,) ",) Ventris, M. and Chadwick, 1. 1973 Documents in Mycenaean Greek. Second editon. Cambridge University Press. Cambridge. Vermeule, E. T. 1967 A Mycenaean Jeweller's Mould. Bulletin of the Museum of Fine Arts, Boston Vo1.LXV, Nr.339: 19-31. Wiener, 1. 1983 Glass Finds and Glassmaking in Mycenean Greece, An Archaeological Study. Dissertation Tilbingen 1983. Los Angeles. Xenaki-Sakellariou, A. 1985 Les tombes a chambre de Myc(mes. Fouilles de ChI', Tsountas (1887-1898). Diffusion de Boccard, Paris, ?o.hick HcGu~(etbb) : Q."7Lc.hi.s-lmy l1,e His-by oC G la.s-!fko.I<i1'l3 leckna~>, GLASS COLORING WORKS WITHIN A COPPER-CENTERED INDUSTRIAL COMPLEX IN LATE BRONZE AGE EGYPT Th. Rehren Deutsches Bergbau-Museum 44787 Bochum, Germany E.B. Pusch and A. Herold Pelizaeus-Museum 31134 Hildesheim, Germany fLrnmiCS' ONJ.. G·~L.2akO\.. Vd... 1lIrt \qq8 ~e AmetlcCl.. Gm.mIL So6eJ.s ABSTRACT Excavation at Qantir, Nile Delta has provided insight into organization and development of an industrial estate of Ramesside Egypt. In founding the new capital of Egypt. Piramesses, during the 19th Dynasty. a huge bronze-casting factory was built. accompanied by a range of other. non-metallic high-temperature industries. Besides abundant production of faience implements, coated with copper-colored glazes, and the manufacturing of Egyptian blue. the coloration of large quantities of red glass also played a major role. In addition. cold working of a wide range of organic and inorganic materials on the site is evident. All crafts appear higWy integrated. probablY entailhlg a sophisticated assembly pattern. A shift in emphasis with time is apparent from changittg installations within-a-few decades. '. / "The production of colored glass-isattested by numerous crucibles. mostly with adhering traces of red glass. and El complete one liter ingot. Blue glasses of different shades are present. but much less abundant{, While evidence of glas~ working by artisans is absent •.there is indication that both the production of raw glass and glass coloring took place. The nature and complexity of high temperature industrial debris found at Qantir suggest a highly specialized labor organization within a framework of shared technologies and skills of closely controlled temperature and redox conditions. Thisucross-craft workshop pattern further reveals a significant level of intra-craft specialization and even spatiaL separation of glass making, coloring and finally working in Late Bronze Age Egypt. To IhI! .'I,nl authorized of The American lne e,p"" 226 Prehistory and History of Glassmaking Technology Cer,mic under th.law, Society, of lne Uniled Stal" Any dupJieouon, wTiuen con<cm of The Am,dc.n Ceramic oi America. reproduClio,. aH copyright or republication inler"ls in this publication of Ih;, publicauon S,,,ieIY or iee paid 10lhe ulpyrighl Clearance Prehistory and History of Glassmaking Technology 'u. Ih, or any parllh".uf. Cemer. propeny wllhonl is prohibited. 227 'j 'j INTRODUCTION Since 1980, the Pelizaeus-Museum, Hildesheim, has carried out excavations in the northern part of ancient Piramesses, the capital of Egypt during the Ramesside period, in the eastern Nile Delta. The excavation is part of a joint mission with the Austrian Archaeological Institute, Cairo, working at Tell cl Dab'a, and extends upon areas excavated earlier this century by Mahmoud Hamza in 1928, Ahmed Abdel-Salam Ahmed in the 1930ies and Labib Habachi in the 1940ies. The major concern of the ongoing excavations is the design and development of settlement pattern in a major urban center of the Late Bronze Age, including social aspects like living conditions, trade relations, and technology transfer. In he context of this paper, we will focus on the central stratigraphic units, called /3 and B/2, of the site Q I (see Fig~ 1). B/3 dates to the late 18th or early 19th ynasty, while B/2 falls into the reign of Ramesses II of the 19th Dynasty. In bsolute dates, the New Kingdom covers the second half of the second millenniill BC, including the reign of Ramesses II for most of the 13th century BC. This eriod reflects the regaining of political andcuhural superiority of Egypt over her eighbors, culminating in the 'Battle of Kadesh' when Ramesses II (allegedly) defeated the other superpower of that period, the Hittites, thus raising a new balance of power. The establishing ofPiramesses as a newcapi~l, in the far north-east of the traditional heartland of Egypt and obviously with an imposing view to the area of interest, forms an integral part of the strategic policy of the ew Kingdom, Of particular significance for the intercultural setting of this capital is the apparently peaceful Presence of Hittites at Piramesses. Several templates or molds of typical Hittite shields ('figure-eight shields') were excavated at Qantir, and are most probably to be seen in the context of the political marriage of amesses II with the daughter of Hatusili 1II in Regnal Year 34 (Pusch 1996), Further evidence for international exchange is given by Mycenaean sherds (Mommsen et al. 1996), contributing to the picture of a city of international 'mportance and far-reaching trade relations. The strata of concern here fall into the period of founding this city and clearly eflect the rapidly changing needs for various materials during this phase. The erection of huge temples, palaces and administrative bUildings required not only millions of mud bricks, but enormous quantities of stone masonry:rai~nce tiles, bronze doors, statues and inlays of all kinds of materials, let alone endless numers of tools, implements and installations. Later, a variety of workshops were ceded to keep things running, for repair, maintenance and gradual enlargement of he infrastructure. \ Figure I: Map of sites Q I and Q IV in the south of to-days village Qantir, Faqus district, Eastern Nile Delta. The main walls in both sites represent the situation in level B with multi·functional workshops in the south of Q I, an open court with pillars in the north (level B/2), and the stables in Q IV (level Bb). L.H. and M.H. refer to excavations by Labib Habachi and Mahmoud Hamza, Drawing Excavation Qantir/J. Lindemann. 'l'lO nMI.:~.~_•. ~~,.I u:~.~_•. ~f' "-1~,,~~nl--;"<r Tp.chnology HE BRONZE WORKING BACKGROUND The site Q I, excavated from 1980 to 1987, covers an area of about 14,50Q, quare meters (see Fig. I). The older stratum with major technical installations for he casting of bronze on an industrial scale is named B/3, followed by the younger ,s.tratum B/2, dominated by multi-functional workshops for hot and cold working of metal, stone and other materials. Details were published elsewhere (Pusch 1990, 1994) and are therefore repeated here only briefly. Prehistorv and Historv of Glassmaking Technology 229 I) ") Figure 2: Reconstruction of the bronze melting complex at site Q I, level 8/3, showing simplified and shortened melting batteries (center), cross furnaces (front) and workshops in the back. Drawing Excavation Qantir/J. Klang. Stratum 8/3, industrial bronze casting This stratum contains a complex bronze casting factory of unique size, adjacent to a workshop area. The southern part consists of a multitude of small rooms. used for the working and finishing of organic and inorganic materials such as wood, leather, bone, stone, and various metals. This compound is enclosed by a thick brick wall. North of this enclosure extends the bronze melting estate, a vast umoofed area, containing two types of installations labeled 'melting batteries' and 'cross furnaces' (Fig. 2). From several field soundings we know that this melting area extends over several 10,000 square meters at least. Melting batteries: These are channels with a length of more than 15 meters and an interior width of about 20 cm. They were built as two parallel rows of mud bricks; embedded into the sandy ground, with no floor or further lining in between. The walls were two bricks high, and carried on top a series of tuyeres. It became evident that four tuyeres were focused on each crucible standing inside the channel. There were always two of these melting batteries running parallel to '1"A n 1.