Journal of Chromatography A, 1034 (2004) 33–40

Rapid multiresidue extraction method of organochlorinated

pesticides from fish feed夽
V. Nardelli a,∗ , C. Palermo b , D. Centonze b,1
a Istituto Zooprofilattico Sperimentale della Puglia e della Basilicata, via Manfredonia 20, 71100 Foggia, Italy
b Dipartimento di Scienze Agro-ambientali, Chimica e Difesa Vegetale, Università degli Studi di Foggia, via Napoli 25, 71100 Foggia, Italy

Received 23 September 2003; received in revised form 6 October 2003; accepted 4 February 2004

Abstract

A rapid multiresidue extraction method for organochlorinated pesticides from fish feed was developed, which is based on the extracted fat
treatment by n-hexane, concentrated sulphuric acid and ENVI-carb, a graphitized non-porous carbon material. The final residue, obtained in
about 50 min, was dissolved in isooctane and analysed by gas chromatography with an electron capture detector (GC/ECD). The presence
of the extracted pesticides was confirmed by gas chromatography–mass spectrometry (GC/MS). Concentration of sulphuric acid and amount
of ENVI-carb were optimized in order to improve analytes recovery, accuracy and detection limits. This simple and relatively fast method
allowed a high recovery of the HCB, Lindane, HEPO, p,p -DDE, p,p -DDD, p,p -DDT residues, with mean values in the range 68–124%
at four fortification levels (12.5, 25.0, 50.0, 100.0 ng/g), and coefficients of variation between 1.9 and 20.2%. Detection limit were equal to
3.0 ng/g, related to fat, for all pesticides, and calibration curves were linear (r > 0.999) in the range of explored concentrations from the
detection limit to 100 ng/g. For all pesticides a good repeatability was obtained (CV% values in the range 0.23–4.16%) when a sequence of
six injections of the isooctane extraction solution was performed. The usefulness of the proposed method has been tested by the analysis of
fish feed samples.
© 2004 Elsevier B.V. All rights reserved.

Keywords: Multiresidue extraction; Fish feed; Cleanup methods; Organochlorine compounds; Pesticides

1. Introduction A particular attention should be addressed to organochlo-

rinated pesticides that, in some cases, are still used in the
The increasing interest for the sea culture farms has worldwide or they are present as persistent residues of previ-
boosted remarkably modern techniques of intensive breed- ous uses. In fact, fish feed, contaminated by such substances,
ing, with a particular focus on the rationalization of feeding represent a potential way of direct introduction into fishes,
systems. Nevertheless, the use of highly nutritious feed re- where they accumulate in various organs especially in adi-
quires accurate sanitary controls of public health. In fact, pose tissue [1], and then into human beings. Therefore, to
in such matrixes it is frequent to find highly toxic chemical prevent the human health risk, a monitoring of these pesti-
products (heavy metals, PCBs, dioxins and pesticides) or cides in fish feed, also to identify their origin, is absolutely
substances not permitted, added for a preservative purpose required.
or as growth promoters. Note that fish feed are very complex matrixes for the
presence, among the other components, of animal origin
products too. Consequently, in the analysis of such matrixes
夽 Presented at the 4th European Pesticide Residues Workshop, Rome, a difficult task is represented by the sample cleanup that
Italy, 28–31 May 2002. should be efficient enough, in terms of analyte recovery,
∗ Corresponding author. Tel.: +39-0881-786309;
elimination of interferences and rapidity, to allow a reliable
fax: +39-0881-786362. screening of contaminated samples.
E-mail addresses: v.nardelli@izsfg.it (V. Nardelli),
centonze@unifg.it (D. Centonze).
Actually, a number of methods are currently used to
1 Co-corresponding author. Tel.: +39-0881-589118; extract organochlorinated pesticides from fatty samples.
fax: +39-0881-740211. The most commons involve adsorption chromatography on

0021-9673/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2004.02.018
34 V. Nardelli et al. / J. Chromatogr. A 1034 (2004) 33–40

