|
Early Iron Production found in the Jordan Valley. The text below appeared in Near Eastern Archaeology in 1999, and is based on my MA research. As a result of ongoing research, some of the conclusions presented here were found to be incorrect. by Xander Veldhuijzen and Eveline van der Steen (published in Near Eastern Archaeology, September 1999, pp. 195-200)
Introduction [Top of Page] In 1996 and 1997, excavations were carried out at Tell Hammeh az-Zarqa in Jordan, a small site on the north bank of the river Zarqa. These excavations, both supervised by Eveline van der Steen, form part of the Deir cAlla Regional Project, a joint project of Yarmouk University of Jordan and Leiden University, co-directed by Dr G. van der Kooij and Dr Zeidan Kafafi. During the excavations large amounts of slag, ash, charcoal, molten clay and remnants of furnace structures, clearly the remains of some form of metal production, were discovered. Most of the research on these remains was carried out at the Archaeometry Laboratory of Yarmouk University, which is headed by Dr Ziad al-Saad. Tell Hammeh az-Zarqa is a small artificial hill located near the point where the Wadi Zarqa opens into the Jordan Valley. The Arabic word ‘hammeh’ means hot spring, and this name must be the result of the thermal springs close to the tell. (Merrill, 1881, p.193 and Blanckenhorn, 1912, p.320.). The proximity of water is a necessity for the production of iron, both in the construction of furnaces (mixing clay) and treatment of the metal during smithing (annealing). Another necessity is availability of ore. Approximately 2.5 kilometres to the north-east of Tell Hammeh is the iron ore deposit of Mugharet al-Warda (cave of the roses), the only iron ore deposit in Jordan. This deposit is a lenticular body of massive haematite iron ore. Survey and excavation near the ore-body have found traces, including medieval furnaces, of iron smelting activities that may date back to Roman times (Coughenour 1976). A third characteristic of Tell Hammeh is formed by the hot eastern winds ('Sharqiya') that blow a larger part of the year. These winds could be providing or augmenting a natural draught in smelting furnaces, although the temperatures necessary to smelt iron seem to make a forced draught more likely. History of the excavations [Top of Page] The 1996 season [Top of Page] A trial excavation was conducted at Tell Hammeh from November 3 until December 5, 1996. The object was to investigate the change in settlement pattern in the Central East Jordan Valley in the transition from the Late Bronze to the Early Iron Age, and the reasons for this change. A few months before the excavation, a large segment at the eastern side of Tell Hammeh was cut off by a bulldozer for agricultural purposes. This created a large section that only needed to be cleaned to gain a first impression of the occupational history of Hammeh. This resulted in Main sections A and B, with a total width of about 10 meters, situated in the central part of the Tell. The pottery collected from these sections indicated the existence at Tell Hammeh of, from the top down: Late and Early Iron Age layers, Late Bronze, Middle Bronze and Chalcolithic layers. Two trenches were made into the section, one 3 metre wide (Square I, into section B), one 4 metre wide (Square II, into section A). Because of the winds, excavation work was severely hampered, and only the Late Iron Age layers could be excavated. In these layers, a mudbrick wall was found, running through both squares. Partly lying against this wall was a layer mixed of slag, ash and charcoal (locus I-14), indicating some sort of metal industry. Samples were taken of the slag. The sampled material was analysed at the Archaeometry Laboratory of Yarmouk University between December 23, 1996 and February 8, 1997. Eleven different lab-samples that were analysed, at least one, HA7, was positively identified as a slag from the smelting of iron, based on its outer appearance and the results of microscopic analyses. The other samples turned out to be ore in various stages of reduction and molten clay. The 1997 season [Top of Page] The possible presence of iron production instigated another season of excavation at Tell Hammeh, from February 18 until March 31, 1997. This was made possible thanks to the generosity of the Department of Antiquities of Jordan, which funded part of this second season. Work on the bulldozer-cut was continued, and sections were created to the south (sections C and D) and to the north (sections E and F) end of the bulldozer cut. Sections E and F revealed remains of what may be