Document 2j4DO8rjawVBrQVD7rDMBa54L

k f the thyroid. lorodibenuv/v 11. Z5, Pocembcr, tin University dde llliyltncad. Set., Tnrlt ipcr Congress, R. 11. (1970). aphthNlrne hi uelenr power* hylenc glycol. dot and Their ilueNtion, and combination t. Ju.xlcol. 8, gem: On the. .tKaj'.inn films i mis Maiden, -lonch. Jtt, .Ishycxirarts !----- ------------------------. , From the v/ ' 1 ELMKK P. \VIIKj; Fd Cotmet. Toxicol. Vol. 9, pp. 81-90. Pcrgamon Press J97I. Printed in Great Britain Problems and Progress hi Analytical Methods* II. ECiAN laboratory ofthe Government Chemist, Cornwall House, Stamford Street, London, S.E.I, England (Received J July 1970) Of all clean birds ye shall cat. But these... ye shall not cat: the eagle, and the ossifage, and the ospray,... (Deuteronomy 14: JI-12) In considering the general analytical background to trace contaminants in foods I pro pose first to review the analytical principles concerned .and then briefly to outline some of the particular areas of interest in this field and the progress being made in solving, them. In the context.of the present meeting, contaminants arc trace components. All trace analyses arc characterized by the parameters of sensitivity and specificity, both of which broadly follow the indications and requirements of the toxicologist, though exceptions can be cited for each. Thus in DDT residue analysis, for example, it is now conventional to report separately p,//-DDT, />./i'*l)l)h and (if present) />,/>'-TDK and often to sum these residues as `total equivalent DDT', even though the main-toxicological consideration, for Homo sapiens if not for Haliaectus leucccephalus, relates to /\/>-DDI or to k Jjim-.il DDJ con taining little or no DDE. This has been discussed elsewhere by Egan (1970) in relation to pcsticidc-rcsiduc levels in the total diet. In looking at the problems and progicss 1 shall consider particularly selected areas of current interest, namely nitrates, nitrites and nitro* samincs, polychloiobiphcnyls and trace metals. In dealing with the general principles I shall consider in particular pcsticidc-rcsiduc analysis, although virtually every consideration in this field applies equally to other fields of trace analysis. Other aspects of nitrosnmines, nitrites and nitrates, lead, mercury and other metals and pesticides will be dealt with by latci cohtribulois. I shall not consider polynuclear nioinalic liydiocmbons, myeotoxins or technological contaminants arising incidentally from the treatment of food during, processing and distribution, such as solvent or fumigant residues, or traces of surfactants, stcrilants or disinfectants. General principles Problems nnd advances in pcsticidc-rcsiduc methodology have been icvicwcd recently by Gunther (1969). hooking, to begin with, at the general principles, we can accept for the moment that the first of these is that the sensitivity of the method used should match Presented at a Sjmposinm on "Chemical Contaminants in foods- Jh/aul or Not?", n'omoici] by the 1-oocl and Print Directorate, Department of National Health and Welfare, Ottawa, Ontario, nnd held in Ottawa on J8-19 June 1970. roonS/l-r HONS 065300 12 It. EGAN the requirements of the toxicological considerations. The general pattern in fact has been for the development of methods sensitive either to about 0-1 ppm or, in more recent years and where the circumstance called for greater sensitivity, to about 01 ppb*, give or take a power of ten either way in each case. The basic operations, illustrated here by pesticide residue analysis, arc common to virtually all methods of trace analysis, although some elements of the sequence may be varied. These operations arc sampling, laboratory subsampling, extraction of the residue, clean-up of the extract from co-cxtraclcd interfering substances, and detection, preliminary identification and estimation of the residue in the clcancd-up extract, l inally, where the identity of the residue is not known in advance of the analysis, the preliminary identification is confirmed using one or more independent methods of analytical approach. The latter is often a distinct operation which may be difficult for residues present in foods at levels ofl ppm and is frcqucutljexccedinglydiflioult at a level of l ppb. Apai l from not always being easy, it is seldom cheap, in terms either of manpower or of equipment, to obtain .an unambiguous positive identification at n level of \ ppm. It is also necessary to have a clear understanding of the recovery factor, which some times assumes major importance when a residue method has to be used near its limit of detection. '1 he exact application of general principles of residue analysis will vary according to individual circumstances; on whether, for example, the analysis rehues to the evaluation of field research, to routine regulatory screening or to the enforcement of legal requirements. With the development of the so-called `multi-detection' or 'multi-residue' techniques based on chromatographic systems for pesticide-residue control, a new difficulty ofcollabornlive study has been recognized in an area in which it becomes necessary