MicrochimActa(2008)162:381–392 DOI10.1007/s00604-007-0862-4 PrintedinTheNetherlands Original Paper Archaeo-chemical analysis of Royal Purple on a Darius I stone jar Zvi C. Koren TheEdelsteinCenterfortheAnalysisofAncientArtifacts,DepartmentofChemicalEngineering, ShenkarCollegeofEngineeringandDesign,Ramat-Gan,Israel Received7March2007;Accepted29August2007;Publishedonline29October2007 #Springer-Verlag2007 Abstract. High-performance liquid chromatography textiledye,andtheonlyexampleyetdiscoveredwhere (HPLC) coupled with photodiode array (PDA) detec- it is the sole paint pigment on such a large royal art tion was used for the microchemical analysis of a object. purpleresidueonthesurfaceofa2500-yearoldstone Keywords: HPLC; Tyrian Purple; Hexaplex trunculus; Bolinus jar. This 30(cid:1)37cm pear-shaped marble vessel con- brandaris;Stramonitahaemastoma;bromoindigo;bromoindirubin; tains carved inscriptions praising the Persian king bromoisatin;Di-MonoIndex;KingDariusI Darius I and is unique due to the use of quadrilingual writings on avessel of that king. The major colorants Royal Purple is a designation used to describe those identified in the purple pigment are 6,60-dibromoin- regal textiles dyed with the purple pigment extracted digo,6-monobromoindigo,and6,60-dibromoindirubin, from the hypobranchial glands of certain Muricidae withnegligiblecontributionsbyindigoand6-bromoisa- sea snails inhabiting the Mediterranean. These purple tin. This analysis establishes that a marine mollusk or violet garments conferred upon the wearer an aura was the source of the purple pigment, which is the of power and sacredness and,thus, thesedressedonly famousRoyalPurpleorTyrianPurpleoftheancients. sovereigns, military generals, eminent officials, and A comparison with the relative dye compositions of high priests. variousMuricidaespecies(Hexaplextrunculus,Bolinus Inthepasttwodecades,thismolluskanpigmenthas brandaris, and Stramonita haemastoma), and with been the focus of accelerated research as it is one of their newly formulated Di-Mono Index values, sug- the most mystifyingly complex of all the natural col- gests that the biological provenance of this ancient orants used inantiquity. Much has been written on its pigment was probably an indigo-deficient Hexaplex general history, initially by classical authors, such as trunculus sea snail. The entire exterior of the vessel – the 4th century BCE (Before the Common Era) Greek including its base – was originally painted purple by philosopher Aristotle [1] and the 1st century CE usingafresco-typetechnique.Thisisonlythesecond (Common Era) Roman historian Pliny [2], and more chromatographic finding of a molluskan purple color- recently by Cardon [3, 4] and Haubrichs [5, 6]. Its ant in use as an ancient paint pigment and not as a chemistry has been reviewed [7, 8], and analytical methods have been developed for multicomponent Correspondence: Zvi C. Koren, The Edelstein Center for the identifications of Muricidae pigments via liquid chro- AnalysisofAncientArtifacts,DepartmentofChemicalEngineer- matography [9–16]. ing,ShenkarCollegeofEngineeringandDesign,12AnnaFrankSt., The three main molluskan species that have been 52526Ramat-Gan,Israel e-mail:[email protected] associated with purple dyeings in the Mediterranean 382 Z.C.Koren region are Hexaplex trunculus (synonyms: Murex nance of the purple residue from the 2500-year old trunculus, Phyllonotus trunculus), Bolinus brandaris Darius vessel. (synonyms:Murexbrandaris,Phyllonotusbrandaris), and Stramonita haemastoma (synonyms: Purpura Description of the stone jar haemastoma, Thais haemastoma) [3–5, 17, 18]. The biochemical procedure of extraction of the dye from The pear-shaped stone jar investigated in this study the living animal and the subsequent textile dyeing (Fig. 1) is exhibited at the Bible Lands Museum in process was one of the most complex of crafts prac- Jerusalem, catalog # BLMJ 1979, and briefly de- ticedbytheancients.Thepigmentisnotpresentinthe scribed in the guide to the museum [29] well as in gland of the living animal, but is spontaneously pro- themuseum’sinternetsite[30].The30cm(maximum ducedoncetheglandanditscontentsareexcisedfrom