Florisil, [2,3] Alumina, [4,5] and Silica gel [6,7]. These ent solvents in the order of HPLC grade water, acetone and
procedures are expensive, owing to the relatively high cost n-hexane.
of adsorbents, and have a low potential for automation.
Alternatively, gel permeation chromatography (GPC) [8,9] 2.2. GC/ECD and gas chromatography–mass spectrometry
is used, which offers a high degree of automation but it is (GC/MS) conditions
a relatively slow sequential sample cleanup, with a large
volume solvent use. Moreover, this technique is time con- 2.2.1. GC/ECD
suming and the relevant analyses of the extracts by gas A Perkin-Elmer (Monza, Italy) Gas chromatograph Mod.
chromatography with electron capture detector (GC/ECD) 8500 with electron capture 63 Ni detector, equipped with
reveal many interfering peaks [10]. Furthermore, previous Perkin-Elmer Software TC4, was used for GC/ECD de-
papers, report either direct treatment with concentrated termination. Chromatographic separations were performed
sulphuric acid [11] or cleanup on Extrelut column, soaked using an Alltech AT-5 (Deerfield, IL, USA) fused-silica cap-
with sulphuric acid [12], before the analysis by GC/ECD. illary column (30 m × 0.25 mm i.d.) with 5% diphenyl-95%
Syhre et al. [10] developed a different cleanup approach for dimethylsiloxane liquid phase, (0.25 ␮m film thickness)
the monitoring of chlorinated compounds in animal feed, with a relative deactivated retention gap.
based on a chromatographic step with ENVI-carb column. The oven temperature has been programmed as follows:
On the best of our knowledge, there are not papers in lit- 80 ◦ C for 2 min, ramped at 10 ◦ C/min to 160 ◦ C and then
erature reporting the extraction and analysis of organochlo- ramped at 5 ◦ C/min to 260 ◦ C. The final isotherm has been
rinated pesticides in fish feed. of 260 ◦ C for 20 min, with a total run of 50 min. The carrier
In the present paper, a rapid multiresidue extraction gas was helium, with a flow of 0.8 ml/min (160 ◦ C) and a
method of organochlorinated pesticides in fish feed has column head pressure of 1.02 atm.
been developed, which involves only the fat extraction from A split/splitless injector with 2 mm i.d. glass-liner has
feed samples by Soxhlet with petroleum ether, and the fol- been used in the splitless mode for 1 min. Injector temper-
lowing cleanup, with a simple procedure based on the use ature was 240 ◦ C and split flow was of 24 ml/min until the
of n-hexane, concentrated sulphuric acid and ENVI-carb, a end of analysis; the injection volume was of 1 ␮l. The elec-
graphitized non-porous carbon material. tron capture detector (ECD) temperature was of 300 ◦ C and
The cleanup has been optimized in terms of ENVI-carb nitrogen was the auxiliary gas with a flow of 55 ml/min.
amount and concentration of sulphuric acid, in order to im-
prove recovery of analytes, accuracy and detection limits. 2.2.2. GC/MS
The proposed method has been employed to detect the A Perkin-Elmer (Monza, Italy) Gas chromatograph Mod.
presence of organochlorinated pesticides in fish feed samples AutoSystem XL with Turbomass Gold Mass Spectrometer,
collected from local fish farms. equipped with Perkin-Elmer Software Turbomass Ver. 4.4.0,
was used for GC/MS confirmation.
Chromatographic separations were performed using a
2. Experimental PE-5 MS (Monza, Italy) fused-silica capillary column
(30 m × 0.25 mm i.d.) with 5% diphenyl–95% dimethyl-
2.1. Reagent and materials siloxane liquid phase, (0.25 ␮m film thickness). The oven
temperature has been programmed as follows: 80 ◦ C for
All solvents were ultra-Resi-Analyzed grade from Merck 2 min, ramped at 10 ◦ C/min to 160 ◦ C and then ramped at
(Darmstadt, Germany). Organochlorinated pesticides (Lin- 5 ◦ C/min to 260 ◦ C. The final isotherm was of 260 ◦ C for
dane, HEPO, HCB, p,p -DDT, p,p -DDE, p,p -DDD) (Su- 20 min, with a total run of 50 min. The carrier gas was
pelco Inc., Bellefonte, PA, USA) (reference standards) were helium with a constant flow of 1 ml/min. A PSS injector
from the collection in our Laboratory. All of them were with 1 mm i.d. quartz liner was used in the following split-
neat compounds (purity >96%). Standard solutions of pes- less mode: time 1 = −0.50 value 0; time 2 = 1.50 value
ticides were prepared in isooctane and stored in freezer at 100; time 3 = 1.80 value 10. The injector temperature
−20 ◦ C. The stock standard solution of each pesticide was was 250 ◦ C, and the injection volume was of 1 ␮l. GC/MS
of 2 mg/l. Intermediate standard solutions were prepared detection functions are summarized in Table 1.
by dilution in isooctane to give working concentrations of
3.0–6.0–12.5–25.0–50.0–100.0 ng/ml. 2.3. Method
ENVI-carb (120/140 mesh, 100 m2 /g) was purchased
from Supelco Inc. (Bellefonte, PA, USA). To remove all 2.3.1. Sample preparation
impurities, it was plentifully washed with different solvents, The fish feed samples were ground in a mill (Foss Teca-
in the following order: n-hexane, cyclohexane and toluene, tor Cemotec Mod. 1090, Hoganas, Sweden) and kept dry
and finally again with n-hexane. until the fat extraction. The fat was extracted from fish feed
All glassware were treated at first with sulphochromic by Soxhlet (Soxtec System: Foss Tecator Mod. 1046 Servit
mixture (Carlo Erba-Milano) and, then, washed with differ- Unit/Mod. System HT2 1045, Hoganas, Sweden) under the
 V. Nardelli et al. / J. Chromatogr. A 1034 (2004) 33–40 35