to define an unusually large amount of subjective experimental detail. Collaborative study of such methods has proved to be possible in limited areas. However, a new feature, which in part has replaced the collaborative study, is the development of practical guidance manuals, and it has been pleasing to sec n number of these developed for pesticide-residue analysis in recent years. ] hose produced in the USA by the Tood and Drug Administration (1967) and the Depart ment of Health, liduention, and Welfare (Burchfield Sc Johnson,' 1965) have now been followed in Canada by the l'ood and Drug Directorate's manual, edited by McLeod, Wales, Graham, Osademik Sc Woman (1969). While these compendia do not necessarily set out to provide full information on all pesticides, most of them give comprehensive analytical guidance on a systematic basis. This is amply illustrated by the Canadian manual, which is designed to meet not only the strictly legal position in relation to established toler ances but also to cover residues arising from any established pesticide usage in Canada. This is done by (low diagrams illustrating the sequence of screening and determination procedures mu! supplementing a fully descriptive experimental text. The emphasis is practi cal throughout and bears particularly on broad multi-residue techniques for detection and estimation. Such manuals represent n substantial contribution to the advance of residue methodology and are not inappropriate to the position in Britain, where enforcement is basically at local government level and where central government provides a Pesticide Residue Analysis Information Service. Chromatographic separation systems arc frequently used not only for clean-up and esti mation but also for ptcliniinary identification of trace constituents. This is usually accept able where the nature of the residue is known in advance, but there are distinct limitations in the extent to which dilfcrcnt chromatographic methods should be regarded as giving *b 10* throughout this paper. MUNS 085389 { i ' j I ; I , \ t J ! ` * ! * | ( indepen aspect a thin-lay rctcntkr snsu-pti tions a) biologic the coni of mikt SCIIMliVi ties ma; amotmi considv stigrexU general mass nr up to 1< inai her, ier-olui: orgaw-t final isi prove p ion fr:r spectrin loss of. ion cut applied Unci umhrst cmim i : the mo1 could 1* the mi Ot ganiv issue re (I gun. phoiu-.chloiiin product of SlICO Of till'.' hound ( (1970) ; l'oljdil As it anal) t'i act hits been rcocut years :ive or tnkc n by pesticide hough some -oratory mibd interfering .'.siduc in (he j advance of independent hich may be ingly difficult mik'either of at n level of xvliich somcr its limit or ry according tc evaluation equiicment.s, ' techniques ofcollahoraan unusually methods has has replaced 1 it has been recent years, the Departc now been ;7 Me I,rod, t necessarily mprehensive lian manual, dished loler' in Ciinadn. 'termination Isis is piactietcction and e of residue OKvment is a IVsticidc *p and rstially accept* limitations I ns giving ( ) PROBLEMS AND PROGRESS IN ANALYTICAL METHODS 83 independent support one 10 another. Elgar (1967) has given particular consideration to this aspect and has suggested four different parameters--gas-chromatographic retention volume, thin-layer chromatographic JiF value (or solvent partition p-vnluc), gas-chromatographic retention lime of a derivative formed hy chemical reaction, and pcsticidnl activity against a susceptible organism- -as alternatives for dealing with pesticide residues. Similar considera tions apply to other types of unknown residues, although there is an obvious limitation in biological response where carcinogens .arc concerned. Not all methods arc of equal value in the confirmation of identity, however. Completely unequivocal confirmation of (he identity of unknown residues is possible by high resolution mass spectroscopy. This is a very sensitive as well as a very specific method of examination, but some purely practical difficul ties may be encountered when this technique is applied to microgram (or sub-microgratn) amounts, for example in conjunction with gas chromatography. It is also necessary to consider the way in which the data should be handled. Widmark (1969), for example, has suggested focusing the spectrometer at certain mass numbers and using it in this way as a general gns-chromatogrAphic detector, if possible with alternation between two or more mass numbers to pin-point specificity. Tcu selected mass numbers can be scanned at a rale of up to 10 swccps/scc. Alternatively the whole mass spectrum can be scanned, using a mass marker, and recorded repetitively. Lovins (1969) lias described the application of highresolution mass spectrometry to the identification of individual organophosphorus and organochlorinc compounds in mixtures. This method cannot at present distinguish struc tural isomers, such as dicldrin aiul endrin or p.p'-DDT and o./f-DPT, but even this may prove possible if in addition an accurate measure of the relative intensities of the various ion fragments is possible. Uiros (1970) has improved the sensitivity available