width)by37cm(height)vesselhasaroundedbottom, the snail or it expires due to other causes. Enzymatic a thick everted rim, and several horizontal natural hydrolysis of the dye precursors followed by photo- bandsencirclingit.Initscurrentstate,thejarconsists chemical oxidative processes occur to produce the of various pieces that have been glued together in the purple or violet pigment. restoration process, but some of the sections from the The discoverersofthis purplepigmentwereproba- body of the vessel are missing. It contains carved bly the ancient Minoans in the Aegean from at least inscriptionsextollingKingDariusIinfourlanguages: the 18th century BCE [19, 20]. However, the earliest Egyptian hieroglyphs (containing a dedication to his direct archaeological and chemical evidence for the 36thregnalyear,whichwashislast,in486=485BCE) use of the purple molluskan pigment in dyeing is and wedge-shaped cuneiform scripts in Old Persian, from Phoenician sites at Sarepta, modern Sarafand, Elamite,andAkkadian[27].Thisisararefind,asthis Lebanon, in the 13th century BCE [21, 22], and at jar is the only vessel of any kind yet discovered be- Tel Akko in today’s northern Israel, in the 13th–12th longing to Darius I that has multilingual inscriptions. centuriesBCE[23,24].AttheseandotherPhoenician sites, potsherds with molluskan purple residues were found.Thesefragmentsbrokeofffromlargeclayvats that could contain several hundred liters of dye solu- tion in which textile dyeing was performed. Hence, this pigment has also been called Tyrian Purple after Tyre, one of the capital cities – and a purple produc- tion center – of the Phoenicians. The vatting process, necessary for dyeing textiles with this colorant, required the dissolution of the pur- ple pigment via reduction to a leuco-form. After the textile was immersed in the dye solution and then removed from it, the dye underwent air-oxidation backtoitsinsolubleoxidizedforminthetextilefibers. The all-natural fermentative reduction that would have conceivably been performed in antiquity has re- cently been independently rediscovered by John Edmonds[25],IngeBoeskenKanold[26],andbythis author [18]. In the current study, the chromatographic method is applied to the analysis of purple stains from the surface of a stone jar belonging to the Achaemenid king Darius I (also known as Darius the Great) who reigned in ancient Persia from 521 – 486=485 BCE [27, 28]. Purple pigments extracted from the three Fig. 1. Darius I stone jar (486=485 BCE) measuring 30cm at molluskan species mentioned above have also been maximum width by 37cm high (Courtesy of the Bible Lands analyzed in order to determine the biological prove- Museum,Jerusalem) Archaeo-chemicalanalysisofRoyalPurpleonaDariusIstonejar 383 This author’s close inspection of the residual mate- rial encompassing the Darius jar revealed that the en- tireoutsideofthestonevesselwasfirstplasteredwith a white chalky material and then painted with the purple pigment in a fresco-type technique. This con- clusion is furthersupportedbythe surprisingfact that even the base of the vessel was painted purple, as residual plaster and purple pigment adhering to it are still visible. The brown stains on the vessel are a result of fouling and soiling of the vessel and are not part of the original coloration. The royal jar’s title-holder, Darius I (Darayavaush in Old Persian) has a colorful history. He is related to two other kings of the Achaemenid dynasty, Cyrus the Great and Xerxes I (the Great), and these three areprobablythemostfamousofPersianrulers.Darius I (the Great) was the son-in-law of Cyrus (Kurush), and the latter’s reign (550-530 BCE) began the Achaemenid dynasty [28]. In the Bible, the name Daryavesh (‘‘Darius’’) appears in Prophets (Hagai and Zecharia) and in Scriptures (Daniel, Ezra, and Nehemia). Furthermore, the biblical Book of Esther provides colorful descriptions of the Persian palace at the capital of Shushan (Susa) that was festooned withpurpleandvioletcloths,whichaccordingtotheir Hebrew etymology were produced from molluskan pigments. That biblical narrative describes how Queen Esther married