Quick viewing(Text Mode)

The First Wine & Beer. Chemical Detection of Ancient Fermented

The First Wine & Beer. Chemical Detection of Ancient Fermented

The First & Chemical Detection of Ancient Fermented Beverages

ease, and materials processing. complex mixture of and or- Rudolph H. Michei and Under certain conditions. in either ganic compounds (Figure l), iuclud- Patrick E. McGovern very dry or waterlogged'environ- ing alcohols, aldehydes, acids, carbo- Museum Amlied Saence Center for menta where microbial activitv and hydrates, , , and Downloaded by UNIV OF PENN on July 16, 2009 Archaeokgy (MASCA) autoxidation are reduced, some or- polyhydroxyaromatics such as tan- University Museum ganic compounds or their degrada- nins, anthocyanins, flavonols, and University of Pennsylvania tion products are preserved in an- catechins, all of which contribute Philadelphia, PA 19104 cient archaeological contexts. We significantly to its color and . previously demonstrated the pres- Phenolic carboxylic acids, which

Published on September 22, 2008 http://pubs.acs.org | doi: 10.1021/ac00056a002 Virginia R. Badier ence of royal purple, one of the most are the degradation pmducta of poly- Department of Near Eastern Studies famous dyes of , on pottery hydroxyaromatics, are relatively sta- University of ble and are a distinguishing group of Toronto. Ontario sherds (Adchem lS86,57,1514A). Canada M5S 1Al In this article, we report our investi- compounds found in grapes that gations of another luxury item of might be sought in an analysis. Even many cultures: wine. This investiga- more specific is tartaric acid and its Organic analysis applied to archaeo- tion, in turn, has led to discoveries salts, which occur in large amounts logical remains, especially the con- about another, perhaps less presti- only in grapes. The survival of the tents of pottery vessels, has devel- gious but no less important, fer- latter compounds, however, is not as- oped rapidly over the past two mented beverage: beer. sured, especially in view of their fi- decades. Lipids, , dyes, per- nite water solubilities. fume ingredients, , and other Wine and archaeology Once a compound such as tartaric compounds-some less than a milli- The chemical detection of a fer- acid or bitartrate is iden- gram in weight and some thousands mented beverage such as wine or tified, there is still the problem of of years old-have been identitied by beer is a greater analytical and ar- how to distinguish between grape using a wide range of chemical and chaeological challenge than that of juice, grape , another fruit juice physical techniques. As these meth- royal purple (6, 6'-dibromoindigotin), (e.g., from dates or pomegranates) ods are further refined, archaeologi- which occurs only in certain marine that has been adulterated with grape cal chemistry promises to open up mollusks and is stable because of its juice, and wine. If written or picto- new chapters in human and environ- resonance structures. Wine is a pro- rial evidence exists, such as inscrip- mental history, including ethnicity cessed beverage made from grape tions on the sides of vessels detailing and genetic development, diet, dis- juice or must. It is an unusually their contents, interpretation is eas-

408 A * ANALMICALCHEMISTRY, VOL. 65, NO. 8, APRIL 15,1993 W03- 2700/93/0365-408Al$04.00/0 0 1993 American Chemical Society Vintugesrenefiom the tomb of Nukht, Western lhebes, New Kingdom. (Dauies, N., lhe Tomb ofNakht at lhebes; reprinted with pennission/mm the Metmpolitun Museum ofArf,New York.)

ier. For prehistoric periods, however, home to one of the oldest literate civ- graphic tablet, and other artifacts there is only the “mute” archaeologi- ilizations in the world the Sumerian were discovered within an oval com- cal record. From this-at best, only a and Elamite city-states. Here, clay plex of well-constructed symmetrical partial record-the researcher at- tablets incised with pictographic rooms enclosing a central courtyard, Downloaded by UNIV OF PENN on July 16, 2009 tempts to infer whether a pottery characters were unearthed. The the Citadel Mound complex. These vessel might have served a special growing population was supported findings, which were contemporary purpose or whether the presence of by irrigation agriculture and horti- with the earliest lowland city-states grape pips (seeds) in a certain con- culture of domesticated plants (e.g., of the Late period, can be un- text points to wine-making, con- , figs, and dates). Traders and derstood within the broader cultural

