Introduction

1The ‘Akko plain is the northernmost coastal plain of modern-day Israel. On the east, it is edged by the highlands of Galilee, which rise to 1,000 m. Warm springs and autumns and hot summers characterize the climate. During the winter, the winds blowing from the Mediterranean Sea decrease temperature and lead to heavy rains. The rainwater as well as numerous springs, feed the streams, running down from the hilly area towards the sea all year long. All these factors render the ‘Akko plain suitable for the cultivation of sugar cane and for sugar production, which is heavily dependent on water for irrigation and for operating the crushing mills.

2Salvage excavations and surveys in the ‘Akko plain, conducted by the Israel Antiquities Authority (hereafter IAA) in the past two decades, revealed five different sites with evidence of sugar production: Lower Ḥorbat Manot (Manueth), el‑Kabri (Le Quiebre), Ben ‘Ami (Le Fierge), Nahariyya (La Noie) and ‘Ein Afeq (Doc and Recordane) (fig. 1). Some of these sites are mentioned in historical sources of the Crusader and Ottoman periods, linking them to sugar cultivation and production (Peled 2009). Sugar production in the Fatimid period is not mentioned directly, but a letter from the Cairo Geniza, dated to 1060, mentions that red sugar was purchased by Jewish Egyptian merchants at Achziv (a‑Ziv), near the Mediterranean port of el‑Kabri (Amar 2000, p. 250; Peled 2009, p. 27).1 At two other urban sites, ‘Akko (Acre) and Ẓefat (Safed), the excavations have confirmed the undeniable evidence of production and storage of the ceramic vessels involved in sugar production.

Fig. 1 – Map of the ‘Akko plain with locations of the sites mentioned in the text (A. Shapiro).

3The archaeological evidence for sugar production in the ‘Akko plain dates to the Fatimid period (11th century), continues through the Crusader (12th-13th century) and Mamluk (14th-15th century) periods to the early Ottoman period (late 16th to early 17th century).

4The POMEDOR project intentionally focused on 11th-13th century sugar molds and molasses jars, represented by material coming from Lower Ḥorbat Manot, el‑Kabri and ‘Akko. The selected vessels underwent typo-chronological, chemical and petrographic provenance studies. Other sites and periods mentioned in this paper were excavated and surveyed by the IAA, and the results of previous studies were used for general assumptions and conclusions, enabling us to encompass a larger chronological frame.

5The well-known shape of the sugar molds that were used in the final stage of sugar production (Stern 2001, pp. 279‑281; Politis 2013, p. 468) is usually conical, with a slightly flattened pierced base. Franken, who in the 1970s studied the sugar molds from Tell Abu Gurdan, Jordan, concluded that the rim variations are similar in all phases of the excavation, and that it is impossible to detect a typo-chronological development of the mold (Franken & Kalsbeek 1975, p. 147). However, this assumption was not confirmed by later typological studies (von Wartburg 2014; Stern 2001; Stern et al. 2015) and has been reconsidered in the present research. Another focus of our research concerned the question of the location of the potters’ workshops that fabricated molds and jars for the sugar production sites in the plain of ‘Akko, and that of the organization and evolution of these workshops over time. This was investigated through chemical and petrographic analysis, performed on a selection of sugar vessels.2

The sites, their structures and pottery related to sugar manufacture

El‑Kabri

6El‑Kabri (fig. 1) is located near four springs, whose waters flow into the Ga’aton stream at the point at which it arrives in the plain. The spring water was used for irrigation and powered flour-mills by means of aqueducts or lifting pools (Frankel & Getzov 2012, site 61).

7The site is identified with the Frankish village of Le Quiebre, mentioned in documents of the 13th century as having been sold by Jean of Ibelin to the Teutonic Order. Although the documents do not mention sugar cultivation or production directly, it is probable that the village was bought by the Teutonic Order because of its interest in the growing industry of sugar production (Frankel 1988, pp. 259, 264; Peled 2009, pp. 134, 165; Frankel & Getzov 2012, site 67). The village and the farm (mazra’a) by the name of al‑Kābira are mentioned in the Crusader-Mamluk treaty concluded in 1283 (Khamisy 2014, pp. 89, 93, no. 7, 8; Barag 1979, p. 203, no. 5).

8In 2000 the IAA salvage excavation (Smithline 2004) revealed remains related to sugar production for both the Fatimid and Crusader periods. An oven, within and around which were discovered numerous vessels associated with sugar production (fig. 2 & 3), was dated to the 11th century, based on the associated table and cooking wares – glazed bowls with polychrome splash-painted sgraffito decoration and cooking pots and baking dishes (fig. 2 for LEV734, and fig. 6 for LEV730 and LEV728 respectively in Stern et al., in this volume; Arnon 2008, pp. 44, 270, Type 252f, pp. 46, 298, Type 752g). This oven is possibly the earliest linked to sugar production so far discovered in the Eastern Mediterranean.

Fig. 2 – 11th century Fatimid sugar molds and molasses jars from el‑Kabri (drawings H. Rozen-Tahan, courtesy of Israel Antiquities Authority, layout A. Shapiro). Lyon laboratory id. are indicated.