Table 1 3. Results and discussion

GC/MS detection functions
Solvent delay (min) 5 3.1. Cleanup optimization
MS pressure (Torr) 7.69e−6
Ionization mode Electron impact (EI+)
Accurate analyses of fish feed samples, which are contam-
Function 1 inated by organochlorinated pesticides, require an efficient
Scan duration (s) 0.10 method of extraction and cleanup.
Interscan delay (s) 0.020
Retention window (min) Always on
Among all the reagents, typically employed on the pesti-
Function type Scan cides extraction cleanup from fat samples, the potential of
Mass range 45–450 combined use of ENVI-carb, a graphitized non-porous ma-
Function 2 terial with adsorbent properties and of the concentrated sul-
Inter channel delay (s) 0.001 phuric acid, have been studied.
Retention window (min) 0.000–50.000 In a preliminary screening the cleanup efficiency was eval-
Function type SIR uated by the treatment of pesticide-free fish feed fat samples
Chan mass 282.00–284.00–286.00–181.00
(in the following referred to blank samples), and the rele-
183.00–246.00–248.00–235.00–237.00
Dwell (s) 0.05 vant GC/ECD analysis was performed only when the extract
solution was colourless and clear.
The choice of an appropriate ENVI-carb amount was es-
sential to obtain a good cleanup in terms of interferences
following conditions: the thimble was loaded with about 10 g elimination. Five different amounts of a such graphitized
(W1 ) of the mixed sample and covered with a thin layer of non-porous material, in the range 0.05–0.4 g, were used in
cotton, previously scoured by petroleum ether, then it was the cleanup procedure of blank samples with a 90% concen-
inserted into the Soxtec HT; the extracted fat collection cup trated sulphuric acid solution; for each amount of ENVI-carb
was dried and pre-weighed (W2 ), 80 ml of petroleum ether the experiment was performed twice. The best results were
were added into the cup and then it was inserted into the obtained using 0.1 g of ENVI-carb, in fact higher amounts
Soxtec HT. The time required for this extraction was 1 h gave pink coloured solutions that clearly indicated an ineffi-
at 103 ◦ C. After solvent evaporation, the cup was released, cient elimination of interferences. Moreover, by using 0.05 g
dried at 100 ◦ C for 3 h, cooled in a glass dryer for 30 min, of ENVI-carb, a faint coloured solution was obtained, which,
and then weighed (W3 ). Percentage of fat was calculated ac- as expected, showed a number of negative and positive inter-
cording to the formula: fat (%) = (W3 − W2 )/W1 × 100. fering peaks in the GC analysis and a high background noise.
The fat residue was transferred to a glass vial and preserved In the extraction cleanup of pesticides from fat samples,
in freezer at −20 ◦ C until the cleanup. the concentration of sulphuric acid plays an important role
in the fat digestion and degradation of organic substances.
2.3.2. Cleanup By using the optimized amount of ENVI-carb, the ef-
The extracted fat was purified in less than 50 min by a fect of sulphuric acid concentration in the range 70–95%
simple procedure based on the use of n-hexane, concentrated has been studied. Best results in terms of fat digestion effi-
sulphuric acid and ENVI-carb, a graphitized non-porous car- ciency, colourless solutions, background noise and absence