iron smelting furnaces. Excavation was continued in the two trenches from 1996: squares I and II. Some stone walls were found from Early Iron Age period, but the excavated area was too small to draw conclusions as to the nature of the occupation. The layer of slag and ash from the iron industry, which was found in square I in 1996, was found to extend into square II. In square I, locus I-14 was excavated further. It contained large amounts of slag, ash, charcoal, and molten clay. This layer appears to be dump of two collapsed furnaces. The first furnace is located in and under the southern section of the square, and has therefore not been excavated completely. It is an apparently round structure, made of mudbricks, and is filled with ash, charcoal, various slag and molten clay. Many pieces of square clay tubes, possibly tuyères, were found in context with this structure. This structure forms the scattered remains of an iron-smelting furnace. All that was left of the second furnace, at the northern end of the square, was a high concentration of fragments of molten clay and mudbrick, connected with slag. This debris layer might be related to the layer of slag and molten clay that became visible in sections E and F. This industrial layer is covered by a washlayer that also contained slag, charcoal, tuyères, fragments of molten clay and one iron arrowhead. These same layers, containing slag, charcoal, ash and clay, were found to extend into square II. Both layers were sampled extensively. These samples were classified and analysed at Yarmouk University in November and December 1997 and form the main part of the range of sampled material. Finally, in order to collect more information on the iron industry a new trench was opened, square IV, to excavate the furnace that was found in section E (IV-13). This furnace was made of heavily burnt mudbricks, and filled with ash, slag and burnt brick fragments. The analyses at Yarmouk University show complete similarity to the slag from I-14. An ashy layer surrounds the furnace structure, which is standing on a layer of heavily burnt material. This layer possibly incorporates remains from an older furnace that was levelled. More charcoal, ash and burnt mudbrick were found in the loci below the furnace structure(s). Therefore, there seems to be a sequence of at least two furnaces here. Unfortunately, the bulldozer had removed the larger part of this furnace, and what was left collapsed almost as soon as it was cleaned. Thereby, only a little slag, ash or charcoal from within this furnace could be sampled. It is as yet only through the archaeometrical analyses of these samples that a connection between the iron-smelting remains of Square I / II and Square IV can be made. No top-plan of the structure could be drawn, but the furnace seems to be circular. From the section, the outside measurements of the furnace seem to be 2 to 2.5 metres across and 60 to 80 cm high. This represents a minimal size, the structure may have been wider, but it will surely have been higher. Two more possible furnace structures, resembling IV-13 were found in Main section F, to the north of section E. In the section they are 1.5 metre across and 40 cm high, again the minimal (outside) size. These structures contain slag, ash, and burnt mudbrick. The most northern structure also contained lumps of haematite iron-ore. Dating [Top of Page] At this moment, the only indication for the dating of the iron production remains comes from associated pottery, which dates from Iron Age II, between 800 and 700 BC. In all squares, the smelting remains are found close to the surface of the Tell. They all seem to belong to the same phase, and no remains prior to Iron Age II have been found so far. However, only a very small area of the site has been excavated. It is possible that iron was produced in the area prior to Iron Age II (Pigott and McGovern, 1982). Further excavation of the site may reveal a long-term centre of iron production, starting earlier than what has been found so far. Laboratory Analyses Method [Top of Page] In two periods, December 23, 1996 to February 8, 1997 and November 10 to December 15, 1997, the Hammeh iron production remains were studied, inventoried and analysed at the Archaeometry Laboratory of Yarmouk University in Jordan by Xander Veldhuijzen. The sampling was supplemented with high-grade haematite (Fe2O3) ore from Mugharet al-Warda, which has been analysed before in the 1960’s (Van den Boom and Lahloub 1962, pp.39-46, Bender 1968, pp.149-151). All material was studied, described and if necessary drawn and/or photographed. A working