•:~ ••...•_ ••..••..• ...J u:"'f- _ f:' ~ ••.••.•...•.•..•..•.•...•••. : •..• .." T~~hn,....lnP"V each other, each containing at least twenty crucibles. The volume of each crucible is estimated to one fifth or one forth of a liter, resulting in a capacity of almost one hundredweight (4 to 5 liters) of liquid copper at a single run for each melting battery. Four contemporaneous melting batteries were identified within site Q I, arranged in two pairs, with one more pair of a slightly earlier date. Several more of them can be inferred from traces in the wider vicinity. Cross furnaces: Each pair of melting batteries is related to a complex type of installation, made up by several brick-lined and floored ditches. Due to the~rmarked pattern, these installations were labeled 'cross furnaces'. A major axis of ca. lO meters in length is intersected in the middle by three slightly shorter axes. The floors of all axes are inclined towards the center, with shallow steps in the main axis. The bricks were lined with a sandy, mud-based 'plaster' containing numerous inclusions of lime gravel. The entire central part of the installations is heavily vitrified, indicating a major and continuous burning in these installations. There is evidence for a platform immediately above the slagged structures, but its extension and exact appearance are unknown. Four such 'cross furnaces' were identified at Q I, but their function remains enigmatic. It is tentatively suggested, based on stratigraphy, orientation and melting debris, that they were used for burning large scale molds and keeping them hot during the casting process. Smal/finds: The northern, open area of the bronze casting factory produced vast amounts of fragments of tuyeres, crucibles and melting debris like charcoal, metal droplets, fused soil particles and the like. The rooms south of the enclosure wall produced a similar range of finds, together with different tools and debris from working various materials. The widespread occurrence of crucible fragments and metal debris is strong evidence for the intimate relation of this workshop area to the bronze melting factory, situated immediately to the north. Stratum BI2, multi-functional workshops All of the installations of stratum 8/3 were short-lived, possibly functioning for only a few years, and soon changed into a different pattern. The workshops in J ~ the south were enlarged and a second enclosure wall was erC()tedjust to the north of the former one. The open area of the melting and casting f~tory was leveled and changed into an open courtyard with octagonal limestone pillars, used as an exercise court for chariots (see Fig. I). } The functional spectrum of the workshops apparently shifted towards servicing these chariots with weaponry, functional and decorative implements utilizing a range of materials. High temperature working remains evident, though on a much less massive scale than before. Several casting molds of steatite, limestone and D •.."'•..•;"•.....••...• , "'.,....• Ut, .• ""ou ,.... r':1..H.,.t"Y'Iol,: •..••.•."T' .........•.... _ .... 1•..••..•. ~. ')'H ) ') clay are proof for small scale bronze casting. Crucible fragments now include two types: thick-walled, hemispherical bronze melting vessels stratum B/3, and a new type of thin-walled, cylindrical glass below). The variety of objects worked, including numerous ze, bone and stone, several molds for the production of identical to those of coloring crucible (see arrow heads of bronmetal appliques for shields, stone and bronze fittings for chariots, clearly demonstrate the setting of the workshops into a high-ranking military context. Chariot troops with archers, protected by leather covered wooden shields with numerous metal applications, formed the backbone of the pharaonic armed forces, and represent represent a change in military technology. Starting in the New Kingdom, they are depicted regularly and in detail on Egyptian tomb paintings and stelae, allowing the unambiguous attribution of tools and weapons from Qantir to these forces. The importance of chariots, not only for Egyptian warfare in general, but for the city of Piramesses in particular, is clearly expressed in the hymnical 'Praise of the Delta "Oh {phm'ao] Bienre-beloved-ofAmun ... / How happy is a day o.lthy time. how sweet was thy voice speaking, when thou didst build the city of PiRamesses-beloved-ofA1I1un, the forefront of evelY foreign land and the end of Egypt, the city beauteous of balconies. radiant with halls of lapis lazuli and malachite. The marshaling place of thy chariotry. The IIll/stering place of thy army." Residence' Additionally matching this, there are stables excavated only 250 meters east of site Q I, at Q IV (Fig. 1), housing at least 418 horses, in a luxurious building complex with pylon-like entrances, porticoes, colwnn halls artd open courts covering an area of approximately 15,000 square meters. nent impurity, averaging about 4,000 ppm, followed by iron with 1,500 ppm. A1I other elements measured were below 1,000 ppm, with bismuth, antimony and tellurium closest to this value. This data is in full accordance with published New Kingdom Egyptian Ongoing MATERIALS scientific investigation of the remains of this industrial and military complex of the Late Bronze Age currently concentrates on artificial high-temperature materials, particularly bronze, glass, and faience. In a1l instances, not only the descriptive analysis of the substances is undertaken, but study also extends into production techniques, as inferred from the analysis of remains and related waste products. Bronze A range of objects and remains, mainly loose metal prills and droplets, were analyzed by ICP-OES. Most objects are low-tin bronzes; tin values scatter around five wt%, with a few samples reaching ten to twelve wt%, and very few with tin below 1,000 ppm. Only four out of 36 have more than ohe percent lead, while on average there is only 0.5 percent lead present. Arsenic is a widespread ?1? Prphi~tnr\1 <Inn j:.fict•..