Shine Like the Sun: Chapter 5: Syria c. 700-1250 INTRODUCTION As in the previous chapter, the pottery discussed in this chapter may be divided into two main groups. This division is based on the technology and to some extent the significance of the pottery concerned. The first group comprises wares with clay bodies and lead glazes. These are less interesting with regard to technology and art-history, and the large number of production centres would require rather more petrographic work than would be fruitful, given the current objectives. Hence, the first group receives less attention than the next group. The second group represents wares with stonepaste bodies, which are usually decorated with lustre, underglaze-paint, or relief decoration. This group is far more important with regard to its technological development and also its art-historical significance; hence, it receives the bulk of attention. For the most part the first group in the "Syrian" region effectively forms a continuum from Egypt to Mesopotamia, with numerous local centres producing coarse wares with splashed lead glazes. These are generally earlier than the stonepaste wares but the slip-incised wares seem to have continued for some time in the region, particularly among the Crusader states. The stonepaste wares are generally later, and their technology appears to have been introduced from Egypt, after which they generally supplant the finer clay-bodied wares. A number of distinctive styles can be identified that are associated with the region of present-day Syria and with the period of the major dynasties that ruled the area, such as the Zangids and Ayyubids (Bosworth 1996:70-75). As such it is really only this second group that makes a distinct "Syrian" chapter meaningful. The typology, technology, and petrography of the few clay-bodied wares available for study will be discussed collectively first, together with a short overview of other Syrian clay-bodied wares. The stonepaste wares will then be discussed in greater depth, with the usual tripartite division into typology, technology, and provenance/characterization, and with a subsequent synthesis. A number of samples of wares which post-date the period covered in this chapter, mostly of Mamluk period (c. 1250-1517; see Bosworth 1996:76-80), will also be considered in the provenance section. CLAY-BODIED WARES There is a reasonable suggestion that certain lead-glazed moulded wares were made in the Syrian region under the Umayyads (Lane 1947:8). If so this would probably represent a continuation of pre-Islamic practice in the region, with well- attested production at sites like Tarsus (Peacock 1982:63). One problem with this is that continuity is less than certain, and at sites such as Antioch, where early Roman and early Islamic wares but no late Roman wares of this type are found (Waagé 1948:81-82, 99), continuity appears completely absent. Production of these wares in Syria in the Islamic period may be supported by fabric descriptions of wares from Antioch, some of which do not appear to be compatible with the Basra Petrofabric, the only body identified in examination of these wares from Iraq, Bahrain, Dhahran, and the Yemen. From published excavation reports a number of styles of wares with colour-splashed lead glazes, often with incised- slip decoration or slip-painted decoration, can be identified. These sites include Antioch (Waagé 1948, figs. 57-77), al- Mina' (Lane 1937), Hama (Riis and Poulsen 1957, figs. 803- 31), Qasr al-Hayr al-Sharqi (Grabar et al. 1978:218-39), and Afamiya/Apamea (Rogers 1972, 1984). Some wares, such as those excavated by Lane (1937) at al-Mina', have reasonable evidence of local manufacture. Judging by the evidence from elsewhere, other wares of this general nature were made in various centres in the Syrian region, while some may even have been imported from Cyprus and the area of Byzantine rule. Further complications will arise with the Crusades and the varieties of wares associated with the Frankish and Italian occupations, including material imported from various parts of the Mediterranean. Under the Crusaders and in association with the Byzantine practices, wares with high-lead glazes slip- incised decoration (sgraffito) continued longer in this region than elsewhere in the Islamic world. In Mamluk Egypt (i.e., after 1250) a revival of the slip-incised technique may be due to its reintroduction from Christian areas. Hence to fully unravel the typology and provenance problems of this group, research at least equal to the entirety of the present study would be required, while the technological significance would be minimal. Two main groups of clay-bodied wares have been available for detailed study. The first is a group of lead-glazed wares from excavations at Qaysariyya/Caesarea, Jerusalem, and the shipwreck of Serce Liman, Turkey (the "Shipwreck" group). Wares of the second group are of various technologies and come from excavations at the site of Tell Aswad at Raqqa (also discussed in Chapter 3). In no way could these two groups be considered representative of the products of the Syrian region as a whole, except perhaps in their typical technology. Results of their study should be presented nonetheless, as some worthwhile conclusions may be inferred from them. The typology, petrology, and what is known of the technology of these wares will be discussed together. "Shipwreck" Group This group of wares came to my attention through the efforts of the excavators of the Serce Liman shipwreck site to find an origin for their finds. These samples plus comparative material from Qaysariyya (Caesarea) were made available, while further comparative material was found in excavated material from Jerusalem (Tushingham 1985), which is stored in the ROM. Typical of the wares in question is a segmental-profile bowl with a simple rim and short, slightly splayed ring-base (see Fig. 5.1). The red body is covered with a pale clay slip, which is then incised through in thin and broad lines to expose the red body, thus creating the decoration; the vessel is then covered with a thin lead glaze (a number have been published by Jenkins 1992). The Serce Liman wreck contains coins in a date sequence culminating in 1024, while the Qaysariyya finds are considered to date to the Crusader period (i.e., after c. 1090 [Pringle 1985]). The Serce Liman date of c. 1024, or shortly thereafter, must be considered to be definitive for the finds from that site. Examples from Qaysariyya may be later than this, as a number imitate the flat-rimmed styles with quadripartite decoration common in Lustre-painted and associated wares in Egypt (see Chapter 4) and Syria (see below) in the twelfth century. All the samples of the above description from Serce Liman, Qaysariyya and Jerusalem are of the same petrofabric, called for the present the "Shipwreck" group. This consists of 10-15% rounded fine quartz sand, 5-7% subangular carbonates, 3-4% opaques, and 1-2% chert. The internal structure of the carbonate is often destroyed by the firing, but for the most part it is probably a finely crystalline limestone. Traces of mollusc shell, including gastropoda and bivalvia up to 0.5 mm in diameter, can also be identified. These shell remains are independent within the clay matrix, and possibly therefore represent organisms deposited in the clay rather than as fossils within the limestone. An unusual feature of this pottery is the slip, which is usually less than 0.1 mm thick, and is composed of a light-coloured clay, with inclusions of rounded quartz and chert actually larger than the thickness of the slip. Evidence of ceramic manufacture exists for Qaysariyya/Caesarea, but it is considerably earlier than the glazed wares we are concerned with, being Byzantine and early Islamic of the seventh to ninth centuries AD. The petrofabrics of these early samples show considerable variation, but indicate utilization of the same raw materials. Similar to the Serce Liman and eleventh-century Qaysariyya samples, this earlier petrofabric consists of rounded sand and carbonates. However, the rounded sand includes grains of feldspars, which are not found in the Serce Liman group, and usually lack chert, while the grainsize distribution also differs. The carbonate also differs, being fine micritic limestone with small gastropod fossils, and is present as large angular grains in the Byzantine samples, and in trace amounts in the early Islamic samples. Hence, although the petrofabrics appear superficially similar it cannot be accepted that these are the same raw materials used for the eleventh-century pottery. However, given the similarities, and given the terrestrial distribution focus for the glazed wares in Qaysariyya and Jerusalem, a provenance in this region for the "Shipwreck" group seems tenable. Tell Aswad Group Some wares from Tell Aswad have already been discussed in Chapter 3, as the pottery reflects Iraqi traditions as well as those of Syria, and is within the Tigris-Euphrates basin. The petrography is more fully discussed in Chapter 3 and also Appendix B, and so will not be probed to any depth here. Pottery known to have been produced at this site includes large storage vessels, unglazed moulded vessels, and lead- glazed wares, but no actual wasters were included for analysis. One of these petrofabrics ("Tell Aswad 4") is the largest group sampled, and includes all the representatives of a type of ware with a yellow glaze, splash-painted with green, including one that was formed by moulding (RQA.01—the "mustard and cress" glazed type known at Afamiya/Apamea [Rogers 1984]), other wares painted with a thin slip, lead- glazed wares, and the only sampled example of a large unglazed vessel (RQA.16). All of these are of what might be called a Syrian derivation. This petrofabric essentially represents Euphrates sediments. Two other petrofabrics represent material from a source more in keeping with the basalt-dominated local terrain, one of which ("Tell Aswad 5") again represents "Syrian" lead-glazed wares and unglazed moulded water jars, while the other ("Tell Aswad 3") is represented solely by opacified wares that show close stylistic similarities to the contemporary Basra wares. This diversity of petrofabrics would support the hypothesis of separate workshops within the site or possibly the region, all with different potting traditions, and each using separate sources of raw materials. Were each group represented solely by different vessels with different functionsthen it might have been that different clays were considered