classification was made, based on the outer appearance of the samples and the various expected technological steps in iron production and their subsequent material remains. Most of the slag showed a ‘tapped’ structure, indicating the practice of running of liquid slag from the furnace during the smelt. Of all selected groups and subgroups, five samples were selected and subjected to various analyses. The non-slag samples, such as the burned clay and tuyères, were analysed for chemical composition. All other samples were cut, hand-polished and cast in epoxy-resin. The mounted samples were then machine-polished to one quarter of a µ. Then they were studied microscopically to determine their mineralogical composition and phases. Microphotographs were made of all samples. Some of the samples were also examined and photographed with a scanning electron microscope (SEM). Chemical elements of certain phases were defined by electron probe microanalysis, to bridge the phase-specific microscopical and the bulk chemical analysis. Various areas of all samples, related to the microscopically studied surface, were pulverised, dissolved in various acids and tested for their chemical (main and trace) elements with Atomic Absorption Spectrometry (A.A.S.). Two samples were later retested using X-ray fluoresence analysis (XRF). Results [Top of Page] The macroscopical analysis divides the samples in five main categories: slag (magnetic and non-magnetic), furnace remains (molten and non-molten), ore (from Hammeh and from Mugharet al-Warda), tuyères, iron objects (from Hammeh and from Deir cAlla, these artefacts have not been analysed yet) and one for odd pieces (most of these have not been analysed). These groups are further divided by characteristics such as magnetism, weight, colour and appearance. The microscopical and chemical analyses confirm this classification. Most slag have a ‘tapping’ structure, indicating that they were run off from the furnace during the smelt. One subgroup of slag has a convex bottom with embedded sandgrains, and patches of rust. These slag appear to have formed at the bottom of the furnace. Visually, most of the molten or vitrified furnace remains appear to form a range of increasingly ‘slagged’ furnace wall. This is borne out by their chemical composition, which shows increasing amounts of iron. Two subgroups, first attributed to the slag category turned out to be ore in different stages of reduction. This identification is borne out by their mineralogical and chemical similarity with the iron ore samples. Microscopically, most of the slag show a rather homogenous glass matrix with very thin wüstite (FeO) dendrites, indicating smelting slag resulting from a high furnace temperature and fast cooling. Many slag showed sharply defined fayalite (Fe2SiO4/2FeO.SiO2) laths in the glass matrix. Some slag showed some magnetite (Fe3O4) blocks and dendrites. The mineralogical phases present in the slag, together with the tapped structure and non-magnetism, point clearly at iron smelting as opposed to smithing. Judging from the often very fine, undeveloped dendrites, both furnace temperature (in excess of 1200oC) and cooling rate of the slag must have been rather high. The chemical analysis identifies the main and the trace elements of all samples. The main elements verify the origin of the samples as a result of iron production, whereas the trace elements serve as ‘fingerprint’ for comparison between slag, ore and artefact. Traces of titanium (Ti /TiO2) and of silver (Ag) and, less clear, traces of Na2O (sodium dioxide) and Zn (zinc), are found in all slag, vitrified furnace wall and ore samples. This strongly suggests a connection between the iron production at Hammeh and the ore deposit of Mugharet al-Warda. The chemical composition of the vitrified end of one of the clay tubes (extremely similar to that of most slag) clearly showed that it had been in contact with the slag. This clearly identifies it as a tuyère used for air supply to the furnace. Conclusion [Top of Page] Combining the results summarised above leads to a tentative reconstruction of the Hammeh iron smelting process. Apparently the smelters at Hammeh used large furnaces, charged with Mugharet al-Warda ore and most likely charcoal and apparently fired to high temperatures. Air was supplied through square unbaked (prior to the smelt) tuyères, 5 by 5 cm in section. Whether bellows were used is not known, but it seems likely, considering the small opening in the tuyères (1 cm in section) and the apparent temperature in the furnace. It is not impossible, however that air supply was