• r" nf r."'o<m<l~;n(7 and promi- Tpr.hnology 1987, Stos-Gale et al. 1995), confir- can be excluded, the most massive metal object yet found is a fragmertt of an oxhide ingot, demonstrating the import of copper metal rather than its local production. The LI composition of this ingot is in full accordance with an origin from the Apliki region on Cyprus (Stos-Gale et al. 1997). Glass Glass was the major new artificial material in New Kingdom Egypt, and played an important part in decorative art not only for vessels, but also as inlays, for bead production etc. The working of glass is attested from various places in New Kingdom Egypt, foremost from Amama (ct: Petrie 1894, Turner 1954, Nicholson 1995), Malkatta and Lisht. Only recently glass melting has been identified at Qantir (Rehren 1995, Rehren 1997, Rehren & Pusch 1997). In considerable contrast to the other sites mentioned, at Qantir we have only evidence for the largescale melting of glass. None of the finds typical for the working of glass to shape were yet found, neithel' drawn rods of colored glass nor semi·finished vessels. Colored glass: The glass worked at Qantir is of the typical Egyptian glass composition, HIGH TEMPERATURE bronze analyses (Cowell ming the idea of an uniformly controlled alloy composition. An important point to be kept in mind is that the majority of metal worked at Piramesses is bronze, while pure or raw copper is very rare. Copper ore smelting which is a soda-lime-silica glass with several percent of magnesium oxide. The vast majority of glass fragments and traces adhering to the melting crucibles (see below) is copper-red glass, colored by five to ten wt% CU20, mainly present as cuprite particles, with only very exceptional samples of copper-blue and cobalt-blue glasses. The latter displays the well-known depletion of potassium, and elevated levels of alumina, iron, manganese, nickel and antimony, in addition to the main colorant cobalt. Details of the glasses studied to date are given in Rehren (1997), and a fuJl account of all glass samples is forthcoming. Glass slags: An important aspect of the glass works at Qantir is the existence of several 'sings' which we believe are directly related to glass production, i.e. the fusion of raw materials to make glass. This is a smelting operation quite different from normal metal smelting, and requires a set of different skills. There are primarily two slag types to be considered, one being fragments of a heavily slagged lining of alternate black and white layers, while the other is lumps of partly fuSed silica-rich material, interspersed with tiny prills of copper metal. n_ ...t...: ..•~..•_~ ••... _ ...1 'LT: ....•....••••.•••• +nJ",~~k / /' •..•....•. 1...•",,,, "'J"}"> ') ,-j _/ Faience and frit bi.; Probably the most prominent, versatile and well-known decorative materials used in Egyptian Art are the different variants of faience and frit. These have been intensively and extensively studied for composition and manufacturing techniques (Kaczmarczyk & Hedges 1983, Vandiver 1983, Tite & Bimson 1986, Tite 1987, Tite et al. 1987, to mention only a few). Faience production: The earliest archaeological research at Qantir was initiated following the fortuitous discovery of 'blue tiles of glazed faience with hieroglyphs in alabaster inlay', in a public road (Hamza 1930: 40). Excavations at the site yielded ample evidence of a faience factory of tremendous size, with not only endless numbers of faience objects and fragments, but more than 10,000 faience molds found during one summer's season alone. These molds make it evident that faience production took place on the site, and not just the use of faience objects (Hamza 1930). The current excavations of the Hildesheim mission continue to secure these molds, though at a less profound rate. Over the years, about 1,300 pieces have been brought to light (Engel in press, Herold & Knauer in press). Ftit: In the Egyptological literature, flit is a term used liberally to describe any siliceous substance which possesses a sandstone-like structure, Le. small, equal-sized mineral grains coagulated by a small amount of glassy binder. It was used for both the body of faience, and colored substances like Egyptian blue and the like (Lucas & Harris 1962). The production of Egyptian blue at Qaptir is evident from the appearance of piano-convex cakes and fractured lumps, ranging in quality, of this material, from whitish, coarse-grained variants full of residual white specks to dark blue, dense and very fme wares. MarI-clay based wares, well represented within the domestic pottery, are entirely absent from the technical ceramic. Bronze melting crtlcibles: The vast majority of crucible fragments from Qantir is related to bronze casting. The clay is rich in organic temper, burned under oxidizing conditions to a bright red. Only the interior surface is typically reduced to a black color where the matrix is totally vitrified. Wall thickness varies from one to five centimeters, and the overall shape of the crucibles is hemispherical, but with possible deviations towards elliptic shapes, and probably with a broad spout (Fig. 3). The crucibles were fired from above, and the ubiquitous green specks in the vitrified surfaces clearly indicate the melting of bronze. Glass coloring crucibles: These crucibles are based on the same clay, but with little organic temper and more sand grains inStead. Whether the sand is really temper or just a contamination of the clay remains open. The sherds are one cm thick, extremely high fired, and vitrified throughout. Bloating, however, is restricted to limited areas at the outside heel of the cylindrical vessels, just above the bottom. Fused parts are very rare, and restricted to collapsed pieces. In summary, the working temperatures appear quite homogeneous throughout the body, and on average much higher than in the bronze melting vessels with steep gradients from "" .,I ,; ,;,;" "" ...-,.- " "" ......- I TECHNICAL CERAMIC AND SLAGS A variety of materials from the excavation at Qantir are directly linked to the production and working of metal, glass and faience. These materials, crucibles, furnace linings and slags, are often more informative about the processes and technologies than the actual materials themselves. Two main classes of these siliceous by-products and waste are considered here, namely the technical ceramic proper, Le. crucibles and tuyeres, and second the large group of slags and linings. Tcchnical ceramic All types of technical ceramic employed at Qantir are based on the local Nile clay. They are, however, blended with different types oftemper for various purposes, and hence differ in appearance and behavior, not only from each other, but also from the usual domestic pottery made of the same raw material (Aston 1997). ~",i n..,.,L: ..••.....•.•• "_rl TT; ...••....•. _ .... C (""'I, •.••.•~ 1_:. __ 'T'........•... _ ....•1"tT\l Figure 3: Bronze melting crucible from Qantir, Stratum H/3 (FZN 83/1118d). The spout is inferred from several fragments, although no matching pieces were found. Reconstruction G. Weisgerber, drawing Excavation Qantir/J. Klang. n_.-J,..: ....•.........••.• ".•..•,t U:"""fA" - - - ------------------------ "f n.1"1""t'n":1t,tnn- T~,..h"nJ"n'<\f \ 'J~" " ) .,,,, ) ,--' inside out Due to this, the cylindrical fragments are much more resistant to wear91/0035b 92/1227 84/1484 and usually less fractured and eroded than the hemispherical ones. At least fort~ 85/0064 individual vessels were reconstructed, demonstrating a high degree of unifonl1it~ in size and shape. Widespread veils of glass, running down the outside, and thicker glass layers sticking occasionally to the insides, are proof for a glass.rela~ ted function of these crucibles. A detailed discussion of size and function of these vessels is published in Rehren & Pusch (1997) and Rehren (1997). 