suitable for different functions, but this is not the case. All include glazed bowls, and some include various unglazed water and storage vessels. If we stretch a reasonable hypothesis to the limit, it is possible to surmise that the workshop of "Tell Aswad 5" represents a workshop reflecting local traditions, perhaps situated on alluvium deposited by the River Balikh or elsewhere off the Euphrates flood plain. Then in the later part of the eighth century a new workshop was founded by potters from Basra, probably somewhere in the same vicinity, but not with access to the same raw materials. This hypothesized centre continued into the ninth century, accompanied by imports from southern Iraq. It might be argued that "Tell Aswad 4" represents the ninth-century products of the region, made in a workshop either situated or utilizing raw materials located nearer the Euphrates. The products of this workshop would seem to imitate the wares of southern Iraq rather than to represent a continuing tradition, and so may represent potters from the workshops of "Tell Aswad 3" and/or "Tell Aswad 5" starting anew. Clay-bodied wares: general discussion As explained above, this is only the beginning of analyses of Syrian clay-bodied pottery based on the methodologies used in this study, but it does provide a few insights. First, on a review of the published excavated pottery, it is clear that the "Shipwreck" group is not widespread beyond Palestine. Each region probably produced its own coarse wares of this type, predominantly for local consumption, although it is significant that the Serce Liman wares were off the coast of Anatolia when their ship wrecked. The Tell Aswad wares certainly show the status of the Raqqa region to be intermediate between the Syrian sphere and the Iraqi sphere. The possibility that these potteries were set up to service the palace and its dependent settlement founded by the caliph Harun al- Rashid at this time would go some way to explaining how Iraqi techniques and technology came into what may have been an area more conformable to Syrian traditions. STONEPASTE WARES Available materials for the study of Syrian stonepaste wares (see Appendix A) include a number of vessels with unknown origins from the ROM, the Ashmolean Museum, and the MMA. Two published profiles of pieces from the Khalili collection were also included for an earlier version of this study (Mason 1997a), as full profiles of this style were not otherwise available until the MMA material was included for the present volume (publication of the Khalili collection includes profiles of a number of Syrian wares, but some show internal evidence of inaccuracy, while they are generally published at too small a scale to be accurately copied, so incorporation of these otherwise useful drawings was kept to a minimum (cf. Mason 1997e). Pottery from alleged locations includes a large number of sherds from the area of Ma‘arrat al-Nu‘man, including a diversity of types, and pottery said to be from Raqqa. Finds from more reliable find spots comprise pottery from recent excavations at Raqqa, Fustat, Jerusalem, Abu Sudeirah in Iraq, Dhiban in Jordan, Ani in Turkey, the ancient site of Nineveh in Iraq, and several sites in northwest Europe. Published wares incorporated into the typology include those from Hama (Riis and Poulsen 1957) and a number of pieces said to be collected in the Tell Minis area (Porter and Watson 1987). Reference will also be made to various other cited publications of Syrian material. STONEPASTE WARES: TYPOLOGY As within the study as a whole (see Chapter 2), the pottery has been classified in a hierarchical taxonomic structure of Class, Type, and Style. The Syrian wares discussed here are all of stonepaste bodies, and may have either an alkali-lead glaze or an alkali glaze. These may effectively be considered to represent two classes, although distinction may be difficult without technical examination. Both of these two Syrian classes include the Lustre-painted type and also the Monochrome-relief group (especially Incised-wares), which is considered to represent a class unto itself in Egypt and Iran (see Chapters 3 and 6). The Underglaze-painted type is found only in alkali glazes, and the Polychrome-relief type (also known as "Laqabi" ware) is found only in the alkali-lead class. Generally this divsion of the Syrian wares into two classes based on glaze chemsitry does not seem particularly to aid study, and it is normally necessary to be cognizant only of the type and style. The types are the above-mentioned Lustre- wares, Polychrome-relief wares, Monochrome-relief wares (Incised and Moulded types), and Underglaze-painted wares (Figs. 5.2, 5.3). The style groups, defined by motif assemblage and form, are presently considered as an entire sequence for all Syrian stonepaste-bodied wares, numbered One to Seven. In the other components of this study the incidence of attributes is tabulated and subjected to seriation. The Syrian material has not, however, been ordered into a seriated sequence, Table 5.1 being used only to relate motifs (Fig. 5.4, labelled the "SS" sequence for Syrian stonepaste), forms (Figs. 5.5, 5.6, 5.7, 5.8, 5.9), pigments, and petrofabrics to the type sequence. The inability to seriate this data at this time is due to a number of factors. In some cases there appears to be few repetitive motifs, insufficient for statistically valid seriation; in other cases those repetitive motifs that do occur do not seem to form links between groups that might suggest chronologically sequential development; and finally, any such arrangement that does exist would be obscured by the large number of unrelated and contemporary groups. The variability of the data may be partly due to the fact that we are dealing with a number of production centres (see below). Following further work on provenance, it might be possible to better define the stylistic attributes of each site and thereby produce a genuine sequence specific for the separate centres. The general methodological conclusions that may be derived from the successful seriation elsewhere in the larger study have been applied to the Syrian wares. For instance, motifs on the reverse and rims of vessels have formed meaningful groups. Forms seem to reflect common approaches not only within Syria, but also in Egypt and Iran, allowing correlation of the different sequences by forms as well as by motifs. Apart from correlation with other sequences, it is also possible to note developments of motifs that reflect a progressive development from one motif to another (see below). Dating of the Syrian sequence is done by reference to archaeological sources and to the sequences developed for Egyptian and Iranian wares (see Chapters 4 and 6, respectively). The only Syrian piece with a datable inscription was described, but not fully published, by Sauvaget (1948). This piece apparently bears the name of al-Malik al-Mujahid Shirkuh II, ruler of Homs between 1186 and 1236. Concrete information from excavation is similarly scant, despite a number of major projects in the area, such as al-Mina' (Lane 1937), Antioch (Waagé 1948), Hama (Riis and Poulsen 1957), and Qasr al-Hayr al-Sharqi (Grabar et al. 1978). At best these excavations have made general suggestions of one ware being relatively "early" and another being relatively "late," with no hard evidence to support even this attribution. A more promising publication of an excavated site is that for the Armenian Garden in the southwest corner of Jerusalem (Tushingham 1985), while further evidence is available from a number of other cited sites. Results of petrographic analysis are more fully reported below, but it is pertinent here to say that petrographic analysis has revealed five major petrofabrics attributable to Syria. These are the "Tell-Minis" Petrofabric, of presently unknown origin but found in the early Lustre-ware style; the "Ma‘arrat" Petrofabric, named after a collection of wasters donated to the Ashmolean Museum (Bartels gift), which the dealer attributed to "Tell Minis near Ma‘arrat al-Nu‘man" (named after the larger adjacent town with quotation marks to indicate the attribution based on the dealer's information); Raqqa-1 Petrofabric attributable to Raqqa on the basis of analysis of wasters purchased by the ROM in the 1920s and "said to be from" Raqqa by the dealer; Raqqa-2 Petrofabric, which is more confidently ascribed to Raqqa on the basis of analysis of wasters excavated by recent expeditions; and finally the Damascus(?) Petrofabric, attributed to that site by inclusion within it of all fourteenth to sixteenth century samples, a period for which there appears to be evidence for production of glazed wares at only one centre: Damascus (see below). The Raqqa-1 Petrofabric remains problematic, as it was not found in any of the wasters sampled from the recent excavations, and so may actually be from a completely different site. It might be suggested that the "true" Raqqa petrofabric (Raqqa-2) would be better so named, and that another name be given to the Raqqa-1 Petrofabric, but as previous publications have labelled the Raqqa-1 Petrofabric as the Raqqa Petrofabric (Mason 1991, press "a," 1994), it was considered preferable to maintain some sort of continuity. Other presently unknown centres may exist, for instance, a small collection of wasters said to be from Halabiya (actually one waster of fragments from three monochrome vessels) has also been found to have a further petrofabric, while a Lustre- ware piece from Nineveh is of yet another (Mason 1995b). However, all decorated pieces of more typical typology have been found to be of one of the major petrofabrics. There also appears to be evidence of stonepaste-bodied pottery production at Qal‘at Ja‘bar (Tonghini 1995, 1996), while there are rumours of such production at Balis/Maskana, but no evidence has been published (e.g., Perez-Arantegui et al. 1995; Raymond and Paillet 1995), and it may remain questionable whether the alleged evidence is relevant to stonepaste pottery production. Production at numerous sites on the Euphrates may make unique characterization for the separate sites difficult or impossible, although it does appear possible on present evidence to separate Raqqa and Halabiya. It is always possible that the Raqqa-1 Petrofabric and conceivably even the "Ma‘arrat" Petrofabric represent material from sites such as these. Although not in Syria, the exhibited material at Konya of wasters of pottery are of styles that could be logically used for subsequent decoration only with