Metalworking at Megiddo during the Late Bronze and Iron Ages

Metalworking at Megiddo during the Late Bronze and Iron Ages NAAMA YAHALOM-MACK, The Hebrew University of Jerusalem ADI ELIYAHU-BEHAR, Bar-Ilan University MARIO A. S. MARTIN, Tel Aviv University ASSAF KLEIMAN, Tel Aviv University RUTH SHAHACK-GROSS, University of Haifa ROBERT S. HOMSHER, Harvard University YUVAL GADOT, Tel Aviv University ISRAEL FINKELSTEIN, Tel Aviv University* Introduction Over the years, the excavations at Tel Megiddo have yielded numerous metal objects, as well as evidence of metallurgical activity. Here we show that metalworking was practiced continuously in the southeastern sector of the tell from the end of the Middle Bronze Age until Iron IIB. During this time, one can trace changes in the production processes, including the introduction of ironworking alongside the continuation of bronzeworking. Dating and contextualizing this significant development is crucial to understanding the social, economic, and geo-political circumstances under which it occurred. The evidence that we present shows that iron production was accompanied by a substantial increase in iron, as expressed on the level of consumption. This occurred during Iron IIA, concurrent with cultural and political changes, expressed, inter alia, in the architec* This study was funded by the European Research Council under the European Community’s Seventh Framework Program (FP7/2007–2013)/ERC grant agreement no. 229418. We would like to thank Yigal Erel, Institute of Earth Sciences, The Hebrew University of Jerusalem for the use of his pXRF analyzer, and Genia Mintz, The Kimmel Center for Archaeological Science, Weizmann Institute of Science, for her assistance with the calibration procedure. tural layout of the city and its cultic practices. Possible evidence that local bronzesmiths may have been partially engaged in iron production already in Iron Age I is found in the form of a hoard recently unearthed at Megiddo, although iron-working debris has not yet been found in contemporary contexts. The fact that metalworking and exploitation of metal resources are strongly tied to political power and social infrastructure is well accepted.1 Especially intriguing is the introduction and gradual preeminence of iron over bronze (which continued to be produced), a process that took place at the turn of the 1 E.g., T. E. Levy and S. Shalev, “Prehistoric Metalworking in the Southern Levant: Archaeometallurgical and Social Perspectives,” World Archaeology 20 (1989): 352–72; T. E. Levy, “Cult, Metallurgy and Ranked Societies – Chalcolithic Period (ca. 4500– 3500 BC),” in The Archaeology of Society in the Holy Land, ed. T. E. Levy (London, 1998), 226–44; T. E. Levy et al., “Reassessing the Chronology of Biblical Edom: New Excavations and 14C Dates from Khirbet en-Nahas (Jordan,)” Antiquity 76 (2002): 425–37; T. E. Levy and M. Najjar, “Some Thoughts on Khirbat en-Nahas, Edom, Biblical History and Anthropology - A Response to Israel Finkelstein,” Tel Aviv 33 (2006): 107–22; I. Finkelstein and E. Piasetzky, “Radiocarbon and the History of Copper Production at Khirbet en-Nahas,” Tel Aviv 35 (2008): 82–95. [JNES 76 no. 1 (2017)] © 2017 by The University of Chicago. All rights reserved. 022–2968–2017/7601–001 $10.00. DOI: 10.1086/690635 53 F 54 Journal of Near Eastern Studies first millennium BC, hand in hand with the emergence of new territorial kingdoms in the southern Levant.2 Increasing evidence for metalworking at Tel Megiddo, a multi-layer site that served as an important territorial/administrative center during both the Bronze and Iron Ages,3 sheds new light on the subject. Although Tel Megiddo has been excavated by different teams for over a century, only the recent implementation of new excavation methods and microarchaeology techniques have brought to light substantial evidence of both bronze- and ironworking.4 This evidence was traced in the recent Tel Aviv University-led excavations in Areas K and Q (Fig. 1).5 Temporal and spatial distribution of metallurgicallyrelated finds indicate that metalworking in the southeastern sector of the tell occurred continuously at least from the end of the Middle Bronze Age (earlier layers have not yet been studied there) until Iron IIB. Within this span of time, a noteworthy development was the introduction and expansion of ironworking alongside continued bronze production (Table 1). Our goals in studying this unparalleled dataset are fourfold: first, J. C. Waldbaum, From Bronze to Iron: The Transition from the Bronze Age to the Iron Age in the Eastern Mediterranean (Göeteborg, 1978). P. M. McNutt, The Forging of Israel: Iron Technology, Symbolism and Tradition in Ancient Society (Sheffield, 1990). J. D. Muhly, R. Maddin, and T. Stech, “The Metal Artifacts,” in Kinneret: Ergebnisse der Ausgrabungen auf dem Tell el-Oreme am See Gennesaret 1982–1985, ed. V. Fritz (Wiesbaden, 1990), 159–75; H. A. Veldhuijzen and T. Rehren, “Slags and the City: Early Iron Production at Tell Hammeh, Jordan and Tel Beth-Shemesh, Israel,” in Metals and Mines - Studies in Archaeometallurgy, ed. S. La Niece, D. R. Hook, and P. T. Craddock, (London, 2007), 189–201; Y. Gottlieb, “The Advent of the Age of Iron in the Land of Israel: A Review and Reassessment,” Tel Aviv 37 (2010): 89–110; S. Bunimovitz and Z. Lederman, “Iron Age Iron: from Invention to Innovation,” in Studies In Mediterranean Archaeology: Fifty Years On, ed. J. M. Webb and D. Frankel (Uppsala, 2012), 103–12; N. Yahalom-Mack and Eliyahu-Behar, “The Transition from Bronze to Iron in Canaan; Chronology, Technology and Context,” Radiocarbon 57 (2015): 285–305. 3 E.g., I. Finkelstein et al., eds., Megiddo V: The 2004–2008 Seasons (Tel Aviv, 2013), and additional bibliography there. 4 N. Yahalom-Mack and S. Shalev, “Metallurgical Activity in Levels K-5 and K-4,” in Megiddo IV: The 1998–2002 Seasons, ed. I. Finkelstein, D. Ussishkin, and B. Halpern (Tel Aviv, 2006); A. Eliyahu-Behar et al., “Metalworking in Area K: A Reevaluation,” in Megiddo V, ed. Finkelstein et al. 5 The Megiddo Expedition is undertaken under the auspices of Tel Aviv University. Consortium Member are The George Washington University, Vanderbilt University, The University of Oklahoma, the Jezreel Valley Regional Project (JVRP), Fuller Theological Seminary, Loyola Marymount University, and Chapman University. The Expedition is directed by Israel Finkelstein (Tel Aviv University), Matthew J. Adams (Albright Institute of Archaeological Research), and Mario A. S. Martin (Tel Aviv University). 2 to demonstrate continuity in metal production in the same area for at least 800 years; second, to investigate when and how the introduction of iron first took place and study its interaction with the existing bronzeworking tradition; third, to understand the socio-political circumstances behind this process; and fourth, to investigate possible connections to cult. In order to meet these goals, we trace the spatial distribution of metallurgical debris (including metal droplets and slag) and metalworking paraphernalia (including pot bellows, crucibles, and tuyères), as well as hoards and cult-related objects, both possibly connected to this industry. We also incorporate the results of the analysis of sediments from potential metalworking contexts, since the use of microarchaeological methods has emerged as most useful for the identification of previously undetected ironworking contexts.6 Results of Lead Isotope Analysis (LIA) performed on several bronze objects are likewise mentioned in order to indicate the use of diverse copper ore sources at various times. Finally, we monitor the increasing numbers of iron vs. bronze tools and weapons in well-stratified contexts at Megiddo as supporting evidence for the crucial technological change of the onset of ironworking and its expression at the level of consumption. Spatial and Temporal Distribution of Metallurgical Remains Bronzeworking Evidence for bronzeworking was found in Area K at the southeastern edge of the mound, in Levels K-11 (Strata XI/X of the Oriental Institute excavations [hereafter: OI]), K-10 (=OI Strata X/IX)7, K-9 (=OI Stratum VIII), K-6 (=OI Stratum VIIA), and K-5 (=OI Stratum VIB). During the late Middle Bronze, Late Bronze, and Iron Ages, this sector of the site functioned as a domestic area (with evidence for bronze production and other industries), with an uninterrupted sequence of superimposed courtyard houses.8 6 See A. Eliyahu-Behar et al., “Iron and Bronze Production in Iron Age IIA Philistia: New Evidence from Tell es-Safi/Gath,” Journal of Archaeological Science 39 (2012): 255–67. 