Egyptian glass is among the finest of the ancient world. Yet how did the ancient Egyptians make it? New work, at the world’s earliest-excavated glass making factory in the A-list site of Tell el-Amarna, is unravelling the mysteries. Here Paul nicholson delves into the archives of the late great Egyptologist, Flinders Petrie, who excavated at Tell el-Amarna in the 1890s; and then takes us to his own excavations, a century later, as field director of the Egypt Exploration Society’s Amarna Glass Project. Here he tells of his excavations, how he undertook a host of fiery experiments, and why his team has shattered a raft of old interpretations.
Tell el-Amarna, some 360 miles south of Cairo, was the capital of the ‘heretic pharaoh’ Akhenaten (1352-1336 BC). Planned as a ‘new town’ by Akhenaten, Tell el-
Amarna was abandoned soon after his death. His town, therefore, offers a rare and significant snapshot of urban Egyptian life and industry in the late 18th Dynasty of the New Kingdom (1550-1075 BC).
The site rose to prominence thanks in part to the work of Flinders Petrie who excavated at the site in 1891-2. Then, in 1912, a German expedition found the famous bust of Nefertiti, the chief wife of Akhenaten. This fixed the site in the public gaze. Following World War I, the archaeological concession passed to Britain’s Egypt Exploration Society – or E.E.S. – who worked there in the 1920s and 1930s and resumed work under Professor Barry Kemp in the late 1970s.
My work, in conjunction with the E.E.S., focussed on glass. At the time of Akhenaten it was still a new material. But how did the Egyptians produce it? Did they make glass from its raw materials (silica, lime and soda) or did they simply import the raw glass from elsewhere? Petrie’s pioneering work was fundamental to understanding early glass production, not only in Egypt but throughout the Near East. However, on closer examination, I found that some of his seminal theories are difficult to reconcile with his finds, which raised questions as to whether or not they were correct. To explore his theories, we need to wind back the clock to 17 November 1891 when Flinders Petrie arrived at Tell el-Amarna.
Petrie at Amarna
Within a matter of hours the unstoppable Petrie had made camp, and within a day had built his excavation house. While this was going on, his trusted workmen were set to work on excavating houses, and Petrie ‘walked over the place, and entered on my heritage’ (Petrie’s Journal now held in the Griffith Institute, Oxford). The scale of the task before him was not lost on Petrie: ‘It is an overwhelming site to deal with. Imagine setting about exploring the ruins of Brighton, for that is about the size of the town and then you can realise how one must feel with such a huge lump of work…this place would need a lifetime to exhaust it properly’ (Petrie Journal). Petrie had no intention of spending a lifetime on the site, he had many other priorities in Egypt, and had specific reasons for working at Amarna as ‘the most promising place for (1) pieces of the finest carving and glaze work of that age, (2) valuable things hidden or lost, (3) any historical objects such as papyri or clay tablets’ (Petrie Journal). Clearly, from the beginning Petrie had an interest in ‘glaze work’. This was no doubt inspired by the numerous pieces of faience (a vitreous material) and glass littered across the site. Indeed, his season at Amarna has become well known for his reconstruction of the glass and faience making process – as well as for his work on the houses and monumental buildings of the site (in particular his discovery of the painted pavement, and the plaster ‘masks’). Although many scholars at the time thought that glass originated in Egypt, Petrie rightly predicted that the origins of glass lay in Mesopotamia, but that the relevant evidence had simply not been found there because of the poor survival conditions offered by the damp ground. However, he also found evidence that made him conclude that the Egyptians in Amarna also made their own glass and were not dependent upon imports from abroad.
His evidence for local glass-making took several forms, and was accumulated gradually. Early on he found a ‘pan’ of frit material which showed large bubbles where gases had escaped during the reaction process, and pieces of unreacted silica. He reasoned that this ‘frit’ [a mixture of quartz (silica), alkali (perhaps natron or plant ash) and lime with a copper or cobalt colourant to give its blue colour] was the first stage in making glass. This was not a fanciful notion on his part, but derived from a sound knowledge of crafts and industries, both ancient and modern.