lustre-pigment, most likely in a typical late "Raqqa" style. In Appendix A may be found attributions of each piece sampled for this part of the study. Here it may be recognized that a great deal of the analysis has been undertaken of wares of unknown origin and of wares with not entirely satisfactory attributions or from sites outside of Syria. This may introduce some biases, for instance, the fact that an overwhelming majority of Underglaze-painted wares are attributed to Damascus may be partly due to the reliance on sites that would have been dominated by trade with Damascus (all sampled Syrian wares from Europe, Egypt, and Turkey are Damascus, as are ten out of eleven pieces from Jerusalem). Syrian Stonepaste-bodied Group One (SSB1)—"Tell Minis" This group includes the "Tell Minis" Lustre-painted style, at least some of the Polychrome-relief or "Laqabi" ware, and other monochrome Incised wares. Study of this group has been aided by a number of profile drawings from the earlier study by Porter and Watson (1987) on "Tell Minis," although these were not published to a uniform scale (they have been rescaled in Fig. 5.5). The drawings include pieces from the original Tell Minis find (see provenance section, below) and from the Hama excavations (TLM and HMA nos., respectively, in Fig. 5.5). Forms in this group include a number of conical bowls (e.g., Fig. 5.5: ROM.14). Bowls of similar form include those with slightly recurved rims (e.g., TMN.01), and bowls with sides that are slightly concave on the interior (e.g., TMN.02). A further form possibly related to the latter has a carination, and a rim that may be recurved or flattened (e.g., ROM.15). Porter and Watson (1987) also include in this group a large bowl with a truncated ellipse profile from the David Collection (DVD.02). Probably one of the later forms is a flat- rimmed bowl (ASH.42), the prototype for what will be an important form in Syria with later groups. Base profiles show general similarities but are not uniform. Generally the foot is a squarish wedge shape and is splayed, while squarer-profiled but equally splayed feet are also known (e.g., ROM.13). A number of pieces have ridges cut into the base, concentric to the foot (e.g., ROM.14), a practice that differs radically from the other pieces, which are essentially plain within the foot- ring. The "ridged" group may have prototypes in Egyptian Incised wares (see Fig. 4.8), and so may be the earlier; alternatively, the "ridged" and "unridged" group may represent the work of at least two different potters. General decorative arrangements for the Lustre-painted wares include a central figure with scrolling foliage, which may be divided into three groups. In the first of these groups (SSB1a) the vessel has a "crescent-rim" pattern (motif SS.1, Fig. 5.4), while the "foliage-fill" decoration (motif SS.2) fills the space around a central figure, which has "Egyptian-band" motifs (SS.3) similar to Fustat Group Two Lustre-painted wares (see Chapter 4). In both Syrian groups the figure and foliage are highlighted by inscription through the lustre pigment (motif SS.4), another typical Fustat Lustre Group Two link. The second and possibly later of these Syrian Group One figural styles (SSB1b) has the figure, but with rather spindly foliage floating in open space (SS.5). Another general decorative style includes a more geometric quadripartite pattern. The final variant of Group One (SSB1c) has coarsely painted pieces with large motifs which have been boldly incised through the lustre (Fig. 5.2: MRT.O7; also published in Porter and Watson 1988, no. 58). This decoration is probably derived from the incised figures of the first "Tell Minis" subgroup and may be argued to be the last expression of the group as a whole. Calligraphic decoration is also known, but it deteriorates into radial lines with intervening squiggles, particularly as a rim design (motif SS.6; see Fig. 5.2: MRT.07, FUA.38). The chronology for Syrian Group One may be suggested by comparison with Egyptian prototypes. Conical vessels are found in Egyptian Lustre Groups One and Two, but apparently not in Groups Three or Four. The slightly splayed rim is not found in Egyptian Group One, but is more characteristic of Group Two or possibly Group Three. The glazes that have been subject to analyses (below), are of the alkali-lead type, very similar to Egyptian Groups Three and Four, but also related to the Incised ware glazes of Egyptian Group Two and later. However, there are also opaque glazes (Porter and Watson 1987: 207, no. A3; also ROM.14), presumably tin- opacified, which would be like the Egyptian Group One and Two wares. Collectively, these data seem to indicate that technological knowledge of lustre-paint and stonepaste was transferred from Egypt (probably by movement of potters) at the period of transition from Egyptian Lustre-painted Group Two to Group Three. Unlike the first Iranian wares (see Chapter 6) there is little discernible difference between some Egyptian prototypes and some of the first Syrian wares, so this would suggest a date for the beginning of Syrian Group One at about 1075. Syrian Groups Two and Three (discussed below) simultaneously share motifs and forms which had developed from Syrian Group One, and derived from influence by Iranian Group Two (dated 1125-1150). This would suggest a date for Syrian Group One of c. 1075-1125, although the bulk of typical "Tell Minis" wares may be more suitable to a shorter range at the beginning of this period. At Hama wares of Group One type are said to commence in the eleventh century (Poulsen 1957, section B.VI, pp. 132- 56), but I have problems assessing the effectiveness of the Hama dating from the published evidence. The possibly very useful sequence at Tille Hoyuk (Moore 1993) puts some diagnostic Group One Lustre-wares (SSB1c) in context with worn Byzantine coins of the mid-eleventh century (but see Appendix C). Fragments of typical "Tell Minis" form, although undecorated, from ‘Ana are found in apparently undisturbed contexts in association with glazed clay-bodied pottery that Northedge (Northedge et al. 1988:102) links with identical wares from Qal‘at ‘Amman, where it was excavated from a destruction horizon dated to an earthquake of about the beginning of the twelfth century. Similar fragments are found at Qal‘at Ja‘bar in the earliest phase defined by Cristina Tonghini (1995), which she feels is possible to date to the late eleventh century based on historical evidence. It is also found at Qal‘at Ja‘bar in association with lead-glazed wares that she dates to the second half of the eleventh century at Tell Shahin. At ‘Acharneh, Group One wares are found in a context that historical sources date to after 1111 (see Appendix C). Blake (1978) describes a number of monochrome-glazed stonepaste pieces from Italy attributed to the eleventh or early twelfth century on the basis of either numismatic association or incorporation into buildings, but without further data they might be Egyptian. Pieces from Genoa which have been attributed to Syria are more likely Egyptian, with a "friable" body, opaque glaze, and also forms typical of Egyptian Group Two Incised wares (Pringle 1977), while others from this site are not illustrated (Pringle 1978). Pieces from Pisa with lead in the glaze are also more likely to be or are certainly Egyptian (Berti and Tongiorgi 1981, nos. 81, 237, 151). A fragment from an Athenian cistern of late eleventh to twelfth century context (Frantz 1938) might also be mistaken as Syrian, but this is also more likely to be Egyptian (see Chapter 4 and Appendix C). Any more exhaustive examination of the literature is plagued by such inabilities to accurately attribute the pieces. Syrian Stonepaste-bodied Group Two (SSB2) This group may be divided into two sub-styles. None of the first group (SSB2a) has been personally studied by myself. They comprise primarily a number of published Lustre- painted pieces from Hama which have decoration and forms primarily derived from Iranian Group Two Lustre-painted practice (Riis and Poulsen 1957, figs. 475-79, 486, 598, 599; see also Arts of Islam 1976 for a turquoise-and-black piece perhaps of this group). The forms of these vessels are dominated by bowls related to the cono-segmental form of the Iranian wares (see Fig. 6.6). Some of these have base forms with splayed foot-rings, probably derived from Syrian Group One ("Tell Minis") wares. Others have vertical feet more in accordance with Iranian practice. Decorative motifs which may be derived from Iranian prototypes include an overall design formed by a pattern of radial lines forming panels in between. Some of the patterns within these panels are also related to Iranian practice, the circular motif in a scribbled field, for instance, may be derived from the "big-eye" motif. More explicit derivatives of the "big-eye" (motif SS.7: see Fig. 5.4) are also found (Riis and Poulson 1957, fig. 486), as are "solid-palmette" designs. Particularly diagnostic are the solid palmettes, like crescents, arranged on each side of a large dot. Motifs of this type are also found on a large blue-glazed and Lustre-painted jar, which is often included with the large corpus of pieces attributed to fourteenth-century Damascus (e.g., Caiger-Smith 1985, pl. IX). Although the provenance would remain debatable without sampling, the decoration of this piece would appear closer to pieces of the twelfth century, as would the form which is closer to Fatimid jars (e.g., Caiger- Smith 1985, pl. 17), with a maximum width lower on the body than the later Damascus jars (e.g., Lane 1947, pl. 24). The second subgroup, known as the "line-rim" style, (SSB2b) are again Lustre-painted wares commonly attributed to Raqqa (Lane 1947, pls. 57B, 59A; see also Riis and Poulsen 1957, figs. 466, 666-67, 669, also possibly 671; and Grabar et al. 1978, G4.10c). Most of the pieces I have come across, and probably also the sampled piece MRT.06, are flat- rimmed bowls (Fig. 5.6). The flat-rimmed bowl may be interpreted as a local development from certain forms in the Syrian Group One typology (Fig. 5.5). That it developed relatively early would be suggested by the forms of the group of wasters said to be from the region of Ma‘arrat al-Nu‘man. As a group of wasters it is reasonable to entertain the thought that they are roughly contemporary. They include some Group One forms (Fig. 5.6: MRT.33), and a large number of flat- rimmed bowl forms, closely parallelled by some of the forms of Groups Two, Three, and Four (those without ridges on the outside, at the