ROMAN COTTON REVISITED J. P. Wild,* F. C. Wild ** and A. J. Clapham*** I WIPSZYCKA 1965: 40-1) and karpasos (Periplus Maris Erythaei 41; ERNOUT and MEILLET 1967: 99), but these can cover other fine vegetable fibres, too. Eventually, in koine Greek the clumsy but unambiguous noun erioxylon, «tree wool», emerges. Theophrastus (HP IV, 7, 7-8) and later Pliny (NH XII, 10-11 (21-22)) report that cotton was grown on a large scale in the islands of the Bahrein group in the Gulf, and finds of seeds and textiles there confirm this (M.TENGBERG, pers. com.; HAERINCK 2002). Indeed, one wonders whether Bahreini cotton was the inspiration behind Sennacherib’s famous experiments with cotton cultivation in his palace garden at Nineveh just after 700 BC (TALON 1986; POTTS 1997: 270-272; TORAY INDUSTRIES 1996: 199-206). Cultivation of cotton in Egypt is referred to in very general terms in the second-century AD by the lexicographer Pollux (VII, 75), who was born in Naucratis. Pliny (NH XIX, 1, 3) mentions its cultivation «on the Arabian side of Upper Egypt», a curious expression given the presence of the Eastern Desert; but he adds Aethiopia (NH XIII, 90), a term which in his day included Nubia and the upper reaches of the Nile as well as Ethiopia proper. More precise information is provided by the documents – papyri, ostraca and wooden boards – found in recent excavations in the oases of Dakhleh and Kharga in the Western Desert. The prime document is a set of fourth-century agricultural accounts written on wooden boards and N this paper we return briefly to the subject of cotton in the Roman world, and specifically to the question of the geography of Roman cotton cultivation. The first writer raised the topic of cotton initially in 1997 in a Festschrift for Hideo Fujii (WILD 1997), but a great deal of fresh evidence has become available since then, especially in the field of archaeobotany. There are three principal sources of information: ancient written documents, finds of cotton textiles and the palaeobotanical record. Much of the recent evidence is unpublished, and we could have made little progress without the help of, and information from, many colleagues, most notably those whose forthcoming work is cited in our bibliography. We are grateful to them all. The written evidence gives us a starting point, and we can draw on the archaeological and palaeobotanical sources as they become relevant to the discussion and offer new and independent insights. There is no specific word for cotton or the cotton plant in Classical Greek. Herodotus (III, 47; III, 106; VII, 65,1) calls it «wool from trees» (eiria apo xylou), with reference to India. The botanist Theophrastus around 320 BC says that the wool-bearing trees resemble a wild rose, with a leaf like a sycamore or vine leaf. He notes that in India cotton plantations look like vineyards from a distance (HP IV, 4, 8; IV, 7, 7-8). Contemporary and later authors use the term byssos (HORNBLOWER and SPAWFORTH 1996: 266; * J.P. Wild, Dept. of Archaeology, University of Manchester. ** F.C. Wild, Archaeological Consultant, Stockport. *** A.J. Clapham, Consultant Palaeobotanist, University of Cambridge. PURPUREAE VESTES. II Symposium Internacional sobre Textiles y Tintes del Mediterráneo en el mundo antiguo (C. Alfaro, L. Karali, eds.), pp. 143-147. 