The next evidence came in January 1892, when Petrie was joined by Howard Carter, then on his first visit to Egypt. Petrie’s initial impression of Carter was as ‘a good-natured lad, whose interest is entirely in painting and natural history …and it is of no use to me to work him up as an excavator’ (Petrie Journal). This was to prove premature, as later journal letters make clear. The two men made an excellent team, and Carter’s arrival coincided with the discovery of what were described as ‘amulet factories’, places where faience was produced. His time at the site also encompassed the excavation of other glass and glaze workshops, one of them discovered by local children. The sum total of Petrie and Carter’s finds of glass and faience led Petrie to attempt a reconstruction of the processes of glass and faience production at Amarna. This was the first time such a feat had been attempted. It has been the standard account of ancient glass production ever since.
So how did Petrie think the Egyptians made glass? He believed that the raw materials were first roasted together in small bowls: this was the ‘fritting’ process. The fritted mass would then be ground up and melted to produce glass. This glass would be free of bubbles since the gases had already escaped during fritting. At Amarna he had also found quartz pebbles, some with drips of glaze on them, or adhering to potsherds. He believed these quartz pebbles had made up the floor of the kiln. He also thought that the continual heating and cooling of the pebbles would have caused them to crack and shatter so that they could subsequently be crushed for use as a silica source in making glass or faience. Despite this, Petrie’s famous diagram of the glass production process (published in his Tell el-Amarna 1894) does not show the quartz pebbles. He also makes it clear that he found no actual trace of kilns or furnaces. These facts are significant.
New reflections on glass
In the 1980s, I had already begun questioning certain aspects of Petrie’s seminal reconstruction of the glass-making process. Elements did not add up. How, I wondered, did hot gases in the furnace pass through the layer of pebbles, and up toward the inverted cylindrical vessels? How were the pebbles balanced on a perforated kiln floor and how did the cylinders stay level enough on them to support the fritted materials, particularly when it would be necessary to stir or rake those materials? What did the elusive furnaces look like? Were faience and glass made together in the same workshops? Some authorities were even suggesting that the Egyptians could not make glass at all, but relied entirely on imports from elsewhere in the Near East for their raw material, which they would then work into vessels and other artefacts in Egyptian style.
To bring some resolution to these questions, my team and I began a careful examination of Petrie’s finds, many of them held among the extensive collection at the Petrie Museum, University College London. Among the objects were cylindrical vessels that were not like other pots; rather they had a calcareous lining on the inside. In the 1950s, Professor W E S Turner had suggested these vessels had been used for the melting of glass. I also examined the pebbles, but it was clear that museum collections include very few of them, though whether this was because they were less common than Petrie’s account suggests, or were not thought worth shipping to Britain was difficult to say. It was clear that the only way to go much beyond the accounts given by Petrie and Turner was to excavate, perhaps re-excavate, at Amarna. The Supreme Council for Antiquities in Egypt generously granted permission for the work. However, the question of where to excavate was not a straightforward one. In his publication of Amarna, Petrie never once states the exact location of any of the ‘three or four glass factories, and two large glazing works’ which he discovered. Nor does his plan of Amarna show the location of these sites. All that is shown is an area marked ‘moulds’ – near to the modern water tower – and the area of the site grid prosaically known as O45.1. We reckoned that ‘moulds’ might refer to an area where moulds for faience production had been found, and since glass and faience production seemed to take place in the same areas, we felt this would be good spot for further investigation. During Barry Kemp’s season of work at Amarna in 1992 (see CWA1), Ian Mathieson helped my search by kindly undertaking a geophysical survey of O45.1.
Encouragingly, this showed spectacularly high readings from the proton magnetometer, consistent with areas of intense burning, probably kilns or furnaces. While we knew that Petrie had found no furnaces, it was clear that he had found moulds for making faience in this area, and it seemed a suitable location for excavation.