1.3 3.9 7.5 1.4 61.2 4.4 3.3 4.0 1.8 7.4 1.3 63.7 14.1 2.3 1.2 12.7 10.2 2.1 3.0 2.7 0.3 0.9 3.2 3.3 1.5 1.6 6.0 8.6 1.1 1.0 68.4 63.6 Table T. Composition MgO Si02 CaO of bronze TiOZ AI203 FeO P20S NaZO K20 (top) and glass (bottom) melting crucibles " Crucible design: There is a remarkable mastery of ceramic technology evi- dent from the two crucible types. Bronze casting crucibles were designed in a way which ensured a maximum of mechanical stability when charged, even at high temperature. Wall thickness and organic temper enhance the ceramic's thermally insulating properties, which keep the heat inside the vessel and restrict the softening of the fabric to the innermost surface layers. The hemispherical shape . offers a great receptive top surface, where the heat was generated, while minimizing wall surface and thus loss by radiant heat. Also, it offers the most stable shape, allowing to move around the crucible for casting. Glass melting crucibles, in contrast, were designed to allow as much heat throughput from outside in as possible, even on expense of mechanical strength. Walls were thin and dense, allowing for the formation of a continuos vitreous phase throughout the body. The cylindrical shape, supported from below and fired from the sides, offered a sumciently large surface for heat uptake. Mechanical stability FZN was of minor impor- ICP analyses by M. Prange & w. Steger, FZN = Qantir find number tance, the vessels were not moved while hot. An interior lining eventually acted as a protective agent to prevent any contamination of the glass by eroding the ceramic ware. Two totally opposite requirements were matcheQ, using the same clay (Tab. 1): highly insulating properties here, maximum thermal conductivity there. Similarly, different needs were satisfied just by choosing appropriate shapes and linings for mechanical and chemical stability, respectively. Furthermore, no parti~ cular clay quality was used to make the crucibles. Minor differences in the chemi, cal composition between glass and bronze melting vessels are due to their different heating modes. The bronze crucibles were heated in direct contact with the charcoal, and thus incorporated phosphorous, soda and potassium as fuel ash components into their vitrified surfaces, while the glass melting vessels were contami· nated to a much lesser extent: They were heated indirectly (Rehren & Pusch 1997) and were effectively separated from the alkali-rich charge by a thin layer of lime (Turner 1954, Rehren 1997). Tuyeres: Large numbers of tuyere fragments were found, related to the melting batteries in stratum B/3 (see above), but also scattered about the entire site. They all belong to a very homogeneous, sophisHcated type with a downward directed air outlet of reduced diameter (Fig. 5), providing the same degree directabilhy for the blast as elbow-shaped tuyeres, but with less susceptibility failure and malfunction. The ceramic used for their manufacture ble from the bronze melting crucibles. Figure 4: Glass melting/coloring crucible from Qantir, FZN 84/0088. The full hight is 13 cm, the external diameter at the bottom is 15 cm. Drawing Excavation Qantir/J. Klang. "'Ott=; n t..~"",.....••~ _.~..J TT: __ 0 ••••• 1 .• _1 , •.,...\f Slags and linings In marked contrast to contemporary metallurgical of to is indistinguisha- metal smelting sites, there are no primary slags at Qantir. Isolated bits of fused crucibles may resemble slags, but are clearly different in their chemistry and morphology such when studied in ) -) 'Glass slag ': These tentatively labeled finds are lumps of a predominantly fused, alkali-rich material with a major proportion of unfused silica. Its color is light yellow to green, and often pieces have a reddish core from minute copper metal inclusions. The size of these lumps varies trom below a cubic centimeter to half fist size. The co-existence of bits of colorless glass, areas dominated by copper prills, and inclusions of quartz and limestone particles imply the association of these to the production of raw glass. However, alkalis so far measured in these slags differ from typical Egyptian soda·lime-silica glass, in that they are potassium-dominated. Cross.jurnace lining: Yct another fused non,metallic substance is the lining of the cross furnaces. Due to a high degree of lime temper in a local Nile mud matrix, the linings fused to a green glassy body! As long as the function of the cross furnaces remains unknown, it is difficult to attribute a particular function to these linings. However, the lime-rich temper is presumed to be significant in that lime is not a typical soil constituent in Qantir, but certainly a by-product or waste of stone working activity. Why this particular material is present in this specific lining still needs to be explored. Lining technology: The use of linings is much older than the Late Bronze Age, but was until then restricted to decorative coatings over ceramic, or low temperature applications like plasters. Among the materials from Qantir, there are three different linings which functioned in a technical manner in high temperature II!! 11 Figure 5: Tuyere from Qantir, Stratum B/3 (FZN 84/1147,2). Vitrification is restricted to the head, where the tuyere was in contact with the charcoal on top of the crucible. Drawing Excavation Qantir/J. Klang. more detail. Apart from this, two main types of slags were defined, and are briefly described here. They probably form a continuum and just represent different steps in a multi-stage process; this will be discussed in more detail later on. Black-and-white slag: This slag consists of dense lumps of mud, fused to a black glassy body, interspersed with white quartz particles. The most spectacular variant of this has a layered structure, with white quartz-rich layers glued by a colorless glass melt, intercalating with bl"ck fused mud layers. The number of layers can reach up to five, and in most instances a white layer forms the sutface. In several cases, this white top layer is covered by a thick film of color/ess glass. The opposite surface is always fractured as if from rupture or erosion, strongly suggesting that this slag type originally formed a kind of lining. Indeed, one piece was found incorporating a brick-like body. The layered variant continuously mcrges into unstructured, black, white-specked bodies. ?1R Prph1<;:tnnl ~nrt J..Jtt::tnr ••, nf nl~t.