ASPECTS OF THE PRODUCTION OF COBALT-BLUE GLASS IN EGYPT* TH. REHREN Institute of Archaeology, University College London, London WC1H 0PY, UK Cobalt-blue glass of the Near and Middle Eastern Late Bronze Age has long been recognized as compositionally distinct from other contemporary glasses (Sayre 1967; Lilyquist et al. 1993). It has been suggested recently by Shortland and Tite (2000) that this chemical distinction re¯ects the use of Egyptian raw materials for making these glasses, different from those used to make glass in Mesopotamia, or its manufacture by Mesopotamian workmen, possibly in Egypt. This assumed that cobalt-bearing alum from the Western Oases and mineral natron from the Wadi Natrun were used for the cobalt-blue glass, while the other, probably Mesopotamian, glasses were made using plant ash as the main alkali source. This note discusses some technical aspects of the possible ways in which the cobalt could have been added to the glass, and how this relates to the likely raw glass used in its making. Combining earlier suggestions by Noll (1981) and Brill in Lilyquist et al. (1993), an alternative explanation of the chemical characteristics is suggested, maintaining that all the glasses under discussion were made using plant ash. Differences in alkali concentra- tions probably re¯ect different soil and plant chemistries, and the colorant was probably added to the glass after being precipitated from the alum as a complex cobalt aluminium hydroxide. KEYWORDS: GLASS, LATE BRONZE AGE, EGYPT, COBALT, PRODUCTION INTRODUCTION It has long been noted (see, e.g., Sayre 1967) that the base glass chemistry of Near and Middle Eastern Late Bronze Age glasses is astonishingly homogeneous across time, space and colour (see Lilyquist et al. (1993) for the concept of the basic glass-forming components, and Rehren (2000a) for a more detailed discussion of this homogeneity). Within this general similarity of the base glass composition, cobalt-blue glasses form a unique subset characterized by signi®cantly higher concentrations in Al 2 O 3 and lower concentrations in K 2 O when compared to other glasses (Lilyquist et al. 1993). The former can be explained as a side-effect of the putative colorant for this particular glass colour, a cobalt-bearing alum found at the Dakhla Oasis (Kaczmarczyk 1986). Analyses of this salt, taken from a distinct layer or vein, were reported in Kaczmarczyk (1986); see also Table 2 in Shortland and Tite (2000, 146). It was found to contain the same suite of trace elements known to accompany the cobalt in the cobalt-blue glasses, in the same relative amounts: manganese, iron, nickel and zinc, each in oxide concentrations of between about 0.2 and 1%. Since then, the relationship between the cobalt-bearing alum and the cobalt-blue glasses has been widely accepted. The general composition of this alum also explained the higher concentration of aluminium in these glasses. How this addition of the colorant, however, could have reduced the potash content in the glass was rarely discussed in the literature (but see Archaeometry 43, 4 (2001) 483±489. Printed in Great Britain * Received 2 October 2000; accepted 23 February 2001. q University of Oxford, 2001