Published on September 22, 2008 http://pubs.acs.org | doi: 10.1021/ac00056a002 sumption of the whole fruit, or some adventurers exploited surrounding developments of the time. It is plau- other activity. areas for precious commodities such sible that Godin Tepe was an early as gold, copper, , carne- Sumerian and/or Elamite trading Wine-making in the highlands of lian, and timber, which were un- post, administrative center, and/or ? Archaeologists under the direction of T. Cuyler Young, Jr., of the in Toronto were ANALYTICAL faced with a puzzling set of evidence during their excavations of Period V APPROACH remains at Godin Tepe, located in the of western available in lowland Greater Meso- military base for protecting the trade Iran. Period V, dating between ap- potamia. route. proximately 3500 B.C. and 2900 B.C. The Zagros Mountains, bordering Other finds are more difficult to (the Late Udperiod), is contempo- lowland Greater , pro- understand. An unusually large fun- raneous with the beginnings of com- vided some of these raw materials nel, about half a meter in diameter, plex urban life as we know it. and were tied into the wider trading and a circular “lid” of smaller diame- Lowland Greater Mesopotamia, network. Godin Tepe is located along ter were recovered from one room of comprising the wide alluvial plains the High Road or the Great Kho- the Citadel Mound comnlex. McGuire of the lower Tigris and Euphrates rasan &ad (the famous Silk Road of Gibson of the Universiiy of Chicago’s rivers in modem Iraq and of Khuz- laler times) in the central Zagros Oriental Institute proposed that the istan in southwestern Iran, was Mountains. Pottery, an early picto- funnel, covered with a coarsely wo-

ANALYTICAL CHEMISTRY, VOL. 65. NO. 8. APRIL 15.1993 e 409 A ANALYTICAL APPROACH

ven cloth (long since deteriorated), appeared to have been a residence is an organic or inorganic residue. might have served as a strainer for where wine might have been con- DRIFTS has the advantage that the grape must. Such large funnels are sumed; the other room might have archaeological sample need not be known to have been used much later been used to distribute wine and damaged, and it eliminates the need in wine-making installations in the other goods to merchants, soldiers, for preparing an optically transpar- Middle East. It is possible that and others through two windows fac- ent potassium bromide wafer. The grapes were heaped into the funnel, ing the central courtyard. spectra from either technique will the lid wa8 used to apply weight and The jars with the red deposits are show the same absorption bands at measure. and the iuice was directed a unique pottery type that has a rope the same wavelengths. &to a ja;. design applied as two inverted U This was a Dlausible scenario. but shapes along opposite exterior sides it lacked har'd evidence. No &ape of each vessel. During the Late Uruk pips, which are often well preserved, period, rope designs often indicated were found in the room, 'and the fun- the placement of real rope. The nel might have been used for another shape and placement of the inverted liquid. The lid might have been just U suggested a means by which these that-a lid for a wide-mouthed ves- vessels could have been supported on sel. The wine-making hypothesis, their sides. In fact, the red deposits however, became more compelling on the interiors of one whole vessel when one of us (VRB) noted a red de- were confined to the base and the posit on the insides of several jars sidewall (Figure 3). The red residue (Figure 2) from two rooms that were was on the same side of the vessel aa directly across the courtyard from the applied rope decoration on the the room where the funnel and lid outside. The deposit was precisely were found. Could the red deposits where materials would have settled have been wine deposits? One room out from a liquid if the vessel had been stored on ita side. The inference that this jar once contained a liquid was supported by another stylistic peculiarity. These jars have relatively narrow mouths and tall necks, compared with other jars of the same period. A narrow- mouthed, tall-necked vessel would Figure 2. Godin Tepe wine jar, late be better suited for storing and pour- Period V. ing liquids. (Counesy William Pran. Royal Ontario Museum.) It also seemed likely that the jars had once been sealed, because two unfired clay stoppers-slightly smaller than the mouth and neck di- ameters of the jars-were found dis- Downloaded by UNIV OF PENN on July 16, 2009 carded in other mms of the complex. The clay could have functioned like a modern-day wine cork, absorbing liquid and expanding to keep out the oxygen, if the vessels had been laid