angle between rim and wall). The (possibly later) flat-rimmed bowl form with external ridge is related to that of Iranian Lustre-ware Group Four (c. 1175-1200, see Chapter 6). The "line-rim" motif itself (Fig. 5.4: SS.8) has parallels with similar motifs on Syrian Group One "Tell Minis" Lustre- painted wares (see Fig. 5.4; also Grabar et al. 1978, G1). A rim-fragment from Qasr al-Hayr al-Sharqi seems to show a transition from the later Group One single line, to a double line, as is found in the current group (Grabar et al. 1978, G4.10a; see also Raymond and Paillet 1995 for a piece from Balis/Maskana, and Mathaf 1922, pl. 97, for an Underglaze- painted variant). In turn, this motif may be ancestral to the "kufesque-band" rim motif (see below) found on Syrian Groups Three and Four wares (in particular, see Lane 1947, pl. 59A). Cristina Tonghini (1994) describes a "narrow lustre band" painted on the underside of the rim of vessels of this style in the Khalili collection, but this was not noted in the MMA pieces I have studied in person. Given the evidence for the dating of Group One, a general dating in the early twelfth century would be suitable for this successive style. Lane (1947, pl. 58) has already pointed out the very close similarities in general design between this group and certain Iranian Lustre-painted wares which in the current study would be attributed to Iranian Group Two (c. 1125- 1150), particularly in the panel-structure of the arrangement (motif SS.9). This dating would also correspond closely to that of Egyptian Group Four (c. 1125-1175), which has very similar flat-rim dish forms (see Chapter 4). Syrian Stonepaste-bodied Group Three (SSB3)—"S-back" This group appears to be mostly represented by polychrome Underglaze-painted wares, and would seem to be predominantly found in bowls with recurved rims, a form perhaps related to Song porcelains imported from China (see Fig. 5.7). The "S-back" motif (Fig. 5.4: SS.10) found on this group would appear to be a development from pseudo-calligraphic decoration found on at least one Syrian piece (Mason 1997c, fig. 12; also Waagé 1948, fig. 51.9). Hence it may be argued that it is derived from the practice of decorating Iranian Groups One and Two wares, which also have calligraphic decoration on the reverse; or are at least derived from a cognate practice. Presence of a distinctive crescent-rim pattern in this group with draped lines descending from the embayments may also be a link with Iranian Groups One and Two, or with the preceding Syrian Group One. A few pieces have the "kufesque-band" rim motif (Fig. 5.4: SS.12), more dominant on Syrian Groups Four and Five (below), and also rarely on Iranian Minai ware of styles equal to Iranian Group Four of c. 1175-1200 (Bahrami 1949, pl. 31). These particular pieces are likely to be at the end of production, possibly transitional with Syrian Group Four. One piece with the "kufesque-band" motif (MMA.38) has a radial design with sectors filled wavy lines that look like a cartographer's symbol for water, and the same pattern in the same arrangement (MMA.12) occurs in an Iranian Group Three Lustre-ware piece (c. 1150-1175). The apparent dominance of the recurved rim style may relate this style to Egyptian Group Three (c. 1075-1125), which shows similarity in the recurved-rim form style (see Fig. 4.6) Collectively all of these parallels would enclose dates of c. 1075-1200, but the main influences are early, while the relationship with Mina'i may be due to influence extending to Iran from the later Syrian wares (see below). Hence, a date of c. 1125-1150 may be more tenable, although considerable flexibility in both directions would be justified. At Tille Hoyuk (Moore 1993) the pottery appears to represent a limited period of occupation, and here some diagnostic "S-back" pieces are found in the third level, where they succeed Syrian Group One wares ("Tell Minis") and appear to overlap with some less diagnostic Group Two pieces (see Appendix C). One piece is recorded from the Jerusalem excavations (Tushingham 1985, fig. 40.11) but this appears to be from a late mixed context. Syrian Stonepaste-bodied Group Four (SSB4)—"arc-back" This group is predominantly represented by polychrome and turquoise and black Underglaze-painted wares, although Lustre-painted examples are also known with the "arc-back" motif. The diagnostic "arc-back" motif of this group (see Fig. 5.4: SS.11; see also Fig. 5.2: JRM.04) consists of a series of pairs of shallow arcs on the reverse. It may be argued that this was derived from practice in Kashan transitional between Iranian Group Two and Group Three or from cognate Syrian practice. It has been suggested for the Iranian groups that a freely painted continuous scroll on the reverse of the bowl, derived from the calligraphy of Iranian Group Two, may have developed into the "knot back" of Group Three. Both of these motifs appear on Syrian examples (see Mason 1997c, fig. 12; see also Logar 1991, Abb. 5.14, and Logar 1992, Abb. 15.12, 15.14, for some calligraphic scroll prototypes of the "arc- back"), in both cases exhibiting strong similarities in approach and motor-habit pattern to the application of the "arc-back" motif. A common motif on some pieces is the "kufesque-band" rim motif (see Fig. 5.4), which is also rarely found on Syrian Group Three and more commonly on Syrian Group Five (below), and also rarely on Iranian Mina'i ware of styles equal to Iranian Group Four of c. 1175-1200 (Bahrami 1949, pl. 31). This is generally considered to be derived from calligraphy, but the possible development from apparently non-calligraphic design elements in Group Two (see above) would suggest otherwise. Unfortunately, the motifs on the reverse of vessels were not considered by the authors of the Hama report (Riis and Poulsen 1957), but a number of pieces have the kufesque rim pattern. A number of pieces were found in the Jerusalem Armenian Garden excavations (Tushingham 1985, figs. 34.34, 40.8, 40.17, 44.4). Most of these examples would appear to be of a conical form, although it is likely that they are of the "proto-biconical" form. The development of the biconical bowl may be hypothetically shown to have originated in a simple conical bowl (e.g., in Syrian Group One, Fig. 5.5: ROM.14). An intermediate stage would seem to resemble a conical wall placed on a flattened bottom, in turn placed on a columnar foot: a proto-biconical bowl (see Fig. 5.7: MMA.21). This would then lead to a true biconical bowl, but the walls of the lower cone would gradually steepen through time. Such a development would hold for Iranian as well as Syrian wares (it is illustrated Fig. 2.2). Fragments of proto-biconical bowl bases at Jerusalem may be associated with the conical rim fragments. One piece that may have an "arc-back" is from a building in Pisa, which together with another piece is dated late twelfth to early thirteenth century (Berti and Tongiorgi 1981, no. 306, fig. 204), but unfortunately this seems to be by reference to the usual literature (Lane's Early Islamic Pottery) rather than to context, which has been very useful on other occasions (see Chapter 4). The type may not be necessarily associated with a particular centre at the moment, although the bulk of sampled types (from Fustat and Jerusalem) are of the petrofabric attributed to Damascus. The association with the transition from Iranian Lustre Group Two to Group Three would suggest a date for this group commencing in c. 1150, and continuing for about fifty years, in order to be contemporary with the Iranian Mina'iwares as well; from this evidence a range of c. 1150-1200 may be suggested with a preference towards the beginning of that range. The material from Jerusalem is from excavations that may provide supplementary dating evidence. There is a wealth of coin evidence from the excavations, with contexts excellently recorded in the Wheeler-Kenyon method (the excavations were begun by Dame Kathleen), and fully reported by Doug Tushingham (1985). The publication does, however, seem to rely more strongly on historical evidence that relates to the city wall, and Tushingham tries to put the entire Ayyubid occupation into the short period of 1212 to 1227. Following re-examination of the evidence, with correlation of coins and pottery to specific strata (see Appendix C) the Jerusalem excavations suggest a similar dating to that proposed above (i.e., c. 1150-1200), with these wares being in an abraded state of tertiary deposition at the time of the make-up associated with Zangid coins and prior to the time of the Ayyubid occupation after 1191. Proto-biconical bowl bases appear to be found in the make-up fill, supporting an early date for the introduction of this form. Syrian Stonepaste-bodied Group Five (SSB5)—"dash-back "The "dash-back" group, named after a common motif on the reverse, includes a large corpus of polychrome bowls with flattened rims, and indeed would appear to be the only polychrome group other than the "S-back" and "arc-back" groups. Other vessels include biconical bowls, and bowls with recurved rims (see Fig. 5.8), many of which are turquoise and black. The "dash-back" motif itself comprises a series of short lines or dashes (Fig. 5.4: SS.13), commonly slanted but in some cases becoming simplified into a broken line. A motif related to the kufesque calligraphy-derived rim motif of the "arc-back" group is also found in this group, but often the design shows simplification. Again this may be related to some examples from Hama, but as designs on the reverse are not reported fully by Riis and Poulsen (1957), it is impossible to say if the examples belong to this or the preceding group. As with the preceding group, the "kufesque-band" can be related to motifs on Iranian Mina'i pieces (Bahrami 1949, pl. 31: the parallel is actually closer with this group; see also Lane 1947, pl. 71B). Other motifs from this group found on Iranian wares include a band of palmettes in circles on Mina'i (Bahrami 1949, pls. 31, 34) and Lustre pieces (Bahrami 1949, pl. 38), and a band of laurel-leaf shaped lozenges (see Soustiel 1985, pl. 133) found in Mina'i (again Bahrami 1949, pl. 31). Foliage sprays found in the field of certain figural pieces (e.g., Lane 1947, pl. 78B) are also related to such motifs in Iranian Group Four Lustre and Mina'i (Lane 1947, pl. 71B). However, certain of the these polychrome figural pieces have more elaborate foliage, suggesting that this motif developed locally in Syria. In most cases it would appear that motifs shared by Iranian and Syrian wares are