143 144 Fig. 1. Map of the Dakhleh, Kharga and neighbouring oases in the Western Desert of Egypt, with access route from Assyut on the Nile (drawn by J.P. Wild). found in a house at Kellis, Asmant el-Kharab in the Dakhleh Oasis (Fig. 1). They record weights of raw cotton paid as dues in kind by a number of peasant farmers with Egyptian names (P. Kell. I, Gr. 61 (WORP 1995); P. Kell. IV, Gr. 96 (BAGNALL 1997)). Two ostraca also refer to crops of cotton (BOWEN 2002: 87). The weight unit cited is called a lith(os), a stone. We cannot calibrate it, but it could be divided into small fractions and Roger Bagnall (1997, 50-51) thinks that it might be over 3 kg, in other words at least 600 cotton bolls. Lith(os) seems only to have been applied to cotton, suggesting that the industry already had its own vocabulary. One lith(os) is said to be «for yarn» (P.Kell. IV, Gr. 96, 558-560; MOSSAKOWSKA-GAUBERT, forthcoming). A clutch of late Roman ostraca from Kysis, Dush in the Kharga Oasis, also give weights of raw cotton measured in lith(oi) (O.Douch IV, 381, 489; V, 596, 600, 634 (WAGNER 2001). One ostracon names women preparing or spinning it (O.Douch I, 51 (CUVIGNY and WAGNER 1986). Cotton was clearly an established crop in both oases by the fourth century (BOWEN 2002: 87). The earliest dated document is a papyrus of c.AD 164-5 from Kharga, which refers to a plantation of olive trees, date palms and cotton plants (P.Iand. VII, 142, II, 8 (KALBFLEISCH 19121938). J. P. Wild, F. C. Wild and A. J. Clapham The written sources are now strongly supported by palaeobotanical evidence in both oases. At Kellis, Dakhleh Oasis, both seeds and cotton bolls have been found regularly in stratified late Roman deposits in domestic structures (THANHEISER and BAGNALL 1997: 39-40; THANHEISER 1999: 91; THANHEISER, WALTER and HOPE 2002: 299-310), indicating that cotton processing was an everyday, presumably seasonal, activity. Spinning, followed by weaving on the warpweighted loom, is attested by spindle-whorls and sunbaked mud loomweights in the same contexts (BOWEN 2002, 87). Recent survey in North Kharga has also brought cotton seeds to light (CLAPHAM, unpublished). There are a few finds of cotton cloth in both oases (Kellis: COOMBS, WOODHEAD and CHURCH 2002; BOWEN 2002: 87; Kysis: JANA JONES, pers.com.); but they take on greater significance when set against the lack of such finds in metropolitan Egypt. However, the documentary papyri indicate that cotton clothing was neither expensive nor rare in Roman Egypt (WILD 1997: 289). There appears to be a gap in the archaeological record here. Outside Egypt the oasis at Jericho in Palestine, watered by the natural spring of Ain Sultan, seems to have been another centre of cotton cultivation, although there is no firm evidence until Gregory of Tours writes about it in c. AD 575 (Liber in Gloria Martyrum 17). He says that he has seen the cottonwool brought back by pilgrims. Interestingly, cotton bolls embedded in beeswax have been found inside relic buckles buried with clerics at Augsburg in Bavaria and at Monnet-la-Ville in eastern France at just that date (WERNER 1977: 173, Taf. 67, 97, 1; 375400). There are also references to local cotton growing in Rabbinic literature (Kilaim 7, 1-2; DANBY 1936: 36). As a crop, cotton was well suited to the agricultural regime of an oasis, as others have commented. It is a thirsty plant, requiring a minimum of 500mm of rainfall or its equivalent by frequent irrigation (LANGER and HILL 1991: 258-263; TOTHILL 1948: 324340). There is ample documentation for irrigation systems and their careful supervision in both the Dakhleh and Kharga oases (CUVIGNY, HUSSEIN and WAGNER 1993, 20-26). In Kharga, moreover, there is archaeological evidence for both well-fed pipelines (WAGNER 