Collaborative study on toxaphene indicator compounds (chlorobornanes) in fish oil

Chemosphere, Vol. 35, No. 7, pp. 1391-1398, 1997 © 1997ElsevierScienceLtd All rightsreserved.Printedin GreatBritain 0045-6535/97 $17.00+0.00

Pergamon

PII: S0045-6535(97)00234-8

COLLABORATIVE

STUDY ON TOXAPHENE INDICATOR (CHLOROBORNANES) IN FISH OIL

COMPOUNDS

Lutz Alder 1, Karl Bache 1, Hans Beck I and Harun Parlar 2 i Federal Institute for Health Protection of Consumers and Veterinary Medicine, PO-Box 330013, D-14191 Berlin, Germany 2 Chair of Chemical-Technical Analysis and Chemical Food Technology, Technical University of Munich, D-85350 Freising-Weihenstephan, Germany (Received in Germany 7 April I997; accepted 23 May 1997) ABSTRACT A validation study of a method for the analysis of four toxaphene indicator compounds in a fatty matrix was carried out. The method of analysis consisted of cleanup by gel permeation and absorption chromatography followed by GC-ECD determination. In total, five materials with indicator compound concentrations in the range from 6 to 60 J.tg/kg and a blank material were analysed. The mean recovery was found to be between 77 and 100%. For three compounds, the analytical results have shown a relative standard deviation of reproducibility of 21 _ 4 %. A higher variation was found for the fourth compound (mean 29 %). The method is recommended for routine analysis in food inspection in Germany, © 1997 Elsevier Science Ltd

1. I N T R O D U C T I O N Chlorinated camphene (ISO common name: camphechlor; in USA, Belgium and Canada: toxaphene) is probably the most complex pesticide and more than one million tons of it have been applied since 1946 [1]. Owing to its widespread use and the environmental stability of some components, toxaphene residues have been identified in diverse environmental matrices, including air, freshwater, sea water, and soil. The highest concentrations of this fat-soluble pesticide have been detected iv. aquatic organisms, particularly fish, dolphins, and whales [2-5]. Determination of toxaphene (camphechlor) residues in environmental matrices and food, is very difficult. Most problems arise from the complex composition of toxaphene and the substantial differences in peak patterns in environmental samples as compared to the technical product. Based on earlier preparations of pure toxaphene congeners, we identified three chlorobomanes (CHBs) which are present in marine fish in relatively high concentrations. These CHBs 1-3 are proposed as indicator compounds for toxaphene residues (Figure 1) [6]. A further chlorobornane (compound 4) is under discussion as an indicator for certain applications. Additionally, these congeners are proposed for the establishment of maximum residue limits (MRLs). Using these individual CHBs as standards, a simple and reproducible determination of toxaphene residues in samples is possible without the need of special cleanup procedures and conventions according to standards, GC methods or quantification procedures. To check whether the congeners can be reliably detected and quantified, a method validation study was conducted on the determination of these toxaphene indicator compounds by a multipesticide cleanup combined with GC/ECD detection. 1391

1392 Figure 1: Stmcture of Indicator compounds for toxaphene residues

CI

CI

CI

C1

CI

CI

1

C1

C1 ~11C1

CI CI

CI

CI CiCI

Indicator Compound 1:

Indicator Compound2:

Indicator Compound3:

Indicator Compound4:

2-endo,3-exo,5-endo, 6-exo,8,8,10,10-octachlorobomaae = T2 = TOX 8 = Parlar's 26

2-endo,3-exo,5-endo,6-exo, 8,8,9,10,10-nonachlorobornane = Tt2 = TOX 9 = Toxicant Ac= Parlar's 50

2,2,5,5,8,9,9,10,10nonachlorobornane = Parlar's 62

2,2,5-endo,6-exo,8,9,10heptachlorobomane = Toxicant B = Parlar's 32

2. M E T H O D

USED FOR CLEANUP AND GC/ECD DETECTION

The cleanup procedure tested in this study was a combination of gel permeation and adsorption chromatography on silica as published earlier [7-8]. This methodology is widely used throughout Europe and included in a proposed European standard for the determination o f pesticides in fatty food [9]. In brief, the lipids from 0,5g oil sample were removed by gel permeation chromatograph, fitted with a 2.5 x 40 cm column containing 50 g Bio-Beads SX-3 (Bio Rad Laboratories, 200-400 mesh). The mobile phase, consisting o f cyclohexane/ethylacetate (1:1), was introduced into the column at a flow rate of 5 ml/min. The toxaphene congeners were eluted together with PCBs, other chlorinated pesticides, such as chlordane components or DDT and metabolites in the range of 95-150 ml. After the removal of solvent, the sample was dissolved in isooctane and then applied to a 0.7 x 23 cm chromatography column packed with 1 g silica gel (Merck No. 7734,

Table 1: GC parameters used Lab. No.