~",.,..".,t,..;f"'O' Tpi""hn0)O>!V processes. A pure lime lining served as a protective agent to keep apart the hot glass charge and the fabric of the cylindrical crucibles (Turner 1954, Rehren 1997). Similarly, but yet less clearly identified, a pure quartz lining acted in raw glass making, probably aided by a fuel ash bed. The cross· furnace lining, identified only on grounds of a high proportion of al'\ unusual, non-local temper, still demands interpretation. Altogether though, these examples prove a highly developed level of high temperature skills and a sound awareness of material issues. PROCESS OUTLINES Of the three major high temperature industries - bronze, glass, and faience the organization of the bronze casting factory was reconstructed to the greatest extend. Together with the technical reconstruction given in Pusch (1990, 1994), the shifting emphasis from large, probably architectural, castings (stratum B/3) to small military implements (stratum B/2, Herold forthcoming) became evident. Our view of the glass industry, forming the central issue of this paper, is developed in some detail below, while the faience and frit production will be dealt with separately in the future. Prehistory and History of Glassmaking Technology 239 ,) ,/) Glass industry Despite much research in ancient glass, we know very little about Late Bronze Age raw glass production. Most assumptions are derived from later, Le. Iron Age and younger archaeological and documentary evidence (e.g. Oppenheim et al. 1970, Moorey 1994), and from the chemical composition of finished objects. Archaeological evidence for Late Bronze Age glass production is scarce. As a result, it is still a matter of debate how the raw glass was actually made, and where glass production took place at all. On the currently available, admittedly limited, archaeological data from Egypt, we propose the following model as a general outline. Raw glass production: Based on chemical analyses of finished products, the making of raw glass during the Late Bronze Age most probably relied on plant ash and locally available desert sands (e.g. Turner 1956a, Brill 1970: 111, Freestone 1987). These were - according to a model developed first by Petrie 1926 with later modifications by Turner 1954 (see Nicholson 1993 and 1995 for details, criticism, and a fuIl bibliography) - fritted in pans, then fused and meIted in crucibles. We assume that the initial fusion of tne raw materials was done in art open hearth or hole in the ground, probably already together with some of the colorants; and not in a crucible or shaIlow tray. The evidence for this is the layered slags, Which apparently represent a former lining and merge, in some cases, into the 'glass slags'. To prevent the forming glass from any soil contamination, in particular iro11,the hearth was lined with a layer of pure sand. This lining is represented in the archaeological record as the layered black-and-white slags (see above). In this model, the repeated layering is interpreted as repair traces from re-use of the hearth, with a layer of clay being added as an adhesive substrate for the new pure sand cover. Further protection was probably given by a layer of fuel ash from preheating the hearth, separating the raw materials from the walls when charged into the hearth. The sand lining was fluxed primarily by potassium-rich fuel ash which is why the glass matrix of the wl1ite layers is dominated by potassium. After this first smelt, which combines the 'frittihg' stage of the later processes and the initial fusing of reactive substances, we assume a series of low temperature operations including the selection of 'good' pieces by hand-sorting, and crushing and washing to remove unreactive salts and compounds. The concentrate thus obtained would have gOne, after drying, to a further melt in a crucible. It should be noted that the washing of glass between subsequent melting steps was practised even in the early Modem Period as a means to separate unwanted salts from the glass (Tumer 1956b: 295T). 240 Prehistory and Historv of nI:N~m,,!t;n(t T"'f'hn"lrwv N020 Wt% Figure 6: System Na20-CaO-Si02 (modified from Roth et al. 1987). The typical Egyptian (reduced) glass composition (dark) falls close to a major eutectic trough leading to more lime-rich compositions. Light grey is the area of possible glass compositions at a temperature of 1050 °C. Glass composition and coloration: Given the variable quality of glass-makers plant ash and local sands, it is highly unlikely that the resulting composition of the raw glass is achieved by weighed-in proportions of raw materials. Instead, we propose that the glasses represent the eutectic composition of reactive components (plant ash and desert sand, plus contamination by fuel ash and the protective layer of the crucibles) at process temperature. Within the reduced system of alkalis, silica and lime (Fig. 6), typical Egyptian glasses fall close to the junction of two eutectic troughs, at a temperature around 1050 °C. If the glass would represent an entirely molten batch, one would expect a wider scatter of glass compositions throughout the area of possible glass compositions. Pr",h;d,w\I <1nrl Hi~tnr" "f n1<>~~m,,!tin<1Tpf'hnnln<1" 241 ) This eutectic melting effect easily explains the stunning homogeneity of Egyptian, but also Mesopotamian and Aegean glasses over space and time (e.g. Sayre 1965: 147) as long as only roughly the same ingredients were used. The assumed crushing and washing of the glass in formation ensured a good homogenization of the melt and also facilitates the removal of gas bubbles and remaining seed, thus reducing the melting time and temperature necessary to obtain a clean glass melt. The main difference to the traditional fritting model lies in the significantly higher temperature which was applied to the batch right from the beginning, resulting in a correspondingly higher lime content ofthe glass. Another aspect of our model is the issue of coloration. The colorant has to be added early enough in the process to be homogeneously distributed throughout the melt. Unless delicate redox conditions required otherwise (like for cuprite red, Ahmed & Ashour 1980), it seems best to add the colorant to the washed concentrate, Le. to the powdered raw glass. An addition of the colorants early it1 raw glass making matches the apparent scarcity of uncolored glass. Otherwise, Le. if glass coJoring would belong to glass working as opposed to glass making, one "Y0uld expect more colorless ingots or chunks of glass in the archaeological record, even if on the object level uncolored glass would not have been frequently used. It may be noted that, in turn, this argument ii11pliesa raw glass production close to the production site of colored glass irtgots. The cylindrical vessels at Qantir are interpreted as a means for producing colored ingots (Rehren 1997), and are hence indirect evidence for glass production in Piramesses. Similarly, Nicholson (1995: 15) mentions that, at el-Amama, the cylindrical 'vessels do not seem 10 have hem discovered on sites where glass was only worked ralher Ihan made. Similarly, the relatively low concentration of glass canes might suggest that glas.fworking was not the main activity in this part of the workshop complex.' The copper content of the glass slags described above indicate, at Qantir, colorant addition already during the initial melting of raw materials. This is due to this particular red colorant, CuzO, with its exceptionally high concentration in the glass and delicate redox requirements during coolirtg to precipitate and grow suitable cupdte aggregates. CRAFT SPECIALIZATION The complexity and close inter-relationship of crafts in Egypt has long been known (Kemp 1989), primarily from idealized tomb drawings which often even contain figure captions. In a situation, however, where the matcrial remains of different crafts are scattered and mixed, it becomes a fundamental task of sciencebased archaeology first to assign individual finds to particular processes, before 'JLI') P,'r.ohi"tnru "~MrI u;,..t,... •.•...... c 1'""'1 ..•...,......-•.•1,: .....•..T .......