Khirbat Nuqayb al-Asaymir and Middle Islamic Metallurgy in Faynan: Surveys of Wadi al-Ghuwayb and Wadi al-Jariya in Faynan, Southern Jordan

Ian W. N. Jones Thomas E. Levy Department of Anthropology Department of Anthropology University of California, San Diego University of California, San Diego 9500 Gilman Drive 9500 Gilman Drive La Jolla, CA 92093–0532 La Jolla, CA 92093–0532 [email protected] [email protected]

Mohammad Najjar Levantine Archaeology Laboratory University of California, San Diego 9500 Gilman Drive La Jolla, CA 92093–0532 [email protected]

Khirbat Nuqayb al-Asaymir, in the Faynan region of southern Jordan, is one of the best-preserved Middle Islamic–period copper production sites in the southern Levant. Two seasons of survey in Wadi al-Ghuwayb and Wadi al-Jariya revealed much about the site and the landscape surrounding it, which during the Middle Islamic period was ­exploited primarily for mining and pastoral activities. The finds from these surveys, focusing on those collected at Khirbat Nuqayb al-Asaymir, are presented here. A new model explaining the revival of copper mining and smelting activities in Faynan dur- ing the Middle Islamic period, linking copper from Faynan to the expanding Jordanian sugar industry, is also proposed.

introduction scholars working on the Islamic periods in Jordan seem to be aware that this revival took place, and those n the early 13th century a.d., the copper min­ who are aware of it have engaged it only on a lim­ ing and smelting industry in southern Jordan’s ited level (e.g., Milwright 2008: 123–24; 2010: 149; I Faynan district was revived, after a hiatus of at Walmsley 1997: 349). least 500 years. Archaeological remains of this revival Recently, though, two synthetic works on the have been known in the region since Glueck’s (1935: ­archaeology of Faynan have been published, both of 30–32) surveys in the early 1930s, yet few attempts which attempt to explain the underlying economic have been made to investigate this phase of industry in motivations behind the Middle Islamic–period re­ Faynan or to explain it in economic terms. Researchers vival of the copper smelting industry (Newson et al. working in Faynan have, somewhat understandably, 2007; Weisgerber 2006). However, neither of these focused on the important and often more productive offers a particularly satisfactory explanation. One earlier periods of copper exploitation. Likewise, few simply dismisses the copper industry in this period as

67 68 JONES, LEVY, AND NAJJAR BASOR 368 unsuccessful and unimportant (Weisgerber 2006: 25– the same time. Likewise, the use of political terminol­ 27), and the other attempts to connect it to a later and ogy can obscure cultural changes at smaller scales. In almost certainly unrelated event in Egyptian monetary place of this, Whitcomb (1992a: 386) argues for the history (Newson et al. 2007: 364–65). Despite these use of a neutral, archaeological periodization based attempts, the economic motivations underlying the re­ on century divisions, rather than dynastic changes. vival of the Faynan copper industry remain somewhat To some extent, this paralleled similar proposals in obscure. history; Hodgson (1974: 234), for example, had di­ This paper seeks to address the problem through vided the 7th–13th centuries into the High Caliphal discussion of Khirbat Nuqayb al-Asaymir (Arabic for and Middle periods. While we agree that neutral, ar­ “Ruin of the Small Black Pass”; henceforth KNA),1 chaeological periods are useful, at least four different the largest and best-preserved Middle Islamic–period archaeological periodization schemes have been pro­ smelting site in Faynan. We present here a preliminary posed for Jordan alone, most based on Whitcomb’s analysis of data collected during intensive, full-cover­ (1992a) proposed chronology (fig. 1). age surveys conducted in 2002 and 2007 by the Edom The first scheme is what we refer to as the “tra­ Lowlands Regional Archaeology Project (ELRAP) ditional” chronology of the Islamic period of central and its predecessor, the Jabal Hamrat Fidan Project and southern Jordan, used by, among others, Lindner (JHF). We then offer a new explanation for the re­ (e.g., Lindner et al. 1998: 238) and Brown (1992: 54). vival of the Faynan copper industry, this one rooted The second periodization is Whitcomb’s (1992a: 386), in the economic history of Jordan between the 13th offered as part of his call for the use of a neutral chro­ and 15th centuries a.d. Specifically, we propose a link nology. This chronology has probably seen the widest between the copper industry of Faynan and large-scale adoption; and since its publication, several attempts sugar production near the Dead Sea and farther north have been made to revise it. One of these revisions is in the Jordan Valley. Before this, however, we provide that used by Walker and LaBianca (2003: 448) at Tall a brief background sketch to situate our arguments. Hisban, which retains the original dates but subdivides the periods to allow for greater specificity. Whitcomb historical and (1997: 106; 2001: 505; 2009: 127) has also offered archaeological background several revisions which differ significantly from his original suggestion, but these have not seen wide­ Periodization and Islamic Archaeology spread use by other archaeologists. It is worth noting that, while many archaeologists In an oft-cited paper published nearly 20 years ago, concur that dynastic chronologies are problematic, not Donald Whitcomb (1992a) problematized the use of all agree that an archaeological periodization is the an­ political terminology to define Islamic archaeological swer. McQuitty (2007: 159), for example, prefers to periods. The argument, briefly summarized, is as fol­ use calendar dates rather than historical or archaeo­ lows. The use of political periodization in archaeology logical periods.2 Despite the common use of calendar implies, falsely, that cultural change inevitably follows dates among archaeologists seeking to avoid political political change, and that the two occur at essentially terminology, Hirschler (2006: ix) warns that “the em­ ployment of ‘neutral’ centuries might lead to a peri­ 1 The site is known by several names in other publications. odization devoid of any analytical value.” Likewise, Glueck (1935; 1940) favors various spellings of Khirbat Nuqayb Walmsley (2001: 634) has argued that for projects in al-Asaymir, “the Ruin of the Small Black Pass,” and Musil (1907) refers to the pass itself using the same diminutive form. The Eng­ their early stages, calendar dates can force an absolute lish-language Jordanian geology map of the area refers to the site as chronology onto the data, where a relative one might Khirbat Naqb Ismayr, “the Ruin of the Black Pass,” instead (Rabbaʿ be more appropriate. At the same time, he also notes 1991). This is complicated, however, by the report on the ʿAraba por­ that Whitcomb’s (1992a) periodization has not been tion of King et al.’s (1989) survey of Byzantine and Islamic sites in universally accepted, in part because the archaeologi­ Jordan. The surveyors were unable to find Glueck’s KNA but were instead told by locals “that the name Nuqayb el-Asaymir is used to cal periodization is arbitrary and somewhat problem­ describe all of the area north-northwest of [Khirbat] Finan” (King atic itself (Walmsley 2001: 634, n. 3). Walker (1999: et al. 1989: 202). In the reports of the Deutsches Bergbau-Museum (DBM) and the Center for British Research in the Levant (CBRL), the site is instead called el-Furn (Arabic for “the Furnace”). We 2 Although in previous publications on Khirbat Faris (notably, follow the diminutive form of the name used in Glueck’s (1935) McQuitty 2005) she has also used Whitcomb’s (1992a) archaeolog­ initial report of the site. ical periodization. 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 69

Whitcomb TraditionalWhitcombHisban Dynastic Revised 500 Byzantine Byzantine Byzantine Byzantine 600 630 7th century 700 Early Early Umayyad Islamic I Islamic I 750 800 EI I Early 900 Early Early EI II ‘Abbasid Islamic Islamic II Islamic II 969 1000 EI III

11th century Fatimid 1100 Middle Middle 1115 Islamic I Islamic I Crusader 1174 1200 MI I MI IIa Ayyubid Middle 1263 1300 Middle MI IIb Late Islamic II Bahri Islamic Islamic II 1400 MI IIc Mamluk 1382 Middle Burji LI Ia Islamic III 1500 Late Mamluk 1516 Islamic I 1600 LI Ib

1700 Late Late Ottoman (Ottoman) Islamic II Islamic IIa 1800

1900 Late Modern 1918 Islamic IIb Modern Mandate

Fig. 1. Visual comparison of chronologies for the Islamic period in Jordan. In this paper, we use the modified version of Whitcomb’s chronology used at Tall Hisban (Walker and LaBianca 2003).

207) also recognizes the advantages of an archaeolog­ (2003) at Tall Hisban. While we have adopted this ical chronology, but due to the poor stratigraphy of archaeological periodization for our project, in this many Islamic-period sites and the difficulty of dating paper we prefer to use it only broadly and for the ar­ later Islamic ceramic types, prefers dynastic terms. chaeological data to which it is best suited, favoring On this issue, we agree with Hirschler (2006: ix) calendar dates (in years a.d.) when possible. Finally, that “[t]he only convincing solution for the moment when discussing historical events, rather than archae­ seems to be a combination of the different possibili­ ological data, we often use dynastic terms (fig. 1).3 ties.” With this in mind, we attempt here to compro­ mise between all three options we have presented. 3 This is not to suggest that dynastic terms are always appro­ For our fieldwork in Faynan, ELRAP has adopted the priate for historical data. As we noted previously, many historians chronological scheme used by Walker and LaBianca have also adopted non-dynastic chronologies. One in particular is 70 JONES, LEVY, AND NAJJAR BASOR 368

Middle Islamic Central and Southern Jordan (Smith, Stevens, and Niemi 1997). As we will discuss in more detail later in this paper, survey in Faynan has The 12th century a.d. saw the beginning of an ex­ found that this general pattern of increased settlement pansion in the strategic and economic importance of in the Middle Islamic period seems to hold true there, central and southern Jordan, which would last until the as well. 15th century, when environmental and political factors led to what is generally seen as a period of “decline” The Middle Islamic Copper Smelting Revival in the lowlands of Jordan (Walker 2007: 181). This in Faynan rise in importance was primarily due to two historical factors. First, increasingly tenuous relations between The exact chronology of copper production at the Islamic and Crusader polities made the inland Cru­ Byzantine Phaino (Khirbat Faynan) is still uncertain. sader strongholds—examples in central and southern Weisgerber (2006: 25) argued that copper production Jordan include Karak, Shawbak, Wuʿayra, and Habis is not attested at all in the early Byzantine period and (see fig. 2)—strategically valuable (Humphreys 1977: that the town of Phaino ceased to be important on any 63). Second, the Ayyubid reorganization of the iqtaʿ level very early in the sixth century a.d. Kind et al. system—the granting to elites of tax revenues from (2005: 192) suggest that copper production at Faynan agricultural land—and the expansion of sugar produc­ ceased, or at least scaled down significantly, between tion into the southern Levant during this period greatly 360 and 370 a.d. Hauptmann (2007: 156) proposes affected the agricultural economy of Transjordan, as it a slightly later date, suggesting that copper produc­ did in other parts of Syria and Egypt (Galloway 1977: tion at Faynan had stopped by the late fifth century. 188–89; Tsugitaka 1997: 42–46). Friedman (2008: 54–56), on the other hand, argues Survey work in southern Jordan has revealed con­ that copper production continued into the sixth century siderable Middle to Late Islamic occupation, although and that the town probably did not decline in impor­ the temporal resolution of these observations is often tance until the late sixth, or possibly even early sev­ not ideal for discussing intra-period variation. This is enth, century. Likewise, Mattingly et al. (2007a: 333, partly the result of difficulties in dating the ubiquitous 348) maintain that copper production continued into handmade ceramics of the Middle and Late Islamic the late Byzantine period, although they admit that it periods, which are discussed later in this paper. None­ is currently impossible to date the end of this phase of theless, some trends are evident in the available sur­ copper production at Faynan with any precision. vey data. In the northern ʿAraba and southern aghwar Regardless of when exactly copper production (plural of ghawr, Arabic for “lowland” or “valley”), ceased in the Byzantine period, it is clear that a hiatus settlement in villages increased during the Middle in copper production, and a decline in occupation in Islamic period, compared with the preceding Early general, followed. While some Early Islamic material Islamic (MacDonald 1988; 1992; MacDonald et al. has been found in the area surrounding Khirbat Faynan 2004). To the south, in the region west of Petra, survey (King et al. 1989: 203) and to the west at ­Khirbat has found similar patterns. Little settlement is evident Hamra Ifdan, the evidence for Early Islamic occupation in the Early Islamic period, but in the Middle Islamic, in Faynan is sparse, and no copper production seems villages were built in defensible locations and, per­ to have taken place during this period. This pattern haps, mostly abandoned again before the beginning of is ­different from that seen in the area around Timna, the Late Islamic period (Hübner 2004; Lindner 1999; ­located ca. 105 km south of Faynan on the Israeli side Lindner et al. 1998). The plateau to the south of Petra of the ʿAraba. There, Early Islamic copper mining is also shows similar patterns, with little evidence for known in Wadi Amram (Willies 1991), and smelting Byzantine and Early Islamic settlement, followed by of the same period occurred at camps near Beʾer Ora an increase in settlement starting in the Middle Islamic (Rothenberg 1988), in a system that Whitcomb (2006) (Hart 1986; Hart and Falkner 1985). In contrast, sur­ argues functioned mostly to provide copper to the bur­ vey in the southern ʿAraba has found very little evi­ geoning city of Ayla (ʿAqaba). There is evidence, in dence for occupation during the entire Islamic period fact, that copper ore was processed in Ayla itself prior except for the region immediately surrounding ʿAqaba to the late eighth century a.d., when the smelting camps ­strikingly similar to the archaeological periodization schemes dis­ were established (Damgaard 2009: 89). cussed here, and breaks the Islamic period into Formative, Middle, In Faynan, on the other hand, post-Byzantine copper Late, and Modern periods (Hirschler 2012). production began again only in the Middle Islamic II 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 71

Fig. 2. Map of sites mentioned in the text. Data from the Digital Archaeology Atlas of the Holy Land (DAAHL: http://daahl. ucsd.edu). 72 JONES, LEVY, AND NAJJAR BASOR 368