Published on September 22, 2008 http://pubs.acs.org | doi: 10.1021/ac00056a002 on their sides. Such precautions make sense if wine were the in- tended final product. Grape juice quickly ferments to wine. The trick is to stop it from be- coming , and this could have been accomplished by a technique still used today: stoppering the ves- sel and storing it on its side. ldentiflcation of tartaric acid Clearly, hypotheses about wine pro- duction and consumption at Godin Tepe stand or fall on whether the jars did indeed contain organic com- pounds derived from grapes. Transmission and diffuse-reflec-

~~~~~ ~~ tame FT-IR spectrometry (DRIFTS) Figure- 1. Compounds found inI wine. are versatile techniques that are Flgure 3. Red deposit (da ~ ~ ~ ~, visible on the interior base and one (a) Delphinidin. an anthocyanidin. (b) Quercetin. particularly useful in an archaeolog- (c) Catechin-3-gallate. (d) Elaglc add. which Is ical chemical investigation when one side of the Godin Tepe jar in Figure 2. eslenfied with m yiek elagaannins. must determine whether the sample (Courtesy William Prm. Royal Ontario Museum.)

410 A * ANALYTICAL CHEMISTRY, VOL. 65, NO. 8, APRIL 15,1993 ~~

The DRIFT spectrum of the sur- gest that the carboxylic acid in the to convert tartaric acid to a com- face scrapings of the interior red de- extracts is tartaric acid. For compar- pound that exhibits green fluores- posit of one Godin Tepe jar revealed ison, we examined the spectrum of cence under W light. Samples from the presence of organics but was the naturally occurring L-(+)isomer the Godin Tepe jars and the Nubian dominated by silicates. More defini- of tartaric acid (Figure 5d). Because , as well as synthetic I,-(+)- tive results were obtained by ex- the three stereoisomers (Figure 6) tartaric acid, all gave the same char- tracting several sherds with boiling have very similar DRIFT spectra, ra- acteristic green fluorescence. acetone. The sample was evaporated cemization of the unknown would Another unanswered question is to , yielding 6-7 mg of - not affect the results. Most signifi- why tartaric acid is present prima- ous solid for both a 17-g sample cant for comparison are the bands at rily as the free acid and not, to any (with a visibly thicker deposit) and a 1445 cm-’ and 1250 cm-’ as well as considerable extent, as its salt. The 73-g sample (with a thinner deposit). the doublet peaks at 1740 and 1720 free acid is more soluble in water Because whole sherds were ex- cm-’, all of which exist because of than the monopotassium salt (the tracted, it is not known whether this carboxylic acid groups in the mole- form in which it crystallizes from relatively small amount of solid de- cule. These values are very close to wine) and would be expected to dis- rives from the reddish deposit, from those of the unknown deDosit and solve and disappear through ground- the interior of the vessel, or from extracts. water percolation. One might conjec- both areas. A control sample from a Althoueh the DRIFT analvsis”. Dro- ture that this dihydroxy dicarboxylic deposit-free area of one jar, however, vides strong evidence for tartaric acid was strongly adsorbed on sili- showed no organics. acid, it does not unambiguously cates by hydrogen bonding. Thus it For a reference sample of an an- identify the compound. Separation could be removed from the hydroly- cient wine residue, a dark-colored and characterization of tartaric acid sate of the bitartrate and preserved deposit from a storage jar or am- by chromatography, followed by in the pottery matrix. The equilib- phora (Figure 4) was chosen from the spectroscopic analysis, is one way to rium between the acid and the salt ancient Nubian site of Gebel Adda, gain greater certainty. Although fu- would shift to the acid side as more now submerged by the of ture investigations may provide such acid was bound up in the ceramic. Lake Nasser behind the Aswan Dam. confirmation, a simpler approach is There was little doubt that this ves- to test the unknown mixture with a implications for wine-making sel once contained wine, because it highly specific Feigl spot test for tar- The chemical authentication of tar- belongs to a special pottery type. taric acid. taric acid in the