7 At the present stage of research, it is clear that Level K-10 was founded during the later years of MB III and continued uninterrupted into LB I. 8 Y. Gadot et al., “Area K (Levels K-5 and K-4): The 1998–2002 Seasons,” in Megiddo IV, ed. Finkelstein, Ussishkin and Halpern; E. Arie and A. Nativ, “Area K (Part II): Level K-6,” in Megiddo V, ed. I. Finkelstein, et al.; M. A. S. Martin, N. Blockman and J. Bidmead, “Area K (Part I): Levels K-8 and K-7,” in ibid. Metalworking at Megiddo during the Late Bronze and Iron Ages F 55 Figure 1—Aerial photo of the southern part of Tel Megiddo, looking northwest, showing excavation Areas K and Q. The picture also indicates the location of adjacent tombs on the Eastern Slope, Schumacher’s smithy, two metal hoards and relevant cultic objects. Table 1. Stratigraphy, relative chronology and sequence of metalworking at Megiddo. According to radiocarbon dates, the transitions occurred as follows: K-10/K-9 between 1560–1505 BCE; K-9/K-8 between 1290–1210 BCE; K-7/K-6 between 1185–1135 BCE; K-5/K-4 between 1110–1060 BCE; and K-4/K-3 between 976–901 BCE (These dates are based on M. B. Toffolo, et al. Radiocarbon 56 (2014): 221–244). Gray boxes and bold letters represent strata with metalworking remains. *OI = Oriental Institute Excavations Period OI Stratum* Area K Iron IIB Late Iron IIA Early Iron IIA IVA VA-IVB VB K-1 K-2 K-3 Late Iron I Early Iron I LB III LB II VIA VIB VIIA VIIB VIIB VIII K-4 K-5 K-6 K-7 K-8 K-9 X/IX XI/X K-10 K-11 MB III-LB I Late MB ** Transitional early/late Iron IIA Hoards 12/Q/76 1739? Ceramic pot bellows in tombs excavated by the OI Area Q Q-3/2 Q-4 (+Schumacher) Q-5** Q-6 Q-7 Q-8 Q-9 Metalworking Iron +bronze Iron +bronze Bronze +iron Bronze Bronze Bronze Bronze Bronze F 56 Journal of Near Eastern Studies Figure 2—Aerial photo of Level K-9, showing spaces with bronzeworking remains. Level K-9 has thus far yielded more significant evidence of metallurgical activity than Levels K-11 and K-10 (Table 2). Bronzeworking debris and scrap were retrieved from various spaces of the large courtyard house unearthed in this level (Squares N/10, O/9–10 and P/9–10; see Table 2). The majority of these remains originated in the central courtyard/hall (08/K/33) of the house, which was surrounded by rooms on all sides, including an adjoining large space to the northwest (10/K/27) (see Fig. 2). Pillar bases were uncovered in both large spaces, suggesting that they were at least partially roofed. Notably, there is no evidence for bronzeworking in Area K from Levels K-8 and K-7, which are contemporary with the 19th Egyptian Dynasty. However, the OI team found ceramic pot bellows (Figure 3: 2–3), one each in Tombs 3 and 1145A, located on the Eastern Slope of the tell, not far from Area K.9 Based on the ceramic assemblages from the tombs, they are roughly contemporary with Levels K-9 to K-7 (=OI Strata VIII–VIIB). P. L. O. Guy and R. M. Engberg, Megiddo Tombs (Chicago, 1938), Pls. 37:7 and 49:22, respectively. 9 The metallurgical remains from Levels K-6, K-5, and K-4 have already been reported.10 The bronzeworking debris in Level K-6 included several crucible fragments found mainly in Square N/10, in an open space to the west of a domestic building (Figure 4). The fragment of a hitherto unpublished pot bellows, recently found in the same area in the course of baulk removal, also belongs to this level (Figure 3:1). In Level K-5, metallurgical debris, including crucible fragments, tuyères, and waste products (see description below), was found in Squares M–N/9–10 (Figure 5:2–4, 6–7). Analysis of the sediments in the western section of the area showed significant concentrations of copper related to this level. Metallurgical activity in Level K-4 (=OI Stratum VIA) is questionable, as the metallurgical debris may have been residual from the Level K-5 activity. In Area Q, located just to the west of Area K (Figure 1), several bronze droplets and an amorphous bronze chunk were found in Locus 12/Q/26, a floor 10 Yahalom-Mack and Shalev, “Metallurgical Activity in Levels K-5 and K-4”; Eliyahu-Behar et al., “Metalworking in Area K: a Reevaluation.” Metalworking at Megiddo during the Late Bronze and Iron Ages F 57 Table 2. Bronzeworking remains from Levels K-11 to K-9 Locus Reg. No. Square Stratum Locus description 10/K/122 LB 6 O/9 K-11? Ashy debris 14/K/95 AR1 O/10 14/K/34 AR3 P/11 10/K/112 AR1 N/10 08/K/33b AR45 O/9 10/K/60 AR15 O/10 08/K/75 AR19 P/10 08/K/75 AR7 P/10 10/K/53 AR3 O/10 10/K/53 AR6 O/10 10/K/29 AR15 P/9 10/K/29 AR23 P/9 10/K/27 AR15 N/10 10/K/68 AR2 O/9 Context Below masonry-built family tomb K-11 Striated surfaces Space in central portion of building K-10 Occupational accumulation Space in eastern portion of building K-10 Striated surfaces Space in northwestern portion of building K-10/K-9 Striated surfaces Courtyard/Hall 08/K/33 in house K-9 Striated surfaces (including heat- Courtyard/Hall altered sediment) 08/K/33 in house K-9 Occupational debris Courtyard/Hall 08/K/33 in house K-9 Occupational debris Courtyard/Hall 08/K/33 in house K-9 Striated surfaces Courtyard/Hall 08/K/33 in house K-9 Striated surfaces Courtyard/Hall 08/K/33 in house K-9 Accumulation on floor Narrow (storage?) Room 10/K/29 in house K-9 Accumulation on floor Narrow (storage?) Room 10/K/29 in house K-9 Striated surfaces Large Room/Hall 10/K/27 in house K-9 Striated surfaces Room 10/K/68 in house of Level Q-5 (=OI Stratum VB, or transitional VB/ VA-IVB) (Figures 6 and 7:d). These finds may suggest that the location of bronzeworking had shifted slightly to the west during this time. It is notable that remains of ironworking were also found on this floor, as detailed below. Ironworking Iron production remains were found in Area Q; here we discuss remains from Levels Q-5 and Q-4. The former features a large, well-constructed pillared building surrounded by domestic units. In the latter phase, this building went out of its primary use and was converted into several rooms and open spaces. The earliest metallurgical remains related to Level Q-5 comprise a single iron slag cake found in Locus 10/Q/68 (Figures 6 and 7:c). This locus marks the continuation of the above-mentioned Floor 12/Q/26, also associated with Level Q-5 (=OI Stratum VB, or transitional VB/VA–IVB), on which several bronze droplets were found, indicating that iron and bronze were worked side by side in Level Q-4 (=OI Stratum VA–IVB). In Description Crucible slag Droplet Droplet Droplet Crucible rim fragment (Figure 5:1) Chunk Droplet Droplet Scrap Droplet Droplet+rods Droplet+rods Droplet+rods Slag addition, in the northern part of Building 12/Q/99, a dark layer of sediments was found, approximately 10 cm thick and containing a considerable amount of charcoal (Locus 12/Q/81). It covered at least three column bases of the initial phase of the building (Q-5), and was therefore associated with secondary re-use of this space (Figure 8a). The excavation of the black layer included systematic sampling of the sediments (Figure 8b). The samples were analyzed using a Bruker (Tracer III-V) pXRF in order to identify possible metallurgical contamination. Copper concentrations in wt% were calculated using a specifically made calibration curve, produced by adding known amounts of copper oxide to typical Megiddo archaeological sediments, and were compared to control samples representing the naturally-occurring background levels. 11 For each of the sediment samples, we also measured the magnetic fraction, i.e., the wt% 11 The pXRF analysis was performed in the Institute of Earth Sciences, The Hebrew University of Jerusalem. For more details of the method, see Eliyahu-Behar et al., “Metalworking in Area K: a Reevaluation.” 