(1200–1400 a.d.) and was probably limited to this (Levy et al. 2003: 259–60). Surveyors spaced no more period. This does leave open the question of why the than 50 m apart (generally ca. 25 m) walked tran­ Crusaders, who controlled the highland plateau of sects and recorded all sites they found—defined very southern Jordan in the mid-12th century, did not exploit broadly and including features as small as scatters of the copper resources of Faynan. Unfortunately, this potsherds or copper slag. While necessarily limiting question has no certain answer at present. It is possible the area that could feasibly be surveyed, this also al­ that excavations of medieval smelting sites in Faynan lowed the team to compile a very complete catalog of and further refinement of southern Jordanian ceramic the sites within the survey boundaries. Artifacts were chronologies will reveal a limited Crusader presence collected and recorded separately from individual fea­ in Faynan, but it is equally likely that the Crusaders tures within sites by surveyors walking transects ap­ simply relied on a different source of copper. proximately 10 m apart. At larger sites, artifacts were Evidence for Middle Islamic smelting comes from also collected separately from the interior and exterior Khirbat Nuqayb al-Asaymir and three small slag of architectural features. This degree of spatial control mounds near Khirbat Faynan, referred to as Faynan was deemed necessary to determine whether there was 2, Faynan 6, and Faynan 7 by the Deutsches Bergbau- any chronological variation in site size and usage. Museum (DBM) team (Hauptmann 2007: 97, 103; Newson et al. 2007: 364). The large, Roman slag Khirbat Nuqayb al-Asaymir mound known as Faynan 1 may also have some con­ nection to Middle Islamic copper production, though Khirbat Nuqayb al-Asaymir is situated between this is less clear (Hauptmann 2007: 127). It is likely Wadi Ghuwayb al-ʿAtshana and Wadi Nuqayb al- that our current understanding of the Middle Islamic Asaymir, ca. 1 km south of the main channel of Wadi period at Khirbat Faynan is still incomplete, due al-Ghuwayb and 1 km southeast of the large Iron Age mostly to the complicated settlement history of that smelting site of Khirbat al-Nahas (fig. 3). The largest site. On the other hand, Khirbat Nuqayb al-Asaymir, building (No. 5300)4 is located at the western end of as a single-period site, presents an excellent opportu­ the small valley that runs through the site, and from nity to investigate Middle Islamic copper production. this building it is possible to look over Wadi Nuqayb al-Asaymir and see parts of Khirbat al-Nahas and Wadi al-Ghuwayb. The site is located on—and named khirbat nuqayb al-asaymir after—an ancient path linking it, and the other sites and related sites in Wadi al-Ghuwayb, with Khirbat Faynan. It is also worth noting that Wadi al-Ghuwayb is one of the easier Survey Goals and Methods routes from Faynan to the plateau to the east (Haupt­ The primary goal of the 2002 JHF survey was to mann 2007: 132), and this was likely an important fac­ explore the Iron Age (ca. 1200–500 b.c.) metallurgi­ tor in the location of KNA. cal landscape of eastern Wadi al-Ghuwayb and Wadi The first western explorer to visit the area around al-Jariya, reflecting the fact that the survey took place KNA was the Czech Orientalist Alois Musil (1868– the same year that excavations at the large Iron Age 1944), although it is not clear from his descriptions smelting site of Khirbat al-Nahas began (Levy et al. if he visited KNA itself or only the surrounding areas 2003: 249). The fortuitous decision to include Khirbat (Musil 1907: 298–99). Later, Nelson Glueck (1900– Nuqayb al-Asaymir and the surrounding area in the 1971) visited KNA as part of his 1934 survey of south­ survey boundaries was, in fact, primarily motivated ern Transjordan. Glueck (1935: 30–32; 1940: 65–66) by the assumption that its ore sources would also have dated the site to the Medieval Islamic period, based 5 been exploited during the Iron Age (Levy et al. 2003: mainly on architecture and pottery. In the early 1980s, 260). The 2007 ELRAP survey was conducted as an extension of the 2002 survey in the northern parts of 4 During the 2002 survey, features were designated by the site num­ Wadi al-Jariya and shared the 2002 survey’s goal of ber followed by a two-digit feature number. Thus, feature 5300 is the investigating the Iron Age metallurgical landscape of first feature identified at WAG 53, the survey number given to KNA. 5 Wadi al-Jariya (Knabb et al. forthcoming). Although Glueck does not specify a period beyond this, his comparison of the architecture of the site to the “desert castle” Both the 2002 and 2007 surveys were conducted of Qasr Kharana (Glueck 1940: 65–66) led at least one reviewer as intensive, full-coverage surveys of the areas within ­(Glidden 1941) to erroneously conclude that KNA dated to the 250 m of either side of the wadis being investigated Umayyad period. 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 73

Fig. 3. Location of Khirbat Nuqayb al-Asaymir in relation to other sites in Faynan, including Khirbat Faynan, the Wadi Sal- mina mines, and several key Iron Age smelting sites previously excavated by ELRAP. Background image: (C) CNES/SPOT Image 1992–1994. 74 JONES, LEVY, AND NAJJAR BASOR 368 the site was surveyed by a team from the Deutsches Little evidence of metallurgical activities was ­Bergbau-Museum, who dated the site to the 13th cen­ found near the other buildings at the site, and it ap­ tury and as “Mamlukisch-türkisch,” based on pottery pears that these instead served domestic or administra­ (Hauptmann, Weisgerber, and Knauf 1985: 190, 192). tive purposes. The exact functions remain somewhat This date has since been revised to Ayyubid–Mamluk unclear, although Buildings 5313 and 5314 seem to (e.g., Hauptmann 2007: 126–27), based on six recently have been watchtowers, based on their positions at the published coins from the site (Kind et al. 2005). The extreme north and south of the site and the views of the DBM team also mentions a small mosque (Hauptmann approaches to the site afforded from them (fig. 7). It is 2007: 126; Weisgerber 2006: 27) which was not re­ possible, especially if slave labor was used at KNA, corded by the Jabal Hamrat Fidan Project team. Finally, that these may have functioned as much to keep the the current project began when the site was visited in workers in as to keep intruders out. 2002 by the JHF survey team, who used a total station6 to map the site’s features in detail (fig. 4). The Ore Sources During the 2002 survey, 15 primary buildings were identified at the site, in addition to numerous smaller Mines are often rather difficult to date, but we features. Copper production activities are associated suggest that two nearby sites represent the main ore primarily with two buildings, designated 5300 (fig. 5) sources for KNA: WAG 57 (fig. 8) and 58. WAG 57 and 5304 (fig. 6). Building 5300, at the western end consists of 11 mine shafts and three mine tailings— of the site, was the center of copper smelting at KNA, the piles of debris left behind by the mining opera­ judging from the large slag piles just outside of the tion—located ca. 200 m northwest of Building 5300 building, its proximity to the mines in Wadi Nuqayb at KNA (fig. 9), between KNA and Khirbat al-Nahas in al-Asaymir (WAG 57 and WAG 58, discussed below), the Wadi Nuqayb al-Asaymir. The site was previously and the results of recent excavations.7 Hauptmann published as dating to the Islamic period, given its (2007: 126–27) has suggested that this building may proximity to KNA, but with Roman/Byzantine-period have functioned as a “shaft-house,” similar to the ceramics (Levy et al. 2003: 252). Only one sherd, an Medieval European structures described by Agricola undecorated lamp fragment, was found at this site, and (1950: 102). Such buildings contained mine shafts and on reexamination it became clear that it could not be a windlass and were the main entrance to the under­ securely dated, although we noted that the fabric is ground mine system. The 2011 ELRAP excavation similar to that of lamps found at KNA. Given the loca­ season, however, produced definitive evidence in the tion of these mines, exploitation during the Roman and form of a relatively complete shaft furnace that smelt­ Byzantine periods is unlikely. The only metallurgical ing was the primary activity that took place in this site in the Wadi al-Ghuwayb system with remains of building. It is still possible that the northernmost room these periods is Khirbat al-Ghuwayba (ca. 3 km north­ was devoted to mining, but this is unlikely. The place­ east of KNA), and copper production there seems to ment of the building was likely dictated by a desire have been limited to the early Iron Age (Ben-Yosef to stoke the furnace using the wind blowing through 2010: 447–48). The only other possibility would be the site from the west and, secondarily, the proximity Khirbat Faynan, nearly 7 km southeast of KNA, where of the location to the ore sources in Wadi Nuqayb al- closer and larger ore sources are known (Hauptmann Asaymir, and was unrelated to the construction of a 2007: 155–56). The other mine site in Wadi Nuqayb shaft-house. al-Asaymir is WAG 58, a cluster of three mines located 120 m northeast of WAG 57. Although the ceramics from this site all date to the Iron Age, it is probable 6 A total station is a piece of surveying equipment that combines that it also served as an ore source for KNA, given its a theodolite (which measures angles) and an electronic distance me­ proximity to both the mines at WAG 57 and to KNA. ter (EDM) to determine the coordinates of a point. For more detailed discussions of this instrument and the “early” digital archaeology There are also several less likely candidates for ore system used during the 2002 survey, see Levy et al. (2001) and Levy sources, identified during the 2007 ELRAP survey in and Smith (2007). the northern part of Wadi al-Jariya: WAJ 609, WAJ 7 A preliminary report of the 2011 excavation season has been 613, and WAJ 626 (fig. 9; see Knabb et al. forthcom­ accepted for publication in a forthcoming volume of the Annual of ing). Sherds of Middle to Late Islamic undecorated the Department of Antiquities of Jordan. This excavation occurred after the initial submission of the present paper for review, and as handmade coarse wares were found associated with such the data are still being analyzed but will be discussed here mine tailings at these sites. It is difficult to date these when relevant. wares very precisely, especially in southern Jordan, 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 75

Fig. 4. Map of Khirbat Nuqayb al-Asaymir based on 2002 Jabal Hamrat Fidan Project survey data. Background image:­ IKONOS satellite imagery courtesy of GeoEye. GeoEye Data is owned by GeoEye, Inc. All rights are reserved by GeoEye, Inc. but it is likely that they date to the Late Islamic pe­ rocky inclusions throughout. Likewise, sherds of a riod. They are considerably coarser than the handmade Late Islamic II (17th–early 20th centuries a.d.; in this wares found at KNA, include sherds from cooking case, likely 19th century) chibouk (tobacco pipe) were pots with large elephant-ear handles, and have large, found at WAJ 601, another site consisting of little 76 JONES, LEVY, AND NAJJAR BASOR 368

Fig. 5. Building 5300 at KNA, the center of copper production at the site. Visible both adjacent to and behind the structure are large piles of black copper slag. Photo: UCSD Levantine Archaeology Laboratory.

Fig. 6. The limited architecture of Building 5304, a secondary metal production building across the site from Building 5300. Photo: UCSD Levantine Archaeology Laboratory. 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 77

Fig. 7. The southern “watchtower” at KNA, Building 5313. The view afforded to the south over Nuqayb al-Asaymir and Wadi Ghuwayb al-ʿAtshana is evident in this photo. Photo: UCSD Levantine Archaeol- ogy Laboratory.

Fig. 8. A filled mine shaft at WAG 57, the primary source of copper ore for KNA. Photo: UCSD Levan- tine Archaeology Laboratory. 78 JONES, LEVY, AND NAJJAR BASOR 368

Fig. 9. Middle and Late Islamic sites recorded during the 2002 and 2007 Wadi al-Ghuwayb and Wadi al-Jariya surveys. Background image: IKONOS satellite image courtesy of GeoEye.

­beyond mine tailings. While it is not possible to en­ related to pastoral nomadic activity at these sites, tirely rule out Middle Islamic re-exploitation of these whether Middle or Late Islamic. This interpretation is primarily Iron Age mines, the distance of these mines also supported by the fact that a campsite was found from KNA makes it more likely that these sherds are among the tailings at WAJ 626 (fig. 10), and most of 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 79

Fig. 10. A Late Islamic “storage” feature, part of the campsite built between the mine tailings at WAJ 626. Photo: Kyle A. Knabb.

the Late Islamic pottery at that site is associated with Approximately 1 km north of KNA, on a talus these campsite features. slope east of Wadi Ghuwayb al-ʿAtshana, is WAG 43 (fig. 11), a small (ca. 30 m2) campsite. The site consists Other Middle to Late Islamic Sites of two oval-shaped structures and two heavily eroded stone features. Two sherds were collected at the site: a WAG 56, a cemetery, sits atop a small hill at the base and a body sherd, both of which come from one southeastern margins of KNA, and the two sites over­ painted, handmade vessel (fig. 12:1). The core is gray, lap to some extent. During the 2002 survey, eight with large, visible inclusions and some chaff temper, graves and two stone circles were identified at the site, and it has a thick, pinkish-red slip painted in broad, though at that time most of the graves had already been linear decorations. The vessel is similar to earlier Late robbed. The cemetery is in close proximity to metal­ Islamic painted wares from central and northern Jor­ lurgical remains from KNA, and it is not clear whether dan described by Walker (2009: 44), and unlike the the cemetery is associated with KNA or postdates the painted wares found at KNA (described below). activities there. The ceramics from this site included Several other Middle to Late Islamic campsites 176 sherds of Middle Islamic handmade coarse wares, were found during the 2002 survey: WAG 42 and 45, at least eight of which were painted, but the question and WAJ 525, 528, and 562. More specific dating of of whether these are associated with the metallurgical these sites is difficult, as the pottery collected from activities represented by the slag mound, or the ceme­ most of them is limited to body sherds of undecorated, tery, or both, is difficult to resolve. handmade pottery. The exceptions are WAG 42 and 80 JONES, LEVY, AND NAJJAR BASOR 368

Fig. 11. Middle to Late Islamic campsite at WAG 43, 1 km north of KNA. Photo: UCSD Levantine Ar- chaeology Laboratory.