bon material. of negative or positive interfering peaks have been obtained
Cleanup procedure was carried out as follows: 2 ml of at concentrations of 90 and 95%.
n-hexane were added to about 0.5 g of fat in a Pyrex tube, In fact, using such two concentrations, the results ob-
and mixed by vortex for 1 min; 0.1 g of ENVI-carb, previ- tained by the analysis of spiked blank samples at 100 ng/g
ously washed with different solvents (in the order n-hexane, fortification level of a standard solution have shown that
cyclohexane, toluene and again n-hexane), were added and for p,p -DDT and its congeners the mean peak areas are
mixed by vortex for 1 min; 2 ml of concentrated sulphuric not significantly different. On the contrary, HEPO, Lindane
acid (90%) were pipetted slowly into the Pyrex tube, avoid- and HCB have shown higher values when a concentration
ing to touch the inner walls. The Pyrex tube was cooled and of 90% has been employed. Under these latter cleanup con-
then the solution was mixed by vortex for 1 min. The mixture ditions, lower noise and higher signal to noise ratios have
was centrifuged for 30 min at 5000 rpm (Tehtnica-Centric been obtained.
322A-Mod. TEH 464000, Zelezniki, Slovenia), and the up- These results can be explained considering a stronger ac-
per clear organic phase solution was pipetted by a Gilson tion of the sulphuric acid at 95%, which probably determines
Microman Pipette (Gilson-Mod. M250 S/N-Villiers-le-Bel, a partial degradation either of the less resistant pesticides
France) and the volume was accurately measured. or of ENVI-carb, with a consequent release of interfering
The solution was carefully evaporated by Bain Marie at substances instead of their elimination.
45 ◦ C under a nitrogen flow, the residue was dissolved in In the light of these results, all the following experiments
1 ml of isooctane, and the obtained solution was analysed have been carried out using 0.1 g of ENVI-carb and sulphuric
by GC/ECD and GC/MS. acid at 90%.
36 V. Nardelli et al. / J. Chromatogr. A 1034 (2004) 33–40

Fig. 1. GC/MS analysis in SIFI of HCB at a 166 ng/g concentration.

The high efficiency of the cleanup of the method was also ion (282 + 284 + 286) for HCB, while “trace b” shows the
assessed by the analysis of standard solutions and spiked extract ion chromatogram (m/z = 284) from the full scan
blank samples in GC/MS, which is the most useful tech- acquisition.
nique in the confirmation of the chromatographic peaks as- Although SIR mode increases efficiency and sensitivity
signment. by a selective scan of individual m/z ratio, for the best identi-
A 166 ng/g standard solution of organochlorinated pesti- fication of each compound the full scan mode has been used
cides has been analysed by GC/MS in selected ion and full to allow the acquisition of library-searchable spectra. The
ion (SIFI) mode acquisition, and in Fig. 1 are shown the full mass spectra of each pesticide present in the extract of
results for HCB. In particular, “trace a” shows the selected the spiked blank sample fitted well with those obtained from
ion recording (SIR) acquisition of the three characteristic the standard solution. In Fig. 2 is reported an example of

Fig. 2. Library search result for p,p -DDE.

