Glass of the Roman World

By Oxbow Books

Glass of the Roman World illustrates the arrival of new cultural systems, mechanisms of trade and an expanded economic base in the early 1st millennium AD which, in combination, allowed the further development of the existing glass industry. Glass became something which encompassed more than simply a novel and highly decorative material. Glass production grew and its consumption increased until it was assimilated into all levels of society, used for display and luxury items but equally for utilitarian containers, windows and even tools.

These 18 papers by renowned international scholars include studies of glass from Europe and the Near East. The authors write on a variety of topics where their work is at the forefront of new approaches to the subject. They both extend and consolidate aspects of our understanding of how glass was produced, traded and used throughout the Empire and the wider world drawing on chronology, typology, patterns of distribution, and other methodologies, including the incorporation of new scientific methods. Though focusing on a single material the papers are firmly based in its archaeological context in the wider economy of the Roman world, and consider glass as part of a complex material culture controlled by the expansion and contraction of the Empire. The volume is presented in honor of Jenny Price, a foremost scholar of Roman glass.

• Language: English
• Publisher: Oxbow Books
• Release date: Jul 31, 2015
• ISBN: 9781782977759

Book preview

Section 1: Technology and Production

1

PRIMARY GLASS WORKSHOPS IN GRAECO-ROMAN EGYPT: PRELIMINARY REPORT ON THE EXCAVATIONS OF THE SITE OF BENI SALAMA, WADI NATRUN (2003, 2005–9)

Marie-Dominique Nenna

For Jenny, guardian angel on my glass path

Introduction

In 1992, an exceptional discovery was made at Bet Eli‘ezer in Israel (Gorin-Rosen 2000): seventeen tank furnaces were unearthed and identified as primary glassmaking furnaces where a mixture of sand and soda was melted in a one-phase process in a reverberatory furnace. The glass slabs were broken up in chunks, which were transported to secondary glass working workshops in which objects were made. The furnaces, laid side by side in rows, are all divided in two parts: two firing chambers of triangular form and a melting tank of 4 by 2m. Only the floors of the firing chambers and of the tank were preserved and based on this, and the fact that part of the remaining walls were built with reused bricks, the excavators assumed that each furnace was dismantled after the firing to allow the glass slab to be broken up. The site seems to have been in use for a short period; then the glassworkers moved to another location. The activity on the site was initially dated to the 6th-7th century AD, but analysis of the glass objects coming from consumption sites suggests this phase could be extended to the beginning of the 8th century (Freestone et al. 2000, 71–72).

This excavation has led to a reappraisal of the interpretation of the 8.8 tonne glass slab (dimensions of 3.40m × 1.95m × 0.50m) discovered in a cistern in Bet Shearim, which had been dated to the 4th-6th century AD (Brill 1967). The new study (Freestone and Gorin-Rosen 1999) has proposed, on archaeological and analytical grounds, dating this glass slab to the beginning of the Islamic period (early 9th century AD). The glass slab was left in situ because it was an enormous defective piece. Other primary glass furnaces of this type were unearthed in Apollonia in the 1950s and more recently (Tal et al. 2004).

All these furnaces are dated to the end of Antiquity and to the beginning of the Islamic period. Indeed, this type of tank furnace was still in use well into the Islamic period, as shown by the study of the furnaces of Tyre (Aldsworth et al. 2002) dated to the 10th-11th century AD. In all cases, except Tyre, only the floors of the furnaces are preserved and there is no information available on the elevation of the walls or on the form of the roof.

These discoveries have led to the development of the concept of two stages of production, with primary and secondary glass workshops which were located in different places during Graeco-Roman antiquity (Nenna et al. 1997). Various shipwrecks of the Hellenistic and Roman period (for example Sanguinaires A, dated to the 3rd century BC; Mjlet to the end of the 1st century AD and Embiez to the end of the 2nd-beginning of the 3rd century AD) have shown that raw glass was shipped during classical antiquity (Foy and Nenna 2001b; Fontaine and Foy 2007; Cibecchini et al. 2012); these represent the first stage of production. Glass coming from the Orient was certainly used in Central-Western European secondary glass working workshops of the Hellenistic period (Němčice: Venclovà 2009; Manching: Gebhardt 2010). However, the sources of raw glass used during the Early Roman period in the western provinces are still in debate from an analytical perspective (Picon and Vichy 2003; Freestone 2006, 211–212; Leslie et al. 2006, 264–266). It should be stressed that, to date, even though Pliny the Elder is explicit about their existence (Freestone 2008), no primary glassmaking furnaces of importance have been unearthed in Western Europe, and that the two main primary production centres are located by the Ancients in the eastern Mediterranean; in Egypt and in the Levant. In the Late Roman period, glass from the Levant (Levantine 1, Levantine 2) and from Egypt (HIMT group) were competing to supply the secondary glass workshops of the Empire (Foy et al. 2003b; Freestone 2005; Nenna 2014).

Fig. 1.1: Map of the northern part of Egypt.