Fig. 3 – 11th century Fatimid sugar molds and molasses jars from el‑Kabri (photos S.Y. Waksman, layout A. Shapiro). Lyon laboratory id. are indicated.

9The most common sugar molds of the 11th century are wide-conical, with a simple, slightly thickened or outward-folded rim, thin walls and a wide rounded convex base with three small holes drilled after the vessel was fired in the kiln (fig. 2: 1‑5; fig. 3: 1‑3, 6‑7). The less common molds are 1/3 lower than the predominant form (fig. 2: 6‑7; fig. 3: 4‑5), or they are narrower, with a concave base and four small square holes pierced before the vessel was fired (fig. 2: 11).3 Among these, one example lacks the holes (fig. 2: 6; fig. 3: 4), and another, of the second type, is deformed, suggesting kiln waste. This may indicate production of the vessels on the site. The molasses jars of the period are semi-ovoid, low and wide (fig. 2: 12‑19; fig. 3: 11‑18), have an out-turned rim and a simple concave base.

10Twenty-five meters south of the oven, the remains of a Crusader-period sugar refinery were unearthed. The massive ashlar building with a vaulted opening and a long ditch in front of its façade, was filled with ashes and many fragments of sugar molds. The stratum apparently dates to the time when the sugar refinery operated.

11The sugar molds from this stratum (fig. 4: 1‑10; fig. 5: 1‑8, 10‑11) were accompanied by material dated to the 12th and/or possibly the early 13th century, including slip-painted and reserved slip bowls, alkaline glazed bowls (fig. 11 for LEV736, fig. 10 for LEV739, and fig. 4 for LEV759 respectively in Stern et al., in this volume), imported Aegean bowls, and undecorated handmade ware (Stern 2012, pp. 44‑48, 65‑69, fig. 4.21: 6‑10, 4.20: 4‑7, 4.79: 1‑4, 4.49: 5, 6 and 4.26: 1 respectively).

Fig. 4 – 12th century Crusader sugar molds and molasses jars from el‑Kabri (drawings H. Rozen-Tahan, courtesy of Israel Antiquities Authority, layout A. Shapiro). Lyon laboratory id. are indicated.

Fig. 5 – Above: 12th century Crusader sugar molds and molasses jars from el‑Kabri; below: 13th century Crusader sugar molds and molasses jars from el‑Kabri, ‘Akko and Lower Ḥorbat Manot (photos S.Y. Waksman, layout A. Shapiro). Lyon laboratory id. are indicated.

12Dated to the 12th century, the sugar molds are conical, with a simple rim, thick straight walls and a slightly rounded bottom with one large hole made when on the potter’s wheel (fig. 4: 9‑10; fig. 5: 10‑11). Several vessels have an incised wavy – or straight – combed decoration beneath the exterior of the rim (fig. 4: 1‑2; fig. 5: 5‑6). Similar sugar molds were found in contexts securely dated to the 12th century at Migdal (Abu ‘Uqsa 2005, fig. 3: 18) and Tiberias (Stern 2018).

13The contemporary molasses jars are ovoid, with an outward-folded rim (fig. 4: 11‑13; fig. 5: 9, 12‑13) and a narrow concave base (fig. 9: 2). They are taller than those of the 11th century and their openings are narrower.

14A small group of molds with thin walls and an outward-folded rim was dated to the 13th century (fig. 6: 6‑8, 13; fig. 5: 26, 28‑29), based on the stratigraphic evidence, typological associations, and similarity to the finds from Lower Ḥorbat Manot (see below). The molasses jars of this period appear in various forms, but are usually low and wide, ovoid or sack-shaped, with a thickened rounded rim, a narrow opening (fig. 6: 14‑16; fig. 5: 23‑25) and a flattened and slightly concave narrow base (Stern 2012, p. 32, fig. 4.11: 4).

Fig. 6 – 13th century Crusader sugar molds and molasses jars from el‑Kabri, Lower Ḥorbat Manot and ‘Akko (drawings H. Rozen-Tahan, courtesy of Israel Antiquities Authority, layout A. Shapiro). Lyon laboratory id. are indicated.

Lower Ḥorbat Manot

15Lower Ḥorbat Manot (fig. 1) is situated close to the K’ziv stream, on the bedrock of the western slope of the Galilean hills. At the site are a large well-preserved building with two rectangular halls, an aqueduct, and the base of a screw oil press hewn in the bedrock (Stern 2001).

16The site is mentioned in several Crusader-period documents, mainly in connection with land transactions. One document dated to 1169 mentions sugar production, while others, dated to the early 13th century, mention that Manuet was sold to the Hospitaller order in the early 13th century (Ellenblum 1998, pp. 199‑203; Peled 2009, pp. 108, 115, 125‑129). The 1238 treaty between the Franks of ‘Akko and the Mamluk sultan lists the village and the farm (mazra’a) of al‑Manawāth as sites that remained in Frankish hands (Khamisy 2014, pp. 88, 93, no. 5‑6; Barag 1979, p. 203, no. 4). It is likely that this farm was, in fact, the sugar production site of Lower Ḥorbat Manot. And the village and the farm by the name of Manawat mentioned in the 16th century Ottoman tax documents’ register, are echoing the two sites mentioned in the 1283 treaty (Rhode 1979, pp. 39, 106, 149).