 V. Nardelli et al. / J. Chromatogr. A 1034 (2004) 33–40 37

Fig. 3. Overlay of a spiked blank sample (a) and a standard solution (b) of organochlorinated pesticides at a 100 ng/g concentration.

library search result, relevant to p,p -DDE, where the exper- obtained for all investigated pesticides, as evidenced from
imental mass spectra is compared with the two best hits of the coefficient variation values in the range 0.23–4.16%.
the Hit List. As it can be seen from the high value of reverse The highest CV% obtained for p,p -DDT can be ascribed to
fits (90%) a good confirmation was achieved for p,p -DDE, the random breakdown of this pesticide in the injector (vide
notwithstanding the real blank sample was spiked with only infra). Even if the HEPO CV% was very low (1.03%), a
80.0 ng/g of p,p -DDE. Same results were obtained for the slight natural degradation was observed by the decrease of
other pesticides. peak areas during the sequence of injections.

3.2. GC/ECD determination 3.3. Recovery study

The GC experimental conditions were optimized in terms The recovery study, performed on blank feed fat sam-
of temperature program that allowed an improvement of the ples spiked with known levels of the six organochlorinated
time and the chromatographic run resolution. Moreover, to pesticides are summarized in Table 2. Six replicates for
avoid the cross contamination between high and low spiked each sample have been carried out at four fortification lev-
blank samples, the sequence of injections was in the follow- els of 12.5–25.0–50.0–100.0 ng/g. and the relevant recov-
ing order: solvent, blank sample, spiked blank sample, and ery results, given as mean values, were in the range of
finally standard solutions. 68–124%. In our opinion, higher recoveries observed for
Typical GC/ECD chromatograms are shown in Fig. 3 p,p -DDE (124%), p,p -DDD (122%), and p,p -DDT (116%)
for a spiked blank sample and a standard solution of are ascribed to matrix effect. Besides, a lower recovery for
organochlorinated pesticides. The gas chromatogram of the p,p -DDT, respect to p,p -DDE and p,p -DDD, is probably
spiked blank sample seems to be free of interfering peaks due to its thermal breakdown in the injector (T = 240 ◦ C).
and it shows a satisfactory overlay, in terms of retention This degradation produces p,p -DDE and p,p -DDD as evi-
times, with that of the standard solution. Quantification denced in the chromatogram of Fig. 4. The recoveries are sat-
has been carried out through peak area comparison with isfactory for all compounds apart for HEPO (not greater than
the external standard technique and a six-level calibra- 76%); this low value can be ascribed to the acidic treatment
tion (3.0–6.0–12.5–25.0–50.0–100.0 ng/g). The calibration that probably converts such epoxide into the correspondent
curves of the analysed pesticides present a good regression
line (r > 0.999) in the range of explored concentrations, Table 2
from the detection limit to 100 ng/g. The detection limit of Results for the recovery experiments of the six organochlorinated pesti-
cides from spiked blank feed fat samples
each persistent pesticide was of 3.0 ng/g, related to fat, cal-
culated automatically by Turbochrom–Perkin-Elmer Soft- Pesticide Average recovery (%) ± S.D.a
ware as a signal to noise ratio of five. This low detection 12.50 25.00 50.00 100.00 Mean
limit was achieved because of the efficient cleanup step that
HCB 91 ± 16 103 ± 7 107 ± 4 113 ± 10 103 ± 9
allows the elimination of possible interfering substances, Lindane 104 ± 13 102 ± 6 103 ± 2 110 ± 9 105 ± 7
obtaining then a low noise value. HEPO 76 ± 2 61 ± 5 62 ± 4 72 ± 12 68 ± 6
The repeatability of GC/ECD determinations was as- p,p -DDE 134 ± 10 121 ± 17 122 ± 10 121 ± 8 124 ± 11
sessed in 5 h period by a series of six replicate injections of p,p -DDD 124 ± 20 129 ± 14 115 ± 9 120 ± 10 122 ± 13
the final isooctane extraction solution from a spiked blank p,p -DDT 119 ± 24 116 ± 11 111 ± 14 118 ± 14 116 ± 16
sample, fortified at a 100 ng/g level. A good precision was a S.D.: standard deviation (n = 6).
38 V. Nardelli et al. / J. Chromatogr. A 1034 (2004) 33–40