closer to original prototypes in the Syrian wares. Also, it may be suggested that Mina'i wares borrow from the full range of Syrian Underglaze-painted ware motifs and also from the contemporary traditions of Iranian Lustre-painted wares. The alternative would be that the Syrians copy only those Mina'i motifs not also found on Iranian Lustre wares, which is unlikely. Hence, it is suggested that, despite the widespread opinion that polychrome Syrian wares are derived from Minai, the opposite is true, and it is this Syrian group that is the closest in style. Dating for this group would therefore be preceding and roughly contemporary with Iranian Group Four, that is, c. 1175-1200, contemporary with Syrian Group Five. Also included in this group are a number of fragments of biconical bowls which unfortunately are represented only by their bases, and do not have the reverse with a diagnostic motif (Fig. 5.8). Traits that may make this group of bases contemporary with Syrian Group Six are the presence of motifs that would suggest their contemporaneity with Iranian Group Four, such as the "chequer-tree" motif (see Fig. 5.2: FUR.44). The reason that these wares may be treated separately, and may indeed form a separate group, is the distinctive nature of their decoration. This consists of various elements in a field of hastily drawn sketchy foliage, and does not appear to be found on any wares with the "dash-back" motif, which are mostly painted in reserve. Such foliage may be noted in some Group Two pieces where it is delimited by a line that isolates a blank area between a central figure (e.g., Fig. 5.2: MRT.06). Unrestrained treatments of this nature are known with the "kufesque" rim (see Grube 1976, no. 217) of Groups Four and Five, and are also known among the Group Six wares (see below); however, the "sketchy-foliage" group includes a large number of polychrome wares, which do not appear to exist in Group Six. Other than polychrome Underglaze-painted, types included in this group are underglaze black with a turquoise glaze, and a number of Lustre-painted pieces with dark turquoise or blue glazes. Samples of the "dash-back" group have been predominantly of the Damascus Petrofabric, but some bichrome examples are of Raqqa-2, including a sample from the recent excavations, while one appears to be of the "Ma‘arrat" Petrofabric (see Fig. 5.3). Syrian Stonepaste-bodied Group Six (SSB6)—"line-back" This group, characterized by the "line-back" motif on the reverse, is also found predominantly in flat-rimmed bowls, although the rim is nearer the horizontal than in the preceding group, and has no exact parallels with Iranian types. Apart from the "line-back" motif, consisting of a sinuous or (perhaps later) straight line painted on the reverse (Fig. 5.4: motif SS.14; see also Fig. 5.3: ASH.35), another characteristic motif is the "lozenge-rim" (SS.15), a series of lozenge-shaped marks or dots arranged along the flat rim (see Fig. 5.3: ASH.35; see also Grube 1963, Abb.12; Soustiel 1985, pl. 131; Grabar et al. 1978, G5.5a). No pieces with red pigment have been found of this group, and the group is at most bichrome (cobalt-blue and chromium-black under a colourless glaze, or more commonly chromium-black under a copper-turquoise glaze) rather than polychrome. The decoration of a number of examples of a beast surrounded by scrolling foliage (see Fig. 5.3: ASH.35) recalls that of Syrian Group One ("Tell Minis") and also Kashan Group One Lustre-wares, but such general designs may continue for some time in regional centres. The "line-back" motif itself might be best interpreted as a development from the broken-line variant of the "dash-back" motif. The "sketchy-foliage" treatment is also found in this group, including isolated foliage spread across a field, similar to the proto-biconical bowls allocated to Syrian Group Five; while at other times it is constrained, as in Group Two (cf. Porter 1981, pl. X). The flat-rimmed bowl form, which appears predominant in this group, is primarily of the variant with the ridge on the exterior at the junction between rim and vessel wall. This also occurs in earlier groups but not with such frequency. This may argue for a later date for this form variant, although it may simply indicate different practices at different centres. In short, a narrow time span for this style may not be suggested at this time, and characteristics may be attributed to periods successive of Group Two (after 1150) to periods successive of Group Five (after 1200). However, were this group made predominantly after 1200, one might expect a larger proportion of biconical bowls, as in Group Seven, although those examples of biconical bowls that do occur are suitable to the thirteenth century. The one piece of stonepaste pottery from the Armenian Garden excavations at Jerusalem that appears to have been in use during the Ayyubid occupation, being practically whole rather than comprising highly abraded sherds (Tushingham 1985, fig. 37.14), is of incised ware, but its decoration is almost identical to a "line-back" piece in the Ashmolean Museum (ASH.37, published in Allan 1981, no. 364). Fragments from an excavation at Harran included wares which may have been of this group (Rice 1952, figs. 14.1, 14.8, 14.11), and these were found together with Group Seven wares which may be dated to c. 1200-1250. Only two pieces of this group have been thin-sectioned to date; one is of Raqqa-2, the second of "Ma‘arrat" Petrofabric. Syrian Stonepaste-bodied Group Seven (SSB7)—"spiral- back" This group appears to represent the bulk of Syrian Lustre- painted wares traditionally attributed to Raqqa (Grube 1963, although most of the examples in Lane 1947 are of Group Two, see above). The "spiral-back" motif itself (Fig. 5.4: SS.19) consists of a series of spiral lines arranged in various complexities as a band on the reverse of the vessel (e.g., Fig. 5.3: ABS.02). There is no obvious ancestor to this motif, but it is conceivable that it developed from the "arc-back." This would really be only a minor adjustment of application for the potter, from opposing arcuate sweeps to a spiral sweep of the brush. Such a hypothesis gains credence by the observation that "arc-back" appears to be the only other motif found on the reverse of "Raqqa"-style Lustre-wares. Other motifs include "dot-pattern" (motif SS.17) and "calligraphy-band" (SS.18). The most common form would appear to be a biconical bowl, although flat-rim bowls (see Fig. 5.9; see also Atil 1973:140) and "hammer-rim" profiles are also known (Riis and Poulsen 1957, fig. 679), together with a variety of closed vessels (Grube 1963). Although the traditional attribution of biconical bowls to solely after 1200 is not supported by the findings of this study, the similar predominance of this form in Iranian groups dated to after 1200 may be a useful parallel. The lack of any Yuan-influenced form would also seem to indicate that this group had ceased production by the date of introduction of the Yuan forms (after 1260, judging by the Iranian evidence). The technique of filling a field with fine spirals (e.g., Grube 1963, Abb. 25, 26) is also correlated to the technique of the spiral-field motif of the Iranian wares. Given the relationship with earlier Syrian wares and contemporary Iranian wares, a date range of c. 1200-1250 might therefore be suggested for this group. Fragments that would appear to be of this group were excavated at the Harran citadel (Rice 1952, fig. 14.2, 14.12), and this has been used towards dating of the pottery elsewhere (Northedge et al. 1988:102). However, I do not find the evidence conclusive (see Appendix C). Petrofabrics in this group comprise both Raqqa-1 and Raqqa-2, the latter being the one reliably attributed to Raqqa. Syrian stonepaste wares summary In summation, the dating of the Syrian stonepaste-bodied wares is as follows: Group One, comprising "Tell Minis" Lustre-painted and Polychrome-relief ("Laqabi") wares attributed tentatively to western Syria, is dated to c. 1075- 1100/25; Group Two, comprising Lustre-painted wares possibly attributed to Raqqa and elsewhere, is dated to c. 1125-1150; Group Three, comprising polychrome Underglaze-painted wares attributed to Damascus, is dated to c. 1125-1150; Group Four, representing Lustre-painted wares from Raqqa, polychrome Underglaze-painted wares from Damascus, and bichrome Underglaze-painted wares from "Ma'arrat," is dated to c. 1150-1175/1200; Group Five, comprising polychrome Underglaze-painted wares from Damascus and Raqqa, is dated to c. 1150/75-1200; Group Six, comprising bichrome Underglaze-painted wares from Raqqa and "Ma‘arrat," is dated to c. 1150/75-1200/50; and Group Seven, comprising Lustre-painted wares attributed to Raqqa, is dated to c. 1200-1250. Many of these are obviously contemporary, but different approaches to forming and decoration in some of the groups, together with the petrographic evidence, would indicate that they were made in different workshops, or even different centres. STONEPASTE WARES: TECHNOLOGY As in Iran, the sudden appearance of stonepaste bodies in Syria, together with Egyptian glazing technology, lustre-paint technology and Egyptian forming and painting practices, suggests the movement of potters from Egypt to Syria. Bodies The Syrian stonepaste bodies are all of the fully developed stonepaste technology, comprising quartz bound together by a network of glass. The earliest Syrian stonepaste pottery, Group One or "Tell Minis" wares, have very finely ground angular quartz grains. The later Syrian wares tend to be much coarser, especially the wares attributed to Damascus (see provenance section, below), which include rounded and uncrushed sand grains of c. 0.3 mm diameter, together with finer, angular grains, presumably representing crushed sand. The chemistry of these bodies (see Table 5.2) tends to form two groups. A number of pieces, including all the Group One ("Tell Minis") samples and possibly one "Ma‘arrat" waster piece (MRT.20, which may actually be considered more transitional between the two groups), appear to have the normal amount of alumina in a typical stonepaste, usually about 2%. However, the rest have very little alumina,sometimes below detection limit. These later pieces, particularly the wares attributed to Damascus, are also relatively highly calcareous, containing up to 8% CaO. This would suggest that rather than using a highly aluminous clay in the stonepaste mix, the later Syrian potters used a highly calcareous clay (unlike some Egyptian examples, there is no evidence