17In 1996-1997, the IAA survey identified the site as the farm and sugar refinery of Manueth, based on the presence of large quantities of sugar molds and the remains of a large building (Frankel & Getzov 1997, p. 106*).

18A very limited salvage excavation (Stern 2001) had exposed in 1995 the western part of this building. Comparison to the excavations in Kouklia and Kolossi in Cyprus (von Wartburg 1995; 2001; Solomidou-Ieronymidou 2007) enabled the reconstruction of the sugar production at the site (Stern 2001, p. 278, plan 1). The K’ziv stream fed the aqueduct leading to the mill, established in the smaller hall of the building. The larger hall housed the refinery, where the juice, extracted from the cane, was boiled in copper cauldrons on a hearth above fire chambers, two of which were visible before the excavation, the third unearthed during it. In front of the main entrance to the large hall a beaten limestone floor was found with numerous fragments of sugar molds – possibly a place where the molds settled on the molasses jars were placed during the crystallization, and tapped to remove the sugar loaves when these were ready (Stern 2001, p. 282, fig. 5). Typologically similar to the contemporary vessels from el‑Kabri (see above), the recovered pottery assemblage (fig. 6: 1, 3‑5, 9, 11‑12, 14‑16; fig. 5: 14, 16‑18, 20‑25) provided secure dating of the earliest sugar molds from Ḥorbat Manot to the 13th century (fig. 9: 3; Stern 2001, pp. 282‑287, fig. 6‑7).

19At about 25 meters to the east of the building, carved into the bedrock, a Byzantine winepress was surveyed (Stern 2001, p. 278, plan 1). This installation was adapted to a screw press and reused for pressing the milled sugar cane. This screw press, called “the Manot Press”, is unique among the eleven similar installations identified in the western Galilee. It was suggested that this type was introduced by the Frankish settlers and should be dated to the Crusader period (Frankel & Stern 1996).

20The good stratigraphic sequence contributed to defining and dating the Mamluk-period sugar vessels (Stern 2001, pp. 289‑292, fig. 12‑13). The capacity of the molds increases from 3.5‑3.8 to 4.0‑4.5 liters. They have a thick folded rim, thick walls and a slightly rounded or flat tip base with a single hole similar to that of 13th century vessels. The molasses jars are ovoid, with an out-turned or thickened rim and a simple convex base; like the molds, they are larger than the Crusader jars.

21The early Ottoman (16th to early 17th century) phase distinguished at the site was dated according to the pottery and the absence of tobacco-smoking pipes (Stern 2001, pp. 297‑298, fig. 18). The sugar molds of this period are smaller than their Mamluk predecessors, with a thick folded rim, and are usually ridged. The Ottoman molasses jars are somewhat similar to the Mamluk ones, semi-ovoid, with a thickened rim, but higher and narrower. No bases are preserved.

Ben ‘Ami

22Ben ‘Ami (fig. 1) is situated to the west of el‑Kabri, on the continuation of the Ga’aton stream. It is detached from the hills, and the nearby ‘Ein Shefa spring apparently powered the mills.

23The archeological site is identified as a Frankish village mentioned in mid‑12th to 13th century documents under similar names: Fiergio, Ferge, Phergia, Fiergis or Le Fierge. The document related to the hospital of the Teutonic Order and dated to the year 1198, refers to the sugar production in the village (Frankel 1988, p. 264; Ellenblum 1998, p. 67; Peled 2009, pp. 119, 134, 165; Frankel & Getzov 2012, site 90). The village and the farm were registered as one of those that remained in Frankish hands in the Crusader-Mamluk treaty of 1283 (Khamisy 2014, pp. 90‑91, no. 31‑32; Barag 1979, pp. 204‑205, no. 24). The name appears later, in the 16th century Ottoman tax-list, in reference to four mills, for which sugar production was not mentioned. However, it is possible that the summer crops included sugar cane and that some of the mills were used for crushing it (Hütteroth & Abdulfattah 1977, p. 193).

24In 1999, an IAA salvage excavation exposed a few remains of the 11th century ashlar building, and two mud-brick ovens accompanied by fragments of sugar molds (Damati 2011, pp. 144‑146, fig. 12‑13). The great similarity of the oven to those excavated at Kouklia (von Wartburg 1995; 2001; Solomidou-Ieronymidou 2007) suggests that this is the sugar refinery mentioned in the historical sources. The Mamluk stratum included a well-constructed building, and the remains of the Ottoman period include a stone pavement with mud bricks and sugar molds (Damati 2011, pp. 143‑144, fig. 8‑9).

25The 2007 salvage excavation revealed an accumulation of black soil containing numerous oven bricks, sugar molds and other domestic vessels dated to the Fatimid, Crusader, Mamluk and early Ottoman periods (Getzov et al. 2016). Because of the poor stratigraphic sequence, the pottery assemblage was classified according to the typology established at el‑Kabri and Lower Ḥorbat Manot (Damati 2011, pp. 147‑156, fig. 16, 18, 20, 22: 7‑12).

Nahariya

26Nahariya (fig. 1) is located within the present city of Nahariya, on top of Giv’at Ussishkin – a low kurkar4 ridge to the south of the Ga’aton stream.

