Bulk analysis shows the pendant is a gold/copper alloy, with only a small amount of silver and falls into the range of Diquis objects analysed by Rovira (1994). From the gold/copper phase diagram (Fig. 3) it is seen that this alloy has a melting point of about 9l5°C, almost the lowest melting temperature possible. This in itself is an interesting fact and might suggest choice of composition by the Diquis metalsmith for ease of melting, although the freezing range of this alloy is only about 5°C from the beginning to the end of solidification. Nevertheless the whole casting is sound with no evidence of defects.

Figure 2: SEM false colour micrograph of a sample from the musicians pendant showing the dendritic cast structure with the gold-rich phase, yellow, and the copper-rich matrix, blue. [Photo: N.D. Meeks, British Museum]

The dendritic cast structure is shown at low magnification (x40) in Fig. 2 which is a false colour SEM image of the sample cut through the false filigree wire where it joins the pendant frame. The microstructure shows segregation into two phases which are of very different composition (Table 1), the yellow, gold-rich interdendritic phase and the blue copper-rich primary dendritic metal. This is unlike the cored structure more commonly seen under normal casting conditions in a gold alloy with 26% copper (eg. fish hook Fig.12). At higher magnification (xl000) the gold-rich phase is seen to be either finely cracked or divided by grain boundaries while the copper-rich primary dendritic metal is not (Fig4b). Figure 4b shows corrosion progressing into the gold-rich phase in some of the outer regions of the sample, while the copper-rich matrix is not corroded. The apparent reversal of normal electrochemical corrosion has been noted in tin-bronze by Tylecote (1985) in which the cathodic and anodic components are reversed due to certain corroding conditions during burial. Corrosion of the gold-rich phase is also found in the anthropomorphic figure described below and this may be a common phenomenon in objects made of these binary phase, copper-rich, depletion gilded alloys. EDX analysis shows the presence of oxygen in the corroded regions of the gold-rich phase, where gold was apparently being lost during burial leaving oxidised copper behind, while the copper-rich primary dendritic phase was apparently cathodic and passive. To fully research this corrosion mechanism is beyond the scope of this paper but it is probable that it is associated with ordering of the gold-rich phase (see below).