WAJ 525, at which rim sherds of Late Islamic–period Gaza Ware were collected (fig. 12:2–3), making it likely that these sites date to the 19th century a.d. or later (but not certain; see Rosen and Goodfriend 1993 for Late Islamic IIa—18th or even late 17th century— dates for Gaza Ware). While there are certainly some unresolved chronological issues, the majority of the easily datable material from the campsites is several centuries later than the dates we propose for KNA, at the earliest. It also seems likely, however, that pastoral nomads made use of the area while KNA was operating. The relationship between pastoral nomads and the miners and smelters is not entirely clear, but nomadic activity in the surrounding wadis may have been part of the rationale behind the watchtowers at KNA (Buildings 5313 and 5314). This is not to suggest the validity of what Johns (1994: 22) calls “the myth of the nomad”; the reality was certainly more complex than Bedouin Fig. 12. Survey pottery from WAG 43 (1), WAG 42 (2), and predation and violent conflict. Nonetheless, it is pos­ WAJ 525 (3). See table 1 for descriptions. Illustration: Caro- sible to recognize that relations between pastoralist line Hebron. and miner were probably mostly mutually beneficial, but also that prevailing attitudes toward pastoralists might have made the overseers of KNA somewhat tural terrace. The dating of the site is far from certain, wary of their presence in Faynan. as the only find was a single body sherd of Middle The final site to discuss is WAJ 576 (fig. 13), ca. 5.5 to Late Islamic handmade pottery. Regardless of the km north of KNA, which was identified as an agricul­ exact date, the urge to associate this terrace with KNA 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 81

Table 1. Description of Late Islamic Sherds Illustrated in Figure 12 No. Site Reg. No. Type Core Int. Ext. Decoration 1 WAG 43 WAG02.222 HMGPW 10YR 8/4 Very pale 10YR 8/2 Very 5YR 7/6 Reddish Paint: Purple-brown brown pale brown yellow 2 WAJ 525 WAJ02.114 Gaza Ware GLEY 2 5/5PB GLEY 2 5/10G GLEY 1 7/5GY Bluish gray Greenish gray Light greenish gray 3 WAG 42 WAG02.135 Gaza Ware 5Y 4/1 Dark gray GLEY 1 4/N GLEY 1 4/N Dark gray Dark gray

Fig. 13. Possible Middle to Late Islamic agricultural terrace at WAJ 576. Photo: Kyle A. Knabb.

should be resisted. The miners, smelters, and adminis­ of the site—as far from KNA as the field systems as­ trators at KNA would likely have obtained their food sociated with Khirbat Faynan—makes it more likely from Shawbak or Karak on the nearby highland pla­ to be associated with pastoral activity in the northern teau, rather than growing it in Faynan. Also, as Alison part of Wadi al-Jariya, although it is presently impos­ McQuitty has stated, “agriculture was and is practised sible to say whether this was during the Middle or Late from a tent” (McQuitty et al. 1997: 183). The location Islamic period. 82 JONES, LEVY, AND NAJJAR BASOR 368

Fig. 14. Frequencies of ware types found at KNA, based on raw sherd counts, N = 1,300.

dating khirbat nuqayb al-asaymir

Pottery

While the large ceramic assemblage discussed here is being published for the first time, small numbers of sherds from KNA have been published twice previ­ Fig. 15. Undecorated (1, 4–5) and decorated (2–3) hand- ously. Glueck (1940: 67, fig. 29) published a black- made ceramics from KNA. See table 2 for descriptions. and-white photo of what he identified as “Mediaeval Arabic” sherds. Later, Hauptmann, Weisgerber, and Knauf (1985: 190–92, fig. 31) published six sherds— situation in southern Jordan has improved less. Few two of undecorated handmade wares, three glazed excavated assemblages have been published, and es­ pieces, and one red corrugated sherd—which they sentially none have been published in final form. This dated to the 13th century, with one exception, which is certainly a disadvantage in trying to analyze a sur­ was more generally dated Mamluk–Turkish. These in­ face-collected assemblage, and our pottery analysis vestigations were followed by our 2002 survey, which should be read with that in mind. retrieved the surface ceramic assemblage described here. Handmade Wares. The KNA survey assemblage After mapping during the 2002 survey, the eight is dominated by handmade wares (these account for members of the survey team spread out in approxi­ 70% of the total raw sherd count; N = 907; fig. 14), mately 10 m intervals and collected pottery and por­ which are unfortunately among the most difficult table artifacts from each building individually. The wares to date. Avissar and Stern (2005: 88) have ar­ locations of sherds collected away from buildings gued, on the basis of excavations in Israel, that undec­ were mapped, as well (fig. 4). A total of 1,300 sherds orated handmade bowls “cannot be dated according to mostly dating to the Middle Islamic period were col­ form, only by stratigraphic context.” The same argu­ lected from the surface of the site (fig. 14). Although ment could be generally applied to most, if not all, the dates presented by Hauptmann, Weisgerber, and undecorated handmade wares (fig. 15.1, 4–5), which Knauf (1985) are sound, the following discussion account for most of the KNA handmade assemblage seeks to offer a more precise date and to present the (93% of the handmade sherds, and 65% of the overall results of preliminary analysis of this much larger raw sherd count; N = 843; fig. 14).8 Undecorated hand­ sample of pottery from KNA. It should be stated at the outset, however, that al­ 8 This number is potentially misleading. First, as noted, these though many advances have been made in the study percentages are based on raw sherd counts and not on estimates of of Middle Islamic ceramics in the last 25 years, the total number of vessels. Second, although the sherds were carefully 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 83

Table 2. Description of Handmade Sherds illustrated in Figure 15 No. Site Reg. No. Building Type Vessel Core Int. Ext. Decoration 1 KNA WAG02.232 5307 HMCW Jar GLEY 1 2.5/N Black 2.5YR 7/8 2.5YR 7/8 Light red Light red 2 KNA WAG02.233 5310 HMGPW Jug GLEY 1 2.5/N Black 2.5YR 6/6 7.5 YR 7/3 Paint: Black Light red Pink 3 KNA WAG02.235 5310 HMGPW Bowl GLEY 1 2.5/N Black 2.5YR 7/6 10YR 7/3 Paint: Black Light red Very pale brown 4 KNA WAG02.236 5312 HMCW Bowl GLEY 1 2.5/N Black 7.5YR 7/6 7.5YR 7/6 Reddish yellow Reddish yellow 5 KNA WAG02.237 5311 HMCW Basin GLEY 1 2.5/N Black 5YR 5/6 2.5YR 6/8 Yellowish red Light red made wares are a reliable indicator of Middle to Late Early Mamluk periods, though Brown (1980: 125) Islamic occupation at a site, but dates more specific points out that this dating is speculative. At KNA, all than this based only on these wares should probably three types of decoration are present, even while the be regarded with caution. “classic” monochrome geometric designs are more The remaining 7% of the handmade assemblage is common (fig. 16). Dating the site more precisely than made up of the decorated coarse wares commonly re­ “Middle Islamic” based on this portion of the assem­ ferred to as “handmade geometrically painted wares” blage is not currently possible, although Micaela Sini­ (HMGPW). The most common identifiable forms at baldi’s (2009) work in the Petra area may soon help KNA are jugs and bowls (fig. 15:2–3). As with un­ to establish a chronology for these wares in southern decorated handmade wares, dating HMGPW is some­ Jordan. what problematic. Brown suggested, based primarily on assemblages at Wuʿayra (Brown 1987: 284) and Middle Islamic Unglazed Wheel-Made Wares. Shawbak (Brown 1988: 232), that an earlier set of The next most common group at KNA, unglazed “linear” designs in red or brown was gradually re­ wheel-made wares, make up roughly 22% of the sur­ placed by the later geometric designs that characterize vey ceramic assemblage (N = 291). The majority of the HMGPW. Johns (1998: 66, 89, fig. 4) has also taken unglazed wheel-made diagnostic sherds from KNA are up this suggestion, noting that simple linear designs from wheel-made cream wares (fig. 17:1–6), primar­ in red seem to be most common in 11th- and 12th- ily jugs made of hard but porous fabrics which would century assemblages, with HMGPW emerging in the keep water cool through evaporation (cf. Milwright second half of the 12th century. Unfortunately, further 2008: 166). Similar vessels are found at Karak (Mil­ work at Wuʿayra has not supported this hypothesis, as wright 2008: 360–61), which was possibly a center for the decoration on this ware in all phases is relatively their production (Milwright 2008: 166), and at Khirbat­ heterogeneous (Vannini and Tonghini 1997: 380). It is al-Shaykh ʿIsa (MacDonald 1992: 232, 235, pl. 31:p). not entirely certain, then, that linear red designs pre­ Milwright (2008: 167–68) cautions that small varia­ date “classic” HMGPW. Additionally, Sauer (1973: tions in profile might not be useful for dating these 62) suggested, based on the pottery at Hisban, that bi­ vessels, but suggests a general date between the 13th chrome HMGPW is characteristic of the Ayyubid and and 15th centuries a.d. It is worth noting, however, that the various characteristically Mamluk deco­ rated versions of these wares (cf. Avissar and Stern examined, it is quite likely that some of the sherds counted as undec­ orated handmade wares may be very worn examples of handmade 2005: 111–12; Walmsley and Grey 2001: 153) are not geometrically painted wares. Third, unlike at Wuʿayra (Vannini and found within the KNA survey assemblage, although Tonghini 1997), undecorated handmade wares were not further di­ two sherds with “incised and punctate” decoration vided into slipped and unslipped groups; it has been suggested that were found near Khirbat Faynan (Tomber 2007: fig. slipped, unpainted wares may be examples of undecorated portions A2.11.106). In addition to the cream wares, the KNA of painted vessels (Vannini and Tonghini 1997: 379), although re­ cent work suggests that—during the 12th century, at least—slipped, assemblage also includes sherds of a more reddish- undecorated vessels may make up a distinct category of handmade orange fabric (fig. 17:7–8). These most likely also wares (Sinibaldi 2009: 456–57). date to the Middle Islamic period. Incised decorations 84 JONES, LEVY, AND NAJJAR BASOR 368

Table 3. Description of HMGPW Sherds in Figure 16 No. Site Reg. No. Building Type Vessel Core Int. Ext. Decoration 1 KNA WAG02.240 5306 HMGPW Body GLEY 1 2.5/N GLEY 1 2.5/N 5YR 7/6 Paint: Red and Black Black Reddish yellow black 2 KNA WAG02.241 5312 HMGPW Body 10YR 7/4 Very 10YR 8/4 Very 10YR 8/4 Very Paint: Red pale brown pale brown pale brown 3 KNA WAG02.242 5312 HMGPW Handle 10YR 8/4 Very 10YR 8/3 Very 10YR 8/3 Very Paint: Purple-red pale brown pale brown pale brown 4 KNA WAG02.243 5311 HMGPW Body GLEY 1 2.5/N GLEY 1 2.5/N 10YR 8/4 Very Paint: Brown (Jug?) Black Black pale brown 5 KNA WAG02.244 5311 HMGPW Body GLEY 1 2.5/N GLEY 1 2.5/N 7.5YR 7/6 Paint: Black (Jug?) Black Black Reddish yellow 6 KNA WAG02.245 5311 HMGPW Body GLEY 1 2.5/N GLEY 1 2.5/N 10YR 8/4 Very Paint: Brown Black Black pale brown

Fig. 16. Handmade Geometrically Painted Wares (HMGPW) from KNA. See table 3 for descriptions.

­appear throughout the Byzantine and Islamic periods, but the decoration of Sherd 228 (fig. 17:8) has parallels in Middle Islamic assemblages—for example, at Tell Abu Sarbut (LaGro 2002: 137–38, ill. 8:25–26).

Mold-Made Oil Lamps. The wares most useful for dating were unfortunately found in much smaller numbers at KNA. The most common of these more diagnostic groups is made up of mold-made slipper lamps, of which only 28 sherds were found during the 2002 survey at KNA. Of these, 13 sherds, from a total Fig. 17. Unglazed wheel-made wares (1–8), and Syrian under- glaze painted stonepaste ware (9). See table 4 for descriptions. of five lamps, were from decorated, and therefore di­ agnostic, portions of slipper lamps. The decorations on these consist mostly of partial naskhi inscriptions and main differences between the earlier (Middle Islamic pseudo-calligraphic designs (fig. 18). Slipper lamps IIa [1200–1250 a.d.]) and later (Middle Islamic IIb–c such as these are common throughout the 13th and [1250–1400 a.d.]) lamps are the shapes of the handles, 14th centuries (Avissar and Stern 2005: 128), and the none of which were found at KNA, and the types of 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 85

Table 4. Description of Wheel-Made Sherds Illustrated in Figure 17 No. Site Reg. No. Building Type Vessel Core Int. Ext. Decoration 1 KNA WAG02.223 5310 Creamware Jug 10YR 6/2 Light 5YR 7/6 10YR 8/4 Very brownish gray Reddish yellow pale brown 2 KNA WAG02.229 5312 Creamware Jug 10YR 6/1 Gray 5YR 7/6 10YR 7/4 Very Reddish yellow pale brown 3 KNA WAG02.225 5311 Creamware Jug 10YR 8/4 Very 5YR 7/6 5YR 8/4 Pink pale brown Reddish yellow 4 KNA WAG02.226 5311 Creamware Jug 10YR 7/1 10YR 7/4 Very 10YR Light gray pale brown 6/4 Light yellowish brown 5 KNA WAG02.224 5313 Creamware Jug 5YR 7/6 7.5YR 7/6 10YR 8/4 Very Reddish yellow Reddish yellow pale brown 6 KNA WAG02.231 5305 Creamware Base 10YR 7/2 10YR 8/3 Very 10YR 7/4 Very (Bowl) Light gray pale brown pale brown 7 KNA WAG02.227 5306 WM Jar 2.5YR 7/6 2.5YR 7/6 5YR 6/6 Notched rim Reddish Light red Light red Reddish orange yellow 8 KNA WAG02.228 5312 WM Bowl? 10YR 8/4 Very 5YR 7/6 5YR 7/6 Incised Reddish pale brown Reddish yellow Reddish orange yellow 9 KNA WAG02.238 5306 Syrian Bowl 7.5YR 8/4 N/A N/A Black paint underglaze Pink under manganese painted purple glaze stonepaste decoration; the early lamps are commonly decorated parent glaze, two are painted in black under a blue with naskhi and pseudo-calligraphic designs, and the glaze, and two are painted in black under a manga­ later lamps with linear and geometric ones (Hadad nese purple glaze (fig. 17:9; fig. 19). In addition to 1999: 217). The lamps from KNA with preserved these, Hauptmann, Weisgerber, and Knauf (1985: fig. decoration are all of the earlier type, which would 31:2–3) found two sherds of this type during their place them in the first half of the 13th century. Simi­ survey of KNA. Of the 18 sherds of stonepaste wares lar examples have been found at Bet Shean (Hadad found during the JHF survey, only two are from the 1999: 217) and Jerusalem (Tushingham 1985: 147), rims of vessels, and both of these sherds are from the and Avissar and Stern (2005: 128) note occurrences same distinctive type of vessel: bowls with incurv­ at several other sites, including unpublished examples ing rims, featuring the simple black-striped motif that from Jerusalem and Acre. Tonghini (1994: 275) notes as a consistent decoration for this shape. Based on a comparison with a com­ Glazed Wares. A total of 20 sherds of glazed plete waster supposedly excavated near Maʿarrat al- pottery were found at KNA in 2002, 18 of which are Nuʿman10 (Porter 1981: 39–40, pl. 27), Porter (1981: from underglaze-painted Syrian stonepaste wares.9 40) suggests that this shape is a “provincial variant” of Of these, ten are painted in black under a turquoise so-called Raqqa wares, although the decoration shows glaze, four are painted in black and blue under a trans­ a connection to Raqqa. Raqqa-type stonepaste wares