Injector type

Injector temp.

Carrier gas

I. Column

2. Column

1 2

splitless splitless

230 °C 230 °C

H2 He

DB-5, 60m X 0.25mm x 0.25~tm DB-5, 60m × 0.32ram x 0.251am

splitless splitless splitless n.r. split n.r. splitless Splitless splitless on'column splitless splitless

235 °C 230 °C 230 °C 240 °C 240 °C 240 °C 230 oc 230°C 220 °C 235 °C 250 °C

N2 n.r. N2 n.r. H2 n.r. HE He H2 N2 H2 H2

DB-5, 30m X 0.32ram × 0.251am DB-5, 60m x 0.32mm × 0.25/am DB-5, 30m x 0.25mm × 0.101am DB-5, 30m x 0.25mrn x 0.25~m DB-5, 30m × 0.25mm × 0.251am DB-5, 60m x 0.32mm × 0.25~tm DB-5, 60m x 0.32mm × 0.251am HP-5, 50m × 0.32mm X 0.52~m DB-5, 60m × 0.25mm × 0.251am DB-5, 60m x 0.25mm x 0.251am DB-5, 30m x 0.32mm x 0.251am DB-5, 60m :,< 0.25mm ::< 0.251am

DB-1701, 60m x 0.25mm x 0.25!am DB-1301, 30m x 0.25mm x 0.251am DB-1301, 30m X 0.32mm X 0.25lam DB-1, 30m x 0.32mm X 0.251am DB-1701, 30m x 0.32mm X 0.151am DB-1701, 30m x 0.25mm X 0.251am none DB-1301, 30m x 0.32mm x 0.251am Rtx-1701, 30m x 0.32mm x 0.251am DB-1301, 30m x 0.32mm x 0.251am DB-1301, 60m x 0.25mm x 0.251am

3 4 5 6 7 8 9 10 11 12 13 14

n.r. - not reported

none DB-1701, 30m x 0.32ram x 0.251am Rtx-20, 60m × 0.25mm x 0.251am

1393 deactivated with 1.5 % water) and with a layer of 1 cm dried sodium sulfate on the top. The toxaphene components, other organochlorine pesticides and PCBs were fractionated by consecutive elution with 8 ml hexane/toluene (65:35 v/v, fraction 1) and 8 ml toluene (fraction 2). The first fraction contained the toxaphene components and, in the case of cod liver oil samples, DDE, DDD, DDT, chlordane components and the PCBs, additionally. The second fraction was used to check the completeness of elution. In contrast to the published method, the final extracts were concentrated to 1 ml. Up to this step, all participants strictly followed the protocol. Gas chromatography was started in each case by using a SE-54 type capillary column of at least 30m length. 65% of participants used 60m capillaries. Generally, columns with an internal diameter of 0.25mm or 0.32mm were used. In most cases, the results were validated with capillaries of higher polarity (DB-1301/DB-1701). Hydrogen or helium were used as carrier gases. The temperature of the splitless injector was in the range of 220 to 250 °C with a median of 230 °C. For more details see Table 1.

3. D E S I G N O F T H E S T U D Y The study was based on the nonreplicate split level design for collaborative tests (Youden pairs) [10]. For the preparation of test samples, two starting materials were used: 1. Corn oil without any toxaphene residues 2. Commercial cod liver oil with organochlorine residues (including toxaphene) typically found in marine fish samples In addition to these unspiked materials pools were prepared from both materials spiked with CHBs both at a lower and at a higher level. For the preparation of the Youden pairs, each of the resulting six materials was used undiluted (= 100 %) and diluted with corn oil to a 90 % level. The final concentrations of indicator CHBs 1-4 were in the range from 6 to 60.2 lag/kg (Tables 3-6). Each participating laboratory received six randomly coded sample ampoules (three Youden pairs). The selection of the samples varied in such a way that not more than two laboratories received the same set of samples and, on the other hand, that each sample was analysed at least by six laboratories. Additionally, each laboratory received a CHB calibration standard, three standard mixtures with PCBs, chlordane components and other chlorinated pesticides found in the cod liver oil, and a separate cod liver oil with specified concentrations of indicator compounds 1-4 for training purposes. All 15 participants received a complete description of the method to be used for the analysis. Individual deviations from the operating procedure were not allowed.