58 F Journal of Near Eastern Studies Figure 3—Pot bellows from Area K, Reg. No. 12/K/130/VS1 (1) and from the tombs on the Eastern Slope, excavated by the OI team (2–3). of magnetic flakes formed in the process of iron forging. These flakes, denoted hammerscales, were collected using a strong magnet. The results showed that the black sediments contained higher concentrations of both hammerscales (i.e., “magnetic fraction”) and copper contamination with respect to control samples (Figure 9), indicating that this layer contained the remains of both iron- and bronzeworking. In order to determine whether this layer represents in situ metallurgical activity, a micromorphological investigation of the sediments was undertaken, a method widely used in archaeological research.12 For this purpose, three blocks of sediment were removed from the northern and western sections of Square C/7 (the northern part of the building). The results showed that this was probably not the original context of deposition of the sediment. Typically, an area where in situ burning takes place is expected to include a substrate showing evidence of heat, and the 12 M. A. Courty, P. Goldberg, and R. Macphail, Soils and Micromorphology in Archaeology (Cambridge, 1989). burnt organic remains (in the case of wood as fuel) are expected to include both charcoal and ash. In the studied feature, however, there was no evidence of a heated substrate, and there was almost no ash despite the presence of large amounts of wood charcoal. The charcoal was randomly oriented and associated with fragments of chalk/limestone (Figure 10). This charcoal and chalk pile also included fragments of unheated bones and occasional fragments of siliceous slag. The overall microstructure of the sediment was highly porous which, considered together with the random orientations, indicate that the black layer may have resulted from secondary dumping of the burnt remains, probably originating from the area east of the building (see below). The micromorphology additionally provided evidence of water-lain fine charcoal micro-layers, in which the black layer thins away from the center of the pile. This pattern indicates that the area was not protected from rain, supporting the stratigraphic field observation regarding the building’s secondary re-use (including partial abandonment) during which the de- Metalworking at Megiddo during the Late Bronze and Iron Ages F 59 Figure 4—Plan of Levels K-6 (a) and K-5 (b). The circles mark the areas where bronzeworking debris was found, including pot bellows, crucible and tuyère fragments, and waste products. funct column bases were covered over and the space remained unroofed. Immediately to the south, but within the same building and level, a couple of tuyère fragments (square in cross-section) were found at approximately the same elevation (Figure 5:7), together with a com- plete ceramic vessel (Figure 5:4). The interior of the vessel was not slagged, but XRF analysis showed copper contamination in the interior rather than exterior of this vessel, suggesting that it was likely used as a crucible. Notably, also at approximately the same elevation, an oven-like installation (10/Q/111) 60 F Journal of Near Eastern Studies Figure 5—Metalworking paraphernalia associated with metalworking in Areas K and Q: Crucible from Level K-10/K-9, Reg. No. 08/K/33/AR45 (1); crucibles from Level K-6, Reg. No. 02/K/44/AR1 (2), Reg. No. 02/K/39/AR2 (3) and Reg. No. 02/K/52/AR6 (4); vessel from Area Q, Reg. No. 12/Q/125/2 (5); tuyères from Area K, Reg. No. 02/K/17/AR6 (6), Reg. No. 02/K/47/AR2.2 (7); tuyère from Area Q, Reg. No. 12/Q/128/2 (8). surrounded by several iron objects was exposed just beyond the southern exterior of the building. In the 2014 excavation season, considerable evidence of iron working was also unearthed just beyond the eastern exterior of Building 12/Q/99 (Figure 6). A sequence of hearths associated with dark industrial debris was exposed in Levels Q-5 to Q-2, mainly in Squares C-D/8, yielding ca. twenty iron slag cakes and fragments from Level Q-5, and ca. thirty from Level Q-4, as well as tuyère fragments, hammerscales, iron prills, and charcoal (Figure 7:b). These hearths were found in proximity to a sequence of large, ca. 1 m-wide ovens. Preliminary results show that the sediment from the hearths was contaminated by copper, confirming that both copper and iron were worked in this area. Based on these findings, we can safely conclude that ironworking took place in this area in Iron IIA, alongside bronzeworking, just outside the Palace 1723 compound, located only ca. 20 m to the west.13 13 Excavated by the OI in their Area B, this compound comprised a large enclosure wall featuring a gate in the north. The enclosure wall surrounded a spacious courtyard with a well-made While iron production started in Level Q-5 (Stratum VB/VA–IVB transition) east of Building 12/Q/99, contemporaneous with its use, the black layer inside the building—a dump which probably originated in the industry adjacent to it—indicates that metalworking in Level Q-4 (Stratum VA–IVB) continued after the building was no longer in primary use. Significantly, ironworking remains in this area were already reported by the first excavator of Megiddo, Gottlieb Schumacher, some one hundred years ago.14 Schumacher identified an iron smithy in the more northern of two rooms (2 m long) just outside the eastern wall of the courtyard of Palace 1723, only ca. 15 m from the spot of the black layer in Area Q. In this room, he recovered a considerable number of iron tools, as well as slag and raw material reported as ashlar structure in the southern end (Palace 1723); see R. S. Lamon and G. M. Shipton, Megiddo I: Seasons of 1925–34: Strata I–V (Chicago, 1939). 14 Attempts to locate these remains in the Vorderasiatisches Museum in Berlin were unsuccessful. Metalworking at Megiddo during the Late Bronze and Iron Ages F 61 Figure 6—Aerial photo of Area Q after the 2014 excavation season. A sequence of hearths and associated tabun-ovens were uncovered east of Building 12/Q/99. lumps of “brown iron ore” and “clay ironstone.”15 The slags and lumps were found piled up together with ash and iron fragments. Based on their elevations and association with the wall of the palace compound, these remains appear to be contemporary with the Iron IIA iron- and bronzeworking in Area Q. Artifacts: Technology and Typology Crucibles Fragments of bronze-melting crucibles from Area K were made of coarse clay tempered with a relatively high proportion of vegetal material, as indicated by elongated cavities.16 Estimated firing temperatures did not exceed 800o C at the crucible’s outer surface, while inside the crucible, a black porous slag-like layer was formed at a higher temperature, resulting from the reaction between the clay and the hot melt. Analysis of 15 G. Schumacher, Tell el-Mutesellim I (Leipzig, 1908), 132, Figs. 191–94, Pls. XXIX and XLII. 16 Eliyahu-Behar et al., “Metalworking in Area K: a Reevaluation.” prills entrapped in the crucible slag-like layer revealed that copper was available, and that the bronzesmiths were not merely involved in bronze recycling, but possibly in alloying copper with tin as well.17 Based on the thick tapering walls of the crucible rim fragments and the thickness of the preserved base (Figure 5:2–4), it appears that most of the Stratum K-6 fragments belong to “flowerpot”-shaped open crucibles. This type of crucible is known from LB III–Iron Age sites,18 and also in Iron II contexts at Tell es-Safi/Gath,19 and Hazor.20 Alternatively, the Level K-10/K-9 fragment in Figure 5:1 appears to have belonged to a hemispherical bowl-shaped crucible. Examples of this type of crucible have been found in the Ibid. N. Yahalom-Mack, Bronze in the Beginning of the Iron Age in the Land of Israel: Production and Utilization in a Diverse EthnoPolitical Setting (Ph.D. dissertation, The Hebrew University of Jerusalem, 2009), 77–82. 19 Eliyahu-Behar et al., “Iron and Bronze Production-Tell es-Safi.” 20 Yahalom-Mack et al., “Metalworking at Hazor: A LongTerm Perspective,” Oxford Journal of Archaeology 33 (2014): 19–45, Fig. 3:6. 17 18 62 F Journal of Near Eastern Studies Figure 7—Bronze- and ironworking remains from Area Q: Iron slag cake (a); magnet with hammerscales collected during the 2014 excavations in Area Q (b); iron slag cake from Level Q-5 (c); bronze droplets from the same context (d). southern Levant in LB II Hazor,21 and in an uncertain context (possibly Stratum VI of the Late Bronze III) at Tel Beth Shean.22 In Cyprus, hemispherical crucibles have been found at such sites as Enkomi, Ambelikou, and Apliki.23 Notably, the hemispherical bowl-shaped examples (ca. 17 cm in diameter and ca. 10 cm deep) are larger than the “flowerpot”-shaped crucibles. Note that the vessel from Area Q (Figure 5:5), identified as a crucible based on the copper contamination of its interior, is unique in shape and also has an extremely low capacity due to its shallow interior. 21 Y. Yadin and A. Shimon, Hazor II: An Account of the Second Season of Excavations, 1956 (Jerusalem, 1960), Pl. CXLVII:11. 22 G. M. Fitzgerald, The Four Canaanite Temples of Beth-Shan (Philadelphia, 1930), Pl. XLIV:12. 