1987: 282 n.5) and for a foggara or qanat network of underground channels (WUTTMANN 2001). A foggarabased irrigation system has also been recognised in the Wadi al-Ajal in the Fazzan of Southern Libya (DRAKE et al. 2004). It is gratifying that Ruth Pelling has now identified cotton seeds in Roman-period contexts there (PELLING 2005: 402, 404, 406). So far, no cotton textiles have come to light. Roman cotton revisited 145 Fig. 2. Findspots of cotton textiles in the western Roman provinces (drawn by J.P. Wild). Fig. 3. Findspots of cotton textiles in the eastern Roman provinces (drawn by J.P. Wild). Can finds of cotton textiles mapped on an empirewide basis help us to pinpoint any more production centres (WILD 1997; LORQUIN and MOULHERAT 20012)? The finds from the western half of the Empire are shown in Fig. 2. Excluded from the map are finds about which we have doubts. Most finds here reflect personal or at least non-commercial importation. A typical example comes from a fourth-century coin hoard at Pachten in the Saarland (WILD 1969), where the purse containing the coins was probably made of recycled cotton cloth. The eastern provinces by contrast, as one might expect, are far richer in cotton (Fig. 3). Palmyra, where a small but measurable proportion of the total textile assemblage was of cotton, lay on a famous caravan route, so the presence of exotica is no surprise (STAUFFER 1995: Abb.100; MILLAR 1998). The graves at At-Tar near Karbala, just west of the lower Euphrates, were also furnished with a number of cotton fabrics (FUJII 1980). There is a cluster of findspots around the Dead Sea, but the total number of items is small. The concentration of finds, however, on the Nile in Lower Nubia is striking, and even more striking are the sheer numbers of cotton finds at each site and their dominance of the textile corpora: at Qasr Ibrim in the first four centuries AD 70-80% of the textiles are of cotton (DRISKELL, ADAMS and FRENCH 1989) and at Gebel Adda, a nearby cemetery, some 75% are cotton (WATSON 1977: 357, 364). Qasr Ibrim is rich in palaeobotanical evidence, currently being examined by Alan Clapham (CLAPHAM and ROWLEY-CONWY forthcoming). Numerous cotton bolls and seed capsules have been recovered from the site in a remarkable state of preservation (Fig. 4). In almost every deposit dated to the first few centuries AD there are crushed cotton seeds with lint still attached; there is S-spun cotton yarn, and even loaded spindles. Cotton plantations cannot have been far away, and new C-14 dates indicate that they were in existence by the early first century AD. The date at which irrigation in this region by means of the saqia water-lifting wheel was developed is still debated, but current views put it later than the first century (EDWARDS 2004: 202-203). At Berenike, a Roman port on the Red Sea coast, roughly half of the textile corpus by the fourth century AD was of cotton (WILD and WILD 2000). Of that, some 50% was probably of Nubian origin, marked by very strongly S-spun yarns. But another 50% was woven exclusively from Z-spun yarn. We regard this as intrusive – but intruded from where? Some pieces, like the blue resist-dyed fabrics and the sailcloth, are surely Indian imports, as we have argued elsewhere (WILD and WILD 2005; WILD and WILD 2001; F.C. WILD 2002). Interestingly, the other major Red Sea port at Quseir/Myos Hormos has also yielded Z/Zspun cotton sailcloth (WILD 1997: 290; HANDLEY 2004). 