from a mass spcctiomctt ie detector by making it possible, to scan the column effluents rapidly without l he loss of accuracy and reproducibility normally associated with the rapid recording of small ion currents; this is done by using a small laboratory time-averaging computer, lie has applied this system to DDT- and d'cldrin-reskhic analysis for qurmlitics down to I /<g. Unequivocal knowledge of the identity of an 'unknown* residue is necessary for a full understanding of its significance before a principle of legal or health importance can ho enunciated or a decision relating to an established principle made. The need for this is all the more apparent in the light of possible alternative (and misleading) constructions which could be put on the results of the simpler residue-analysis techniques. 'I he limitations of the original pesticide-residue methods some 20 yems ago. based for example on tote) organically-bound chlorine, wore well recognized, allhough it is still necessary today to issue rcmimlm from lime to time th.nl nou-spccilic procedures should be used with camion (I'.gan, 1970). At the same time, general methods continue to have some use. Organopbosphorns-pcsticide residues arc in general much less stable than those of the persistent mganochlorine compounds and their detection, together with the detection of their breakdown products, presents a different and in some ways more difficult problem. There arc a number of successful chromatographic approaches to this, but it is dearly unnecessary to use any of these ifthc substrate of intciest enn readily be shown to contain little or no organically bound (or total) phosphorus. This approach was used hy Abbott, Crisp, Tarrant & '1 alien (1970) in total-diet studies in Iiritain. It is also of course applicable to mercury. Volychlorobiphenyls As recognized by McCully & McKinley (196-1), some of the early chromatoeraphie analytical systems failed to distinguish adequately between dicldrin ami p,;/-J)Dl\ Inter- MONS 085390 I) ) jhALiAk* > 94 H. LOAN fcrence in organochlorinc-residiic analysis by compounds of the polychlorobiphcnyl (PCB) I)pe was recognized by Koburn in 1965. Later Jensen (1966) found interference in pike to be due to these inntciinls, and similar findings were reported by Holmes, Simmons Sc Tntton (1967) in predatory terrestrial birds and by Iloldcn tfcMarsdcn (1967) in seals and porpoises. The PCB compounds have been widely used in industry, for example as Dansformer oils and as plasticizers for paints, lacquers and resins. For this reason they may present a special problem to the residue analyst; traces can derive even from simple packag ing materials such as cardboard, as was recently pointed out by Bailey, Bunyan & Fislnvick (1970). These authors were examining samples of ground cashew nuts, one of which appeared to contain about 10 ppm PCB; the same sample was found on closer examination to be the only one packed in a lacquered cardboard box. PCB compounds arc technical mixtures which contain many similar polychlorinated compounds. The chromatograms show a pattern of multiple peaks, a number of which may also correspond to those given by sonic of the commonly occurring organochlorinc-pcstitide residues, as was shown by Reynolds (1969). Such considerations led Kisebrough, Reichc & Olcotl (1969), for example, to conclude that some measurements of DDT resi dues reported earlier were too high. For a given PCB composition, the relative intensities of the peaks in the pattern may be substantially different. This problem can be tackled by wpaiating PCBs (along with /;,//-DDP) from pesticide residues using a silica-gel column. To a first approximation the total PCB detector response as measured by the aiea under the chart peaks can Ik taken as the same as that for a similar amount of/>.//-DDF, and results can Ik expressed in this way. Kocnian, Ten Noever dc Brmiw & dc Vos (1969) based semiquantitative PCB measurements on the height comparison with the peak rt -- 1*45 using n paiticular commercial compound as the standard. Clearly it is possible only to estimate the order of the PCB residue, or of that PCB mixture which, after weathering or other manner of degradation, is present. It is nevertheless desirable to consider uniformity in the reporting procedure for PCB residues and this in fact has been the subject of recent proposals by the Organization for Fconomic Co-operation and Development. Trace metals Arsenic and lead have long been the subject of concern in relation to food, in which residues may arise both from natural and other scnuccs. Both elements have been used in crop prelection for over 100 years and the need to limit trace levels in materials used in the food industry was recognized more than 50 years ago. The general residue position has improved considerably in recent decades but a fuller understanding of the significance of (he persistence of residues of some of the synthetic insecticides has stimulated a need to be aware of the possible build-up of inorganic residues where these materials arc in regular use for pesticide, herbicide or defoliant purposes. The interest has