23 R. F. Tylecote, “Metallurgical Crucibles and Crucible Slags,” in Archaeological Ceramics, ed. J. S. Olin and A. D. Franklin (Washington, DC, 1982), 96. Bellows The pot bellows mentioned above from the tombs on the Eastern Slope and from Area K are all tall vessels (rather than squat), which would have been operated by hand. Those from the tombs (Figure 3:2–3) have rounded shoulders and a depression below the rim. Similar objects have been found in Israeli, Lebanese, and Syrian coastal sites, as well as in northern Syria.24 Although no ceramic pot bellows have been found in Area Q, the presence of tuyères suggests the use of bellows. It is entirely plausible that during the Iron Age, the latter were made of perishable material since, to our knowledge, no such ceramic items have ever C. J. Davey, “Some Ancient Near Eastern Pot Bellows,” Levant 11 (1979): 101–11; N. Yahalom-Mack et al., “Metalworking in Area G,” in Excavations at Dor, Final Report, Volume II: Area G: The Late Bronze And Iron Ages, ed. A. Gilboa, I. Sharon, and J. R. Zorn (Jerusalem, forthcoming), Fig. 22:4. 24 Metalworking at Megiddo during the Late Bronze and Iron Ages Figure 8—(a) Aerial photo of Building 12/Q/99 in Area Q where evidence of iron working was found. The extent of the “black layer” is indicated in red. The locations of the square tuyère fragments (Figure 5:8) and the vessel/crucible (Figure 5:5) are marked as well. (b) An installation like a tabun-oven was exposed south of the building. A photo of the “black layer” in the course of sampling. F 63 F 64 Journal of Near Eastern Studies Figure 9—Analyses of sediments from the “black layer” in Building 12/Q/99 in Area Q obtained by pXRF. The plot shows copper concentrations against the magnetic fraction (in weight percent), measured for both sediment samples collected from the “black layer” as well as control samples. been found in association with ironworking at Megiddo or elsewhere in the southern Levant. Tuyères The tuyères from Area K (Levels K-6 to K-4) are coarsely made of calcareous clay with varying amounts of foraminifers, and are also tempered with a high proportion of coarse vegetal matter (mostly straw).25 XRF analysis revealed that most of the examined tuyères were similar in composition to the crucibles from Level K-6.26 This similarity may suggest some regularity in the production of the refractories used in the Levels K-6 to K-4 bronze workshop, at the very least in the choice of material. The tuyères from Area K are round in cross-section (ca. 4.5–5.0 cm in diameter), bent, and tapered down to a pointed tip (Figure 5:5–6). The central perforation (bore hole) tapers from back to tip, in order to prevent hot air from being sucked in from the furnace. 25 26 Eliyahu-Behar et al., “Metalworking in Area K: A Reevaluation” Ibid. There are no signs of vitrification and slag formation at any of the preserved tips, suggesting no direct contact between the tuyère and the fire. Their firing temperature has been estimated to be lower than 800o C.27 The round, bent tuyères are known from many bronzeworking contexts related to the LB III–Iron IIA, at such sites as Tel Dan, Hazor, Yoqneam, and Tell es-Safi/Gath. A single tuyère of the bent type was recovered at LB II Sarepta.28 The tuyères from Area Q are square in cross-section, which is typical of ironworking (Figure 5: 8). Similar tuyères have been found in ironworking contexts related to the Iron IIA at Hazor, Tel Beth Shemesh, Tell es-Safi/Gath, and Tell Hammeh in Jordan.29 Ibid. J. B. Pritchard, Sarepta: A Preliminary Report on the Iron Age (Philadelphia, 1975), Fig. 29:8 (=62:4). 29 H. A. Veldhuijzen, Early Iron Production in the Levant; Smelting and Smithing at Early 1st Millennium BC Tell Hammeh, Jordan and Tel Beth-Shemesh, Israel (Ph.D. diss., University College London, 2005); Veldhuijzen and Rehren, “Slags and the City”; Yahalom-Mack et al., “Metalworking at Hazor.” 27 28 Metalworking at Megiddo during the Late Bronze and Iron Ages Figure 10—(a) Scan of a thin section produced from the central part of the black sediment feature (width is 4.2 cm). The black particles are wood charcoal and the white particles are chalk. Note the aggregated nature of the sediment. (b) Scan of a thin section produced from the edge of the black sediment feature (width is 4.2 cm). The black feature is at the center of the image, overlying and overlain by dense brown sediment with anthropogenic debris. Note the absence of large chalk fragments and general laminated appearance. (c) Microphotograph representing the general microscopic appearance of the black sediments in the center of the feature (frame length is 7.2 mm). Note the angular charred wood fragments. Gray colored material is chalk powder (possibly burnt) and gray-brown fragments are burnt soil and/or pottery. Note the black capping of fine charcoal on top of the angular brown (possibly pottery fragment) indicating small-scale water flow within the black sediment. The white areas are voids, i.e., the material is dominated by coarse fragments of burnt wood, chalk, and soil that are not bridged or encapsulated within a sediment groundmass. This spongy microstructure, together with the angularity of charcoal and chalk fragments, indicates little if any transport/reworking. The absence of wood ash indicates that burning did not occur in situ, suggesting this material is the result of dumping of coarse by-products of metallurgical activity that took place elsewhere. (d) Microphotograph representing the general microscopic appearance of the black sediments at the edge of the feature (frame length is 7.2 mm). Note that the charcoal fragments are very small relative to those found at the center of the feature, forming thin black stringers. The stringers are encapsulated by groundmass sediment that is dominated by calcite. These observations suggest that the edge of the feature formed through hyperconcentrated flow of fines (i.e., very low flow energy) from the center of the feature to the edge where sedimentation of the fines formed alternation of calcite-rich and charcoal-rich microscopic laminae. The evidence for flow in this locality further indicates that the black sediment was exposed to rain in an unroofed area. F 65 F 66 Journal of Near Eastern Studies Waste Products Analyses of copper-based waste products using a scanning electron microscope demonstrated that spherical droplets are essentially bronze drops, with an as-cast microstructure, whereas amorphous pieces are heterogeneous in their microstructure, exhibiting different zones in each sample.30 The latter include partial oxidation products shown as cassiterite crystals and cuprite dendrites, which precipitated from the melt during or at the end of the melting process, and were therefore unintentionally produced. Both the droplets and the amorphous pieces are typically formed in remelting activities involving the use of open crucibles. Iron production remains at Megiddo include hammerscales, iron prills, and slag cakes. The latter are formed at the bottom of hearths associated with the forging of blooms or iron objects, and therefore could indicate both smelting and smithing activities.31 Quantification of Bronze and Iron Tools and Weapons In order to trace the development of iron use, we have compiled an up-to-date database of bronze and iron tools and weapons from Iron Age Megiddo retrieved up through the 2014 excavation season.32 The data include only well-stratified and dated material from the OI excavations, as well as published and unpublished material from the recent Tel Aviv University excavations. Although Megiddo has yielded numerous metal objects from the Iron Age, this strategy of focusing on secure contexts alone significantly narrowed the list to a total of 450 objects. The fact that bronze was often recycled, and that iron sometimes corrodes to an unidentifiable form, are two of the reasons why such an inventory can only provide a very general and relative picture, and should only be taken as a broad indication of the process. Figure 11a shows the count of all the objects, beginning in Iron I, as no iron objects at Megiddo predate this period (which commenced in the late 12th century BC). Figure 11b shows only objects that were catalogued as tools or weapons. According to both diagrams (Figures 11a and 11b), iron appeared during Iron I in relatively small Eliyahu-Behar et al., “Metalworking in Area K: a Reevaluation” A. Eliyahu-Behar et al., “Iron Smelting and Smithing,” Journal of Archaeological Science 40 (2013): 4319–30. 