The twelfth and thirteenth centuries constitute a high point in the history of Islamic ceramics, with finely pot ted and decorated fritwarrs produced in both Syria and Iran. The widespread production of fine-lxdicd fril wares (also known as artificial paste or stonepaste-bodicd ceramics) and techniques either new (nnderglac paint ing. ajourc. and mina^i) or refined to new heights (luster and molded wares) base focused deserved attention on these two areas. Here we will examine the production of glazed ceramics in Svria and thejazira in the twelfth and thirteenth centuries by examining pan of a sample of glazed ceramics found during excavations at the south eastern Turkish site of Gritille. Until its recent flooding by the waters of the Ataturk Dain, Gritille was a small mound on the right bank of the Luphratcs River soon af ter it came down from the Anatolian highlands.
1 It was located some five miles tq»streain from Samsat, the prin cipal site in the region, which served to guard the major river crossing on the route from Syria to the eastern Ana tolian plateau, and dominated a small vallev surround ing it. The Sarnsat region was peripheral to both Anato lia and Syria, although it had participated more in the history of north Syria and the Ja/ira. Gritille itself was pe ripheral to Samsiit. in that it guarded a minor river cross ing at the uppet end of the valley that wasSamsafs hin terland. Gntillc was first foruficd in the medieval period in the mid eleventh century and soon abandoned, only to be rcfortificd about a century later near the end of the Cru sader county of Edcssa. Thereafter, from the mid twelfth century through the early to mid thirteenth cen tury. Gritille scivcd as a rural agrRu1lui.il settlement with largely domestic architecture. To examine the production of glazed ceramics, a sub set of 168 sherds from medieval levels at Gritille were sampled anti chemically analyzed using instrumental neutron activation analysis (INAA).
This study corre- lates visual categories of Islamic glacd fritwarcs (derived mainlv from decorative technique) with categories of chemical composition to answer questions relating to glazed ceramic production in Syria and the Jaira in the twelfth and thirteenth centuries. The project is still in progress, but it has already yielded an unexpec ted result that we will prevent in this paper. Of the 168 glazed ceramics sampled. 37 were found not to Ikfiitwares at all. but instead had calcareous clay bodies as determined by chemical analysis. Figure I contains drawings of indi cator sherds (rims and bases) in the calcareous clay ceramic sample. The sherds consututing this sample wcre of small open form vessels. The one exception to this is a vessel (ID # GT14S) whose shac and glaze indi cated that it belonged to a closed form vessel The indi cator sherd drawings show these open form vessels to be small bowls with rim diameters of approximately ten cen timeters. The sample contains two distinct rim shapes, a flared everted rim and a squared head rim. Ihose sherds large enough to show more body shape point to carina tion a few centimeters below the rim. There are also tw4 liases, one of which belongs to an oil lamp. Most of the sherds arc of turquoise or blue-green glaze, with three manganese purple glazed sherds and two with turquoise glaze on the exterior and manganese on the interior.
Chemical analysis.
The sherds were sampled for chemical analysis by drilling with tungsten carbide drill bits. Approximately 200 milligrams were extracted, dried, and 100 milligram siibsamplcs taken for analysis. Analy sis was conducted by INAA at the Smithsonian Institu tion Conservation Analytical laboratory’s INAA facility using the National Institute of Standards and Technolo gy’s 20 megawatt research reactor. The analytical proto col was similar to that described in Blackman 1984 For the calcareous clay bodied ceramic 28 elements were quantifiable. The chemical data from the 37 calcareous clay sherds were first processed with the hierarchical aggregative clustering program AGCI,US using a nearest neighbor clustering algorithm on a squatrd mean Euclidean dis tance matrix based on the elements listed in figure. Clustering analysis divided the sample into five groups with one outlier displayed in figure 2. Chemical groups 1 through 4 all contained calcium in excess of 10 percent Ml body colors ranged from pale yellow through white. Chemical group 5, containing only two samples, aver aged about 7.5 percent calcium and the bodies displayed a brown color. Group 5. distinctive in both chemistry and body color, and probably better classified as an earthen ware. was excluded from further statistical analysis. Principal components analysis was carried out on the remaining four groups using the elements Na. Cr. Fe. Rb. and la. Figure 3 is a plot of the second and third principal components for the four groups, with the ellipses around the samples in each group representing the 93 percent confidence intervals.
This figure readily shows the four groups distinguishable f rom one another in terms of their chemical composition at better than the 93 percent confidence level. The implications of the sta tistical analysis of the chemical data are that the calcare ous glazed ceramics were produced using four different clay sources. Although it is not possihle conclusively to prove multiple production sites without an extensive geological clay sample collet lion fiom the region, it serins reasonable to presume that four chemically dis tinctive clay sources would not have been accessible to or utilized bv a single ceramic workshop. Therefore, as many as four production locations may be indicated by the chemical data.
