ABSTRACT
Introduction
Taxonomy of the plant
Impurities of plant origin
The analysis of coca leaves
Clandestine extractive process and cocaine impurities
Stereochemistry
Presumptive tests
Sample preparation
Note on chromatographic methods
Conclusions
Acknowledgment
Author: H. L. SCHLESINGER
Pages: 63 to 78
Creation Date: 1985/01/01
Literature concerning the taxonomy and alkaloid content of the coca plant is examined in this article. The process used to extract cocaine alkaloids from the plant is described and information relating to the constituents of illicitly processed cocaine reported. Investigations into the stereochemistry of cocaine resulting from United States laws controlling such drugs are reviewed. Advances in presumptive tests for cocaine are described and some comments are made relative to trends in chromatographic analysis.
The continuing world-wide problem of cocaine abuse is reflected in a considerable body of scientific literature relating to the analysis of materials and to the development of more comprehensive schemes for identifying cocaine in mixtures. This article describes some of the results of recent work of particular relevance to forensic scientists who examine seized materials. Toxicological aspects of cocaine chemistry are not dealt with. The emphasis is on work published since 1975, but earlier publications are considered where it is desirable to provide background material on a particular point. The findings presented in this article supplement those of earlier in-depth studies [ 1] [ 2] [ 3] . [ 4]
Impurities in cocaine may arise from constituents of the plant, from the manufacturing process or from adulteration. The taxonomy of the plant is of interest because plants with distinctive physical characteristics may also possess distinctive alkaloid contents. Unfortunately, the taxonomy of the coca plant is confusing to the non-specialist. The literature, especially the older literature, contains a bewildering array of names of species and varieties of plants whose leaves contain significant levels of cocaine. The subject is complex even to specialists. For example, it has been reported tha "many of the species of Erythroxylum are difficult to distinguish owing to their small, inconspicuous flowers, a lack of well defined taxonomic characters and the great variability observed in certain characters. One of the greatest obstacles to identifying the American species is their sheer number" [ 5] . Even among experts, there is a division of opinion as to whether only one species ( Erythroxylum Coca Lam.) [ 6] [ 7] or two closely related species (E. Coca and E. Novogranatense (Morris) Hieron.) [ 5] [ 8] are cultivated for their cocaine content. The more recent investigations favour the latter view, upon which the following discussion is based.
E. Coca is a plant "native to the Andean montana, a region of the eastern Andes between about 500 and 1,500 meters elevation, consisting of wet, tropical montane forests ... The leaves are characteristically large and thick, broadly elliptic in shape, more or less pointed at the apex and dark green in colour" [ 5] . The plant has been under cultivation for a long period of time and truly wild specimens of E. Coca are now rare. Hybridization between cultivated E. Coca and other species of Erythroxylum is common [ 5] . Machado identifies four cultivars of E. Coca, Among distinguishing features of the leaves of these plants are differences in size, shape, colour and texture. He also reports differences in cocaine content in these varieties [ 8] .
Plowman describes a variety of E. Coca ( E. Coca var. ipadu) that is found only under cultivation in the Amazon valley of Brazil, Colombia and Peru. This variety is distinguished by containing significantly lower levels of cocaine than plants of E. Coca itself. The leaves are usually broadly elliptic and rounded at the apex [ 9] .
E. Novogranatense is principally found today in the more mountainous regions of Colombia, but was at one time much more widely distributed, and plants of this species were cultivated commercially for a time in Java. The plant is similar in many ways to E. Coca, but has narrower, thinner, bright yellow leaves which are usually rounded at the apex. E. Novogranatense is better adapted to growth in hot, dry climates than E. Coca [ 5] .
The names "Trujillo coca" "small-leaved coca" or "Peruvian coca" are applied to a distinctive variety (E. Novogranatense var. truxillense) grown on the desert coast in the region near Trujillo, Peru. Plants of this variety have leaves that resemble those of plants classified as E. Novogranatense, but may lack the lines parallel to the central vein that are found in many specimens of Erythroxylum species. It is the decocainized leaves of this plant that are the source of a widely used beverage flavouring. Plants cultivated for use as flavouring agents have higher levels of essential oils than are found in other plants of this variety [ 10] .
Fifteen other species of the genus Erythroxylum have been found to contain cocaine [ 5] [ 11] although at levels much below those reported for E. Coca and E. Novogranatense.