extended to some other elements such as copper and mercury; and the fact that some or these now feature in organometallic combination has also meant that part of the interest may relate to the form in which the residue is present. Analytical chemistry has advanced considerably since the turn of the ccntmy, and while it would be wrong to dismiss lightly the ability of analysts in the nineteenth century, the practical application of trace analysis to every day need lias improved enormously. It is perhaps worth remembering that, writing in 1819, Accum gave simple directions for the detection of 1 ppm lead in water. Today the method of choice for arsenic is based on the spcctrophotomctric procedure of Winkler (1962) which uses the reaction ! | j ! ' - > ; : 1 ; , - J j J I , j j > ; 1 < j ! I { , j ; HONS 085391 between Bode Si ncccssar is siiilah by Mori on litterfeeds ha. found lv pig liver rcmaiiii. Scp.ir.Ui; widely organic. Trace Of tills I copoei.i ns a rc:i the lint! chloi-al' use of r in indiv poimJs cbroi.iat of org.r been dc order * dcslrnct Widuun less sea illnlil;, t and tn< based o di(hi/'<ui by shak more >e graphic Si I alto I'rovi' spcclroj a numb quired. range o .medium applied here llv; inf pin} to a pa; hcnyl (PCI!) ce In pike lo Simmons A i in scats awl pie as trails-. >n they may nplc pnekagi A TUhwick irh appeared ion to be the vchlormntcd f which may loiinc-pcstiKiscbrour.il, T DDT' rcsit intensities ? tackled by pel column, a undo the . and results based semi- M5 mini', lo estimate 11* or other rmity in the it proposals I, in which .Tii used in used in the .'Mtion has lilinmrr of need to tie in regular 'omc other in orpartohe fotm in ee the turn lysis in Die improved ave simple for aiscnic >e reaction f) PJIODLEMS AND PROGRESS IN ANALV1ICAL METHODS B5 between arsine and the silver salt of diethyl dithiocarbamatc. As shown more recently by l)odc A Hncbmnnn (1968), a silver salt concentration of at least 20/g/J0mI pyridine is necessary, the red complex formed being arsenic lri(diethvTdilhiocarbamatc). The method is suitable for total nTscnic estimation and has been used for the measurement or residues by Morrison (1969), who concluded that the levels in poultry tissues, soils and crops grown on litter-treated soils were unaffected by the use of litter from poultry' houses where arsenical feeds had been used. Peden (1970), in a direct survey in Britain of 120 pig and other lively found less than 0*1 ppm arsenic in all of the lamb, ox and calf livers and some 85% of Die pig livers examined, with up to 1-0ppm arsenic in about 10% and over 10 ppm in the remaining 5% of the pig livers. Some work has been done on chromatographic methods for separating organically-bound from inorganic arsenic but such methods have not been widely employed. Unlike the position with mercury, the general indication seems to be that organically-bound arsenic is the less toxic, at least to avian and mammalian species. Trace mercury analysis has also long been a subject of interest in view ol the toxic nature of this metal, although, somewhat surprisingly, it has in former times featured in pharma copoeia preparations. The interest in mercury has increased sharply in recent years, initially ns a result of industrial use of orgnnomcicurial compounds ns fungicides in articulnnc and the timber industry, and more recently by the recognition of its long-established use in the chlor-alkali industry (Somers & Smith, 1971). Serious local problems have arisen from the use of orgtmomercuvials for wood-pulp manufacture in Sweden and from their preseme in industrial-manufacturing diluents in Japan. The high toxicity of mclhylmercury com pounds, in particular, has encouraged the development of analytical methods based on the chromatographic separation and identification of individual organomcrcuriaU (or gioups of organomcrcuriaU), and both tlun-laycr and gas-liquid chromatographic systems haw been described. Atomic absorption spectroscopy can also be used, with a scnsitiv itv of the order of J0/ig./ml or better. Cross checks by neutron activation analysis, a useful non destructive method available, for a number of additional trace elements, arc also possible Widmark (1968) has investigated the use of mass spectroscope. A chtomatographic approach less sensitive than mass spcctiotnetry blit nevertheless of considerable value because of us ability to distinguish the various forms of organically-bound mercury one from anuihei and from inorganic mercury has been developed by Tatton A WaysinBc (1969). This i* based on the separation of organomcrcnrial diihizonates and is sensitive to 2/<g. The dithi/onates, which arc self-indicating on thin-layer cluomaloplates, are. formed simplv by shaking the extracted organomcrcurials in chloroform solution with dithi/one. An even more sensitive procedure, down to I pg or less of mercury, is based on r.w.