32 N. Yahalom-Mack et al., “The Development of Iron Use: An Updated Count of Bronze and Iron Objects from the Southern Levant,” Tel Aviv (forthcoming). Figure 11—Number of bronze versus iron objects from different periods at Megiddo; all objects (a) Tools and weapons only (b). numbers, while bronze was still dominant. During Iron IIA, iron increased considerably, specifically in the “tools and weapons” category. Figure 11b shows that in Iron IIA, iron comprised nearly 60% of all such objects. It can be clearly seen that in both the absolute number of iron objects and in the dominance of iron in relation to bronze, there was a considerable decrease after Iron Age IIA. This is apparent both from the general count (Figure 11a) and from the count of tools and weapons alone (Figure 11b). Although based on a small database, the strategy of utilizing only contextually secure samples allows us to safely suggest that iron production during Iron IIA was directly related to the developing local administration. The data at hand is insufficient to determine the nature of production during Iron IIB (8th century BC). It is clear, however, that following the violent destruction of the city at the end of this period by the Assyrians, iron production did not continue at the same level of intensity, and bronze regained its dominancy. 30 31 Metal Hoards Two metal hoards were found in proximity to each other near the southern edge of the tell. One is a col- Metalworking at Megiddo during the Late Bronze and Iron Ages F 67 Figure 12—Hoard 1739 being uncovered by the OI expedition. lection of bronze objects found by the OI team in their Area CC, Locus 1739, dated to Iron I.33 Another (12/Q/76), a hoard of bronze and iron objects unearthed during the 2012 season in Area Q, is related to Level Q-7 (=Stratum VIA).34 Based on the wealth of metal objects and the proximity of these two hoards to the metalworking contexts in Areas K and Q, we view them as possible products of this metallurgical activity. Hoard from Locus 1739 A collection of bronze objects was found in a domestic area of Area CC, Locus 1739 (Figure 12). According to the find register, there were at least thirty-two copper/bronze artifacts in this locus, including nineteen bowls, two strainers, five jugs, three spearheads, two lugged axes, and a single double axe.35 The excavators attributed Locus 1739 to Stratum VI, without distinguishing between its two phases (VIB and VIA). Lilly Gershuny examined the vessels and concluded, based on stylistic considerations, that they were likely produced by a single workshop during the time of Stratum VIB, while Ora Negbi assigned the hoard to Stratum VIA on the basis of two other metal hoards that were retrieved from this stratum (a hoard found by Yigael Yadin in the 1960s, and a silver hoard from Locus 2012 of the OI excavations).36 As the hoard from Locus 1739 is comprised of only bronze objects (in contrast to Hoard 12/Q/76 from Stratum VIA described below, and to the Yadin hoard that also contained iron artifacts37), it is more likely contemporary Loud, Megiddo II, 150, with references to plates. L. Gershuny, Bronze Vessels from Israel and Jordan (München, 1985), 41–42; O. Negbi, “The Continuity of the Canaanite Bronzework of the Late Bronze Age into the Early Iron Age,” Tel Aviv 1 (1974): 167. 37 See A. Zarzecki-Peleg, Yadin’s Expedition to Megiddo, Qedem 56 (Jerusalem, 2016), 305–18. 35 36 33 G. Loud, Megiddo II: Seasons of 1935–1939 (Chicago, 1948), 150, Pls. 173: 11–13, 183: 15–17, and 189: 4–11. 34 E. Hall et al. “An Early Iron Age Metal Hoard from Area Q,” in Megiddo VI. The 2010–2014 Seasons, eds. I. Finkelstein, E. Cline, and M. A. S. Martin (Tel Aviv, forthcoming). 68 F Journal of Near Eastern Studies unworked semi-precious stones. Various small bronze objects were also present, as were silver earrings. Nine iron knives, including four with bronze rivets, were stacked directly beside the bronze bowls.38 Taken together, the wealth of metals, the unworked stones, the use of metal to chain the beads, and particularly the location of the hoard in a traditional metalworking area of the tell, all suggest that the metal items in the hoard were possibly produced in the nearby bronze and iron smithy. Although this cannot be proven at this point, see below concerning the source of the copper in the bronze rivets belonging to an iron knife from this hoard. Origin of the Copper Figure 13—Selected objects from Hoard 12/Q/76: Bi-metallic knife (a); bronze bowls and scale pans as recovered (b); stone beads (c), some chained with a metal wire (left). Note the chunk of unworked carnelian among the beads on the right. with the considerable bronzeworking remains of Level K-5 (=Stratum VIB), as suggested by Gershuny. Hoard 12/Q/76 This hoard was exposed in Square I/1 very close to the edge of the tell and bordering on Area CC of the OI excavations, where Hoard 1739 was found. It included two stacked bronze bowls with two scale pans covering the contents of the bowls (Figure 13). The larger exterior bowl was covered with remains of cloth, indicating the deliberate action of hoarding. The contents of the bowls included numerous beads made of carnelian and other materials, some chained with a metal wire, as well as a few pieces of Lead Isotope Analysis (LIA) was performed on several bronze artifacts from Areas K and Q.39 The results showed a clear dichotomy between the analyzed objects dated to Iron I and those dated to LB II. Seven out of eight artifacts from Stratum VIA (=late Iron I) were consistent with Arabah (Wadi Faynan in Jordan or Timna in Israel) ore deposits, as was one of three artifacts from Stratum VIIA (=LB III). Three artifacts dated to LB II, however, had lead isotopes consistent with Aegean ores. Surprisingly, only one of the sampled objects, dated to the Iron I, had lead isotopes consistent with copper ores from Cyprus, despite the fact that the latter is considered the likeliest source of copper for the southern Levant during the Late Bronze Age.40 While based on a small and not representative sample, this pattern indicates the shift from the use of western copper sources during the Late Bronze Age to the use of copper from the Arabah during Iron I. Notably, the sampled objects from Stratum VIA include one of the bronze bowls and bronze rivets from two different iron knives of Hoard 12/Q/76. The source of the raw material of the rivets indicates that the bimetallic knives from the hoard, although typologically closely paralleled in Cyprus,41 were not riveted there, and most likely not produced there either. While we Hall et al., “An Early Iron Age Metal Hoard from Area Q.” N. Yahalom-Mack, I. Segal, and I. Finkelstein, “Provenancing Copper-based Objects from the Bronze and Iron Ages in the Southern Levant” (forthcoming). 40 N. Yahalom-Mack et al., “New Insights into Levantine Copper Trade: Ingots from the Bronze and Iron Ages in Israel,” Journal of Archaeological Science 45 (2014): 159–77. 41 E.g., S. Sherratt, “Commerce, Iron, and Ideology: Metallurgical Innovation in 12th–11th Century Cyprus,” in Proceedings of 38 39 Metalworking at Megiddo during the Late Bronze and Iron Ages cannot know if the knives were imported to Megiddo as an end-product, or whether they were produced at the site (both the iron blade and the riveting of the blade to the handle) as an imitation of the Cypriot original, we can most likely conclude that this was a Canaanite rather than a Cypriot product. Discussion The accumulated evidence presented here indicates continuous metalworking in the southeastern sector of Tel Megiddo from the late Middle Bronze until Iron IIB (Table 1). No other area of the tell produced a similar metalworking sequence. Bronzeworking is represented by the distribution of metallurgical debris in Area K in Levels K-11, K-10, K-9 (Middle Bronze III to Late Bronze IIA), and in Levels K-6 and K-5 (Late Bronze III and early Iron I, respectively). The occurrence of ceramic pot bellows in the adjacent tombs on the eastern slope, which probably date to the time of Levels K-8 and K-7 (Late Bronze IIB), suggests that there was possibly some bronzeworking during LB II. The question of iron production in Iron I remains, in the meantime, open. In Area Q, we have evidence for the production of both bronze and iron in Levels Q-5 and Q-4, dating to Iron IIA; however, excavation did not proceed below these levels and thus it is so far unknown whether or not earlier iron production remains might be uncovered. Notably, excavation in Iron I levels in Area K revealed evidence of bronzeworking, but none of iron