The alkaloids cis- and trans-cinnamoylcocaine, methylecgonine, tropa-cocaine, pseudotropeine, truxilline, hygrine, cuscohygrine, and nicotine have been identified in extracts of coca leaves [ 7] [ 12] [ 13] [ 14] . Not all of these compounds have been found in all varieties. Hegnauer reports that differences exist in the relative amounts of some alkaloids among coca leaves from certain varieties [ 7] . In particular, he finds that leaves from E. Coca tend to contain high levels of cocaine and cuscohygrine and lower levels of other alkaloids, while leaves of E. Novogranatense var. truxillense contain high levels of cocaine and cinnamoylcocaines and lower levels of other alkaloids [ 7] . More recently, cis-cinnamoylcocaine, trans-cinnamoylcocaine, and cocaine were quantified in a number of single leaves from E. Coca, E. Novogranatense, and E. Novogranatense var. truxillense [ 14] . An important finding is that, regardless of variety, younger leaves have higher levels of cinnamoylcocaine than older leaves. Ratios of the cis to trans isomers appear to be variable among leaves of a single type. The data also suggest that E. Coca may have lower levels of cinnamoylcocaine relative to cocaine than the other types analysed.
Rivier suggests that because a base was used in the extraction of hygrine alkaloids, a question exists as to whether these substances exist as such in the leaf or whether they are artefacts from the extractive procedure [ 14] . The alkaloid profile of a sample can be dramatically changed by the extractive process. Of the alkaloids that have been found in coca leaves, only the cinnamoylcocaines, ecgonine methyl ester, and cocaine itself are usually detected in processed cocaine. Levels of the cinnamoylcocaines and ecgonine can be affected by the manufacturing process. If potassium permanganate is used to purify cocaine, the cinnamoylcocaines will be oxidized and ecgonine methyl ester is a product of the reaction.
Proper preservation of the sample is especially important if quantitative analysis of leaf material is to be undertaken. It is reported that Indians who live in hot and humid environments, and who consume coca leaves, prepare fresh material daily because of the rapid loss of potency that occurs [ 9] under such conditions. It has been shown that sun-drying coca leaves is superior to alternative modes of preservation [ 15] . Dried coca leaves maintained under low humidity can retain detectable levels of the alkaloid for decades [ 11] . It has been observed in laboratory that undried leaves will decompose into a viscous mass containing little or no cocaine. A number of procedures have been employed for extracting alkaloids from leaves. Rivier examined six extractive procedures and reports that a short immersion in boiling ethanol ensures the quantitative extraction of ecgonine-type alkaloids and minimizes the breakdown of cocaine [ 14] . Extraction with boiling ethanol had been previously recommended [ 16] [ 17] [ 18] for the determination of cocaine and cinnamoylcocaines in leaves. Turner and others examined seven extractive procedures and also found refluxing with ethanol to be superior to alternatives in terms of total cocaine extracted and reproducibility [ 17] . Extraction with hot methanol has also been shown to be effective [ 19] .
Analysis of the alkaloids extracted from leaves has been accomplished by gas-liquid chromatography in combination with mass fragmentography and stable isotope dilution [ 16] , and by gas-liquid chromatography alone [ 17] [ 18] [ 19] . Quantification of cocaine and both cinnamoylcocaine isomers can be accomplished in a single procedure by mass fragmentography and gas-liquid chromatography methods [ 18] .
Reliable first-hand information regarding practices employed in the illicit production of cocaine is generally unavailable because of the clandestine nature of the work. The general outlines are given below [ 20] , but individual operators may vary:
Leaves are mixed with water and a material such as calcium carbonate or lime (calcium hydroxide) that will produce an alkaline reaction in the resulting pulp, and the mixture is crushed. Kerosene (or an equivalent hydrocarbon) is added and the mixture stirred;
The kerosene is recovered and the coca leaf pulp discarded. Acidified water is mixed with the kerosene, extracting alkaloids into the aqueous phase. The kerosene is removed. If coca paste is to be produced, the water is made basic with lime, ammonia, or some other alkaline substance, precipitating the more basic alkaloids. The precipitate, which often contains mixed inorganic salts as well as crude cocaine, is isolated and dried. Some illicit operators do not produce coca paste, but treat the acidified cocaine solution as described in the next step, beginning after the point where the coca paste is dissolved in dilute acid;
Coca paste is dissolved in dilute sulfuric acid. Potassium permanganate may be added at this stage until the solution remains pink. The solution is allowed to stand, then filtered. The filtrate is made basic with ammonia, precipitating cocaine base and other alkaloids. The precipitate is recovered by filtration, washed with water, and dried;
The crude cocaine base is dissolved in ethyl ether. The ether is filtered. Concentrated hydrochloric acid and acetone are added to the filtrate. The resulting precipitate of cocaine hydrochloride is collected by filtration and dried.
Since cinnamoylcocaine isomers are detected in many seized cocaine hydrochloride samples [ 21] [ 22] and permanganate destroys these com- pounds, potassium permanganate is not used by all operators, or at least not in quantities sufficient exhaustively to oxidize the material being processed. Besides the cinnamoylcocaine isomers, the following compounds have been identified in unadulterated illicitly produced cocaine: ecgonine, ecgonine methyl ester, benzoylecgonine, trans-methyl cinnamate, cis-methyl cinnamate, ecgonidine (anhydroecgonine) methyl ester, methyl benzoate, benzoic acid, acetone, ether, ethanol, and water [ 23] - [ 29] . The presence of ecgonidine in commercially produced coca paste has been reported [ 30] . Its absence in illicit cocaine samples may arise from differences in processing conditions. The presence of ecgonidine methyl ester in some seized cocaine samples [ 26] is of interest in this connection.