9 We use this term to refer to the group of ceramics sometimes 10 As Porter and Watson (1987: 192) note, the Bartels Gift, called “Raqqa wares.” We prefer this more unwieldy name, as it which includes this waster, is somewhat problematic. It was donated points to the fact that these wares were certainly produced in places to the Ashmolean Museum after being purchased from a Syrian other than Raqqa—Mason (2004: 106) argues, on the basis of pe­ dealer who claimed it came from an illicit excavation near Maʿarrat trographic analyses, that “Damascus would appear to be the most al-Nuʿman, but some skepticism is necessary. The stories told by significant Syrian production center”—and in fact continued to be dealers do not always reflect reality; for example, Mason (2004: produced after Raqqa fell to the Mongols in 1259. We note, how­ 106–7) has demonstrated through petrographic analysis that a group ever, that our term must also be considered a shorthand, as it is of wasters allegedly from Raqqa in the Royal Ontario Museum in likely that underglaze-painted stonepaste wares of the same type fact belong to a separate group compared to wasters known to be were also produced in Egypt (Brown 2006: 381; Scanlon 1984: from Raqqa. Even if this dealer’s story is true, a single waster from 118–19; T­ onghini 1998: 50). an illicit excavation is a poor chronological anchor for this type. 86 JONES, LEVY, AND NAJJAR BASOR 368

Fig. 18. Early 13th-century a.d. mold-made slipper lamps from KNA, all bearing naskhi inscriptions or pseudo-inscriptions. can date anywhere between the late 12th and 13th cen­ center) in the late 12th century (Mason 2004: 98–99), turies (Porter 1981: 10), but Porter (1981: 41) argues, much earlier than Porter’s (1981: 40–41) date for the based on comparison with similar vessels from Persia, rim type. Although this is an uncommon form, exca­ that this particular rim shape dates to the second half vated comparanda are known in Syria at Qalʿat Jaʿbar of the 13th century. (Tonghini 1998: fig. 67), in the group Tonghini (1998: More recently, however, Jenkins-Madina (2006: 46–50) calls “fritware 2,” dated to the late 12th and 62, 92–93) has identified four wasters with this profile 13th centuries. Stonepaste wares with this profile are in the Museum of Turkish and Islamic Art in Istan­ also found in Phase H at Tell Abu Qaʿdan in the Jordan bul, one of which is known to have come from early Valley (Franken and Kalsbeek 1975: 132). Although 20th-century excavations of Raqqa,11 suggesting that Franken and Kalsbeek (1975) give only a site-specific stonepaste vessels with this form were produced there. chronology for Tell Abu Qaʿdan, Sauer (1976: 94) has If this is the case, the second half of the 13th century suggested that Phase H is Ayyubid. is an unlikely terminus post quem for the introduc­ One of the other two glazed sherds is of a gritty, tion of this shape, as the Mongols destroyed Raqqa light reddish-brown fabric with a thin, unevenly ap­ in 1259 a.d., effectively ending pottery production plied white slip and thin greenish-yellow glaze. It is there (Porter 1981: 10). Likewise, Mason (2004: 109) a small body sherd and, as such, is not particularly cautions against assuming that Iran always provided diagnostic, but some discussion is warranted here, the influence for Syrian wares, pointing out that, in especially since Hauptmann, Weisgerber, and Knauf some cases, the opposite was likely true. He places (1985: fig. 31:6) found a bowl base with a similar the “arc-back” motif evident on the large manganese- glaze at KNA, which they date to the 13th century. glazed sherd from KNA (figs. 17:9 and 19: bottom- Monochrome glazed bowls are common in the Levant from the 12th through the 15th centuries, with a peak 11 These four wasters, labeled in her publication as W48, W105, in the 13th and 14th centuries (Brown 1988: 237; Mil­ W106, and W107, are all housed at the Museum of Turkish and wright 2008: 188–89), and their use continues through Islamic Art in Istanbul (Jenkins-Madina 2006: 62, 92–93). W48 is known to have come from the 1906 excavations of Raqqa, while the at least the 18th century, possibly continuing into the other three are listed as more generally coming from “kaza de Rakka 20th century (Milwright 2000: 195–96). Sherds simi­ et de Maareh” (Jenkins-Madina 2006: 92–93). lar to our KNA example, made of a sandy orange fab­ 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 87

Fig. 19. Late 12th- to 13th-century Syrian underglaze-painted stonepaste wares (“Raqqa wares”) from KNA.

ric and glazed in green over a white slip, have been this is not necessarily the case.12 The most common found in a 16th-century context at Khirbat Faris (Mc­ of these is an often corrugated, brownish-red ware, Quitty and Falkner 1993: fig. 21:41–42). Despite some commonly identified as Byzantine. As mentioned similarities to the Khirbat Faris examples, the long life above, it is possible that these are unglazed sherds of this ware makes it difficult to assign the KNA sherd of brittle glazed cooking pots. Likewise, based on to any specific period, especially as it lacks any diag­ concentrations of the same ware in Middle Islamic nostic features. assemblages at Burj al-Ahmar (Pringle 1986), ʿAfula The remaining glazed sherd is from a brittle type (Dothan 1955), and Karak, Milwright (2008: 170) has of cooking pot with a reddish-brown, sandy fabric and argued that unglazed corrugated wares may continue a small bit of brown glaze. It is quite likely that other into the Middle Islamic period. Additionally, Schaefer sherds of this ware from unglazed portions of vessels (1989: 38) suggests that “red and buff wares” similar are present in the KNA assemblage, as the fabric is to those of the Byzantine and Hellenistic periods but similar to the brownish-red corrugated “Byzantine” with a higher amount of sand temper, continue into the wares discussed below. At Gharandal, this ware is Middle Islamic period at sites in the Negev. Another dated as Fatimid–Ayyubid, or Middle Islamic I–IIa possibility is that they were being reused at the site, (Walmsley and Grey 2001: 153). perhaps in masonry, as Nabataean, Roman, and Byz­ antine sherds were at Wuʿayra (Vannini and Tonghini Nondiagnostic and Early Sherds. A small num­ 1997: 381). Based only on survey data, however, it ber (N = 53) of nondiagnostic sherds, collected from buildings that also yielded diagnostic Middle Islamic 12 An earlier occupation currently seems unlikely, as the 2011 sherds, possibly belong to earlier periods. These may excavations in Area X (Survey Building 5300) revealed only a represent earlier periods of occupation at the site, but single phase of use. 88 JONES, LEVY, AND NAJJAR BASOR 368

Discussion

Considering both the ceramic and numismatic evi­ dence, the primary occupation of the site (or at least the occupation that produced strongly diagnostic ma­ terial) seems to have been in the Middle Islamic IIa, or the first half of the 13th centurya .d. The form of the underglaze-painted stonepaste wares found at KNA has been interpreted as late 13th century (Porter 1981: 40–41), but the presence of wasters with this rim type at Raqqa (Jenkins-Madina 2006: 62, 92–93) makes a post-1259 a.d. date for its introduction untenable. In fact, Mason’s (2004: 98–99) dating of the “arc-back” decorative motif places some of the stonepaste wares found at KNA as early as the late 12th, rather than late 13th, century. How long KNA was occupied is diffi­ cult to say without further excavation, especially since many of the handmade sherds found at KNA could date from anywhere in the Middle Islamic period, but Fig. 20. Corroded copper fals found in Building 5308 at two scenarios are plausible. Considering these sce­ KNA. Photo: Aaron Gidding. narios will also require some discussion of the dating of Middle Islamic copper production activities in the is not possible to determine the exact significance of vicinity of Khirbat Faynan. this pottery. Judging from the artifacts collected on the Currently, this dating relies on numismatic evi­ site in conjunction with the architecture, there is little dence and a single radiocarbon date obtained by the doubt that all the structures visible on the site surface Wadi Faynan Landscape Survey team from Faynan at KNA were constructed and used during the Middle 7 (WFLS Site WF5741), a slag mound near Khirbat Islamic period. Faynan identified by Hauptmann (2007: 103) as Iron Age. Five coins of the mid-14th century were col­ lected near Khirbat Faynan, although it is not certain The Coins that they are all associated with the smelting activ­ Numismatic evidence for dating KNA comes pri­ ities, as opposed to some other use of the area (Kind marily from the five copper fulus (sing. fals) and a sil­ et al. 2005: 188). A radiocarbon sample from Faynan ver half-dirham described by Kind et al. (2005). They 7 yielded a calibrated date between the early 15th found five Syrian Ayyubid coins and one Rum Seljuq and early 17th centuries a.d. (Mattingly et al. 2007b: coin,13 all dating to the period between 1200 and 1235 fig. 9:17). The WFLS team views this date as likely a.d. (Kind et al. 2005: 188), which provide additional 15th century (Newson et al. 2007: 364), but, given the support for KNA being active primarily during the context, the date is surprising, a point we will discuss early 13th century. In 2002, the JHF team found an­ in more detail later. Based on the evidence currently other copper fals in Building 5308 (fig. 20), but this available, Middle Islamic copper exploitation in the was too corroded to be identified as anything more area around Khirbat Faynan seems to have begun later specific than an Islamic-period coin. than at KNA, starting perhaps as late as the 14th cen­ tury and possibly continuing into the 15th century. The two possible scenarios regarding the length of the Middle Islamic occupation at KNA, then, are (1) 13 While Kind et al. (2005: 188) imply that they found the coins that KNA continued to be an active copper produc­ at KNA during survey, they also state that “most of the coins [de­ tion site following the beginning of copper production scribed in the 2005 paper] were bought at modest prices from the activities at Khirbat Faynan, or (2) that metallurgical few local bedouins and in particular from their children, who stayed with their herds during those years near Faynan” (Kind et al. 2005: activities shifted from KNA to Khirbat Faynan, per­ 170). It is difficult to determine which coins were found by the DBM haps in order to more easily exploit the copper mines team and which were purchased from local Faynan residents. of Wadi Salmina as opposed to those in Wadi al-Ghu­ 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 89 wayb, which had already been heavily exploited dur­ wares. This demonstrates at least some connection ing the Iron Age (Ben-Yosef et al. 2010; Levy et al. with the trade routes to Syria, which is not surprising, 2003). A gap in the industrial occupation of Faynan as copper from the site was certainly bound for places at some point in the 13th century is also possible, per­ beyond Faynan. More importantly, this also suggests haps related to the unstable political situation of the the presence of people—often political elites—who mid-13th century (Humphreys 1977) or the Mongol could afford imported glazed pottery, or perhaps the raids on southern Syria in a.d. 1260 (Jackson 1980), valuable items that may have been transported in it, at though it is unlikely that the Mongols had much ef­ the site (Milwright 1999: 511; 2006: 21). fect on southern Jordan directly.14 However, it does Similarly, over 20% of the ceramic assemblage seem that the Ayyubid prince al-Mughith’s relations consists of wheel-made vessels, which is rather high with the Mongols played some part in the decision of for a rural Middle Islamic site in Jordan. As Brown Sultan Baybars to execute him and place Karak under (2000: 93) notes, unglazed wheel-made vessels would direct Mamluk control in 1263 (Amitai-Preiss 1997: have been less expensive than glazed wares to produce 10–11). Based on the available evidence, it is difficult and purchase, but still would have required a larger in­ to determine which of these models is correct, but the vestment than similar handmade forms. Wheel-made lack of evidence for post-13th-century occupation at wares are thought to be much more common at urban KNA does suggest a shift in production from KNA to sites and regional centers such as Karak Castle, but Khirbat Faynan in the late 13th century. The biggest uncommon at most rural sites of the period in Trans­ piece of evidence against this is the position of KNA jordan (Brown 1992: 259–63). The large amount of on Nuqayb al-Asaymir, the small pass linking the site wheel-made pottery found during the excavation of a to Khirbat Faynan, which suggests that the two sites cistern at rural Khirbat Faris (65% of the sherd count) may have been active at the same time. On the other has challenged this view, however, and it is possible hand, it is not entirely certain that this pass was actu­ that further excavations at other rural sites will reveal ally in use during the Middle Islamic period. higher percentages of wheel-made forms than seen during survey (McQuitty et al. 1997: 189, 207). None­ anthropological implications theless, the surface ceramic assemblage from KNA includes a high number of wheel-made forms for a The Organization of Production Middle Islamic–period site, especially one in the area south of Wadi al-Hasa, which Milwright (2006: 20) The survey ceramic assemblage (fig. 14) allows for identifies as the “least economically developed part of some preliminary suggestions to be made about the or­ Jordan” during the Ayyubid period. ganization of copper production at KNA. Glazed wares At the same time, the presence of elites at the site make up only about 1.5% of the total number of sherds is hardly surprising. The slag mounds at KNA contain collected during the 2002 survey. Compared with rural an estimated 1,000 tons of slag (Hauptmann 2007: 126), sites on the Karak Plateau, where the mean percent­ which corresponds to 65–100 tons of copper produced.15 age of glazed wares is 14%, this is rather low (Brown While this is relatively small compared with Iron Age and 2000: 87–88, figs. 3, 4). The average on the Karak Pla­ Roman production in Faynan (Ben-Yosef 2010: 933–39; teau is increased by the 18 high-density sites, mostly Hauptmann 2007: 147), it is still a significant amount near Karak Castle, with percentages of glazed wares as of copper. Likewise, the site itself required a fairly sub­ high as 56%, compared with the 40 sites with 0–14% stantial architectural investment. Although the buildings glazed wares (Brown 2000: 88, fig. 4), but KNA still has a relatively low percentage of glazed wares. It is 15 interesting to note, however, that the Karak Plateau The amount of copper metal produced relies on a double es­ timate. First, the amount of slag in the mounds is estimated, and sites are dominated by monochrome-glazed wares, then the amount of metal produced is estimated by assuming a 1:10 where 18 of the 20 glazed sherds collected at KNA to 1:15 ratio of metal to slag (Hauptmann 2007: 147). Ben-Yosef come from underglaze-painted Syrian stonepaste (2010: 761–63) has demonstrated for the Iron Age that the slag con­ tent of most slag mounds is no more than 40% and suggests that 14 A Mongol governor of Karak was appointed in 1260, fol­ previous estimates of the amount of metal production represented by lowing the submission of al-Mughith, the Ayyubid prince who a slag mound may be rather high. These estimates have not yet been controlled Karak, to the Mongols, but this official had not yet left investigated for the Islamic period, but similar caveats are likely to Damascus prior to the Mongol defeat at the Battle of ʿAyn Jalut apply. This does not change our interpretation but is useful to keep (Amitai-Preiss 1997: 7–8). in mind. 90 JONES, LEVY, AND NAJJAR BASOR 368 are constructed of local materials and were clearly in­ copper mining tended to be functional rather than aesthetically pleasing, and the sugar industry it is very likely that they were all constructed during the Middle Islamic period to facilitate copper exploitation. Explaining the Revival of the Faynan Copper It is unlikely that copper production on this scale in a Industry place as marginal as Faynan would have been organized without some sort of government involvement. Two explanations for the resumption of copper pro­ duction in Faynan in the Middle Islamic period have The Layout of the Site been put forth, but neither is entirely satisfactory. The first, proposed by Gerd Weisgerber, is that “[t]he re­ The spatial layout of KNA also gives some clues as sumption of metal production [at Khirbat Faynan] in to the organization of copper production. Smelting oc­ Mamluk times was probably not a government initia­ curred only at the eastern and western margins of the tive, but an ad hoc, trial-and-error enterprise of no great site, although some other production activities were or long-lasting significance” (Weisgerber 2006: 27). likely going on in other parts of the site. This concern He argues that Middle Islamic metallurgical activities with keeping smelting away from domestic and ad­ around Khirbat Faynan consisted mostly of “recycling ministrative areas is distinct from what is seen nearby of ore dumps and slag” and that small-scale mining during the Iron Age at Khirbat al-Nahas, Khirbat al-Ja­ probably only took place at KNA (Weisgerber 2006: riya, and Khirbat al-Ghuwayba (see Levy, Ben-Yosef, 26). Clearly, this assertion does not explain the revival and Najjar 2012 for descriptions of these sites), where of copper production at KNA. As already discussed, smelting occurred throughout the site, and shows at it is unlikely that production at the scale of KNA was least some concern for the comfort of the people living organized without any government involvement. at KNA. Likewise, this layout is very different from However, even if this assertion is limited to Khirbat the independent production units evident during the Faynan, it is inadequate. It is likely that recycling of Early Islamic period at ʿArja in Oman, where each fur­ ancient slag did occur, and the relatively late radiocar­ nace, roasting installation, and slag heap tends to be bon date from Faynan 7 may indicate that experiments accompanied by a house, suggesting that copper pro­ with recycling occurred after mining in Faynan had ducers there maintained some independence (Weisger­ ceased, but evidence is also emerging that produc­ ber 1987: 158). In contrast, the layout of KNA, with tion at Khirbat Faynan did involve small-scale min­ only two smelting areas and watchtowers over the en­ ing. During the 2009 ELRAP field season, a survey trances to the site, suggests more centralized control conducted in Wadi Salmina revealed the presence of over production. four previously unknown mines (fig. 21). The pottery found there consists mostly of Middle to Late Islamic Production Systems handmade wares. We have already mentioned the dif­ ficulties in dating mines based on surface pottery, but A key fact in the study of production systems is this does suggest that they are associated with Middle that these systems have multiple components, and un­ Islamic copper exploitation in Faynan. Given their lo­ derstanding what these components consist of and the cation, roughly 5 km east of Khirbat Faynan (fig. 3), way they interact with one another is essential for un­ it seems most likely that ore from these mines would derstanding the system as a whole (Costin 2001: 277). have been transported to Khirbat Faynan rather than to Up to this point, we have focused primarily on the KNA. This operation would still have been small, es­ components of the Middle Islamic copper industry lo­ pecially if Faynan 2 and Faynan 6 are the only Middle cated in Faynan—in other words, the extractive side of Islamic slag mounds near Khirbat Faynan—together, the production system. While it is impossible to offer they likely contain less than 60 tons of slag (Haupt­ a complete reconstruction of this system at the pres­ mann 2007: 97, 103)—but this shows that Middle Is­ ent stage of our research, it is also necessary to look lamic smelting at Khirbat Faynan was probably not at other aspects of the production system in question. limited to recycling Roman and Iron Age waste ma­ In the following section, we offer some preliminary terials. thoughts on the way in which the Faynan copper in­ This assertion also fails to account for why the dustry interacted with the broader economic and po­ Faynan copper industry was revived in the 13th cen­ litical systems of Middle Islamic Transjordan. tury. Whitcomb (2006: 242) has suggested that the 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 91