• hn,..JoOV /j these tcchnical processes can be studied in more detail. Of coursc, a re-assessment of assignations may become necessary over time. As shown above, combined archaeological and scientific evidence demonstrate the massive working of bronze, glass and faience in intimate relation to each other at and around sites Q I and . Q IV, and probably of Egyptian blue and gold on a lesser scale also. It is notewor· thy that no indications for pottery kilns exist in this area as yet, although local pottery production has definitely taken place at Piramesses. This raises the question of the degree and level of intra- and cross-craft specialization. Intra-craft special ization Specialized skills: Within the three major industries, we can assign various classes of finds and instaJlations to specific processes. In regard to bronze, we ..may at Qantir distinguish between melting in tuyere-blown CrUciblesin the melting channels, and casting, probably related to the 'cross furnaces' and mold fragments. After cooling, a whole range of low temperature crafts were utilized in smoothing, polishing, engraving, gilding etc. of the finished objects, with tools like punches, hammers and smoothing pebbles, etc., found in the material record of adjacent workshops. Before melting, however, smelting and alloying of the metal was necessary. This splits the metal workers already into three mairt 'unions', with the possibility for further specialization within each group. The same holds true for glass. The production of raw glass certainly required skills different from those necessary to work the hot glass to objeets, and again different from those used for any cold working operations. Similarly, in faience and frit production the preparation of the ingredients will have been done by people other than those assembling the fmished items to composite tiles, etc. Regional specialization: Beside the diversity of skills needed, other factors were also in favor of intra-craft specialization. In a Late Bronze Age setting, metal smelting was geographically linked to the sources of ores and fuel, while casting and finishing of objects was undoubtedly done nearer to the consumers. In between were trading forms, such as bars, oxhide ingots or semi·finished products. This spatia] separation, implying long distance trade routes, of course furthers the intra-craft specialization. Center specialization: Another type of geographically controlled specialization lead to the formation of specialized centers, covering the same range of teclmi" ques and materials as other centers, but witb slightly different emphases. As an example of this we propose here the specialization of centers in producing glass ingots of different color, namely cobalt blue, copper blue, and cuprite red. Only for cuprite red glass ingots we are able to present a geographic origin, namely in Qantir. 'The sheer number of glass coloring crucibles from that site, together with ) ) I .1 IIHJi, the overwhelming predominance of red glass and the lack of debris related to final glass working allows us to postulate such a red glass ingot factory for Piramesses (Rehren & Pusch 1997, Rehren 1997). Only one complete red glass ingot from Piramesses survived, and is now on display at the Egyptian Museum in Cairo. However, an evaluation of about 40 reconstructed crucible diameters from Qantir (Fig. 7) allows to assign an average ingot diameter of about 13 cm to the hypothetical type-ingot from that site. Large quantities of blue glass ingots were found on board the trading vessel wrecked off the Turkish coast near Ulu Burnn (Bass 1986, Pulak 1988). They date roughly to the same period as the crucibles from Qantir, being probably only 50 to 100 years older. The origin of these ingots is a constant matter of debate, and almost everything from Mesopotamia to the Mediterranean region appears possible. Very recently, Nicholson et al. (1997) were able to measure thickness and diameter of almost twenty of these ingots and to correlate the data to the colorant as determined by visual inspection. On average, the twelve Co-colored ingots are ,, , ,':;',;,; ".' .. ..~.1·~'".:!7 -::jJl " ':::'::::::::;;J ~, ~'. '.': '''~:~::.:J " I~.~. , abundance of cylindrical vessels in el-Amarna well have hosted one such center. (Nicholson 1995: 15), this city may , - ."'••......,...,.,..,--I'-'~-' ,'.,,-/ .. , ..,. ingots is a function of the amOlrnt of glass charged into the vessel, and as such dependent on the craftsmcn's recipes. This seems to have been more strictly followed for the copper-blue ingots from utu Burun than for the cobalt-blue ones, according to the data given by Nicholson et al. (1997). The slight, though apparent differences in thickness and diameter (Fig. 8) may thus reflect two different workshops as sources of the Co-blue and Cu-blue ingots. Although at present we can not assign a specific location to either of these workshops, it appears only reasonable to assume a certain geographic distance between them. In view of the " '.:..;.>~:" '.,:' ',': . more than 15 percent larger in diameter, and almost twice as thick as the five Cucolored ones. Obviously, ingot diameter is a function of the internal crucible width, with apparently little variation among the diameters of the two ingot types from Vlu Burun (Nicholsoll et al. J 997). Such a narrow scatter in crucible - and hence ingot - diameter is already attested from Qantir (Fig. 7). While ingot diameters therefore reflect the standardization of crucible sizes, thc thickness of the Red illgot thickness (ill cm) 10 '2J .•• 9antir ", Ingot '-'.~ ---rn;--' ra::=-:::J 5 I Figure 7: Histogram of 40 individual crucible diameters from Qantir/Piramesses (in cm). The majority has an external base diameter of 14 to 16 cm, corresponding to an estimated ingot diameter of 12 to 14 cm. Original data and further discussion are given in Rehren & Pusch (1997). The insets give two examples of such crucibles, demonstrating the range in internal crucible profile. Crucible drawings Excavation Qantir, S. Groteliischen. 5 10 Ingot diameter (ill cm) 15 I 20 Figure 8: Diagram of crucible diameter vs. thickness for 5 Cu-blue and 12 Co-blue glass ingots ftom the Vlu Burun wreck (data from Nicholson et al. 1997). The red glass ingot from Qantir (inset) is much thicker, but has a similar diameter. Ingot drawing Excavation Qantir/K. Enge!. ) If' Although no chemical analyses are yet published for the Ulu Burun ingots, it is a well established fact that Egyptian cobalt-blue glasses have a base glass che. mistry quite different from that of glasses of other colors (Lilyquist & Brill 1993). The most significant discrepancies are among alumina and potassium, with the former being present up to four times as much in the cobalt-colored glasses, and the latter being present only at a fourth of the normal rate. Other differences OCClIr with iron, manganese, nickel, zinc and antimony, all being enriched in the cobaltblue glasses. Dr. Robert Brill, of The Coming Museum of Glass, has been able to analyze several of the cobalt-blue ingots from Ulu Burun, and also such glasses from Amarna and Mycenaean finds. According to his data, these glasses all have the same higher Alz03 and lower KzO values in common. Although they span a wide geographical and chronological range, it appears not unreasonable to assume that they were all made from similar raw materials following similar batching and glassmaking processes (Brill 1997). The increased amounts of transition elements and alumina are probably due to the cobalt-bearing