146 Fig. 4. Roman-period cotton bolls from Meroitic Qasr Ibrim, Lower Nubia (photo by J.P. Wild). It may be too hasty, however, to regard all Z-spun cotton as of Indian origin. What is one to make of a fustian from Qasr Ibrim (QI05 0260), for instance, with flax warp spliced S in the Meroitic style and flattened Z-spun cotton weft? Is it a local product? Less than 10% of the Roman-period Meroitic cottons which we have recorded recently at Qasr Ibrim are of Z-spun yarn, but an Indian connection to this remote site would be very hard to argue on any other grounds. Should we be looking for another Z-spinning cotton-rich area or areas? It is now obvious that spin direction alone will not answer questions of origin. Much more specific characterisation of the fabrics will be required, with the problems of group definition in mind. But that is a task for the future. Bibliography BAGNALL, R.S. (1997): The Kellis Agricultural Account Book, Dakhleh Oasis Project Monograph 7, Oxbow Monograph 92, Oxford. BOWEN, G.E. (2002): «Textiles, basketry and leather goods from Ismant el-Kharab» in HOPE, C.A.and G.E.BOWEN (eds): Dakhleh Oasis Project: Preliminary Reports on the 1994-1995 to 1998-1999 Field Seasons, Oxford, 87-104. CLAPHAM, A.J. and P.A. ROWLEY-CONWY (forthcoming): «The archaeobotany of cotton (Gossypium sp.L) in Egypt and Nubia with special reference to Qasr Ibrim, Egyptian Nubia». COOMBS, C.E., A.L WOODHEAD and J.S. CHURCH (2002): «Report on the characterization of three fabric samples from Ismant el-Kharab» in HOPE, C.A. and G.E. BOWEN J. P. Wild, F. C. Wild and A. J. Clapham (eds): Dakhleh Oasis Project: Preliminary Reports on the 1994-1995 to 1998-1999 Field Seasons, Oxford, 115-119. CUVIGNY, H., A. HUSSEIN and G. WAGNER (1993): Les Ostraca grecs d’Aïn Waqfa - Oasis de Kharga, Documents de Fouilles XXX, IFAO, Cairo. CUVIGNY, H. and G. WAGNER (1986): Les Ostraca grecs de Douch, Fascicule I, Documents de Fouilles XXIV, IFAO, Cairo. DANBY, H. (1936): The Mishnah, Oxford. DRAKE, N., A. WILSON, R. PELLING, K. WHITE, D. MATTINGLY and S. BLACK (2004): «Watertable decline, springline desiccation and the early development of irrigated agriculture in the Wadi al-Ajal, Libyan Fazzan», Libyan Studies 35, 95-112. DRISKELL, B.N., N.K. ADAMS and P.G. FRENCH (1989): «A newly discovered temple at Qasr Ibrim: preliminary report», Archéologie du Nil Moyen 3, 11-55. EDWARDS, D.N. (2004): The Nubian Past: An Archaeology of the Sudan, London. ERNOUT, A. and A. MEILLET (1967): Dictionnaire de la Langue Latine, 4th ed., Paris. FUJII, H. (1980): «Special Edition on the Studies on Textiles and Leather Objects of Al-Tar Caves, Iraq», AlRa–fida–n I. HAERINCK, E. (2002): «Textile remains from Eastern Arabia and new finds from Shakhoura (Bahrain) and ed-Dur (Umm al-Qaiwan, UAE)», Arabian Archaeology and Epigraphy 13, 246-254. HANDLEY, F.J.L. (2004): «Quseir al-Qadim 2003: the textiles», Archaeological Textiles Newsletter 38, 27-30. HORNBLOWER, S. and A. SPAWFORTH (1996): The Oxford Classical Dictionary, 3rd ed., Oxford. KALBFLEISCH, C. (1912-1938): Papyri Iandanae, Leipzig/ Berlin. LANGER, R.H.M. and G.D. HILL (1991): Agricultural Plants, Cambridge. LORQUIN, A. and C. MOULHERAT (2001-2): «Corpus des vestiges de coton recensés pour l’antiquité tardive en Syrie, Egypte, Nubie et au Soudan», Aquitania XVIII, 186-192. MILLAR, F. (1998): «Caravan cities: the Roman Near East and long-distance trade by land» in AUSTIN, M., J. HARRIES and C. SMITH (eds): Modus Operandi: Essays in Honour of Geoffrey Rickman, Bulletin of the Institute of Classical Studies Supplement 71, London, 119-137. MOSSAKOWSKA-GAUBERT, M. (forthcoming): «Quelques