Fig. 21. View from inside an adit—a horizontal entrance to a mine—in Wadi Salmina. Photo: Erez Ben-Yosef.

Early Islamic inhabitants of Zughar may have had 2007: 365). This would make sense in light of Haupt­ mining and smelting camps similar to those set up near mann’s (2006: 125) assertion that neither Faynan nor Timna to supply Ayla. The most likely place to find Timna contains enough ore, compared with sources these would be Faynan, and yet there is no evidence such as Cyprus, to be sustainably important at an in­ of Early Islamic smelting there. It is important to de­ ternational level. termine what socioeconomic factors made Faynan an One problem with this argument relates to the in­ attractive source of copper in the 13th century but not crease in the circulation of copper coins in Egypt. New­ in the 8th or 9th century. son et al. (2007: 365) note parenthetically that Schultz The second explanation, proposed by the Wadi (1998) was probably unaware of the possible connec­ Faynan Landscape Survey team (Newson et al. 2007), tion to Faynan. While this is true, it should be noted that attempts to address the question left open by Weis­ the historical sources he cites do not mention Levantine gerber’s (2006) model: why was Faynan an attractive copper either. Al-Maqrizi (1994: 71), Schultz’s main source of copper only in the Middle Islamic II? They source, explicitly links the increase in circulating cop­ point to the work of Schultz (1998: 135–38), who dem­ per coinage to the deliberate manipulation of European onstrates that the weight of fulus increased by about trade by the supervisor of the royal treasury, Mahmud 1 gram in the mid-14th century, and argue that this ibn ʿAli. Ashtor (1976: 305) has criticized this account, increase could be due to the new availability of copper which places the blame on a single person, as being from Khirbat Faynan and KNA (Newson et al. 2007: simplistic, and argues that the primary reason for the in­ 365). They also address the dependence of the Mam­ crease in copper coinage was simply the growing avail­ luks on copper imported from Europe at the end of the ability of European copper. While this may be true, the 14th century a.d., suggesting that this could reflect the important point is that al-Maqrizi does not mention any instability of Faynan as a copper source (Newson et al. connection to Levantine copper sources. 92 JONES, LEVY, AND NAJJAR BASOR 368

A second problem relates to the increase in the indigo as the main agricultural product of the Jordan weight of fulus. As Schultz (1998: 148) notes, the in­ Valley (LaGro 2002: 26). The 13th and 14th centuries crease in the weight of copper coins actually indicates were the most active periods for the sugar industry in that copper had become more scarce. Copper coins Jordan, although production continued into the 15th were accepted at face value in small transactions but century (Milwright 2008: 121). The main locus of the in larger ones would be exchanged by weight for sil­ industry was the Jordan Valley, where 34 sugar pro­ ver coins. Therefore, an increase in the weight of cop­ duction sites are known from the eastern side of the per coins suggests an increase in the value of copper Jordan River (Abu Dalu 1995: 38, fig. 1), and about against silver (Schultz 1998: 146–48). Considering 8 are known from the western side (LaGro 2002: 27– this, the increase in the weight of copper coins is actu­ 28). Closer to Faynan, three sugar mills are known in ally the result of copper becoming more scarce, not the southern aghwar, at least two of which—one in more plentiful. It is unlikely that changing standards Ghawr al-Safi and one in Ghawr al-Dhraʾ—are known for the weight of fulus in 14th-century Egypt were re­ as Tawahin al-Sukkar (Whitcomb 1992b: 114–15). lated to copper production in Faynan. Sugar production also possibly took place near Shaw­ A final issue to consider is the fact that it is un­ bak (Benvenisti 1970: 253; Brigitte-Porëe 1995: 419; necessary to look as far as Egypt to find demand for Milwright 2006: 15; Walker 2011: 104), although copper. Copper and copper alloys were used to make archaeological remains have been more difficult to a wide variety of objects during the Middle Islamic find there. Brooker and Knauf (1988: 185) suggest period, ranging from intricately decorated vessels (cf. that a mill in nearby Wadi al-Bustan is an Ayyubid al-ʿImari 1967; Rice 1952) to weights (Balog 1970) sugar mill but also point out that Shawbak’s Mediter­ to more utilitarian objects, some of which have been ranean climate is not well suited to growing sugarcane. chemically linked to Faynan/Timna copper (al-Saaʾd Milwright (2006: 15, n. 72) notes, however, that the 2000).16 Demand for copper in Jordan would have suggestion is interesting in light of 14th-century ref­ been high, and it is likely that Faynan copper would erences to sugar from “Cranco di Monreal.” Even if have served a more local need. sugar production did not take place near Shawbak, the Here we propose a third model that takes this into known sugar mills in the southern aghwar are still in account. It is rooted in more local economic processes close proximity to Faynan (fig. 22). that may be more appropriate to the available data. The process of sugar production has been summa­ Specifically, we suggest that the revival of the Faynan rized by a number of scholars (e.g., Abu Dalu 1995: copper industry in the first half of the 13th century 39–40; Burke 2004: 112; LaGro 2002: 30–34; von may be linked to the emergence of the south Levan­ Wartburg 1995: 85–87), often relying on al-Nuwayrī’s tine sugar industry, which was rapidly becoming more (1931) detailed 14th-century account. The following important during this period. brief summary of the process is derived from LaGro’s (2002: 30–34) doctoral dissertation. The sugarcane Copper and the Sugar Industry was cut in December and transported to the sugar factory, where it would be chopped and then crushed When exactly sugarcane began to be grown in the under a stone. After this, the cane was put in baskets southern Levant is not entirely clear. In some places, and squeezed under a press, and all of the juice sieved there is evidence that production began as early as the and taken to a boiling area. The juice was then boiled, late 10th century a.d. (Damati 2011; LaGro 2002: 25; sieved, and boiled again to achieve the right amount Taha 2009: 181), although on a fairly limited scale. of reduction; roughly 90% of the water must be evapo­ By the early 13th century, sugarcane had surpassed rated out of the juice before the sugar can crystallize, according to Von Wartburg (1995: 86). After boiling, 16 On the other hand, the use of lead-isotope analysis for ore the reduced liquid was poured into ceramic molds to sourcing has been criticized on the grounds that it is difficult to crystallize. These molds were also placed on top of come up with a good signature for most ore sources and that objects jars, which would collect the syrup draining out of the are often made of copper from a variety of sources due to trade and crystallizing sugar. Depending on the quality of the recycling (Budd et al. 1996). It is also worth noting that it is not final product, the crystallized sugar might also have possible to distinguish copper from Faynan and Timna using lead- isotope and other chemical methods, because these sources formed been mixed with water and milk and boiled again. together and were later separated by tectonic shifts (Hauptmann The relevant part of this process for our discus­ 2007: 68–69). sion is the boiling phase, as the large cauldrons in 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 93

cast bronze—but it seems likely that dusut were gen­ erally made of copper rather than an alloy. Burchard of Mount Sion, a German monk who traveled through the Levant in the late 13th century, wrote a brief de­ scription of the sugar production process and noted that “[t]he juice squeezed out of them [the crushed sugarcanes] is boiled in copper boilers” (Burchard of Mount Sion 1896: 99). Likewise, al-Nuwayrī refers to “cooking [sugar] in dusut of copper” (al-Nuwayrī 1931: 270). While historical sources make this point fairly clear, they have less to say about how and where dusut were produced. For this we must turn to data from excavation. Evidence for local production, or at least repair, of dusut has been found in excavations of several sugar mills. At Tawahin al-Sukkar in Ghawr al-Safi, a smith­ ing hearth bottom was surface collected (Photos-Jones et al. 2002: 606), and a possible Mamluk-period cop­ per-working installation has been found at Khirbat al- Shaykh ʿIsa (Zughar), Tawahin al-Sukkar’s associated settlement (Politis, O’Hea, and Papaioannou 2007: 206–7). Farther from Faynan, at Tawahin al-Sukkar near Jericho, a hoard of copper objects, ranging from hinges, chains, and needles to jewelry, was found (Taha 2009: 188). Additionally, excavations in the por­ tion of that site known as Tell al-Nuhas—which un­ fortunately had been bulldozed by the property owner prior to the excavation—revealed the foundations of a furnace and a large amount of copper slag (Taha 2001: 69). Likewise, small pieces of metal and slag, which were interpreted as evidence for the manufacture and repair of copper cauldrons, were found in 13th-cen­ tury contexts at Lower Horbat Manot, located in the western Galilee (Stern 2001: 300). These finds sug­ gest, although the evidence is somewhat sparse, that dusut were being manufactured or repaired either at or very close to sugar production sites, and that at many sites at least some of the metal used was obtained by recycling older copper objects.17 The extent to which