Caution is needed when interpreting results of analysis in cases where ecgonidine methyl ester is detected in a cocaine sample by gas-liquid chromatography or by gas-liquid chromatography mass spectrometry. Ecgonidine methyl ester can be formed in the injection port of a gas-liquid chromatography instrument from the decomposition of cocaine [ 28] .
Pseudococaine is reported to have been found as an artefact in an extract of coca leaves [ 31] and the analogous pseudoecgonine methyl ester is produced in a considerable yield by heating cocaine in alkaline solution [ 23] . Neither compound has been identified in illicitly produced cocaine.
Until its recent revision, United States law controlling cocaine did not explicitly mention the drug, but referred to "Coca leaves ... and any salt, compound, derivative, or preparation there of which is chemically equivalent or identical with any of these substances ..." [ 32] . Many states have adopted drug laws modelled after the federal law. Some courts ruled that under the law, only l-cocaine was controlled, because only this enantiomer occurs in coca leaves. As a result, forensic laboratories examining cocaine exhibits for trial in some jurisdictions were required to identify the enantiomer present in cocaine exhibits. Courts also requested forensic chemists to demonstrate that the analyses they had performed were not only capable of distinguishing between l-cocaine and d-cocaine, but also between cocaine and its diastereo-isomers. In 1976 the laboratories of the United States Drug Enforcement Administration adopted a policy requiring the identification of the enantiomer present in all samples containing cocaine. Many other forensic laboratories adopted similar policies. As a result, considerable attention was focused on research to develop methods for improving the efficiency of cocaine enantiomer identification as well as to develop or validate methods for distinguishing among the cocaine diastereoisomers. Some results of that research are discussed below.
Four pairs of enantiomers are predicted from the accepted structural formula of cocaine. Each member of a pair of enantiomers has a diastereoisomeric relationship to members of all the other pairs. All diastereoisomers have been synthesized [ 33] [ 34] and their configurations and conformations determined by proton and C-13 magnetic resonance spectro-metry [ 35] [ 36]
Siegel and Cormier described the preparation of d-pseudococaine, a diastereoisomer of cocaine, and demonstrated that it could be distinguished from l-cocaine on the basis of thin-layer chromatography, polarimetry, infra-red spectrophotometry, melting points, proton magnetic resonance spectrometry, the gold chloride microcrystalline test, gas-liquid chromatography and gas-liquid chromatography/mass spectrometry [ 37] . They found that distinctions made by gas-liquid chromatography and gas-liquid chromatography/mass spectrometry were more ambiguous than those made with the other techniques tested.
Three publications [ 38] [ 39] [ 40] describe methods for the separation of all four cocaine diastereoisomers by high-performance liquid chromatography. One publication [ 39] , also reports the results of examination of the diastereoisomers by gas-liquid chromatography and chemical ionization mass spectrometry. Gas-liquid chromatography separations are found to be complicated by the instability of two diastereoisomers (allococaine and allopseudococaine). These compounds exhibit peak broadening, apparently a result of decomposition on the column. It is also noted that allococaine decomposes on standing in chloroform solution. Decomposition of the two diastereoisomers also complicates attempted analyses by gas-liquid chromatography/Chemical ionization mass spectrometry, although the results obtained are of use in that they can be interpreted to identify the decomposition products of the diastereoisomers as benzoic acid and ecgonidine methyl ester. Direct-probe chemical ionization mass spectro-metry data show that the technique can be used to differentiate the diastereoisomers. A thorough characterization of cocaine enantiomers and diastereoisomers [ 40] , gives procedures applicable to forensic work for differentiating among diastereoisomers by infra-red spectrophotometry, proton magnetic resonance spectrometry and electron impact mass spectrometry. The differences observed among diastereoisomers when examined by electron impact mass spectrometry are small, but have been confirmed by more recent work using a high resolution instrument and mass-analysed kinetic energy spectrometry [ 41] . Allen and others [ 40] give separative procedures for diastereoisomers through thin-layer chromatography, gas-liquid chromatography and high-performance liquid chromatography. Methods for identifying cocaine enantiomers by crystal tests, infra-red spectrophotometry and mixed melting points are also presented. There is general agreement that discrimination among diastereoisomers by gas-liquid chromatography methods is not as unambiguous as other methods tested [ 37] [ 39] [ 40]
The crystal test suggested by Allen and others [ 40] involves addition of a known cocaine enantiomer to a sample and then an acidic gold chloride reagent is used to form characteristic crystals. If the cocaine enantiomer added is different from the one originally present, crystals characteristic of the racemate will form rather than those characteristic of the pure enantiomer. It is important when performing this test that the amount of cocaine added is at least roughly the same as the amount present in the sample. It may be difficult to achieve this when dealing with a highly adulterated sample. Should any difficulty arise in discriminating between crystal types, an infra-red spectrophotometry spectrum can be obtained that may facilitate the identification [ 42] .