27In 2013, the IAA salvage excavation (Lerer 2014) unearthed the scanty remains of a wide wall and ceramic finds, including sugar molds, which were securely dated to the 12th century, in concordance with the finds from el‑Kabri. Thereby, it was suggested to recognize the excavation site as La Noie, mentioned in Guérin’s description of the Galilee (Guérin 1880, pp. 24‑25), which had previously been identified to the south along the same kurkar ridge (Frankel & Getzov 2012, site 9). The Crusader village of La Noie is mentioned in numerous documents from the 12th century onwards, often in the context of sugar production (Frankel 1988, pp. 267‑268; Ellenblum 1998, p. 176; Peled 2009, pp. 112‑115; Frankel & Getzov 2012, site 6). The village and the farm are also mentioned in the Crusader-Mamluk treaty concluded in 1283 (Khamisy 2014, pp. 90, 93, no. 29, 30; Barag 1979, p. 204, no. 23). The involvement of the Teutonic and the Hospitaller Orders in sugar production at La Noie is referred to in the historical sources more often than for any other site in the ‘Akko plain, indicating that this must have been a major sugar production site (Peled 2009, pp. 124‑135, 269‑270). However, this stands in contrast to the scarcity of the archaeological finds. Due to its position on the ridge, above the Ga’aton stream, we believe that Lerer’s excavation exposed the remains of the village La Noie, that the sugar refinery existed closer to the stream, and that its remains were not identified during construction in the 1950’s.

‘Ein Afeq

28‘Ein Afeq (fig. 1) is situated in the swampy area created by dozens of springs that feed the source of the Na’aman stream and irrigate the nearby fields.

29Numerous documents from the early 13th century to 1265 mention the legal dispute between the Templars and the Hospitallers on the rights to the water that powered the mills of Doc and Recordane, which at some point required papal intervention. They also mention aqueducts that conducted water to the Hospitaller’s sugar cane plantation, indicating that at least one of the mills was for crushing sugar cane (Frankel 1988, p. 261; Ellenblum 1998, pp. 206‑208; Shaked 2000, pp. 62‑66; Boas 2006, pp. 83‑84; Peled 2009, p. 130). Both Doc and Recordane are included in the 1283 treaty between the Crusaders and the Mamluks, and mentioned as having a village and a mill: Daʿūq and Kurdāna (Khamisy 2014, p. 91, no. 35‑36, 37‑38; Barag 1979, p. 205, no. 27 and 28, respectively).

30In 1990, the archaeological survey revealed that the mill today called Recordane is the Templars’ Doc, with an architectural phase dating it to the Crusader period. The location of the Hospitaller Recordane is some 370 meters to the south, in the area that is now partly flooded (Shaked 2000, pp. 66‑71; Boas 2006, pp. 83, 85).

‘Akko

31‘Akko (fig. 1) was as an important seaport beginning in the 11th century, which in the 13th century became the capital of the Latin Kingdom of Jerusalem.

32Although there is not any archaeological or written evidence that a sugar refinery operated there, large quantities of sugar molds were unearthed at three of the excavations within the borders of the Crusader city (by E. Stern in 1997, A. Tatcher in 1998, and A. Boas in 1999 and 2000).

33The excavation conducted by A. Tatcher discovered a 12th century building, a well and a shallow pool, accompanied by dozens of accordingly dated sugar molds (Tatcher 2000; Stern 1999b, pp. 76‑77, fig. 52).

34The excavations conducted by A. Boas exposed public and private buildings and industrial installations of the 13th century Teutonic quarter of the Crusader Acre, in one of which a large quantity of 13th century sugar molds was found (Boas 2006, pp. 61‑63).

35During the 1997 excavation season at the Hospitaller compound (fig. 7: 1), conducted by E. Stern (Stern 1999a, pp. 13*-14*, 16), hundreds of shattered sugar molds and molasses jars were unearthed (Stern 2012, pp. 3, 21, pl. 3.1). They were found in two adjacent halls, under a thick layer of burnt debris and stone blocks from collapsed vaults (fig. 7: 2, 3). The molds were stacked rim-to-base and laid on the floor in rows. On top of each two rows, another row was placed. The molasses jars were stored in same way. Separating the vessels from the floor to prevent breakage there appears to have been a straw padding, which was burnt or subsequently charred. Only one third of each hall was excavated – seven rows of sugar molds (about 80 specimens in a row) in one of the halls, and molds and molasses jars in the other (Stern 1999b, fig. 51), a total estimated to about 6,500 vessels.

Fig. 7 – (1) View of the excavations of the Hospitaller compound (‘Akko, 1992); (2) Stack of sugar molds in Hall 7 of the Hospitaller compound (‘Akko, 1997); (3) Sugar mold standing on top of the molasses jar (‘Akko, 1997); (4) View of the Mamluk citadel (Ẓefat, 2002); (5) Waster from a Mamluk sugar mold (Ẓefat, 2003); (6) Pile of locally manufactured Mamluk bowls (Ẓefat, 2003) (photos H. Smithline (‘Akko) and N. Getzov (Ẓefat), courtesy of Israel Antiquities Authority).