Following the discovery of Bet Eli‘ezer, in 1996 I initiated a project on primary glass workshops in Graeco-Roman Egypt, searching for sites which could have held such structures and hoping to define the characteristics of the glass industry in the country to which the Ancients attributed a wealthy and successful glass industry, but in which the archaeological remains connected to this craftsmanship were very scant at the time. From field surveys in the region around Alexandria, the Mareotid, and in Wadi Natrun, a depression with salt and natron lakes between Alexandria and Cairo, we could identify five primary glass workshops sites (Fig. 1.1), two in Mareotid (in Taposiris Magna and Marea-Philoxenité) and three in Wadi Natrun (in Zakik, Bir Hooker and Beni Salama). We were also able to recognize that the raw glasses produced in those two regions differed in composition and also differed from the raw glass widely distributed in the western provinces of the Roman Empire (Nenna et al. 1997; 2000; 2005, Thirion-Merle et al. 2003). The identification of these sites was made possible by the presence of surface finds of tiny raw glass fragments, of larger blue-green glassy chunks of poor quality, of bricks covered with a thick vitreous material which had a fractured surface suggesting that glass has been broken off, and others bearing an irregular blue-green vitreous layer. These characteristics are evidence for the presence on these sites of primary tank furnaces in which a glass slab was made and broken up in chunks for transportation, like those discovered at Bet Eli‘ezer.

We chose to concentrate on the Wadi Natrun sites, which seemed to be the more promising, and conducted geophysical surveys. In Zakik, a site already known by scholars of the Expedition d’Égypte, and by the Egyptologist Gardner Wilkinson (Nenna et al. 2000, 100–102), the geophysical survey showed no structures, as the site had been reoccupied at the beginning of the 19th century by natron workers and the upper antique layers disturbed. However, surveys of Bir Hooker (Nenna et al. 2005) and Beni Salama (Fig. 1.3) revealed similar and very distinctive magnetic anomalies shaped like staples, approximately 8m long, with two 2m appendages at right angles. A systematic survey of the pottery, glass and faience from the sites showed that in Zakik, in addition to the Ottoman material, other finds are dated to the Hellenistic and Early Roman periods. At Bir Hooker, the findings of the survey and a small excavation we led in 2004, indicated that the occupation spanned the period between the 3rd century BC and the end of the 2nd century AD (Nenna 2005, 194–198; Marquié 2007).

The site of Beni Salama (Fig. 1.2) situated on the southern edge of the village of the same name, at the eastern entrance to the Wadi Natrun depression, is one of the few ancient sites of the region which is known through archaeological literature because of the presence of the remains of a temple of the Middle Kingdom (Fakhry 1940) and because of an initial study in the 1970s of the remains of the glass industry (Saleh et al. 1972). It is divided into two parts; in the northern area are two mounds, showing magnetic anomalies, comprising the waste from the glass workshops. In the southern area, on a higher plateau, there is an occupation zone, with the remains of a fortress and of a temple of the XIIth dynasty, which was in use from the Middle Kingdom until the 7th century AD, indicated by an examination of the pottery, glass and faience (Marquié 2007).

Beni Salama Excavations

The excavations in Beni Salama took place in 2003 and between 2005 and 2009. We present here the preliminary results, and for further details refer the reader to the exhaustive publication, in progress, which will appear in the Études Alexandrines (for preliminary reports, see Nenna 2007a; 2009; 2010 for the 2005–2007 campaigns; the reports of the succeeding campaigns are in press in the ASAE; see also Nenna 2007b, 127–130; Nenna 2008, 61–62; Orientalia 77.3, 2008, 248–250).

On the southern hill of the industrial zone, we opened two excavation areas around the two magnetic anomalies (Fig. 1.3): sector 1 in the eastern part (416m²), sector 2 (260m²) in the western part. In sector 1, the magnetic anomaly is in the form of a rectangle: after digging, we understood that it corresponds to two different tank furnaces orientated in approximately the same direction: in the eastern one, the western wall of the tank was dismantled to break up the glass slab, in the western one, the eastern wall was dismantled. In sector 2, the staple-shaped magnetic anomaly corresponds to a tank furnace, the western wall of which was dismantled.

The configuration of the surface layers of the two sectors are different (Fig. 1.4). The furnaces of sector 1 are situated in a hollow in the hill; opposite, the furnace of sector 2 is on the slope of the hill. Sector 1 has thus been transformed into a sort of basin, the upper layers of which are impregnated with salt and petrified (natural and man-made) to a depth of up to 1.20m. In contrast, erosion, probably due to the wind and high daily fluctuations in humidity, has removed the upper part of brick structures in sector 2, particularly in the north of the sector, and the petrified layer here is thinner.

The excavation of sector 1 has taken a long time because of the complexity of the structures, which include:

•  the presence of two furnaces built side by side; the western one, and possibly the eastern one too, shows three phases of activity.

•  a weakness in the structure of the north-eastern angle of the western tank, which has forced the glassworkers to build and rebuild supports.

•  the partial reoccupation of the sector after the glassmaking phases while the vault of the last western furnace was still partly standing, which has caused deep destruction to the northern part of the western furnace, but has allowed us to reach the foundation levels of the tank.

•  the transformation of the northern part of the zone into a brick dump during the phase of abandonment.

•  the presence of a thick salty layer covering all the remains and embedding some of them; it seems that this layer might be connected to activities related to the treatment of salt and natron.

In sector 2, the excavation has shown that roughly the same location was reused four times for building the furnaces, and that an earlier now very fragmentary furnace existed on the eastern side. The erosion of the northern part of the main furnace has prevented us from understanding the full plan of the firing-chamber, and in the tank the hardness of the floors has not allowed us to reach the foundation levels. Much later (around the 6th c. AD), within a phase of cessation of glass activities, the southern part of the sector was used as a rubbish dump, rich in amphorae, pottery and basketry, probably coming from the cleaning of Early Roman domestic contexts, situated on the