It has been reported that (+)-di- para-toluoyl-(-)-tartaric acid reacts with cocaine to form an insoluble tartrate salt and that this can be used to separate cocaine from other alkaloids [ 43] . Since the tartrates of d- and l-cocaine are diastereoisomeric, an infra-red spectrophotometry spectrum obtained from the precipitate differentiates between l- and d-cocaine. The characteristic crystal habits of the cocaine tartrates themselves are the basis of another suggested crystal test [ 44] .
Bowen and Purdie report that identification of the enantiomer [ 45] and quantification of cocaine [ 46] can be accomplished by use of circular dichroism spectropolarimetry. Identification of the enantiomer is done by complexing cocaine with an optically active sugar derivative, inducing a structural dissymmetry in the drug molecule. Cocaine concentration is determined by measuring the ellipticity of the cocaine solution and comparing the result with a standard plot. For this procedure to provide accurate results, cocaine must be optically pure and free from optically active contaminants.
A method for enantiomer identification has been developed that involves adding pure l-cocaine to cocaine recovered from a developed thin-layer chromatography plate [ 47] and measuring the melting point of the mixture. If the material removed from the plate was not l-cocaine, the mixed melting point will differ from the melting point of l-cocaine alone. Another thin-layer chromatography procedure for identifying cocaine enantiomers [ 48] consists of hydrolyzing the sample to ecgonine, then re-esterifying with optically pure 2-octanol. The octanyl esters of ecgonine formed from l- and d-cocaine are diastereoisomeric and can be resolved chromatographically.
A procedure has been described for the simultaneous identification of cocaine and enantiomer determination by proton magnetic resonance spectrometry [ 49] . A spectrum of cocaine dissolved in an optically active solvent is obtained by proton magnetic resonance spectrometry. Inter-molecular forces between cocaine and the solvent will shift the resonance peaks of cocaine in a manner that can be used to identify the enantiomer present in the sample.
A new law passed in the United States in 1984 is more explicit than its predecessor and controls "cocaine, its salts, optical and geometric isomers, and salts of isomers" [ 50] . It is expected that in the near future isomer identifications will no longer be necessary. The effort expended in research regarding isomer identification will have had lasting value in improving the specificity of cocaine analyses and will remain useful in instances where it is suspected that a particular sample of cocaine might have been synthesized rather than extracted from the plant [ 51] . It is of interest in this connection that a stereospecific synthesis which produces racemic cocaine free from diastereoisomers [ 52] has been reported.
Since its introduction by Young, a 2 per cent cobalt thiocyanate solution has been a widely used reagent for the presumptive identification of cocaine [ 53] . Young recommends the use of a second solution containing stannous chloride, the purpose of which is to distinguish between cocaine and the common adulterant procaine. Young identified some compounds that give false positive results with the test, but in the hands ofan experienced forensic chemist who uses it in conjunction with other presumptive tests, the cobalt thiocyanate test is very useful for rapidly screening samples [ 54] . A more detailed history of the test has been described elsewhere [ 55] .
An elucidation of the reaction mechanisms involved has been published [ 56] . lt assigns the formula
(alkaloid H) 2 [Co(SCN) 4] to the relatively water-insoluble blue complexes formed in neutral-to-basic solutions, and the formula
[Co(alkaloid) 2] (SCN) 2 to the more water-soluble, brownish-red to pink complexes formed in acid-to-neutral solutions. Solubilities in water and chloroform are given for complexes formed with 29 alkaloids and nine metal thiocyanates.
The relative simplicity of the cobalt thiocyanate test made its use feasible outside the laboratory and created a demand for a variation with increased specificity. One group examined five variations of the cobalt thiocyanate reagent. Maximum specificity was obtained with one containing orthophosphoric acid. This reagent gives positive tests with methadone and a number of local anaesthetics [ 57] . Other variations designed to improve specificity were introduced [ 58] - [ 60] . The version known as the Scott test [ 60] is marketed commercially in the United States. It employs a cobalt thiocyanate reagent in a 1 : 1 water: glycerine solution that forms the characteristic blue precipitate with cocaine, followed by addition of concentrated hydrochloric acid to form a clear pink solution, then of chloroform which turns blue as the complexed cocaine is partitioned into the organic phase. Methadone does not give a positive test, nor do the local anaesthetics tested other than cocaine. However, phencyclidine and phenothiazine derivatives [ 61 ] - [ 63] , especially in combinations with other drugs, give positive reactions to this test. Use of a fourth reagent that contains ferric chloride in a perchloric acid and nitric acid solution is found to increase specificity [ 61] . Alternatively, specificity can be improved if compounds giving a positive reaction when subjected to the Scott test are tested with Marquis reagent [ 63] .