Godin Tepe jars Such jars were imported in quantity In this test, P,V-dinaphthol and agrees with the archaeological sce- from Upper Egypt into Nubia during concentrated sulfuric acid are used nario that the jars originally con- the late Byzantine period (fourth to tained a grape product. But although sixth centuries A.D.). Whole villages the presence of tartaric acid is the in Nubia were composed largely of crucial link in the chain of argu- , and empty amphorae, like ments, it is insufficient for establish- the one tested, were piled in heaps ing whether the grape product was a around the taverns. liquid and, if so, whether it was fer- The DRIFT spectra of the Godin mented to wine. For these infer- Tepe jar extracts (Figure 5a and 5b) ences, we must rely on the archaeo- and the ancient Nubian reference logical evidence. The available data

Downloaded by UNIV OF PENN on July 16, 2009 sample of wine (Figure 5c) have (e.g., the narrow mouths and elon- many similarities, including an in- gated necks of the jars and the red tense C-H peak around 2900 cm-’; a deposits that had formed from mate- sharp, intense carbonyl peak, with a rials settling out) point to the jars shoulder, at 1720-1740 cm-’; and having been filled with liquids. Un-

Published on September 22, 2008 http://pubs.acs.org | doi: 10.1021/ac00056a002 carboxylic acid-related bands be- der normal conditions and at room tween 1385 and 1470 cm-’ and 1240 temperature, grape juice easily and and 1250 cm-’. Of particular interest quickly ferments to wine. is the 1720-1740 em-’ carbonyl Grape must proceeds band, which is at the upper end of rapidly because a variety of wild the range for carboxylic acids (1680- , primarily the genera Saccha- 1740 em-’) and is characteristic of romyces and Candida, are present in carboxylic acid groups located near the “bloom” of the grape skins. The strong electron-withdrawing groups, ability of yeasts to ferment grape such as halogens and carbonyls. In sugars depends on the presence of addition, the bands at 1600-1650 specific . Oxygen respiration cm-’ suggest the presence of carbox- is required for the growth and multi- ylic acid salt(s). The hydroxyl bands plication of the cells and conse- at 3400-3500 cn-’ are broad and re- quently for the course and degree of sult from the presence of several fermentation. The continued avail- types of hydroxyls. The strong band ability of oxygen after fermentation at 2900 em-’ and the numerous ab- has ceased leads to the multiplica- sorption bands from 1550 to 800 tion of the bacteria (Acetobacfer)that em-’ indicate a variety of organics, are responsible for the conversion of possibly some of those listed above to . The precau- as wine or grape components. Figure 4. Nubian wine amphora tions that were taken in storing the The DRIFT results strongly sug- (Courtesy William Pran, Royal Ontario Museum.) liquid contained in the Godin Tepe

ANALYTICAL CHEMISTRY, VOL. 65, NO. 8, APRIL 15.1993 411 A ANALYTICAL APPROACH

jars-stoppering the vessels and lay- places. Even the development of spe- vessel served as a beer container. ing them on their sides-makes the cialized pottery vessels and storage is a principal most sense if this liquid were wine. techniques for wine-making were component of ‘%beerstone,”which set- Air was prevented from entering and not beyond the expertise of village or tles out at the bottom and along the nourishing the bacteria that convert small-scale household vintners. sides of beer fermentation and stor- wine to vinegar. Although the Godin Tepe jars very age tanks, and it characterizes - The actual processing of grapes probably contained wine, it is still ley beer. This salt is soluble in water into wine need not be very complex, uncertain whether the beverage was to only 6 mglL at 18 “C, and the as the textual and pidorial evidence produced on site from grapes grown compound would be expected to accu- for wine-making from the ancient locally. The evidence for a local wine- mulate on the inside of a storage world clearly shows. In Egyptian making installation is only a large vessel, especially one with a greater wine-making, the grapes were tread funnel and a putative lid for pressing surface area caused by incising. on