4. S T A T I S T I C A L T R E A T M E N T O F D A T A The resulting data were first inspected for major discrepancies. ,,Less than" and ,,not detected" values were considered as zero. Due to complications in the cleanup of samples (contamination problems) all results from one laboratory were excluded from the data set. All remaining concentration data are listed in Table 2. Outliers and method performance values were calculated from raw data using the AOAC guidelines [9]. Single Gmbbs' test (2-tail; P = 0.01) and pair value test (overall P = 0.01) were considered to detect extreme averages from one laboratory. Cochran's test (applied as a 1-tail test at a propability of 0.01) was used to check the homogeneity of within-laboratory variances. The single Grubbs' test detected the data of laboratory 4 for compound 3 in medium spiked corn oil as outlier. Both other tests never have been significant, i.e. from a total of 318 valid concentration

1394 Table 2a: Results of toxaphene indicator determination in materials I-3 Material 1 (corn oil. not spiked)

Split level

'True' concentration 7

10

11

12

13

14

Mean concentration found

Concentration found in laboratories (~g/kg)

Indicator compound 1

100% 90%

0.0 0.0

n,d. n.d.

n.d. n.d.

n.d. n.d.

0.3 0.6

n.d. n.d.

n.d. n.d.

0.1 0.1

Indicator compound 2

100% 90%

0.0 0.0

n.d. n.d.

n.d. n.d.

n.d. n.d.

1.5 1.8

n.d. n.d.

n.d. n.d.

0.3 0.3

Indicator compound 3

100% 90%

0.0 0.0

n.d. n.d.

n.d. n.d.

n.d, n.d.

1.5 1.5

n.d. n.d.

n.d. n.d.

n.d. n.d.

Indicator compound 4

100% 90%

0.0 0.0

a) a)

n.d. n.d.

n.d. n.d.

b) b)

n.d. n.d.

n.d. n,d.

n.d. n.d.

Material 2 (corn oil. medium spiked)

Split level

'True' concentration

Mean concentration

Concentration found in laboratories (~g/kg) 1

2

3

4

7

8

9

10

12

11.4 9.5

9.7 9.4

9.0 11.0

9.0 8.0

9.3 8.7

7.3 7.8

5.3 5.8

8.6 8.2

14.7 28.4 1 3 . 3 22.0

14.0 9.0

1 1 . 0 11.0 10.0 10.2

11.8 13.0

5.1 7.1

13.1 11.5

Indicator compound 1

100% 90%

9.4 8,46

8.0 7.0

8.6 7.4

Indicator compound 2

100% 90%

12,4 11,16

11.0 10.0

10.8 9.2

Indicator compound 3

100% 90%

12.2 10.98

10.0 9.0

9.5 8.3

Indicator compound 4

100% 90%

14.4 12.96

9.0 ll

Material 3 (cornoil. highspiked)

Split level

'True' concentration

20.5 18.9

1 2 . 7 16.6 12.8 13.9

7.9 6.5

8.0 6.0

11.0 10.0

c) c)

17.0 16.0

3.5 4.2

10.9 9.9

13,4 1046

a) a)

13.0 12

17.0 15.9

12.2 13.4

b) b)

13.4 12.8

Concentration found in laboratories (pg/kg) 5

6

8

found

9

11

13

14

Mean coneentration found

Indicator compound 1

100% 90%

28.2 25.4

24.5 21.7

28.1 24.6

24.0 24.0

30.4 21.0 2 7 . 1 17.6

20.5 21.7

33.8 30,0

26.0 23.8

Indicator compound 2

100% 90%

37.2 33.5

31.2 26.9

35.8 31.5

30.0 28.0

37.5 30.9

26.9 24.7

26.7 27.8

49,2 4Z6

33.9 30.3

Indicator compound 3

100% 90%

36.6 32.9

29.9 24.8

31.7 26.9

31.0 30.0

c) c)

33.4 29.1

25.4 26.6

34.3 31.6

31.0 28.2

Indicator compound 4

100% 90%

43.2 38.9

37.9 35.8

30.7 27.3

36.0 34.0

57.7 48.4

32.5 28.6

32.8 34.3

58.2 52.6

40.8 37.3

n.d. - not detected a) incomplete chromatographic separation)

b) unaceptable high bias of calibration curve c) unsufficient detector response for determiation