the Persian king Akhashverosh, probablyXerxes(Khashayarsha)isintended,whowas the son of Darius the Great. An analysis of the elemental composition of the stone jar was performed by Dr. S. Ilani of The GeologicalSurveyofIsrael,Jerusalem[27].Themeth- odwasbasedonscanningelectronmicroscopy(SEM) with an energy dispersive system (EDS). It was re- ported that the jar’s material is made of Ca, C, and O, and that it reacts with HCl (5%), which is indi- cative of calcite (CaCO ) [27], and signifying that 3 the stone jar’s material is marble. The white chalky coat onto which the purple pigment adhered was also analyzed in that same laboratory with the result that thewhitematerialconsistedofAl,Si,andO,withthe Al and Si contents almost equal (about 15%), a com- 2 Fig. 2. Molecularstructures, PDA-producedUV=Visspectra, and abbreviated names of the standard molluskan dyes: isatinoids (IS isatin, 6BIS 6-bromoisatin), indigoids (IND indigo, MBI 6-mono- bromoindigo, DBI 6,60-dibromoindigo), and indirubinoids (INR indirubin, 6MBIR 6-monobromoindirubin, 60MBIR 60-monobro- moindirubin,DBIR6,60-dibromoindirubin) 384 Z.C.Koren position indicative of kaolinite (Al O (cid:2)2SiO (cid:2)2H O) USA; www.mallinckrodt.com). The HPLC-gradeeluents consisted 2 3 2 2 of methanol and water, both supplied by J. T. Baker (Deventer, [27]. Holland; www.jtbaker.com), and 85% ortho-phosphoric acid by Although the museum’s website [30] describes the Fluka(Buchs,Switzerland;www.fluka.com).Theconcentratedacid residual pigment on the Darius jar as ‘‘Purple Dye wasdilutedwithHPLC-gradewatertoprovidea5%w=vsolution withapHof1.50at25(cid:3)C. (Murex)’’, the pigment has never been scientifically Thestandardreferencedyesweresynthesizedandkindlyprovid- identified as such prior to this current work. This de- edbyDr.ChrisCooksey(UniversityCollegeLondon,UK)[7,8], scription was simply based on a visual assessment exceptforindigo,whichwasobtainedas‘‘indigorein’’fromBASF of the hue of the pigment. Recently, this purple stain (Germany; www.basf.com). The common and abbreviated names and the molecular structures of the nine dyes analyzed are given was mineralogicallyanalyzed,butwithanambiguous inFig.2. determination that the reddish colorant might be due to red ochre, hematite, Fe O [27]. 2 3 HPLCsystem Flora and fauna sources for purple and pink pig- ments from other than molluskan origins are varied. The ambient-temperature reverse-phase chromatographic system manufactured by Waters (Milford, MA, USA; www.waters.com) For example, purple dyes (or pigments) can be pro- consistedofa600EControllerpumpanda996PDAdetector,each duced by mixing the blue indigo dye (extracted from controlledbythe Millennium-32 software.Thestationaryphase the leaves of the indigo or woad plants) together with consistedofa3.0(cid:1)150mmC Symmetrycolumn(WatersPart 18 No.WAT054200)with5mmand100A˚ particleandporediameters, a red dyestuff (extracted, for example, from the roots respectively. A 20-mL sample loop was used. The ternary mobile of the madder plant, or from scale insects) [31]. phase system consisted of water, methanol, and 5% w=v H PO. 3 4 Further, pink and violet mineral pigments were found Two linear gradient elution methods were utilized, and these are in an artist’s palette and on the frescoed walls at the shown in Table 1: Method 1 consists of a constant flow rate of 0.8mLmin(cid:4)1,whileMethod2usesanincreasingflowratetowards 1st century BCE King Herod’s palace at Jericho. In the end of the run. Both methods are identical up to 21min of those examples, the pink pigment originated from elution. heating white kaolinite containing a small amount of yellow goethite to 850(cid:3)C, thereby transforming goe- SamplepreparationforHPLCanalysis thite to red hematite, which fused with the kaolin- Theextractionsofthearchaeologicalandsnailpigmentsandtheir ite, and the violet pigment was a mixture of that subsequentfiltrationswereperformedundersubduedlightingcon- pink colorant with Egyptian blue [32]. Similarly, ditions in order to prevent the photo-debromination and thus de- Vitruvius, the 1st century BCE Roman architect and gradation of any brominated colorant that may be