36The destruction layer is dated to the conquest of ‘Akko by Ashraf el‑Halil in the year 1291 (Stern 2008, pp. 207‑209, fig. 21.3‑4); as expected the sugar molds were dated to the 13th century (fig. 6: 2, 10; fig. 5: 15, 19) and are of same type as contemporary forms from Lower Ḥorbat Manot (fig. 6: 1, 3‑5, 9, 11‑12, 14‑16; fig. 5: 14, 16‑18, 20‑25) and Ben-’Ami (Getzov et al. 2016).

Typological features of the sugar wares

37The study of the typo-chronological development of the conical sugar molds provides clues for understanding the technological evolution of this industry, and enables dating of the archaeological strata, particularly when no other dating material is present. We present here the typological features of the main shapes.5

38The chronological span begins with the 11th century Fatimid period sugar molds (fig. 8: 1), all conical, with a simple, slightly thickened or folded rim and thin walls. The base, either wide and concave or narrower and convex, presents three to four small holes that were drilled after the vessel was fired in the kiln or pierced before firing. Vessels belonging to this phase were found at el‑Kabri, where they were well dated, and at Ben ‘Ami. The Fatimid period sugar molds may reflect the beginning of the sugar industry, when the potters worked in the traditional manner, with the intention of finding the “correct” shape of the vessel for evaporating the thick syrup.

39The 12th century sugar molds of the Early Crusader period (the first Latin Kingdom of Jerusalem) (fig. 8: 2) are very wide-conical, with a simple rim, thick straight walls and a slightly rounded bottom, one hole being made while the mold was formed on the potter’s wheel.

40The 13th century sugar molds of the Late Crusader period (the second Latin Kingdom of Jerusalem) (fig. 8: 3) are conical like those of the 11th century, have a folded or thickened rim, walls that are slightly thinner than those of 12th century molds, and a base similar to those of the latter. The Crusader period (both 12th and 13th centuries) sugar molds present features of average consumer goods – fast mass production, minimal quality and maximal quantity, reflecting the progressive development of the cane sugar industry.

41The 14th-15th century, Mamluk period sugar molds (fig. 8: 4) show further modification. The vessels increase in size from an average capacity of 3.8 liters to 4‑4.5 liters. They are wide-conical with a massive out-folded rim, thick walls and a single hole in a rounded or flat tip base. The fabric of the molds is coarse and the surfaces are not smoothed, which may reflect the boom in the mass production of sugar.

42The 16th to early 17th century sugar molds of the early Ottoman period (fig. 8: 5) are small and coarse, and their folded rims are thick and usually ridged. This change appears to reflect the decline of the sugar industry in the region during this period.

Fig. 8 – The main shapes of the typo-chronological development of sugar molds and molasses jars from the 11th to the early 17th century (drawings H. Rozen-Tahan, courtesy of Israel Antiquities Authority, layout A. Shapiro).

43The main typological changes of the sugar molds over time are the following:

• the thickness of the wall changes from about 5‑7 mm for the Fatimid samples to 11‑14 mm for the Crusader ones, in which the thin walls are often accompanied by a folded or thickened rim, which reinforces the vessel;6
• the size of the vessel, which increases from the Fatimid to the Mamluk period (see above), and decreases during the Ottoman period;
• the development, that occurred in the plain of ‘Akko between the Fatimid and the Crusader periods, from several small holes in the bottom (fig. 2: 1‑2; fig. 3: 1‑2) to one large hole (fig. 4: 9‑10; fig. 5: 10‑11, 14‑15; fig. 6: 1‑2, 4‑5).

44The latter feature would require further examination. Sugar molds with one hole were known as the only type in use in the southern Levant (modern-day Israel, Palestinian territories and Jordan) and Cyprus (see in Taha 2009; Abu Dalu 1995; von Wartburg 2001; Solomidou-Ieronymidou 2007; Stern 1999b). However, sugar molds with three holes, dated to the end of the 12th and early 13th centuries, were found at a sugar refinery that was excavated at Susa, Iran (Kervran 1979, p. 180, fig. 64: 1‑3, pl. XVII: c). In addition, Al‑Nuwairi (1279-1333), who served in various state offices during the regimes of al‑Malik and an-Nasir, wrote in the early 14th century an encyclopedia which includes a description of sugar production processes in Egypt. He states that at the bottom of each earthenware mold (abulg), into which the boiled juice is poured, are three holes, which are plugged with pieces of sugar cane (Deerr 1949, p. 91). This information clearly indicates that the use of three holes in a sugar mold did not cease at the beginning of the 12th century, and that the process of sugar production developed differently at the various locations. It is possible that the Franks, when they arrived in the southern Levant in the 12th century, learned to operate the local sugar industry, which they developed, introducing changes that included the model of a mold with a single hole. At the same time, in the Muslim regions (Egypt and Persia) the use of three holes continued well into the 14th century. Sugar molds with one hole were used also in Cyprus from the 13th century onwards, as the production of sugar was developed there by Europeans (von Wartburg 2001; Solomidou-Ieronymidou 2007), especially after the fall of the Latin Kingdom of Jerusalem.

45The typo-chronological changes listed above represent the ceramic vessels used in the crystallization stage of sugar production in one restricted geographical region – the plain of ‘Akko, between the 11th century and the early 17th century. The development of the cane sugar industry appears to be well reflected in the size of the sugar molds, which continuously increased from the Fatimid period up to the Mamluk period. At the end of this period, sugar production became degraded, which is mentioned in the written sources (Ashtor 1977) and indicated in the decrease in size of the Ottoman vessels.