in crushing vats to extract the the grapes. Nevertheless, if one were A Feigl spot test (which is sensi- juice or must. A bag press was also searching for an early center of wine tive to IO-‘ g) provided definitive ev- used to squeeze out any juice re- production, Godin Tepe-centrally idence that the pale yellowish depos- maining in the solid residue. Pottery located on a major trade route in the its in the incisions of the ancient jars were then filled with the juice, central Zagros Mountains-would be Iranian sherds contained an oxalate, which quickly fermented to wine by an excellent place to begin. The peo- most probably . Re- natural processes. The jars were ple from the Transcaucasian , duction of the oxalate in an acidic completely stoppered only after most approximately 600 km to the north medium to glyoxalic acid, followed by of the fermentation had occurred. where the wild grapes still grow to- reaction with phenylhydrazine and A similar series of steps was prob- day, likely were in contact with Go- hydrogen peroxide, gives a distinc- ably used to produce the wine con- din Tepe in Period V. The emerging tive pinkish red color. Another an- tained in the Godin Tepe jars and, cities of lowland Greater Mesopota- cient beer sample, as well as pure indeed, this process of wine-making mia, especially their elite classes, calcium oxalate, were run as con- might well have been discovered at represented a large potential market trols. Scrapings from an Egyptian many times and in many different for wine. The entrepreneurial and New Kingdom blue-painted “beer- experimental spirit of the period, ,” belonging to a type used in a quite evident in the extensive trade “ and beer” ritual and clearly connections, urban development, ag- intended for beer (according to tomb riculture, and horticulture of the paintings and reliefs), tested posi- lowlands, might well have been ap- tive. We also tested modern beer- plied to viticulture and viniculture in stone from a brewer’s vat, which the highlands. gave a pinkish red color. In addition to beerstone, Another fermented beverage occur naturally in relatively large Recently, one of us (VRB) observed a amounts (- 5-10% by fresh weight) curious crisscross pattern of long in- in spinach and rhubarb, species of cisions on the interior of a Late Uruk which grow in the Iranian high- double-handled jug from one of the lands today. Compared with cere- same moms containing the wine jars als, however, these plants make up

Downloaded by UNIV OF PENN on July 16, 2009 at Godin Tepe. The incisions were a minor part of the human diet, and filled with a pale yellowish residue. it is unlikely that they or their de- The early Sumerian sign for beer rived products were stored or pro- shows a pottery vessel with internal cessed in the ancient incised vessel. linear markings. Together, the ar- Minor sources of or ox- chaeological and pictographic evi- alate salts, which are widely distrib- Published on September 22, 2008 http://pubs.acs.org | doi: 10.1021/ac00056a002 dence strongly suggests that this uted in nature, were ruled out. Feigl

dgure 5. Spectra of extracts and tartaric acid. (a. b) DRIFT spectra of Godin Tepjar extra& mixed with KBr. (c) DRIFT spectrum of Nubian amphora wine deposit mixed with KBr. (d) IR transrnision spearurn of 1-(+)-tartaric acid. Figure 6. Stereoisomers of tartaric acid,

412 A - ANALYTICAL CHEMISTRY, VOL. 65, NO. 8, APRIL 15,1993 spot tests of exterior scrapings of the vessel and from the interior of one of the wine jars gave negative results. The discovery of oxalate inside the Godin Tepe jug confirmed the ar- chaeological and pictographic evi- dence that the vessel was a beer con- tainer. This result should not be surprising, given that the site was in close contact with lowland Greater Mesopotamia, where beer was the preferred . The beer at Godin Rudolph H. Michel received his B.S. de- Tepe was likely made from locally gree from the City College of New York grown because carbonized six- and his Ph.D.from the UnivenityofNotre row barley predominated in the ar- Dame. In 1985, aper retiringfrom E. I. chaeobotanical material recovered du Pont de Nemoun as a Research Fel- from Period V. low, he volunteered his time