Fig. 4. Example of thermal breakdown of p,p -DDT, obtained by the injection of 100 ng/g standard solution.

diol or sulphate, which are much more soluble in the water ing substances when used in the extraction of pesticides from
phase and then not extractable by n-hexane. In fact, as it can vegetable based animal feed [10].
be observed in Fig. 5, the recoveries for HEPO became ac- On the other hand, the use of sulphuric acid is a simple,
ceptable (85%) when the sulphuric acid was replaced with fast and efficient cleanup method when applied to foods
the HPLC grade water, during the cleanup procedure. How- [11,12], but in the case of feedingstuffs for poultry [11] a
ever, during the cleanup, an occurring natural degradation of number of coeluting peaks in the chromatogram affect the
this less resistant pesticide, giving water-soluble products, determination of PCBs.
can not be ruled out (vide ante). In the same way, even if vegetable based feed are pro-
The combined use of sulphuric acid and ENVI-carb pow- cessed, the use of cartridges home-packed with ENVI-carb
der represents a valid alternative cleanup procedure over [10] gives chromatograms showing a high background
the existing methods suggested in the European Committee noise and a number of interfering peaks that could affect
for Standardization (CEN) guide-line [13], as well as other the determination of pesticides when real-life samples are
methods proposed in the literature [10–12]. analysed. Moreover, the fortification of reference matrix
In addition to the high cost of adsorbent materials, the is carried out on the extract from accelerated solvent ex-
sample cleanup based on Florisil columns requires the use traction (ASE) system, which has not yet been validated
of a large volume of different eluting solvent mixtures to [14]. Finally, the cartridges present overloading problems,
recovery all pesticide species. GPC based methods are time so that the amount of the ASE extract to be purified should
consuming procedure and they require expensive instrumen- be optimized dependently on the particular real matrix
tation, without assuring an efficient elimination of interfer- processed.

Fig. 5. Overlay of two procedural blanks: (a) procedural blank of HPLC grade water, spiked with HEPO; (b) procedural blank of concentrated sulphuric
acid, spiked with HEPO.
 V. Nardelli et al. / J. Chromatogr. A 1034 (2004) 33–40 39

Fig. 6. Chromatogram of a naturally contaminated fish feed sample after fat extraction and cleanup.

Table 3
Content of organochlorinated pesticides in contaminated fish feed samples
Sample Fish feed fat (%) Average concentration (ng/g fat) ± S.D. (n = 3)

HCB Lindane HEPO p,p -DDE p,p -DDD p,p -DDT

1 20.3 7.56 ± 0.19 13.29 ± 0.39 10.78 ± 0.38 6.32 ± 0.40 7.81 ± 0.33 7.02 ± 0.27
2 22.6 8.10 ± 0.18 14.26 ± 0.25 <LODa 5.74 ± 0.39 6.75 ± 0.46 7.37 ± 0.34
3 23.2 7.92 ± 0.37 15.10 ± 0.31 10.18 ± 0.49 6.25 ± 0.28 6.79 ± 0.40 7.68 ± 0.41
a LOD: limit of detection (signal-to-noise ratio of 5).