Divussian.
At Gritille. calcareous clay glazed vessels were recovered in levels dated bv coin finds from the sixth decade of the twelfth century through the third decade of the thirteenth century.’ The squared bead rim is found in three of the four major compositional group ings, and examples from the earlirst to latest phases are also found in these groups. This suggests two tilings: first, a local production, exemplified in Group 1, that lasted well over seventy vears but was confined to a lim ited shape and glaze range, and second that this local production imitated imported examples. Our candidate for lot al production is the town of Samsat. the largest set tlement in the region and the center of local administra tion. ‘Ihe division of the calcareous clay glazed ceramics from Gritille into groups constituting local production and imports is based on frequency of occurrence. Group I, our candidate for local production, contains 16 of the 37 sherds sampled. The chemical composition of these sherds is very homogeneous. In addition. 12 of the Ifi sherds have a glaze that is distinct from other hues of blue. Instead of the bright, evenly litcd turquoise found on other samples (and indeed on some of this self-same group), these 12 sherds have a glaze of blue-green, opaque but not evenly so. with the color varying from spot to SK on the vessel.
This suggests a firing temper ature lower than that of the turquoise glaetl wares. Group 3. a small group consisting almost entirely of the same everted rim type (possibly even from the same vessel — these sherds were found in the same excavation area and belong to the same phase) points to a produc tion entirely different in shape, quality of glaing or fir ing (these sherds arc all turquoise glazed), and chemical composition. Group 2 also has a different chemical com position, although the rim shapes and glazes are similar to Group 1. Two outliers, GT 27 and GT 34B. both of darker fabric, point to an overlap between the produc tion of regular earthenwares and these calcareous ear thenwares. GT 27 has a blue-green glaze similar to many of the samples, and GT 34B a similar squared bead rim and carinalcd body. An extreme outlier. GT 143. is the only closed form vessel sherd of the entire sample. Frit wares found at Gritille also have the same beaded nm. carinated body, and turquoise or blue-green glaze as the calcareous clay vessels, demonstrating an merlap with fritware production, too. (lomftaranda. An analogous ware was found bv excavators of the slightly later medieval site of Asvan Kale to the north of Gritille.
There, concurrent with the on-site pro duction of sgraffiato and other earthenwares excavators noticed that “one other very distinctive type of glazed pot occurs frequently enough to suggest a local origin.”11 This type was a howl made with whitish or gray clay and covered with dark blue or manganese glaze. Both size and shape recall the Gritille calc areous pottery, although there is just enough difference (especially in the pres ence of vertical ridges around the pots) to preclude a shared origin. Similarly, to the south of Gritille, excavations at Anti och produced types of glazed ceramics common enough that one can suppose them to he of local manufacture. Two types of this pottery were characterized by opaque turquoise glaze on a yellow body and greenish-blue glaze on a thick yellow body, similar in glaze, shape, and body color to some of the Gritille sample presented above. The differing compositional groups of the Gritille cal careous clay ceramics and the examples cited a bine from sites to the north and south of Gritille together lead one to suppose that the use of this kind of clay body was not just confined to an isolated area. The presence of calcareous clay in fine vessels is testified from earlier eras and other parts of the Islamic world, namely Abba sid Iraq and Fatimid Egypt,” so this is not a new ceramic technique.
Quite the contrary, it represents the survival of an earlier technique in an era when a new- technology had seemingly carried the field. With the rise and spread of fritwarc production in the eleventh or twelfth centu ries,4 the making of fine vessels with calcareous clay bod ies fell out of favor, as the whiteness of the fabric and fineness of potting available from an artificial paste body made it more attractive. The knowledge and exploita tion of beds of white clay would not have ceased, how ever. despite the overwhelming popularity of fntwares in the medieval and late medieval Middle Fast. White calca reous day continues to be used as a small pan of the frit ware process.” Interestingly, however, the dominance of fritwarc in inland Syria and the Jazira in the twelfth cen tury is threatened at the turn of the century, not by a resurgence of interest in calcareous clay vessels, but by another kind of pottery production — glazed earthen ware and sgrafTiato.
In our opinion, «hen. the observable wide variety in fineness of polling, glazing, firing, and decoration of latc-twcltth and earlythirtccnthentury Syrian ceramics must be due to many production cen ters, in as small and peripheral areas as the Samsat valley. The production of calcareous clay vessels is part of the wide variety of ceramic production that goes with this decentralization. Until activation analysis is performed on them, these vessels, with calcium contents of between 10 and 17 percent, are visually indistinguishable from the lower end of the spectrum of fntware production occur ring at the same lime.