and consider- Wine and beer connoisseurs may able experience to begin a new career in disagree about many matters, but archaeochemical research at MASCA. neither group can as yet claim that Maqnetic wine or beer was the earliest fer- mented beverage. The inhabitants of Godin Tepe in the Late R&nance of obviously appreciated the nutritional and psychotropic benefits of both Carbonaceous . Solids We thank T. Cuyler Young, Jr., and Edward J. Ked of the Royal Ontario Museum in Toronto ocusing on new magnetic resonance for pmviding samples fmm the Godin Tepe ves- methods for characterizingsolid sels, and Nicholas B. , also of the Royal F Ontario Museum. for providing samples from :arbnaceous fuels, this new volume the Egyptian vessels. A modern sample of beer- covers 2-D methods. high-resolution I3C stone was scraped out of a fermentation vat at Patrick E. McCovern received his A.B. de- and 'H NMR (CRAMPS) analyses. the Deck Street Company of Philadel- gree in chemisttyfrom Cornell Uniuenity phia, PA. BNW Chase and his staff.in partieu- dynamic nuclear polarization tech- lar N. Rapposelli and E. Matthews. at the Er- and his Ph.D. in Near Eastern archaeol- niques. temperature-dependent perimental Station of The Du Pont Co. in ogy and literaturefrom the University of relaxation measurements, chemical Wilmingon, DE, ran the DRIFT and transrnis- Pennsyluania. He joined MASCA (IS a re- methods to improve quantitative sion spectra. Vernon Singleton, professor of enology at the University of California-Davis, search scientist in 1979. He engages in a reliability, ultrafast MAS methods. and kindly advised us on the eharaeteristie compo- wide range of research in archaeological new signal enhancement techniques. nents of wine and beer as well as the appmpri- chemistry, including radiocarbon dating, Also discussed are topics of current ate analytical techniques for their identifiea- cesium magnetometer surveying, colorant interest in the area of ESR. including

Downloaded by UNIV OF PENN on July 16, 2009 tion. Andrew Waterhouse of the University of analysis of ancientglmes, pottety technol- California-Davis recently consulted with us on spin-echo methods 1-D and 2-D pulsed ehmmatographie analyses. The barley from Pe- ogy, and organic analysis of vessel con- ESR techniques. relaxation time riod V st Godin Tepe was identified by Nsomi tents and dyes. measurements. and ultrahigh frequency F. Miller of MASCA. measurements. In addition. measurements are made Suggested reading Published on September 22, 2008 http://pubs.acs.org | doi: 10.1021/ac00056a002 on a set of Argonne Premium Coal Orgnntc Contenb of Ancient Vessels: Materials samples to compare different magnetic Analysis and Archaeological Investigation; resonance techniques on a homogenous Biers. W. R.: MrCavern. P. E., Eds.: MASCA, The University MuseuG of Ar: set of carbonaceous fuel samples. chncology and Anthropology. Philadel- hia, 1990, MASCA Research Papers In Robert E Botto. Argonne National cirnce and Archaeoluev. Vol. 7 Note LalkJratory, Editor esmciallv the article."'Drink and Be Yuzo Sanada. Hokkaido University,Editor hl;rry. Infrared Spectroncapy and An- cient Near Eastern Wine." p 11.f Advances in Chemistry Series No. 229 Farka5. J. Terhnology and Ll~ochemtstryof 666 pages (1993)Clothbound IVtne: Gordon and Breach Scrence: New ISBN 0-8412-1866-8 York; 1988; 2 vols. Virginia R. Badler received her B.A. de- $149.95 Feigl, F. Spot Tests in Organic Applications; Elsevier: New York, 1954; pp. 260-61, gree in 1970from the University of Cali- *e" a0 OYI-OY. fornia-Santa Barbara and her M.A. de- American Chemical Society Lesko. L. €I.Kina Tut's Wine Cellar: B. C. gree in 1972 from the University of DiStrlbUtiOn Office Dept 56 Toronto, both in art histoty. Currently she 1155Sixteenth Street. NW is working on her F'h. D. in West Asian ar- Washington. IX 20036 chaeology at the University of Toronto. Or CALL TOLL FREE Her resenrch interests include typological, functional, and chemical analysis ofpot- 800-227-5558 (In Washington. Dc. 202-872-4363) and use tety, as well (IS computer reconstruction of Your credit card! New York,'1993 old and new mcavations.

ANALYTICAL CHEMISTRY, VOL. 65, NO, 8, APRIL 15,1993 413 A

Top View