Abstract Neoarchean rocks of the Tropicana Zone, including granites with subduction-zone affinities, formed in a terrane adjacent to, or on the margin of, the Yilgarn Craton at the commencement of a long-lived, amphibolite to granulite facies event – the 2722–2554 Ma Atlantis Event. Early stages of this event overlap with extensive komatiite emplacement within the Eastern Goldfields Superterrane (Yilgarn Craton), suggestive of a plume-related rift environment, which was followed by 2660–2630 Ma greenschist facies, orogenic gold mineralization. This indicates differences in the tectonic evolution of the Tropicana Zone compared with within the craton, although isotopic data show similarities in crustal sources. At c. 2520 Ma, the Tropicana Zone was retrogressed to greenschist facies as it was thrust onto the Yamarna Terrane (Yilgarn Craton), forming a northwesterly directed fold-and-thrust belt above the flat-lying Plumridge Detachment. This fold-and-thrust belt is host to the c. 2520 Ma, Tropicana gold deposit. The Plumridge Detachment may extend north to the Yamarna greenstone belt, linking to the Yamarna Shear Zone – the boundary between the Burtville and Yamarna Terranes. The fertility of the Tropicana Zone is related to its Neoarchean geodynamic setting within a continental arc environment, implying that deformed margins of Archean cratons may be prospective for Neoarchean Au deposits.

In geochemical diagrams, granitoids define ‘trends’ that reflect increasing differentiation or melting degree. The position of an individual sample in such a trend, whilst linked to the temperature of equilibration, is difficult to interpret. On the other hand, the positions of the trends within the geochemical space (and not the position of a sample within a trend) carry important genetic information, as they reflect the nature of the source (degree of enrichment) and the depth of melting. This paper discusses the interpretation of geochemical trends, to extract information relating to the sources of granitoid magmas and the depth of melting. Applying this approach to mid-Archaean granitoids from both the Barberton granite–greenstone terrane (South Africa) and the Pilbara Craton (Australia) reveals two features. The first is the diversity of the group generally referred to as ‘TTGs’ (tonalites, trondhjemites and granodiorites). These appear to be composed of at least three distinct sub-series, one resulting from deep melting of relatively depleted sources, the second from shallower melting of depleted sources, and the third from shallow melting of enriched sources. The second feature is the contrast between the (spatial as well as temporal) distributions and associations of the granites in both cratons.