17 Copper recycling was by no means limited to the sugar in­ dustry. Several examples should suffice to illustrate the widespread Fig. 22. The locations of key sugar production and copper pro- nature of the practice. It is known that during the 11th and 12th cen­ duction sites discussed in the text. Data from Digital Archae- turies, recycling was part of the trade in copper objects between the ology Atlas of the Holy Land (DAAHL: http://daahl.ucsd.edu). Middle East and India. The Cairo Geniza contains references to the fact that “old or broken vessels and implements of all descriptions were sent from Aden to India and worked there into new utensils which the sugar was boiled, called dusut (sing. dast) according to order” (Goitein and Friedman 2008: 16). Likewise, a in Arabic, were made of copper. Dusut are often de­ large 11th-century hoard found at Tiberias was likely meant for re­ cycling, given the large amount of scrap it contained (Ponting 2008: scribed interchangeably as copper or bronze—Agno­ 59–60). Allan (1984: 91–92) notes that al-Jawhari and al-Maqrizi letti (2009: 144–45), for example, describes a dast in both describe an early 15th-century copper shortage during which the Karak Archaeological Museum as being made of copper coins were sent from Egypt to Yemen to be melted down 94 JONES, LEVY, AND NAJJAR BASOR 368 recycled copper could be used in the manufacture of It is also important to know how many dusut would new dusut is not entirely clear, however. If the purity have been needed at an average sugar factory. Al- of the metal was a concern, whether for heating of Nuwayrī (1931: 270) states that each millstone required the crushed sugar, the durability of the vessels, or per­ eight copper kettles in which to cook the pressed sug­ haps even the taste of the final product, then recycling ar.19 In terms of the Faynan copper industry, the best would not have been ideal, as contamination can occur place to look for archaeological confirmation of this during this process (Ponting 2008: 49–51). The likeli­ would be Tawahin al-Sukkar in Ghawr al-Safi. Exca­ hood that the revival of the Faynan copper industry vations there have revealed well-preserved remains of can be linked to the production of these vessels thus a mill house and its water supply, which have been in­ depends to some extent on the amount of copper that vestigated quite thoroughly (Photos-Jones et al. 2002: would have been required by the average sugar mill. 599–605; Politis et al. 2005: 319–25), but evidence for The obvious question to ask is, how much copper the boiling and refining stage of the process is limited was in a dast? Unfortunately, dusut are not often de­ to the results of a small trench in a waste dump in the scribed in detail in historical descriptions of the sugar northern part of the site. Surface finds did include a production process, nor are they often found archaeo­ “kiln wall” waster, which likely came from the wall of logically, likely owing to the recycling processes de­ a boiling hearth rather than from a pottery kiln (Pho­ scribed previously. One historical source that does tos-Jones et al. 2002: 606, 612), but little can be said describe these vessels, however, is the German trav­ about the layout and organization of the boiling room. eler Arnold von Harff, who, in his description of the Better evidence comes from well-preserved boil­ late 15th-century Egyptian sugar industry, explained ing rooms at Kouklia Stavros, Episkopi Serayia, and that sugar is boiled in a “great long kettle holding Kolossi in Cyprus. The boiling installations at all three nine or ten pails18 full [of ground sugarcane]” (von sites follow the same basic plan: a long hall with round Harff 1946: 99). Unfortunately, this account is both hearths along the wall for boiling cane juice, with later than and geographically removed from the 13th- vaulted fire chambers underneath the hearths (Solo­ century Levantine sugar industry, and his description midou-Ieronymidou 2007: 77). While the basic plan is of dusut suggests a different vessel from the one used the same, the number of hearths is somewhat variable. in the Levant and Cyprus. A better indicator of the At Kouklia Stavros, it is likely that up to eight hearths size and shape of these vessels in the 13th-century were in use at any one time (von Wartburg 1995: 104), Levant is one of the few preserved examples, a dast but at Kolossi, eight fire chambers were found, each in the Karak Archaeological Museum which was “ac­ of which would have had two hearths (Solomidou- cidentally discovered” at Tawahin al-Sukkar in Ghawr Ieronymidou 2007: 79). Some inferences can also be al-Safi in the 1980s and which likely dates to the 14th made for other sites. The boiling installations at Lower century (Nashef 2009: 141; Politis et al. 2005: 324). Horbat Manot were not actually revealed during ex­ This dast is circular, 102 cm in diameter, and esti­ cavation, but Stern (2001: 303) suggests that the site mated to weigh 150–250 kg (Nashef 2009: 141). Addi­ would have had 12 hearths based on a comparison with tional support for this being a standard size and shape Kouklia Stavros. Eight hearths, therefore, may be an was found during the excavations at Kouklia Stavros average, or even a minimum, and is a reasonable num­ in Cyprus, which revealed that sugar was boiled on ber to use in our estimates. round fireplaces ca. 1 m in diameter (von Wartburg Based on these numbers, a sugar production site and Maier 1989: 182). We can use this as an estimate would have required at least 1.2–2.0 tons of copper for dusut in general, and assume they were around 1 m for making dusut, in addition to any copper required in diameter and weighed somewhere between 150 and for repairs or replacements. Problems in dating sugar 250 kg. mills make it difficult to estimate the amount of cop­ per the sugar industry would have required in the early 13th century, but we can estimate more broadly. and recast into utilitarian objects. This process has also been docu­ mented ethnographically. In the mid-1960s, Wulff (1966: 22–23) If only known sugar mills within the modern bound­ observed Iranian copper workers purchasing and remelting scrap aries of Jordan are considered, 46–78 tons of copper copper. would have been required by the sugar industry for 18 A newer German translation (von Harff 2007: 111) gives “Ei­ mer” in place of “pails.” If this is correct, and refers to the same unit of measurement used for wine, von Harff is describing a very large 19 LaGro (2002: 29, n. 107) translates this as one large and eight vessel. Eight to nine eimers is roughly 450–500 liters. small kettles, which would make our estimates slightly low. 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 95 dusut.20 At the high end, this is more than the low-end netians, under pressure from the Byzantine emperor estimates for the total amount of copper produced at Leo V, attempted to ban trade with Egypt and Syria KNA. Even if much of this demand could be met by (Jacoby 2009: 372). Likewise, trade in iron, arms, and recycling scrap, a reliable source of copper in Jordan wood was again banned following the fall of Acre in would have been beneficial. 1291, and this prohibition was repeated by the Church, It is possible, then, that the exploitation of copper with limited success, throughout the early 14th cen­ resources in Faynan beginning in the 13th century was tury (Ashtor 1976: 298). Nonetheless, Lopez (1987: spurred on by the demand for copper for use in the 350) argues that the Italian cities did, at times, have to local sugar industry, especially given the proximity of somewhat reluctantly enforce these bans. Given this, Faynan to several sugar production sites in southern it is possible that, in the early 13th century, the Ayyu­ Jordan. Historical sources from the 16th and 17th cen­ bid elite were attracted to Faynan as a reliable, local turies recommend producing pottery related to sugar source of copper, similar to the situation proposed production on-site to reduce transport costs (von by Coughenour (1976: 75) for the resumption of iron Wartburg 1983: 299)—a practice that was apparently production at Mugharat al-Warda near ʿAjlun in the common throughout the Middle Islamic period, as ev­ Middle Islamic period. idenced by wasters of these vessels found at Tawahin Whether or not these prohibitions were responsible al-Sukkar and Khirbat al-Shaykh ʿIsa in Ghawr al-Safi for the revival of Faynan as a copper source in the (Photos-Jones et al. 2002: 612)— and a local source of Middle Islamic period, the expanding Jordanian sugar copper for dusut may have been desirable for the same industry is still the likely underlying motive. This is reason, especially since von Wartburg (1995: 100, n. not to say that all of the copper produced in Faynan 23) notes that many historical sources complain about would have been used to make dusut. As we have the high cost of these vessels. mentioned, a wide variety of copper utilitarian ob­ The availability of European copper at the begin­ jects were common during the Middle Islamic period, ning of the 13th century should also be considered. In and some Faynan copper would probably have gone the late 12th and 13th centuries, the Catholic Church toward their manufacture. Likewise, it is likely that made attempts to limit certain kinds of trade with Mus­ many dusut were recycled as scrap after the decline of lim countries. This began in 1179 with the Third Lat­ the sugar industry in the late 15th century. Nonethe­ eran Council, which specifically banned trading iron, less, the demand for copper by the sugar industry is a wood, and arms. A decade later, Pope Clement III viable explanation for the revival of the Faynan copper extended this embargo to include all trade with Mus­ industry in the 13th century. lims, during times of both war and peace (Freidenreich 2011: 46–47). Although these restrictions were eased conclusion by Pope Innocent III (Mažeika 1994: 66), the Fourth Lateran Council in 1215 again banned all trade with At this initial stage of our research at Khirbat Muslims, although with a set time limit of four years Nuqayb al-Asaymir, it is possible to make several (Schroeder 1937). It is unclear how often these bans tentative conclusions about the site. First, analysis of were actually obeyed, however. Ashtor (1976: 240– the 2002 survey pottery suggests a date in the Middle 41) notes that the Pisans made treaties with the Ayyu­ Islamic IIa, or the first half of the 13th century, which bids in 1173 and 1215—the year of the Fourth Lateran supports previously published numismatic evidence Council—with the goal of importing iron and wood to (Kind et al. 2005) and mostly agrees with the earlier Egypt; and the Venetians made similar treaties with report of pottery from the site (Hauptmann, Weisger­ the Ayyubids in both Egypt and Aleppo throughout ber, and Knauf 1985). We hope that excavations at the the early 13th century. These prohibitions—and the site will allow us to determine how long this occu­ tendency to ignore them—have precedents dating pation lasted. Second, the survey pottery assemblage back at least to the early ninth century, when the Ve­ is atypical for rural sites of the period, but it is also important to keep in mind that KNA is not a typical 20 This assumes that most sugar production sites contained only rural site; it was set up to produce a rather valuable one press, but this is not always the case. Tawahin al-Sukkar in and somewhat scarce commodity. As such, centralized Ghawr al-Safi has three presses (Politis 2010: 4). This makes it the largest sugar production site known in the southern Levant, but our organization and elite supervision are suggested, not present knowledge of many other sugar factories remains somewhat only by the pottery assemblage but by the nature and patchy. function of the site itself. 96 JONES, LEVY, AND NAJJAR BASOR 368

Moving to the Faynan copper industry in general, else, we hope our suggestions about this connection recent syntheses of the archaeology of Faynan (New­ will at least prompt further discussion of this aspect of son et al. 2007; Weisgerber 2006) do not adequately the Middle Islamic economy of Jordan. explain the underlying reasons for the reemergence of The ELRAP investigations of KNA, and the study copper production in the 13th century. It seems likely of the Middle Islamic copper industry in general, are that this revival was actually spurred by the sugar in­ still at an early stage. However, future excavation at dustry’s demands for copper, which were high enough KNA and at Middle Islamic Khirbat Faynan will cer­ to make a reliable, local source desirable, especially tainly add to our understanding of the copper industry if papal embargoes on European trade with Egypt and in Faynan during this period—not only of the organi­ the Levant were being enforced—or even if it was zational, technological, and economic aspects but also simply feared that they would be enforced. If nothing of the day-to-day lives of those who worked there.

acknowledgments

We are grateful to the Department of Antiquities of Jordan Adams and Jim Anderson for their participation during the and especially the late Dr. Fawwaz al-Khrayshah for his sup­ Jabal Hamrat Fidan project; Kyle Knabb, who supervised the port of the surveys described here. The fieldwork described 2007 Wadi al-Jariya survey; and Dr. Erez Ben-Yosef, who here was made possible by a generous grant awarded to co- supervised the 2009 ELRAP survey of Wadi al-Salmina and author TEL from the C. Paul Johnson Family Foundation and offered insightful comments on an earlier draft of this pa­ the National Geographic Committee on Research and Ex­ per. The peer reviewers of both submitted drafts of this paper ploration. This work was supported by the National Science also offered many helpful comments which have greatly im­ Foundation under IGERT Award #DGE-0966375, “Training, proved the final version, and we warmly thank them, as well. Research and Education in Engineering for Cultural Heritage Additionally, we would like to thank Dr. Bethany Walker of Diagnostics” at the University of California, San Diego’s Missouri State University for allowing co-author IWNJ ac­ Center of Interdisciplinary Science for Art, Architecture and cess to the MSU ceramic study collection, and for productive Archaeology (CISA3), California Institute for Telecommu­ discussions during his stay in Springfield. All errors in this nications and Information Technology (Calit2) to co-author work are, of course, the fault of the authors. We would also IWNJ. We are also grateful for the logistic support provided like to acknowledge GeoEye for generously providing the by the American Center of Oriental Research in Amman; its UCSD Levantine Archaeology Laboratory with the IKONOS previous director, Dr. Pierre Bikai; current director, Dr. Bar­ satellite imagery used in this paper. Finally, we are grateful to bara Porter; and associate director, Dr. Chris Tuttle. We would the Royal Society for the Conservation of Nature (RSCN) in also like to thank Lisa Soderbaum, who supervised the 2002 Jordan for their help and cooperation. Our site is located in UC San Diego survey at Khirbat Nuqayb al-Asaymir; Russell the Dana UNESCO Biosphere protected area.

references

Abu Dalu, R. A. Allan, J. W. 1995 Taqniyyat Maʿāṣir al-Sukkar fī Wādī al-Urdunn 1984 Shaʿbān, Barqūq, and the Decline of the Mamluk Khilāl al-Fitrāt al-Islāmiyya. Pp. 37–48 (Arabic Metalworking Industry. Muqarnas 2: 85–94. section) in Studies in the History and Archaeology Amitai-Preiss, R. of Jordan, Vol. 5: Art and Technology throughout 1997 Hülegü and the Ayyūbid Lord of Transjordan the Ages, ed. K. ʿAmr, F. Zayadine, M. Zaghloul, R. (More on the Mongol Governor of al-Karak). Schick, J. Jabajy, and N. R. Tawfiq. Amman: De­ Archivum Eurasiae Medii Aevi 9: 5–16. partment of Antiquities of Jordan. Ashtor, E. Agnoletti, S. 1976 A Social and Economic History of the Near 2009 The Sugar Bowl. Pp. 144–45 in Da Petra a East in the Middle Ages. Berkeley: University Shawbak: Archeologia di una frontiera, ed. G. of ­California. Vannini and M. Nucciotti. Florence: Giunti. Avissar, M., and Stern, E. J. Agricola, G. 2005 Pottery of the Crusader, Ayyubid, and Mamluk 1950 De Re Metallica. Trans. H. C. Hoover and L. H. Periods in Israel. IAA Reports 26. Jerusalem: Hoover, from Latin. New York: Dover. Israel Antiquities Authority. 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 97

