal to 0.994 was obtained. The total average recovery for TC in fish muscle samples from six different fish matrices, fortified with TC at 50, 100, and 200 microg/kg levels, was 75.7% with average relative standard deviations (RSDs) ranging from 2.0 to 7.7%. RSDs ranged from 2.5 to 5.8% and from 5.2 to 7.6% for in-day and interday repeatability, respectively. The detection and quantification limits in fish muscle matrices were 16 and 53 microg/kg of TCs, respectively. The newly developed CCD-SSF method has been applied to the screening of the TC residues in fish muscle samples. The method has been demonstrated to bear some advantages, such as its simplicity, high throughput, low cost, use of fewer pollutants, and reasonable sensitivity.     

Determination of flubendazole and its metabolites in eggs and poultry muscle with liquid chromatography-tandem mass spectrometry

The optimization of a quantitative and sensitive LC-MS/MS method to determine flubendazole and its hydrolyzed and reduced metabolites in eggs and poultry muscle is described. The benzimidazole components were extracted from the two matrices with ethyl acetate after the sample mixtures had been made alkaline. The HPLC separation was performed on an RP C-18 column with gradient elution, using ammonium acetate and acetonitrile as mobile phase. The analytes were detected after atmospheric pressure electrospray ionization on a tandem quadrupole mass spectrometer in MS/MS mode. The components were measured by the MS/MS transition of the molecular ion to the most abundant daughter ion. The overall extraction recovery values for flubendazole, the hydrolyzed metabolite, and the reduced metabolite in eggs (fortification levels of 200, 400, and 800 microg kg(-1)) and muscle (fortification levels of 25, 50, and 100 microg kg(-1)) were, respectively, 77, 78, and 80% and 92, 95, and 90%. The trueness (fortification levels of 400 and 50 microg kg(-1), respectively, for eggs and muscle), expressed as a percentage of the added values for these analytes, was, respectively, 89, 100, and 86 and 110, 110, and 98%. The proposed MS detection method operating in the MS/MS mode is very selective and very sensitive. The limits of detection for flubendazole and its hydrolyzed and reduced metabolites in egg and muscle were, respectively, 0.19, 0.29, and 1.14 microg kg(-1) and 0.14, 0.75, and 0.31 microg kg(-1). The limits of quantification were, respectively, 1, 1, and 2 microg kg(-1) and 1, 1, and 1 microg kg(-1). The discussed method was applied to a pharmacokinetic study with turkeys. Residue concentrations in breast and thigh muscle of turkeys orally treated with flubendazole were quantified. Medicated feed containing 19.9 and 29.6 mg kg(-1) flubendazole was provided to the turkeys for seven consecutive days. For the trial with the recommended dose of 19.9 mg kg(-1), one day after the end of the treatment, the mean sum of the flubendazole plus hydrolyzed metabolite residue values in thigh and breast muscle declined to below the maximum residue limit (50 microg kg(-1)) and were, respectively, 36.6 and 54.1 microg kg(-1). The corresponding values with the higher dose of 29.6 mg kg(-1) were, respectively, 101.7 and 119.7 microg kg(-1).     

Determination of low levels of perchlorate in lettuce and spinach using ion chromatography-electrospray ionization mass spectrometry (IC-ESI-MS)

A sample preparation method was developed to quantify environmentally relevant (low micrograms per liter) concentrations of perchlorate (ClO4(-)) in leafy vegetables using IC-ESI-MS. Lettuce and spinach were macerated, centrifuged, and filtered, and the aqueous extracts were rendered water-clear using a one-step solid-phase extraction method. Total time for extraction and sample preparation was 6 h. Ion suppression was demonstrated and was likely due to unknown organics still present in the extract solution after cleanup. However, this interference was readily eliminated using a Cl(18)O4(-) internal standard at 1 microg/L in all standards and samples. Hydroponically grown perchlorate-free butterhead lettuce was spiked to either 10.3 or 37.7 microg/kg of fresh weight (FW), and recoveries were between 91 and 98% and between 93 and 101%, respectively. Five types of lettuce and spinach from a local grocery store were then analyzed; they contained from 0.6 to 6.4 microg/kg of FW. Spike recoveries using the store-bought samples ranged from 89 to 100%. The method detection limit for perchlorate in plant extracts is 40 ng/L, and the corresponding minimum reporting limit is 200 ng/L or 0.8 microg/kg of FW.     