present in the extractedsolution.Eachsolidmicro-sampleanalyzed(aboutamil- engineer, described imitation purple pigments for ligram)wastreatedwith200mLofdimethylsulfoxide(DMSO)and wall paintings [33]. heated for 5min at 100(cid:3)C. The resulting mixture was allowed to The importance of this current work was therefore cooltoroomtemperaturefor15min,andthenfilteredina0.45-m to determine the nature and source of the ancient micro-spinpolypropylenecentrifugetubewithnylonfilter(Alltech, Part No. 2490; www.alltechweb.com=product) and immediately Darius purple pigment by microchemical means. injectedintotheHPLCsystem. Experimental Archaeologicalsampleanalyzed Thestonejarconsistedofmanyouterareascontainingresidueswith Reagentsanddyes alargevarietyofpurplehues.Forthecurrentstudy,minutesamples The extracting solvent, dimethyl sulfoxide (DMSO), was spectro- werescrapedoffthestonejarfromtwoareas.Eacharchaeological photometric grade and supplied by Mallinckrodt (Paris, Kentucky, sample analyzed consisted of a mostly white powdery material Table1. LinearHPLCgradientelutionmethodsfortheanalysisofmolluskandyes:constant(Method1)andincreasing(Method2)flowrate methods Method Time(min) Flowrate(mLmin(cid:4)1) Methanol(%) Water(%) 5%H PO (pH¼1.50)(%) 3 4 Methods1and2 0–3 0.8 30–75 60–15 10 3–20 0.8 75 15 10 20–21 0.8 75–100 15–0 10–0 Method1 21–30 0.8 100 0 0 Method2 21–24 0.8–1.4 100 0 0 24–25 1.4–2.1 100 0 0 25–27 2.1 100 0 0 Archaeo-chemicalanalysisofRoyalPurpleonaDariusIstonejar 385 Table2. DescriptionoftheMuricidaesnailsamplesusedinthisstudy Snailspecies Provenance Sampleabbreviation Sampledescription usedinthefigures Hexaplex Akhziv, Tr.A. Theglandsfrom20snailswereexcisedonthebeachofAkhziv, trunculus northernIsrael Israel,onanovercastmorningofFriday,May14,1993 ThreeDMSOextractionsfromthissamplemixturewere preparedforanalysis Hexaplex Spain Tr.S. TheexcisionanddryingoftheglandswereperformedinSpain trunculus andobtainedfromJ.Guberman TenDMSOextractionsfromthissamplemixturewereprepared foranalysis Bolinus Fiumicino, Br.F. Theglandularcontentsweresmearedasastainonafilterpaper brandaris Italy andobtainedfromC.Porter TwoDMSOextractionsfromthissamplemixturewereprepared foranalysis Stramonita Israel Hm.I. ThesnailswerecollectedalongthecoastofIsraelandplacedinthe haemastoma author’slabaquarium.TheglandswereexcisedonTuesday,June1,1993 andplacedinaPetridish.Thecolorwasveryslowtodevelopinthe labandthefinalpurplecolorappearedafteranumberofdays ThreeDMSOextractionsfromthissamplemixturewerepreparedforanalysis mixedwithsomepurplepigment.Underanopticalmicroscope, it The structures of the standard dyes and their ac- was seen that some of the pigment particles adhered to thewhite companying UV=Vis spectra produced by the PDA plasterparticleswhileotherswereseeminglypure.Threeextractsof thearchaeologicalsampleswereanalyzed. detector are shown in Fig. 2. As expected, dyes in the same chemical group show similar visible wave- lengths at maximum absorption, (cid:1) , which for the Preparationofthemolluskanpigments max two isatinoids is in the range of 410–418nm, for the The shells of Hexaplex trunculus, Bolinus brandaris, and three indigoids 598–615nm, and for the four indiru- Stramonitahaemastomaseasnailswerestrategicallybrokenwith binoids 530–544nm. ahammerblowtoexposeeachsnail’shypobranchialgland,which wasthenexcisedfromtheanimalandexposedtothenaturallight The chromatograms for the archaeological and andtemperatureofthesurroundings.Thepigmentproducedfrom modern snail pigments are shown in Fig. 3 and are the glandular fluids, whose final color was purple, was allowed discussedbelow.TherespectiveUVwavelengthscho- tonaturally dry.Eachsuchsample consisted of remnantsof the glandular meat and the pigment itself. The snails used for this sen for displaying these chromatograms are the ones study are described in Table 2.Several DMSO-extractions were that clearly show all the detectable colorants. The performedfromeachsampleandtheexactnumbersaregivenin other peaks appearing in the chromatograms are only thetable. UV-absorbing substances and thus are non-colorants. Chromatographicproperties,suchasretentiontimes, Results and discussion t , and absolute and relative integrated peak areas for R the dyes in the archaeological and molluskan sam- HPLC analyses ples were calculated at three relevant wavelengths and areshowninTable3.Thewavelengths chosen for The constant flow-rate elution method (Table 1) uti- each dyecorrespondtoastandarduniform288nm,at lized in this study was previously successfully used which all the dyes have significant absorption, and to for the production of a calibration chromatogram for the UVand visible (cid:1) for the specific dye. Though the detection of ten reference dyes composed of indi- max theareasarenotactualconcentrations,semi-quantita- goid,indirubinoid,and isatinoidcolorants[13].Inthe tivecomparisonscan nevertheless be madewith them current study, a second method was also used. These by using a standard wavelength of 288nm, for exam- two methods are identical for the elution of the first ple, for all peak areas. nine dyes that may be present in molluskan purple pigments, but in order to hasten the exit of the last HPLC analyses of the archaeological sample component(DBIR)theflowratewasincreasedforthe secondmethodataratethatwouldyieldcolumnhead The chromatogram of the extracted Darius pigment pressures below 4000 PSI. is shown in Fig. 3a and the corresponding chro- 386 Z.C.Koren matographic properties of each identified dye in that pigment are given in Table 3. The major dyes that were detected in extracts from the archaeological purple residue were DBI, MBI, and DBIR with UV-Vis absorption spectra and retention times clear- ly matching the respective standard dyes. Traces of IND and 6BIS were also detected. The identification of DBI in such a sample is definite evidence that the purple used in painting the stone vessel was the Tyrian Purple or Royal Purple of molluskan origin. The DBI dye is a chemical biomarker for such a marine pigment. The finding of archaeological fibers dyed with a realmolluskanpurplepigmentisarareeventbyitself as these textiles were reserved for the most powerful and holy. In the current study, the unique use of this colorant as a painting pigment – and not as a textile dye – is a major discovery. The only other definitive chromatographic detection of the use of purple as a paint pigment is from the Late Bronze Age (17th cen- tury BCE) wall paintings of Akrotiri (the ancient name of the Greek island of Santorini) at Thera [14, 16, 34]. These purple paint pigments may be the pur- purissum, a processed mixture consisting of a chalky substance mixed with a molluskan pigment, and de- scribed many centuries later by Pliny [2, 16]. HPLC analyses of the molluskan pigment samples Archaeometric analyses often involve elemental anal- ysesofarchaeologicalobjectssuchasclayware,glass, ceramics, etc. in order to determine the geographical provenance of the object. In this current work, an at- tempt was made to determine the biological prove- nance of the Darius purple pigment by analyzing various purple colorants from different Muricidae speciesthatwerelistedinTable2.Thecorresponding chromatogramsofextractsfromthesesnailsareshown in Figs. 3b–e, and the relevant chromatographic prop- ertiesgiveninTable3.Forclarityofcomparisons,the percent areas measured at 288nm for all samples analyzed are also plotted in Fig. 4. 1 Fig.3. HPLCchromatogramsoftheDMSO-extractedpurplepig- mentsfrom:(a)theDariusjar(Method1ofTable1)at300nm;(b) Hexaplex trunculus snails collected at Akhziv (Method 2) at 290nm; (c) H. trunculus snails collected in Spain (Method 1) at 300nm;(d)BolinusbrandariscollectedinFiumicino,Italy(Meth- od 2) at 288nm; (e) Stramonita haemastoma collected in Israel (Method2)at288nm Archaeo-chemicalanalysisofRoyalPurpleonaDariusIstonejar 387 2) a,m Table %are288n 2.80 20.65 0 0 0.66 0 65.44 0 10.46 o el) ents(seeals astoma(Isra Absolutearea(PDAunits) 69,63376,6319,662 513,5094,640,068179,169 000 000 16,3725,322 000 1,627,6171,662,4251,270,390 000 260,184368,338136,329 m m nailpig S.hae (cid:5)(cid:5)tR(min) 3.02 7.19 12.94 17.52 24.97 s nd a,m ariusa no) %are288n 0.08 37.45 0 0 0.85 0 59.27 0 2.35 D ci dyeforthe aris(Fiumi Absolutearea(PDAunits) 1,04520,704754 506,7843,051,471117,834 000 000 