expressions grecques liées à l’aspect technique de la production des tuniques en Égypte» in MATHIEU, B., D. MEEKS and M. WISSA (eds): L’Apport de l’Égypte à l’Histoire des Techniques, IFAO, Bibliothèque d’Études 142, Cairo 2006. PELLING, R. (2005): «Garamantian agriculture and its significance in a wider North African context: the evidence of the plant remains from the Fazzan Project», Journal of North African Studies 10, 397-411. POTTS, D.T. (1997): Mesopotamian Civilisation: The Material Foundations, London. STAUFFER, A. (1995): «Kleider, Kissen und bunte Tücher: Einheimische Textilproduktion und weltweiter Handel» Roman cotton revisited in SCHMIDT-COLINET, A. (ed): Kulturbegegnung im Grenzbereich, Mainz, 57-71. TALON, P. (1986): «Le coton et la soie en Mesopotamie?», Akkadica 47, 75-78. THANHEISER, U. (1999): «Plant remains from Kellis: first results» in HOPE, C.A. and A.J. MILLS (eds): Dakhleh Oasis Project: Preliminary Reports on the 1992-1993 and 1993-1994 Field Seasons, Oxford, 89-93. THANHEISER, U. and R.S. BAGNALL (1997): «Crops» in BAGNALL, R.S. (ed): The Kellis Agricultural Account Book, Dakhleh Oasis Project Monograph 7, Oxbow Monograph 92, Oxford, 35-45. THANHEISER, U., J. WALTER and C.A. HOPE (2002): «Roman agriculture and gardening as seen from Kellis» in HOPE, C.A. and G.E. BOWEN (eds): Dakhleh Oasis Project: Preliminary Reports on the 1994-1995 to 1998-1999 Field Seasons, Oxford, 299-310. TORAY INDUSTRIES (1996): «Report on the analyses of textiles uncovered at the Nimrud tomb-chamber», AlRa–fida–n XVII, 199-206. TOTHILL, J.D. (1948): Agriculture in the Sudan, London. WAGNER, G. (1987): Les Oasis d’Egypte à l’Epoque grecque, romaine et byzantine d’après les Documents grecs: Recherches de Papyrologie et d’Epigraphie grecques, Cairo. WAGNER, G. (1999): Les Ostraca grecs de Douch, Fascicule IV (356-505), Cairo. WAGNER, G. (2001): Les Ostraca grecs de Douch, Fascicule V (506-639), Cairo. WATSON, A.M. (1977): «The rise and spread of Old World cotton» in GERVERS, V. (ed): Studies in Textile History, Toronto, 355-368. 147 WERNER, J. (1977): Die Ausgrabungen in St. Ulrich und Afra in Augsburg 1961-1968, Münchner Beiträge zur Vor- und Frühgeschichte 23, München. WILD, F.C. (2002): «The webbing from Berenike: a classification», Archaeological Textiles Newsletter 34, 9-16. WILD, F.C. and J.P. WILD (2001): «Sails from the Roman port at Berenike, Egypt», International Journal of Nautical Archaeology 30 (2), 211-220. WILD, J.P. (1969), note in P.R.FRANKE: «Pachten: Der älteste Baumwollfund nördlich der Alpen (300 n.Chr.)», 16. Bericht der Staatlichen Denkmalpflege im Saarland, 161-163. WILD, J.P. (1997): «Cotton in Roman Egypt: some problems of origin», Al-Ra–fida–n XVIII, 287-298. WILD, J.P. and F.C. WILD (2000): «Textiles» in WENDRICH, W.Z. and S.E. SIDEBOTHAM (eds): Berenike 98: Report of the 1998 Excavations at Berenike and the Survey of the Egyptian Eastern Desert including Excavations in Wadi Kalalat, Leiden, 251-274. WILD, J.P. and F.C. WILD (2005): «Rome and India: early Indian cotton textiles from Berenike, Red Sea coast of Egypt» in BARNES, R. (ed): Textiles in Indian Ocean Societies, Oxford, 11-16. WIPSZYCKA, E. (1965): L’Industrie textile dans l’Egypte romaine, Wroclaw. WORP, K.A. (1995): Greek Papyri from Kellis I, Dakhleh Oasis Project Monograph 3, Oxbow Monograph 54, Oxford. WUTTMANN, M. (2001): «Les qanats de ’Ayn-Manawir» in BRIANT, P. (ed): Irrigation et Drainage dans l’Antiquité‚ Qanats et Canalisations Souterraines en Iran, en Egypte et en Grèce, Paris, 109-135.