-chromuio graphic separation of the oiganomcrcuria) dilhi/onalo>. Recent total diet studies by Abbott A Tattoo (1970) in Britain show mcrcuiy levels of krs< than 0 02 ppm, as in Canada. Provided the interest relates to the total residue. X-ray spectrometry, ultraviolet emission spcclrography and atomic absorption spectrometry are all available for lead, copper and a number of other metals, the method of choice depending mainlv on the sensitivity re quired. X-ray spectrometry is the least sensitive of the three methods but covets a wide range of elements, from sodium lo uranium; the creates*, sensitivity is for lead and the medium-heavy metals such ns chromium, copper, iron, manganese and zinc It can he applied directly to foods, for example to orange-juice concentrate for trace tin cMimaiion; hoc lire lower limit of detection is about 10 me 'kg, prov iding a method suitable for scieeiv ing put poses. It can also be applied lo the ash, either diteetlv oi in solution impregnated on (o a paper disc. Surface-layer examination sensitivities are of the order of a few intcroeuun HONS 085392 I1 AAL 86 II. tCAN for copper, iron and zinc. Ultraviolet emission spcctrography, in which the sample material is subjected to an electric arc discharge and the emitted ultraviolet radiation is detected as A series of spectra) lines on a photographic plate, can be applied to the estimation in food of all heavy metal residues with a limit of sensitivity of about 1 pg, Much of this work is now done by atomic absorption spectrometry, although the technique is still useful when several metals arc simultaneously of Interest. Atomic absorption spectrometry is very sensitive and is particularly suited to trace metal determinations in food. Ihc sample is first ashed in order to prepare (he necessary solution. Representative limits of detection are About 0 01 /ip/ml for manganese, OOilpg/inl for beryllium, cadmium and zinc, Q-2/ig/ml for copper, chromium and lead and 1-2 /ig/ml for aluminium, molybdenum and vanadium, the limitation usually being the amount of other salts present in the spray solution. Even greater sensitivities can he achieved if the element of interest is first selectively extracted from the ash solution, for example by chelation. Under these conditions concentration limits of 0-25 /ig/ml for tin, about 0-0U4 pg/ml for chromium, nickel and lead and about 0-0005 pg/ml for cadmium and copper arc possible. Incidentally, Gorsuch (1970) has recently produced an excellent book on the destruction of organic mallet, in which he deals individually with botli wet oxidation ami the ashing process for a wide range of metals and other elements. Cia.s- liquid chromatography also ofieis useful possibilities if suitable volatile organomclnllic derivatives can be prepared, foreman, Gough Sc Walker (1970) for example, have used beryllium /5-diketoncs such as beryllium 1,1,1-triP.uoroacetylacetonc to determine this metal at levels down to I np/ml in urine. The complex, which is thermally stable, is cstimated using an electron capture detector. The possibility of using dicyclopenladicnyltitanium chloride in a similar way has also been investigated and it should be possible to extern! the technique to other metals of interest such as chromium. Nitrates, nitrites and nitrosnimnes Several different procedures arc available for the estimation of traces of nitrite. The Gricssllosvay method for nitrite, based on dinzolizntion of sulphauilic acid and coupling of the product with 1-naphlhyktmine, has been modilied by Kamm, McKcown Sc Smith (1965) and subjected to close study by Kamm, Bray and Collin (I96S), who found it to give results which were statistically consistent amt precise for various food substrates. A similar modi fication has been extended to nitrate in feeds by Schnll & Matcher (1969) using A'-(Inaphlhylclhylcne diamine) after reduction to nitrite, and an automated version of this technique for water analysis has been described by Henriksen (1965). Since l-nnphthylaminc presents a carcinogenic hazard, the basic method lias been further modified by Crosby (1967), who used Ihc non-carcinogcnic l-napluhyhmine-7-sulphonic acid (CJcvc's acid) instead. Waters (1964), in a critical review oflhc direct methods available for the determinalion of nitrate in water, considered that no one method stood out as pre-eminent, a view also reached by Bunion & Crosby (1969), who extended the study to include ion-spccific electrode methods. The latter authors concluded that the ion-spccific electrode method gives more consistently satisfactory recoveries than the other methods. The diversity of proccdnres for nitrate is also reflected by the variety of official methods. Those in Britain include reduction to ammonia (titration) and nitration of xylcnol (colorimetric). T he American method, which depends on the formation of brucine nitrate (colorimetric), is identical with (hat recommended by the Wot Id Health Organization, but is subject to some interference by chloride; n second tentative method is based on ultraviolet spectrophotometry. In com- MQNS 085393 | ' \ ; f ; j ' ; I ! . : ; ! 1 ( . I j | j ! I j ; | ; . j I , I I paring th electrode Ue deter emiti u| c Nituit.and ns t. from tl.e pO'ithui al:n.>\( <<. analytic..' cUaiKtp partition, (.! on? in v inures!, iaeluito \ a:\! clc. sir.ue (?;* for the l r\t>;ul. i a r.-iisf.-. `1 l:i`. at. the . I ha r bo p:u;\ !