Provenance studies

46Within the framework of the POMEDOR project, chemical and petrographic analyses were carried out on sugar molds and molasses jars dated to between the 11th and the 13th century from el‑Kabri, Lower Ḥorbat Manot and ‘Akko (tabl. 1). The results are presented below, in order beginning with the earliest samples, together with the outcome of previous petrographic analyses (Shapiro 2001; 2012). Chemical analysis by X-ray fluorescence was carried out in Lyon (CNRS UMR 5138), and hierarchical clustering analysis was used in the interpretation of the chemical data (fig. 9, tabl. 2, see e.g. Waksman 2011 for details). Chemical analyses of samples of other categories of pottery manufactured in ‘Akko and its vicinity (see Stern et al., in this volume) were also taken into account.

Fig. 9 – Classification according to the chemical compositions of samples of Fatimid and Crusader sugar wares from Acre, el‑Kabri, Lower Ḥorbat Manot, and of other Crusader wares of the Acre region (based on the concentrations of 17 elements, see Waksman 2011).

Tabl. 2 – Chemical compositions of samples of Fatimid and Crusader sugar wares from Acre, el‑Kabri, Lower Ḥorbat Manot, and of other Crusader wares of the Acre region, ranked as in the classification fig. 9.
Major and minor elements are indicated in oxides weight %, trace elements in ppm (part per million). Elements between brackets were not taken into account in the classification.

The Fatimid period

47The Fatimid period (11th century) samples form two main petrographic and chemical groups (tabl. 1 and 2, fig. 9), which do not appear to be correlated with miscellaneous typological features.

Group 1, ‘Akko Hamra (fig. 10: 1)

48The nine samples attributed to this group are characterized by a ferruginous, sometimes slightly calcareous matrix with some quartz silt. Non-plastic material comprises 15‑17% of the sherd’s volume. This is a coastal sand, with particles of 0.2‑0.3 mm, in which quartz is dominant; shell debris and chert are present in lower quantities than quartz; limestone, lumps of ferruginous clay, heavy minerals and plagioclase are sporadic. Some of the limestone inclusions measure up to 2 mm. The firing temperature is estimated to be 750°C.

49The raw materials most probably used for this petrographic group are Hamra clay and quartz sand. Hamra clay appears as lenses within the kurkar ridges along the northern Israeli coast, including an area close to the Old City of ‘Akko (about 2 km to the north-east, personal observation), and as a large outcrop at Liman, a modern settlement about 15 km to the north of ‘Akko (personal observation; Sneh 2004); the quartz sand could have been mined from the large dune about 3 km to the southeast of ‘Akko (Karcz & Sneh 2011).

Group 2, ‘Akko to Rosh ha-Niqra (fig. 10: 2)

50The nine samples composing this group are characterized by a slightly calcareous ferruginous matrix with some quartz silt. The dominant non-plastic material is calcareous sand, particle size 0.3‑0.4 mm, including fossil and fresh marine shells and their fragments, and Amphiroa sp. algae (a fingerprint of the northern Israeli and southern Lebanese coast). Quartz is present in smaller quantities than shells. In some of the samples there are large (1.2 mm) rounded grains of kurkar. The estimated firing temperature is the same as for Group 1.

51For this petrographic group, the Hamra clay (see above) was probably admixed with other more calcareous soil, and the quartz-calcareous sand from the coast between ‘Akko and Rosh ha-Niqra was used as temper (Sneh 2004; Karcz & Sneh 2011).

Fig. 10 – Microphotographs of thin sections of sugar molds, polarized light.

52The classification based on chemical data (fig. 9, tabl. 2) indicates at least two different clayey materials within the Fatimid samples of el‑Kabri. One of these corresponds to low calcareous pastes, having fairly high contents in iron, titanium and aluminium. These samples all belong to petrographic Group 1, rich in quartz temper. The other chemical group corresponds to high-calcareous pastes (about 30% CaO), and includes mostly samples of petrographic Group 2, characterized by temper with a high content of calcareous shell material. Chemical analysis appears to distinguish to some extent these calcareous Fatimid sugar wares from those of the Crusader period (fig. 9).

The Early Crusader period

53The Early Crusader period (12th century, 1st Latin Kingdom of Jerusalem) vessels both continue the pottery tradition of the Fatimid period, and include new materials. The main technological difference between the 11th century sugar molds and those dated to the 12th and 13th centuries is that almost all the latter have a well expressed surface and internal vitrification. This implies that sea water was involved in the clay dough preparation, including “washing” the clay and dipping the dry vessel into salt water before firing.

54Of the twenty vessels examined by petrography, four belong to petrographic Group 1, eight to Group 2 (see above), and eight to two new petrographic groups, 3 and 4 (tabl. 1).

Group 3, ‘Akko to Rosh ha-Niqra 2 (fig. 10: 3)

55Three samples are characterized by a ferruginous and calcareous matrix containing plenty of foraminifers and some quartz silt. The non-plastics are rare, 0.2‑0.3 mm rounded quartz grains and fossil shell debris including Amphiroa sp. algae.