1395 Table 2b: Results of toxaphene indicator determination in materials 4-6

Material 4 (cod liver oil not spiked)

Split level

'True' concentration

Mean concentration found

Concentration found in laboratories (lag&g) I

2

3

4

5

6

8.6 7.6

8.7 7. I

6.9 6.4

5.0 4.1

6.7 5.9

24.4 39.6 1 9 . 8 27.4

17.1 17.4

22.0 20.1

23.3 20.4

16.7 22.8

9.4 12

14.3 13.0

14.3 15.3

10.1 8.0

n.d. n.d.

6.2 5.9

2.7 2.3

Indicator compound 1

100% 90%

6.6 6.0

5.0 4.0

6.1 6.3

Indicator compound 2

100% 90%

23.0 20.7

20.0 19.0

16.6 18.6

Indicator compound 3

100% 90%

15.6 14.0

12.0 11.0

10.1 23.3 1 2 . 3 20.6

Indicator compound 4

100% 90%

0.0 0.0

n.d. n.d.

Material 5 (cod liver oil, low spiked)

Split level

'True' concentration

n.d. n.d.

n.d. n.d.

Mean concentration found

Concentration found in laboratories (jag/kg) 1

8

9

10

Indicator compound 1

100% 90%

16.0 14.4

12.0 I1.0

15.0 13.0

1 4 . 6 11,2 1 3 . 5 10,4

Indicator compound 2

100% 90%

35.4 31.9

26.0 26.0

26.0 23.0

37.4 31.4

Indicator compound 3

100% 90%

27.8 25.0

19.0 23.0 21.0 22.0

c) c)

Indicator compound 4

100% 90%

14.4 13.0

11.0 11.0

Material 6 (cod liver oil. high spiked)

Split level

13.0 11.0

'True' concerttration

1t 9.9 9.5

12.5 11.5

26.0 25,2

25.5 21.9

28.2 25.5

20.6 20.4

17.7 18.9

20.1 20.6

1 4 . 6 1 5 . 4 10.8 13.5 15.4 9.9

13.0 12.2

Mean concentration

Concentration found in laboratories (lag&g) 2

3

4

5

6

7

12

13

14

found

Indicator compound I

100% 90%

34.8 31.3

29.0 31.9

32.8 29.7

31.8 32.6

31.3 25.8

29.2 23.7

26.0 28.0

40.6 33.1

30.2 27.9

36.8 35.2

32.0 29.8

Indicator compound 2

100% 90%

60.2 54.2

49.5 54.5

52.8 51.3

74.4 60.6

46.9 38.6

59.7 3 5 . 0 46.5 41.0

29.3 31.2

43.5 41.3

73.4 69.8

51.6 48.3

Indicator compound 3

100% 90%

52.2 47.0

35.7 45.0

52.8 51.3

42.2 40.2

3 8 . 1 51.6 33.0 39.0

1 5 . 5 22.3 21.2 13.7

74.9 58.9

40.7 37.8

indicator compound 4

100% 90%

43.2 38.9

40.8 36.3

52.2 46.9

54.4 48.8

36.9 22.2

49.6 47.2

44.2 38.0

n.d. - not detected a) incomplete chromatographic separation)

38.2 29.0

33.0 38.0 a) a)

b) b)

37.4 35.3

b) unaceptable high bias of calibration curve • c) unsufficient detector response for determiation

1396 data of 14 laboratories passing the initial review, only 2 had been rejected. Outlier tests and summary statistics (repeatability and reproducibility standard deviations st/s R and their relative values) were calculated by a computer program according to the AOAC statistics for split level design, where repeatability (r) and reproducibility limits (R) are simple multiples of the above measures of precision expressed as standard deviations (a factor of 2.8 based on a statistical probability of 95 % was used).