11,5086,976 000 801,977839,431647,431 000 31,74443,92914,109 d h n foreac B.bra (cid:5)(cid:5)tR(min) 2.95 7.28 12.88 18.35 25.09 s e valumax %area,288nm 0.87 0.59 38.91 3.19 39.49 1.84 4.06 1.14 9.90 (cid:1) UVandVis ulus(Spain) Absolutearea(PDAunits) 166,0131,894,86028,015 111,922436,53723,438 7,406,7258,098,0885,090,386 608,002612,170225,819 7,517,9666,279,090 350,777436,200209,274 773,655684,167667,782 217,697237,97285,908 1,883,7974,982,726127,947 e c h n ndatt H.tru (cid:5)tR(min) 3.00 7.42 10.29 12.24 13.37 18.37 19.89 20.18 27.58 a 1. (288nm) %area,288nm 0 0 0.35 0 7.36 0.73 67.89 0 23.68 Table1.ofTable astandardwavelength H.trunculus(Akhziv) (cid:5)(cid:5)tAbsoluteR(min)area(PDAunits) 000 000 10.107,5859,6487,254 000 12.75160,281101,620 16.9515,94419,54210,460 18.311,479,3731,411,7131,096,127 000 25.07516,030745,782275,715 method,Method1,ofatemethod,Method2, esat a,m rateowr Table3.Integratedpeakareasandretentiontim Dariuspigment (cid:5)(cid:1)DyetAbsolute%areR(nm)(min)area(PDA288nunits) IS2880024104180 6BIS2887.121,6240.2125511,901410958 IND2889.941,4410.192861,4806152,711 INR2880029005400 MBI28812.60114,68214.8160770,942 6MBIR2880029805300 DBI28818.14638,23182.43303640,405598480,404 0MBIR28800629405440 DBIR28826.9318,2642.3630026,45654010,293 (cid:5)Retentiontimesbasedontheconstantflow(cid:5)(cid:5)Retentiontimesbasedontheincreasingfl 388 Z.C.Koren of H. trunculus snails in 2001 [13, 35]. However, the two most glaring differences between these snails is that the Akhziv sample has a negligible quantity of IND (<0.4% peak area at 288nm), whereas the Spanish mollusk shows nearly 40% peak area (Table 3). On the other hand, the relative DBI compositions of the two snails are quite the opposite withAkhzivatalmost70%peakareaandtheSpanish at only 4%. BoththeAkhzivandSpanishH.trunculuspigments contain significant levels of MBI, with the Akhziv sample at about 7% and the Spanish sample at nearly 40%. Other studies havealso shown that H. trunculus snails contain significant amounts of MBI [9, 10, 12, 13, 15]. The first HPLC detection of DBIR in H. trunculus snails was in 1995 [12] and the relative quantity of this dye is perhaps an unexpected result. The Spanish sample shows about 10% peak area whereas the Akhziv sample shows about 24%. These results indicate that the common denomi- nator for all H. trunculus snails studied so far is the existence of significant levels of MBI. However there are indigo-rich (DBI-poor) and indigo-poor (DBI- rich) H. trunculus snails. The indigo-rich mollusks, such as the ones from Spain of this investigation Fig. 4. Comparative plot of the percent peak areas measured at and from Tel Dor in north-central Israel [12], as well 288nmforninepossibledyesfromtheDarius(‘‘Dar.’’)andsnail as the snails previously studied [9, 10], produce bluer samples(seeTable2foranexplanationofthesnailabbreviations) purple (violet) pigments. The pigment produced from indigo-poor H. trunculus snails, as the ones from The striking results for several snails of H. truncu- Akhziv, is a redder purple. lus of different geographical origin analyzed show The differences between B. brandaris, Fig. 3d, and that the adage of ‘‘not all H. trunculus snails are cre- S. haemastoma, Fig. 3e, snails are only slight. They ated equal’’ [13] is again quite apparent in this study. both do not contain any IND, and negligible levels H. trunculus snail pigments can have radically dif- of MBI. As expected, their DBI levels are high with ferentcompositionsandthushues – fromreddishpur- both at about 60% peak area, and they contain some ple to a bluish purple (or violet), depending on the DBIR from 2 to 10%, respectively (Table 3). This relative quantities of redder or bluer colorants consti- generaltrendisalsocorroboratedbyotherresearchers tuting the pigment. There is a misconception regard- [9,10,15].However,theSpanishH.trunculussample ing H. trunculus pigments in that it is often thought and the S. haemastoma snail have nearly the same that they all tend to have bluer (violet) tones as com- relative peak area for DBIR at about 10% each. The pared with other species. However, the results for most dramatic and