of limestone in the sand). This change in the type of clay used seems to be related to chronology rather than provenance, as it can be noted in a number of Syrian petrofabrics. The switch is closely linked with a change from the alkali-lead glazes of Group One to the alkali glazes of the subsequent groups, which may explain the need for this technological change. Since being first reported (Mason 1994) this change of body chemistry has also been noted in analysis of pottery from Qal‘at Ja‘bar (Franchi et al. 1995), and in pottery from Balis/Maskana, Raqqa, and the Quwayq valley (Perez-Arantegui 1995). Glazes and surface zone Opaque glazes have been noted by others for some of the Syrian wares described above, and has been observed on one occasion by the author (a Group One piece), but they are exceedingly rare. Syrian stonepaste wares with clear glaze, can be divided into two main groups: alkali-lead glazes and alkali glazes. The first group is represented by samples of Group One ("Tell-Minis") Lustre-painted wares and associated Polychrome-relief or "Laqabi" ware (see Table 5.3). The Group One alkali-lead glazes are similar to those of Egyptian Groups Three and Four wares and not too dissimilar from those of the Egyptian Incised wares (see Table 4.3). The clear nature of the glazes required that the Egyptian red clay body was lightened by a wollastonite slip. This would not be a problem with the white stonepaste-bodied Syrian wares. Porter and Watson (1987:207, no. A3) describe some "Tell Minis" wares with opaque glazes, presumably tin-opacified, and I have noted one in the ROM collection (ROM.14). If this is so, then this would also provide a technological link with the common practice for Lustre-wares in Egypt. Since being first reported (Mason 1994) the predominant alkali-lead chemistry of the Group One glazes has also been noted in analysis of pottery from Qal‘at Ja‘bar (Franchi et al. 1995), but not in glazes from Balis/Maskana, Raqqa, and the Quwayq valley, although some "Laqabi" wares were apparently included (Perez-Arantegui 1995). The remaining stonepaste-bodied wares, including Lustre- painted and Underglaze-painted wares, have alkali glazes. These are predominantly soda (13-22% Na2O) with little potash (1-3% K2O), although some are also quite highly calcic (5-10% CaO). The origin of this glaze technology is problematic. No alkali glazes were found in the Egyptian wares, so presumably the potters coming from Egypt would have "discovered" the technology in Syria. This may be particularly relevant given the production at Raqqa, which is physically within the sphere of Mesopotamia, and may have provided access to the ancient alkali-glaze technology of this region. Pre-Islamic Mesopotamian glazes analysed in this study are also predominantly sodic with high lime; however, early Islamic alkali glazes contain roughly equal amounts of soda, potash, and lime (see Table 3.4). Whatever the origin of this glaze technology, it would appear to be similar to that used in Iran at the same time (see Table 6.5). The presence of magnesium in the glaze, often cited as evidence for use of plant ash as a source for alkali, would appear to be low for the Syrian wares, which contrasts not only with the earlier Mesopotamian glazes but also with the contemporary Iranian glazes (see Chapter 8). The Underglaze-painted wares are decorated with true pigment-paints, with no medium of clay or stonepaste in the application (Pl. 5.1). Colorants include chromium black, cobalt-blue, copper-green, copper-turquoise, and iron-oxide red. Even the thick red ochre pigment is entirely composed of grains of iron oxide, with no other mounting medium (see Pl. 5.2). The dating of these wares (Group Three, c. 1125-1150; Group Four, c. 1150-1200; Group Five, c. 1150-1200) would put them considerably earlier than the first Iranian true Underglaze-painted wares, which would seem to date to well into the thirteenth century (see Chapters 6 and 8). STONEPASTE WARES: CHARACTERIZATION AND PROVENANCE In presentation of the data for the provenance of the Syrian wares we will first discuss the historical and archaeological evidence, and then the petrographic data. The abundance of selected inclusions is provided in Tables 5.4 and 5.5. Grainsize-distribution histograms for these wares are provided in Figures 5.10 and 5.11. Production centres Given the shared technology of the stonepaste potters and glassmakers, it may be worth noting that Tyre and particularly Aleppo were known as glassmaking centres in Islamic times (Porter 1981:9). A possibly important observation was made in the thirteenth century by Yaqut that the Aleppo glassmakers obtained their quartz as sand from the Jebel Bishri, south of Raqqa (Porter 1981:9). Recent preliminary analysis of the stonepaste pottery from the citadel at Aleppo indicates a local source of quartz. It remains to be seen whether the glassmakers and potters were using completely different sources, or Yaqut was just completely wrong. Historical references to pottery production seem mostly to talk about centres around Ma‘arrat al-Nu‘man. Yaqut in c. 1225 mentions ceramic production at Armanaz (about halfway between Aleppo and Ma‘arrat). Abu ‘l-Fida (d. 1331) in c. 1321 mentions ceramic production at Kafar Tab, just to the south of Ma‘arrat. The Yemeni historian al-Khazraji (d. 1409) describes pottery from Shayzar being used in the circumcision ceremonies of the son of the Sultan (r. 1377-1401) (Porter and Watson 1987:205). Shayzar is the well-known fortress-city on the Orontes near Hama, but the district of Shayzar is described by Yaqut as being near Ma‘arrat al-Nu‘man (Porter and Watson 1987:206). None of these sources, apparently, refers to glazed wares, and two of them date to after 1250. This date is significant because it delimits the scope of the present study, and because after this date it would seem that the only production centre for glazed wares that can be recognized is Damascus (see below). Archaeological evidence for production in the region of Ma‘arrat al-Nu‘man consists of a large group of wasters said to be from the area (see below). Other documentary evidence exists for Damascus. Lane (1957:15-16, pl. 7) describes a large high-shouldered jar, with blue glaze and lustre-painted decoration, and an inscription saying it was the work of one Yusuf in Damascus. Circumstantial evidence is available from a number of European sources, for instance, the "eleven pots of Damascus" listed by an apothecary in Valencia in 1329; a further 59 pots and 28 bowls listed by another apothecary in Barcelona in 1364; and numerous smaller numbers mentioned in France and Italy in the fourteenth and fifteenth centuries (Lane 1957:17). A further 40 Damascus jars were unloaded in London in 1445 (Childs, in press). Abundant archaeological evidence for production at Damascus also exists, including published wasters (Carswell 1979). All of this evidence is for material that post-dates the period covered in this study. The significance of this evidence lies in its application to the characterization of a Damascus Petrofabric (see below). Raqqa has traditionally been the most popular suggestion for the originating centre for most of the pottery described above (Tonghini and Grube 1989). Evidence for Raqqa consists of a number of descriptions dating to the first half of the twentieth century of wasters and kilns (given in full by Porter 1981:10-11) by traveller Gertrude Bell, Islamicist Jean Sauvaget, and Eustache De Lorey, Director of the French Institute in the 1920s. A great deal of pottery and wasters sold through the art market allegedly came from the diggings described in these accounts. This includes a group of wasters said to be from Raqqa, and currently stored in the ROM (see below). Unfortunately these differ from wasters excavated at Raqqa by the expedition led by the late Michael Meineke, so this attribution is highly suspect. Sauvaget (1948) published a typology of "Raqqa" wares based on excavations at the site, but these pieces are mostly fragmentary, and actual wasters are not published. There also appears to be evidence of stonepaste-ware production at Qal‘at Ja‘bar (Tonghini 1995, 1996), and there are rumours of such production at Balis/Maskana, but no evidence has been published (e.g., Perez-Arantegui 1995; Raymond and Paillet 1995) and it may otherwise remain questionable whether the alleged evidence is relevant to stonepaste pottery production. Recently, wasters of Underglaze-painted ware have been found in the excavations of the Aleppo Citadel (Gonnella in press), and petrographic analysis has shown these to be of a unique petrofabric. Apart from the attested centres there has traditionally been some attribution to Rusafa, particularly of polychrome Underglaze-painted wares. Even beyond this group there may be previously and currently unknown centres, as indicated for instance by the wasters from Halabiya (see below). Pottery stylistically related to later "Raqqa" style Lustre-wares is thought to have been produced at Konya, although the attribution of all wares with simple blue designs such as those found on the Konya wasters remains problematic because it is likely that they were all destined to be Lustre-wares of the late "Raqqa" style. Chemical analysis of a group of Syrian Lustre- wares has also produced a number of distinct groups (Redford and Blackman 1997). Damascus(?) Petrofabric Evidence for production of pottery at Damascus is formidable for the fourteenth century; indeed, the only unequivocal evidence for stonepaste-bodied wares in this period seems to be for this centre (see above). Petrographic analysis of fourteenth-century pottery from Syria, including Lustre- painted wares, Underglaze-painted wares of Yuan style, and also tiles (Mason 1996; cf. also Carswell 1979; Toueir 1973, pl. 11, B, b-c), has revealed a single petrofabric. The occurrence of evidence for a single centre combined with the existence of a single petrofabric is not considered a coincidence, but a direct correlation. The Damascus(?) Petrofabric has therefore been tentatively defined by analysis of pottery of Mamluk date. The Damascus(?) Petrofabric itself consists overwhelmingly of clear quartz, with small amounts of slightly cloudy quartz, and a general absence or trace amount of any other inclusion. This composition easily separates it from any other petrofabric (see Appendix B). The largest grains are well-rounded, indicating a raw material comprising a sand with average grainsize diameter of about 0.3 mm (see Fig. 5.2). This is entirely unlike