An interesting analogue of the cobalt thiocyanate test is described by Travnikoff, who has developed a complexing reagent composed of an aqueous phase containing a solution of copper sulphate and potassium thiocyanate and an organic phase consisting of chloroform [ 64] ; addition of cocaine causes the chloroform layer to turn brown. Cocaine purity can be roughly estimated by observing the depth of colour in the chloroform layer. A limited number of common drugs were examined. No interfering compounds were found.
Methyl benzoate is among the volatile components of illicitly produced cocaine samples and has a characteristic odour that is sometimes detected in such samples [ 29] . A test for cocaine in the older literature involves heating the sample in acid, then adding ethanolic sodium hydroxide to form ethyl benzoate [ 65] . A simpler procedure is described in which transesterification to methyl benzoate is accomplished at room temperature with a reagent consisting of sodium hydroxide in methanol [ 66] .
Household bleach (sodium hypochlorite solution) has been used for some time by drug-law violators as a presumptive test for cocaine as well as a test for the detection of some common adulterants. The specificity of the test has been studied and results compared with those obtained with the cobalt thiocyanate reagent. The comparison was performed under simulated field conditions. The bleach reagent gave superior results [ 67] .
In all analytical work, careful attention must be paid to obtaining a representative sample for analysis. A special problem has been reported to exist with cocaine because its particles are often larger than the adulterant particles [ 68] . The problem is minimized if large aliquots are taken or if the sample is ground before sampling.
Cocaine samples adulterated with sugars tend to discolour in storage. An explanation of the phenomenon is provided by Dugar and others, and a description given of an especially troublesome sample containing cocaine, procaine, aspirin and lactose which decomposed rapidly into a black, tarry mass [ 69] . Furfural and 5-hydroxymethylfurfural are found in discoloured cocaine samples. These compounds and formylbenzocaine are detected in mixtures of cocaine, benzocaine, and lactose prepared in the laboratory and which discoloured rapidly [ 70] . It is clearly impossible to perform a quantitative analysis that accurately reflects the original composition of a sample undergoing rapid decomposition, and probably wise not to allow any cocaine sample to remain unanalysed for long periods of time.
It is generally known that cocaine in solution will hydrolyze if exposed to strong acids or bases. Some more subtle phenomena may cause problems during cocaine analyses. Cocaine in aqueous solutions with pH above 5.5 will decompose into benzoylecgonine and methanol. A solution with pH 5.8 exhibits a 10 per cent loss of cocaine after 13 days storage at room temperature [ 71] , and a solution with pH 8 loses 35 per cent of its cocaine content after 4.5 hours [ 72] .
As mentioned earlier, if cocaine is to be quantified by gas-liquid chromatography the sample should be converted to the free base prior to injection to minimize decomposition in the injection port [ 73] .
A publication addressing the question of non-reproducibility of quantitative analyses of cocaine by gas-liquid chromatography [ 74] offers evidence that the phenomenon may result from the differences in the relative solubilities of cocaine and the internal standard. If the material to be quantified is less soluble in the injection solvent than the internal standard, the less soluble material may precipitate on the syringe needle, with the result that the ratio of sample to internal standard introduced onto the column varies from injection to injection. It is recommended that the material to be injected should be pulled back from the tip of the needle into the body of the syringe and that relatively dilute injection solutions should be employed.
Thorough reviews of separative and quantitative methods for cocaine have been published recently [ 75] [ 76] . A trend towards the development of increasingly efficient methodology may be discerned through study of the body of work reflected in these publications. Procedures are described that can accomplish the separation of many substances quickly and efficiently-This is viewed as a general trend in forensic analysis, but is of special relevance to the analysis of cocaine, where multiple adulterants are common, as are samples that contain no cocaine at all.
The method described by Baker and Gough [ 54] is an excellent example of a versatile method that presents an integrated scheme of analysis for cocaine samples, in which presumptive testing with colour tests is conditionally followed by thin-layer chromatography and gas-liquid chromatography analyses. Fourteen local anaesthetics are shown to be separable from cocaine. Among high-performance liquid chromatography methods, three seem of special interest. Noggle and Clark use the technique of absorbance ratioing to improve the specificity of the analysis for cocaine samples adulterated with local anaesthetics [ 77] . The same instrumental parameters are used to analyse typical mixtures that are sold on the illicit market as cocaine, but that do not contain any controlled substances. Both Jane and others [ 78] and Gill and others [ 79] report on the development of high-performance liquid chromatography methods that can detect cocaine in the presence of the common adulterants and some adulterants that are very rare. Cocaine is separated from cinnamoylcocaines and benzoylecgonine. Gill and others use a single column and two mobile phases to cover the range of polarities encountered in the materials analysed. Lurie [ 80] , [ 81] describes a method that involves only two mobile phases, and which can be used for the analysis of heroin, barbiturate, amphetamine and cocaine samples.
Courts in the United States have come to expect the forensic chemist to identify with the highest degree of scientific certainty the drug of abuse found in evidentiary material. To some extent the cocaine isomer issue has reinforced the requirement of the courts for methods of high specificity.