Simultaneous gas chromatographic determination of carbofuran, metalaxyl, and simazine in soils

A method for extraction, cleanup, and simultaneous gas chromatographic detection of carbofuran, metalaxyl, and simazine in soils has been developed. Pesticide residues were extracted from soil with acetone containing 10% 0.2M HCl-KCl buffer (pH 2.0), cleaned up with methylene chloride-carbonate buffer (pH 10.7) solvent partitioning and solid-phase extraction on disposable silica gel columns, and quantitated with gas chromatography using a Supelcowax 10 megabore capillary column and a nitrogen-selective detector. Method limits of detection were 0.02 microgram/g for the 3 pesticides in surface soils (0-30 cm depths) and 0.02, 0.02, and 0.005 microgram/g in soils below 30 cm (subsoils) for carbofuran, metalaxyl, and simazine, respectively. Recoveries for carbofuran, metalaxyl, and simazine were 92.6 +/- 5.5, 93.6 +/- 5.0, and 88.4 +/- 6.7%, respectively, when soil samples were spiked with pesticide concentrations ranging from 0.02 to 2.0 micrograms/g.     

Capillary gas chromatography with atomic emission detection for pesticide analysis in soil samples

A method for the simultaneous determination of 10 pesticides (organochlorines, organophosphorus compounds, and pyrethrins) in soils using capillary gas chromatography with atomic emission detection (GC-AED) is reported. Soil samples are first "cleaned-up" with 25 mL of an ascorbic acid solution (pH 2.15). The aqueous phase is extracted with ethyl acetate, and the solid residue is then extracted twice with 10 mL of ethyl acetate. The three resultant organic extracts are combined, concentrated to dryness, and reconstituted in 1 mL of acetone. The pesticides are selectively detected by monitoring chlorine and bromine in the first run and sulfur emission line wavelength in the second run. Each chromatographic run takes 19 min. Detection limits are between 25 and 75 pg, depending on the compound, which corresponds to 1.7 and 5.0 ng/g in the soil samples, respectively. Recoveries of the pesticides from spiked preparations result in an overall mean recovery of 95.3% (n = 120) at fortification levels ranging from 10 to 60 ng/g, depending on the compound. The method is reliable and can be useful for routine monitoring in soils.     

Liquid chromatographic determination of natural and synthetic colorants in lyophilized foods using an automatic solid-phase extraction system

Five synthetic and five natural colorants were determined in lyophilized dairy products and fatty foods using an automatic method based on lixiviation and a solid-phase extraction process that includes cotton and RP-C(18) columns for the sequential retention of synthetic colorants and natural colorants, respectively. The lyophilization of the sample coupled with the separation procedure provides clean extracts despite the complexity of the matrices studied. In addition, the lyophilization process preserves the sample for at least 2 months without changes in the concentrations of the colorants. Identification and determination of synthetic and natural colorants were carried out using a liquid chromatograph equipped with a diode array detector. The detection limits achieved for all of the colorants (0.03-75 microg/g of lyophilized sample) allowed their determination within the limits established by the European Union, with good precision (approximately 4.5%). In addition, colorants spiked to different foods provided average recoveries (spiked at three concentration levels in four types of dairy samples and in three types of fatty foods) near 94 +/- 4%.     

Capillary gas chromatographic method for determination of benzylpenicillin and other beta-lactam antibiotics in milk

A capillary gas chromatographic method is described for determining residues of beta-lactam antibiotic residues in milk, with specificity for benzylpenicillin (penicillin G), phenoxymethylpenicillin, methicillin, oxacillin, cloxacillin, dicloxacillin, and nafcillin. Residues are extracted from milk with acetonitrile. Samples are cleaned up by partitioning between aqueous and organic phases at different pH values. The penicillin residues are methylated with diazomethane to render them amenable to determination by gas chromatography on a methyl silicone fused silica column. Samples are introduced by split/splitless injection using a programmed temperature vaporization injector and are detected by nitrogen-selective thermionic detection. Internal standardization is used for quantitation. The limits of detection for all penicillins are well below 1 microgram/kg. Recoveries of spiked samples at 3 and 10 micrograms/kg are in the range of 42-85% (coefficients of variation 2-5%) and 41-92% (coefficients of variation 3-7%), respectively.