surprising aspect of the dye analy- thepurpleH.trunculuspigmentfromAkhzivinnorth- sis is that both the B. brandaris and S. haemastoma ern Israel, for example, show that this is not always samples show large quantities of 6BIS, almost 40% the case. peak area for the B. brandaris to about 20% for the The H. trunculus pigment from Akhziv, Fig. 3b, S. haemastoma. This is the firsttime that this dye has containsfiveoftheninedyesthatwerestudied,where- been found in these species, certainly at such high as the Spanish H. trunculus pigment, Fig. 3c, has all levels, and is in sharp contrast to the H. trunculus of them. The two brominated indirubin isomers and snails,whichhavevirtuallynone.Thus,thesignificant the two isatinoids were first detected in pigments presence or lack of this 6BIS dye can also be used to Archaeo-chemicalanalysisofRoyalPurpleonaDariusIstonejar 389 differentiatetheB.brandarisandS.haemastomasnails cma from H. trunculus ones. masto [10] The findings regarding the relative dye composi- e m. tions of Muricidae snails can be applied to determine S.ha H 003916 the biological provenance of the purple pigment used a in the painting of the Darius stone jar. However, such m o anecaorlmypaanriismonponseseibdlsetsotabteistcicaaultimouisslsyioanptpoliaenda,layszeitailsl oted) S.haemast(Israel)Hm.I. 000.8685.4813.66 ctra. purple-producingsnailspeciesinastandardizedman- n e p e s noficlolernugdrgd.ieecerTraplththroeieanreqntgudetaahorleecgitshraeiaentpminsvhnuieicacmlayiallelraplaenoibngcdlvmaeitqriepuoonaanntrnsmao.tmieftTnatehthttaieeevlsreseslcnytovaoniaclrdschi,iaoetbtinmholteenricossbaliilnooliny-f- udyexceptwher B.dbrandaris Br.[15] 0–1 1–297–981 =mtheirUVVis atttoaa2cahhhrcnof8eeeescdm8aeiiorgsm.xnptminhAtmzohoriepenaafiss)sabtcecdonttniaaphnyitytonnarelesiotyntdn,bDpgiaDHaoanrmagbof.BclrielfiltoigIteursuhulhudeasecontnsiniaftncpndpnog.itgoidghfsgMfimnpecmdFaoMoBaesiiesngnltnIsBnsed,.atitibI(iaf4thlflofitsr,(eaor,or>oasnimttmpmsh7tor,oie%anoastietnnldarh)tya-yluc,emsp.ceotripwigatexFfnagdl,roagra>ieoetponhmmnt9pndphet7e)aoea%Tamtpjlhornlpaouseardbusptttlsrithdeeohpte(oyahsnaal3dkeeestt, dernsnailpigments(valuesfromthisst B.B.ccbrandarisbrandarisnculus(Fiumicino)Br.Br.[10]F.[10] 00001.36394.88833.7614 ptivitiesofthesedyesascalculatedfro cfroolmorainntdiogno-tphoeorD(aarnidusDjBarI-irsichth)eHp.igtrmunecnutlupsrosdnuacielsd, dmo mes H.tru Tr. 5573530 absor slyimTrihelaperorrettoseudtlhtseatoofnstpehesecfirrfieoclmatwivAaevkpehelzeainkvg.atrhesasbtyhaottwheerreHcPleLaCr- eologicalan breviatedna H.trunculus(Spain)Tr.S. 40.723.3441.334.2510.36 therelative smamPaThdtnnoyhhueeddweoednnireietntaaesntvrghsie[eoce1wrvnoig4aaefis,ptnvrhaa1eeMaois6nnnuiu,cpdttlieor3odnimt4tcphasT]ibibhtud.gaeiesaTbTmnrelhueetdeehldessnie4nsKatfd.warnpaoiTfailcbimrtslghaohrymeiumsasteintteienhhsondoeAuitonnssafkhceolr7yrtlarchuottshhfhedtariaueoesrKecrnimonaeadilnt(beopISttrrgsiuuhigasrirrinoapcpysfentailrogleleosrmBtm[ippvun1Ceiiadi2ggntE)ya-]--t, offivedyesat288nmforarcha Modernsnailsamplesandab H.H.btrunculustrunculus(Akhziv)Tr.Tr.[12]A. 4.050.350017.797.4160.0068.3918.1623.85 ommunication.ginal600nmvaluesbymeansof onalcommunication. tdofaurd(TaioimnnufaunshudtakrrgooteerelhybudDy,s(ecig.sarlaBaoeehesAansmsdIndtRpuwthapchhwc,eliooeaeystayemnlissnloreaeecrepdna,nesoasrtsssreprasputrodisaneeeslmiwitrtdsordfisffpfoeneuooeoortnewrrrieanxamemmmodifrnrdryeeeuetooatdsdlpocpmntdoiahhgbosdrrafiiybnttioamhseffe)mfledtoaleih,fttlarmtohaoaeh(tprrennoInoelrNr,dgytorsthereRrcIeeaeeNaex)lvpsasnutDeahdwrcsdtaaliiyic,unacocruecesaMgtnircihsln.aoasomBegtlnoiInerssnIesdo,pp,trbiishrraeDtfeooooiotvothnBmcctdaaneeeellI)lyes-r--rl,. e4.Relativepeakareasasapercentage Archaeologicalsamplesandabbreviatednames abDariusTheraKabri Dar.Th.[14]Kb.[12] 0.190.713.93014.8425.9024.0782.6166.2954.75R2.367.1017.25 Karapanagiotis,Ref.[14]andpersonalcKoren,Ref.[12],convertedfromtheoriWouters,Ref.[10].Karapanagiotisetal.,Ref.