production. Bronzeworking Based on the remains from Area K, it appears that the bronzeworking activity was characterized by the use of open crucibles and hand-operated pot bellows. While we cannot determine the scale of production, notably, these metallurgical remains were found on floors within domestic units, rather than in public or industrial venues. The relatively small number of artifacts specified in Table 2 may be related to the high value of bronze; floors may have been routinely cleaned, and bronze fragments collected for remelting, as is performed even today in traditional metal smithies in the International Symposium: Cyprus in the 11th Century B.C., ed. V. Karageorghis (Nicosia, 1994). F 69 India.42 Based on these and other ethnographic observations, the evidence may suggest that the bronzeworking activity was conducted at a household level.43 The analysis of crucible fragments from Level K-6 and the remains from Level K-5 allow reconstructing the bronzeworking activity there. Crucibles were filled with metal (either bronze scrap, or copper and tin separately) and charcoal, placed in shallow pits, and then heated. Ceramic pot bellows were placed nearby. A top made of perishable material such as skin was tied around the vessels, probably utilizing the depression below the rim.44 By lifting the top and pressing it down, air was sucked into the vessel and forced down through the hole at the bottom and into a pipe made of perishable material. The pipe carried the air from the bellows and through a ceramic end-piece—the tuyère. The tuyère in turn received the air at ground level and directed it downwards onto the crucible. Once the metal was molten, the crucible was lifted and the hot melt was poured into a mold. Notably, since no molds were found in Area K, we assume that they were made of unbaked clay or sand. The lack of substantial evidence of metallurgical activity from Levels K-8 and K-7 in the form of droplets or other remains may be related to the contingencies of the excavation, but could have some significance since these levels are related to the time of the 19th Egyptian Dynasty (=LB IIB), during which time evidence of bronzeworking activity in Canaan is generally scarce. The evidence includes remains from Tel Rehov and possibly Tel Zeror, as well as along the northern coast at Tel Dor, and possibly Tel Nami.45 Significantly, at Tel Rehov the small workshop was identical to bronze workshops uncovered at 42 T. E. Levy et al., Masters of Fire – Hereditary Bronze Casters of South India (Bochum, 2008). 43 E.g., C. L. Costin, “Craft Specialization: Issues in Defining, Documenting, and Explaining the Organization of Production,” Archeological Method and Theory 3 (1991): 1–56. 44 Davey, “Some Ancient Near Eastern Pot Bellows.” 45 Yahalom-Mack, Bronze in the Beginning of the Iron Age, 102– 109. Tel Rehov: N. Yahalom-Mack, “Metals and Metal-working,” in The Excavations at Tel Reḥov, 1997–2012, ed. A. Mazar and N. Panitz-Cohen (Jerusalem, forthcoming). Tel Zeror: K. Ohata, Tel Zeror II: Preliminary Report of the Excavation, Second Season 1965 (Tokyo, 1967), 29. Tel Dor: Yahalom-Mack et al., “Metalworking in Area G,” Fig. 22:4. Tel Nami: M. Artzy, “Nami: A Second Millenium International Maritime Trading Center in the Mediterranean,” in Recent Excavations in Israel: A View to the West, ed. S. Gitin and W. Dever (New York, 1995), 17–41; M. Artzy, “Cult and Recycling of Metal at the End of the Late Bronze Age,” in Periplus Festschrift für Hans-Günter Buchholtz zu seinem achtzigsten 70 F Journal of Near Eastern Studies Qantir/Pi-Ramesses in the Nile Delta, utilizing canals rather than pits to hold the crucible in place, as well as unique tuyères fabricated differently than the Canaanite ones.46 During the time of the 20th Egyptian Dynasty (=LB III), bronzeworking remains are also relatively infrequent; they have been found at Tel Dan, Akko, and at the Egyptian outposts at Beth-Shean and Tel Mor.47 Significantly, the bronzeworking at Tel Mor also included the use of canals like those at Qantir/ Pi-Ramesses, suggesting Egyptian rather than Canaanite bronzeworking traditions.48 Could this frequency and character of the metalworking remains in LB IIB-III suggest some Egyptian restriction over the working of bronze? In fact, a significant increase in bronzeworking occurred after the withdrawal of Egypt from Canaan, as evidenced by remains of this activity detected at twelve different sites located in various geographical regions.49 At Megiddo, this is evident in Level K-5, within which significant bronzeworking remains were found. The rich collection of bronze objects found in Hoard 1739 was possibly the product of this activity. LIA of three bronze artifacts from Area K at Megiddo shows that during the LB II, copper was imported from the Aegean, while during the LB III–Iron I transition in the 12th century BC, bronze smiths were using mainly copper from the Arabah. Excavations Geburstag am 24. December 1999, ed. P. Åström and D. Sürenhagen (Jonsered, 2000), 27–32. 46 N. Yahalom-Mack, “Egyptian Bronzeworking Practices in Late Bronze Age Canaan,” BASOR 374 (2015): 103–14. 47 Tel Dan: D. Ilan, Northeastern Israel in the Iron Age I: Cultural, Socioeconomic and Political Perspectives (Ph.D. diss., Tel Aviv University, 1999); R. Ben-Dov, “Craft Workshops at Tel Dan,” Eretz-Israel 30 (2011): 77–94 (in Hebrew); Akko: M. Dothan, “Notes and News, Akko,” Israel Exploration Journal 31 (1981): 110–12; M. Dothan, “The Significance of Some Artisans’ Workshops along the Canaanite Coast,” in Society and Economy in the Eastern Mediterranean (c.1500–1000 B.C.): Proceedings of the International Symposium held at the University of Haifa from the 28th of April to the 2nd of May 1985, ed. M. Heltzel and E. Lipinski (Leuven, 1988); Tel Mor: T. Barako, Tel Mor: A Late Bronze Age Egyptian Outpost in Southern Canaan. The Excavations Directed by Moshe Dothan 1959–1960 (Jerusalem, 2007). Beth Shean: N. Yahalom-Mack, “The Typology of the Metal Objects,” in Excavations at Tel Beth-Shean 1989–1996, Volume III: The 13th–11th Century BC Strata in Areas N and S, ed. N. Panitz-Cohen and A. Mazar (Jerusalem, 2009), 581–82. 48 Yahalom-Mack, “Egyptian Bronzeworking Practices.” 49 Ilan, Northeastern Israel in the Iron Age I, 220–30. YahalomMack, Bronze in the Beginning of the Iron Age, 102–109; Ben-Dov, “Craft Workshops at Tel Dan.” at Faynan and Timna show that major mining and smelting activities were undertaken in the Arabah valley already during Iron I.50 The availability of copper from the Arabah certainly contributed to the spread of bronzeworking during this time, and it is hypothesized that against this backdrop, bronzesmiths began to pursue the working of iron.51 Ironworking Until recently, there was insufficient evidence for a full reconstruction of the ironworking activity at Megiddo, despite the “black layer” in Building 12/Q/99 which, according to the micro-morphological examination, was re-deposited. However, the 2014 excavation season unearthed hearths and industrial refuse which appear to indicate in-situ iron working, yielding two preliminary conclusions. First, the main activity took place in an open space east of Building 12/Q/99, not far from Palace 1723 of Iron IIA. This reinforces Schumacher’s observations concerning ironworking in two small rooms adjoining the exterior of the eastern wall of the palace compound. Second, sediments contaminated with copper show that bronzeworking took place here alongside the ironworking. Association with Cult The long-acknowledged association between metalworking and cult (for Canaan, note for example the finds at Hazor, Tel Rehov, and Tell es-Safi/Gath52) appears to be distinctly manifested at Megiddo. In fact, for the Late Bronze, Iron I, and early Iron IIA Ages, there is a clear spatial overlap between structures fully devoted to cult (to differ from the household level) 50 T. E. Levy et al. “Early Bronze Age Metallurgy: A Newly Discovered Copper Manufactory in Southern Jordan,” Antiquity 78 (2004): 865–79; T. E. Levy et al., “High Precision Radiocarbon Dating and Historical Biblical Archaeology in Southern Jordan,” Proceedings of the National Academy of Sciences 105 (2008): 16460–65; E. Ben-Yosef et al., “A New Chronological Framework for Iron Age Copper Production at Timna (Israel),” BASOR 367 (2012): 31–71. 51 N. Yahalom-Mack and I. Segal, “The Origin of the Copper Used in Canaan during the Late Bronze - Iron Age transition,” in Papers in Honor of B. Rothenberg, ed. E. Ben-Yosef (Tel Aviv, forthcoming). J. J. Bimson and J. M. Tebes, “Timna revisited: Egyptian chronology and the copper mines of the southern Arabah,” Antiguo Oriente 7 (2009). 