The chemical taxonomy of the coca plant is important to the forensic chemist because an interfering substance may be present in the extract from a particular variety of plant. Similarly, information regarding the extractive process used by illicit cocaine producers is relevant to the design of analytical methods. Knowledge concerning cocaine enantiomers and diastereoisomers is needed to ensure that the methods used distinguish cocaine from closely related compounds. Elimination of any source of error that arises from an inadequate procedure for sample preparation or from artefact formation is clearly essential to the task of forensic analysis.
In order for a laboratory to meet its needs for specificity while maintaining an acceptable level of analytical efficiency, more efficient and versatile methods have been developed. Advances in presumptive testing and in chromatographic methods have made it possible to screen samples of varying composition rapidly. The introduction into routine use in forensic laboratories of equipment for gas-liquid chromatography/electron impact mass spectrometry, proton magnetic resonance spectrometry or similarly advanced analytical techniques has also improved the possibilities for detection and identification of cocaine in seized samples.
The assistance of fellow members of the Special Testing and Research Laboratory staff was essential to the preparation of this report. My appreciation is extended to Andrew C. Allen, Donald A. Cooper, Joseph E. Koles, Ira S. Lurie, James M. Moore, Stanley P. Sobol, and Albert R. Sperling.
M. L. Bastos and D.B .Hoffman, "Comparison of methods for detection of amphetamines, cocaine and metabolites", Journal of Chromatographic Science , vol. 12, 1974, pp.269-280.
02M. L. Bastos and D. B. Hoffman, "Detection and identification of cocaine, its metabolites, and its derivatives", Cocaine: Chemical, Biological, Clinical, Social and Treatment Aspects , S. J. Mule, ed. (Cleveland, CRC Press, 1976), pp. 33-5 8.
03R. Hawks, "Cocaine: the material" , National Institute of Drug Abuse Research Monograph No. 13 (Washington, D.C., Government Printing Office, 1977), chapter 3.
04S. Archer and R. L. Hawks, "The chemistry of cocaine and its derivatives", Cocaine: Chemical, Biological, Clinical, Social and Treatment Aspects , S. J. Mule, ed.(Cleveland, CRC Press, 1976), pp. 59-70.
05T. Plowman, "Botanical perspectives on coca", Journal of Psychedelic Drugs , vol. 11, 1979, pp. 103-117.
06R. T. Martin, "The role of coca in the history, religion and medicine of South American Indians", Economic Botany , vol. 24, No. 4(1970), pp.422-438.
07R. Hegnauer and L. H. Fikenscher, "Untersuchungen mit Erythroxylum Coca Lam", Pharmaceutica Acta Helvetiae , vol. 35, 1960, pp. 43-64.
08E. Machado, "Determination of varieties and cultivars in Peruvian coca", Proceedings of the Interamerican Seminar on Medical and Sociological Aspects of Coca and Cocaine , F. R. Jeri, ed. (Pacific Press, 1980), pp. 239-245.
09T. Plowman, "Amazonian coca", Journal of Ethnopharmacology , vol.3, 1981, pp.195-225.
10T.Plowman, "The identity of Amazonian and Trujillo coca", Botanical Museum Leaflets (Harvard University), vol. 27, 1979, pp. 45-68.
11G. H. Aynilian and others, "Cocaine Content of Erythroxylum species", Journal of Pharmaceutical Sciences , vol. 63, No. 12 (1974), pp. 1938-1939.
12T. A. Henry, The Plant Alkaloids , 4th ed. (Philadelphia, Blakiston, 1949), pp. 92-105.
13E. P. Claus, Pharmacognosy (Philadelphia, Lea and Febizer,1970).
14L. Rivier, "Analysis of alkaloids in leaves of cultivated Erythroxylum and characterization of alkaline substances used during coca chewing", Journal of Ethnopharmacology , vol. 3, 1981, pp.313-335.
15M. J. Balick, L. Rivier and T. Plowman, "The effects of field preservations alkaloid content of fresh coca leaves (Erythroxylum spp.)", Journal of Ethnopharmacology , vol. 6, 1982, pp.287-291.
16B. Holmstedt and others, "Determination of cocaine in some South American species of Erythroxylum using mass fragmenentography", Phytochemistry, vol. 16, 1977, pp. 1753-1755.
17C. E. Turner,C. Y. Ma and M. A. Elsohly, "Constituents in Erythroxylum coca: I. Gas-chromatographic analysis of cocaine from three locations in Peru",
00Bullet in on Narcotics (United Nations publication), vol. 31, No. 1 (1979), pp. 71-76.
18C. E. Turner, C. Y. Ma and M. A. Elsohly, "Constituents of Erythroxylum coca: I. Gas-chromatographic analysis of cocaine and other alkaloids in coca leaves", Journal of Ethnopharmacology , vol. 3, 1981, pp. 293-298.
19E. Solon and A. Sperling, personal communication ,1984.