[15]andpers amenable to interpretation if more components were Tabl Dye INDINRMBIDBIDBI a b c d 390 Z.C.Koren Provenance and function of the stone jar The hieroglyphic inscriptions carved on the stone jar areindicativethatthese – andthevesselitself – were probably produced in Egypt, which was at the time part oftheAchaemenidEmpireandvirtuallytheonly placewheretheuseofEgyptianhieroglyphshadprac- tical meaning [27]. However, part of the jar’s proces- sing was almost certainly produced in an area other thanEgypt,anargumentthatcanbededucedfromthe natureofthepaintpigment – molluskanpurple.There is no known association of molluskan purple pigment inPharaonicEgypt[36],neitherasatextiledyenoras a paint pigment. The purple craft probably did not Fig. 5. Di-Mono Index (D.M.I.) values at 288nm for pigments from Muricidae snails and on archaeological objects (see Tables enter Egypt before the Greek Macedonian conquest 2 and 4 for explanations of the sample abbreviations; values ob- of Egypt by Alexander the Great and the establish- tainedfromthisstudy,exceptwherenoted) ment of the Hellenistic Ptolemaic dynasty at the end of the4th centuryBCE after thedemiseofAlexander. Hence,itisalsomostprobablethattheplasteringand reportedupon,neverthelessapatterncanbespottedin painting of the vessel were performed in a different these cases. As was the case with the Darius jar, the part of the Persian Empire. The painting would have TheraandKabripigmentshaveaconsiderableamount been performed at a geographical location where the of MBI, which is negligible in the B. brandaris and expertiseinextractingthepurplepigmentfromMurex S.haemastomasamples.Thus,itismostprobablethat snails existed. Finding such a site is not difficult as the archaeological Thera and Kabri pigments were during the reign of Darius I the Persian Empire con- producedfromsimilarindigo-poorH.trunculussnails, trolledmuchoftheMediterraneanbasin[28],hometo as was the case with the Darius jar. the purple-producing Muricidae snails. This study suggests that only a few important dyes An obvious conjecture regarding the use of this canhelpdeterminethezoologicalsourceofthedye,at stone jar is that it was somehow associated with con- least in the case of H. trunculus. Thus, a simple pre- taining the purple pigment – either for storage to be dictiveindextodeterminewhetherH.trunculussnails used later or as a dyeing vessel. These assumptions, were used for the production of a particular pigment however,can be convincinglydisproved fora number may be the ratio of peak areas for the di- and mono- of reasons. The interior walls of this jar are relatively brominated indigoids, A =A , at 288nm or an- DBI MBI clear and do not show any residual signs that the other wavelength with significant absorptions for vessel contained any purple (or other) pigment inside both. This ‘‘Di-Mono Index’’ (D.M.I.) for each pig- it, unlike the ancient dyeing vat potsherds that were mentisdepictedinFig.5.Thissimplemodelneedsto found in the Levant [12]. This clean-look does not be tested further of coursewith more snails analyzed. appear to be a result of any deliberate cleaning pro- However, the trend in values appears to be that pig- cess in antiquity or by modern-day conservators. ments from H. trunculus species have D.M.I.’s in the Furthermore, due to the delicate nature of the exter- rangeoflessthan1andupto10,withtheindigo-rich nallypaintedjar,therewouldbenosensiblereasonto snails at the low end of the scale and the indigo-poor use this vessel as a dye vat. Heating of the purple vat onesatthehigherend.TherangeofD.M.I.valuesfor was described by Pliny [2], and its use can also be B. brandaris are from about 30–70, and for the S. deduced from the char marks still visible on the out- haemastoma about 30–100. Thus, there are clear dif- sides of archaeological potsherds [12]. Heat was nec- ferencesbetweenthevaluesforH.trunculusandthose essaryfortheeffectivedissolutionviareductionofthe for the other two species. Based on the respective pigment in the vat’s liquid mixture and would have D.M.I.values,itcanbededucedthatthethreearchae- obviously damaged the royal purple painting. It also ological pigments mentioned above were probably does notseemreasonable tosupposethat thecontain- producedfromsimilarindigo-poorH.trunculussnails, er held cosmetics, for instance, as smaller containers which reaffirms the previous conclusion.