any other petrofabric defined for Syria, or indeed any other region, perhaps the closest being that of Rayy (see Chapter 6 and Appendix B). Of course the importance of Damascus in the Mamluk period is not of particular relevance to this study; however, the Damascus(?) Petrofabric is also found in wares of our period. Damascus would seem to have been the predominant production centre for all Group Three "S-back", Group Four "arc-back," and Group Five "dash-back" styles, which includes the majority of the sampled polychrome wares, and all the wares with sketchy decoration assigned to Group Six, including the Lustre-painted wares. As such Damascus would appear to be the most significant Syrian production centre. However, this appraisal may be biased in that a large number of the samples come from sites that would have been recipients of the export market, such as Fustat, Ani, and northwest Europe, or locally dominated by Damascus, such as Jerusalem and Dhiban. "Tell Minis" Petrofabric The traditional attribution of the "Tell Minis" style has been based on numerous finds of saleable vessels (in both the ancient and the modern pottery markets) in western Syria, including a cache of thirteen vessels found at Tell Minis itself (Porter and Watson 1987). Further association with western Syria was made by the Hama excavations, where a large number was found, while other finds were made in Damascus and Ma‘arrat al-Nu‘man (Porter and Watson 1987). Nothing is known of the find-spot of the ROM bowls (see Appendix A,) but they were purchased from a Beirut dealer. On archaeological sites distribution is quite wide, with wares of this style (see above) found in Turkey (Tille Hoyuk) and Iraq (‘Ana), and pottery of this petrofabric found in eastern Syria (Raqqa) and Egypt (Fustat). Four diagnostic samples of the "Tell Minis" Lustre style were initially available for analysis (see Table 5.4), predominantly of the substyle comprising large motifs with lines inscribed in them. One is from the group said to be from the Tell Minis/Ma‘arrat al-Nu‘man area (see under Ma‘arrat al-Nu‘man in Appendix C), one is from Fustat, and two are unprovenanced. These last two samples are from the Ashmolean Museum and actually have the same accession number: 1978.2217, as they were formerly put together by a dealer to form a single vessel (see Appendix A). They are clearly from separate vessels originally, and the rim is actually of the "Ma‘arrat" Petrofabric (see below). A fifth sample is of Polychrome-relief or "Laqabi" ware, and is also from the group of wares said to have been found at Ma‘arrat al- Nu‘man. Two further samples, apparently undecorated but of diagnostic form, were added from the recent excavations at Raqqa (see Appendix B). The petrofabric of these wares (except that attributed to "Ma‘arrat") is generally very fine, with a maximum grainsize diameter of 0.2 mm (see Fig. 5.2). The quartz is predominantly slightly cloudy, with significant cloudy and clear quartz, and some very cloudy quartz. There is also some cryptocrystalline quartz, possibly chert, up to 5% in some cases, with little or no sheared or polycrystalline quartz, but this may be a function of the fine grainsize. Any correlation with the "Ma'arrat" Petrofabric (below) does not seem tenable. Although the cloudiness of the quartz is similar, and the relative lack of "sheared" and polycrystalline quartz in "Tell Minis" may be due to the finer grainsize, there are also distinct general differences, specifically, the higher feldspar in the Ma'arrat Group and the higher cryptocrystalline quartz in the "Tell Minis" group. It may, however, be reiterated that one example of the Group One "Tell Minis" type is actually of the "Ma‘arrat" Petrofabric. Raqqa-1 Petrofabric Characterization of this petrofabric was initially based on a group of vessels in the ROM, including wasters, said to be from Raqqa by the dealer from whom they were originally purchased. As wasters were included in the group, they were considered to define the Raqqa Petrofabric, and this was so named in earlier reporting (Mason 1991a, 1994, although with uncanny prescience I warned of relying on dealer's attributions). Unfortunately, subsequent analysis of wasters excavated at Raqqa were of a different petrofabric; indeed, the original Raqqa Petrofabric was not found in any of the sherds excavated from Raqqa. It might therefore be logical to use the name Raqqa Petrofabric for this subsequent group, and to rename the original petrofabric ("Bogus Raqqa" might be suitable). However, in the interests of consistency, the former Raqqa Petrofabric has been renamed the Raqqa-1 Petrofabric, while the body that may be reliably attributed to that site is known as Raqqa-2. The Raqqa-1 Petrofabric is characterized by a predominance of clear quartz, although it has significant slightly cloudy quartz and a very high content of "sheared" quartz. It is easily distinguished from all the other petrofabrics, but is very different from some "Ma‘arrat" Petrofabric examples (which, intriguingly, were also attributed to their site by a dealer). In these cases the "Ma‘arrat" Petrofabric contains feldspar which is absent in Raqqa-1, and "Ma‘arrat" is consistently coarser (see Fig. 5.3). Given that this petrofabric can no longer be attributed to Raqqa, it remains problematic as to where this group was actually produced. Apart from the turquoise and black wasters, it also includes Lustre-painted wares, specifically, the one example of Group Two available for analysis (MRT.06) and one example of Group Seven (MRT.23). "Ma‘arrat" Petrofabric The "Ma‘arrat" Petrofabric is defined by analysis of a group of wasters said to have come from Tell Minis, near Ma‘arrat al-Nu‘man, and donated to the Ashmolean Museum in 1980 by Dr. H. Bartels (see Appendix C). This is again a dealer's attribution, and hence may be suspect. As the petrofabric has subsequently been shown to differ from the "Tell Minis" Petrofabric as first defined (above), it was named after the nearby large town; because it is a dealer attribution, the name has quotation marks. As a collection of wasters the "Ma‘arrat" group does serve one very useful purpose. It is reasonably certain that the large group of wasters were all made in the same place. As such it provides an excellent opportunity to explore the variability that may occur in the products of a single site. Therefore, a far larger number of these were sampled than would otherwise be necessary. This has been of great utility not only as a methodological exercise, but also because the group covers the mid-range of quartz cloudiness (see Table 5.5). A number of other petrofabrics also cover this mid- range, and could easily be confused with the "Ma‘arrat" group if it were not so well defined. The petrofabric itself is quite coarse, with a maximum grain diameter commonly up to 0.5 mm. The quartz itself is mostly slightly cloudy, with a significant amount of cloudy grains, although occasionally there are more cloudy than slightly cloudy grains. Polycrystalline quartz is always present, and "sheared" quartz is generally very common. Cryptocrystalline quartz, probably chert, is occasionally found, while feldspars are always present, up to 3% of content. The occasionally very high "sheared" quartz content is similar to the Raqqa-1 Petrofabric, but Raqqa-1 has predominantly clearer quartz and no feldspar. The Halabiya Petrofabric is very close to many samples of the "Ma‘arrat" group, and differs significantly only in the lack or trace amount of feldspar. The Raqqa-2 group in many cases differs only in the finer grain size and in the trace or absence of feldspars. Apart from the waster group, a figural "line-back" style piece and a turquoise and black "dash-back" piece have been included in this group, while an "arc-back" piece is among the waster group (see Fig. 5.3: MRT.29). Halabiya Petrofabric This petrofabric is defined by analysis of pottery found at Halabiya, north of Dayr al-Zor on the Euphrates. The samples are two out of three warped sherds which are fused together, and hence are a waster. The sherds, although distorted, represent rim fragments from a flat-rimmed vessel and from a conical, or biconical, bowl. Both are very similar to many of the wasters attributed to Ma‘arrat al-Nu‘man (see above), and prior to analysis it was considered that this may have been the true origin of these pieces. However, the petrofabric does seem discernibly different. The Halabiya petrofabric is characterized by predominantly slightly cloudy to cloudy quartz, with some "sheared" quartz, in coarse grains up to 0.5 mm in diameter. As such it is very similar to some of the "Ma‘arrat" samples; however, the consistent feldspar content of "Ma‘arrat" is absent or in trace amounts. Had such large numbers of "Ma‘arrat" samples not been analysed this single parameter might not have been seen as sufficient to separate the two groups, but the consistently high feldspar content of the "Ma‘arrat" group justifies separation. Raqqa-2 Petrofabric This petrofabric was initially defined by analysis of pottery that was considered to be typologically of the "Raqqa" style, and as such represented more of the sampled "Raqqa" Lustre- wares than the petrofabric previously thought to be attributable to Raqqa (Mason 1994). Subsequent analysis of a group of wasters from recent excavations at Raqqa found that this petrofabric was the same as that of the bona fide Raqqa wasters, and is hence the true Raqqa Petrofabric. As a Raqqa Petrofabric had already been defined (see above), this petrofabric was renamed the Raqqa-2 Petrofabric, although sensu strictu it is the only petrofabric reliably attributed to that site. Apart from the Raqqa wasters, this group includes Lustre- painted sherds from the excavations at Abu Sudairah in Iraq (see Fig. 5.3), excavated by Gerald Reitlinger (1935). Generally this group consists predominantly of slightly cloudy to cloudy quartz, often with significant amounts of very cloudy quartz, and some polycrystalline and "sheared" quartz, in a fine grainsize. In many regards this is not too dissimilar from the "Ma‘arrat" group, although the generally coarser grainsize, the presence of feldspar, and the generally relatively clearer quartz of the latter group distinguishes it. The Halabiya wasters are even closer, and in some cases only the coarser grainsize of the Halabiya samples distinguish them. The samples of this group, although generally attributed to Raqqa, are quite varied. The Lustre-painted wares include a Group Seven "scroll-back" piece, a "dot-pattern" piece, a typical "line-back" piece, a typical "dash-back" piece, and a relief ware piece. The "dash-back" piece is the only polychrome Underglaze-painted sample not attributed to Damascus (or more recently Aleppo), and apart from its petrofabric it would be perfectly typical of its group. The Polychrome-relief or "Laqabi" ware sample (one of two available for thin-sectioning, the other being of "Tell Minis" Petrofabric) has also been included in this petrofabric, although the quartz is rather cloudier than the other samples, and there is hardly any "sheared" quartz. Petrography discussion As mentioned above, the thirteenth-century writer Yaqut provided an account of how the glassmakers of Aleppo acquired their quartz source, a sand, in the Jabal Bishri. The postulated centre of Rusafa is nearby the source; Raqqa would have been en route to Aleppo, Halabiya a short distance to the east (closer than Aleppo), and Tell Minis and/or Ma‘arrat al- Nu‘man a relatively short distance further than Aleppo. Hence, the question of long-distance transportation of raw materials for making pottery is a problem which must be addressed. If long-distance transportation was pursued on a large scale it would make the petrofabrics of all the centres in this part of Syria exactly the same. Fortunately, this is not what we find. Whenever it has been necessary to separate wares, such as wasters from Raqqa, Halabiya, and Ma‘arrat, or Damascus wares of Mamluk date, it has been possible to separate them. Further, when it has been necessary to group together wares that must have been made in the same place, such as the group of Ma‘arrat wasters, they have formed a cohesive and largely distinct group. The Aleppo glassmakers made a very valuable product, commonly traded to Europe and as wondrous in China as porcelain was in the west. Hence, although it might be argued that bringing raw materials such a long distance would be justified for the glassworkers, it would certainly be less justified for the pottery-makers. Besides, the glassmakers were probably very jealous of their resource, as are all traditional craftsmen, and would have taken measures to prevent others using their sand. And all this is assuming Yaqut knew what he was talking about. Currently ongoing research on pottery production in Aleppo indicates that the potters used local sources of quartz for their stonepaste (for Aleppo production see Gonnella, in press). Another point worth noting is the apparent inability to link one stylistic group uniquely to a petrofabric, although in many cases there are strong associations. Judging by the dichotomies of cost between certain wares, such as finely painted polychrome wares against more loosely painted bi- or monochrome wares, there probably would have been a hierarchy of wares. This hierarchy often has better links to centres than do the motif-defined groups; for instance, polychrome Underglaze-painted wares aremade almost solely at Damascus. Hence, the phenomenon that we are observing may be the copying of styles of a higher-status centre by lower-status centres. This is indicated by the Underglaze- painted wares in the Ma‘arrat waster group which show general similarities to the main type groups, but fail in the details. In particular, note the heavy "arc-back" of MRT.29 (Fig. 5.3), and the attempt at a "kufesque-band" pattern on the rim interior which fails miserably. Although copying would certainly appear to be a relevant factor, the movement of potters may be another. These centres are all relatively close to each other, and this may have enabled greater movement of potters, perhaps necessitated by the instability of the region (see Chapter 7). SYNTHESIS The very earliest Syrian glazed wares may have been moulded and lead-glazed wares showing a marked continuity in technology and tradition from local Romano-Byzantine practice. This was rapidly supplanted by influence from elsewhere in the Islamic world, particularly Iraq. Iraqi- influenced wares were particularly important in the Mesopotamian part of Syria, but the coarse slip-incised and slip-painted high-lead glazed wares became widespread. Although there may have been use of the proto-stonepaste technology at Raqqa in the ninth century, it did not lead on to true stonepaste, as it did in Egypt, and early Syrian wares remained technically and artistically undistinguished. This situation ended with the introduction of stonepaste and lustre-painting technology in the late eleventh century, and the first production of Syrian Group One ("Tell Minis") wares. As in Iran, the forms and decorative motifs, coupled with the lustre-painting and stonepaste-body technologies, indicate that potters actually came from Egypt to Syria in about 1075. The glaze technology of these first Syrian wares was of soda-lead chemistry, very similar to that used in Egypt after c. 1075, but also related to glazes in Egypt made before that date. Most of the later Syrian stonepaste-bodied wares had alkali glazes. The presence of alkali glazes in the Ma‘arrat waster group, which in some cases has the same forms as the Syrian Group One ("Tell Minis") style wares, would suggest that the switch to alkali glazes was undertaken quite early, perhaps by 1100. This would not necessarily be earlier than in Iran, where alkali glazes were often used on the reverse of vessels in this early period. However, the Syrian potters showed a major commitment to the technology by using solely alkali glazes in subsequent wares. Alkali glazes do not seem to have been used in Egypt, where the stonepaste body and lustre-painting had originated, so they may have been "rediscovered." This revival of alkali glazes may indeed have occurred in Syria, or perhaps in Mesopotamian Syria (Raqqa) where the alkali glazing tradition may have been continuous. Why this switch took place is unknown. Reducing conditions to turn oxide into metal are relatively the same for copper as for lead, so high-copper lustres would run the risk of having black glazes if they contained lead, a problem found on a number of early Syrian lustre vessels in the Ashmolean Museum (although the glaze chemistry of these pieces is unknown). As this period seems to see a switch from greenish silver-rich lustres to brownish copper-rich lustres, the need to deal with this problem may have encouraged the reliance on alkali glazes (further analytical work on this issue is currently underway). Group One appears to have had a short-lived period of output, and there may have been a gap in production subsequently in Syria. However, the revival of production in Syria by about 1125 does show some continuity of forms and techniques. This is represented by Lustre-painted Group Two. The link between "Tell Minis" and these early twelfth century Lustre-wares, traditionally attributed to Raqqa, seems clear. However, the Group Two wares appear to have had quite a short period of production, and the link with the next group of Syrian Lustre-painted pottery, Group Seven, is more tenuous. No links have been found at all between the forms or motifs of these two Lustre-painted groups. This may be explained by the time interval of fifty years between these two groups, which is suggested by the chronology of this study. During this intervening period, Syrian production appears to be dominated by Underglaze-painted wares. Underglaze-painted wares would appear to have commenced production by 1125, with Group Three. This group is predominantly of Damascus(?) Petrofabric, and includes polychrome Underglaze-painted wares, in chromium- black, copper-turquoise, cobalt-blue, and a thick iron-oxide bole-red pigment. From the beginning this represents true underglaze-painting technology, with the decoration applied as pigment oxides with no mounting medium. This contrasts with the early Iranian wares, in which the pigment was applied in a stonepaste-slip (see Chapter 6). There have been suggestions that Syrian polychrome Underglaze-painted wares were influenced by Iranian Mina'i-style overglaze-painted wares. The technology of the Syrian wares is certainly in no way derived from Iranian practice; indeed, the opposite is probably true. The designs of the earliest wares must also be considered to be unrelated to the Mina'i styles, and also probably to predate them, perhaps by as much as fifty years (Mina'i wares may be dated to after 1175). The later figural styles of Group Six have motifs that can easily be related to those found on Mina'i wares, indicating that they are roughly contemporary. However, even here it could be argued that the influence was from Syria to Iran, as the Syrian motifs are often closer to pre-existing traditions, and show more variability than the formularized Mina'i motifs. There appears to be a number of production centres in this period, producing superficially similar wares. This situation would continue into the following century. In the last half of the twelfth century there no large groups of Lustre-painted wares appear to have been made in Syria. Underglaze-painted wares would be cheaper to produce (no tricky "second firing"), and the colourful polychrome Underglaze-painted wares may have eclipsed Lustre-painted wares in the fashion market at the time. A similar situation may have occurred in Iran where the Mina'i technique, which was probably more expensive than lustre-painting, seemed to have seized the high-end of the market for a while. In Syria, some Lustre-painted wares may have been produced during this period. Pieces with dark blue or turquoise glazes are commonly attributed to Damascus of the Mamluk period (i.e., after 1250). Although analysis of a number of wares seems to confirm the provenance attribution, the actual motifs and forms of some of these pieces suggests an earlier date). Such an early date would also correlate with the dark blue glazes of some early twelfth century Iranian Lustre-painted wares. It might therefore be possible to detect some continuity among these wares, which presumably continued into the Mamluk period, as wares with typical Mamluk motifs are also known. The next major period of Lustre-ware production in Syria would seem to date to after 1200. During this period Underglaze-painted wares declined in quality, with apparently no polychrome wares at all. At the end of our period of interest Syrian wares seem to have gone through another minor period of reduced output. However, in the fourteenth century production would again be among the most significant in the Islamic world, this time based entirely in Damascus. Return to Table of Contents