RESUMO
The success of early Chinese blue-and-white porcelains relied heavily on imported cobalt pigment from the West. In contrast to art-historical concept, which contains both typological evidence and literature records, it is assumed that imported Sumali blue was completely superseded by domestic Chinese asbolane ore based on the analytical results of the Fe/Mn ratio in imperial productions from the Xuande reign (1426 to 1435 CE) onward. Using a focused ion beam transmission electron microscopy technique to reassess this hotly debated question, we have identified two classes of residual submicron pigment particles in the blue glaze with diagnostic differences in morphology, chemical composition, and distribution behavior. Compared with the microstructural features of the blue-and-white porcelains of the Yuan and Qing dynasties, we show that a mixture of imported and domestic cobalt pigments was used for aesthetic reasons, indicating that the overseas supply chain of imported pigment remained consistent and adequate even though the authorities had terminated official trade and tributary activities after the death of Admiral Zheng He. This discovery further suggests that the globalized trading network and cross-regional industrial chain had been extensively established in the 15th century. Moreover, we provide analytical evidence against the fundamental assumption of the current Fe/Mn provenancing criteria, implying that the failures of previous chemical analyses can be attributed to elemental differentiation between the silicate glaze and the arsenic pigment. We propose an innovative method for directly assessing original mineralogic information from submicron residual pigment particles that provides a more reliable way to trace cobalt circulation and holds great promise for provenance studies.
RESUMO
Seventeen glass vessels and twenty glass beads recovered from the excavations at the ancient city of Malindi and the archaeological site of Mambrui in Kenya, east Africa were analysed using electron probe microanalysis (EPMA) and laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS). The results show that all of the glass samples are soda-lime-silica glass. They belong to the high alumina -plant ash glass type, characterised by high alumina and relatively low calcium contents, widely distributed in eastern (10th- 16th centuries AD) and southern Africa (13th - 15th centuries AD), Central Asia (9th- 14th centuries AD) and southeast Asia (12th- 13th centuries AD), made with plant ashes and sands. This is an understudied glass type for which previous research has indicated there were three types. When compared with published research on such glasses using Zr, Ti, Ba, Cr, La, Li, Cs, Na2O, MgO and CaO we have identified at least four different compositional groups of v-Na-Al glass: Types A, B, C and D. By comparing the results with contemporary v-Na-Al glass vessels and beads from Central Asia, Africa, and southeast Asia we show that most of the Malindi and Mambrui glass share similar characteristics to the compositions of Mapungubwe Oblate and some of the Madagascar glass beads from southern Africa. They belong to Type A v-Na-Al glass which is characterised by an elevated level of Ti and Ba and a relatively high ratios of Cr/La, relatively low Zr concentrations and low ratios of Zr/Ti. Differences in Zr, Li, MgO and Na2O concentrations in Type A glass indicates that there are subgroups which might derive from different glass workshop(s) specialising in Type A v-Na-Al glass production. Comparison with the chemical compositions of glass from Ghazni, Afghanistan and Termez, Uzbekistan, and by using lead isotope analysis, we suggest v-Na-Al glass was manufactured in Central Asia and possibly worked into vessels and beads there.