caused by natural draught, or that this draught was augmented by the strong local wind. The iron production debris (slag, ash, charcoal, ore, tuyères, furnace remains and various fragments of burnt and vitrified clay), that were found at Tell Hammeh constitute a quite complete range of iron-smelting remains. This indicates some form of iron production centre, using the ore from nearby Mugharet al-Warda. Archaeologically, there is no indication of any other activity than smelting. Indicators of forging such as hammer-scale, hammering stones or anvils have not been identified. Besides being of considerable importance for the reconstruction of very early iron smelting techniques, the finds at Tell Hammeh constitute the oldest centre of iron production found so far, in the region if not worldwide. To compare, the only other find of comparable age, from Tell Afis, Syria, is supposedly dated to approximately 700 BC and consists of a few slag found on the floor of a room (Matthiae, 1979, pp2-4; Ingo et al. 1992, pp.273-275). The next oldest finds stem from Agadez, Niger (approx. 500 BC) (Posnansky and McIntosh 1976, p.184; Calvocoressi and David 1979, p.10; Childs and Killick 1993; Miller and van der Merwe 1994). The first instance of slag together with furnace structures in the entire Mediterranean region stems from Baratti Beach, Italy and is dated to 170 BC ±70) (Craddock and Hughes,1995, p.265). References [Top of Page] Abu-Ajamieh, M.M. and Bender, Friedrich K. and Eicher, R. N., 1988. Natural Resources in Jordan, Inventory-Evaluation-Development Program. Natural Resources Authority, Amman. Bender, Friedrich K., 1968. Geologie von Jordanien. In: Beiträge zur Regionalen Geologie der Erde, Band 7, Berlin. Blanckenhorn, M., 1912. Naturwissenschaftliche Studien am Toten Meer und im Jordantal. Friedländer, Berlin. van den Boom, G. and Lahloub, M., 1962. The Iron-Ore Deposit 'Warda' in the Southern Ajlun District. Unpublished report, Federal Institute of Geoscience and Natural Resources (Unveröffentlichte Bericht deutsche geologische Mission Jordanien, Archiv Bundesanstalt für Bodenforschung), Hannover/ National Resources Authority, Amman. Calvocoressi, D. and David, Nicholas, 1979. A New Survey of Radiocarbon and Thermoluminescence Dates for West-Africa. In: Journal of African History 20-1, pp.1-29. Childs, S. Terry and Killick, David, 1993. Indigenous African Metallurgy: Nature and Culture. In: Annual Reviews of Anthropology 22, pp.317-337. Coughenour, R. A., 1976. Preliminary Report on the Exploration and Excavation of Mugharet el Warda and Abu Thawab. In: ADAJ 21, pp.71-76. Craddock, Paul T. and Hughes, M.J. (eds), 1995. Early Metal Mining and Production. Edinburgh. Ingo, G. M., Scoppio, L., Mazzoni, Stefania, Mattogno, G. and Scandurra, A., 1992. Application of surface and bulk analytical techniques for the study of iron metallurgy slags at Tell Afis (N/W Syria). In: Vandiver et al., 1992, pp.273-275. Matthiae, Paolo, 1979. Sondages a Tell Afis (Syrie) 1978. In: Akkadica 14, pp.2-4. Morton, G.R. and Wingrove, J., 1969. Constitution of Bloomery Slags: Part I-Roman. In: Journal of the Iron and Steel Institute 207, pp.1556-1564. 1972. Constitution of Bloomery Slags: Part II-Medieval. In: Journal of the Iron and Steel Institute 210, pp.478-488. Pigott, Vincent Charles and McGovern, Patrick E. and Notis, M.R., 1982. The earliest steel from Transjordan. In: Journal of the Museum Applied Science Center for Archaeology (MASCA J) 2(2), pp.35-39. Philadelphia. Posnansky, Merrick and McIntosh, Roderick, 1976. New radiocarbon dates for north and west Africa. In: Journal of African History 17, pp.161-195. Rostoker, William and Bronson, Bennet, 1990. Pre-Industrial Iron, Its Technology and Ethnology. Archaeomaterials Monograph 1, Philadelphia. van der Steen, Eveline J., 1996. The central East Jordan Valley in the Late Bronze and Early Iron Ages. In: BASOR 302, pp.51-74. 1997. Excavations at Tell el-Hammeh. In: Occident and Orient 3, pp.12-14. 1999. Excavations at Tell el-Hammeh. In: Studies in the History and Archaeology of Jordan VII (forthcoming) Vandiver, Pamela B., Druzik, James R., Wheeler, George Segan and Freestone, Ian C. (eds), 1992. Material Issues in Art and Archaeology III. Materials Research Society, Pittsburgh, Pennsylvania. Veldhuijzen, Harald Alexander, 1998. Early Iron Smelting. Analysis and Interpretation of Late Iron Age Iron-Smelting Remains from Tell Hammeh az-Zarqa, Jordan. Waldbaum, Jane Cohn, 1989. Copper, Iron, Tin, Wood, the start of the Iron Age in the Eastern Mediterranean. In: Archaeomaterials 3, pp.111-122. [Top of Page] |
||||||||||||||