Digging Amarna’s heart of glass
We began excavation in 1993, and within days revealed a kiln. However, it was quite obvious that this small kiln was of the type well known as pottery kilns from Amarna, and although it was surmised that the construction of a glass furnace might be similar, we were unconvinced. Shortly afterwards a larger, thicker, curving wall began to emerge. This was covered by a layer of tumbled brickwork and was initially considered to be too large to be a kiln or furnace of any kind. However, in this thinking we had fallen into the same trap as later critics of the work: namely we assumed that the earliest glass kilns to be excavated in Egypt would be small. We assumed this on no better evidence than early must equal small. We were in for a surprise.
As the excavation progressed it became clear that this structure was indeed a kiln or furnace. Labelled Kiln 2, its brickwork formed a complicated pattern, and was some half a metre thick, far more than the 16cm usually associated with pottery kilns. In addition, the pattern was completely unlike those of pottery kilns, and was clearly intended to withstand considerable heat and thermal stress. As the halfsectioning of Kiln 2 continued, a third structure – Kiln 3 – began to emerge. The first trace of this was a sloping brick, next to which was a pool of solidified ‘slag’ (in fact, melted furnace lining rather than true slag). This new structure was only partly excavated in 1993, and there followed a year of debate about these structures before our return in 1994.
With eager anticipation, the 1994 season saw Kilns 2 and 3 completely excavated. They were spectacular structures with an internal diameter of c.1.5 metres, and surviving to several courses high below ground. Their uppermost parts had been broken away when the later building, designated O45.1, had been built but on Kiln 3 enough remained to show that this had been domed. Not only did some of this curving dome remain, but loose bricks from the site showed ‘stalactites’ of slag which hung from them at an angle. Given that gravity in the 18th Dynasty appears to have obeyed the same laws as that today, the bricks must have been angled in order for the slag to drip vertically downward, indicating a domed structure.
The kilns were not the only focus of the excavation. As they progressed, so too did work on what was clearly a potter’s workshop, complete with a pit for preparing the clay. Around the kilns we made finds of pieces of raw glass, pieces of faience, and the remains of the aforementioned cylindrical vessels that Turner had suggested might be for melting glass. We were now convinced that Turner was correct to suggest that these vessels were involved in glassmaking. Indeed, we had already found a vessel of this type, with glass adhering to it, in the Palace. Dumps, an area of the site with which Petrie was familiar as producing the remains of glass- making.
Interestingly, Petrie had found human bones in these same Palace Dumps, along with glass-making debris, and our excavation revealed that, before the construction of the O45.1 glassworks, there had been a cemetery on the site. We might tentatively suggest that some of the human remains, along with glass-making debris, might have been carted to the dumps at the time later building at O45.1 took place, or when the furnaces themselves were rebuilt during the life of the industrial installation.
The furnaces were an important discovery, but caution was necessary in determining their use. Might they just be for making lime, or for burning charcoal? Were they metallurgical? Our investigations suggested that they were for none of these and that they might be for glass production. This suggestion was generally met with the view that the size of such structures would preclude them reaching the temperatures necessary for the making of glass, because of that old assumption that early glass kilns ‘must be small’.
In reconsidering our hypothesis that these might be glass furnaces, we realised two things. First: no comparably early glass furnaces were known from anywhere in the world. Second: although Petrie’s reconstruction of glass-making still appeared in almost every book on glass history, many glass historians still did not believe that the ancient Egyptians could actually make, rather than simply work, glass. In other words, Petrie’s reconstruction of glass-making methods was seen as correct, but it did not apply in Egypt – where he had actually found his evidence!
This view had come to pass by an accident of archaeological history. For a number of issues had blinkered scholars’ view of Egyptian glass. First, the discovery – as Petrie had always predicted – that glass was made earlier in Mesopotamia than in Egypt (an idea confirmed in the early 20th century). And second, the fact that a few of the famous Amarna Letters (an archive of correspondence on clay tablets, mostly diplomatic, found at the site) talk of importing glass to Egypt.