20Drug Enforcement Administration, unpublished documents, 1984.
21M. Morales-Vaca, "A laboratory approach to the control of cocaine in Bolivia", Bulletin on Narcotics (United Nations publication), vol. 36, No. 2 (1984), pp. 33-43.
22I. Jane and others, "Quantitation of cocaine in a variety of matrices by high-performance liquid chromatography", Journal of Chromatography , vol. 214, 1981,pp. 243-248.
23C. C. Clark, personal communication ,1978.
24C. C. Clark,personal communication ,1983.
25P.Majlat and I. Bayer, "Separation of cocaine, benzoylecgonine and ecgonine", Journal of Chromatography , vol. 20, 1965, pp. 11-17.
26F. Medina, personal communication, 1979.
27J. M. Moore, "Identification of cis-and trans-cinnamoylcocaine in illicit cocaine seizures", Journal of the Association of Official Analytical Chemists , vol. 56, 1973, pp. 1199-1205.
28T. Lukaszewski and W. K. Jeffery, "Impurities and artifacts of illicit cocaine", Journal of Forensic Sciences , vol. 25, No. 3 (1980), pp. 499-507.
29C. C. Clark, personal communication, 1981.
30F. Toffoli and U. Avico, "Coca paste: residues from the industrial extraction of cocaine. Ecgonine and anhydroecgonine", Bulletin on Narcotics (United Nations publication), vol. 17, No. 4 (1965), pp. 27-36.
31J. A. Siegel and R. A. Cormier, "The preparation of d-pseudococaine from l-cocaine", Journal of Forensic Sciences , vol. 25, No. 2 (1980), pp. 357-365.
32Code of Federal Regulations (Washington, D. C., Office of the Federal Register, Government Printing Office, 1984), part 21, paragraph 1308.12b.(4).
33S. P. Findlay, "The three-dimensional structures of the cocaines: I. Cocaine and pseudococane", Journal of the American Chemical Society , vol. 76, No.11 (1954), pp. 2855-2862.
34S. P. Findlay, "The three-dimensionaI structures of the cocaines: II. Racemic allococaine and racemic allopseudococaine", Journal of Organic Chemistry , vol. 24, 1959, pp. 1540-1550.
35A. Sinnema and others, "Configurations and conformation of all four cocaines from NMR spectra", Recueil des travaux chimiques des Pays-Bas , vol. 87, No.10 (1968), pp. 1027-1041.
36F. I. Carrol, M. L. Coleman and A. H. Lewin, "Syntheses and conformational analyses of isomeric cocaines: a proton and C-13 nuclear magnetic resonance study", Journal of Organic Chemistry , vol. 47, 1982, pp. 13-19.
37J. A. Siegel and R. A. Cormier, "The preparation of d-pseudococaine from l-cocaine", Journal of Forensic Sciences , vol. 25, No. 2 (1980), pp. 357- 365.
38C. Olieman , L. Maat and H. C. Beyerman, "Analysis of cocaine, pseudococaine, allococaine and allopseudococaine by ion-pair reverse-phase high-performance liquid chromatography", Recueil des travaux chimiques des Pays-Bas , vol. 98, 1979, pp. 501 - 502.
39A. H. Lewin, S. R. Parker, and F. I. Carroll, "Positive identification and quantitation of isomeric cocaines by high-performance liquid chromato-graphy", Journal of Chromatography , vol. 193, 1980, pp. 371 - 380.
40A. C. Allen and others, "The cocaine diastereoisomers", Journal of Forensic Sciences , vol. 26, No. 1 (1981), pp. 12 -26.
41R. H. Shapiro and others, "Mass spectral analysis of cocaine and pseudo-cocaines", Spectroscopy International Journal , vol. 2, 1983, pp. 227- 231.
42C. L. Flo, personal communication, 1982.
43G. J. Sorgen , personal communication, 1976.
44R. Ruybal, personal communication, 1982.
45J. M. Bowen and N. Purdie, "Identification of cocaine and phencyclidine by solute-induced circular dichroism", Analytical Chemistry , vol. 53, 1981, pp. 2239 -2242.
46J. M. Bowen and N. Purdie, "Determination of cocaine by circular dichroism", Analytical Chemistry , vol. 53, 1981, pp. 2223 -2239.
47P. L. Morgan , V. Ingram and D. Francois, personal communication, 1978.
48D. Eskes, "Thin-layer chromatographic procedures for the differentiation of the optical isomers of cocaine", Journal of Chromatography , vol. 152, 1978, pp. 589 - 591.
49R. Gelsomin o and J. K. Raney, personal communication, 1979.
50Comprehensive Crime Control Act of 1984, Public Law No. 98-473.
51D. A. Cooper and A. C. Allen, "Synthetic cocaine impurities", Journal of Forensic Sciences , vol. 29, No. 4 (1984), pp. 1045 - 1055.
52J. J. Tuforiello and others, "A stereospecific synthesis of d, l -cocaine", Tetrahedron Letters , No. 20, 1978, pp. 1733 - 1736.
