Oceanic volcanics, erupted through thin, young lithosphere, represent a window on the asthenosphere and deep mantle. In contrast, continental basalts and mantle xenoliths, emplaced through thick, old lithosphere, may tell us about the nature of the deep crust and the lithospheric mantle, as well as the evolution of magmas during their ascent to the surface. Isotopic data represent a powerful tool for such studies, firstly because of their ability to date geological events, and secondly because of their usefulness as tracers of complex mixing processes.

Unfortunately, continental igneous rocks are difficult to interpret. This is because they can derive an enriched elemental and isotopic signature from three possible sources: mantle plumes, sub-continental lithosphere and the crust. Resolving these components from one another in continental volcanics and plutons has been a major subject of discussion in geochemistry for several decades. Much progress has been made, but the large number of variables tends to make each case a unique example; or as Read (1948) put it, there are ‘granites and granites’. This makes a generalized approach to continental magmas difficult, and forces us to adopt a case study approach as an attempt to illustrate underlying principles.

Mantle xenoliths provide a more direct means of sampling the sub-continental lithosphere. Their texture provides evidence of a solid source, while the peridotite (= lherzolite) petrology of the commonest types is readily distinguished from crustal xenoliths (which will not be dealt with here). Therefore, our approach in this chapter will be firstly to study the lithospheric mantle by means of xenoliths, secondly to examine crustal contamination processes and lastly to look at some classic case studies in the genesis and evolution of continental igneous rocks.

Mantle Xenoliths

The sub-continental lithosphere is distinguished from the underlying asthenosphere by its non-convecting, rigid state. Hence it was termed the ‘tectosphere’ by Jordan (1975, 1978). Jordan argued from seismic and heat-flow evidence that this tectosphere was 200–300 km thick under shield areas. Evidence for 3.5 Ga garnet inclusions in diamonds (Richardson et al., 1984) suggests a similar thickness of continental lithosphere in the Archean.