56The matrix for this group originated from either Brown red sandy soils covering the kurkar ridges, or more inland Brown Alluvial or Rendzina soils (Ravikovitch 1969). Coastal sand, similar to that for Group 2, was used as tempering material. This combination may indicate that some of the 12th century sugar molds were produced by pottery workshops which used inland soil and coastal sand, and were located on the shore or not far from it (Sneh 2004; Karcz & Sneh 2011).

57Two additional samples from Ben Ami, dated to the 12th century, can be attributed to petrographic Groups 1 and 3 (Getzov et al. 2016).

Group 4, Green vitrification, apparently Galilee

58This group includes five samples, three of which are strongly vitrified (to the green stage of vitrification).7 Another two have a calcareous matrix that is partially vitrified with some quartz silt and tiny ore nodules. The rare non-plastic inclusions are sub-rounded to sub-angular 0.2‑0.3 mm quartz grains and opaque dots of ore mineral. Such a lithology may suggest an inland provenance, but the high stages of the matrix vitrification do not allow better pinpointing.

The Late Crusader period

59The Late Crusader period (13th century, 2nd Latin Kingdom of Jerusalem) vessels from Lower Ḥorbat Manot and ‘Akko (tabl. 1) partly continue the potters’ technologies of the 11th and 12th centuries. Of the eleven samples, three belong to petrographic Group 1, four to Group 2 and the others to Group 3; all have well-expressed internal and surface vitrification, the same as for the 12th century assemblage (Shapiro 2001, pp. 312‑314; 2012, pp. 104‑105).

60Dated to the 13th century, the samples from Ben Ami belong to petrographic Group 3 (Getzov et al. 2016).

61The sugar vessels of the Crusader period examined through chemical analysis present a large range of chemical compositions, particularly in their calcium contents. However, they are included in the large, heterogeneous, group of common wares – especially “Acre bowls” and other “Acre wares” – manufactured during the Crusader period in ‘Akko and its vicinity (fig. 9, tabl. 2, and Stern et al., in this volume). According to petrographic analysis, they may be related to locations around Acre and along the coast to the north. Most of them are attributed to petrographic Groups 1 and 2 (tabl. 1).

The Mamluk and the Ottoman periods

62The Mamluk (14th-15th centuries) and the Ottoman (16th-17th centuries) periods were investigated during previous research, and the results obtained by petrography are briefly presented below to complete the chronological span.

63Of seventeen examined Mamluk samples (8 from Lower Ḥorbat Manot and 9 from Ben Ami), five may be attributed to petrographic Group 3 – inland soil (Brown red sandy soils or Rendzina soils from valleys and Taqiye formations) and coastal sand, but without any vitrification. These five samples could be of coastal origin, but the lack of vitrification may indicate that the pottery workshops were not so close to the coast, and, thus, to an access to seawater. Others are clearly of inland origin, because Rendzina or other inland soil mixed with calcareous inclusions was used as raw material (Shapiro 2001; Getzov et al. 2016). Therefore, for all of the Mamluk vessels, an inland origin may be suggested, even for those containing coastal sand, as the sand could have been transported to a pottery workshop located inland.

64The lithology of the early Ottoman samples of the 16th-17th centuries (two from Lower Ḥorbat Manot and one from Ben Ami) is similar to that of the Mamluk material, which is undoubtedly of inland provenance (Shapiro 2001; Getzov et al. 2016).

Discussion

65Petrographic and chemical analyses provide information on developments in the manufacture of sugar vessels, including the organization of production.

66We still have little evidence for the Fatimid period. Sugar wares found at the site of el‑Kabri were made from two different clayey materials, with no apparent correlation with the different types (fig. 2‑5 and 9). They suggest little standardization as far as the choice of the raw materials is concerned. Besides, at this point of research, there is not enough archeological evidence to determine whether sugar wares were manufactured together with other categories of pottery, thus we cannot conclude that they correspond to a specialized production. The raw materials may be related through petrography to coastal formations present in the vicinity of the site.

67A variety of clayey materials was also in use during the Crusader period, on the three sites studied within the POMEDOR project. However, this corresponds mostly to the range of materials exploited in ‘Akko and its vicinity8 to produce not only sugar wares, but also various common wares, including bowls, jars, kraters… known as “Acre wares” (Stern 1997; 2012; Stern et al., in this volume; Shapiro 2012). The main technological “fingerprint” for all these wares, including the sugar vessels, is that, irrespective of the type or petrographic group, almost all of them were tempered with coastal sand, and present vitrification that is well expressed on the surface and internally, easily recognizable under the petrographic microscope. The latter feature implies that sea water was involved in the clay body preparation, including “washing” the clay and dipping the dry vessel into brine before firing (Shapiro 2012, pp. 104‑105). This process may have bestowed an important functional property upon the sugar molds, as the vitrified (glassy) surface would have made tapping out the sugar loaves much easier. The association of sea water and coastal sand is indicative of a coastal provenance.

68The Mamluk and Ottoman periods show a change in fabric composition. Chalk and limestone are used as temper rather than coastal sand, and the inland clay replaces the coastal one. Sea water was no longer used in the preparation of the clay body. Because the raw materials used correspond to the geological formations present in most areas of the Galilee, it is impossible to accurately define the place of production, but they clearly demonstrate an inland provenance.