>i. . further : noco.'-ar.. slr.tlc-i. v *.-p.:r.t\tV-uil'..1 1 ivy ; ` by ulita. the saw eVorid.- , ritio-av. cd spot : rori.ull. may bo : Ic'I.il .V wave at . i; :l!n \\ . (1970), u samMio . niirwu;. pie mnfcrial , detected ns (ion in food (hit work is useful when ctry is very ie sample is (election arc c, 0-2 pti/ml I vanadium, utfon. liven ly extracted mcculrntion t and about deMiuclion I the ashing tile organo* ample, have (ermine this able, is esti:tcnlndicnyl> possible to 1 tie Ciriesspjiog of the milh (1966) ' give results milar modiusinp. A(lMOii of this ihthylaminc i by Crosby leve's acid) dclcrmiiianent, n view ion-specific. >cthod gives ty of piece* Main include . Ameiican icntirnl with interference try. In com- U ) PROBLEMS AND PROGRESS IN ANALYTICAL METHODS 87 paring the various methods, Bunton & Crosby (1969) concluded that the ion-spccific electrode method is rapid and relatively inexpensive and can be used with confidence for the determination of nitrate in both natural and polluted waters, particularly in routine control examination for which the highest degree of accuracy may not be required. Nitrate is an example of a pollutant which is partly of natural origin and partly otherwise; and as far as run-off nitrate is concerned, as much may be of natural origin as is derived from the use of uitratc fertilizer, even in arable farmland areas. It is less easy to state the position regarding the A'-nitrosamines, although it is clear that where they occur this is almost certainly as the result of interaction between free secondary amine and nitrite. ~l he analytical approach to traces of nitrosnmincs is basically classical and starts with extraction, clcan-up of the extract by steam distillation, ion-exchange chromatography and solvent partition, followed by separation, detection, estimation and provisional identification by gas-liquid chromatography or polarography. This is a difficult area of trace analysis: it is one in which from 1 to 10 ppb of specific but not easily identifiable substances arc of interest, when present in a very large excess of complex organic substrate which itself may include very many other largely unrelated trace as well as principal constituents. Extraction and clean-up processes in general call for individual consideration, according to the sub strate of interest. Dichloiomcthnnc, and to a lesser extent acetone ami acetonitrile are used for the extraction of volatile niirosnmincs, but may give some difficulty wth plant tissue; extracts must be carefully dried before bulk is reduced by evaporation. Steam distillation is a satisfactory clcan-up technique but will obviously not apply to non-volatile nitrosaminc. This also raises the question as to which individual compounds arc of interest and how far the interest extends to non-volatile nitrosamines. The unequivocal confirmation of the identity of the material estimated is particularly difficult. Polarography lacks specificity, although, like thin-layer chromatography, it can be particularly useful in a negative way. The use of two different gas liquid chromato graphic columns giving concordant results should be required; it might then be po-Mbh further to identify the peaks by un independent technique, such as the use of two dillvrem detectors. Tor unequivocal identification, however, detection by mass spectrometry is necessary. Several of the earlier reports of the detection of nitrosaminc in biological sub strates were based on Die evidence of thin-layer chromatographic methods in winch the separated chromatographic spots were rendered visible by a spray reagent. 1 lie methods described by Preussmnnn (1964) and by Preussmann, Neuralh, Wulf-Lorentzen, Daibci & Hengy (1964) have been widely used. In the first of these the compounds arc decomposed by ultraviolet irradiation and sprayed with sulphanilic ncid/i-nnphthylanimc solution; in the second the spots are sprayed with aqueous cthnnolic diphcnylamim: and palladium chloride, and then irradiated with ultraviolet light. Neither reagent is entirely specific for niirosnmincs; both however arc very useful in the negative sense that the absence of a colour ed spot is a pood indication of the absence of nitrosaminc. Spcctrophotomctric methods are normally limited to pure solutions and give a measure of the A'-ni(roM compounds ulmh may be present. Polarography, used by Heath & Jarvis (1955), also gives a measure of the total A'-nitroso compounds that may be present and is non-specific, although the reduction wave at about - 0-9 v is given by all nitrosamines tested. It is useful as a rapid sciccning method and can be made somewhat more specific if, as shown by Wallers, Johnson i< Kay (1970), it is applied befoic and after ultraviolet photolysis, which decomposes the mlu>samine with the elimination of the polarographic wave. Photolysis rales differ for various nitrosamines, however. Dniber & Preussmann (1964) used this reaction ns the basis lor MQNS 0*5394 (I ) . *kW' b/.MM. A-*11Ul II II. EGAN nitrosnmine estimation, measuring spectropholomclricnlly the increase in extinction of Gricss-lloxvny rengent. Gas-!iquiil