52 N. Yaholom-Mack and Adi Elyahu-Behar, “The Transition from Bronze to Iron in Canaan,” Radiocarbon 57/2 (2015). Metalworking at Megiddo during the Late Bronze and Iron Ages and metallurgical activity. Evidence for cult activity beyond the domestic sphere in the southeastern sector of Megiddo (Areas K and Q and their vicinity) includes three main features.53 First, there is Monumental Temple 2048, which functioned throughout the Late Bronze and Iron I, until the destruction of city in the late Iron I.54 Second, we have a massive, public, late Iron I building, with two small, well-cut pillars still standing in-situ, seeming matzevah stones. The building was uncovered in 2014 in the southern sector of Area Q, close to the spot where a late Iron I hoard was found in 2012. And third, a unique, large, and well-constructed structure dating to Level Q-5 (the transition phase from the early to late Iron IIA) has recently been uncovered at the highest point of the mound. The central hall of the building is divided into four aisles by three rows of pillars, six in each row. The two side rows are made of monolithic pillars, ca. 2 m. high. A large number of cult stands, chalices, and figurines were unearthed in the vicinity of this building.55 A cache containing horned stone altars and shrine models, uncovered by the OI team between the present Areas K and Q (Figure 1), may have originally belonged to this structure.56 It is noteworthy that ironworking in Level Q-5 took place immediately adjacent to the east of this building, and that Schumacher’s smithy is located a few meters to its west. A change in cult activity at Megiddo is noticeable starting in the late Iron IIA: elaborate structures fully devoted to cult disappear, and the focus of cult activity shifted to rooms within administrative buildings. This seems to coincide with the strengthening of central authority of the Northern Kingdom at Megiddo (e.g., construction of ashlar palaces). A. Kleiman et al., “Cult Activity at Megiddo in the Iron Age: New Evidence and a Long-Term Perspective,” Zeitschrift des Deutschen Palästina-Vereins (forthcoming). 54 A. Kempinski, Megiddo: A City-State and Royal Centre in North Israel (Munich, 1989), 77–83. 55 Kleiman et al. “Cult Activity at Megiddo in the Late Bronze and Iron Ages.” 56 H. G. May, Material Remains of the Megiddo Cult (Chicago, 1935), Pl. XII–XV. D. Ussishkin (“Schumacher’s Shrine in Building 338 at Megiddo,” Israel Exploration Journal 39 [1989]: 149–72) associated this cache with Building 338, located to the north of Area K, yet, recent excavations confirmed the dating of the structure to Stratum IVA, as originally proposed by the OI team. 53 F 71 Bronze/Iron Transition The transition from bronze to iron has traditionally been understood as a trajectory of improvement, with the latter metal replacing the former due to its advantages, particularly its strength. However, other factors played a role in this development, mostly technological, but also social and cultural. A simple linear evolution of replacement is not an accurate scenario; a more complex situation existed, as evidenced from the finds at Megiddo, as well as other sites. The question of whether iron was superior to bronze is not a trivial one. As iron produced in the bloomery process during the Iron Age was not liquefied, and therefore could not be cast into shape, bronze was retained as a complementary metal. Iron would have been stronger than bronze only with specific heat treatments that, as far as we know, were not systematically employed during the Iron Age.57 However, it may be assumed that the use of iron was advantageous, at the very least because of the availability of its raw material and the control over the entire production process.58 Systematic iron production began in the southern Levant during the late 13th–early 12th centuries BC, occurring simultaneously in the Ajlun region in modern Jordan; likely in the area of the Warda iron ore deposit, with the production of iron/steel bracelets (found, for example, in the Baqa’a burials in Jordan59); and in Cyprus, with the production of bi-metallic knives.60 Notably, the former iron products were of a prestige nature; the latter were not purely utilitarian, but also apparently had prestige value. Utilitarian Yaholom-Mack and Elyahu-Behar, “Transition from Bronze to Iron in Canaan”; A. Eliyahu-Behar and N. Yahalom-Mack are preparing a study on this topic. 58 Yaholom-Mack and Adi Elyahu-Behar, “Transition from Bronze to Iron in Canaan.” 59 M. R. Notis et al. “The Metallurgical Technology: the Archaeometallurgy of the Iron IA Steel,” in The Late Bronze and Early Iron Ages of Central TransJordan; The Baq’ah Valley Project, 1977–1981, ed. P. E. McGovern (Philadelphia, 1986), 272–78. 60 S. Sherratt, “Commerce, Iron, and Ideology: Metallurgical Innovation in 12th–11th Century Cyprus,” in Proceedings of the International Symposium: Cyprus in the 11th Century B.C., ed. V. Karageorghis (Nicosia, 1994), 59–108; S. Pickles and E. Peltenburg, “Metallurgy, Society and Bronze/Iron Transition in the East Mediterranean and the Near East,” Report of the Department of Antiquities, Cyprus (1998): 67–100; J. D. Muhly, “Texts and Technology: the Beginning of Iron Metallurgy in the Eastern Mediterranean,” in Ancient Greek Technology: Proceedings of the 2nd International Conference on Ancient Greek Technology (Oct. 2005), ed. T. P. Tassios and C. Polyvou (Athens, 2006), 19–31. 57 72 F Journal of Near Eastern Studies iron objects (tools and weapons) are not found before Iron Age I (11th century BC). The most intriguing questions are how this new technology began, subsequently spread, and what were the specific social conditions for its adoption. The latter two aspects can be examined with the Megiddo metallurgical dataset. The metalworking evidence from Megiddo paints a complex picture, which can be examined in terms of “continuity” and “discontinuity.”61 Bronzeworking can be generally associated with processes of “continuity” throughout most of the second millennium BC and into the first millennium BC. However, ironworking is associated with “discontinuity,” expressed in the appearance of large-scale centralized iron production in the early first millennium BC. This “discontinuity”—when ironworking remains per se first appear and the number of iron objects increase dramatically—occurs at Megiddo during the transition from Stratum VIA to VB, signifying the change from cultural traditions and political organization of second millennium Canaan to those of the early first millennium kingdom of Israel. This transformation included a shift from city-states dominating limited territories to large-scale territorial kingdoms, which impacted the long-term history of the site.62 A major alteration occurred in the layout of the settlement and in almost every aspect of its material culture.63 The particular area of the tell where metalworking remains dating to this period were found—the southeastern sector of the settlement—changed from residential to palatial. While bronzeworking is associated with the earlier form of settlement, systematic iron production is related to the latter. The strongest evidence for the association between iron production and the central Israelite administration is the sharp decrease in iron 61 “Continuity” refers to gradual changes, while “discontinuity” refers to social and/or population change, occurring as longer-term historical changes. See F. Braudel, “Histoire et Sciences sociales: La longue durée,” Annales: Histoire, Sciences Sociales 13 (1958): 725–53; V. Roux, “Spreading of Innovative Technical Traits and Cumulative Technical Evolution: Continuity or Discontinuity?” Journal of Archaeological Method and Theory 20 (2013). 62 I. Finkelstein, “City States and States: Polity Dynamics in the 10th-9th Centuries B.C.E.,” in Symbiosis, Symbolism and the Power of the Past: Canaan, Ancient Israel, and their Neighbors, ed. W. G. Dever and S. Gitin (Winona Lake, IN, 2003), 75–83. 