In the 1980s 1 defined an assemblage of metal objects (the Initial Metal Group) whose distribution stretches all the way from northern Colombia to Costa Rica (Cooke and Bray 1985; Bray 1992, see Fig. 1). These artifacts seem to be the oldest metal specimens in Caribbean Colombia and Central America, and, on present evidence, can be dated mainly to the early centuries after Christ. Most are items of personal jewelleri, made of gold or of gold-copper alloy with surface-enrichment, though copper may occasionally be used alone. Among the Initial forms are double-spiral ornaments made of hammered sheet metal, and also a variety of cast items, including spread wing bird pendants (both single and double) and sets of conjoined animals with raised taus (Fig. 1). The Early Tairona double-spirals seem closely related to the Initial ones but may be a little later, though they clearly fall within the first millennium A. D. (Falchetti 1987: 5).

Fig. 2. Distribution of  selected categories of International metalwork

By A. D. 600 the Initial Metal Group had evolved, in the same general area, into the International Group (see Fig. 2 for sorne of the characteristic forms and their distributions). This Group, in turn, disappeared from the archaeological record by about 900-1000. It must be emphasized that the Initial and International Groups constitute a single une of development. The division between them, typological and chronological, is an arbitrary one, and the two groups may eventually have to be combined.

The key area for this process of development is Caribbean Colombia, and in particular the eastern shore of the Gulf of Urabá (Fig. 4) where Uribe (1988) has described a mass of jewellery that is early in date and eclectic in style. From the Quimbaya region of central Colombia came the idea of making lime flasks, as well as a liking for human figures with 'dreamy' faces, multistrand necklaces and forward-pointing crowns. To these were added local Caribbean forms: double spirals, frogs, simple spread-wing birds, single and multiple animals with recurved tails, ornaments made of hammered wire, and embossed plaques made from sheet metal. This blend of influences produced a distiNctive Urabá subgroup within the Initial-International tradition as a whole.

Urabá lies at the gateway to the Isthmus, and items in this specific Urabá substyle were carried to Panarna and Costa Rica, along with more generalized Caribbean forms such as Darien pendants and human figures with elaborate headdresses. Ceramic associations and radiocarbon dates indicate that metalworking spread from Colombia to the Isthmus during the 4th or 5th century A.D. (Cooke and Bray 1985; Cooke et al. 1994), and by about A. D. 500 gold had begun to replace jade as a prestige material in some areas of Costa Rica. These dates seem to apply throughout the region, from Cerro Juan Díaz in central Panama (with Initial double spirals and fragments of an eagle pendant; R. Cooke, personal cornmunication), to the looted cemetery at Guácimo in Costa Rica, said to have yielded a double-spiral ornament, creatures with recurved taus, and two Urabá items (a frog and a Quimbayoid figure), associated with pottery dated A.D. 400-600 (Stone and Balser 1965; Snarskis 1985: 28). What is not yet clear from the few available dates is whether this south-to-north diffusion was a rapid process or a more gradual one. By juggling the error margins of the C14 determinations one can argue either way, for a hundred years or so, or for a substantially longer period.

The Initial and early International material represents a rather heterogene ous 'horizon style', with nothing particularly 'Chibchan' un either íts distribution or its iconography. The next episode of the story is one of fragmentation. In the Isthmus, during the centuries between A.D. 600 and 900, the International Style broke up into the series of local styles known as Conte (Cocle), Veraguas, Chiriquí, Diquís, etc. (Fig. 4). In these areas, too, the old technology of enclosed casting over a core was largely replaced by open-back casting. A similar process of breakup took place in northern Colombia, where the International metalwork gaye way to the mature Sinú and Tairona styles (Fig. 4). It is at this stage of evolution that the metallurgical data become relevant to the Gran Chibcha question.

Once established in the Isthmus, the new metallurgical technology was soon employed to satisfy local tastes, and we begin to see significant changes. The simple, and ideologically neutral, iconography of the International Metal Group was replaced by the culture-specific imagery of each local region, elements of which were already present in other media before the arrival of metals. Within each of these regional cultures the same images tend to occur in all materials (metal, pottery stone, shell, etc.). I am calling these sets of regional images Iconographic Clusters, and each of them has its own internal history and stylistic development. I am also assuming that the im agery is not merely decorative, but incorporates statements on matters social and religious, in the broadest sense of those words (see, for example, Linares de Sapir 1976, 1977). To put it another way, each regional Icono graphic Cluster is a visual manifestation of the local world-view or behief system.