69Vessels of a fabric very similar to the Mamluk sugar molds are abundant in the ceramic assemblages from Ẓefat (Dalali-Amos & Getzov 2019). Previous provenance studies defined these as local (Shapiro 2019), and it is possible that many of the Mamluk period sugar vessels were produced in the vicinity of Ẓefat.

70In summary, the petrographic and chemical analyses reveal the transfer of sugar mold fabrication from the coastal area to the mountain region, corresponding to the transition of rule from the Franks to the Mamluks. At the same time, the sugar refineries remained on the coastal plain, where the conditions were appropriate for cane cultivation.

71We would like to propose, that when during the 12th and 13th centuries ‘Akko was a main seaport, then the capital of the second Latin Kingdom of Jerusalem, the pottery workshops which made the sugar molds were concentrated along the seashore.

72After the fall of ‘Akko in 1291, when the Mamluk capital of the region moved to Ẓefat (fig. 1), the sugar ware workshops would have been relocated inland, where they would have remained until the early Ottoman period. This hypothesis is supported by the archaeological evidence.

73As mentioned above, the 13th century sugar molds produced in the ‘Akko vicinity were stored in the city, including more than 6,500 vessels in one of the main halls of the Hospitaller compound (fig. 7: 1‑3), and in the presumed location of the Teutonic quarter. It is likely that they were stored there before the distribution to the sugar refineries in the countryside, this evidence being one of the ways to determine the quantities of the sugar loaves ordered. These finds are confirmed by historical sources (Boas 2006, pp. 93‑94; Peled 2009, pp. 124‑138), thus linking the 13th century Crusader military orders and sugar production.

74During the excavations of the Mamluk stratum in the castle of Ẓefat (Barbé & Damati 2004) (fig. 7: 4), sugar molds and molasses jars were found among the ceramic assemblage, including numerous jars, jugs and bowls (fig. 7: 6), all of similar fabric, defined as local (Stern, forthcoming; Shapiro 2019). One of these sugar molds presents firing marks on a break (fig. 7:5), indicating that it broke while in the kiln and was thus produced at the site. It is not likely that kiln waste would have been brought into the fortress, therefore, we can assume, that the mold, even cracked, still could have been used. As Ẓefat is not located in a geographical region suitable for sugar cane cultivation, it can be proposed, that, as at ‘Akko, the sugar vessels were probably produced in close proximity of the castle, and then distributed to sugar refineries all over the coastal plain.

75Archaeological and archaeometric evidence from Acre and Ẓefat clearly suggest direct control by the regional capital of sugar and sugar wares production in the Crusader and Mamluk periods.

76An additional example for direct governmental involvement in cane sugar production comes from Jordan. In the palace of the governor, in the Mamluk regional capital of Hisban, a storeroom containing numerous molasses jars and molded glazed bowls with Arabic inscriptions was found (Walker 2003, p. 258, fig. 3; 2004, pp. 132‑133; 2010, pp. 145‑149). Like Ẓefat, Hisban is situated in the mountains, at some distance from the Jordan Valley – an area where sugar production flourished under Mamluk rule (Abu Dalu 1995; Strange Burke 2004, pp. 109‑111).

Conclusions

77In this paper, we have presented five sugar production sites excavated and surveyed in the plain of ‘Akko. Within the POMEDOR project, the archaeometric studies were carried out for those sites ranging in date from the Fatimid to the Crusader period; the results of previous research, used for the typo-chronological investigations, discussions and assumptions, have enabled an extension of the time span up to the Early Ottoman period. It appears that there are no similar regions in the Eastern Mediterranean in which sugar refineries are so densely present, and where the corresponding ceramic finds have been studied so intensively.

78The conical ceramic molds, which are the main objects involved in the process of cane sugar production, serve as direct evidence for changes that occurred in this industry over the 600 years of its existence in the plain of ‘Akko.

79The intention of our study was to examine and interpret the data by distinguishing and characterizing the changes over time in the shapes and fabrics of the sugar molds. The archaeological and archaeometric data have provided insight into the organization of the production and distribution of sugar vessels, and could be used as sources of historical information. At least from the Crusader period onwards, the evidence points to a dissociation between the locations of sugar manufacture and those of sugar vessel manufacture. The vessels appear to be directly associated with the successive centers of political power in the region, where they were stored before being redistributed to the sugar refineries. For example, on the two urban sites, Crusader ‘Akko (Acre) and Mamluk Ẓefat (Safed), the evidence from the excavations clearly confirms government control over the sugar industry.

80We have shown here that the displacement of the center of government, from the coast to the mountains, is reflected in the fabric of the sugar molds, and this emphasizes the deep involvement of the central government in sugar production.9

Acknowledgements

81This study was funded by the French National Research Agency (ANR) through the POMEDOR project, and we acknowledge the support of the ANR under reference ANR-12‑CULT-0008. We thank the staff of the analytical facilities, CNRS UMR 5138 Lyon, the Israel Antiquities Authority who kindly provided permission to sample pottery from its excavations, the excavators of the Israel Antiquities Authority Eliezer Stern, Danny Syon, Howard Smithline, and the curators Ayala Lester and Giulia Roccabella.