Experimental glass-making and laboratory analyses
The distinction between glass-making and glassworking is a critical one. No one would deny that the Egyptians could work glass with great skill by the time of Akhenaten. This requires temperatures of only 800-900°C (depending on composition) rather than the 1100-1150°C necessary to make glass from its raw materials. Were our Kilns 2 and 3 simply for glass-working, without being capable of reaching temperatures sufficient to actually make glass?
The only solution was to reconstruct one of the kilns at full scale and to try to manufacture glass in it from local raw materials. In 1996, Caroline Jackson, of Sheffield University, and I attempted to reconstruct Kiln 3 at its original size. A mixture of Amarna desert sand, with plant ashes (in fact seaweed ash was used as a substitute) and a little cobalt to give the characteristic blue colour was used in the experiment. Although lime is necessary to make a stable glass which will not deteriorate over time, we added none to our experimental batch materials. This is because the Amarna sand naturally contains lime, and we judged that this might be sufficient, and a further reason why a glass industry might have existed at the site. The experimental firing reached a maximum temperature of around 1150°C, and spent several hours at c.1100°C. We expected that this firing would produce a well-reacted frit, but in fact it produced an ingot of transparent blue glass, free of both gas bubbles or unreacted quartz! This revealed two things. Firstly, it was clear that even in the hands of inexperienced experimenters Kiln 3 was capable of temperatures sufficient to cause the reaction of glass making materials. And secondly, the fritting stage in glass production may not have been necessary even in ancient times. Experimental archaeology cannot prove that Kilns 2 and 3 were for glass-making, it can only demonstrate that they may have been. However, when combined with other evidence the case becomes still stronger.
Shortly after the experiment had taken place, Professor Tite and Dr Shortland of Oxford University, further demonstrated that the frit found by Petrie did not directly match the composition of the glass. This meant that while it was possible that such frit might have been used in colouring glass, or faience, it was not itself the essential first step in glass-making. This supports the notion that fritting may have been unnecessary as part of glass manufacture. So, we had worked out how glass was could be made at Amarna and where it was made. We were soon to discover more about ‘Turner’s vessels’ (those cylindrical pots of the type represented in the Petrie Museum that Turner had thought were used in glass-melting), and just how important Egypt’s place might have been in the story of ancient glass.
Flicking through the pages of an old National Geographic, published in 1987, my interest was caught by a picture of glass ingots from a shipwreck at Uluburun, off the Turkish coast. The ingots boasted an oddly familiar Amarna potlike pattern of grooves and ridges on them, as if made with the fingers. Since dipping ones fingers into molten glass is inadvisable, these ridges must be casts from the vessel in which the glass was made. Moreover, the virtually cylindrical ingots look just like the vessels associated with glass-making first identified by Turner.
Mystery solved?
Caroline Jackson and I started to investigate. We discovered that the ingots were the same size as the Amarna vessels – indeed, a cast of one of the ingots fitted neatly into some of the vessels from Amarna. As for the lime-rich layer on the inside of the vessels, it would have served as a good parting agent when the vessels were broken open. It would also have prevented the iron-rich Nile clay from discolouring the glass, something that would have been particularly important for glasses that were not strongly coloured by cobalt.
Together with Dr Robert Brill of the Corning Museum of Glass, and Dr Jackson, we carried out lab analysis of the Uluburun glass. With great anticipation we discovered the ingots had the same composition as the glass from Amarna. Egypt must have been the source of the Uluburun ingots, which might have been on their way to the Mycenaean world, off to be turned into new objects, rather than Egypt being the recipient for these glass ingots. So while there is no doubt that Egypt did import some glass, as the Amarna Letters make clear, this does not mean that all glass was imported; just as the import of some foreign plastic into Britain does not mean that Britain cannot produce its own plastic.
This article is an extract from the full article published in World Archaeology Issue 28. Click here to subscribe