Balog, P. 2006 Late Islamic Ceramic Sequences from el-Lejjûn: 1970 Islamic Bronze Weights from Egypt. Journal of Stratigraphic and Historical Contexts. Pp. 373– the Economic and Social History of the Orient 99 in The Roman Frontier in Central Jordan: 13: 233–56. Final Report on the Limes Arabicus Project, Ben-Yosef, E. 1980–1989, Vol. 2, ed. S. T. Parker. Dumbarton 2010 Technology and Social Process: Oscillations Oaks Studies 10. Washington, DC: Dumbarton in Iron Age Copper Production and Power in Oaks Research Library and Collection. Southern Jordan. Ph.D. dissertation, University Budd, P.; Haggerty, R.; Pollard, A. M.; Scalife, B.; and of California, San Diego. Thomas, R. G. Ben-Yosef, E.; Levy, T. E.; Higham, T.; Najjar, M.; and 1996 Rethinking the Quest for Provenance. Antiquity Tauxe, L. 70: 168–74. 2010 The Beginning of Iron Age Copper Production in Burchard of Mount Sion the Southern Levant: New Evidence from Khirbat 1896 Burchard of Mount Sion: A Description of the al-Jariya, Faynan, Jordan. Antiquity 84: 724–46. Holy Land. Trans. A. Stewart, from Latin. Pp. Benvenisti, M. 1–136 in Library of the Palestine Pilgrims’ Text 1970 The Crusaders in the Holy Land. Trans. P. Fitton, Society 12. Reprinted New York: AMS, 1971. from Hebrew. Jerusalem: Israel Universities. Burke, K. S. Brigitte-Porëe, P. 2004 A Note on Archaeological Evidence for Sugar 1995 Les Moulins et fabriques à sucre de Palestine et Production in the Middle Islamic Periods in Bilād de Chypre: Histoire, géographie et technologie al-Shām. Mamlūk Studies Review 8: 109–18. d’une production croisée et médiévale. Pp. 377– Costin, C. L. 510 in Cyprus and the Crusades: Papers Given 2001 Craft Production Systems. Pp. 273–327 in at the International Conference “Cyprus and the ­Archaeology at the Millennium: A Sourcebook, Crusades,” Nicosia, 6–9 September, 1994, ed. ed. G. M. Feinman and T. D. Price. New York: N. Coureas and J. Riley-Smith. Nicosia: Society Kluwer Academic/Plenum. for the Study of the Crusades and the Latin East: Coughenour, R. A. Cyprus Research Center. 1976 Preliminary Report on the Exploration and Ex­ Brooker, C. H., and Knauf, E. A. cavation of Mugharat el-Wardeh and Abu Tha­ 1988 Review of Crusader Institutions, by J. Prawer. wab. Annual of the Department of Antiquities of Zeitschrift des Deutschen Palästina-Vereins ­Jordan 21: 71–78, 186–89. 104: 184–88. Damati, E. Brown, R. M. 2011 A Sugar Industry Site from the Fatimid to the 1980 The 1976 ASOR Soundings. Pp. 105–32 in The Ottoman Periods at Tell Umm al-Faraj, Western Excavations at Araq el-Emir, Vol. 1, ed. N. L. Galilee. ʿAtiqot 65: 139–59 (Hebrew), 77* (En­ Lapp. Annual of the American Schools of Orien­ glish summary). tal Research 47. [Ann Arbor]: American Schools Damgaard, K. of Oriental Research. 2009 A Palestinian Red Sea Port on the Egyptian Road 1987 A 12th Century a.d. Sequence from Southern to Arabia: Early Islamic Aqaba and Its Many Transjordan: Crusader and Ayyubid Occupation Hinterlands. Pp. 85–97 in Connected Hinter- at el-Wuʿeira. Annual of the Department of An- lands: Proceedings of Red Sea Project IV, Held tiquities of Jordan 31: 267–88. at the University of Southampton, September 1988 Summary Report of the 1986 Excavations: Late 2008, ed. L. Blue, J. Cooper, R. Thomas, and Islamic Shobak. Annual of the Department of J. Whitewright. Society for Arabian Studies Antiquities of Jordan 32: 225–45. Monographs 8; BAR International Series 2052. 1992 Late Islamic Ceramic Production and Distribu­ Oxford: Archaeopress. tion in the Southern Levant: A Socio-Economic Dothan, M. and Political Interpretation. Ph.D. dissertation, 1955 The Excavations at ʿAfula. ʿAtiqot 1: 19–74. Binghamton University. Franken, H. J., and Kalsbeek, J. 2000 The Distribution of Thirteenth- to Fifteenth- 1975 Potters of a Medieval Village in the Jordan Val- Century Glazed Wares in Transjordan: A Case ley: Excavations at Tell Deir ʿAlla: A Medieval Study from the Kerak Plateau. Pp. 84–99 in The Tell, Tell Abu Gourdan, Jordan. North-Holland Archaeology of Jordan and Beyond: Essays in Ceramic Studies in Archaeology 3. Amsterdam: Honor of James A. Sauer, ed. L. E. Stager, J. A. North-Holland. Greene, and M. D. Coogan. Studies in the Ar­ Freidenreich, D. M. chaeology and History of the Levant 1. Winona 2011 Muslims in Western Canon Law, 1000–1500. Lake, IN: Eisenbrauns. Pp. 41–68 in Christian-Muslim Relations. A 98 JONES, LEVY, AND NAJJAR BASOR 368

Bibliographical History, Vol. 3: 1050–1200, ed. 2012 “Catching the Eel”—Documentary Evidence for D. Thomas and A. Mallett. History of Christian- Concepts of the Arabic Book in the Middle Pe­ Muslim Relations 15. Leiden: Brill. riod. Journal of Arabic and Islamic Studies 12: Friedman, H. A. 224–34. 2008 Industry and Empire: Administration of the Ro­ Hodgson, M. G. S. man and Byzantine Faynan. 2 vols. Ph.D. dis­ 1974 The Venture of Islam: Conscience and History in sertation, University of Leicester. a World Civilization, Vol. 1: The Classical Age Galloway, J. H. of Islam. Chicago: University of Chicago. 1977 The Mediterranean Sugar Industry. Geographi- Hübner, U. cal Review 67: 177–94. 2004 Qurayyāt el-Manṣūr und Ḫirbet el-Faiḍ in Süd­ Glidden, H. W. jordanien. Zeitschrift des Deutschen Palästina- 1941 Review of The Other Side of the Jordan, by Vereins 120: 141–56. N. Glueck. American Journal of Archaeology Humphreys, R. S. 45: 118–20. 1977 From Saladin to the Mongols: The Ayyubids of Glueck, N. Damascus, 1193–1260. Albany: State Univer­ 1935 Explorations in Eastern Palestine, II. Annual sity of New York. of the American Schools of Oriental Research al-ʿImari, A. A. 15. New Haven: American Schools of Oriental 1967 Studies in Some Islamic Objects Newly Discov­ ­Research. ered at Qus. Annales islamologiques 7: 121–38. 1940 The Other Side of the Jordan. New Haven: Jackson, P. American Schools of Oriental Research. 1980 The Crisis in the Holy Land in 1260. English Goitein, S. D., and Friedman, M. A. Historical Review 95: 481–513. 2008 India Traders of the Middle Ages: Documents Jacoby, D. from the Cairo Geniza (“India Book,” Part One). 2009 Venetian Commercial Expansion in the Eastern Études sur le Judaïsme Médiéval 31. Leiden: Mediterranean, 8th–11th Centuries. Pp. 371–91 Brill. in Byzantine Trade, 4th–12th Centuries: The Ar- Hadad, S. chaeology of Local, Regional and International 1999 Oil Lamps from the Abbasid through the Mam­ Exchange: Papers of the Thirty-Eighth Spring luk Periods at Bet Shean, Israel. Levant 31: 203– Symposium of Byzantine Studies, St John’s Col- 24. lege, University of Oxford, March 2004, ed. Hart, S. M. M. Mango. Publications of the Society for the 1986 Some Preliminary Thoughts on Settlement in Promotion of Byzantine Studies. 14. Burlington, Southern Edom. Levant 18: 51–58. VT: Ashgate. Hart, S., and Falkner, R. K. Jenkins-Madina, M. 1985 Preliminary Report on a Survey in Edom, 1984. 2006 Raqqa Revisited: Ceramics of Ayyubid Syria. Annual of the Department of Antiquities of Jor- New York: Metropolitan Museum of Art. dan 29: 255–77. Johns, J. Hauptmann, A. 1994 The Longue Durée: State and Settlement Strate­ 2006 Mining Archaeology and Archaeometallurgy gies in Southern Transjordan across the Islamic in the Wadi Arabah: The Mining Districts of Centuries. Pp. 1–31 in Village, Steppe and State: Faynan and Timna. Pp. 125–33 in Crossing the The Social Origins of Modern Jordan, ed. E. L. Rift: Resources, Routes, Settlement Patterns, Rogan and T. Tell. London: British Academic. and Interaction in the Wadi Arabah, ed. P. Bien­ 1998 The Rise of Middle Islamic Hand-Made Geo­ kowski and K. Galor. Levant Supplementary metrically-Painted Ware in Bilad al-Sham (11th– ­Series 3. Oxford: Oxbow. 13th Centuries a.d.). Pp. 65–93 in Colloque 2007 The Archaeometallurgy of Copper: Evidence international d’archéologie islamique: IFAO, from Faynan, Jordan. New York: Springer. le Caire, 3–7 février 1993, ed. R.-P. Gayraud. Hauptmann, A.; Weisgerber, G.; and Knauf, E. A. Textes arabes et études islamiques 36. Cairo: In­ 1985 Archäometallurgische und bergbauarchäolo­ stitut français d’archéologie orientale. gische Untersuchungen im Gebiet von Fenan, Kind, H. D.; Gilles, K. J.; Hauptmann, A.; and Weisger­ Wadi Arabah (Jordanien). Der Anschnitt 37: ber, G. 163–95. 2005 Coins from Faynan, Jordan. Levant 37: 169–95. Hirschler, K. King, G. R. D.; Lenzen, C. J.; Newhall, A.; King, J. L.; 2006 Medieval Arabic Historiography: Authors as Deemer, J. K.; and Rollefson, G. O. Actors. SOAS/Routledge Studies on the Middle 1989 Survey of Byzantine and Islamic Sites in Jor­ East 5. London: Routledge. dan, Third Preliminary Report (1982): The Wadi 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 99

ʿArabah (Part 2). Annual of the Department of Lopez, R. S. Antiquities of Jordan 33: 199–215. 1987 The Trade of Medieval Europe: The South. Pp. Knabb, K. A.; Jones, I. W. N .; Levy, T. E.; and Najjar, M. 306–401 in The Cambridge Economic History Forth- Patterns of Iron Age Mining and Iron Age Settle- of Europe, Vol. 2: Trade and Industry in the coming ment in Jordan’s Faynan District—the Wadi Middle Ages, ed. M. M. Postan and E. Miller. al-Jariya Survey in Context. New Insights into 2nd ed. Cambridge: Cambridge University. the Iron Age Archaeology of Edom, Southern MacDonald, B. Jordan: Surveys, Excavations and Research 1988 The Wadi el Ḥasā Archaeological Survey, 1979– from the Edom Lowlands Regional Archaeology 1983, West-Central Jordan. Waterloo, Ontario: Project (ELRAP), ed. T. E. Levy, M. Najjar, and Wilfrid Laurier University. E. Ben-Yosef. Los Angeles: Cotsen Institute of 1992 The Southern Ghors and Northeast ʿArabah Ar- Archaeology. chaeological Survey. Sheffield Archaeological LaGro, H. E. Monographs. Sheffield: Collis.. 2002 An Insight into Ayyubid-Mamluk Pottery: De­ MacDonald, B.; Herr, L. G.; Neeley, M. P.; Gagos, T.; scription and Analysis of a Corpus of Mediaeval Moumani, K.; and Rockman, M. Pottery from the Cane Sugar Production and Vil­ 2004 The Tafila-Busayra Archaeological Survey 1999– lage Occupation at Tell Abu Sarbut in Jordan. 2001, West-Central Jordan. American Schools of Ph.D. dissertation, Universiteit Leiden. Oriental Research Archaeological Reports 9. Bos­ Levy, T. E.; Adams, R. B.; Anderson, J. D.; Najjar, M.; ton: American Schools of Oriental Research. Smith, N.; Arbel, Y.; Soderbaum, L.; and Mu­ al-Maqrizi, A. niz, A. 1994 Mamluk Economics: A Study and Translation of 2003 An Iron Age Landscape in the Edomite Low­ al-Maqrīzī’s Ighāthah. Trans. A. Allouche, from lands: Archaeological Surveys along Wādī Arabic. Salt Lake City: University of Utah. al-Ghuwayb and Wādī al-Jāriya, Jabal Ḥamrat Mason, R. B. J. Fīdān, Jordan, 2002. Annual of the Department 2004 Shine Like the Sun: Lustre-Painted and Asso- of Antiquities of Jordan 47: 247–77. ciated Pottery from the Medieval Middle East. Levy, T. E.,; Anderson, J. D.; Waggoner, M.; Smith, N.; Bibliotheca Iranica, Islamic Art and Architec­ Muniz, A.; and Adams, R. B. ture Series 12. Costa Mesa, CA: Mazda. 2001 Interface: Archaeology and Technology—Digi­ Mattingly, D.; Newson, P.; Creighton, O.; Tomber, R.; Grat­ tal Archaeology 2001: GIS-Based Excavation tan, J.; Hunt, C.; Gilbertson, D.; el-Rishi, H.; Recording in Jordan. The SAA Archaeological and Pyatt, B. Record 1/3: 23–29. 2007a A Landscape of Imperial Power: Roman and Levy, T. E.,; Ben-Yosef, E.; and Najjar, M. Byzantine Phaino. Pp. 305–48 in Archaeology 2012 New Perspectives on Iron Age Copper Produc­ and Desertification: The Wadi Faynan Land- tion and Society in the Faynan Region, Jor­ scape Survey, Southern Jordan, ed. G. Barker, dan. Pp. 197–214 in Eastern Mediterranean D. Gilbertson, and D. Mattingly. Wadi Faynan Metallurgy and Metalwork in the 2nd Millen- Series 2; Levant Supplementary Series 6. Ox­ nium bc: A Conference in Honour of James D. ford: Oxbow. Muhly, Nicosia, 10th–11th October 2009. ed. Mattingly, D.; Newson, P.; Grattan, J.; Tomber, R.; V. Kassianidou and G. Papasavvas. Oxford: Barker, G.; Gilbertson, D.; and Hunt, C. Oxbow. 2007b The Making of Early States: The Iron Age and Levy, T. E., and Smith, N. G. Nabataean Periods. Pp. 271–303 in Archaeology 2007 On-Site GIS Digital Archaeology: GIS-based and Desertification: The Wadi Faynan Land- Excavation Recording in Southern Jordan. Pp. scape Survey, Southern Jordan, ed. G. Barker, 47–58 in Crossing Jordan: North American D. Gilberston, and D. Mattingly. Wadi Faynan Contributions to the Archaeology of Jordan, ed. Series 2; Levant Supplementary Series 6. Ox­ T. E. Levy, P. M. M. Daviau, R. W. Younker, and ford: Oxbow. M. Shaer. London: Equinox. Mažeika, R. Lindner, M. 1994 Of Cabbages and Knights: Trade and Trade Trea­ 1999 Late Islamic Villages in the Greater Petra Region ties with the Infidel on the Northern Frontier, and Medieval “Hormuz.” Annual of the Depart- 1200–1390. Journal of Medieval History 20: ment of Antiquities of Jordan 43: 479–500. 63–76. Lindner, M.; Knauf, E. A.; Hübner, U.; and Hübl, J. McQuitty, A. 1998 From Edomite to Late Islamic: Jabal as-Suffaha 2005 The Rural Landscape of Jordan in the Seventh– North of Petra. Annual of the Department of An- Nineteenth Centuries ad: The Kerak Plateau. tiquities of Jordan 42: 225–40. Antiquity 79: 327–38. 100 JONES, LEVY, AND NAJJAR BASOR 368