chromatography is at present sensitive to about 1 ppm niUosnminc but current work is designed to improve this, for example by using better clcttn-up procedures (01 more concentrated sample extracts) or mote selective detectors. Nitrogen-specific detectors arc obviously of interest here. Sen (1970) has recently taken advantage of the great sensitivity of the electi on-capture detector to dimcthylnilraminc, the oxidation product of dimcthylnitrosnminc. Conclusion In the opening paragraph of this paper, I referred to the fact that while both sensitivity and specificity in trace analysis broadly follow the indications of toxicologists, this is not Invariably so, and I used OPT and DDK as on example of specificity. Whereas 15 20 years ago the need in pesticide residue analysis was for greater sensitivity, this has. broadly speak ing. now been adequately achieved, the main example being the persistent organochlorinc insect (cities. Indeed, there lias on occasion been criticism that some of the analytical methods now available are too sensible, and it is quite true, for example, ihnt with routine chromato graphic systems there is virtually no substrate in which 10 ppb DDI, if not 10/1012, cannot be detected. Hut it is possible that such orders of sensitivity may be required, or indeed may be inadequate, if the final decision on the safely of any chemical to which man is exposed is to l>c based on direct observation in man. This clearly is a subject which raises serious technical and ethical problems, as Magee (1970) has indicated. At the same lime the Sub committee on hnvironmcntal Improvement, Committee on Chemistry ami Public Affairs, of the American Chemical Society (1969) has indicated that the identification and measure ment of limiting nutrients in water would be greatly eased by improved analytical methods for low levels of phosphorus and various forms of nitrogen compounds. For the detection of some compounds and their metabolites there may in fact be a call for the development or even more sensitive anah tical methods than arc currently available. "Il wasn't the words which frightened (he birds, ltiil U>c hori ibk double entendre." (Longford Reed) . Abbnll, I). C,, Crisp, S.. Tarrant, K. R. & Tolton, J. 0`G. f 1970). Pesticide residues in the total diet in I:ii(;laiul and Wales, 1966-1967. Ill,--Oigatiophosphorus jKslicidc residues in the total diet. J'csiic. Sci. 1, 10. Abbott, D, C. ft Tntton, J.O'O. (1970). Pesticide residues in the total diet in England and Wales, 1966-1967. IV.-- Mercury content of the total diet. Pesiic. Sci, !, 99. Accum, P. (I8I9V Chemical Announcements, pp. 91-92. Thos. Boys, London. American Chemical Society (1969). "Cleaning Our Environment--The Chemical Basis for Action". ACS, Washington, D.C. Hailey, S,, Dimyati, P. J. ft Fishwick, F. D. (1970). Polychlorinated biphenyl residues. Chany bid. p. 705. Biros, I'. J. (1970). Pnlianceincnt of mass spectral data by means of a limcitvciasingcompuici. Analyt. Client. 42, 537. Bode, II. u. Ilaihmann, K. (196S). Ober die Rcaklion zwischcn Arsenunsserstofi" und Sillodintliyldilhio- cnrhamidal. /. anuiyt. (hem. 241, 18. Btmton, N. G. A Crosby, N. T. (I%9). The determination of nitrate in waters and effluents using specific ion cltcltodcs. i'roc. Sac. HVj/, Tietif. I.xtnn. 18, 338. lluichOckl, 11. P. A: Johnson, IV K, (1965). Cuttle to the. Analysis of Pesticide Residues. Vols I A II. U.S. Department of Health, l-dneation, and Welfaie, Washington, D.C. Crosby, N. T. (1967). The determination of nitrite in water using Clcvc's acid. Proc. Soc. II'at. Treat. Lxam, 16,51. MGNS 085395 Daiber, f) Vetb I'pan. tl. ; pO.-i.-r-; .stm:. : IVm. It I I, .ii. k , ( / l'Oid i a:,.i v. l oi,..,- and r.: (ioisrc\ i Gunt' I Ilea-;., |). ti-? Ho;n('... i An . Ho:.:.i I.','!. 11,0 A'.,-. . Jen- K.-'i.r .. i K;n'r--, I ih- h'r*-,,:.. and ; . Imr, . j /./( - i MC( l:' 's ' I Met, i it fn i\ i U;t.- ' ( M.t - o. p. \ Mi ne J J.iU I'c.lcii, .' | > />">'!. ; Ph-i.-.i - Aim,' Priii . ;i IVi.-m ' t Reyno'd,. 1 Hi.tf t Kisvhi, . . . n.ued ;. Rohuui, J. i: ! - in i Sclwll. h. J> J An r-t Sen. ,\. P i j at p c.', .. Someis, | . .. 7/. Q. lalii'ii. J i * < K-sV, 1-. > t W.ihi i s, i . i A.run * ,. Wains. | . W'iifm'-k.f; i rxIimMion of nboiil 1 |>pm utiiiR belter vc defecton;, ccnlly token Jtramme, the Ih sensitivity s, this is not 15-20 yenrs oruJIy RjH'iikanot hlorine icnl methods ic eh t omatoI0,,l cannot .1, oi indeed tn is exposed tises serious lire the Sub* blip AfVoiis, ul nu'itMtte* ml methods lie detection cvclopmcnl | total Oici in I. Pcstir, Sci. X I960 1967, rtkm'\ AC'S, * hid. p. 705. Unity/, ( hem. Mlltyldilhiousing MH-i'ific I A )]. u.s. heat, l..\um. PROBLEMS AND PROGRESS IN ANALYT.at METHODS I) Puiber, D. u. Prtuasmann, R. (1961). Quantitative colorimelrische Bestimmunc organische A'-Nitroso* Yerbinbungen dutch ptiotochcmischc Spalliing dev Nttfosaininbindung. Z. analyt. Chem. 206, 344. Egan, H. (1969). Pesticide Residues in Food: The Situation Today. Proceedings of an International Sym posium on Chemical and Toxicological Aspects of Environmental Quality, p. 105. G. Thieme Verlag, Stuttgart. npfltt, H. (1970). Pesticide assay. Chew, tn Britain 6, 39. Elgar, K. E. (1967). Oiganochlorinc pesticide residue analysis. Confirmation of identity. /. Ass. of. Analyt. Cheat. 