On the contrary, in the analysis of more complex matrices to a t-test at 95% confidence level, with those obtained
as fish feed, by the synergic effect of ENVI-carb and sul- by GC/MS.
phuric acid chromatograms with low noise levels and with-
out interfering peaks have been obtained. In spite of the
simplicity of operations and the low-cost of materials and 4. Conclusions
instrumentations, pesticide recovery values, detection limits
and reproducibility compared-well with those obtained by The developed rapid multiresidue extraction method is
the other methods. suitable for monitoring of organochlorinated pesticides in
fish feed. The combined use of concentrated sulphuric acid
3.4. Analyses of real samples and ENVI-carb allows a quantitative cleanup extraction of
the analytes. No complicated apparatus are required and
Recently, the European Parliament and the Council have other advantages are the need of a low organic solvent vol-
published [15] a new directive, where have been fixed ume and a non-intensive manual labour requirement. By this
the limits of several residues in feedingstuffs including method good results are obtained over a wide range of ana-
organochlorinated pesticides. To assess the potential of the lyte concentrations in terms of precision, linearity, accuracy,
proposed sample cleanup method 20 fish feed samples, detection limits and recovery.
collected from local fish farms, have been analysed. The The usefulness of the proposed method has been demon-
results have evidenced the presence of pesticides in three strated by the determination of organochlorinated pesticides
of the samples but at concentrations below the legal limits; in feed fish samples collected from local fish farms.
Fig. 6 shows a typical chromatogram obtained for a nat- This low-cost and simple procedure, based on rapid and
urally contaminated fish feed. Quantification of pesticides safe operations, may be a useful tool in routine analysis of
has been carried out through peak area comparison with the the organochlorinated pesticides, in place of the currently
external standard technique and the correspondent concen- used conventional techniques.
tration values are given in Table 3 together with the percent
of the fat content. The elimination of interfering substances
and the low background noise, obtained by the efficient Acknowledgements
cleanup step, have permitted a high precision (CV% in the
range 1.7–6.8%) and accuracy in the determination of each The authors thank Dott. M. Santoro for his helpful dis-
compound; in fact the results compared-well, according cussions and contributions in GC/MS work. The authors
40 V. Nardelli et al. / J. Chromatogr. A 1034 (2004) 33–40

are grateful to Dott. D. Palermo for his invaluable teaching [5] P.A. Greve, W.B.F. Grevenstuck, Med. Rijksfac. Landbouwwet. Gent
and encouragement, which were very useful in the develop- 40 (1975) 1115.
[6] H. Steinwandter, H. Schluter, Fresenius Z. Anal. Chem. 286 (1977)
ment of the present work. Furthermore, we thank the Istituto 90.
Zooprofilattico Sperimentale della Puglia e della Basilicata [7] H. Steinwandter, H. Schluter, Dtsch. Lebensm. Rundsch. 74 (1978)
for the financial support. 139.
[8] W. Specht, M. Tillkes, Fresenius Z. Anal. Chem. 322 (1985) 443.
[9] P. Armishaw, R.G. Millar, J. AOAC Int. 76 (6) (1993) 1317.
[10] M. Syhre, G. Hanschmann, R. Heber, J. AOAC Int. 81 (3) (1998)
References 513.
[11] C. von Holst, A. Müller, E. Björklund, E. Anklam, Eur. Food Res.
[1] D.L. Murphy, J.W. Gooch, Arch. Environ. Contam. Toxicol. 29 Technol. 213 (2001) 154.
(1995) 297. [12] A. Di Muccio, A. Santillo, R. Dommarco, M. Rizzica, L. Gambetti,
[2] B. McMahon, J.A. Burke, J. AOAC Int. 61 (1978) 640. A. Ausili, F. Vergori, J. Chromatogr. 513 (1990) 333.
[3] T. Suzuki, K. Ishikawa, N. Sato, K.-I. Sakai, J. AOAC Int. 62 (1979) [13] EN 1528-2, 1996.
681. [14] EN 1528-3, 1996.
[4] A.V. Holden, K. Marsden, J. Chromatogr. 44 (1969) 481. [15] European Commission, Off. J. Eur. Commun. L140 (2002) 10.