2007 Khirbat Fāris: Vernacular Architecture on the Politis, K. D.; Kelly, A.; Hull, D.; and Foote, R. Karak Plateau, Jordan. Mamlūk Studies Review 2005 Survey and Excavations in the Ghawr as-Safi 11: 157–71. 2004. Annual of the Department of Antiquities McQuitty, A., and Falkner, R. of Jordan 49: 313–26. 1993 The Faris Project: Preliminary Report on the Politis, K. D.; O’Hea, M.; and Papaioannou, G. A. 1989, 1990 and 1991 Seasons. Levant 25: 37–61. 2007 Ghawr as-Safi Survey and Excavations 2006– McQuitty, A. M.; Sarley-Pontin, A.; Khoury, M.; Charles, 2007. Annual of the Department of Antiquities M. P.; and Hoppe, C. F. of Jordan 51: 199–210. 1997 Mamluk Khirbat Faris. ARAM 9: 181–226. Ponting, M. J. Milwright, M. 2008 The Scientific Analysis and Investigation of a 1999 Pottery in the Written Sources of the Ayyubid– Selection of the Copper-Alloy Metalwork from Mamluk Period (c. 567–923/1171–1517). Bulle- Tiberias. Pp. 35–61 in Tiberias: Excavations in tin of the School of Oriental and African Studies the House of the Bronzes, Final Report, Vol. 1: 62: 504–18. Architecture, Stratigraphy and Small Finds, ed. 2000 Pottery of Bilad al-Sham in the Ottoman Period: Y. Hirschfeld and O. Gutfeld. Qedem 48. Jerusa­ A Review of the Published Archaeological Evi­ lem: Hebrew University of Jerusalem. dence. Levant 32: 189–208. Porter, V. 2006 Central and Southern Jordan in the Ayyubid 1981 Medieval Syrian Pottery (Raqqa Ware). Oxford: Period: Historical and Archaeological Perspec­ Ashmolean Museum. tives. Journal of the Royal Asiatic Society 16: Porter, V., and Watson, O. 1–27. 1987 “Tell Minis” Wares. Pp. 175–248 in Syria and 2008 The Fortress of the Raven: Karak in the Middle Iran: Three Studies in Medieval Ceramics, ed. J. Islamic Period (1100–1650). Islamic History and Allan and C. Roberts. Oxford Studies in Islamic Civilization, Studies and Texts 72. Leiden: Brill. Art 4. Oxford: Oxford University. 2010 An Introduction to Islamic Archaeology. Edin­ Pringle, D. burgh: Edinburgh University. 1986 The Red Tower (al-Burj al-Ahmar): Settlement Musil, A. in the Plain of Sharon at the Time of the Cru- 1907 Arabia Petraea, Vol. 2: Edom. 2 vols. Vienna: saders and Mamluks a.d. 1099–1516. British Hölder. School of Archaeology in Jerusalem Monograph Nashef, K. Series 1. London: British School of Archaeology 2009 Bronze Cauldron. P. 141 in Da Petra a Shawbak: in Jerusalem. Archeologia di una frontiera, ed. G. Vannini and Rabbaʿ, I. M. Nucciotti. Florence: Giunti. 1991 The Geology of the Al Qurayqira (Jabal Hamra Newson, P., Mattingly, D.; Daly, P.; Tomber, R.; el-­Rishi, Faddan): Map Sheet 3051 II. 1:50,000 Geologi­ H.; Gilberston, D.; Grattan, J.; Hunt, C.; cal Mapping Series; Geology Bulletin 28. Am­ McLaren, S.; and Pyatt, B. man: Royal Jordanian Geographic Centre. 2007 The Islamic and Ottoman Periods. Pp. 349–68 Rice, D. S. in Archaeology and Desertification: The Wadi 1952 Studies in Islamic Metal Work. Bulletin of the Faynan Landscape Survey, Southern Jordan, ed. School of Oriental and African Studies 14: G. Barker, D. Gilberston, and D. Mattingly. Wadi 564–78. Faynan Series, 2; Levant Supplementary Series Rosen, S. A., and Goodfriend, G. A. 6. Oxford: Oxbow. 1993 An Early Date for Gaza Ware from the Northern al-Nuwayrī Negev. Palestine Exploration Quarterly 125: 1931 Nihāyat al-arab fī funūn al-adab, Vol. 8. Cairo: 143–48. Maṭbaʿat Dār al-Kutub al-Miṣrīya. Rothenberg, B. Photos-Jones, E.; Politis, K. D.; James, H. F.; Hall, A. J.; 1988 Early Islamic Copper Smelting—and Worship— Jones, R. E.; and Hamer, J. at Beer Ora, Southern Arabah (Israel). Institute 2002 The Sugar Industry in the Southern Jordan Val­ for Archaeo-Metallurgical Studies Newsletter ley: An Interim Report on the Pilot Season of Ex­ 12: 1–4. cavations, Geophysical and Geological Surveys al-Saaʾd, Z. at Tawahin as-Sukkar and Khirbat ash-Shaykh 2000 Technology and Provenance of a Collection ʿIsa, in Ghawr as-Safi. Annual of the Department of Islamic Copper-Based Objects as Found by of Antiquities of Jordan 46: 591–614. Chemical and Lead Isotope Analysis. Archae- Politis, K. D. ometry 42: 385–97. 2010 Ancient Landscapes of the Ghor es-Sāfī: Sur­ Sauer, J. A. veys and Excavations 1997–2009. ACOR News- 1973 Heshbon Pottery 1971: A Preliminary Report on letter 22/2: 1–5. the Pottery from the 1971 Excavations at Tell 2012 MIDDLE ISLAMIC METALLURGY IN FAYNAN 101

Ḥesbân. Andrews University Monographs 7. Faynan Landscape Survey, Southern Jordan, ed. Berrien Springs, MI: Andrews University. G. Barker, D. Gilberston, and D. Mattingly. Wadi 1976 Pottery Techniques at Tell Deir ʿAllā . Bulletin of Faynan Series 2; Levant Supplementary Series the American Schools of Oriental Research 224: 6. Oxford: Oxbow. 91–94. Tonghini, C. Scanlon, G. T. 1994 The Fine Wares of Ayyubid Syria. Pp. 248–93 1984 Mamluk Pottery: More Evidence from Fustat. in Cobalt and Lustre: The First Centuries of Muqarnas 2: 115–26. ­Islamic Pottery, ed. E. J. Grube. Nasser D. Kha­ Schaefer, J. lili Collection of Islamic Art 9. London: Nour 1989 Archaeological Remains from the Medieval Is­ Foundation. lamic Occupation of the Northwest Negev Des­ 1998 Qalʿat Jaʿbar Pottery: A Study of a Syrian Forti- ert. Bulletin of the American Schools of Oriental fied Site of the Late 11th–14th Centuries. British Research 274: 33–60. Academy Monographs in Archaeology 11. Ox­ Schroeder, H. J. ford: Oxford University. 1937 Twelfth Ecumenical Council: Lateran IV 1215. Tsugitaka, S. Internet Medieval Sourcebook. http://www. 1997 State and Rural Society in Medieval Islam: Sul- fordham.edu/halsall/basis/lateran4.asp. tans, Muqtaʿs and Fallahun. Islamic History and Schultz, W. C. Civilization, Studies and Texts 17. Leiden: Brill. 1998 Mahmud Ibn ʿAli and the “New Fulus”: Late Tushingham, A. D. Fourteenth Century Mamluk Egyptian Copper 1985 Excavations in Jerusalem, 1961–1967, Vol. 1. Coinage Reconsidered. American Journal of British Academy Monographs in Archaeology 2. Numismatics 10: 127–48. Toronto: Royal Ontario Museum. Sinibaldi, M. Vannini, G., and Tonghini, C. 2009 The Franks in Southern Transjordan and the 1997 Mediaeval Petra: The Stratigraphic Evidence Contribution of Ceramic Studies: A Preliminary from Recent Archaeological Excavations at al- Report on the Pottery Assemblages of al-Bayda Wuʾayra. Pp. 371–84 in Studies in the History and Wadi Farasa. Annual of the Department of and Archaeology of Jordan, Vol. 6: Landscape Antiquities of Jordan 53: 449–64. Resources and Human Occupation in Jordan Smith, A. M., II; Stevens, M.; and Niemi, T. M. throughout the Ages, ed. G. Bisheh, M. Zagh­ 1997 The Southeast Araba Archaeological Survey: A loul, and I. Kehrberg. Amman: Department of Preliminary Report of the 1994 Season. Bulletin Antiquities of Jordan. of the American Schools of Oriental Research von Harff, A. 305: 45–71. 1946 The Pilgrimage of Arnold von Harff, Knight: Solomidou-Ieronymidou, M. From Cologne, through Italy, Syria, Egypt, Ara- 2007 The Crusaders, Sugar Mills and Sugar Pro­ bia, Ethiopia, Nubia, Palestine, Turkey, France duction in Medieval Cyprus. Pp. 63–81 in Ar- and Spain, which He Accomplished in the Years chaeology and the Crusades: Proceedings of 1496 to 1499. Trans. M. Letts, from German. the Round Table, Nicosia, 1 February 2005, ed. Works Issued by the Hakluyt Society, Second P. Edbury and S. Kalopissi-Verti. Athens: Pier­ Series 94. London: Hakluyt Society. ides Foundation. 2007 Rom, Jerusalem, Santiago: Das Pilgertagebuch Stern, E. J. des Ritters Arnold von Harff (1496–1498), ed. 2001 The Excavations at Lower Ḥorbat Manot: A and trans. H. Brall-Tuchel and F. Reichert. Co­ Medieval Sugar-Production Site. ʿAtiqot 42: logne: Böhlau. 277–308. von Wartburg, M.-L. Taha, H. 1983 The Medieval Cane Sugar Industry in Cyprus: 2001 The Excavation of Tawaheen es-Sukkar in the Results of Recent Excavation. The Antiquaries Jordan Valley. Orient Express 2001/3: 68–71. Journal 63: 298–314. 2009 Some Aspects of Sugar Production in Jericho, 1995 Design and Technology of the Medieval Re­ Jordan Valley. Pp. 181–91 in A Timeless Vale: fineries of the Sugar Cane in Cyprus: A Case Archaeological and Related Essays on the Jor- Study in Industrial Archaeology. Pp. 81–116 in dan Valley in Honour of Gerrit van der Kooij on Paisajes del azúcar: Actas del Quinto seminario the Occasion of His Sixty-Fifth Birthday, ed. E. internacional sobre la caña de azucar, ed. A. Kaptijn and L. P. Peti. Archaeological Studies Malpica. Granada, Spain: Diputación Provincial Leiden University 19. Leiden: Leiden University. de Granada. Tomber, R. von Wartburg, M.-L., and Maier, F. G. 2007 Classical and Islamic Pottery. Pp. 445–64 in 1989 Excavations at Kouklia (Palaepaphos), 15th Pre­ Archaeology and Desertification: The Wadi liminary Report: Seasons 1987 and 1988. Report 102 JONES, LEVY, AND NAJJAR BASOR 368

of the Department of Antiquities of Cyprus 1989: Weisgerber, G. 177–93. 1987 Archaeological Evidence of Copper Exploitation Walker, B. J. at ʿArja. Journal of Oman Studies 9: 145–72. 1999 Militarization to Nomadization: The Middle and 2006 The Mineral Wealth of Ancient Arabia and Its Late Islamic Periods. Near Eastern Archaeology Use I: Copper Mining and Smelting at Feinan 62: 202–32. and Timna: Comparison and Evaluation of 2007 Sowing the Seeds of Rural Decline? Agriculture Techniques, Production, and Strategies. Arabian as an Economic Barometer for Late Mamluk Jor­ ­Archaeology and Epigraphy 17: 1–30. dan. Mamlūk Studies Review 11: 173–99. Whitcomb, D. S. 2009 Identifying the Late Islamic Period Ceramically: 1992a Reassessing the Archaeology of Jordan of the Preliminary Observations on Ottoman Wares Abbasid Period. Pp. 385–90 in Studies in the from Central and Northern Jordan. Pp. 37–65 in History and Archaeology of Jordan, Vol. 4, ed. Reflections of Empire: Archaeological and Eth- S. Tell. Amman: Department of Antiquities of nographic Studies on the Pottery of the Ottoman Jordan. Levant, ed. B. J. Walker. Annual of the American 1992b The Islamic Period as Seen from Selected Sites. Schools of Oriental Research 64. Boston: Amer­ Pp. 113–18 in The Southern Ghors and North- ican Schools of Oriental Research. east ʿArabah Archaeological Survey, ed. B. Mac­ 2011 Jordan in the Late Middle Ages: Transformation Donald. Sheffield Archaeological Monographs of the Mamluk Frontier. Chicago Studies on the 5. Sheffield: Collis. Middle East 8. Chicago: Middle East Documen­ 1997 Mamluk Archaeological Studies: A Review. tation Center. Mamlūk Studies Review 1: 97–106. Walker, B. J., and LaBianca, Ø. S. 2001 Umayyad and Abassid Periods. Pp. 503–13 in 2003 The Islamic Quṣūr of Tall Ḥisbān: Preliminary The Archaeology of Jordan. ed. B. MacDonald, Report on the 1998 and 2001 Seasons. Annual R. Adams, and P. Bienkowski. Levantine Ar­ of the Department of Antiquities of Jordan 47: chaeology 1. Sheffield: Sheffield Academic. 443–71. 2006 Land behind Aqaba: The Wadi Arabah during Walmsley, A. G. the Early Islamic Period. Pp. 239–42 in Crossing 1997 Land, Resources and Industry in Early Islamic the Rift: Resources, Routes, Settlement Patterns Jordan (Seventh–Eleventh Century). Current and Interactions in the Wadi Arabah, ed. P. Bi­ Research and Future Directions. Pp. 345–51 in enkowski and K. Galor. Levant Supplementary Studies in the History and Archaeology of Jor- Series 3. Oxford: Oxbow. dan, Vol. 6: Landscape Resources and Human 2009 Ayla at the Millennium: Archaeology and His­ Occupation in Jordan throughout the Ages, ed. tory. Pp. 123–32 in Studies in the History and G. Bisheh, M. Zaghloul, and I. Kehrberg. Am­ Archaeology of Jordan, Vol. 10: Crossing Jor- man: Department of Antiquities of Jordan. dan, ed. F. al-Khraysheh, R. Harasheh, S. Nsour, 2001 Restoration or Revolution? Jordan between the H. Taher, and S. Khouri. Amman: Department of Islamic Conquest and the Crusades: Impres­ Antiquities of Jordan. sions of Twenty-Five Years of Archaeological Willies, L. Research. Pp. 633–40 in Studies in the History 1991 Ancient Copper Mining at Wadi Amram, Israel: and Archaeology of Jordan, Vol. 7: Jordan by An Archaeological Survey. Bulletin of the Peak the Millennia, ed. G. Bisheh. Amman: Depart­ District Mines Historical Society 11: 109–38. ment of Antiquities of Jordan. Wulff, H. E. Walmsley, A. G., and Grey, A. D. 1966 The Traditional Crafts of Persia: Their Develop- 2001 An Interim Report on the Pottery from Gharan­ ment, Technology, and Influence on Eastern and dal (Arindela), Jordan. Levant 33: 139–64. Western Civilizations. Cambridge, MA: M.I.T.