colorant (Kaczmarczyk 1986). The significantly and constantly lower level of potassium, however, cannot be explained by a simple addition of such material, but strongly suggests a specific glass base being used for this particular color. The combination ofthese two completely independent lines of argument, namely base glass composition and ingot size, is considered sufficient evidence to postulate two distinct, though similar, centers for the production of cobalt-blue and copper-blue glass ingots. Cross-craft specialization Despite the probable extent of intra-craft specialization in Late Bronze Age societies, there is a high degree of cross-craft specialization as well. Most recently, P. Nicholson (1996: 18) suggested this for a 'generalised vitreous materials Industry. Similarly, finishing techniques of metal and glass, like grinding, polishing and driUing, have much in common with traditional lapidary techniques. The proximity of such skills, employed at different materials, becomes evident e.g. from the joint occurrence of anow heads made of bone, stone and metal, all being produced in the same workshops of Q I, level B/2. As yet another example of crosScraft specialization, we here propose the mastering of closely controlled redox conditions at high temperatures for copper. It is certainly not simply the control over temperature which characterizes this industrial complex. In this case, one would also expect to find pottery production close-by. In our case, the delicately adjusted composition of the hot gasses involved in the various processes is the crucial 'aspect, in other words, to prevent the fonnation of too much dross and scale on the molten bronze, and to get the copper oxide into the right, monovalent, oxidation state in the glass, in the faience glazes and frit bodies. ') CONCLUSIONS We have been able - tlrst by archaeological and then scientific work - to identify three major high temperature industries within a complex workshop area of ancient Piramesses, and to attribute most of the finds from sites Q I and Q IV to one of these industries. In view of the general degree of cross-craft and intra-craft specialization of that time, the possible structure behind the apparent mix of crafts can be considered. There are several potential advantages to combining different crafts in one industrial complex, including a joint demand for a range of products and services, the necessity to jointly use particular installations and infrastructures, and the recourse to specialized know-how in common need for different crafts. In the case of the industrial complex evident at Qantir, the common thread 'linking these industries was not solely the joint use of copper, or high temperatui res, or the range of mechanical skills, but a combination of these factors. Were it the application of high temperatures, with the furnaces, fuel supply and so on necessary for this, one would expect to have pottery kilns close by also. These, however, are absent from the areas studied so far. The sole use of bronze, on the other hand, does not sufficiently describe the range of activities evident, nor do the low temperature workshops and techniques. To conclude, the industrial complex described here entails a specialized knowledge of high temperature working for a range of materials, all grouped around the mastery of closely controlled redox conditions for copper at various temperatures and in different chemical environments. This quintessential heart of the complex is then supplemented by a wider range of subordinate, less specialized and demanding crafts. Up to the level of semi-finished trade objects, aU necessary factors were present, including the casting of bronze and the production of glass ingots and Egyptian blue for export. The final level of crafts, however, set up to make finished oJ>jects,were orily present according to the local demand, Le. for military equipment, but apparently not for luxury items like glass vessels. The existence of large potteries at ancient Piramesses can be taken for granted, and the working of hot glass to vessels etc. may also be assumed to have existed somewhere in this royal city. This only strengthens our argument that regional craft specialization was a major and complex issue in the Ncw Kingdom, of which we now have only the tlrst indication. ACKNOWLEDGMENT Our thanks are due to all authorities of the Supreme Council of Archaeology, Egypt, for their continuous support throughout the years, and to the members of the various excavation campaigns at Qantir. Special thanks are heartly expressed to Dr. Paul Nicholson and Dr. Robert Brill, for providing access to their data on ) ") Ulu Burun material prior to publication, and for allowing us to make use of it here. Dr. Brill's analyses will be published soon in some publication relevant to the Ulu Burun finds, and they will also be included in a major monograph on chemical analyses of ancient glasses scheduled to appear late in 1998. Two referees gave valuable comments on an earlier version of this text, and Jonathan Golden and Andrew Kramer are thanked for improving the text linguistically. REFERENCES Ahmed, A. & Ashour, G., 1980, "Effect of heat treatment on the crystallisation of cuprous oxide in glass", Glass Technology 22,24-33. Aston, D., 1997, Die Keramik des Grabung.fplatzes QI, Teil 1 - CO/pus of Fabrics, Wares and Shapes, Forschungen in del' Ramses-Stadt 1, Zabern, Mainz. Bass, G., 1986, "A Bronze Age shipwreck at Ulu Burun (Kas): 1984 campaign", American.Joul'tlal of Archaeology 90,269-296. Brill, R., 1970, "The chemical interpretation of the texts", in: A. Oppenhcim et al., Eds., Glass and Glclssmaking in Ancient Mesopotamia, The Coming Museum of Glass Monographs Vol. III, 105-128. Brill, R., 1997, Letter dated May 30, 1997. Cowell, M., 1987, "Scientific Appendix 1. Chemical Analysis", in: V. Davies, Ed., Catalogue of Egyptian Antiquities in the British Museum, VII, Tools and Weapons, I, Axes, British Museum Publications, London, 96-118. Engel, E.M., in press, "Modeln zur Fayence-Produktion aus Qantir", Agypten lmd Altes Testament 10/2. Freestone, I., 1987, "Composition and microstructure of early opaque red glasses", in: M. Bimson & I. Freestone, Eds., Early Vitreous Materials, British Museum Occasional Paper 56, London. Hamza, M., 1930, "Excavations of the Department of Antiquities at Qantir (Faqus district)", Annales du Service des Antiquites de I 'Egypte 30,31-68. Herald, A., forthcoming, "High temperature industries in the Late Bronze Age capital Pil'amesses Qantir: Workshop news from the "House of Ramesses, beloved of Amun", QantirlPiramesses, site Q I, stratum B/2", in: K. Barakat & J. Merkel, Eds., Book of Proceedings, Ancient Egyptian Mining and Metallwgy and Conse/'vation of Metallic Artifacts, Cairo. Herold, A. & Knauer, N., in press, "Anhang: Gesamtliste del' Modeln zur Fayence-Produktion aus del' Ramses-Stadt, Fundjahrgange 1980-1994", A'gyptel1 Kaczmarczyk, & Hedges, R., 1983, Ancient Egyptian Faience, Aris & politan Museum of Art, New York. Lucas, A. & Harris, J., 1962, Ancient Egyptian J'lIatel'ials and Industries, Arnold Ltd., London. Moml11sen, !