63 N. Franklin, “Revealing Stratum V at Megiddo,” BASOR 342 (2006): 95–111; E. Arie, “In the Land of the Valley”: Settlement, Social and Cultural Processes in the Jezreel Valley from the End of the Late Bronze Age to the Formation of the Monarchy (Ph.D. diss., Tel Aviv University, 2011). objects, both in absolute numbers and in relation to bronze, alongside the significant increase in bronze objects following the destruction of the Stratum IV city by Assyria in 732 BC. The location of the iron (as well as bronze) production venue, just east of Palace 1723 and its related courtyard, may suggest that the ironworking activity was attached to the central administration, although alternative explanations may be considered. The location of bronzeworking during most of the second millennium in courtyard houses suggests a different, domestic-centered production mode. Analogous processes may be noted at other sites. For example, ironworking was associated with an administrative building at Beth Shemesh during Iron IIA, while bronzeworking during Iron I was carried out in courtyards and between houses.64 While it seems that the adoption of ironworking was the result of a deliberate act on the part of the central authority, the mechanism by which iron was first introduced during the initial phase in Iron I is more ambiguous. Although iron objects have been found in many Iron I sites in Canaan (Megiddo among them), none of the actual iron production remains unearthed thus far predate Iron IIA.65 We have mentioned earlier that during Iron I, the political vacuum left after the departure of the Egyptians provided favorable conditions for a floruit of bronzeworking; it apparently also opened new vistas for innovation and provided a fertile platform for the absorption of new technologies. It remains to understand how the new technology of ironworking reached Megiddo. Was it prompted by the imitation of iron objects that came in through trade? Was the knowledge acquired from itinerant (iron?) smiths? Another pertinent question is whether the production of iron at its inception in Iron Age I was at all related, as a processually evolutionary development, to that of Iron Age IIA. Were the differences merely quantitative or also qualitative? We suggest that Hoard 12/Q/76 provides some indication that iron was in fact produced at Megiddo during Iron I. LIA showed that bronze rivets from two of the iron knives from the hoard were made of copper from the Arabah, indicating that the knives Bunimovitz and Lederman, “Iron Age Iron”; E. Grant and G. E. Wright, Ain Shems Excavations (Palestine), Part V (Haverford, PA, 1939), 56. 65 Yaholom-Mack and Adi Elyahu-Behar, “Transition from Bronze to Iron in Canaan”: 290–93. 64 Metalworking at Megiddo during the Late Bronze and Iron Ages were more likely locally produced than imported from Cyprus. There is always a possibility that the blades were merely riveted at Megiddo; however, according to LIA, bronze rivets from similar blades found at Khirbet Qeiyafa were also made from Arabah copper.66 The unworked carnelian and other semi-precious stones, the use of metal to string the beads, and the location of the hoard in a traditional metalworking area all suggest that the objects are associated with their immediate locus of production. If we are correct in our interpretation of this hoard, and since both bi-metallic knives of Cypriot style and iron bracelets were present in the assemblage, we would argue that local bronzesmiths adopted iron technology through the “process of copy” and/or contact with itinerant smiths.67 The involvement of bronzesmiths with ironworking in Iron I is strengthened by the association between bronze and ironworking during Iron II. This is evident from the metallurgical remains and sediments in Area Q at Megiddo and at other sites, such as Tell es-Safi/Gath68 and Hazor,69 as well as from the continuity in the location of metalworking during the Bronze and Iron Ages, and the continuity in bronzeworking during Iron IIA. Technological innovation is complex and related to varying reasons and circumstances. The borrowing of one or two new technical traits into a pre-existing technological system can be the result of the expertise of a sole craftsman with the ability to influence other practitioners.70 However, the adoption and embedding of an entirely new technology, such as ironworking, would entail other factors. The introduction of ironmaking involved the adoption of a new technology, a new chaîne opératoire, the acquisition of resources from hitherto untapped locations and methods, and a different organization of production. 66 Yahalom-Mack, Segal, and Finkelstein, “Provenancing Copper-based Objects from the Bronze and Iron Ages.” 67 M. J. O’Brien and R. A. Bentley, “Stimulated Variation and Cascades: Two Processes in the Evolution of Complex Technological Systems,” Journal of Archaeological Method and Theory 18 (2011): 309–37. V. Roux, “Spreading of Innovative Technical Traits”: 312–30. 68 Eliyahu-Behar et al., “Iron and Bronze Production at Tell es-Safi.” 69 Yahalom-Mack et al., “Metalworking at Hazor.” 70 V. Roux, “Technological Innovations and Developmental Trajectories: Social Factors as Evolutionary Forces,” in Innovation in Cultural Systems. Contributions from Evolutionary Anthropology, ed. M. J. O’Brien and S. J. Shennan (Cambridge, MA, 2010), 217–34. F 73 It is possible that, in our case, the Iron I bronzeworkers who adopted the innovative iron technology in addition to their usual practices, engaged only in the forging of iron. Subsequently, during Iron IIA, when ironmaking came to dominate, the complete chaîne opératoire of iron production, including the smelting of iron from its ore, was performed. A caveat to this suggestion, of course, is that to date, no iron production remains from Iron I at Megiddo have been found. Since the Iron IIA metallurgical remains may have been the byproducts of either iron forging or smelting, this hypothesis remains to be explored with further excavation and analysis of slag. Such a scenario would suggest that while some ironworking traits were adopted by individuals during Iron I, the technology as a whole was fully adopted during Iron IIA, the impetus likely being the agency of the central authority and its ability to allocate the necessary manpower.71 Anthropological investigation has clearly shown that knowledge of a new technology is not a sufficient cause for its adoption and that complex social, cultural, political, and ideological factors play a role as well.72 In the case of Megiddo, the evidence at hand indicates that the direction of change was initially (Iron I) from the “bottom up” (the initiative of local craftsmen), and subsequently (Iron IIA) from the “top down” (initiative of the kingdom’s administration), part and parcel of far-reaching political and social changes in this period, as seen in all aspects of material culture at Megiddo and beyond.73 Conclusions The unique sequence of metalworking in the southeastern sector of Tel Megiddo indicates both continuity of metallurgical activity from the late Middle Bronze to Iron IIA and discontinuity indicated by the adoption of iron technology during Iron IIA. It 71 V. C. Pigott, “The Emergence of Iron Use at Hasanlu,” Expedition 31 (1989): 67–79. 72 M. B. Schiffer, “Transmission Processes: A Behavioral Perspective,” in Cultural Transmission and Archaeology: Issues and Case Studies, ed. J. M. O’Brien (Washington, D.C., 2008), 102–11. 73 R. Torrence and S. E. van der Leeuw, “Introduction: What’s New About Innovation,” in What’s New? A Closer Look at the Process of Innovation, ed. S. E. van der Leeuw and R. Torrence (London, 1989), 12–13; C. Costin et al. “The Impact of Inca Conquest on Local Technology in the Upper Mantaro Valley, Perú,” in ibid., 107–39. 74 F Journal of Near Eastern Studies also supports the notion that bronze- and ironworking were closely related. The Megiddo finds shed new light on the introduction of iron to the southern Levant. Indirect evidence suggests that iron was already worked by bronze smiths during Iron I. This metallurgical shift would have occurred after the Egyptian administration was no longer present in the region, but concurrent with the floruit of the Arabah mining and bronzeworking. Our finds also suggest that the increased production of iron at Megiddo was related to the fundamental change in the character of the site from a late-Canaanite city-state in Iron I to an administrative center of the Northern Kingdom of Israel in the Iron IIA. This alludes to the possibility that even though the know-how existed previously, actual iron production was likely a socio-cultural adaptation, probably related to new administrative/political attitudes toward metallurgy. Notably, with the demise of the Northern Kingdom at the end of the 8th century BC, the number of iron objects decreased alongside a renewed floruit of bronze items.