50, 1069. , Food and Drug Administration (1967). Pesticide Analytical Manual. Department of Health, Education, and Welfare, Washington, D.C. Foreman, J. K., Gough, T. A. St Walker, E. A. (1970). The determination of traces of beryllium in human and rnt mine samples by gas chromatography. Analyst, Land. 95. 797. Got such, T. T. (1970). The Destruction of Organic Matter, l'crpamon Press, Oxford. Gunther, )'. A. (1969). Current status of pesticide residue methodology. Ann. S.)'. Acad. Sei. 160, 72. Heath, D. F. St Jarvis, J. A. E. (1955). The polarographic determination of dtmeihvlniuo>aminc in animal tissue. Analyst, Land. 80, 613. Ilcnrikscn, A. (1965). An automatic method for determining nitrate and niliiie in fresh and saline water*. Analyst, Ijtnd. 90, 83. Holden, A. V. A Marsden, K. (1967). Organochlorine pesticides in seal* and porpoises. Nature, tend. 216. 1274. Holmes, D. C., Simmons, J. H. A Tatton, J. O'G. (1967). Chloiinatcd hydrocarbons in British wildlife. Nature, Ix/nd. 216, 227. Jensen, S. (1966). Report of a new chemical hazard. A'ctr Seient. 32, 612. Knmm, I.., McKeown, G. G. A Smith, D. M. (1965). New colorimetric method for the determination of the nitrate and nitrile content of baby foods. J. Ass. off. agrie. Chem. 4S, $92. Kamm, 1.., Bray, D. F. A Coffin, D. Is. (1968). Assessment of the naphihyla/cmaphihylatnine method for (tic determination of nitrate and nitrile. J. Ass. off. analyt. Chem. 51, 140. Kocman, J. H., Ten Nocvci dc Drauw, M. C. A dc Vos, R. 11, (1969). Chlorinated biphenyl* in Ash, mussel!, ami birds from the River Rhine and the Netherlands coastal area. Nature, land. 72), 1126. Ixwins, R. F.. (I960). Identification of pesticides In mixtures by hirh-rooluiion mass spectrometry. J. agiie. Pd Chem. 17, 663. MeCnity. K. A. A McKinley, W. P. (196-1). Determination of chloiinatcd pesticide residues in f.it by electron capture gas chromatography. J. Ass. off. agrie. Chem. 47, 652. Met .end, H. A., Wales, R. A., Graham, R. A., Osadetnik, M. A Bluman, M. (1969). Analytical Metho d for Pesticide liesiducs in lootls. Food and Drug Directorate, Ministry of National Health and WcJf.uc, Ottawa, Ontario. Magee, I*. N. (1970). Toxicology and certainly. AVn- Seient. 46, 61. Morrison, J.(1969). Distribution of arsenic from poultry litter in broiler ducks, soil and crept. J. agiic. I'd Client. J7, 1288. Fcdeii, J. D. (1970). A survey of the arsenic, copper and lead contents of pig and other animal liver,. J. Ast. puUI. Analysts 8, J-1. Bivussmunn, R. (1961). 7um oxydativen Alihan von Nitrosaminen mil Fnaynvfteicn Modoll-Systcmen. Arrnrmiftel-1 m a lt. 14, 769. Breussmann, R., Ncurath, G.. Wulf-Lorvntzen, G., Daiber, D. u. Money, 11. (1961). Anfj-rKv.eth'dcn ut.d DUnnsc-lnchl-Chromatographic von orprmischcn A'-Nilt osovciivm1.inr.cn. Z. analyt. (Inr:. 202, IS7. Reynolds, L. M. (1969). Polychlorobiphonyls(BCB's)and their imcrfciencc with pesticide residue an;d>.-i . Bull. ene. contain, A Toxicol. (U.S.) 4, 128. Hischioiiph, H. W., Rcichc, B. A Olcolt, 11. S. (1969). Current psogress in the determination of polyrlilu::- nnlcd biphenyls. Pull. rue. contain. A Toxicol. {U.S.) -1, 197. Koburn, J. (1965). A simple conccntmtion-ccl) technique for determining snudl amounts of 1 slide ions m-J its use in t)>c determinulion of residues of org.tnoehlminc pesticides. Annc.-st. I.oul. 90. 467. Schall, K. 1). A Hatcher, 1). W., Sr. (I960). Colorimetric determination ol nitrate:, tint! nitrites in livd*. J. Ass, off, analyt. CVrrw. 51, 763. Sen, N. B. (1970). On*- liquid chromatographic determination of dimclhylnitrosamine ns riirctl:>lnitr.i'i.i''i: at picogram levels. J. Chromat. 51, 301. Somers, 1;. A Smith, D. M. (1971). Source and occurrence of environmental contaminant,. 1,1 Crura. Toxicol. 9, In press. Tatton, J. 0'C. A WagslafTc, B. J. (1969). Identification and determination of oreanomercuiial fungicide residues by thru-layer and ras chromaiogtaphy. J. C hr,,mat. 44, 2S4. Walters, <\ Johnson. 1-. M. A Ray, N. (I97U). Separation and detection of volatile and non-vo!a;.l.` A'-nitiosamincs. Analyst. Loud. 95. 485. Waters, W. F. (1961). The determination of nitrate in water. Pun. Sor. Hur. Tirnt. Turn:. 13. 793, Widmaik.G. (1968). Or gnnomercury fungicide residue ana fysis. 1 valuation. 7. Ass. off. analyt. Chon, 51, .*'w HONS 085396 90 II. EGAN Wlclmark, C. <1969). OrpnnooMnrinc compounds. Unambiguous identification of unknown residues. J. Am. off. nnalyt. Chon. 52, 30ft. Winkler, \V. O. (1962). Identification and estimation of the arsenic residue in livers of rats ingesting arsenical*. J. Ass. off. agric. Chou. 45, 80. I | . f ^ Cosm't. MQNS 085397 In nut-' County, I of accidci S|uum?!y. with liftUr fora lileli nldrin am cnnlly Ur. cmUc-isic JlMlilt an- In a si.\ dicKltiit(? imlicci wakhin, .wi associate' meat and blood am pesticide tosicity. '1 he cor peak com either was by the pn singly or i total DO fat. In fen the pcsti.'i 01,05.0* t J'HscntcJ si*omoi Oni.uio JMatimi; a.