53J. L. Young, "The detection of cocaine in the presence of novocaine by means of cobalt thiocyaate", American Journal of Pharmacy , vol. 103, 1931, pp. 709 - 710.
54P. B. Baker and T. A. Gough, "The rapid determination of cocaine and other local anesthetics using field tests and chromatography", Journal of Forensic Sciences , vol. 24, No. 4 (1979), pp. 847- 855.
55M. J. de Faubert Maunder, "The rapid detection of drugs of addiction", Medicine, Science and the Law , vol. 14, No. 4 (1974), pp. 243 -249.
56C. Stainier, "L'utilisation du thiocyanate de cobalt dans l'analyse des bases organiques", Il Farmacco , vol. 29, No. 3 (1974), pp. 119 - 135.
57G. V. Alliston and others, "An improved test for cocaine, methaqualone and methadone with a modified cobalt (II) thiocyanate reagent", Analyst , vol. 97, 1972, pp. 263 - 265.
58C. Ruybal, personal communication, 1973.
59R. Ruybal, personal communication, 1972.
60L. J. Scott, personal communication, 1973.
61S. K. Lorch, personal communication, 1974.
62C. L. Winck and Eastly, "Cocaine identification", Clinical Toxicology , vol. 8, 1975, pp. 205 -208.
63J. D. Prall, personal communication, 1975.
64B. Travnikoff, "Semiquantitative screening test for cocaine", Analytical Chemistry , vol. 55, 1983, pp. 795 -796.
65S. P. Sadtler and N. Evers, "Cocaine", Allen 's Commercial Organic Analysis , 5 th ed., C. A. Mitchell, ed. (Philadelphia, P. Blackistones, 1929), pp. 509 -549.
66F. W. Grant, W. C. Martin and R. W. Quackenbush, "A simple sensitive specific field test for cocaine based on the recognition of the odour of methyl benzoate as a test product", Bulletin on Narcotics (United Nations publication), vol. 27, No. 2 (1975), pp. 33 -35.
67R.W. Samuels, "Evaluation of sodium hypochlorite as a screening agent for the identification of cocaine", Clinical Toxicology , vol. 12, 1978, pp. 543 -550.
68B. Kratochvil and B. Brown, "Sampling errors in the determination of cocaine in seized drugs"', Journal of Forensic Sciences , vol. 29, No. 2 (1984), pp. 493 -499.
69S. Dugar, T. Catalano and R. Cerrato, "Discoloration effect of diluents in contraband cocaine"', Journal of Forensic Sciences , vol. 19, No. 4 (1974), pp. 868 - 872.
70J. M. Moore, personal communication, 1973.
71V. Das Gupta, "Stability of cocaine hydrochloride solutions at various pH values as determined by high-pressure liquid chromatography", International Journal of Pharmaceutics , vol. 10, 1982, pp. 249 -257.
72S. M. Fletcher and V. S. Hancock, "Potential errors in benzoylecgonine and cocaine analysis", Journal of Chromatography , vol. 206, 1981, pp. 193 -195.
73C. C. Clark, "Gas liquid chromatographic quantification of cocaine HCl in powders and tablets: collaborative study", Journal of the Association of Official Analytical Chemists , vol. 61, 1978 , pp. 683-686.
74J. C. Roberson, "Reproducibility of the internal standard method in gas chromatographic quantitation of cocaine", Analytical Chemistry , vol. 50, 1978, pp. 2145 -2146.
75T. A. Gough and P. B. Baker, "Identification of major drugs of abuse using chromatography", Journal of Chromatographic Science , vol. 20, 1982, pp. 289-329.
76T. A. Gough and P. B. Baker, "Identification of major drugs of abuse using chromatography: an update", Journal of Chromatographic Science , vol. 21, 1983, pp. 145 - 153.
77F. T. Noggle and C. R. Clark, "Liquid chromatographic analysis of samples containing cocaine, local anesthetics, and other amines", Journal of the Association of Official Analytical Chemists , vol. 66, 1983, pp. 151 - 157.
78I. Jane and others, "Quantitation of cocaine in a variety of matrices by high-performance liquid chromatography", Journal of Chromatography , vol. 214, 1981, pp. 243 -248.
79R. Gill, R. W. Abbott and A. C. Moffat, "High-performance liquid chromatography system for the separation of local anesthetic drugs with applicability to the analysis of illicit cocaine samples", Journal of Chromatography , vol. 301, 1984, pp. 155 - 163.
80I. Lurie, "Application of reverse phase ion-pair partition chromatography to drugs of forensic interest", Journal of the Association of Official Analytical Chemists , vol. 60, 1977, pp. 1035 - 1040.
81I. Lurie, "Improved isocratic mobile phases for the reverse phase ion-pair chromatographic analysis of drugs of forensic interest", Journal of Liquid Chromatography , vol. 4, 1981, pp. 399-408.