5. R E S U L T S The checked cleanup resulted in sufficiently resolved chromatograms. The PCBs included in the extracts of cod liver samples did not complicate the determination of indicator CHBs. Some participants would have preferred a cleanup, separating PCBs from the slightly more polar CHBs. Only in some laboratories, coelution of CHBs with other compounds has been observed (critical pairs: indicator compound 3 and diethylhexylphthalate or mirex; compound 4 and cis-nonachlor or pp'-DDD or PCB 153). With a second capillary the resolution of interesting peaks could always be achieved. The influence of the injector temperature on the response of the indicator compounds 2 and 3 has been confirmed by the participants. Injector temperatures higher than 240 °C should be avoided. Sometimes, such capillaries working well in most laboratories have shown drastically reduced intensities of particular indicator compounds. Decomposition on active sites has been discussed as possible reason for this phenomenon. A graphic representation of concentrations found versus true concentrations is shown in Figure 2. No significant systematic error (bias) was found in the data submitted. The summary statistics calculated after removal of the two outliers are presented in Table 3. Reproducibility standard deviations were found comparable to the Horwitz equation (RSD R = 2(1 - 0.5 log c), which is dependent from analyte concentration (c). This equation had been derived empirically from an examination of more than 3000 method performance studies. It has been proposed that reproducibility standard deviation values found within a range of 0.5-2 times the RSD R may be considered as an acceptable precision ofrnethod performance between laboratories[11]. Horwitz' RSD R are used in Table 3 as ,,acceptable relative standard deviation". Higher reproducibility standard deviations than twice the Horwitz' RSD R were never calculated from the data of this collaborative study.

6. C O N C L U S I O N S The analytical method for the determination of the proposed toxaphene indicator compounds in a fatty matrix was shown to be quantitative even at concentrations of 101ag/kg fat. Recovery from a vegetable and a fish oil was uniform and reasonably precise. Based on the range of fat content in commercially significant fish species (1-20 %) and the assumption of quantitative extraction of toxaphene residues together with the fish fat, a limit of determination of about 1-21ag/kg can be achieved on a wet weight basis. These results demonstrate the validity of the analytical method tested and the applicability of the toxaphene indicator compound concept.

6 5 9

Cod liver oil, not spiked Cod liver oil, low spiked Cod liver oil, high spiked

6 5 9

Cod liver oil, not spiked Cod liver oil, low spiked Cod liver oil, high spiked

6 4 9

Cod liver oil, not spiked Cod liver oil, low spiked Cod liver oil, high spiked

4 7 7

6 5 7

Corn oil, not spiked Corn oil, low spiked Corn oil, high spiked

Cod liver oil, not spiked Cod liver oil, low spiked Cod liver oil, high spiked

Toxaphene indicator compound 4

6 8 6

Corn oil, not spiked Corn oil, low spiked Corn oil, high spiked

Toxaphene indicator compound 3

6 8 7

Corn oil, not spiked Corn oil, low spiked Corn oil, high spiked

Toxaphene indicator compound 2

6 9 7

Corn oil, not spiked Corn oil, low spiked Corn oil, high spiked

Toxaphene indicator compound 1

No. of labs with valid results

12 10 14

8 14 14

12 8 18

12 16 12

12 10 18

12 16 14

12 10 18

12 18 14

No. of valid results

Table 3: Statistical results of method validation

0.0 14.4 43.2

0.0 14.4 43.2

15.6 27.8 52.2

0.0 12.2 36.6

23.0 35.4 60.2

0.0 12.4 37.2

6.6 16.0 34.8

0.0 13.0 38.9

0.0 13.0 38.9

14.0 25.0 47.0

0.0 11.0 32.9

20.7 31.9 54.2

0.O 11.2 33.5

6.0 14.4 31.3

0.0 8.5 25.4

x2

xi

0.0 9.4 28.2

'True' conc. at 90% level

'True' conc. at 100% level

0.0 13.7 41.0

0.0 13.7 41.0

14.8 26.4 49.6

0.0 11.6 34.8

21.9 33.6 57.2

0.0 11.8 35.3

6.3 15.2 33.1

0.0 8.9 26.8

(xl+x2)/2

Mean 'tree' conc.

2.5 12.6 41.1

n.d. 13.1 39.1

14.8 20.3 39.2

n.d. 10.4 29.6

21.9 26.8 50.0

0.3 10.7 32.1

6.3 12.0 30.9

0.1 8.5 24.9

x

92% 100%

96% 95%

77% 79%

90% 85%

80% 87%

91% 91%

79 % 93 %

95 % 93 %

Conc. Mean found recovery (mean of both levels)