-L, Beier, Th., Hein, A., Podzuweit, Ch., Pusch, E.B. & Eggebrecht, A., 1996, "Neutron activation analysis of Mycenaean sherds from the town of Ramesses II near Qantir and Greek-Egyptian trade relations", in: S. Demirci, A. (her & G. Summers, Eds., Archaeometry 94, The Proceedings of the 291h International Symposium 0/1 Archaeometry, Ttibitak, Ankara, 169-178. Moorcy, R., 1994, Ancient Mesopotamian Materials and Industries, Clarendon Press, Oxford. Nicholson, P., 1993, Egyptian Faience and Glass, Shire Egyptology Series, 19, Buckinghamshire. Nicholson, P., 1995, "Glassmaking (Egypt)", in: AnlJales du 13e Congres de l'Association l'Histoil'e du ji(,rre, 11-19. Nicholson, P., Jackson, in Qantir-Piramesse/Nord", Agypten & Levante 1, 75-113. Pusch, E., 1994, "Divergiel'ende Verfahren del' Metallverarbeitung und Qantir? Bemerkungen 145-170. zu Konstruktion IT ~p("hnnlo2'Y POUI' Egyptians in Theben und Technik", Agypten & Levante 4, E., 1996, ""Pi-Ramesses-Beloved-of-Amun, Prehistorv and History of Glassmaking (;1 I'1V(·n"".1V;r'I Jnternationale campaign". Amel'ican Joumal of Archaeology 92, 1-37. Pusch, E., 1990, "Metallverarbeitende Werkstatten del' frtihen Ramessidenzeit ?d~ ",f Some new C. & Trott, K., 1997, "The Ulu Burun glass ingots, Kaczmarczyk, A., 1986, "The source of cobalt in ancient Egyptian pigments", ih: J. Olin & M. Blackman, Eds., Proceedings of the 24th Intemational Al'chaeometl'y Symposium, Smithonian Institution Press, Washington, 369-376. tH1fJ U;C'tr\t"tf at Amama: cylindrical vessels and Egyptian glass", JOllrnal of Egyptian Archaeology 83 Oppenheim, A., Brill, R., Barag, D. & von Saldern, A., 1970, Glass and Glassmaking i/1 Ancient Mesopotamia, Coming Museum of Glass, Coming, NY. Petrie, W.M.F., 1894, Tell el-Amama, London. Pctrie, W.M.F., 1926, "Glass in the Early Ages", JOlmwl of the Society of Glass Technology 10,229-234. Pulak, C., 1988, "The Bronze Age shipwreck at Ulu Burun, Turkey: 1985 Chariotry" Prphtt.'tnr" and glassworking work", J01lmol of Glass Sttidies 37, 11-19. Nicholson, P., 1996, "New evidence for glass and glazing at Tell el-Amarna Pusch, und AlIes Te.l·tament10/2. A. Phitlips, Warminster. Kemp, B., 1989, Ancient Egypt, Routledge, London. Lilyquist, C. & Brill, R., 1993, Studies in Early Egyptian Glass, The Metro- Headquarters of thy and Hittites in the Delta Residence of the Ramessides", in: Pelizaells-Musellm Hildesheim, The Egyptian Collection, Zabern, Main, 126-145. TechnoJogy 249 '\ / ',) Rehren, Th., 1995, "High temperature industries in the Late Bronze Age capital Piramesses Qantir: Bronze and glass production and processing", Inte/'l1ationClI Conference on Ancient Egyptian Mining and Metallurgy Aletallic Artifacts, Cairo, 10-12 April 1995. Rehren, Th., 1997, "Ramesside 355-368. and Conservation glass colouring crucibles", of' Archae01l1ef/:F 39. Rehren, Th. & Pusch, E., 1997, "New Kingdom glass melting crucibles from Qantir-Piramesses", JOl/rnal of Egyptian Archaeology 83, Roth, R., Dennis, J. & McMurdie, H., 1987, Phase Diagrams for Ccmmi.\'ts VI, The American Ceramic Society, WesterviJIe. Sayre, E., 1965, "Summary of the Brookhaven program of analysis of ancient glass", in: Application of Science in Examination of Works of Art, Museum afFine Arts, Boston, 145- 54. THE INTERDEPENDENCE OF GLASS AND VITREOUS FAIENCE PRODUCTION AT AMARNA A J Shortland and M S Tite Research Laboratory for Archaeology and the History of Art, University of Oxford, 6 Keble Road, Oxford, OXl 3QJ, UK I Stos-Gale, Z., Gale, N. & Houghton, J., 1995, "The origins of Egyptian copper: Lead-isotope analysis of metals from el-Amarna", in: V. Davies & L. Scho- INTRODUCTION field, Eds., Egypt, the Aegean 127-135. Excavations at the glass factory area at Amarna, undertaken in 1993 and 1994 by Paul Nicholson on behalf of the Egyptian Exploration Society as part of Bany Kemp's Amarna Project, revealed two large furnaces with fused clay adhering to their walls and with large quantities of fused clay both within and around them (NichoJson 1995). Also, found in the vicinity of the thrnaces was a high concentration of industrial debris associated with the production of vitreous materials and pottery. Amarna was the 18111 Dynasty capital of Amenophis IV later known as Akhenaten (1353-1337BC). It was rapidly built on a virgin site in Middle Egypt, but was inhabited for only a relatively short period of time, the royal court relocating back to Thebes during the reign of Tutankhamen (1336-1327BC). Although the site was extensively robbed of almost everything of value in antiquity, it still represents a unique snapshot of an 18th dynasty city during one of the most interesting and controversial of all the periods of Egyptian history (Aldred 1988, Kemp 1989). and the Levant, British Museum Press, London, Stos-Gale, Z., Maliotis, G., Gale, N. & Annetts, N., 1997, "Lead isotope elmracteristics of the Cyprus copper ore deposits applied to provenance studies of copper oxhide ingots", ArchaeometlY 39, 83-123. Tite, M., 1987, "Characterisation 29,21-34. of early vitreous materials", Archaeo1l1ef1:V Tite, M. & Bimson, M., 1986, "Faience: An investigation of the microstructures associated with the different methods of glazing", Archaeometly 28,69-78. Tite, M., Bimson, M. & Cowell, M., 1987, "The technology of Egyptian blue", in: M. Bimson & 1. Freestone, Eds., Early Vitreous Materials, British Museum Occasional Paper 56, London, 39-46. Turner, W.E.S., 1954, "Studies in ancient glasses and glassmaking processes. Part I. Crucibles and melting temperatures employed in ancient Egypt at about 1370 B.C.", JOl/rnal of the Society of Glass Technology 38, 436-444T. Turner, W.E.S., 1956a, "Studies in ancient glasses and glassmaking processes. Part IV. The chemical composition of ancient glasses", Journal of the Society of Glass Technology 40, 162-186T. Turner, W.E.S., 1956b, "Studies in ancient glasses and glassmaking processes. Part V. Raw materials and melting processes", JOlll'11al of the Society of Glass Technology 40, 277-300T Vandiver, P., 1983, "The manufacture of faience", in: A. Kaczmarczyk Hedges, Ancient Egyptian Faience, Aris & Phillips, Warminster, A I-A 144. & R. This industrial debris, taken together with that trom the Amarna excavations undertaken by Petrie at the end of the last century (Petrie 1894) and now held in the British Museum (London), the Petrie Museum (University College London) and the Ashmolean Museum (Oxford), provides clear evidence for the inter-relationship between the full range of vitreous materials being produced in Egypt during the 18111 dynasty. Thus, in addition to evidence for the well-established technologies for the production of ordinary faience (fragments, moulds) and Egyptian blue and related flits (fragments, residues adhering to ceramic vessels and "tiles"), there is clear evidence for the new technology for glass production (spills, rod and vessel fragments, drips adhering to "cylindrical vessels"). Also found were fragments of a cobalt-blue flit which was probably associated with the production of the distinctive cobalt-blue To Iho •• Ient,uthcri,.d of The American the •• prc •• wrillen ')<:() Prph'ctn •.." mvl W;dnr" I"\f r:.1'U'''''''·\,1Vi"n 'TprhnnJoe:V orul:r the l,ws Cmmic conIC"' of the UnilCd States of Amcri". Society. Any duplication. of The American Ceramic reproduction, ,11 copyright er rcpubli"lion interests in this publication of this publication SCdClY or fce paid 10 the Copyright Clc",ncc Pr"hiotnrv ~nrl Hi!;tnrv of GlassmakilH! Technolo!!v are the prop<rty or any part thereof. without C'nter. is prohibited. 251