0.6 3.1

1.3 2.3

2.3 1.0 6.1

0.6 1.8

3.6 1.7 5.1

1.5 2.0

0.4 0.4 2.6

0.8 1.4

Sr

1.7 8.7

3.6 6.4

6.4 2.7 17.0

1.6 4.9

10.0 4.7 14.3

4.1 5.5

1.2 1.2 7.1

2.3 3.9

2,8 x Sr

4.7% 7.6%

9.9% 5.9%

15.5% 4.7% 15.5%

5.4% 5.9%

16.3% 6.3% 10.2%

13.7% 6.1%

6.8 % 3.5 % 8.3 %

9.8 % 5.6 %

Sr/x

2.2 9.0

2.3 10.9

5.1 1.9 15.9

5.3 2.9

6.5 4.5 13.9

2.5 7.0

1.6 2.0 4.1

1.6 4.5

SR

6.0 25.1

6.5 30.4

14.3 5.2 44.6

14.7 8.0

18.3 12.5 38.9

7.0 19.5

4.5 5.5 11.4

4.5 12.6

2,8 × SR

17.2% 21.8%

1"7.7% 27.8%

34.6% 9.2% 40.6%

50.5% 9.6%

29.8% 16.7% 27.8%

23.4% 21.7%

25.2 % 16.4 % 13.2 %

19.2 % 18.0 %

SR/X

39.40% 30.90% 25.90%

30.70% 26.10%

30.20% 28.80% 26.00%

31.80% 27.20%

28.4% 27.6% 25.1%

31.7% 26.8%

34.3 % 31.1% 27.0 %

32.8 % 27.9 %

2~1-0,~loso)

R e pe at ab iIi ty R e p r o d u c i b i 1i t y Acceptstandard limit relative standard fimit relative able deviation r standard deviation R standard reprodeviation devia- ducibility tion (RSD R)

....d

1398 ACKNOWLEDGEMENT We express appreciation to the following collaborators who participated beside the authors in the study: Mr. S. L. Ali, Zentrallaboratorium Deutscher Apotheker, Eschbom Mr. B. Fetterroll, Chemische Landesuntersuchungsanstalt Karlsmhe Mrs. Hanh, Staatliches Chemisches Untersuchungsamt Braunschweig Mr. Ch. Hemmerling, Veterin~ir-und Lebensmitteluntersuchungsamt Frankfurt/Oder Mr. G. Kempe, Landesuntersuchungsanstalt ftir das Gesundheits- und Veteriniirwesen Sachsen Prof. Dr. A. Kettrup, Institut fur Okologische Chemie, GSF Oberschlei6heim Mr. R. Kruse, Staaliches Veterin~untersuchungsamt fiir Fische und Fischwaren Cuxhaven Mrs. K. Kypke-Hutter, Chemisches Landesuntersuchungsamt Freiburg Mr. G. Lach, Handels- und Umweltschutzlaboratorium Dr. Wiertz-Eggert-Dr. J6rissen, Hamburg Mr. Langst~idtler, Landesuntersuchungsamt fiir das Gesundheitswesen Nordbayem Mr. H.-A. Meemken, Chemisches Landesuntersuchungsamt NRW, Miinster Mr. G. Rimkus, Lebensmittel- und Veterin~iruntersuchungsamt Neumiinster Mr. R. D. Weeren, Dr. Specht & Partner Chemische Laboratorien GmbH, Hamburg

REFERENCES [1] [2] [3] [4] [5] [6] [7]

Voldner EC, Li YF (1993) Chemosphere 27:2073-2078 Saleh MA (1991) Rev Environ Contam Toxicol 118:1-85 DeBoer J, Wester PG (1993) Chemosphere 27:1879-1890 Stern GA, Muir DCG, Ford CA et al (1992) Environ Sci Technol 26:1838-1840 Van der Valk F, Wester PG (1991) Chemosphere 22:57--66 Alder L, Vieth B (1996) Fresenius J Anal Chem 354:81-92 Manual of Pesticide Residue Analysis, Vol. I (1987); Thier H-P, Zeumer H (Eds), VCH, Weinheim, pp. 383--400 [8] Specht W, Tillkes M (1980) Fresenius Z Anal Chem 301:300-307 [9] European Committee for Standardization (CEN) prEN 1528, 1994 [10] AOAC (1989) J. Assoc. Off. Anal. Chem. 72:694-704 [11] IUPAC: Analytical, applied and clinical chemistry divisions; Interdivisional working party for harmonization of quality assurance schemes for analytical laboratories (1990): Harmonized protocol for the adoption of standardized analytical methods and for the presentation of their performance characteristics. Pure Appl. Chem. 62, 149-162